Beal Roy E.eds. Engine Coolant Testing Volum PDF

S T P 1192
Engine Coolant Testing: Third

Volume 3
Roy E. Beal, editor
ASTM Publication Code Number (PCN)

04-011920-15
ASTM
1916 Race Street
Philadelphia, PA 10103
ASTM Publication Code Number (PCN): 04-011920-15
ISBN: 0-8031-1851-1
ISSN: 1050-7523
Copyright 9 1993 AMERICAN SOCIETY FOR TESTING AND MATERIALS, Philadelphia, PA. All
rights reserved. This material may not be reproduced or copied, in whole or in part, in any printed,
mechanical, electronic, film, or other distribution and storage media, without the written consent of the
publisher.
Photocopy Rights
Authorization to photocopy items for internal or personal use, or the internal or personal use of
specific clients, is granted by the AMERICAN SOCIETY FOR TESTING AND MATERIALS for users
registered with the Copyright Clearance Center (CCC) Transactional Reporting Service, provided that
the base fee of $2.50 per copy, plus $0.50 per page is paid directly to CCC, 27 Congress St., Salem,
MA 01970; (508) 744-3350. For those organizations that have been granted a photocopy license by
CCC, a separate system of payment has been arranged. The fee code for users of the Transactional
Reporting Service is 0-8031-1851-1/93 $2.50 + .50.
Peer Review Policy
Each paper published in this volume was evaluated by three peer reviewers. The authors addressed
all of the reviewers' comments to the satisfaction of both the technical editor(s) and the ASTM
Committee on Publications.
The quality of the papers in this publication reflects not only the obvious efforts of the authors and the
technical editor(s), but also the work of these peer reviewers. The ASTM Committee on Publications
acknowledges with appreciation their dedication and contribution to time and effort on behalf of ASTM.
Printedin Ann Arbor,MI

May 1993
Contents
Overview
A Review of Automotive Engine Coolant Technology--H. J. HANNIGAN
Corrosion Protection of Aluminum Heat-Transfer Surfaces in Engine Coolants

Using Monoacid/Diacid Inhibitor TechnoIogy--JEAN-PmRRE MAES AND
P A U L V A N DE VEN 11
Discussion 22
Fleet Test Correlations of Original Equipment Coolant Pump Failures and Engine
Coolant Formulations--JEFFREY M. BURNS 25
Discussion 40
An Investigation of Carboxylic Acids as Corrosion Inhibitors in Engine Coolant--

W I L L I A M C. M E R C E R 44
Discussion 58
Fleet Test Evaluation of Engine Coolants Using Sebacic Acid Inhibitor

Technology--NORMAN C. ADAMOWICZ AND DANIEL F. FALLA 63
Discussion 82
Corrosion Testing for Aluminum Alloy Selection in Automotive Radiators--

R O Y E. BEAL A N D R E F K I E L - B O U R I N I 83
An Overview of Cavitation Corrosion of Diesel Cylinder Liners--

R I C H A R D D. H E R C A M P 107
Discussion 126
Phosphate-Molybdate Supplement Coolant Additives for Heavy Duty Diesel

Engines--R. D. HUDGENS 128
Discussion 148
Toxicity and Disposal of Engine Coolants--R. DOUGLAS HUDGENS AND

R. B. B U S T A M A N T E 149
Discussion 164
Test Strips for Rapid On-Site Analysis of Engine Coolants--PAUL g. HEMMES,

T H O M A S H. KREISER, S A R A H VALLE, A N D R I C H A R D D. H E R C A M P 165
Discussion 179
Application of Inductively Coupled Plasma (ICP) Emission Spectroscopy and
Laser Ablation-ICP for Problem Solving in Coolant Systems--
W A L T E R Z A M E C H E K AND DALE A. McKENZIE 180
Discussion 189
The Chemistry of Oxalic Acid Cleaning of Engine Cooling Systems--

PETER M. WOYCIESJES 190
Investigation of Deposits on Water Pump Seal F a c e S - - R A N D A L L J. STAFFORD 205

Discussion 213
The Relationship Between Sealing Performance of Mechanical Seals and

Composition of Coolants for Automotive Engines--KENJI KmYU,
OSAMU HIRATA, AKIRA YOSHINO, KEN OKADA, AND
HIROSHI HIRABAYASHI 215
Characterization of Used Engine Coolant by Statistical Analysis--

STEPHEN M, WOODWARD AND ALEKSEI V. GERSHUN 234
Discussion 246
Coolant Maintenance and Extension of Coolant Life for Light Duty Vehicles--
R I C H A R D D. H E R C A M P AND ROBERT A. REMIASZ 248
Discussion 257
A Multi-Stage Process for Used Antifreeze/Coolant Purification--

ROBERT C. RICHARDSON 258
Discussion 274
An Evaluation of Engine Coolant Recycling Processes: Part I - -

WAYNE H. BRADLEY 276
Discussion 287
Heavy Duty Diesel Engine Coolants: Technology U p d a t e - - F . A. KELLEY 289
Index 297
STP1192-EB/May 1993
Overview
Engine coolant usage continues to increase on a worldwide basis as the overall vehicle pop-
ulation becomes larger. Many off-highway vehicles and stationary equipment facilities also use
engine coolant. Water preservation and environmental concerns are reflected in a gradually
expanded use of coolant for industrial cooling applications. Vehicle cooling predominates and
is the major concern for the symposium. Average vehicle size and coolant capacity has recently
reduced in the United States. More efficient engine designs tend to use less coolant volume for
equivalent heat rejection purposes. Modern automobiles are made with newer and lighter
weight materials. The importance of aluminum alloy protection by engine coolant has become
evident, together with an increased use of composite plastics. Meanwhile, the average age of
vehicles on the highway has increased, and these older vehicles require engine coolant replace-
ment at regular intervals. The engine coolant specialist has therefore many technical chal-
lenges and the technology is developing sufficiently that a meeting to present advances and
discuss current problems was needed.
The first symposium was held in Atlanta, GA, in 1979. It was well supported and resulted
in ASTM Engine Coolant Testing: State oftheArt, (STP 705), which still provides a good ref-
erence. Success led to a second conference in 1984, held in Philadelphia, PA, at which the rapid
changes in material usage and testing requirements were expounded upon by many of the
authors. This symposium resulted in ASTM Engine Coolant Testing." Second Symposium,
(STP 887), and probably the most important development was the basis of a new standard for
evaluating hot surface protection for aluminum engine alloys that has now become an inter-
national standard for the coolant industry. Propylene glycol was introduced as an alternate
base fluid for coolants. Electrochemical test methods were evaluated and discussions of spe-
cific needs for heavy duty engines highlighted.
The third symposium was held in Scottsdale, AZ, Engine Coolant Testing." Third Sympo-
sium, 6-8 Nov. 199 l, and was well attended with presentations from European, Japanese, and
United States authors.
Papers presented at the conference covered advances in the development, testing, and appli-
cation of engine cooling fluids for automobiles and heavy duty engines that have occurred
since the last meeting.
A keynote opening address by Hannigan set a good tone to the conference by presenting a
brief history of ethylene glycol engine coolant. Ethylene glycol was first suggested for use as an
engine coolant in military aircraft in England in 1916. Other aircraft applications followed,
with the Curtiss Hawk PIA in 1926 being of particular note. Use of ethylene glycol in auto-
mobiles began in 1927. Wide adoption occurred in the period 1949 through 1955 as a factory
fill in place of methanol. Developments have continued, and Hannigan presents the highlights
bringing us up to the present time.
Four authors presented papers on new families of engine coolant that operate in a medium
pH range. Maes and Van Den Ven's work in Europe on the use of low depletion monoacid
diacid inhibitor technology reveals good high-temperature corrosion protection of aluminum
when acids are properly balanced. An evaluation program included ASTM Test Method for
Corrosion of Cast Aluminum Alloys in Engine Coolants Under Heat-Rejecting Conditions (D
Copyright9 1993by ASTMInternational www.astm.org

2 ENGINECOOLANT TESTING: THIRD VOLUME
4340) hot surface tests, a dynamic heat transfer test, and a coolant aging test. These were used
with top quality commercial engine coolants and the monoacid dibasic acid technology, for
direct comparisons. Static heat transfer testing gave good results with all technologies.
Dynamic heat transfer testing was more discriminating and favored the new monoacid and
diacid combinations for aluminum corrosion protection under hot surface conditions with
apparently better heat transfer at the metal/coolant interface.
Burns carried out an extensive fleet test with a carboxylic acid long life coolant formulation
with very good results. Two hundred and three Ford Crown Victoria Taxi cabs were used. The
chief objective of the program was to evaluate coolant pump failure with respect to the new
carboxylic long life coolant, when compared to more traditional formulations. Coolant instal-
lation was color coded and pump failures from each group identified. Four conditions and a
factory fill were involved. The new carboxylic formulation resulted in the lowest pump failure
rate, although reasons why could only be speculated upon. A new coolant pump bench test is
recommended for comparative study of coolant formulations.
Mercer examined experimental carboxylic acid inhibitor formulations, to test their effi-
ciency in both laboratory and fleet tests. Problems were encountered with lead based solder
and aluminum alloy corrosion, although other metals were adequately protected. Aluminum
protection required high levels of acids present. Compatibility of the carboxylic acid formu-
lations and phosphate buffered coolant was poor, and mixtures resulted in reduced protection.
No inhibitor depletion was observed, but this did not prevent corrosion of the aluminum and
high lead solder alloys in fleet testing.
Extended life coolant with sebacic acid was compared to current high silicate alkaline phos-
phate coolant prevalent in the United States in a three-year municipal fleet test by Adamowicz
and Falla. Results demonstrated no particular advantage with the sebacic acid formulation
over current North American coolant. They also concluded that factory fill coolant life can be
extended far beyond previous expectations. Metal coupon corrosion losses were minimal for
either coolant throughout the test. The relatively high cost of the sebacic acid coolant precludes
its use on an economic basis.
Durability of aluminum alloy automotive radiators in service depends on the alloy selected
and the expected engine coolant environment. Beal and E1-Bourini investigated accelerated
testing procedures for alloy development with appropriate coolant conditions. New alloys are
continually under development to improve radiator service life. The challenge is to find testing
methods that correlate with service experience without resorting to long-term vehicle exposure
trials. A combination of electrochemical studies and simulated service has demonstrated a via-
ble approach. Unless related to a particular engine coolant environment, serious mistakes can
be made in aluminum alloy radiator materials chosen.
Heavy duty diesel engines use significant quantities of coolant and emphasis on long oper-
ation periods continues as engine design changes, resulting in higher efficiencies. Hercamp
presented an historical overview of cavitation corrosion in diesel cylinder liners, relating var-
ious factors that are involved with liner pitting. A major problem in the 1950s, much work has
been done since to identify causes and develop solutions. The paper covers scientific back-
ground and theories, and not all workers agree on the damage mechanism. Coolant effects and
the use of supplement coolant additives (SCAs) are covered with some reference to engine
related factors. Education of maintenance personnel is important to follow prescribed proce-
dures in coolant and SCA practice. Testing can assist in avoiding trouble and is cost effective
in checking coolant condition.
Hudgens briefly covered the history of supplemental coolant additives used in heavy duty
diesel engines, and then went on to describe a new family of phosphate molybdate packages
that are designed to perform better with aluminum components and cause less problems if
OVERVIEW 3
overtreatment occurs. A test scheme involved most ASTM standards and the German FVV
Test, in addition to bench cavitation work. The phosphate molybdate formula provides better
protection in hard water, and the ability to reduce nitrite levels is beneficial to solder protec-
tion. Liner pitting is prevented at lower overall inhibitor addition.
Toxicity and disposal of engine coolants is a topical subject that was reviewed and discussed
by Hudgens and Bustamante. Properties of ethylene and propylene glycols and major addi-
tives used in engine coolants are included. Propylene glycol is not toxic and provides an envi-
ronmentally acceptable coolant base. However, inhibitors used have varying degrees of tox-
icity, and after use, when heavy metals are dissolved into the coolant, the resultant fluid is
definitely toxic, whether propylene glycol or ethylene glycol are used as the base fluid. Present
laws and regulations are referenced, and a discussion on the hazardous concept is included.
Used coolant may or may not be hazardous depending on its condition when tested against
EPA threshold values. Both ethylene and propylene glycols are biodegradable. 400 million gal
( 1514 million L) of coolant are sold every year. 10% of coolant may be lost by leakage, 25% or
more by improper disposal, and the remainder generally handled according to regulations.
Recycling is becoming a commercial feasibility and is being done in the western United States
in particular on a large scale. Total volume of recycled coolant is still low compared to coolant
sold per year. The paper gives a good overview of facts and concerns regarding handling and
disposal of engine coolants.
Test strips have been developed for rapid on-site analysis of engine coolants for some specific
attributes. Hemmes et al. described their efforts. Strips for nitrite and molybdate measure pro-
tection for cavitation erosion. One has been developed for MBT in conjunction with chloride
level identification. Test strips for pH and RA also are available. Measurement of freeze point
has been carried out for over ten years. The wider range of test strips assist in maintenance
programs and for identifying when SCAs should be used in heavy duty vehicles.
Engine coolant analysis techniques use standard equipment with particular procedures for
accurate results. Advances coincide with new analytical instrumentation. Problems in coolant
systems can be solved by application of inductively coupled plasma (ICP) emission spectros-
copy and laser ablation ICP according to Zamechek and McKenzie. Coolant analysis by ICP
is enhanced by specially developed software for interferences and data reduction. Aqueous
standards are used with 50-fold dilution of the analytical samples. Preparation methods are
described. The laser ablation system was used for spacial mapping of elements on the surfaces
of water pump seals. A uniquely adapted sample cell and target area was devised with optical
focusing and alignment.
Oxalic acid cleaning of engines removes inhibitors, rusts, and other deposits. Some concern
has been expressed on the post cleaning effect of the process when vehicles are used for coolant
testing. Woyciesjes reviewed the chemistry involved. Oxalates form a variety of complexes
with typical metals in the engine circuit. Ferrous oxalate can be 10 #m or more in thickness.
Borate conditioning removes some of the oxalate. Oxalates can affect subsequent coolant
properties by having a detrimental influence on pH, RA, and inhibitor levels. A high pH,
borate conditioning fluid minimizes the consequences, and in new vehicles the effect is small.
Vehicles exhibiting heavy corrosion should not be used for coolant testing, because cleaning
will not be totally effective.
Pump seal failures are a contemporary problem with disagreements on causes and solutions.
This topic was received with much interest by the attendants. Deposits on water pump seal
faces were examined by Stafford from heavy duty diesel on-highway engines. Coolant leakages
were traced to deposit films built up on the siliconized graphite seals. Surface analyses of the
buildup revealed elements from the coolant, coolant additives, corrosion metals from the
engines and calcium from hard water. Mileage at pump removal ranged from 28 000 to
199 000 miles (45 060 to 320 251 km). A calcium-iron-phosphate complex precipitated dur-
ing nucleate boiling episodes was determined as the cause of seal leakage because of seal face
separation caused by the deposit.
Kiryu et al. examined the effect of coolant on water pump mechanical seals in a very thor-
ough investigation. There is an urgent problem attributed to coolant formulation contami-
nation and an increase in engine operation condition severity. Leakage occurs by deposit for-
mation and growth of the film, which creates a gap at the seal face. Oxygen-rich conditions at
150"C can cause inhibitor solidification that deposits on the seal. Test work confirmed that
high-temperature seal operation causes deposits related to silicates, when they are present. In
triethanolamine coolant, copper and iron salts were the culprit, usually from breakdown acids
promoting corrosion of copper parts. A third coolant formulation resulted in precipitation of
dibenzothiazyl disulfide on seal ring surfaces leading to leakage. All the deposition problems
were solved by designing seals with lower interface operating temperature, controlling mate-
rials used, and reducing surface roughness at the seal face.
Depletion of engine coolant inhibitors, contamination, and breakdown of the glycols occurs
during the use of engine coolant in service. Vehicle makers provide recommendations on
changing coolant on a regular basis. These changes provide the waste stream that can be used
for recycling. Statistical analysis of used coolants gathered from New England through Georgia
was performed by Woodward and Gershun. A total of 2500 vehicles was reviewed in the
results. Standard laboratory techniques were used for the analyses with appropriate conditions
for accuracy of data collected. A wide range with nonnormal distribution was found for resid-
ual inhibitors. Corrosive contaminants, such as chloride and sulfate, varied widely with chlo-
ride levels similar to ASTM corrosive water and sulfates significantly higher. Degradation of
the glycol to acetates, glycolates, and formates depletes the reserve alkalinity. A prediction is
made that 20% of used engine coolant will have lead in excess of 5 ppm, and thus be regarded
as hazardous waste. Suspended solids are found regularly with over 25% of those coolants
tested having 500 ppm or more. Recycling needs careful consideration because of variations
in fluid conditions and the need for a balanced product.
Extension of coolant life in automobiles is feasible when a three-step examination is made
that determines coolant has not been used for 65 000 kin, is not oily, murky, or rusty, or is
less than 25% concentrated with a reserve alkalinity of less than 3. Under these conditions
Hercamp and Remiasz show that a supplemental coolant additive package can provide at least
a further year of life to the coolant. Standard ASTM tests were used for verification, and field
experience has been satisfactory. The additive is used in conjunction with a closed-loop cool-
ant flushing system attached to the vehicle.
Richardson described a recycling process for used coolant that involves a multistage process
with dual bed deionization. The process purifies the coolant removing contaminants and par-
ticulates. The resulting fluid has very low concentrations of all species providing a clean fluid
for reinhibition. The process used and data obtained are described. The author considers that
efficient removal of contaminates to a low total dissolved solids level is necessary for a consis-
tent finished product.
Recycling processes were discussed by Bradley. The paper reviewed several different
approaches to providing the service. Awareness of environmental issues in the disposition of
spent engine coolant prompted a study to examine the efficiency of various systems. A refer-
ence coolant was utilized that was collected from many vehicles, resulting in a mixture of sev-
eral types of inhibitor packages and degradation products. All recycling processes used the
same coolant for test purposes. Processes evaluated were filtration, filtration-flocculation-
coagulation, deionization, reverse osmosis and vacuum distillation. Some systems were com-
binations of these processes. These systems are described. Off site coolant recycling is per-
formed on a large scale typically by fractional distillation, and these companies are included
OVERVIEW 5
in the G.M. approval program. Recycled coolants must meet or exceed GM 1825 M coolant
specification. A progressive test program was undertaken. Coolants failing any test in the
sequence were rejected. Physical tests, followed by ASTM Test Method for Corrosion Test for
Engine Coolants in Glassware (D 1384) and ASTM D 4340 hot surface evaluation were per-
formed. Only those passing proceeded to pump cavitation and simulated service. Results were
not available at the conference and will be published later.
Two keynote papers were invited covering automotive and heavy duty vehicles technolo-
gies. The objective was to educate the newcomers and remind the veterans of coolant tech-
nology development over the years to the present time. Both papers were timely and a success
at the symposium. Hannigan covered automotive cooling technology in which he has been
personally involved over many years making a good presentation of history and finishing with
highlights of present challenges. A summary of current heavy duty technology in coolants was
ably addressed by Kelley with discussion on liner pitting, silicate drop-out, water pump seal
leakage, and other problems. He discussed the value of ASTM standards and new require-
ments for the future with a good overview.
The symposium was a success and reflected advances in coolant technology and present
areas of concern. A special thanks to all the authors, the symposium subcommittee, chairmen
of the individual sessions, and the ASTM staff is warmly given. Jenny Beal, Denise Steiger,
and Gloria Collins deserve specific mention for the organization of the conference and social
events. This volume will make a valuable contribution to publicly available information on
coolant technology.
Roy E. Beal
Amalgamated Technologies Inc., Suite 208,
13901 N. 73rd St., Scottsdale, AZ 85260;
symposium chairman and editor.
H. J. H a n n i g a n ~
A Review of Automotive Engine Coolant

Technology
REFERENCE: Hannigan, H. J., "A Review of Automotive Engine Coolant Technology," Engine
Coolant Testing: Third Volume, ASTMSTP 1192, Roy E. Beal, Ed., American Society for Test-
ing and Materials, Philadelphia, PA, 1993, pp. 6-10.
ABSTRACT: A brief history of ethylene glycol application as an engine coolant is presented.
Concurrent engine cooling system corrosion problems and coolant corrosion inhibition require-
ments are reported. Comments on current engine system corrosion problems are provided.
Resistance to corrosion failure is shown to depend upon the correct combination of cooling sys-
tem design, materials, coolant inhibition, and coolant retention.
KEYWORDS: ethylene glycol, engine coolants, engine cooling systems, corrosion, corrosion
inhibition
Automotive engine coolant technology probably began in 1885 when Karl Benz invented
and patented the first automotive radiator to provide recirculation cooling for the water-
cooled engine that he built for his first horseless carriage. The radiator was developed to elim-
inate the problem with evaporative cooling, which boiled away one gallon of water each hour
of operation of the single cylinder engine.
It is interesting to note that ethylene glycol, propylene glycol, and their derivatives were first
synthesized in 1859 by Charles Wurtz, a French chemist. It was not until World War I that a
commercial process for making ethylene glycol from alcohol was developed in Germany for
use in explosives.
The first engine coolant application of ethylene glycol was suggested in England in 1916 for
high performance military aircraft engines. In the United States, the initial experimental glycol
coolant applications took place in 1923. Shortly thereafter, a liquid cooled Curtiss Navy Racer
captured the world seaplane speed record. The monoplane was powered with a Curtiss
inverted V 12 aluminum engine cooled with an ethylene glycol-water solution. The cooling
system used heat exchanger panels in the wings to combine cooling with aerodynamic effi-
ciency. By 1926, this engine, equipped with an underslung radiator and cooled with an eth-
ylene glycol-water solution, was the power plant for the Curtiss Hawk P 1A, which became the
standard U.S. Army Air Corps pursuit plane. Parallel development of glycol cooling was
underway in Great Britain and Europe. The recognized advantage of glycol-water coolants was
the high boiling point, which permitted high temperature cooling with reduced frontal area.
Also, the lower vapor pressures raised the threshold of coolant pump cavitation, enabling oper-
ation at higher altitude. In the following years, coolant inhibitors were developed to control
corrosion and coolant degradation at coolant temperatures as high as 275~ (135~ in normal
pressurized operation. Triethanolamine phosphate MBT inhibitor compositions were even-
tually specified for military aircraft coolants both in England and the United States.
Although a U.S. patent was issued in 1918 for the use of ethylene-glycol to lower the freezing
Consultant, International Copper Assoc., Ltd., Dallas, PA.
6
HANNIGAN ON A REVIEW OF ENGINE COOLANT TECHNOLOGY 7
point of water in automobile cooling systems, large scale production of ethylene glycol did not
begin until 1925. At that time, Union Carbide and Carbon Corporation was the sole volume
producer of ethylene glycol until the beginning of World War II. Initially, small quantities of
uninhibited ethylene glycol were sold for automotive antifreeze use. Very soon it became
apparent that uninhibited glycol solutions could become corrosive to cooling system metals.
Engine cooling systems were commonly afflicted with large amounts of residual water jacket
rust, even when new. Substantial coolant pump air induction was to be expected and cylinder
heat joint exhaust gas leakage was a fact of life.
Prestone brand ethylene glycol antifreeze was first marketed by the National Carbon Co. in
1927. It was the only commercial glycol based antifreeze brand. Its competition was glycerol
and the older existing brands of methanol and alcohol volatile antifreezes.
By 1930, the National Carbon Co. and the Linde Research Laboratories jointly developed
a two-phase chemical and insoluble oil inhibitor system to prevent corrosion, rust loosening,
and rust transport to the radiator. Mr. Daniel H. Green of the National Carbon Co. was
appointed as manager of the automotive engineering department to direct coolant research,
applications engineering, and field technical support. This position presented a real challenge
of major proportions. Here was a new brand, a new product, with no track record in the auto-
motive after-market, or the automotive industry. It was priced at $5.00 per gallon, against alco-
hol brands selling for about $1.00 per gallon in a depression economy. Like the good news and
bad news stories, the good news might have been that things couldn't be any worse. Then
again, maybe they could.
In many instances, automobile factory engineering personnel, service representatives, and
after-market service dealers firmly believed that ethylene glycol solutions "leaked more than
water" and "ate through" cylinder head gaskets and coolant pump seals, and "would find a
leak where water wouldn't." The high visibility of leakage residue with glycol solutions con-
tributed to leakage misconceptions. Water and alcohol solutions would quickly evaporate at
the point of leakage. In one situation where internal coolant leakage and engine seizures were
at issue, a laboratory supervisor at a leading car manufacturer declared, "The only way to keep
ethylene glycol out of the crankcase is to keep it out of the cooling system." The fact was that
cylinder head gaskets were leaking frequently due to lack of recommended retorquing of the
cylinder head bolts after the engine was operated for a short time. Cylinder head gaskets were
relatively thick and conformable with very little torque retention.
The internal coolant leakage problem disappeared in the middle 1950s with the advent of
numerous higher performance overhead valve V8 engines. These were designed with much
improved cylinder head joints to withstand increased compression ratios and higher combus-
tion pressures. Between 1949 and 1955, the U.S. car manufacturers recognized the need for
increased coolant temperatures to reduce engine sludging and wear. Consequently, ethylene
glycol base coolant was widely adopted for factory fill in place of methanol. Thermostat open-
ing temperatures were increased to 180~ (82~ Water with corrosion inhibitors was still
installed at the factory during the warm-weather months.
Methanol and ethanol were steadily losing market share. Little or no inhibitor research and
development were dedicated to alcohol antifreezes. The usual inhibitors were borate or nitrite,
offering marginal, if any, corrosion protection to aluminum.
Despite the limited concern about aluminum corrosion, some of the older L head engines
were equipped with aluminum cylinder heads. Examples were the 1934 and 1935 Ford V8s
and the 1937 and later Lincoln Zephyr V 12s. Other car manufacturers offered optional higher
compression ratio cylinder heads of aluminum. Crevice corrosion at the cylinder head joint
was a frequent problem. Consequently, the car manufacturers reverted to cast iron cylinder
heads.
From the post war period to about 1960, commercial North American glycol coolant inhib-
8 ENGINECOOLANTTESTING: THIRD VOLUME
itors were usually borate buffered with small amounts of various supplementary inhibitors,
such as sodium salts of MBT, arsenite, nitrite, molybdate, or phosphate. Second phase insol-
uble polar oils were also used as supplementary inhibitors.
In Europe and England, the British Standards Institute composition specifications domi-
nated the markets. BS 3150 Type A was a triethanolamine orthophosphate, sodium MBT
inhibitor system. BS 3151 Type B consisted of sodium benzoate and sodium nitrite. BS 3150
Type A was used with aluminum component engines and BS 3151 Type B with cast iron
engines. Glycol concentrations were held to the minimum in Europe and England. Ethylene
glycol was widely regarded as an inferior heat transfer agent, to be removed as soon as possible
at the arrival of warm weather.
Up to the 1950s, there was little or no uniform performance testing among the car factories
and the antifreeze industry. Each company, if indeed it did coolant testing or research, or both,
devised its own in-house tests. These usually consisted of a glassware test method, a bench
circulating test on occasion, and short-term vehicle tests using subjective evaluation of engine
cooling system components, such as coolant hoses, coolant pumps, thermostats, and coolant
outlets.
The Linde Laboratory of Union Carbide pioneered the full-scale engine dynamometer test
method with an interesting approach. The L head Ford V8, in production from 1932 through
1953, was designed with a coolant pump for each cylinder bank along with separate outlet and
inlet radiator hoses on each side. By modifying the upper and lower radiator tanks with a cen-
ter sealing baffle the engine could be operated having identical twin cooling systems with
exactly same coolant temperatures, temperature differentials, heat load, and heat flux concen-
trations. This unusual feature enabled performance evaluation of two different coolant for-
mulations simultaneously under precisely same conditions.
Beginning in 1958, radiator core solder alloys for original equipment radiators were changed
from 70% lead, 30% tin to about 97% lead, 3% tin composition. Radiator solder corrosion
increased dramatically. Voluminous corrosion consisting of lead hydroxide and occluded
inhibitors would obstruct the radiator tube openings causing restricted coolant flow and over-
heating. Coolant suppliers were caught unaware of this situation. Manufacturing methods
were also involved. If the radiator tube-to-header joints were soldered externally, little or no
tube opening obstruction occurred. Where the tube-to-header joints were soldered internally,
excess solder would enter the tube opening by capillary action and cause corrosion clogging.
Residual flux containing chlorides also increased the rate and volume of corrosion. To prevent
the problem, insoluble oils and emulsifiable oils were installed with the factory fill coolant on
the assembly line by some car manufacturers. High lead solder corrosion proved diIticult to
control with single phase soluble salts inhibitor systems. Short term laboratory tests could
identify more aggressive inhibitor systems, but could not predict long term protection.
Extended vehicle tests lasting 12 months or more became necessary to reliably indicate satis-
factory in situ high lead solder inhibition. Recent field surveys of radiator corrosion modes
show solder bloom in 5-year-old and older cars to still be prevalent. In contrast, Japanese radi-
ators soldered with 60% to 40% or 70% to 30% lead-tin solder do not encounter solder bloom
even after long-term use and marginal coolant maintenance.
In 1962 and 1963, the U.S. car factories adopted year-round installation of inhibited eth-
ylene glycol coolants. This was done to eliminate freeze damage to left over new car inventories
and to provide increased corrosion protection. To assure adequate freeze protection, mini-
m u m glycol concentration was set at 44% (-20*F) ( - 2 9 ~ in the United States and 53%
(-40*F) ( - 4 0 ~ in Canada. The antifreeze industry changed its recommendation from 33%
minimum glycol concentration to 50% to ensure year-round adequate corrosion protection.
This period marked the introduction of die cast aluminum coolant pumps and mating die
cast aluminum timing chain covers on most of the then current production V8 cylinder
HANNIGAN ON A REVIEW OF ENGINE COOLANT TECHNOLOGY 9
engines. Before long, the pumps and covers were encountering perforation from cavitation-
erosion-corrosion in police fleets and other high engine speed duty such as ambulance oper-
ation. Existing circulating tests, engine dynamometer tests, and company fleet tests did not
produce the attack. Special aluminum pump cavitation-erosion-corrosion tests were devel-
oped by the car factories and were adopted by the coolant suppliers to develop improved inhib-
itor systems. Increased phosphate inhibitor concentration, along with adequate coolant con-
centrations provided control of the problem.
Cooling system design factors showed a significant effect on pump cavitation corrosion inci-
dence and severity. Pump impeller clearances and configuration in relation to housing scroll
configuration and internal pressure differentials could improve or worsen cavitation corrosion
attack. Today, North American factory fill inhibitor systems provide very satisfactory inhibi-
tion of aluminum pump cavitation even with marginal designs.
In the mid 1960s, there were early clues to the potential problem of aluminum cylinder head
heat rejecting cavitation corrosion and transport deposition at the radiator and heater core.
U.S. rebuilt P51 fighter plane cooling systems were filled with high phosphate inhibited auto-
motive coolant. The fighters were flown to South American countries for military duty. By the
time they arrived, the engines were overheating due to aluminum corrosion deposition in the
radiator. This did not happen when the Air Force specification triethanolamine phosphate
inhibited coolant was used.
Occasionally an imported car aluminum cylinder head would leak coolant into an exhaust
port due to perforation of the exhaust port wall from the coolant side. Usually, the failure was
accompanied by a partially clogged radiator core. Certain model U.S. cars equipped with alu-
minum cylinder heads and factory fill high RA borate phosphate formulas would develop
overheating from radiator deposition when the cars were driven from Michigan to Arizona for
proving ground track testing. Other lower RA differently balanced inhibited coolants did not
cause the problem. At the time, none of these situations appeared significant enough to gen-
erate a demand for new inhibitor technology.
In 1973, one of the U.S. car manufacturers elected to use an aluminum cylinder head engine
in a new line of cars to be introduced in 1974. Durability track testing revealed radiator depo-
sition caused engine overheating about halfway through the test schedule. The coolant inhib-
itor system in use was the factory specified high RA, high phosphate formula. Further com-
parative testing established that an existing stabilized silicate supplemented coolant
composition widely available in the after-market effectively controlled the problem. This cool-
ant brand was approved for the new car factory fill in 1974. Subsequent coolant research and
development testing revealed the mechanism of cavitation erosion corrosion in relation to
nucleate boiling at aluminum heat rejecting surfaces.
It became apparent that cylinder head design could be a critical factor in relation to cavi-
tation corrosion. Coolant jackets having very high surface to volume ratios, siamesed exhaust
ports, drilled passages in attempt to relieve "hot spots," and other marginal coolant flow dis-
tribution conditions were much more prone to cavitation corrosion and transport deposition.
Coolant composition has played an interesting role in designing the engine cooling system
to eliminate or minimize heat transfer deficiencies. In the early 1960s, International Harvester
Engine Div. developed an engine dynamometer test to identify high heat flux areas and inad-
equate coolant flow. A magnesium borate inhibited coolant solution provided a reliable
graphic in situ pattern of heat transfer in the engine cylinder head and cylinder block. The
magnesium borate inhibitor deposited progressively on increasingly higher temperature heat
rejecting surfaces. This technique is currently used by some U.S. car manufacturers to confirm
satisfactory cooling in a new engine design.
It is worth noting that the discovery of this invention was serendipitous. In 1950, magne-
sium borate was used in a Canadian after-market antifreeze. The product was used in a road-
10 ENGINE COOLANT TESTING: THIRD VOLUME
grader engine that suffered cylinder head cracking in the combustion chamber area. Excessive
inhibitor deposition was concentrated at the high heat flux areas. The engine manufacturer
and the antifreeze supplier were at issue whether the coolant caused the cracking or the cyl-
inder head design was deficient. Subsequently, the engine manufacturer increased the by-pass
flow rate through the engine and the coolant supplier discontinued the magnesium borate
inhibitor system.
At present, contemporary brazed aluminum radiators and heater cores have shown down-
tube random pitting to be a significant problem for the coolant inhibitor system, in the south-
ern tier and west coast of the United States. Inhibitor depletion due to loss of glycol concen-
tration continues to be a c o m m o n condition in this southern region with three year old and
older passenger cars. Also, the lack of freezing weather provides a market for bogus "engine
coolant" labeled products which are little or no more than colored water. The use o f these
products in a brazed aluminum radiator will program the radiator for tube pitting perforation
failure. Perforation can occur as early as 6 weeks and certainly within 12 months, depending
on the chloride ion concentration in the coolant water source. In the middle U.S. and northern
frontier areas premature pitting perforation failures have been relatively rare. None-the-less,
the presence of high chloride concentration may initiate pitting despite adequate glycol con-
centration and normal inhibitor condition.
In retrospect, these experiences confirm that coolant technology is driven by material and
design changes in the engine cooling system. In terms of engine durability, coolant side cor-
rosion is a critical factor. Long term resistance to failure from corrosion is inescapably depen-
dent upon the combined effects of coolant inhibition and the materials and design of the total
engine cooling system.
Jean-Pierre Maes ~and Paul Van de Ven
Corrosion Protection of Aluminum Heat-

Transfer Surfaces in Engine Coolants Using
Monoacid/Diacid Inhibitor Technology
REFERENCE: Maes, J.-P. and Van de Ven, P., "Corrosion Protection of Aluminum Heat-
Transfer Surfaces in Engine Coolants Using Monoacid/Diacid Inhibitor Technology," Engine
Coolant Testing: Third Volume. ASTMSTP 1192, R. E. Beal, Ed., American Society for Testing
and Materials, Philadelphia, 1993, pp. 11-24.
ABSTRACT: A low depletion monoacid/diacid corrosion inhibitor technology that provides

superior high-temperature aluminum corrosion protection combined with excellent heat-trans-
fer characteristics has been identified. The etficiency of the inhibitor technology was evaluated
gravimetrically on aluminum specimens under static and dynamic heat-transfer conditions.
Testing techniques and results are discussed. The thermal properties of the protective film and
the coolant were evaluated under dynamic heat-exchange conditions. Information on compo-
sition and structure of the protective film formed under conditions of heat-transfer was obtained
through the use of microscopy and Fourier transform infrared spectrometry (FTIR).
KEYWORDS: engine coolants, corrosion, inhibitors, aliphatic acid, heat-transfer, monoacids,

diacids, aluminum, heat-exchange surface
The physical and, more particularly, the heat-transfer properties of a coolant are largely
determined by the choice of the freezing point depressant. Specific heat, thermal conductivity,
fluidity, freezing point, and boiling point relate directly to the nature and the concentration of
the diol used. The chemical and corrosion properties can however be controlled by the use of
additives. Reserve alkalinity and hard-water stability can be improved by the use o f a pH buffer
and stabilizer, respectively. Additionally coolants may include an antifoaming compound, a
dispersant to avoid deposit formation, and a dye for identification.
Corrosion in an engine cooling system will have two main effects: (1) deterioration of the
metal component either by uniform wastage or localized attack (pitting, crevice corrosion) and
(2) the production of insoluble corrosion products that can block radiators, thermostat valves,
and so forth, and impede heat-transfer by deposition on the heat exchange surfaces. The major
additive constituents of engine coolants are therefore corrosion inhibitors.
Because of the increased use of aluminum and other light metals for engine and cooling part
components, the inhibition of localized forms of corrosion, such as pitting and crevice cor-
rosion, has become crucial. Several engine manufacturers have introduced test requirements
that relate directly to protection against localized corrosion of aluminum. More specifically,
the galvanic pitting corrosion test [1] was introduced as a measure of the long-term effective-
ness of coolants in preventing pitting corrosion attack in aluminum heat exchangers. Together
Research group leader and development chemist, respectively, Texaco Research and Development--
Ghent, Technologiepark--Zwignaarde 2, Ghent, Belgium.
11
with some conventional tests, cyclic polarization techniques can be used to determine suscep-
tibility to pitting and crevice corrosion [2,3].
Early inhibited coolant formulations in Europe contained a combination of triethanol
amine, phosphoric acid, and sodium mercaptobenzothiazole (MBT). Triethanol amine phos-
phate (TEP) coolants are still used, but most European coolants are now based on a borate-
benzoate package containing varying amounts of nitrite, nitrate, silicate, and triazole. In
recent formulations salts of organic aliphatic acids have been substituted for nitrite.
In the United States, coolants currently on the market may contain a relatively large amount
of phosphate and varying amounts of nitrate, silicate, and triazole. Other formulations with
less phosphate, contain borate in addition to nitrate, silicate, and triazole.
As experience was growing [4], some adverse properties of the used inhibitors and inhibitor
combinations have come to light. Sodium mercaptobenzothiazole, for example, relatively
insoluble at low temperatures, which can form insoluble calcium salts in hard waters, has been
largely replaced by benzotriazole or tolyltriazole. In the absence of nitrite, benzoate is rela-
tively ineffective for the protection of steel and cast iron. If not correctly stabilized, silicate can
form gels on standing. Borate can cause corrosion of aluminum alloys at high temperatures
[5] as well as phosphate that can also be precipitated by hard waters.
Silicate is still considered to be an effective inhibitor of aluminum corrosion, and silicate
stabilizers have alleviated some of the problems associated with gel formation and precipita-
tion of this chemical.
While these disadvantages were being discovered, encouraging progress was being made in
inhibitor development. Based on work by Butler and Mercer [6], disodium sebacate, the
sodium salt of decandioic acid, was found to be an effective corrosion inhibitor for both ferrous
and aluminum alloys. Various other monobasic and dibasic carboxylic acids have also been
identified as effective corrosion inhibitors by Hersch et al. [ 7].
Where conventional inhibitors, such as nitrite and silicate, are subject to rather rapid deple-
tion in service, this is not the case for organic monoacids or diacid salts [8]. The introduction
of formulations that use organic acid salts as the major inhibitor constituent has opened new
prospectives for the development of long-life coolants.
Work by Darden et al. [9] showed that inhibitor combinations of monoacid and diacid salts
provide improved protection of aluminum, particularly against pitting and crevice corrosion.
The cyclic polarization scans for different combinations of monoacids and diacids are shown
in Fig. 1.
The higher pitting potentials found for aluminum in synergistic combinations of the acid
salts indicate improved protection against pitting [2,9]. The difference between the pitting and
protection potentials, as determined by cyclic polarization, can be correlated directly with sus-
ceptibility to crevice corrosion [3]. Reduced hysteresis indicates improved ability to repassi-
vate active corrosion cells.
Most current coolant formulations include silicate for high-temperature corrosion protec-
tion of aluminum heat-transfer surfaces. However, because of the limited solubility of silicate
in propylene glycol, monoacid-diacid inhibitor technology also appears to be a preferred
choice for incorporation in less-toxic propylene glycol based coolants.
These silicate-free inhibitor formulations offer not only improved corrosion protection and
a stability that is easier to maintain but also contribute to the durability of water pump seals
[lO].
The etficiencies and heat-transfer properties of monoacid- diacid inhibitors have been eval-
uated on aluminum coupons under static and dynamic heat-transfer conditions. Test results
have been compared to the results obtained with more traditional coolant solutions.
2H00 '"'""I '"""I '"""I '"""I '""'I '"''I '"~" ,"v,,] '"I"'] "q'"] '"r"] '"r"j '"p"] '"p"
'"'"1 '""'1 '"~"1 '"'"1
ALLiHINUH ALLiHINUH ALUH INUH
t68~
3
z
W
2qO
/ .'..
r
w
>
m
>
z
o
>
z
_<r
o
m
<
m
-45~ z
0
z
c)
0
- 1200 , ,,i,d , .,h,,J .,,t,.,J , ,,~,,,,i , ,,I.J ,,,~,,,J ,.I . . . . . I..I , .h.J ,,,h.] , ,,h,,.[ , ,,I,,.I ,, ,h,,] t,,I,,, ....,,I , ,,,,,.1 , ,,.,,l , ,,.,,,I , ,,,,,.I ,,,.,J , ,,.,
-2 -1 '~ 1 2 .3 -2 -1 0 1 2 3 4 -2 -1 0 1 2 3 0
10 10 1~ 16 10 10 1~ 10 10 t0 10 10 t0 10 10 10 10 10 10 10 10
I (UA/CH^2) I (U~/CH^2) 0
I (UA/CH.,.2) z
"0
F I G . l - - E l e c t r o c h e m i c a l corrosion test results on aluminum from left to right." increasing amount of
dibasic acid and decreasing amount of monoacid. 0
-I
m
(3
co
Experimental Procedures
Performance Tests
Evaluating the heat transfer and high-temperature corrosion properties of an engine coolant
requires specific testing. Test conditions have to be controlled and recorded for analysis, and
repeatability has to be ensured. Two computer controlled test rigs were constructed for this
purpose. The performance of the coolant at high temperature was evaluated by weight loss
examination of the test specimens, inspection of the test rig components, and analysis of the
cooling fluid.
ASTM Test Method for Corrosion of Cast Aluminum Alloys in Engine Coolants under
Heat-Rejecting Conditions (D 4340) covers a laboratory screening procedure for evaluating
the effectiveness of engine coolants in preventing corrosion of aluminum castings under static
heat-transfer conditions.
The dynamic heat transfer test evaluates corrosion protection at heat-transfer surfaces on
metal specimens under dynamic conditions. Similar test rigs have been described in the lit-
erature [11,12]. The test apparatus (Fig. 2) consists of a circulating rig comprising a pressur-
ized container with a cooling element, a pump, and two corrosion specimens mounted in a
high- temperature corrosion chamber with a heating block, inducing a constant heat input into
one of the specimens. The test fluid was a 20% volume mixture of the engine coolant in de-
ionized water.
The aging test simulates the aging process of a coolant in use by alternately heating and
cooling the coolant while circulating it through the pressurized test rig (Fig. 3). Again a 20%
volume solution of the engine coolant in de-ionized water is used. Corrosion during the aging
cycle was monitored by weight loss of metal specimens, similar to those used in the ASTM
Test Method for Corrosion Test for Engine Coolants in Glassware (D 1384). The aged coolant
was re-used in the dynamic heat transfer test to assess performance of the aged coolant.
Results
High Temperature Corrosion Tests
Some of the advantages of monoacid-diacid inhibitor combinations versus conventional
inhibitor technologies were put forward earlier. The performance testing at high temperature
of the different inhibitor technologies included:
9 Monoacid-dibasic acid technology (Sample A)

9 Nitrite-free formulations (Sample B)
9 Conventional benzoate-nitrite-silicate technology: high and low silicate (Samples C and
D, respectively)
9 Inorganic phosphate technology (Sample E)
9 Triethanol amine phosphate technology (Sample F)
All tested formulations are top quality commercial engine coolants available on the European
or U.S. market. Table l provides a listing of the inhibitors used in these coolants. The test data
were developed mainly for ethylene glycol based coolants. Test results for propylene glycol (1,2
propanediol) based coolants are included for the monoacid-diacid formulation only (Sample
Ap). The additive packages in Sample A and Ap are identical (Tables 2 and 3).
Two different aluminum alloys were used: SAE 329 (UNS A03190) and alloy 6082 (DIN
AIMgSil). SAE 329 is the cast aluminum alloy used in ASTM D 1384 and D 4340. The micro-
structure of this material usually shows segregation (dendrites) of silicon-rich phases and some
MAES AND VAN DE VEN ON CORROSION PROTECTION 15
T~ST PARAMETERS
Aluminum
Controlled
N o m i n a l h e a t input 2000 W
Input temperature* 85 ~
Coupon Surface 25.7 cm 2
Pressure 1.5 atm.
Flow 3.5 i/min
Fluid c o n t e n t 1 1
Monitored
Block temperature 240 ~

Output temperature 92 ~
T e m p e r a t u r e of the 92 ~
u p p e r coupon
T e m p e r a t u r e of the 160 QC
lower coupon
T e m p e r a t u r e in the 86 ~
c o o l i n g vessel
PT-100 is a Resistance Temperature Device
FIG. 2--Dynamic heat transfer test.
TABLE 1--Engine coolant inhibitors.
Inhibitor A B C D E F
Borax X X X X X
Sodium Benzoate X X X
NAN02 X X
NaNO3 X X X X
Triazole X X X X X X
NaSiO3.5aq. X X X X
Phosphoric acid X X
Triethanol amine X
Monoacid X X
Dibasic acid X
NOTE: X is presence of inhibitor shown by analysis.

TEST PARAMETERS
Controlled
Temperature in the 115 ~

heating vessel
cooling vessel
Pressure 2.5 atm.
Flow 3.5 i / m i n
F l u i d content 6 1
Test duration 504 h
Monitored
Nominal heat input 5000 W

c o u p o n vessel
FIG. 3--Aging test.
cast porosities. The wrought alloy 6082 material has a more uniform microstructure and con-
tains about 1% magnesium.
Results o f the ASTM D 4340 tests are reported in Table 2. The tests were run under standard
ASTM conditions for a 25% solution of the different engine coolants. Passing results (ASTM
Specification for Ethlyene Glycol Base Engine Coolant, for Automobile and Light Duty Ser-
TABLE 2 - - T h e A S T M D 4340 corrosion test.
Corrosion Results, mg/cm2/week

Test A Ap B C D E F
SAE 329 0.30 0.32 0.39 0.21 0.07 0.00 0.29

Alloy 6082 0.09 -0.01 6.73 --0.05 --0.06 -0.05 0.76
NOTE: Negative values indicate a weight gain. ASTM D 3306: maximum 1 mg/cm2/week.
MAES AND VAN DEVEN ON CORROSION PROTECTION 17
vice [D 3306] for ASTM D 4340: m a x i m u m 1 mg/cm-'/week) were obtained for all the tested
coolants. This confirmed that the selected coolants are of high quality and representative for
their technology group. Good results were also obtained with alloy 6082, in monoacid-diacid
and in the conventional silicate and phosphate coolants.
Different corrosion rates for the tested coolants were found in the dynamic heat transfer test
for SAE 329 test specimens (Table 3). Low silicate conventional coolants and the phosphate
coolants showed high corrosion rates, even after a test duration of only 48 h. The high silicate
conventional coolant performed well, but high corrosion rates were found when the test dura-
tion was extended to 69 h. The monoacid-diacid coolant (Sample A) showed consistently low
corrosion, even when the test duration was extended to 116 h and aged test solutions were
used. Corrosion during "aging" for 504 h (not shown in Table 3) ranged from 69 mg/coupon
for Coolant A to 384 mg/coupon for Coolant C (SAE 329 corrosion coupons).
Heat- Transfer Properties

Coupon temperatures were monitored during the dynamic heat transfer tests. A constant
heat input was maintained (2000 W). Figure 4a shows the mid-section temperature rise of the
heated a l u m i n u m coupon (SAE 329) as recorded for the best performing conventional cool-
ant, Sample C (high silicate), and the monoacid-dibasic acid coolant, Sample A. This temper-
ature pattern can be related to the heat transfer through the protection/corrosion layer at the
metal/liquid interface.
The coupon/protective layer system can be considered as serial steady-state heat conduc-
tors. The heat flux through each layer is the same and equal to the power received by the flow-
ing coolant solution (the heat flux actually induced will be slightly higher as there is some loss
of power by radiation). The power received by the fluid can be calculated as:
P = &,cAT
where
P = received power,
dm= mass flow,
c = heat capacity of the fluid, and
AT = temperature difference of the fluid before and after heating.
TABLE 3--Dynamic heat transfer test.
Corrosion Results, mg/coupon
Test Duration, h A Ap B C D E F
SAE 329
48 10.7 4.2 512 - 18.0 607
69 26.0 10.3 ... 261 1"2"6"0 1243 1"5"1"0
116 21.6 46.5 . . . . . . . . .
aged 48 . . . . . . 255 . . . . . .
aged 116 '~8"4 "'" ALLO'Y6082 . . . . . . . . .
48 ... 2130 ... 1600
116 23.() "8.7 . . . . . . . . . . . .
aged 116 34.1 . . . . . . . . . . . . . . .
NOTE: Negative values indicate a weight gain.
18 ENGINE COOLANT TESTING: THIRD V O L U M E
2OO
1gO J
if
R
|~ I r r ~ I t
0 10 20 .~0 40 60 eO 7O 80
"nine (hi
O.6
0.8
,p
02
.,%,**.
.,<
<1 0,1
sua~A
0 ' I L l J L i I
0 10 ~0 ~0 40 ~ 80 70 80
Tlme (h)
FIG. 4--(a) Coupon temperature versus test duration and(b) heat transfer characteristics (Axp,) versus
test duration.
Actual values for the power received by the fluid based on the monitored temperature differ-
entials are approximately 1850 W for a controlled total input power of 2000 W.
The heat flux ff is given by
~=Sq
where S is the surface through which heat is transferred and q is the heat-current density. The
heat-current density can be calculated using Fourier's law [13]
MAES AND VAN oE VEN ON CORROSION PROTECTION 19
q = - - X - - AT
Ax
where
= heat transfer coefficient,
AT = temperature drop through the metal, and
s = metal thickness.
Knowing the heat transfer coefficient for SAE 329 [14]
)~ = 1.09 10 2Jm IK-Js-I
and the temperature at a point xT, we can calculate the temperature profile through the alu-
minum coupon and deduce the temperature of the protective layer at the interface with the
aluminum surface.
When we assume the temperature of the layer surface at the interface with the fluid to be at
the boiling temperature of the coolant solution ( 123~ for a 20% mixture of ethylene glycol in
water), we can calculate the temperature drop over the protective layer. This temperature dif-
ference is a direct measure for the ratio of thickness of the layer over the heat transfer coeffi-
cient (Fig. 4b).
Microscopic Examination
Figure 5 shows micrographs (dark field illumination) of surface layers formed on the heat-
transfer surfaces of the heated SAE 329 coupons after dynamic heat transfer tests in the high
silicate conventional coolant (Sample C, Fig. 5a) and the monoacid-diacid coolant (Sample
A, Fig. 5b), respectively. The conventional coolant (Sample C) produces a uniform surface or
reaction layer on the coupon surface while the aliphatic acids form a structured film covering
only certain areas that might be associated with segregation phases (silicon-rich dendrites) or
material defects (cast porosity, inclusions) on the metal surface while the remaining surfaces
are apparently uncovered.
Analyses of Protective Film

The University of Antwerp (Universitaire Instellingen Antwerpen) was contacted to analyze
the film formed on the heat-transfer surfaces in coolant A. FTIR studies of the coupon surfaces
indicated the presence of organic matter, while X-ray fluorescence analyses of scrapings indi-
cated the presence of silicon and aluminum. Although this work is not decisive in identifying
the film composition, it indicates that an interaction between the organic acids and the metal
surface occurs, as suggested by Darden et al. [9]. Further work, which includes scanning elec-
tron microscopy (SEM) and FTIR mapping of the coupon surface, is ongoing.
Discussion
High- Temperature Corrosion Protection
Different inhibitor technologies have been evaluated. Phosphate and conventional silicate
coolants provide good protection of aluminum heat-transfer surfaces under the conditions of
the ASTM D 4340 static heat-transfer test. No significantly different weight losses are
observed. The corrosion rates are well within the specification limits.
The dynamic heat transfer test clearly distinguishes between the different inhibitor tech-
FIG. 5--Protective layer on the aluminum surface from Sample C (top) and Sample A (bottom).
nologies. High corrosion rates are found for the phosphate containing coolants under tests
conditions of high fluid flow and heat flux. The low silicate conventional and nitrite-free cool-
ants provide better protection, but corrosion rates are still high. High silicate coolants will pro-
vide protection in the short duration heat transfer test (48 h). Silicate coolants and particularly
high silicate coolants generally perform well in short duration tests [ 11,12]. Silicate adsorbed
on the metal surfaces and retained in cast porosities will protect the metal surface and con-
tribute to some weight gain. However, silicate depletion eventually results in high corrosion
rates when test duration is extended, or when aged coolant is used. High corrosion rates are
also found with the low silicate containing alloy 6082.
The synergistic [9,15J combination ofaliphatic monoacid and diacid salts provides lasting
protection even when test duration is extended. Under heat-transfer conditions a thin organic
containing film or reaction layer is apparently formed on selective areas of cast microstructure.
Chemical cleaning after the test, according to the FVV cleaning procedures [12], generally
MAES AND VAN oE VEN ON CORROSION PROTECTION 21
does not remove this protective layer. After the 48-h tests, a weight loss is observed for the
different aluminum coupons. Extending the test period does not significantly increase the
measured weight losses. This becomes even more manifest when alloy 6082 is used.
Tentative protection mechanisms for aluminum provided by the aliphatic acids have been
given by Darden et al. [9]. In case of film formation, the inhibitor anion would complex with
the metal while it is still bound to its solid lattice. No bulk layer is formed, rather a layer of
microscopic thickness is formed at the anodic sites of the metal surface. This is confirmed by
our current microscopic and analytical work.
Heat- Transfer
The heat transfer characteristics of the silicate film and aliphatic acid film differ widely. The
difference in temperature drop over the protective layers can be explained on the basis of the
postulated protection mechanisms. If we assume that the film heat-transfer coefficient X is a
constant factor over the temperature interval, the shape of the film thickness over heat transfer
coefficient (Ax/X) curve (Fig. 4b) provides an indication of protective film buildup and decay.
The silicate film (Sample C) grows fast initially. After 10 to 15 h, the growth rate stabilizes.
The effect of inhibitor depletion, marked by increasing corrosion rates between 48 and 69 h
of test duration, is not evident from Fig. 4b. This is not unusual, as corrosion products will
gradually substitute the decaying silicate film. As such, the heat-transfer properties of the "pro-
tective" film are those of the combined silicate/corrosion layer gradually shifting to those of a
corroded metal surface.
For the monoacid-dibasic acid coolant (Sample A), Fig. 4b shows only a small initial growth
of a protective layer, after which the heat-transfer properties related to ~ / X remain practically
constant. The excellent heat-transfer characteristics can also be explained by the selective way
the film is formed on anodic areas in the microstructure (Fig. 5b), leaving sufficient blank areas
for good heat transfer. On wrought materials, such as alloy 6082, a more uniform but lighter
film was formed as the microstructure of the metal is more homogeneous.
Further work, which should provide more insight in the actual protection mechanisms, is
planned.
Conclnsion
Synergistic inhibitor combinations of monoacid and diacid salts have been shown to pro-
vide good corrosion protection at high-temperature heat-transfer surfaces. Improved heat-
transfer properties are found compared to conventional coolants. These inhibitor formula-
tions also provide good corrosion protection against localized forms of corrosion, such as
pitting and crevice corrosion.
The inhibitors are virtually depletion-free and can be used in less-toxic propylene glycol
coolants. As development work is continuing, further advantages of this technology come to
light making it an ideal candidate for use in future world-wide, long-life engine coolants.
References
[1] Wiggle, R. R., H ospadaruk, V., and Stylogou, F. A., "The Effectiveness of Engine Coolant lnhibitors
for Aluminum," Corrosion 80, National Association of Corrosion Engineers Conference, Houston
Paper 69, 1980.
[2] Bond, A. P., "Pitting Corrosion-A Review of Recent Advances in Testing Methods and Interpreta-
tion," Localized Corrosion-Cause of Metal Failure, STP 516, M. Henthorne, Ed., American Society
for Testing and Materials, Philadelphia, 1971, pp. 250-261.
[3] France, W. D., Jr., "Crevice Corrosion of Metals," Localized Corrosion-Cause of Metal Failure,
STP 516, M. Henthorne, Ed., American Society for Testing and Materials, Philadelphia, 1971, pp.
164-200.
[4] Mercer, A. D., "Corrosion Inhibition of Engine Coolants," Corrosion, Protection and Finishing
Techniques, The Institution of Mechanical Engineers, Automotive Division, International Seminar,
1985.
[5] Wiggle, R. R., Hospadaruk, V., and Tibaudo, F. M., "Corrosion of Cast Aluminum Alloys under
Heat-Transfer Conditions," SAE paper 810038, Society of Automotive Engineers, Warrendale, PA,
Feb. 1981.
[6] Buter G. and Mercer A. D. ``nhibitr Frmuatins fr Engine Cants British Crrsin Jur-
nal, Vol. 12, No. 3, 1977, pp. 171-174.
[ 7] Hersch, P., Hare, J. B., and Sutherland, S. M., "An Experimental Survey of Rust Preservatives in
Water: II. The Screening of Organic lnhibitors," Journal of Applied Chemistry, Vol. 11, 1961, pp.
261-271.
[8] Burns, J. M., Field Test Data, to be published.
[9] Darden, J. W., Triebel, C. A., Maes, J. P., and Van Neste, W., "Monoacid-Diacid Combinations as
Corrosion Inhibitors in Antifreeze Formulations," Worldwide Trends in Engine Coolants, Coolant
System Materials and Testing, SP-811, Paper 900804, Society of Automotive Engineers, Warren-
dale, PA, 1990.
[10] Burns, J. M., "Long-Life Engine Coolants for Improved Water Pump Seal Durabilityand Extended
Change Intervals," Engine Coolant Testing: Third Volume, STP 1192, American Society for Testing
and Materials, Philadelphia, 1993, pp. 25-43 (this publication).
[11] Frey, G., Liebold, G., and Starke, K., "Anforderungen an moderne Kiihlstoffe (Requirements for
Modern Engine Coolants)," Mineral6l Technik, Beratungsgesellschaft fiir MineraliSI-Anwendung-
stechnik mbH, Hamburg, Germany, Nov. 1985.
[12] Forschungsvereinigung Verbrennungskraftmachinen e.V., "Prtifung der Eignung yon Ktihlmittel-
zus~iten ftir die KtihlflOssigkeiten von Verbrennungsmotoren (Testing of Coolant Additive Proper-
ties to be used in Engine Coolants)," FVV, Heft R443, Frankfurt, Germany, 1986.
[13] Knudsen, J. G. "Heat Transmission," Chemical Engineers' Handbook, 5th. ed. Section 10, R. H.
Perry and C. H. Chilton, Eds., McGraw-Hill, New York, 1973.
[14] American Society for Metals, "Properties of Aluminum Casting Alloys," Metals Handbook, 8th. ed.,
ASM, OH, 1961, p. 956.
[15] Darden, J. W., Triebel, C. A., Maes, J. P., and Van Neste, W., U.S. Patent No. 4,647,392,3 March
1987.
DISCUSSION
William P. Weeks ~ (written discussion)--In a 1990 SAE paper, Fleetguard, Inc. showed
lower aluminum weight loss for propylene glycol compared to ethylene glycol using an iden-
tical, conventional, inhibition package. Scanning your data on this new inhibition technology,
I see the same trend. Can you comment or speculate on why?
J.-P. Maes and P. Van de Ven (authors' response)--Based on these results, the results you
quote, and similar findings it appears that it is the corrosiveness of propylene glycol, or better
the glycol/water mixture, compared to monoethylene glycol that plays a role. Differences in
physico-chemical characteristics are the most likely cause of the observed phenomena.
So far, no thorough investigation to determine the cause o f these differences has been per-
formed. So any answer from the author is mere speculation. The cause could be the difference
in dissolved oxygen; propylene glycol has a lower oxygen solubility than monoethylene glycol.
The oxygen reduction reaction is the c o m m o n cathodic reaction for corrosion processes.
Reducing the oxygen a m o u n t in the solvent with a similar diffusion rate would probably
reduce the corrosion rate.
Arco Chemical Co., 3801 West Chester Pike, Newton Square, PA 19093.
DISCUSSIONON CORROSION PROTECTION 23
John Arnold 2 (written discussion)--The cyclic polarization test used to optimize monoacid/
diacid ratios were run at room temperature. Since passivation potentials change with temper-
ature and since corrosion is minimal at room temperature, are your coolants optimized for
room-temperature operation, instead of the corrosive thermal region where engines operate?
J.-P. Maes and P. Van de Ven (authors" response)--We have run tests both at room tem-
perature and high temperature (88 ~ Although the actual values of the different parameters
did change with temperature, the overall conclusions on the corrosion protection for the dif-
ferent metals did not. Therefore running the test at higher temperature does not significantly
influence the optimizing process.
Performance-test results, obtained at conditions that exist in actual engines, further prove
the efficiency of these optimized formulations. The dynamic heat transfer test discussed in the
paper is just one example of a high-temperature performance test.
John Arnold2 (written discussionJ--You stated that carboxylic acids protect aluminum by
bonding to the silicon rich dendrites. The support for this conclusion was observation of car-
bon on the dendrites by S E M / E D A X and the observation of some organic on the surface by
FTIR. Please explain how the carbon at the dendrites was determined to be due to carboxylic
acids and not due to the carbon normally present at dendrites in heat tempered aluminum
alloys? Also, explain how the organic observed by F T I R was determined to be carboxylic acid
and not residual glycol.
J.-P. Maes and P. Van de Ven (authors' response)--The proposed mechanism is only ten-
tative, as is stated in the paper. However when looking at the combination of all the results,
the proposed mechanism becomes more than just speculation:
9 Microscopic examination: dendrite structure coinciding with discolored areas.

9 SEM microprobe analysis: dendrite structure coinciding with silicon rich phases. Carbon
was not determined.
9 F T I R analysis: Organic material in conjunction with the discolored areas.
In a blank test the coupon showed no F T I R signal nor did microscopic examination indicate
any discolored areas. Although none of this is conclusive, it would be rather shortsighted to
just discard them. The authors do agree however that further investigation is necessary to give
a definite answer.
Peter Woyciesjes 2 (written discussion)--It was suggested in your presentation that current
North American coolants can have a larger effect on heat transfer properties of aluminum in
the dynamic heat transfer test as compared to Coolant A. Can the results of this test be related
to vehicle performance since there are no heat transfer problems observed in the field with
these coolants? Has it been demonstrated that this small change in heat transfer properties has
any effect on vehicle performance?
J.-P. Maes and P. Van de Ven (authors'response)--No field test data are yet available, how-
ever, the dynamic heat transfer test simulates the heat transfer conditions of the hottest areas
in an engine. We must also stress the fact that an effective heat transfer fluid is one of the most
important requirements of an engine coolant. The trend toward smaller, more efficient
engines that operate at higher temperatures cannot be denied. Our research aims at the devel-
opment of coolants for this next generation of engines that will definitely operate under more
stringent heat exchange conditions.
William Mercer 2 (written discussion)--You have stated that deionized water and not cor-
rosive water is used in many of your tests. Is deionized water also used in your electrochemical
2 First Brands Corporation, Danbury, CT 06810.

tests? If so, please explain why pitting would be expected to occur in the absence of chloride
ions.
J.-P. Maes and P. Van de Ven (authors' response)--The electrochemical test used for this
work involved the dilution of the engine coolant with corrosive hard water of the following
composition (per litre deionized water):
148 mg Na.,SO4
165 mg NaCI
138 mg NaHCO3
275 mg CaCI2
William Mercer 2 (written discussion)--Can you comment on the use of Rapid Cyclic Polar-
ization scans to measure corrosion since they are more properly used to measure capacitative
changes of surface layers?
J.-P. Maes and P. Van de Ven (authors'response)--The cyclic potentiokinetic polarization
technique has previously been described to determine the susceptibility of various metals and
materials to localized corrosion in various environments. The pitting potentials and repassi-
vation potentials, thus determined, can be directly related to the pitting tendency of the mate-
rial in the particular environment. Correlation between measurements and the occurrence of
crevice corrosion has been established. References 1 to 7 elaborate on the experimental tech-
nique and the interpretation of the experimental data.
William Mercer 2 (written discussion)--Can you explain in more detail how synergism
between the monoacid and the diacid is shown by electrochemical techniques?
J,-P. Maes and P. Van de Ven (authors' response)--The synergism between the monoacid
and diacid was not the subject of this paper. It has been established before and has been exten-
sively discussed in Refs 9 and 15. This synergism involves an improved protection against
localized corrosion phenomena as indicated by changes in pitting potentials and repassivation
potentials.
References
[ 1] Bond, A. P., "A Review of Recent Advances in Testing Methods and Interpretation," Localized Cor-
rosion-Cause of Metal Failure, STP 516, American Society for Testing and Materials, Philadelphia,
1972, pp. 250-261.
[2] Pourbaix, M., "Theoretical and Experimental Considerations in Corrosion Testing," ASTM Annual
Meeting, Toronto, Canada, Cebelcor RT. 186, Brussels, 22 June 1970.
[3] France, W. D., Jr., "Application and Limitations of Controlled Potential Corrosion Tests," Cebelcor
RT. 186, Brussels, 1970.
[4] Turnbull, A., "The Solution Composition and Electrode Potential in Pits, Crevices and Cracks," Cor-
rosion Science, Vol. 23, No. 8, 1983, pp. 833-870.
[5] Pourbaix, M., "Corrosion Localisre: Fonctionnement et Mrchanismes de Protection," Cebelcor
RT.276, Brussels, 1984.
[6] Kruger, J., "Current Ideas on the Initiation of Localized Corrosion," Cebelcor RT.279, Brussels,
1984.
[7] Pourbaix, M., "M~chanismes, Enseignement et Recherches en Corrosion Localis~e," Cebelcor
RT.280, Brussels, 1984.
Jeffrey M . B u r n s I
Fleet Test Correlations of Original Equipment

Coolant Pump Failures and Engine Coolant
Formulations
REFERENCE: Burns, J. M,, "Fleet Test Correlations of Original Equipment Coolant Pump
Failures and Engine Coolant Formulations," Engine Coolant Testing: Third Volume, ASTM
STP 1192, R. E. Beal, Ed., American Society for Testing and Materials, Philadelphia, 1993, pp.
25-43.
ABSTRACT: Automobile coolant pump failures can be minimizedby choosing an engine cool-
ant formulation that has a relatively insignificant effect on the seal as well as the other materials
used in the construction of the pumps. A fleet test involving 203 1990 Ford Crown Victoria taxi
cabs provided data demonstrating that the number of pump failures experienced by the group
of taxis employing a unique experimental coolant free of most traditional corrosion inhibitors
was much lower than the number experienced by any of the other four test groups employing
different, more traditional, formulation variations.
Sixty to eighty-five percent of the original equipment coolant pumps on the four groups of
vehicles employing the more traditional coolant formulations had to be replaced during the test
period, which nominallycovered the vehicles' first 160 000 km. Only 15% of the original coolant
pumps installed on the experimental coolant test group required replacement.
KEYWORDS: coolant pump, engine coolant formulation, fleettest, corrosion inhibitor, silicate,
carboxylic acid, long-life coolant, heavy-duty coolant, universal coolant, notched box and
whisker plot
Global competition and increasing environmental awareness are creating markets for prod-
ucts that reduce consumer maintenance and impact the environment more favorably. One
method of achieving both goals is to produce products with a longer service life. This can not
be done unless there is no adverse effect on the durability of the system components with which
the product interacts.
Recent fleet testing has yielded data comparing the compatibility of several different coolant
formulations with coolant p u m p component materials. A short-term, high-mileage taxi test,
employing four different engine coolant formulations, revealed differences between tradi-
tional, current, and potential future inhibitor technologies with respect to coolant pump life
expectancy and replacement rate.
New technology for corrosion inhibitors has yielded engine coolant formulations that are
potentially life-time fill due to the slow depletion rate of the inhibitors in service [1]. The data
derived from this fleet testing indicate that there is also improved compatibility with the cool-
ant pump. This technology is a step toward the goal of a maintenance-free system for auto-
motive service, which will also reduce the environmental impact by generating less waste.
Senior engineer, Texaco Chemical Company, Austin Research Laboratories, 7114 N. Lamar Blvd.,
Austin, TX 78752
25
Background
A fleet test, involving four groups of 20 1990 Ford Crown Victoria taxis, was established in
the fall of 1989 to develop inhibitor depletion data, corrosion rates from coupon bundle
results, and general applicability information for four different coolant formulations. These
cars were equipped with the Ford 5.0 liter V8, which has a cast iron block and combustion
chamber heads. The cooling systems employed brass heavy-duty radiators with the exception
of a group of 24 cars, which had experimental aluminum radiators installed as part of a com-
ponent evaluation program. The vehicle identification numbers in the test groups were not
sequential and were randomly distributed throughout the fleet.
The taxi company also operates 123 new vehicles and 12 older model standby vehicles that
were not included in the coolant test or sample evaluation program. These vehicles retained
the factory-fill coolant and were monitored as a representative population. The data from
these cars are presented for comparison as being representative of the situation experienced in
the field. After several months, a high number of coolant pump replacements were required
throughout the entire fleet.
This fleet is an ideal test to track part failure data because the city of New York
requires that every taxi must be inspected once a quarter, and the fleet maintenance supervisor
maintains excellent maintenance records. The initial signs of component failure are noted
early.
The fleet has long-time experience with lubricant evaluation, and recently added engine
coolant to their evaluation program. The maintenance personnel have demonstrated excep-
tional ability to follow proper procedures and to generate extremely reliable data.
The cabs were filled with five different coolant formulations which included the factory-fill
coolant and a commercial, traditional, high silicate, American formulation, commercial Euro-
pean and American universal formulations, and an experimental long-life coolant based on
patented inhibitor technology [2]. The commercial formulations were all obtained through
commercial outlets while the experimental coolant was blended at the Texaco Chemical Com-
pany's Austin, TX, research laboratories.
Because of the logistics of fleet testing, each coolant used was a different color except for the
factory-fill coolant, which was also green. This was the commercial circumstance at the time
the fleet test was initiated. The colors will be used to identify the individual products while the
factory-fill coolant will be identified as FF.
Yellow and Factory-Fill (FF)

The high-silicate formulation tested is the single most popular formulation sold in the
United States. It will be identified as the yellow test coolant. It includes both phosphate and
borax as the buffer system and contains nitrate. The factory-fill coolant was very~similar. It will
be identified as the factory-fill (FF) test coolant. These coolants are representative of the tra-
ditional technology intended for automotive service where aluminum protection is required
and are the types of coolants employed as factory-fill for light-duty engines by all three major
American manufacturers.
These coolants meet the ASTM Specification for Ethylene Glycol Base Engine Coolant for
Automobile and Light Duty Service (D 3306) and most of the automobile original equipment
manufacturers' (OEMs) specifications [3]. The factory-fill systems were topped-off with a
c o m m o n aftermarket formulation corresponding to the General Motors GM6038-M formu-
lation [4].
BURNS ON FLEET TEST CORRELATIONS 27
Purple and Green

The universal formulations meet the corrosion inhibition performance specifications for
both light- and heavy-duty cooling systems and provide adequate a l u m i n u m protection while
meeting the low silicate definition, having 0.1 wt% sodium metasilicate (250 ppm as silicon)
or less as specified in the ASTM D 4985. The low-silicate level is the common factor that char-
acterizes all of the common so-called universal formulations. The European universal coolant
is a borax buffered formulation incorporating some carboxylic acid technology. This coolant
will be identified as the purple fluid. The American universal product is a phosphate buffered
formulation. It will be identified as the green test coolant.
Blue
The long-life coolant tested incorporates a combination of monobasic and dibasic carbox-
ylic acids to produce a product with a longer service life due to slow inhibitor depletion rate
[ 1]. This product will be identified as the blue test coolant. It contains no silicate, borax, phos-
phate, amines, nitrite, or nitrate. Like all of the coolant formulations, it contained tolyltria-
zole. The constituents in each of these six formulations are detailed in Table 1.
Test Procedure
The test coolants were installed in the taxi cooling systems as soon as practical for the fleet
after the new automobiles arrived at the fleet garage. The cooling system was drained, flushed
with tap water, and filled with the test coolant prediluted to 50% with deionized water. Startup
mileage in the test vehicle groups ranged from 0 km on all of the factory-fill vehicles to an
average of over 9750 km for the purple coolant test group.
Stop-leak additive, typically carboxymethyl cellulose, is added to the automotive cooling
system at the factory to plug minor gasket leaks. It is possible, but not likely, that most of this
additive was flushed out when the cooling systems were drained during the conversion of the
four test groups to the test coolants. However, it was retained in the factory-fill coolant cooling
systems that remained intact throughout the test period. No stop-leak additive was added to
the coolant test group vehicle systems when they were charged with the test coolants.
TABLE 1--Engine coolant inhibitor comparison for the test coolant formulations, the factory-fill
coolant, and the top-off coolant for the factory-fill systems.
Inhibitor FF Yellow Green Purple Blue GM6038-M
Borax yes yes no yes no yes

Phosphate yes yes yes no no yes
Nitrate yes yes yes yes no yes
Nitrite no no no no no no
Silicate Level high high low low none low
(Low = <250 ppm
High -- >250 ppm)
Tolytriazole yes yes yes yes yes yes
Carboxylic acids no no no yes yes no
Amines no no no no no no
Silicate stabilizer yes yes yes yes no yes
Antifoamant yes yes yes yes yes yes
Initial pH > 10 > 10 > 10 <9 <9 > 10
The vehicles were maintained according to standard operating procedure by the taxi com-
pany. The overflow bottles were topped-off, with the appropriate coolant prediluted 50% with
deionized water, by the maintenance personnel. The concentration was monitored by mea-
suring the freezing point with a hand held refractometer and adjusted, if necessary, to - 37~
by adding either deionized water or coolant concentrate, which was the standard procedure in
this fleet.
The factory-fill cars were topped-off with a coolant corresponding to the GM6038-M [4]
formulation specification as was their practice before initiating this fleet test. Although the
GM6038-M is not approved for the service fill for these systems, it is inexpensive, and is com-
monly used by this type of fleet. It was the coolant purchased for topping-offby this fleet before
the inception of this test. GM6038-M is a low-silicate, phosphate, borax, heavy-duty coolant
that meets ASTM D 4985, but does not meet ASTM D 3306 because o f poor performance in
ASTM Test Method for Corrosion of Cast A l u m i n u m Alloys in Engine Coolants Under Heat-
Rejecting Conditions (D 4340) [3].
Coolant samples were taken at every oil change, which were at approximately 3000- to 5000-
mile (4828- to 8047-km) intervals. One hundred millilitres of sample was captured each time.
Full chemical analysis of the majority of the samples, with respect to inhibitor and contami-
nant concentration, was undertaken to generate inhibitor depletion rate data and monitor the
quality of the cooling system in each vehicle.
The results of the analysis also served as an indication of the care being taken in maintaining
the cabs.
The taxis are subject to a quarterly inspection by the city of New York Limousine and Taxi
Commission. A m o n g other things, each car is inspected for fluid leaks. Any fluid dripping
from the bottom of the engine compartment constitutes a failure. The cab company conducts
a thorough preinspection to reduce potential of the cabs not passing the city inspection. Com-
ponents that would jeopardize the audit are repaired or replaced. If the coolant pump is leaking
from the weep hole, it is subjected to pressure and dynamic testing. It is replaced if the leak is
steady and noticeable.
Results
Coolant p u m p replacements throughout the test fleet provided the results that are the basis
for this paper. Three of the test coolants, the yellow, purple, and green coolants, produced
roughly the same results. Seventy-five to eighty-five percent of the original equipment coolant
pumps on the 60 cabs filled with these three test coolants were replaced during the test period
due to leakage. The factory-fill control group experienced a 60% failure rate. The blue coolant
test group, however, required only 15% of the pumps to be replaced.
The pumps that did not fail either survived until the end of the test or were removed with
the engine during an overhaul procedure that was part of lubricant testing that this fleet is also
used for. The number of pumps that survived to the end of the test or were removed with the
engines that were overhauled reveals as much about the effect of the coolants on the pump as
the failures do.
Nine of the original equipment pumps (45%) installed on the blue coolant test cars were
retired on the cabs at the end of test. Twenty-two of the original pumps (18%) on the factory-
fill cabs reached the end of test. Three of the purple coolant cabs' original pumps (15%) sur-
vived. There were two surviving pumps (10%) among the green coolant test group. None o f
the original pumps (0%) survived the testing with the yellow coolant.
The pumps that were removed at overhaul were on engines involved in a lubricant evalua-
tion program. The engines were removed for wear and deposit measurement. The coolant
pumps were replaced with new, original equipment, pumps during the rebuilding process. The
engines that were chosen for removal were picked because of the oil formulation tested, for
reasons completely unrelated to the coolant test. It was not possible to predict which engines
would be chosen for inspection because the lubricant formulations being tested were randomly
distributed throughout the fleet.
An average of 9 (45%) engines were removed from each test of the four coolant test groups
while 50 (41%) of the 123 engines in the factory-fill test group were removed. The only engine
overhauls that matter to the data presented here are those on which original equipment pumps
were still operational. Test groups with a large number of early failures had few pumps that
survived to the point at which the engine was removed from service whereas test groups that
provided longer service life for the p u m p lost more of the original equipment pumps at this
point.
The number of original pumps removed with these engines was dependent on the number
of early life pump failures. The blue coolant that had few failures had 8 of the original pumps
removed at overhaul (40%). The other 4, more traditional coolant formulation test groups had
a smaller percentage that did not complete the test for this reason. The yellow coolant had 3
pumps (15%) removed in this manner. The purple coolant lost 2 pumps (10%) to this process.
Only 1 of the original pumps (5%) was removed from the green coolant test group because of
the lubricant testing. Because the n u m b e r of factory-fill coolant vehicles was greater, 27 of the
original pumps were removed for the engine inspection process, but the overall percentage
(22%) was not significantly different than that of the other three traditional formulations.
Table 2 details n u m b e r and percentage of the original pumps that failed, were removed with
engines intended for overhaul, or survived the test period.
Removal of engines for this overhaul and inspection process reduced the failure data that
was available by removing pumps from the test before they had a chance to fail. Most of the
engines were removed between 96 500 and 112 500 km. This event could not be avoided
because lubricant testing is this fleet's primary function.
Tables 3 through 7 list the original pump failures in each of the test groups, the pumps that
survived, and the pumps that were removed at the engine overhaul. The final disposition of
each original pump can be obtained here.
The number of repeat failures in each group was also quite different. With the blue and green
coolant groups none of the new pumps that replaced the failed original equipment failed again.
The purple coolant group had 3 repeat failures. The factory-fill cars had 11 repeat failures out
of the 74 cars that had their original pumps replaced, while the yellow coolant cars had 8 repeat
TABLE 2--Original equipment water pump dispos#ion for the 5 coolant test groups.
Original Pumps Removed at

Coolant Original Pumps Overhaulfl % of Total, Total Original Pumps
Formulation, that Failed Number of Engines that Survived
Total Number During Test, % Removed for Overhaul from Cab Life, % of
of Pumps of Total Group Total
FF, 123 74, 60% 27, 22%, 50 22, 18%

Yellow, 20 17, 85% 3, 15%, 9 0, 0%
Green, 20 17, 80% 1, 5%, 8 2, 10%
Purple, 20 15, 75% 2, 10%, 10 3, 15%
Blue, 20 3, 15% 8, 40%, 9 9, 45%
aThe number of original equipment pumps removed at overhaul depended entirely on the number of
original equipment pumps still operating at that mileage. The test groups with earlier failures lost fewer
pumps for this reason. Each test group had approximately the same number of engines remover for
overhaul at approximately the same mileages.
TABLE 3--Factory-fill coolant test group original equipment coolant pump final disposition
information.
Car Actual Test Coolant Mileage at Date at
Number Mileage, km Mileage, km Used Fill, km Failure
FACTORY-FILL CABS THAT HAD FAILURESOF THE ORIGINAL COOLANT PUMP DURING TEST PERIOD
766 20 246 20 246 FF 0 10 March 90
744 36 638 36 638 FF 0 25 April 90
761 40 607 40 607 FF 0 25 April 90
757 43 589 43 589 FF 0 29 April 90
743 47 431 47 431 FF 0 16 May 90
706 50 366 50 366 FF 0 22 May 90
889 52 833 52 833 FF 0 28 April 90
756 54 648 54 648 FF 0 16 June 90
800 56 303 56 303 FF 0 08 July 90
714 56 910 56 910 FF 0 21 May 90
801 57 870 57 870 FF 0 15 July 90
707 58 484 58 484 FF 0 27 May 90
948 60 025 60 025 FF 0 25 May 90
910 60 288 60 288 FF 0 26 May 90
758 61 168 61 168 FF 0 11 July 90
764 62 179 62 179 FF 0 18 July 90
767 62 433 62 433 FF 0 05 July 90
797 62 486 62 486 FF 0 03 Aug. 90
900 63 093 63 093 FF 0 08 July 90
763 64 010 64 010 FF 0 07 July 90
941 64 290 64 290 FF 0 04 May 90
932 64 617 64 617 FF 0 04 May 90
723 65 272 65 272 FF 0 02 July 90
746 65 748 65 748 FF 0 03 July 90
795 66 427 66 427 FF 0 19 July 90
936 68 415 68 415 FF 0 22 July 90
894 68 674 68 674 FF 0 03 July 90
739 70 330 70 330 FF 0 14 July 90
796 70 517 70 517 FF 0 21 Aug. 90
919 70 608 70 608 FF 0 22 July 90
909 70 758 70 758 FF 0 14 July 90
768 71 065 71 065 FF 0 06 Aug. 90
754 71 186 71 186 FF 0 31 July 90
742 72 150 72 150 FF 0 09 July 90
741 72 429 72 429 FF 0 15 July 90
748 74 255 74 255 FF 0 27 July 90
762 74 513 74 513 FF 0 30 Aug. 90
805 75 061 75 062 FF 0 07 Sept. 90
898 75 171 75 171 FF 0 06 Nov. 90
802 76 416 76 416 FF 0 03 Sept. 90
794 77 091 77 091 FF 0 30 Aug. 90
916 77 165 77 165 FF 0 06 Aug. 90
755 77 857 77 857 FF 0 25 Aug. 90
905 77 870 77 870 FF 0 27 Aug. 90
899 78 489 78 489 FF 0 16 Oct, 90
753 78 557 78 557 FF 0 18 Aug. 90
938 78 909 78 909 FF 0 22 June 90
912 79 717 79 717 FF 0 16 Nov. 90
738 80 498 80 498 FF 0 16 Aug, 90
897 81 727 81 727 FF 0 02 Sept. 90
765 86 953 86 953 FF 0 03 Oct. 90
720 87 616 20 418 FF/green 67 198 08 Sept. 90
737 90 249 90 249 FF 0 14 Nov. 90
TABLE 3--Continued.

FACTORY-FILL CABS THAT HAD FAILURES OF THE ORIGINAL COOLANT PUMP DURING TEST PERIOD
915 90 778 90 778 FF 0 20 Sept. 90
745 91 781 91 781 FF 0 29 Sept. 90
914 93 511 93 511 FF 0 05 Aug. 90
747 96 108 96 108 FF 0 30 Sept. 90
901 96 232 96 232 FF 0 08 Nov. 90
906 98 550 98 550 FF 0 05 Sept. 90
749 98 585 98 585 FF 0 02 Sept. 90
917 101 175 101 175 FF 0 18Oct. 90
946 102 327 102 327 FF 0 25 Sept. 90
799 102 382 102 382 FF 0 07Nov. 90
911 103 177 103 177 FF 0 18Nov. 90
769 103 817 103 817 FF 0 29Oct. 90
939 103 894 103 894 FF 0 24 Sept. 90
760 109 207 109 207 FF 0 20Nov. 90
903 113 859 113 859 FF 0 26 Aug. 90
933 119 500 119 500 FF 0 24 Sept. 90
895 125 695 125 695 FF 0 18 Sept. 90
943 133 793 133 793 FF 0 09 Jan. 91
803 139 945 139 945 FF 0 14Feb. 91
719 152 831 80 813 FF/green 72 018 17Feb. 91
710 170 608 95 083 FF/green 75 525 0 6 M a r c h 91
Average 77 t74 76 241 933
FACTORY-FILL CABS THAT HAD ORIGINAL PUMPS THAT SURVIVED TEST DURATION
940 79 226 79 226 FF 0 27 Feb. 91
907 106 131 106 131 FF 0 01 April 91
751 111 650 111 650 FF 0 24 March 91
949 117 799 117 799 FF 0 26 March 91
703 128 421 128 421 FF/blue 0/54 492 25 Feb. 91
750 131 944 131 944 FF 0 24 March 91
893 132 269 132 269 FF 0 27 Feb. 91
704 135 100 135 100 FF/blue 0/54 713 25 Feb. 91
942 138 965 138 965 FF 0 06 Feb. 91
711 141 251 141 251 FF/green 0/71 939 05 March 91
944 141 854 141 854 FF 0 06 Feb. 91
937 143 597 143 597 FF 0 05 Feb. 91
721 146 413 146 413 FF/green 0/70 443 12 March 91
717 146 611 146 611 FF/green 0/66 300 11 March 91
701 149 601 149 601 FF/blue 0/68 784 25 Feb. 91
798 149 748 149 748 FF 0 15 April 91
715 154 231 154 231 FF/green 0/65 874 11 March 91
712 155 569 155 569 FF/green 0/71 646 06 March 91
913 155 801 155 801 FF 0 14 Feb. 91
708 168 436 168 436 FF/blue 0/75 810 27 Feb. 91
888 176 411 176 411 FF 0 08 Feb. 91
934 194 814 194 814 FF 0 21 Feb. 91
Average 141 175
FACTORY-FILL CABS THAT HAD THE ORIGINAL PUMP REMOVED WITH THE ENGINE AT OVERHAUL
702 105 404 105 404 FF/blue 0/64 090 24 Oct. 90
705 108 687 108 687 FF/blue 0/54 480 04 Dec. 90
709 97 214 97 214 FF/green 0/67 100 01 Oct. 90
713 96 625 96 625 FF/green 0/76 468 31 Aug. 90
716 105 443 105 443 FF/green 0/69 023 03 Nov. 90
718 96 553 96 553 FF/green 0/72 960 12 Sept. 90
722 102 480 102 480 FF/green 0/67 240 17 Oct. 90
TABLE 3--Continued.

FACTORY-FILL CABS THAT HAD THE ORIGINAL PUMP REMOVED WITH THE ENGINE AT OVERHAUL
724 108 092 108 092 FF/green 0/64828 07 Nov. 90
725 97 510 97 510 FF 0 27 Sept. 90
726 101 178 101 178 FF 0 26 Sept. 90
740 98 690 98 690 FF 0 28 Sept. 90
752 108 088 108 088 FF 0 05 Nov. 90
759 112 240 112 240 FF 0 09 Nov. 90
804 123 491 123 491 FF 0 04 Feb. 91
806 107 767 107 767 FF 0 03 Dec. 90
807 103 622 103 622 FF 0 26 Dec. 90
890 96 577 96 577 FF 0 14 Sept. 90
891 99 268 99 268 FF 0 03 Aug. 90
892 96 960 96 960 FF 0 13 Aug. 90
896 96 487 96 487 FF 0 20 Aug. 90
902 101 704 101 704 FF 0 15 Nov. 90
904 99 868 99 868 FF 0 05 Oct. 90
908 96 789 96 789 FF 0 04 Sept. 90
918 97 074 97 074 FF 0 17 Sept. 90
935 101 983 101 983 FF 0 01 Aug. 90
945 98 434 98 434 FF 0 30 July 90
947 98 823 98 823 FF 0 31 July 90
Average 102 113 . . . . . . . . . . . .
TABLE 4--Yellow coolant test group original equipment coolant pump final disposition information.

YELLOW COOLANT CABS THAT HAD ORIGINAL COOLANT PUMP FAILURESDUR1NG THE TEST
PERIOD
885 25 365 25 365 yellow 0 19 May 90
879 43 530 43 530 yellow 0 27 June 90
886 49 378 49 378 yellow 0 25 July 90
874 52 159 52 159 yellow 0 05 Aug. 90
869 55 350 55 350 yellow 0 02 Aug. 90
882 58 068 56 327 yellow 1741 20 Aug. 90
870 59 261 59 261 yellow 0 19 Aug. 90
871 60 442 60 442 yellow 0 06 Sept. 90
875 66 168 66 168 yellow 0 12 Sept. 90
876 67 279 67 279 yellow 0 11 Sept. 90
877 76 940 76 940 yellow 0 12 Oct. 90
880 79 804 79 804 yellow 0 30 Oct. 90
872 90 622 90 622 yellow 0 10 Feb. 91
868 91 593 91 593 yellow 0 08 Nov. 90
873 120 026 120 026 yellow 0 26 March 91
883 130 582 130 582 yellow 0 24 April 91
887 138 516 138 516 Y to G 0 31 May 91
Average 74 417 74 314 ... 102
YELLOW COOLANT CABS THAT HAD ORIGINAL PUMPS THAT SURVIVEDTEST DURATION
NO PUMPS IN CABS WITH YELLOW COOLANT SURVIVED THE TEST
YELLOW COOLANT CABS THAT HAD THE ORIGINAL PUMP REMOVED WITH THE ENGINE AT
OVERHAUL
878 118 652 118 652 yellow 0 18 Feb. 91
881 124 132 122 011 yellow 2121 28 Feb. 91
884 122 371 122 371 yellow 0 12 March 91
Average 121 718
TABLE 5--Green coolant test group original equipment coolant pump final disposition information.

GREEN COOLANTCABSTHAT HAD ORIGINALCOOLANTPUMP FAILURESDURINGTHE TEST PERIOD

831 45 580 25 717 green 4 192 05 July 90
844 54 805 29 560 green 7 232 20 Aug. 90
829 55 508 29 438 green 8 132 22 July 90
838 56 414 29 827 green 8 412 31 July 90
840 73 153 39 283 green 9 933 10 Sept. 90
846 73 158 41 404 green 6 524 09 Oct. 90
841 76 091 42 767 green 7 265 07 Oct. 90
837 88 593 48 408 green 10 688 05 Nov. 90
833 91 380 49 713 green 11 375 02 Nov. 90
843 93 913 53 889 green 7 187 24 Nov. 90
835 10l 134 57 629 green 8 390 20 Dec. 90
834 102 254 57 153 green 10 276 23 Nov. 90
830 102 924 59 304 green 7 483 23 Nov. 90
826 103 838 58 339 green 9 951 22 Jan. 91
832 121 723 69 406 green 10 025 03 Feb. 91
847 122 424 70 569 green 8 855 23 Jan. 91
845 142 112 83 527 green 7 688 14 April 91
Average 93 641 52 710 8 812
GREEN COOLANTCAnS THAT HAD ORIGINALPUMeSWHAT SURVIVEDTEST DURATION
839 153 387 143 510 green 9 877 08 May 91
Average 95 310 ...
GREEN COOLANT CABS WHICH HAD THE ORIGINAL PUMP REMOVED WITH THE ENGINE AT
OVERHAUL
828 107 842 97 612 green 10 231 07 Dec. 90
842 116 999 107 995 green 9 004 28 Jan. 91
Average 112 421 . . . . . . . . . . . .
failures among the 17 cabs that had the original pump replaced once. Both of these two groups,
the factory-fill and the yellow, had replacement pumps that also subsequently failed.
Figure 1 is a multiple notched box-and-whisker plot [5] of all five groups of vehicles dis-
cussed in this report. It is a graphical comparison of the data generated by each group of vehi-
cles. The mileage of the original equipment coolant pumps that failed during the test and that
survived the entire test period were included in this figure. The mileage of the pumps removed
because of the overhaul and inspection process were not.
This plot is a modification of the standard box-and-whisker plot. A notch is added to each
box corresponding to the width of a confidence interval for the median, while the width of the
box is proportional to the square root o f the number of observations in the data set. The con-
fidence level on the notches is set to allow pair-wise comparisons to be performed at the 95%
level by examining whether two notches overlap [5].
The whiskers are drawn to include data points that fall outside of the box but fall within a
distance one and one half times its total length. Points not included in the whiskers are drawn
as dots on the figure and can be considered outliers that are not representative of the popula-
tion. The pumps that survived the test in the purple and green cars can be considered flyers
while the early failures in the blue coolant test group fall into this category.
This plot clearly demonstrates as shown in Table 8 that the blue coolant provides a much
longer service life for the coolant pump than any of the other four formulations.
Table 9 compares the numerical average mileage at startup and at failure for each of the
groups. The average listed in this table is the numerical average of mileage at which the original
TABLE 6--Purple coolant test group original equipment coolant pump final disposition information.

PURPLE COOLANTCABSTHAT HAD ORIGINALCOOLANTPUMP FAILURESDURINGTHE TEST

PERIOD
816 65 813 55 336 purple 10 477 21 July 90
808 71 350 59 005 purple 12 345 01 Aug. 90
813 72 412 61 089 purple 11 323 10 Aug. 90
827 78 575 69 168 purple 9 407 29 Sept. 90
810 79 595 62 848 purple 16 747 03 Sept. 90
812 84 495 73 368 purple 11 127 02 Oct. 90
819 88 054 77 282 purple 10 771 20 Nov. 90
818 91 361 80 926 purple 10 435 16 Oct. 90
809 97 116 85 529 purple 11 587 01 Nov. 90
823 97 592 89 703 purple 7 889 22 Nov. 90
825 98 091 91 002 purple 7 089 20 Nov. 90
814 98 239 86 588 purple 11 652 14 Sept. 90
817 102 750 93 313 purple 9 437 21 Nov. 90
811 104 086 92 652 purple 11 434 29 Nov. 90
826 113 365 103 067 purple 10 298 21 Dec. 90
Average 89 526 78 725 10 801
PURPLE COOLANT CABS WHICH HAD ORIGINAL PUMPS THAT SURVIVED TEST DURATION
815 159 066 148 000 purple 11 066 02 May 91
821 156 522 146 394 purple 10 128 03 May 91
824 154 946 147 640 purple 7 306 03 May 91
Average 156 845 ...
PURPLE COOLANT CABS WHICH HAD THE ORIGINAL PUMP REMOVED WITH THE ENGINE AT
OVERHAUL
820 74 218 59 702 purple 14 516 04 Jan. 91
822 66 500 60 287 purple 6 213 13 Dec. 90
Average 70 359 . . . . . . . . . . . .
pumps failed. It does not take into account pumps that are still in service. Again, the fact that
there were only 3 occurrences out of a possible 20 make calculated numerical averages for the
blue coolant test group statistically meaningless. This is why utilizing statistical techniques,
such as the notched box and whisker plot is crucial to understanding the data.
The data on which this report is based can be obtained by writing to the author.
Discussion
This data represent the distribution of original equipment coolant pump failures that would
be expected over the service life of vehicle populations employing the coolants tested. This
discussion does not attempt to address the failures that occur in the first 12 months/24 000 km
of service that have been the topics of other papers [6, 7]. The failures experienced by this fleet
appear to be directly related to coolant formulation by incidence and by distribution of
occurrence.
There is currently no theoretical mechanism offered to interpret these failures. The pumps
that failed were the coolant pumps present in the cars as delivered from the factory. No deter-
mination of their general quality or consistency could be made. It is assumed that this popu-
lation represented a normal distribution of the total population of similar pumps. Information
concerning the c o m m o n causes of pump failure in vehicle populations is available through the
TABLE 7--Blue coolant test group original equipment coolant pump final disposition information.

BLUE COOLANT CABS THAT HAD ORIGINAL COOLANT PUMP FAILURES DURING THE TEST PERIOD
867 29 548 22 957 blue 6 590 16 May 90
865 86 193 79 603 blue 6 590 20 Oct. 90
857 132 869 131 443 blue 1 426 11 May 91
Average 82 870 78 001 4 869
BLUE COOLANT CABS THAT HAD ORIGINAL PUMPS SURVIVEDTEST DURATION
849 156 243 152 641 blue 3 602 14 May 91
852 147 372 140 092 blue 7 281 14 May 91
853 146 211 139 026 blue 7 184 14 May 91
854 138 780 135 581 blue 3 199 15 May 91
856 149 786 143 166 blue 6 621 16 May 91
860 157 207 150 870 blue 6 338 17 May 91
861 154 980 150 863 blue 4 117 21 May 91
862 155 302 150 469 blue 4 833 21 May 91
866 140 029 133 516 blue 6 513 21 May 91
Average 149 546
BLUE COOLANT CABS WHICH HAD THE ORIGINAL PUMP REMOVED WITH THE ENGINE AT
OVERHAUL
848 115120 107 440 blue 7 680 13 Dec. 90
850 118 116 111083 blue 7 033 21 Jan. 91
851 111 125 104 192 blue 6 933 21 Dec. 90
855 121 699 115 478 blue 6 220 29 Jan. 91
858 118 882 112 514 blue 6 368 06 March 91
859 114 469 106 683 blue 7 786 25 Jan. 91
863 113 385 106 421 blue 6 964 27 Dec. 90
864 120 324 113 515 blue 6 809 14 Feb. 91
Average 116 640 ...
TABLE 8--Average pump life expectancy and standard

deviation by notched box and whisker plot.
Size of 95%
Confidence
Service Life Interval Notch, Failures Per
Coolant Mean, km km Total a
Factory-fill 77 250 17 700 74/96

Yellow 67 600 24 000 17/17
Green 91 750 20 900 17/18
Purple 98 200 29 000 15/18
Blue 148 000 22 500 3/12
"The total number of pumps is the sum of the original equip-

ment pumps that failed during the test period and the original
equipment pumps that survived to the end of the test. The original
equipment pumps that were removed as part of the lubricant eval-
uation program are not included or considered in the service life
evaluations.
[ I I i I I I
24.1 . . . . . . . .
"7.
19.3
"o
"~ 14.5
o. 9.6
o-
I.
= 4.8
0
E
0.0
I B I l r I
Coolant Test Group

FIG. 1--Notched box and whisker plot comparing coolant pump service life distribution of the 5 coolant
test groups.
seal manufacturer [8]. The coolant pumps used to replace the original pumps were obtained
from an original equipment parts dealer and were not remanufactured.
The chemical differences between the coolant formulations are cited as one possible expla-
nation of the failure distribution. All of the test coolants were ethylene glycol based and
blended roughly 50% with deionized water. The factory-fill formulation was mixed according
TABLE 9--Average kilometres at failure of the original

equipment pumps in each test group.
Average
Kilometres at Failure
Average Start-up of Occurrences Per
Kilometres Failures/ Total Pump
Coolant at Failure Whole Group Population
Factory-fill 77 174 933/0 74/123
Yellow 74 416 102/193 17/20
Green 88 530 8 447/8 636 17/20
Purple 89 526 10 802/10 547 15/20
Blue 82 870 4 868/6 030 3/20
to assembly plant practices by the manufacturer with water of unknown quality. The average
water concentration was found to be approximately 47% by weight.
The factory-fill coolant in these vehicles was quite similar to the yellow coolant, and it may
have been made by the same manufacturer. The difference in failure distribution between
these two groups may be attributable to the stop-leak additives used by the auto manufacturer
at initial fill, or it could be attributable to the fact that these cars were topped-offwith a different
formulation, GM6038-M.
The stop-leak additives are used by the automobile manufacturer to reduce warranty prob-
lems caused by variable head bolt and gasket sealing quality throughout the engine. They are
added to the cooling systems of most domestically produced cars. The coolant pump seal man-
ufacturer believes that the stop-leak additives help rather than hinder the service life of the seal
[9]. The additives appear to aid in the reduction of filming and wear, but it is not understood
how. It is possible, but not likely that most of these stop-leak additives originally in the 80 test
vehicles employing the 4 test coolants were flushed out when the cars were converted to the
test coolants.
Another phenomenon may be responsible for the few longer life pumps that do not fit the
distribution predicted in the notched box and whisker plot o f the failure occurrences of the
majority of the units in the factory-fill, yellow, green, and purple coolant test groups. One of
the cars filled with the yellow coolant that had the original pump, cab 887, was flushed out and
filled with fresh green coolant at 130 672 km. The pump was found to be leaking at the next
inspection, 7844 km later.
In the field, it is c o m m o n for coolant p u m p seals to leak soon after the coolant is changed,
especially if the change interval was longer than recommended because of neglect. The cause
of this behavior is not known, but one explanation is that the higher pH of the fresh coolant
redissolves precipitates that have formed in the system as the alkalinity decreased.
Solids can also be formed by inhibitor drop-out, most notably silicate, products of corrosion,
and hard water ion salts. The coolant change may simply remove solids by the mechanical
action of the flush. The precipitates or solids may have acted as a stop-leak for the coolant
p u m p seal preventing excessive amounts of coolant from escaping through leak paths on an
already damaged seal surface.
The universal coolant formulations, with lower additive levels, may not be as prone to this
phenomenon. Additionally, the purple and blue coolants have a much lower initial p H than
traditional American coolant formulations. Addition of a sequestering agent to the formula-
tions employing the more traditional inhibitor technology may reduce this effect, but it would
predict a reduced average durability of the p u m p according to this mechanism.
A group of 24 cabs, numbered 701 to 724, which originally were part of the factory-fill test
group, were switched over to two of the test coolants for a component test (Table 3). The aver-
age mileage at this conversion was approximately 66 000 km for all 24 cabs. Eight of the cars,
cabs 701 to 708, were switched over to the blue test coolant. The other 16 cabs were filled with
the green test coolant. The top-off coolant was not used because it does not meet the original
equipment manufacturer's service fill requirements.
Handling these cars posed some difficulties since they do not fit the profile of the fleet at
large. The majority of the failures experienced by this group o f cars occurred before the con-
version from factory-fill to the test coolants used for the radiator test.
The cabs that were switched to the blue coolant experienced two failures among the eight
units before the switch occurred. It should be noted that none of the coolant pumps in the
group of cabs numbered 701 to 708 (Table 10) failed after they were flushed and refilled with
the blue coolant.
The 16 cabs that were flushed and replaced with the green test coolant saw 2 failures before
the switch and 3 additional failures after the switch. The failure rate among the 24 cars
TABLE lO--Radiator test group failures.
Cab Numbers/ Average Failures Before Failures After

Coolants/Number of Kilometres at Change~/Average Changeb/Average
Cabs Flush and Fill Kilometres Kilometres
701 to 708/factoiy-fill 63 355 2/54 425 0/does not apply

to blue/8 cabs
709 to 724/factory-fill 69 907 2/61 091 3/137 018
to green/16 cabs
a The failures before the change over are included in the factory-fill group (see Table 3).
b Only the failures of original equipment pumps were considered.
involved in this radiator test was lower than in the general taxi population. The percentage of
failures was considerably lower than any o f the other test groups except for those cabs filled
with the blue coolant initially.
The data from these two test groups are included with the rest of the factory-fill cabs for
purposes of calculating durability of the factory-fill test group. The average results for the fac-
tory-fill vehicles are, consequently, probably somewhat better than would normally be
expected because a lower percentage from this group failed at a higher average mileage. The
two highest mileage failures reported as factory-fill cabs were from this group of vehicles. If
these vehicles were omitted from the data presented for the factory-fill test group, the average
life expectancy of the coolant p u m p would decrease slightly.
These data may indicate that proper maintenance of the cooling system can extend the life
of the cooling system or that replacing the factory-fill coolants, which are typically of the high
silicate type, with low-silicate universal coolants improves the chances of pump survival. The
three failures that occurred in the group of cabs that were switched to the green coolant, num-
bers 709 to 724, occurred at mileage nearly twice the average for the rest of the factory-fill
group.
Experimental Procedure
From Table 1, the only c o m m o n differences that could explain the distribution of pump
failures are that the blue coolant lacks nitrate, silicate, and silicate stabilizer, whereas all of the
other coolants tested contained these three additives.
A test of silicate, nitrate, and a silicate stabilizer with exposed seal halves was performed to
evaluate inhibitor effect on the seal material. The seals used for this evaluation were not the
same seals that are used in the Crown Victoria coolant pump. They are composed of the same
phenolic material, but incorporate a bronze ring into the sealing surface and are of the design
used in many General Motors coolant pumps. These seals were used because they were easily
accessible and because no supply of the correct seals could be obtained.
The individual halves of the coolant p u m p seal were immersed in a 50% solution of coolant
mixed with deionized water and placed in an oven at 100~ The phenolic portion of the cool-
ant p u m p seal, the stationary side, caused colorless coolant, formulated without dye, to turn
yellow-brown in a hot soak test.
This phenomenon only occurred when the seal halves were soaked in the formulations
based on traditional inhibitor technology, not the experimental, long-life coolant formula-
tions. When the individual inhibitors were tested, only the silicate, and, to a lesser degree, the
silicate stabilizer, reproduced the results with respect to color hue and intensity.
The solution containing the nitrate changed color slightly, but it did not resemble the results
produced by soaking the seal in the colorless coolant solution. The alkalinity of the solution
may also come into play [7], but the results produced by the purple coolant, which has the
lowest initial pH, were the same as that produced by the higher p H coolant formulations in
about the same amount o f time in the oven.
The nature o f this change is not known. It is possible that there is an interaction between
the silicate inhibitor, silicate stabilizer, or silicon in solution and the phenolic surface. It is also
not known if this potential interaction is related to the failures experienced by this fleet, but it
is one possible explanation. Reactions involving silicon that can take place in engine cooling
systems are discussed by Kiryu et al. [ 7]. If silicate is involved in the failure mechanism, a
linear relationship between the number of failures in the coolant test group and the relative
silicate level of the traditional inhibitor formulations can not be shown from these data.
Conclusions
The goal of the original equipment manufacturers is to produce products with reduced
maintenance, improved performance, and lower cost. Consumers' quality perception is based
on the durability of and the amount of inconvenience encountered with products.
Because of the number of factors that affect the coolant pump, failures can not be elimi-
nated, but their frequency can be reduced by employing alternative coolant formulations.
Engine coolant formulations should provide corrosion inhibition for the metals used in engine
construction, but must also be inert to the other materials.
Although the mechanism of failure is unknown at this time, the data presented clearly indi-
cate that the blue coolant can provide a much longer useful life for the coolant pump. It is not
believed that the blue coolant is beneficial for the pump, but rather that the adverse effect on
the p u m p component materials is minimized.
A bench test to quantify the effect that coolant formulations have on the coolant pump
would be a valuable tool for developing formulations with less adverse impact on the coolant
pump. The ASTM Committee D-15 on Engine Coolants should undertake the development
of such a test and should include favorable performance in the test in the standards covering
engine coolants.
Acknowledgments
The author would like to thank Mr. Leonard M. Richmond and the Midtown Operating
Corporation for the excellent control of this fleet and the level of organization that the data
that they provided exhibited. The author would also like to thank Mr. Bradford S. Brown for
his guidance on the statistical evaluations of the data, and Dr. James L. Jackson for the use of
his statistical graphics software, his personal computer, and his time. Additionally, the tech-
nical and analytical assistance of Dr. Tristan A. Laurian, Mr. Charles F. Taylor, Mr. Stanley
E. Whitfield, and Ms. Renee T. Bade was much appreciated.
References
[1] Darden, J. W, Triebel, C. A., Maes, J. P., and Van Neste, W., "Monoacid/Diacid Combination as
Corrosion lnhibitors in Antifreeze Formulations," SAE Paper 900804, Society of Automotive Engi-
neers, 1990.
[2] Darden, J. W., Triebel, C. A., Maes, J. P., and Van Neste, W., U.S. Patent 4,647,392, Monobasic-
Dibasic Acid/Salt Antifreeze Corrosion Inhibitor, 1987.
[3] ASTM Committee D-15, 1990 Annual Book of ASTM Standards, Section 15 General Products,
Chemical Specialties, and End Use Products, Volume 15.05 Engine Coolants; Halogenated Organic
Solvents; Industrial Chemicals, American Society for Testing and Materials, Philadelphia, 1990.
[4] Automotive Engine Coolant, Concentrate--Ethylene Glycol Type, Specification GM6038-M, Gen-
eral Motors Engineering Standards.
[5] Statistical Graphics Corporation, Statgraphics Statistical Graphics System Users Guide, 1987.
[6] Kinisky, T. G. and McKenzie, D. A., "Investigation of Premature Water Pump Seal Failures," SAE
Paper 900806, 1990.
[ 7] Kiryu, K., Tsuchiya, K., Shimomura, T., Yanai, T., Okada, K., and HirabayashL H., "The Effect of
Coolant Additives and Seal Composition on Performance of Water Pump Seals of Automotive
Engines," SAE Paper 890609, 1989.
[8] Wisniewski, F., Seizer, J., Sehnal, J., and Casler, J. C., Inc., Report to ASTM Subcommittee D15.11
on Heavy Duty Coolants, 17 Aug. 1989, Minutes of Subcommittee Meeting, 1989.
[9] Poznicek, J., private communication, John Crane, Inc., 1990.
DISCUSSION
Norman Adamowicz j (written discussion)--How often were oil changes done?

J. M. Burns (author's response)--Approximately every 3000 to 5000 miles (4828 to 8047
km).
Norman Adamowicz ~(written discussion)--How large were the size of the coolant samples?
J. M. Burns (author's response)--Approximately 100 mL.
Norman Adamowicz ~(written discussion)--You flushed the cooling systems of all test cars
before filling with test coolant but did not flush the factory-fill systems. Why did you not flush
their coolant and refill with factory-fill coolant?
J. M. Burns (author's response)--The assumption was made that factory-filling the vehicles
is a preferred way to introduce the test coolant into these systems. We believed that the system
as delivered from the manufacturer was as good as it could be and any treatment that we tried
would be poor by comparison. The directionally poorer performance of the yellow coolant
relative to the factory-fill coolant seems to support this assumption, If it were possible to fac-
tory-fill all of the systems with the test coolants, it would have been done.
Norman Adamowicz ~(written discussion)--Why did you mix a nonapproved AIF W OEM
D/F? Actually then, what you tested was not coolant formulations in water pump seals but
coolant formulations of the cooling systems. Therefore your data/info are not supportive of
the title o f the paper.
J. M. Burns (author's response)--The practice of the fleet was adhered to and no changes
were made initially. Based on the data presented in the notched box-and-whisker plot, this
variable did not introduce considerable bias. The factory-fill coolant group data are presented
as anecdotal information that represents the situation in the field today. The similarity
between the data yielded by this portion of the fleet and the data produced by three of the four
test coolant groups that were established under comparable parameters indicates that the dif-
ference in top-off coolant from the factory-fill formulation did not introduce a great deal of
bias. The performance of the four conventional coolant formulations was essentially the same.
Peter Woyciesjes 2 (written discussion)--You have stated that there was "tremendous top
off" required in the test fleet. What was the amount of coolant added to the vehicles over the
course o f the test. What was the silicate level observed in the vehicles with silicate coolants?
Given the known effects on components like water pumps by overdosing inhibitors, can you
comment on the effects of "tremendous top off?."
J. M. Burns (author's response)--The silicate level (measured as silicon in solution) in the
yellow coolant portion of the fleet ranged from a low of 3 parts per million (ppm) to a high of
J Ford Motor Company.

DISCUSSION ON FLEET TEST CORRELATIONS 41
364 ppm (although only two samples had more than 180 ppm) with an average of 90 ppm.
The silicate level in the green coolant portion of the fleet ranged from 15 to 176 ppm with an
average of 59 ppm, and the silicate level in the purple coolant portion of the fleet ranged from
less than 1 p p m to 150 pprn with an average of 56 ppm. This certainly does not appear to be
over concentration of silicate.
The comment about there having been tremendous top offwas based on the fact that a large
volume of coolant was used, not on the inhibitor levels found in the systems. There are other
reasons for coolant usage.
I do not believe that overconcentration of silicate was the major cause of the incidence of
p u m p replacements, but that silicate is involved in the failure mode. The blue coolant had a
higher percentage of inhibitor additive, which is sometimes interpreted as total dissolved sol-
ids, but produced much better results. The differences in the silicate level accounts for only
0.036 wt% of the total solution. Even if overconcentration is involved, it would merely lend
support to the conclusion that a silicate-free coolant have improved compatibility with the
coolant pump.
Peter Woyciesjes 2 (written discussion)--It has been shown that silicate protects metals by
quickly forming a protective layer. The remainder of the silicate remains in solution to rein-
hibit any corrosion sites that may develop. The protective film cannot be removed by water
washing the cooling system.
Would not the drain and refill process with fresh silicate coolant lead to overconcentration
effects described in earlier presentations? Would not a nonsilicate coolant benefit from this
precharge of silicate coolant?
J. M. Burns (author's response)--Overconcentration of silicate is one possible explanation
for the data that were generated. The measured silicon levels in the coolant samples averaged
56 to 59 parts per million for the low silicate, universal formulations and 90 ppm for the high
silicate formulation.
These levels do not seem sufficient to explain the results through a mechanism related to
overconcentration of silicate, although the conclusions of this study would remain unchanged
even if they did. A silicate-free coolant would still have improved compatibility with the cool-
ant pump.
The original equipment manufacturers recommend replacing the factory-fill coolant after
a certain number of years to miles, and the consumer is likely to use the products that are
available on the market which are high silicate and universal formulations. Your argument
would suggest that a high silicate coolant should never be used in cooling systems other than
for factory-fill. Although I would certainly agree with this, the data presented here does not
suggest that simply lowering the silicate level provides much of an improvement.
The universal formulations, the green and the purple, which have half the initial silicate level
of the typical high silicate formulations, and the factory-fill systems that were topped off with
a low silicate formulation performed directionally better than the high silicate formulations,
but the results were of the same population statistically.
For the purposes of this study, there is no way that the silicate precharge could be a benefit
to the silicate-free coolant if the data are to be believed. The silicate would have a detrimental
effect on the coolant pump from the point of initial fill. No corrosion data were presented, and
it was not the author's intent to discuss that performance aspect in this publication. That has
been covered in other publications and may be the topic for a future publication.
William Mercer 3 (written discussion)--High lead solder losses for coolants with monoacids
are known to be very high and have been shown to be due to the formation of lead carboxy-
lates. Lead carboxylates were also shown to be transported throughout the cooling system and

are known to be excellent lubricants. You have stated that these vehicles were equipped with
specimen bundles that contained high lead solder. What were the high lead solder losses with
the blue coolant?
What evidence do you have that the extended p u m p life observed with the blue coolant is
not serendipitous and due to the formation of an excellent pump seal lubricant by corrosion
of high lead solder?
J. M. Burns (author's response)--High lead solder weight losses for the coolants with
monoacids are not known to be very high; they are merely alleged to be high in formulations
that may not be representative of this technology. The high lead solder weight losses for the
different coolant formulations were 155 mg for the blue coolant, 106 mg for the purple cool-
ant, 62 mg for the yellow coolant, and 46 mg for the green coolant. Although these are not
excellent results for the blue coolant, neither are they excellent results for any of the other
coolants tested, and it certainly does not support the theory that monoacids are generally
aggressive to the high lead solder alloy.
One hundred twenty two of one hundred sixty five coolant samples had no measurable lev-
els of lead in solution. The highest level of lead measured was 23 ppm while the average was
less than 5 ppm. The levels of lead found in the 5 of the 20 cars that had corrosion coupons
were not statistically different than the levels found in the other 15. Samples from the other
coolant test groups were not tested for lead.
Based on this fleet test, there is no support for the monoacid/lead corrosion theory presented
at this Symposium.
The data suggest that these formulations provide a lower degree of protection to the specific
lead alloy in question. To determine if formulations containing monoacids are aggressive to
lead solders, a performance comparison should be made with an uninhibited water/ethylene
glycol mixture.
The allegation that carboxylic acids are generally corrosive to lead solder alloys are not sup-
ported by the data available. Two recent patents, U.S. 5,085,791 and U.S. 5,085,793 suggest
that certain monoacids can even improve protection of the high lead solder alloy employed in
these tests.
The formation of lead carboxylate is one explanation, but that has not been proven. The
low depletion of the carboxylic acid suggests that this is not occurring. There is no evidence to
suggest that lead carboxylates are actually formed in these systems. Only 5 of the 20 vehicles
contained coupon bundles, and hence a source of the lead for formation of the lead carbox-
ylate. This mechanism does not explain the difference in results between the various
formulations.
The suggestion that lead carboxylates provide lubrication for the pump seal, thus explaining
the differences in p u m p compatibility noted in this study is an unlikely explanation at best.
Lead carboxylates have not been found on any of the coolant p u m p seal surfaces. The blue
and the purple coolants both contain the same monoacid, yet produced radically different
results. The purple coolant produced results that were statistically similar to the three formu-
lations that contained no monoacids.
If this mechanism is involved with the modes o f failure, then it is to a much lesser degree
than the silicate is involved.
William Mercer 3(written discussion)--The coolants were not randomized over initial start-
ing mileage. Were they randomized over the fuel and lubricant tests run concurrently on these
vehicles? What was the effect of the fuel and lube test products on vehicle performance? What
is the effect of lack of randomization on the statistical analysis of the results?
J. M. Burns (author's response)--The effect of the different starting mileages of the cars is
believed to be minimal. Figure 1 does not indicate that there is a considerable distinction
between three of the test coolants, the yellow, the green, and the purple, and the factory-filled
DISCUSSION ON FLEET TEST CORRELATIONS 43
vehicles. The blue coolant test vehicles were filled at a average mileage between that of the
factory-fill and the yellow, both o f which represent low initial mileages, and that of the green
and purple, both of which represent somewhat higher initial mileages, but produced statisti-
cally significantly better results than any of the other formulations involved.
Based on the data in Tables 3 to 7, no discernable bias was introduced by this variable. The
coolant test was transparent to and independent of the lubricant tests conducted in this fleet.
The coolant test groups were not related to the lubricant test groups. Different lubricants were
distributed randomly throughout the coolant test groups. Evaluation of the data for the cool-
ant testing has indicated no bias resulting from the lubricant testing which was concurrently
conducted in these vehicles.
W i l l i a m C. M e r c e r I
An Investigation of Carboxylic Acids as

Corrosion Inhibitors in Engine Coolant
REFERENCE: Mercer, W. C., "An Investigation of Carboxylic Acids as Corrosion Inhibitors in
Engine Coolant," Engine Coolant Testing. Third Volume, ASTM STP 1192, R. E. Beal, Ed.,
American Society for Testing and Materials, Philadelphia, 1993, pp. 44-62,
ABSTRACT: The ability of carboxylic acids to inhibit corrosion in automotive cooling systems
was first reported more than 40 years ago. Aromatic carboxylic acids, such as benzoic acid and
its derivatives, have been extensively studied and are used in Europe as corrosion inhibitors in
commercial coolants. There are few studies of corrosion inhibitionby aliphatic carboxylic acids
in the open literature. Aliphatic dibasic acids, such as l, 10-decanedioic,and aliphatic monobasic
acids, such as octanoic, have recently been proposed as inhibitors in automotive coolant for-
mulations. To date, ASTM performance test data have not been available.
This paper presents a comparison of the performance of selected carboxylic acid coolants with
each other and with a traditional North American coolant in laboratory and fleet tests. Protec-
tion of cooling system metals, inhibitordepletion, and mixed coolant effects were investigated.
KEYWORDS: corrosion, engine coolants, automotive coolants, organic acids, octanoic acid, 2-
ethylhexanoic acid, l, 10-decanedioicacid, tolyltriazole, silicate
Alternative coolant formulations for improved corrosion inhibition have been an objective
of research in coolant laboratories for many years. While many formulations are proposed,
few are acceptable commercial engine coolants. The performance of proposed coolants is
judged in laboratory and fleet tests against a standard. A m i n i m u m is ASTM Specification for
Ethylene Glycol Base Engine Coolant for Automobile and Light Duty Service (D 3306).
Another standard is the performance of commercially available coolants. Proposed formula-
tions that perform better than the standards are considered superior formulations and are
likely to be commercialized.
Engine coolants must inhibit corrosion of cooling system metals. These include steel, cast
iron, copper, brass, lead solders, cast aluminum, and wrought aluminum. Cooling system met-
allurgy is in a state of flux. In the last 15 years engine metallurgy has changed from primarily
cast iron to a mix of cast a l u m i n u m and cast iron. Heat exchangers in newer cars are primarily
wrought aluminum, but there are still many cars with soldered copper/brass radiators. Because
of the variation in cooling system metallurgy, coolants must contain inhibitors for all metal
combinations found in vehicles.
A consumer survey has shown that as many as 41% of cooling systems are neglected to some
degree [ 1]. Automobiles are operated with little or no coolant in the overflow bucket, with low
coolant concentrations, and coolant beyond its useful life. Cooling systems that have been
neglected are difficult to reinhibit because of increased surface area and the presence of acidic
corrosion products. Coolants must provide good performance in both new systems and cor-
roded systems.
1 Group Leader, Coolant Group, First Brands Automotive R&D Laboratory, 55 Federal Road, Dan-
bury, CT 06810.
44
MERCER ON CARBOXYLIC ACIDS AS CORROSION INHIBITORS 45
The introduction of a new coolant should not cause compatibility problems with existing
coolants. Coolant mixing can occur when systems are topped off or when coolant is changed.
If coolant mixing causes degraded corrosion inhibition, the coolants are incompatible.
Aliphatic carboxylic acids have been under study in our laboratories for several years. Cor-
rosion inhibition in automotive cooling systems by aliphatic carboxylic acids was first pat-
ented in the United States in 1945 [2]. There was little interest in the efficacy of aliphatic car-
boxylic acids until Butler and Mercer [3] published on the subject in 1977. This interest was
followed by technical activity [3] and some recent marketing activity. For example, the Saturn
factory fill contains an aliphatic carboxylic acid as a minor inhibitor.
Silicate based coolants with phosphate/borate or phosphate buffer systems have dominated
both factory fill and the aftermarket in North America since their introduction in 1972. A
representative silicate coolant (GM 6043M) was chosen as the standard for an evaluation of
some aliphatic carboxylic acid coolants. A series of coolants with aliphatic carboxylic acid
inhibitors was evaluated in laboratory and fleet tests. Protection of cooling system metals,
inhibitor depletion, and mixed coolant effects were investigated. This paper presents the
results of these investigations.
Experimental Procedures
Test coolants prepared in the laboratory were analyzed before and after all tests. Analyses
included inductively coupled plasma emission spectroscopy (ICP) for inorganic ions, gas chro-
matography for the organic acids, and ion chromatography for sodium nitrate, sodium tolyl-
triazole, and sodium mercaptobenzothiazole.
Coolant Composition
Two experimental designs were used to determine the composition of monoacid/diacid
combinations for future testing. Test coolant compositions are shown in Tables 1 and 2. Cool-
ants A - L used combinations of octanoic acid and 1,10-decanedioic acid. Coolants M - X
included nine combinations of 2-ethylhexanoic acid and 1,10-decanedioic acid and three cool-
ants that used only 2-ethylhexanoic acid.
TABLE 1--Composition of coolants A-L.
Sodium Sodium
Octanoic 1,10-Decanedioic Tolyltriazole, Hydroxide, Balance,
Coolant Acid, wt% Acid, wt% wt% wt% wt%
A 0.38 0.4 0.25 0.24 98.73

B 0.76 0.4 0.25 0.34 98.25
C 1.52 0.4 0.25 0.53 97.30
D 2.28 0.4 0.25 0.73 96.34
E 0.38 0.8 0.25 0.39 98.18
F 0.76 0.8 0.25 0.49 97.70
G 1.52 0.8 0.25 0.68 96.75
H 2.28 0.8 0.25 0.87 95.80
I 0.38 1.2 0.25 0.53 97.64
J 0.76 1.2 0.25 0.63 97.16
K 1.52 1.2 0.25 0.83 96.20
L 2.28 1.2 0.25 1.02 95.25
NOTE: Balance includes ethylene glycol, water, dye, and antifoam.
TABLE 2--Composition of coolants M-X.
Sodium Sodium
2-Ethylhexanoic 1,10-Decanedioic Tolyltriazole, Hydroxide, Balance,
Coolant Acid, wt% ~ Acid, wt% wt% wt% wt%
M 0.5 0 0.1 0.13 99.27

N 2.0 0 0.1 0.55 97.35
O 3.5 0 0.1 0.97 95.43
P 0.5 0.2 0.1 0.21 98.99
Q 2.0 0.2 0.1 0.65 97.05
R 3.5 0.2 0.1 1.03 95.17
S 0.5 0.8 0.1 0.44 98.16
T 2.0 0.8 0.1 0.87 96.23
U 3.5 0.8 0.1 1.26 94.34
V 0.5 1.4 0.1 0.69 97.31
W 2.0 1.4 0.1 1.10 95.4
X 3.5 1.4 0.1 1.49 93.51
NOTE: Balance includes ethylene glycol, water, dye, and antifoam.
The composition of coolants 1 and 2 were developed based on results obtained with coolants
A - L and M - X , respectively. Table 3 shows the composition of coolants 1,2, and GM 6043M.
Laboratory Tests
Laboratory tests included ASTM Test Method for Corrosion Test for Engine Coolants in
Glassware (D 1384), ASTM Test Method for Corrosion of Cast A l u m i n u m Alloys in Engine
Coolants Under Heat Rejecting Conditions (D 4340), and ASTM Test Method for Cavitation
Erosion-Corrosion Characteristics of A l u m i n u m Pumps With Engine Coolants (D 2809).
There were several deviations from the ASTM D 1384 standard. These included the addition
of a high lead solder specimen (97.0% lead, 2.5% tin, and 0.5% silver), a modified beaker, and
the use of fine pumice for cleaning the a l u m i n u m specimens. The results presented for ASTM
D 1384 are averages of triplicate runs, and those for ASTM D 4340 are averages of duplicate
runs. ASTM D 2809 was run once for each fluid.
The high lead solder specimens were coated with a heavy deposit after ASTM D 1384 when
monoacids were present in the coolant. Since the standard solder cleaning procedure did not
remove all of the deposit, a modified cleaning procedure was used. The samples were heated
TABLE 3--Composition of Fleet test coolants.
Inhibitor, wt% Coolant 1 Coolant 2 GM 6043M
(85%) Phosphoric acid . . . . . . O.15

Sodium tetraborate . . . . . . 0.40
Sodium Silicate 0.30
(500/0) Sodium tolyltriazole Oi51J 0146 0.20
(50%) Sodium mercaptobenzothiazole . . . . . . 0.50
Sodium nitrate . . . . . . O.10
Sodium molybdate ...
Octanoic acid 115 "'"
2-Ethylhexanoic acid 315 "'"
1,10-Decanedioic acid ()18 0.2 iii
in toluene for 2 h at 100~ rinsed with acetone, and recleaned according to the standard pro-
cedure. This cleaning procedure was not required for high lead solder specimens removed
from fleet vehicles. There was no deposit on these specimens.
Linear Regression Analyses

The contributions of the individual inhibitors in coolant A - X to the data obtained in the D
1384 tests were evaluated by linear regression techniques. The independent variables were ini-
tially evaluated by stepwise regression techniques. Stepwise regression is an iterative process.
The experimental variables are evaluated individually to determine the best straight line
through the data. The effect o f the best fit variable is subtracted from the data. The program
then fits a line through the remaining variables using the residual data. The variables identified
as important in the stepwise regression analysis are used in the multiple linear regression
analysis.
The model used in the stepwise analysis was
specimen weight loss = constant + [monoacid]

+ [diacid] + [monoacid] 2
+ [diacid]-' + ([monoacid]*[diacid])
The model accounted for contributions by the individual acids in each array. The squared
term accounted for effects of the acids where the contribution changes more rapidly than pro-
vided for by the linear term. Interactions between the acids that cause more than additive
effects were also examined. The linear regression provided coefficients for each term. These
coefficients show whether the variable increased (positive coefficient) or decreased (negative
coefficient) specimen weight loss.
Fleet Tests
Selected aliphatic carboxylic acid coolants were evaluated in fleet tests. Coolants 1 and 2
were tested in older vehicles (Series 1). Coolant 2 and G M 6043M were tested as factory fill
(Series 2) in new vehicles. Coolants were evaluated by metal specimen weight loss and solution
analysis. A description of the fleet test procedures is in Table 4.
A bypass (heater circuit) specimen capsule was used in both tests as in ASTM Practice for
TABLE 4--Fleet test procedures.
Vehicle Odometer Test

Type Initial, Duration, Engine Engine Engine Coolant
Fleet Coolant (Number) km km Type Metallurgy Treatment Dilution
Series 1 Voyager 48 000 48 000 3.0L V6 l water wash yes
1 (2)
Series 2 Voyager 96 000 48 000 3.0L V6 l water wash yes
1 (2)
Series 2 Grand 0 118 000 3.1L V6 2 factory fill no
2 Prix (5)
Series GM 6043M Grand 0 118 000 3.1L V6 2 factory fill no
2 Prix (5)
ENGINE METALLURGY: I Cast iron block, aluminum cylinder head, and brass radiator. 2 = Cast
=
iron block, aluminum cylinder head, and aluminum radiator.

Testing Engine Coolants in Car and Light Truck Service (D 2847). The specimen bundles con-
tained the same seven metals used in ASTM D 1384. New radiators were put in all test vehicles
at the beginning of the test. These radiators were tested to determine their thermal and flow
characteristics before and after the test. Other cooling system components were examined and
replaced if necessary.
Series 1 Tests
The fleet tests of coolants 1 and 2 were run consecutively in the same vehicles. The 48 000-
km tests took approximately 10 months to complete. Specimen bundles were removed at
8000-km increments between 16 000 and 48 000 km. Solution samples were obtained at 8000-
k m intervals.
The coolant concentration was reduced during the Series 1 tests from 50 to 25 vol%. ASTM
corrosive water was used for coolant dilution. Coolant dilution mimicked consumer neglect.
Dilution stressed the coolant and provided an indication of the ability to protect under severe
duty conditions.
Series 2 Test--The Series 2 test is scheduled for three years and 116 000 km. The vehicles
were factory filled with the test coolant. Coolant dilution will remain at 50 vol% with deionized
water throughout the test. Specimen bundles are removed at 19 000-km intervals. Coolant
samples are removed at 16 000-km intervals.
Results and Discussion

A S T M D 1384 of Monoacid/ Diacid Combinations
Two sets of twelve coolants were used to evaluate the performance of aliphatic carboxylic
acids in ASTM D 1384 and to choose combinations ofmonoacids and diacids for further test-
ing. The weight loss data are shown in Tables 5 and 6. The weight loss data and the results of
linear regression analyses are discussed below.
High Lead Solder--There are two weight loss columns presented for high lead solder
(Tables 5 and 6). The first column is the weight loss after cleaning according to the procedure
for solder in A S T M D 1384, and the second column is the weight loss after the modified pro-
cedure described in the Experimental Section. The special procedure was used because the
specimens were coated with a layer of corrosion product after the initial cleaning. The removal
of the corrosion product resulted in higher specimen weight losses (6 to 282 mg). The alter-
native cleaning procedure was not the cause of the weight loss. Similar treatment of specimens
from tests of traditional North American coolants showed no weight loss change. The weight
of untested specimens was unaffected by the alternative treatment.
The use of the alternative cleaning procedure was supported by fleet test results (see below).
The corrosion product was never seen on these specimens despite high weight losses. The tol-
uene soluble layer was sufficiently adherent to remain on the specimen in the no flow condi-
tions of a beaker (ASTM D 1384) but did not remain on the specimen in a real cooling system.
No standards have been set for high lead solder weight loss in ASTM D 3306. The data
presented here must be evaluated with the standard for ASTM solder of 30 mg and the con-
sideration that approximately 1000 mg of solder was originally coated on the brass specimen.
The ASTM solder standard can be used. Our experience is that high lead solder is more difficult
to protect than ASTM solder. Therefore, a rule of t h u m b is that below 30 mg weight loss, high
lead solder is well protected, and above 100 mg (10% of available solder) weight loss, a serious
high lead solder corrosion problem exists.
Only 3 of the 24 coolants had weight losses less than 100 mg per specimen after the standard
TABLE 5---ASTM D 1384 weight loss data coolants A-L, rag~specimen.
High Lead High Lead ASTM Cast Cast

Coolant Solder (1) Solder (2) Copper Solder Brass Steel Iron Aluminum
A 182 247 2.2 1.2 2.5 0.6 --0.8 123.4

B 262 296 2.6 2.4 3.0 1.2 -- 1.0 100.2
C 113 229 1.7 1.6 2.4 0.5 -1.2 9.4
D 122 338 1.8 1.1 2.3 0.0 - 1.3 48.8
E 86 288 1.8 2.7 3.0 1.1 -0.7 20.1
F 70 306 2.5 2.1 3.4 0.1 -0.6 1.7
G 282 449 2.7 2.7 3.5 0.9 0.1 13.2
H 224 393 2.2 3.1 3.0 0.8 -0.9 4.4
I 154 436 1.1 3.2 1.5 --0.4 - 1.8 4.3
J 184 507 2.8 12.0 3.5 0.7 - 1.0 2.5
K 166 509 3.0 5.5 3.6 2.3 1.3 11.5
L 162 534 3.1 6.3 3.9 0.6 -0.7 0.0
Upper limit none none 10 30 10 10 10 30
(D 3306)
NOTE: High lead solder (1)--weight loss after standard cleaning procedure. High lead solder (2)--
weight loss after modified cleaning procedure.
c l e a n i n g p r o c e d u r e . O n l y c o o l a n t M h a d a weight loss less t h a n 100 m g after the alternative

c l e a n i n g p r o c e d u r e . N o n e o f t h e c o o l a n t s h a d high lead solder s p e c i m e n weight losses less t h a n
30 mg. C o o l a n t s J - L h a d weight losses greater t h a n 500 rag. T h i s was a p p r o x i m a t e l y 50% o f
t h e available high lead solder. N o n e o f the test c o o l a n t s p r o v i d e d good p r o t e c t i o n for high lead
solder.
T h e c o n t r i b u t i o n o f t h e i n d i v i d u a l i n h i b i t o r s to t h e weight losses was d e t e r m i n e d using lin-
ear regression t e c h n i q u e s . T h e m o d e l for h i g h lead solder corrosion derived from this analysis
was the s a m e for b o t h sets o f c o o l a n t s
high lead solder weight loss = Co + c~[monoacid]
TABLE 6--ASTM D 1384 weight loss data coolants M-X, rag~specimen.
High Lead High Lead ASTM Cast Cast

Coolant Solder ( 1) Solder (2) Copper Solder Brass Steel Iron Aluminum
M 73 79 3.2 7.3 3.9 6.6 388 61.9

N 128 138 3.4 4.8 5.0 3.9 238 73.8
O 160 212 3.5 1.0 5.5 1.2 -0.6 39.1
P 112 134 3.6 0.8 4.9 1.4 1.6 10.9
Q 192 223 3.3 1.6 4.6 1.4 - 1.1 6.9
R 218 310 3.6 2.2 3.7 1.4 -0.8 12.8
S 163 201 3.7 5.7 5.6 0.8 - 1.0 11.9
T 292 337 3.6 4.6 5.2 0.9 -0.5 10.8
U 332 378 3.4 3.3 3.6 1.0 - 1.4 8.2
V 193 215 4.3 10.7 4.2 1.1 -0.7 9.5
W 267 324 6.0 3.9 2.5 1.3 -0.1 7.6
X 202 237 4.3 2.8 3.2 1.1 --0.6 12.2
Upper limit none none 10 30 10 10 10 30
(D 3306)
NOTE: High lead solder (l)--weight loss after standard cleaning procedure. High lead solder (2)--
weight loss after modified cleaning procedure.
For coolants A - L this model accounted for 95% of the data, and for coolants M - X the model
accounted for 54% of the data (Table 7). The monoacid coefficient was positive. Therefore,
increased monoacid concentration was accompanied by increased high lead solder weight loss.
There was no effect of the diacid or an interaction between the acids observed for either set of
coolants. The effect of the 2-ethylhexanoic acid was clearly seen in coolants M, N, and O where
there was no diacid present. The high lead solder weight loss increased from 79 to 378 mg as
the 2-ethylhexanoic acid level increased from 0.5 to 3.5 wt%. The monoacids attacked high
lead solder in ASTM D 1384.
A S T M Solder--All coolants provided adequate protection to ASTM solder in this test. The
weight loss range for coolants A - L was 1 to 12 mg and for coolants M - X was 0.8 to 10.7 mg.
This was a large enough range that linear regression techniques were used as above to deter-
mine the effect of inhibitors on weight loss. There were no observable effects of either 2-ethyl-
hexanoic acid or 1,10-decanedioic acid on the corrosion of ASTM solder in coolants M - X .
There was an effect of octanoic acid on the corrosion of ASTM solder in coolants A-L. The
model was
ASTM weight loss - Co + c~[octanoic] + c2[octanoic]2
This model accounted for 73% of the data (Table 8). As the concentration of octanoic acid
went up the corrosion of ASTM solder went up. The positive second order term indicated that
at higher concentrations of octanoic acid the corrosion losses increased more rapidly than
accounted for by the linear term.
The protection of ASTM solder was adequate in the presence of the more active high lead
TABLE 7--High lead solder multiple linear regression.
Coolants A-L: High Lead Solder, N = 12, r2 = 0.951

Analysis of Variance
Source Sum of Squares df Mean Square F Ratio Probability
Total 127385.64 11
Regression 121206.62 1 121206.62 196.16 0.0000
[monoacid] 121206.62 1 121206.62 196.16 0.0000
Residual 6179.02 10 617.90
Regression Coefficients
Standard Standard
Variable Coefficient Error T Value
Constant 377.7 7.2 52.64

[monoacid] 137.7 9.8 14.01
Coolants M-X: High Lead Solder, N = 12, r E = 0.541
Total 62355.43 11
Regression 33748.02 1 33748.02 11.80 0.0064
[monoacid] 33748.02 1 33748.02 11.80 0.0064
Residual 28607.4 t 10 2860.74
Standard Standard
Constant 194.4 15.4 12.59

[monoacid] 43.3 912.6 3.43
TABLE 8--ASTM solder multiple linear regression.
Coolants A-L: ASTM Solder, N = 12, r2 = 0.729

Total 103.15 11
Regression 75.16 2 37.58 12.08 0.0028
[monoacid] 63.78 1 63.78 20.51 0.0014
[monoacid]2 11.38 1 11.38 3.66 0.0881
Residual 27.99 9 3.11
Standard Standard
Constant 2.3 0.88 2.57

[monoacid] 2.6 0.75 3.51
[monoacid] z 2.6 1.36 1.91
solder. However, the corrosion of ASTM solder was accelerated by octanoic acid. More testing
of these systems needs to be done to demonstrate that the protection of ASTM solder will
continue to remain adequate in the absence of high lead solder.
Cuprous Metals--Copper and brass protection was adequate with coolants A - X . All weight
losses were 6 mg or less, well below the 10 mg limit is ASTM D 3306-89. Statistical analysis
of the weight losses showed no effect of acid concentration for either metal.
Ferrous Metals--Except for coolants M and P the protection of the ferrous metals was very
good. Coolants M (0.5%) and P (0.70/o) suggest that a m i n i m u m acid level (>0.7%) in the cool-
ant concentrate was required to protect cast iron. Protection was very good once that level was
exceeded. There were no other effects due to either acid or an interaction between them
observed in the linear regression analysis.
Aluminum--Aluminum weight losses ranged from 0 to 123 mg in coolants A - L and 7 to
74 mg in coolants M - X . Nine of the twelve coolants in each array had weight losses lower than
the ASTM D 3306 specification (30 mg) for aluminum weight loss in ASTM D 1384. There
also seemed to b e a m i n i m u m acid level required to protect aluminum. The aluminum losses
were above 30 mg when there was less than 1.35% acid present in the coolant concentrate.
The effect of the acids and interactions between them was determined by linear regression
techniques. For coolants A - X the model that described corrosion was
aluminum weight loss = co + c~[monoacid] + ~[monoacid*diacid]

+ c3[monoacid] 2 + c4[diacid]
This model accounts for 87% of the data (Table 9) for coolants A-L. The coefficients for
octanoic acid and 1,10-decanedioic acid were both negative, indicating that corrosion
decreased as the concentration of each acid increased. The positive second order term for octa-
noic acid indicates that weight losses were higher at higher acid concentrations than accounted
for by the linear term. An interaction term between the acids was also present. The coefficient
for the interaction term was positive. Therefore, the interaction between the acids resulted in
greater aluminum corrosion. Thus, this interaction, if real, represents a negative synergism.
This suggests that the acids would be more effective when used alone. A molar ratio [3] term
was included in the evaluation. It had no significant effect.
The same model describes aluminum corrosion in coolants M - X . The model accounts for
90% o f the data (Table 9). The acids protect aluminum by the same mechanism even though
TABLE 9--Aluminum multiple finear regression.
Coolants A-L: Aluminum, N = 12, r 2 = 0.867

Souse Sum of Squares df Mean Square F-ratio Probability
Total 18888.17 11
Regression 16377.06 4 4094.26 11.41 0.0035
[diacid] 3718.24 1 3718.24 10.37 0.0147
[monoacid] 7023.68 1 7023.68 19.58 0.0031
[monoacid] 2 2545.38 1 2545.38 7.10 0.0323
[mono*di] 3089.77 1 3089.77 8.61 0.0219
Residual 2511.11 7 358.73
Regression Coetficients
Standard Standard
Variable Coetficient Error T-Value
Constant 7.6 9.5 0.8

[diacid] - 53.9 16.7 - 3.22
[monoacid] - 41.0 8.0 - 5.09
[monoacid] 2 38.8 14.6 2.66
[mono*di] 67.3 22.9 2.93
Coolants M-X: Aluminum, N = 12, r 2 = 0.898
Source Sum of Squares df Mean Square F-ratio Probability
Total 5880.85 11
Regression 5285.51 4 1320.63 15.45 0.0014
[diacid] 582.68 1 582.68 6.82 0.0349
[monoacid] 2745.40 1 2745.40 32.12 0.0008
[monoacid] 2 798.11 1 798.11 9.34 0.0184
[mono*di] 1156.33 1 1156.33 13.53 0.0079
Residual 598.33 7 85.48
Regression Coetficients
Standard Standard
Variable Coel~cient Error T-Value
Constant 10.6 4.6 2.29

[diacid] - 12.7 4.9 - 2.61
[monoacid] - 12.3 2.2 - 5.67
[monoacid]: 7.7 2.5 3.06
[mono*di] 14.6 3.9 3.68
the m o n o a c i d has been changed. This is not surprising since 2-ethylhexanoic acid is an isomer
o f octanoic acid. There did not appear to be any effect of the change in chain length from
octanoic acid to 2-ethylhexanoic acid as was postulated [3].
Laboratory Testing: Coolants 1 and 2

T w o coolants were selected for further laboratory testing based on the A S T M D 1384 results
with coolants A - X . Coolant 1 contains 1.5 wt% octanoic acid and 0.8 wt% 1,10-decanedioic
acid. C o o l a n t 2 contains 3.5 wt% 2-ethylhexanoic acid and 0.2 wt% 1,10-decanedioic acid. The
coolant compositions were selected to give the best balance o f metal specimen protection
available within the tested concentration range.
A S T M D 4340--The data from A S T M D 4340 testing of these coolants is shown in Table
10. Coolant 1 failed the test marginally. Coolant 2 passed the test. The coolant 2 corrosion rate
was slightly higher than that o f G M 6043M.
TABLE IO--ASTM D 4340 data.
Aluminun Corrosion
Coolant Rate, mg/cm 2 week Result
1 1.1 fail
2 0.3 pass
GM 6043M 0.1 pass
Upper limit (D 3306) 1.0 ...
A S T M D 2809--Water p u m p ratings for these coolants were in the 9 to 10 range. The alu-
m i n u m surface was discolored but not damaged by coolants 1 and 2. The coolants provided
adequate protection for aluminum water pumps.
Compatibility o f Coolant Mixtures--Mixing of coolants can occur in several ways. Virtually
all cooling systems require makeup at some time because of small leaks. Component failures
can lead to large coolant losses. The periodic replacement o f coolant also can lead to mixing
since it is difficult to remove all of the coolant from the system. Since mixing will occur, it is
imperative to assure compatibility between coolants. The corrosion inhibition provided by a
mixture of coolants should be no worse than the m i n i m u m the performance of the individual
coolants.
Mixing effects were examined by combining coolant 2 with two conventional North Amer-
ican coolants. ASTM D 4340 was used to evaluate performance. Representative samples of
the North American coolants were obtained in the aftermarket. The North American coolants
were phosphate buffered and had both high and low silicate levels. These coolants passed
ASTM D 4340 when tested alone (0 vol% coolant 2, Fig. l). In all cases there was a range of
mixtures with corrosion rates greater than the specification for ASTM D 4340 in ASTM D
3306 (1.0 mg/cm2/wk).
Adding small amounts of the high silicate phosphate buffered coolant caused the corrosion
inhibition provided by coolant 2 to break down. At the lowest level tested, 90% coolant 2, the
corrosion rate was greater than 4.0 mg/cm2/wk. The data in Fig. l suggest contamination by
as little as 2.5 vol% conventional coolant will cause corrosion o f heat rejecting aluminum.
These levels easily can be reached when topping offa factory fill of organic acid coolant with
Corrosion Rate (mg/cm2/wk)

10 r
ASTM
Spec-~
25 40 60 75 90 100
Coolant 2 (Volume %)
-e- High Si/Phosphate -=- Low Si/Phosphate
FIG. 1--Coolant mixture corrosion rates in ASTM D 4340.
conventional coolant. The corrosion rate continued to increase as the concentration o f coolant
2 decreases. The corrosion rate with the high silicate phosphate buffered coolant reached a
m a x i m u m at 75 vol% coolant 2. A corrosion rate below 1.0 mg/cmZ/wk was not reached until
the conventional coolant accounted for nearly 60% of the mixture.
The situation with the low silicate phosphate buffered coolant was worse because the cor-
rosion rate was unacceptably high over the entire mixture range (Fig. 1).
The solution chemistry of the test coolants was adversely affected by mixing. Complete
depletion of silicate was observed, and the phosphate buffered coolants had large decreases in
the phosphate concentration. There was no inhibitor depletion when the phosphate buffered
coolants were tested alone. The solutions above the aluminum test specimens were opaque
because of suspended corrosion products. In a vehicle, suspended corrosion products like these
are transported to the radiator and deposited, leading to radiator plugging and cooling system
failure [19].
The exact cause of the corrosion is unknown. It was reported that low levels of some anodic
corrosion inhibitors provided worse protection than not having the inhibitor in the formula-
tion [20]. Because ASTM D 4340 was run at 25% dilution, the levels of corrosion inhibitors
from the conventional coolants were at very low levels in these mixtures. The organic acid
coolants can provide acceptable aluminum protection when not contaminated. However, they
cannot overcome the aggressive condition caused by the presence o f small amounts o f con-
ventional corrosion inhibitors. The high silicate coolant is robust enough to protect heat reject-
ing aluminum when diluted by up to 40% by coolant 2. The low silicate coolants are not com-
patible with coolant 2 at any level.
Fleet Testing--The fleet tests addressed coolant performance in both used (Series 1) and
factory filled (Series 2) vehicles. Coolant 1 was evaluated in used cars only, while coolant 2 was
also tested in used vehicles and selected for a longer term test in new vehicles. Coolant 2 was
selected for the longer term test because o f improved performance in laboratory tests and in
the Series 1 fleet test. G M 6043M is the standard factory fill for the test vehicle and so was used
as the control. The fleet test data are shown in Tables 11 and 12.
Solder Protection ofSeries I Fleet Tests--Coolants 1 and 2 were unable to protect high lead
solder (Table 11). At 16 000 k m the high lead solder specimen in coolant 1 had lost approxi-
mately 50% (480 mg) of the solder originally on the specimen. At the end of 48 000 km the
specimen was almost completely stripped of solder (947 mg). The performance o f coolant 2
was better but still unacceptable. The coolant 2 high lead solder weight loss at 16 000 km was
240 mg and increased to 727 mg by the end of the test. In both coolants the weight losses
increased throughout the test. The high lead solder specimens removed from the fleet vehicles
were free of the corrosion product observed on the specimens following ASTM D 1384.
After the fleet tests, the radiators were removed from the vehicles, sectioned, and individual
TABLE 11--Series 1fleet test weight loss data, rng/specimen.
High Lead ASTM Cast Cast

km Solder Copper Solder Brass Steel Iron Aluminum
COOLANT I
16 000 480 10 4 6 3 --2 32
32 000 521 8 8 12 4 -2 38
48 000 947 6 6 14 1 -- 1 100
COOLANT 2
16 000 239 11 0 8 0 0 32
32 000 356 6 --2 7 0.5 0 42
48 000 727 8 2 7 6 --2 54
TABLE 12--Series 2fleet test weight loss data, rng/specimen.

km Solder Copper Solder Brass Steel Iron Aluminum
COOLANT 2
19 000 194 42 --0.3 12 --0.3 -- 1.4 2
38 000 865 48 0.6 18 0.4 -- 0.6 3
GM 6043M
19 000 11 44 0.4 9 --0.3 3 0
38 000 126 47 --0.8 13 --0.6 2 0
tubes split open for examination. The inside of the coolant 1 and 2 radiator tubes was coated
with a deposit identified as a lead carboxylate by infrared spectroscopy and inductively cou-
pled plasma (ICP) emission spectroscopy. The carboxylate has not been identified. However,
it is probably the monoacid given the dependence o f high lead solder corrosion on monoacid
concentration in ASTM D 1384 tests of coolants A - X . Lead carboxylate in the radiator sug-
gests that the lead corrosion product is soluble in the coolant and can be transported to cool
areas where it precipitates.
Coolants 1 and 2 protected A S T M solder well. There was no loss higher than 8 mg for either
coolant.
Cuprous~Ferrous Metal Protection of Series 1 Fleet Tests--Coolants 1 and 2 protected the
cuprous metals and ferrous metals well.
Aluminum Protection of Series 1 Fleet Tests--The coolant I aluminum weight loss at
16 000 k m was 30 mg and increased to 100 mg by 48 000 km (Table 11). There was a continued
increase in weight loss with no signs of leveling out toward the end of the test. The weight loss
was due to general surface corrosion. There was no localized corrosion observed.
Coolant 2 performed slightly better. The initial weight loss was 30 rag. The final weight loss
was 54 mg and again showed no sign of leveling off.
Solder Protection of Series 2 Fleet Tests--The longer term fleet test shows similar perfor-
mance (Table 12). G M 6043M provided better high lead solder protection at 38 000 km with
a 126-mg average weight loss. Coolant 2 continued to provide poor protection for high lead
solder with solder losses greater than 860 rag.
In this fleet, the high lead solder weight losses were accompanied by very high soluble lead
levels. Unlike the Series 1 tests where there was no detectable lead in solution, the 40 000 km
chemical analysis for the coolant 2 vehicles showed 70 parts per million (ppm) lead in solution.
This level exceeds the EPA limit of 5 ppm and means that this coolant must be disposed as
hazardous waste. There was no detectable lead in the G M 6043M samples.
Cuprous~Ferrous Metal Protection of Series 2 Fleet Tests--Cuprous and ferrous metal pro-
tection by coolant 2 and G M 6043M was equivalent through 38 000 kin.
Aluminum Protection of Series 2 Fleet Tests--The Series 2 test has passed the 38 000-km
mark (Table 12). There was no aluminum specimen weight loss with G M 6043M. Coolant 2
had a 3-rag weight loss.
The performance of coolant 2 was much better in the Series 2 fleet than in the Series 1 fleet.
There were two major differences between the fleets. One was the age of the vehicles. The sec-
ond was that deionized water rather than ASTM corrosive water was used in the Series 2 test.
The use of deionized water means that there was no chloride present to accelerate aluminum
corrosion in the Series 2 fleet. It appears that coolant 2 performance may be a function of the
condition of the cooling system and will require more care in the selection of water for dilution.
The effect of corrosive ions, such as chloride, on the performance of coolant 2 is being exam-
ined in our laboratories.
TTZ Concentration (mg/L)

1200
1000
80C
600
40(
200
I
25 50
Kilometers (Thousands)
~-Coolant 2 (Series 1) =-Coolant 2 (Series 2) -)(-GM 6043M

FIG. 2--Tolyltriazole depletion in fleet tests
Inhibitor Depletion--One of the claimed advantages of aliphatic carboxylic acid coolants is

that they will have extended service lives compared to current coolants. These claims are based
on constant acid levels in solution during testing. It has been confirmed that there is no observ-
able change in the acid concentration during a 48 000-km fleet test. However, constant acid
concentration has not meant constant corrosion inhibition throughout the fleet test. Both alu-
minum and solder weight losses increased throughout fleet tests.
The other major inhibitor in coolants 1 and 2 is tolyltriazole. The tolyltriazole level
decreases throughout the course of the test. While a conventional coolant is reported to lose
about 10% of the initial charge of tolyltriazole [21], organic acid coolants regularly lose 80%
of the tolyltriazole in 48 000-km tests (Fig. 2). The depletion rate is higher in factory filled
vehicles. In Series 1 tests the coolant 2 tolyltriazole level at 48 000 km was about 21% of the
initial level. In the Series 2 tests the coolant 2 tolyltriazole level is below 14% of the original
level at 40 000 km (Fig. 2). The tolyltriazole level in the GM 6043M at 40 000 km is 84% of
the original level (Fig. 2).
The cause of the tolyltriazole depletion is unknown. One report on tolyltriazole depletion
showed that the tolyltriazole lost from solution could be recovered from the metal surfaces or
from solid residues [22]. The authors concluded that tolyltriazole depletion did occur, but was
not due to degradation or chemical reaction. They felt that the tolyltriazole was complexed to
metal surfaces or present within residues and corrosion products on the metal surfaces. The
observation of depletion only in organic acid coolants suggests that the organic acid plays some
role in the depletion.
Conclusion
Selected carboxylic acid formulations have been evaluated in laboratory and fleet tests. The
standards by which they have been judged are ASTM D 3306 and comparisons with GM
6043M.
Experimental design sets tested in ASTM D 1384 gave a good overview of the ability of
aliphatic monoacids and diacids to protect c o m m o n cooling system metals. Aggressive behav-
ior of octanoic and 2-ethylhexanoic acid toward solder was observed in these tests. Cuprous
metal protection was good. The design sets suggested that m i n i m u m levels of the acids must
be exceeded before a l u m i n u m and cast iron were adequately protected.
Design set results were used to select coolants for further testing. Coolant 1 was a combi-
nation of octanoic acid and 1,10-decanedioic acid. Coolant 2 was a combination of 2-ethyl-
hexanoic acid and 1,10-decanedioic acid. Of these two, coolant 2 offered better performance
in both laboratory and fleet testing and was selected for further testing. Coolant 1 would not
meet ASTM D 3306 standards because it failed ASTM D 4340.
ASTM D 4340 showed that mixtures of coolant 2 and silicate containing aftermarket cool-
ants gave unusually high a l u m i n u m corrosion rates. The exact mechanism for the corrosion
has not been determined, but it is thought that coolant 2 is not able to overcome the corrosive
effects of very low levels of conventional corrosion inhibitors. It is not known whether this
corrosion will be observed in vehicles with mixed coolants but mixing effects should be studied
in vehicles.
Fleet testing of coolants 1 and 2 showed several interesting effects. There was continued
aggressive behavior toward high lead solder. Tolyltriazole depleted from solution prematurely.
Despite the loss of tolyltriazole from solution, cuprous metal protection remained good.
Finally, a comparison of the a l u m i n u m protection by coolant 2 in the Series 1 and 2 fleet tests
showed differences. This could be attributed to the age of the vehicle, the type of vehicle, or to
the absence of corrosive ions in the Series 2 fleet. All of these have implications in the use of
coolant 2 in the aftermarket where vehicle age, type, and water quality cannot be controlled
by the coolant manufacturer.
Acknowledgments
The author thanks Angela L. Gelsi for preparing all of the coolants and running the labo-
ratory tests described in this paper. I also thank the Applications Engineering and Analytical
Groups of First Brands Corporation for their expert work.
References
[1] Cooper, N. R., Hannigan, H. J., and McCourt, J. C., "A One Thousand Car Assessment of the U.S.
Car Population Cooling Systems," Proceedings of the Second Automotive Corrosion Prevention Con-
verence, P-136, Dearborn, MI, 1983, pp. 121-130.
[2] Barnum, E. R., U.S. Patent No. 2,369,640, 20 Feb. 1945.
[3] Butler, G. and Mercer, A. D., "InhibitorFormulationsfor Engine Coolants," British Corrosion Jour-
nal, Vol. 12, No. 3, 1977, pp. 171-174.
[4] Boreland, W., Dossor, J. M.,and Jaques, G. S., U. S. Patent No. 4,382,008, 3 May 1983.
[5] Perrot, P. H., French Patent No. 2,489,355, 3 May 1982.
[6] Darden, J. W., U. S. Patent No. 4,561,990, 31 Dec. 1985.
[7] Yeakey, E. L., Speranza, G. P., Triebel, C. A., and McCoy, D. R., U. S. Patent No. 4,578,205, 25
March 1986.
[8] Duranleau, R. G., Triebel, C. A., and Lambert, C. L., U. S. Patent No. 4,584,119, 22 April 1986.
[9] Darden, J. W., U. S. Patent No. 4,587,028, 6 May 1986.
[10] Darden, J. W. and McEntire, E. E., U. S. Patent No. 4,592,853, 3 June 1986.
[I1] Darden, J. W., Triebel, C. A., Van Neste, W. A., and Maes, J. P., U. S. Patent No. 4,647,392, 3 March
1987.
[12] Darden, J. W., U. S. Patent No. 4,657,689, 14 April 1987.
[13] Van Neste, W. A., Maes, J. P., Darden, J. W., and Triebel, C. A., U. S. Patent No. 4,759,864.
[14] Van Neste, W. A., Maes, J. P., and Vereeken, R., U. S. Patent No. 4,851,145, 25 July 1989.
[15] Keil, H. S. H., Alfs, H., and Schulze, K., U. S. Patent No. 4,414,125, 8 Nov. 1983.
[16] Matteodo, J. B., Tucoulat, M., and Pascal-Mousselard, M., U. S. Patent No. 4,869.841, 26 Sept.
1989.
[17] Falla, D. J. and Berehtold, P. H., U. S. Patent No. 4,946,616, 7 Aug. 1990.
[18] Darden, J. W., Triebel, C. A., Maes, J. P., and Van Neste, W., "Monoacid/Diacid Combinations as
Corrosion Inhibitors in Antifreeze Formulations," Proceedings of the Worldwide Trends in Engine
Coolants, Cooling System Materials and Testing Conference, SP-811, Detroit, Michigan, 1990, pp.
135-151.
[19] Beard, L. K., Conville, J. J., and Wilson, J. C., "A Demonstration of Aluminum Transport Depo-
sition," Engine Coolant Testing, STP 887, R. E. Beal, Ed., American Society for Testing and Mate-
rials, Philadelphia, 1984, pp. 44-60.
[20] Hudgens, R. D. and Hercamp, R. D., "Test Methods for the Development of Supplemental Addi-
tives for Heavy Duty Diesel Engine Coolants," Engine Coolant Testing, STP 887, R. E. Beal, Ed.,
American Society for Testing and Materials, Philadelphia; 1984, pp. 189-215.
[21] Meloy, G. K., and Derickson, R. M., "Cobratec* TT-100 in Sebacic Acid Engine Coolant Formu-
lations," PMC Specialties Group, April 22, 1987.
DISCUSSION
J. M. Burns ~(discussion question)--How many vehicles were used in each fleet test?
W. C. Mercer (author's response)--In the Series 1 test there were two vehicles used to test
coolants 1 and 2 and five vehicles for coolant 3. In the Series 2 tests, 5 vehicles were used for
coolant 2 and G M 6043M.
J. M. Burns 1 (discussion question)--Were the same vehicles used in more than one test?
W. C. Mercer (author's response)--The Series 1 vehicles were used for more than one test.
The Series 2 vehicles are used once.
J. M. Burns ~(discussion question)--Were the fleet tests run in parallel or series, if the same
vehicles were used?
W. C. Mercer (author's response)--The Series 1 fleet tests were run in series.
J. M. Burns ~ (discussion question)--How were the cooling systems prepared? Was oxalic
acid used as is recommended in ASTM Practice for Testing Engine Coolants in Car and Light
Truck Service (D 2847)?
W. C. Mercer (author's response)--The cooling systems for tests of coolant 1 and 2 were
water washed between tests. Series 2 fleet vehicles were factory filled.
J. M. Burns ~ (discussion question)--Was the lead carboxylate precipitate that was found
positively identified as a product formed by the carboxylic acids used in the formulation, or
could it have been composed of the degradation products of ethylene glycol, such as glycolic
acid, formic acid, or acetic acid, or a cleaning agent such as glycolic acid?
W. C. Mercer (author's response)--It was not conclusively identified as caused by the mono-
acid salt. However, coolants using only diacids do not cause high lead solder corrosion. There-
fore, there are no precipitates seen in the radiators. It is unlikely that the observed product
could be a result of glycol degradation products. F o r this to be the case, ethylene glycol deg-
radation would have to be much higher in the coolants with monoacids than in traditional
coolants. There were no unusual levels of glycol degradation seen with any of the coolants.
J. M. Burns ~ (discussion question)--Were any lead solder alloys, other than the two stan-
dard ASTM test coupon materials, tested to determine if the acids were actually detrimental
to lead in general?
W. C. Mercer (author's response)--No.
J. M. Burns ~ (discussion question)--How were the experimental formulations of those
tested developed? Were they subjected to any type of strategy of formulation?
Texaco Chemical Company, 7114 Lamar Blvd., Austin, TX 78752.

DISCUSSION ON CARBOZYLIC ACIDS AS CORROSION INHIBITORS 59
W. C. Mercer (author's response)--Coolants 1 and 2 were selected for fleet testing based on
the ASTM D 1384 results with coolants A - L and M - X , respectively. The acid levels used in
coolants A - L and M - X cover a wide range of the concentrations available to a coolant for-
mulator. The use of an experimental design like that used for coolants A - X allows the for-
mulator to explore many combinations quickly. The chemist can extrapolate from observed
performance to expected performance within the range tested.
J. M. Burns ~(discussion question)--Do you have any data from compatibility testing with
Japanese engine coolant formulations?
W. C. Mercer (author's response)--Compatibility is assumed based on the absence of
reported problems despite extensive mixing in the aftermarket. We do not have ASTM D 4340
results for combinations other than the ones discussed in the paper.
J. M. Burns ~(discussion question)--What was the standard deviation of the losses reported
on the lead solder coupons?
W. C. Mercer (author's response)--A m a x i m u m standard error (95%) of plus or minus 15%
of the mean weight loss was observed for the high lead solder coupons.
J. M. Burns ~(discussion question)--What was the effect of the cleaning method used for the
lead solder and aluminum coupons on the weight loss?
W. C. Mercer (discussion question)--Since the alternate cleaning procedure removed cor-
rosion product from the high lead solder coupons, the weight loss was higher following this
procedure. This weight loss represents the true weight loss since the layer is a corrosion product
and not a protective layer. There was no effect of this procedure on specimens from tests of
traditional coolants or on virgin specimens.
All aluminum specimens were cleaned by the same procedure. It is not anticipated that the
aluminum cleaning procedure affects specimens from one type of coolant and not another.
J. M. Burns ~(discussion question)--Data were presented that indicated that many consum-
ers neglect their cooling systems. It was stated that as much as 35% o f the U.S. fleet have mod-
erate to heavy rust in the engine blocks and heads while up to 41% of the U.S. fleet have heavy
to moderate sediment in the cooling systems from corrosion products and inhibitor dropout
such as silica gel. In your opinion, would an engine coolant with a longer service life and less
required consumer maintenance reduce these problems and provide the consumer with a
more reliable product?
W. C. Mercer (author's response)--First, there was no indication in the reference that any
of the sediment was due to inhibitor dropout. There were some sources of sediment identified:
core sand, corrosion products, water impurities, stop leak particles, and incompatible
additives.
Use of coolant beyond its service life was only one of the problems identified in that survey.
Others identified included operation with less than 1 in. (25.4 mm) of coolant in the overflow
bucket and coolant concentration well below 50%. Neither of these issues will be addressed by
a coolant with a longer service life. In fact, coolants 1 and 2 will perform worse than standard
coolants under these conditions. This was clearly shown in the Series 1 fleet tests.
There was no indication in the survey that consumers recognize the recommended change
interval and ignore it. The more general attitude is that consumers have the freeze protection
checked. If acceptable, the cooling system is ignored. I do not think that a longer life coolant
would reduce consumer neglect. Education would reduce consumer neglect. Coolant suppliers
and the automotive manufacturers need to convince consumers that corrosion protection is
important. All coolants have a service life. Use beyond that service life will damage engines.
D. Miller 2 (discussion question)--Ifthe formulations were not properly optimized, then this
would explain the observation of high aluminum weight losses. Were the formulations opti-
mized? If so, by what method were they optimized?
2 Texaco Chemical Company, P.O. Box 15730, Austin, TX 78761.

W. C. Mercer (author's response)--Coolants 1 and 2 were selected for fleet testing based on
A S T M D 1384 results with coolants A - L and M - X , respectively. The acid levels used in cool-
ants A - L and M - X cover a wide range of the concentrations available to a coolant formulator.
They should give an idea of the chances for optimization of these fluids. I do not think these
coolants will provide the balance of properties required for an optimized coolant.
D. Miller 2 (author's response)--Was synergism observed between the monoacid and the
diacid?
W. C. Mercer (author's response)--Coolants D, H, and K had monoacid/diacid ratios iden-
tified in the literature as synergistic. There was no synergism observed with these coolants in
the ASTM D 1384 testing. There was a potential interaction between the 2-ethylhexanoic acid
and the 1,10-decanedioic acid observed in coolants M - X . The net effect of the interaction was
to make aluminum corrosion higher than expected. This could be described as a negative
synergism.
D. Miller 2 (discussion question)--How many cars were included in the fleet test? What are
the statistics (range and standard deviation) for the data obtained from the fleet tests?
W. C. Mercer (author's response)--In the Series 1 test there were two vehicles used to test
coolants 1 and 2 and five vehicles for coolant 3. In the Series 2 tests, 5 vehicles for coolant 2
and G M 6043M. Data from coolant 2 tests are presented as examples of the data obtained in
fleet tests (Table 1).
D. Miller 2 (discussion question)--When comparing tolyitriazole (TT) depletion from the
carboxylic acid coolant to G M 6043M, it should be pointed out the G M 6043M contains mer-
captobenzothiazole (MBT). Since MBT provides protection to essentially the same metals as
TT, the presence of MBT would likely affect the depletion rate of TT. Furthermore, the initial
TT concentrations in the two coolants are very dissimilar. The carboxylic acid coolant starts
with twice the concentration of T T as G M 6043M. An initially rapid depletion of high TT
concentrations is commonly observed in many coolants. Any claims regarding suspected cop-
per and brass protection for these metals was not included in the presentation.
W. C. Mercer (author's response)--It has been our experience that tolyltriazole in tradi-
tional North American coolants remains at very high percentages of the initial level. This is
true whether mercaptobenzothiazole is present or not. We have not seen traditional North
American coolants with high levels of TT show an initial rapid depletion. The rapid loss of TT
from solution is a property of the carboxylic acid coolants. As stated in the text, the use of
cuprous metal protection has been adequate so far. It is still not clear that coolants with very
low levels of cuprous metal inhibitors will protect these metals in the long term. This question
has yet to be answered.
D. Miller 2 (discussion question)--If the carboxylic acid coolants are corrosive to lead
through complexation, then why do they protect 70/30 solder? ASTM D 1384 tests that we
have performed show that 70/30 solder is still protected by carboxylic acids even when the
TABLE 1--Data from coolant 2 test.

Parameter Solder Copper Solder Brass Steel Iron Aluminum
SERIES I (48 000 KM)
Mean 727 8 2 7 6 -- 2 54
High 785 12 6 8 20 --4 71
Low 685 4 --2 6 --3 -- 1 45
SERmS 2 (38 000 KM)
Mean 865 48.3 0.6 17.6 0.4 --0.6 3.3
Std error (95%) 117 5.0 7.4 2.5 0.6 1.0 1.2
DISCUSSION ON CARBOZYLIC ACIDS AS CORROSION INHIBITORS 61
high lead solder coupon is not included. ASTM D 2570 simulated service testing of carboxylic
acid coolant, meanwhile, even shows excellent protection of high lead solder. Therefore, the
conclusion that carboxylic acids attack lead appears to be incorrect. Could there be another
reason for high lead solder corrosion?
W. C. Mercer (author's response)--As shown in the text, octanoic acid corrodes high lead
solder and ASTM (70/30) solder when both are present in the specimen bundle. 2-ethylhex-
anoic acid corrodes high lead solder. In our laboratories, high lead solder corrosion with cool-
ants 1 and 2 is very high in all tests except engine dynamometer tests. In the dynamometer
tests, high lead solder corrosion is moderate and ASTM solder corrosion is very high. Com-
plexation is the most likely reason as monoacid salts of lead have been identified in fleet vehicle
radiators.
D. Miller2 (discussion question)--Although you have stated that existing coolant formula-
tions are fully compatible with each other, no data to support this were presented. What are
the ASTM D 4340 results for a Japanese nonsilicate coolant combined with an American sil-
icate coolant? What about other types of combinations? Meanwhile, carboxylic acid formu-
lations have been used in Europe for more than two years without any reported compatibility
problems.
W. C. Mercer2 (author's response)--Compatibility is assumed based on the absence of
reported problems despite extensive mixing in the aftermarket. We do not have ASTM D 4340
results for combinations other than the ones discussed in the paper. Since the carboxylic acid
formulations currently used in Europe are very similar to the coolants already in use, it is not
surprising there have been no problems reported.
E. McWilliams 3 (discussion question)--Was analysis of voluminous ppt in heat done by
aluminum test ASTM 4340? Was analysis of corrosion blacked during testing done?
W. C. Mercer (author's response)--We have generally attributed black aluminum corrosion
to oxides, but have seen high carbon levels with auger analysis.
D. Berg4 (discussion question)--What analytical method was used to determine the concen-
trations of tolyltriazole (TTZ) in your work?
W. C. Mercer (author's response)--We use a Dionex ion chromatograph with an ALC-18
column.
J. A. Lima 5(discussion question)--Bill, you speak as though "long life" coolants are already
available using silicone/silicate technology. If this is true, could you comment on any advan-
tages the aromatic carboxylates, as found in many current European formulations, may have
over aliphatic carboxylates, or over current North American coolants?
W. C. Mercer (author's response)--We have not evaluated the performance of coolants that
use aromatic carboxylic acids.
J. P. Maes 6 (discussion question)--Dr. Mercer, can you please comment on the size of the
fleet, namely, the numbers of vehicles used in your fleet test?
W. C. Mercer (author's response)--The fleet tests were run in 2 Plymouth Voyagers and 5
Pontiac Grand Prix for each coolant.
J. P. M a e s 6 (written discussion)--Your presentation alleges concerns about the ability of
monobasic/dibasic acid coolant technology to protect aluminum and soldered copper/brass
radiators. In Europe, optimized monobasic/dibasic acid coolants that contain the acids men-
tioned in your presentation have been the subject of qualification testing by a major car man-
ufacturer. The coolants were fleet tested over 100 vehicles, for up to 100 000 miles (160 930
3 Caterpillar, 100 N.E. Adams St., Peoria, IL 61614.

4 Mobay Corporation, Mobay Road, Building 14, Pittsburgh, PA 15205.
5 Houghton Chemical Corporation, Allston, MA.
6 Texaco Research and Development--Ghent, John Kennedylann 31, B9042 Ghent, Belgium.
km) and time periods of four years. Included in this fleet were a number of vehicles fitted with
copper/brass radiators. The others were fitted with aluminum radiators. At no stage in the fleet
testing program were there any reported problems related to radiator corrosion. Likewise, all
radiators examined at the end of test were totally compatible with the manufacturers' sign-off
criteria. In summary, monobasic/dibasic acid coolants have been qualified by a major Euro-
pean car manufacturer after exhaustive laboratory testing and four-year fleet testing with no
concerns on aluminum, solder, or copper/brass corrosion and no reported problems over two
years of in-service use throughout all plants of this manufacturer.
Norman C. Adamowicz' and Daniel F. Falla 2
Fleet Test Evaluation of Engine Coolants

Using Sebacic Acid Inhibitor Technology
REFERENCE: Adamowicz, N. C. and Falla, D. F., "Fleet Test Evaluation of Engine Coolants
Using Sebacic Acid Inhibitor Technology," Engine Coolant Testing. Third Volume, ASTM STP
1192, R. E. Beal, Ed., American Society for Testing and Materials, Philadelphia, 1993, pp. 63-
82.
ABSTRACT: A three-year, twelve-car municipalfleettest was conducted to compare the relative

merits of an extended life sebacic acid coolant and current high silicate alkaline phosphate cool-
ant used in North America. The test was conducted under rigidly controlled conditions in accor-
dance with ASTM Practice for Testing Engine Coolants in Car and Light Truck Service(D 2847).
This paper will show the ( 1) depletion rates for chemical inhibitors used in the coolants and
(2) actual photographs of cooling system component parts removed from the vehiclesat the ter-
mination of this test,
KEYWORDS: fleet, sebacic, phosphate, corrosion, aluminum, inhibitors, silicate, depletion,

radiators
Background
Current high silicate alkaline phosphate engine coolants were developed in the early 1980s
to especially protect new a l u m i n u m cooling system components against corrosion. These par-
ticular components and the type of corrosion are as follows:
1. Cast aluminum:
9 Cylinder h e a d - - h o t surface corrosion.
9 Water p u m p - - c a v i t a t i o n corrosion.
9 Thermostat housing--cavitation corrosion.
2. Wrought aluminum:
9 Radiator--hot surface corrosion deposits and tube pitting.
9 Heater core--header plate surface corrosion and header plate crevice corrosion.
Continued protection for all other metals previously used in the cooling system is also
required. The newly developed inhibitor package was successful in preventing all metallic cor-
rosion in vehicle cooling systems.
Rationale for a World Wide Engine Coolant

There are numerous styles of chemical inhibitor packages used successfully with ethylene
glycol world wide. However, the mixing of different styles may accelerate the long-term cor-
Design engineer, Ford Motor Company, 21500 Oakwood Blvd., Dearborn, M148121.
2 Development chemist, Dow Chemical Canada, Inc., Sarnia, Ontario, Canada.
63
rosion rate of metals. With the advent of increased trans-oceanic shipment of cars and trucks
it became important to find a c o m m o n coolant inhibitor technology.
Sebacic acid corrosion inhibitor technology (designated as AL-39) [1 ] was developed by the
British National Physical Laboratory [2] for use by the British Military in a variety of vehicles
world wide using a wide range of water quality.
Preliminary confidential test results conducted by the British National Physical Laboratory,
the British Military, and by Ford of England, indicated that this new technology had the fol-
lowing potential advantages over high silicate alkaline phosphate inhibitors:
9 Reduced hot surface corrosion of aluminum cylinder heads.

9 Reduced heat transfer corrosion deposits in radiators.
9 Reduced engine hot spot scale deposits that influence heat transfer efficiency.
9 Low depletion rate of inhibitors.
9 Extended coolant change intervals to lessen the environmental impact and reduce the
cost o f ownership.
Development Issues
Sebacic acid is made by processing the oils extracted from the castor bean. It is a commodity
material controlled by the Brazilian government and other third world countries. As such, the
4.5% concentration defined by the AL-39 formulation is too expensive for commercial use.
Formulations were, therefore, developed that substituted other dibasic or monobasic acids for
a portion of the sebacic acid. Additives such as these are referred to as extenders or extender
acids.
As a further cost reduction it was decided to reduce the number of component chemicals in
the inhibitor package to the fewest number required to provide total metallic corrosion pro-
tection. The sequence of tests conducted below were conducted satisfactorily and confirmed
that these changes were viable.
9 ASTM D 1384, Test Method for Corrosion Test for Engine Coolants in Glassware.
9 ASTM D 4340, Test Method for Corrosion of Cast A l u m i n u m Alloys in Engine Coolants
Under Heat Recycling Conditions.
9 Ford Motor Company Simulated Service Corrosion Test for Aluminum Radiator.
9 Ford Laboratory Test Method BL 2-2, Engine Dynamometer Test.
The inhibitor package was then made up as follows:
9 Total acids (sebacic acid and dodecane dioic acid).

9 Tolyltriazole (TTZ).
9 Silicate.
9 Buffer.
Accordingly, this is a simpler formulation than trying to balance the seven or eight com-
ponents of a typical alkaline phosphate coolant.
Fleet Testing
Fleet testing is diffcult to do (at best) for these reasons:
9 Locating a facility that agrees to test parts at no charge to you and allows you to disrupt
their work week.
MAES AND VAN oE VEN ON FLEET TEST EVALUATION 65
9 The knowledge that you may lose 40% o f your test cars to accidents.
9 Campaigned vehicles that further eliminates test cars.
9 Accidental contamination with other coolants.
9 Re-assignment of vehicles to other locations.
9 Stolen cars.
9 Change in fleet management.
9 Lost coolant samples.
Before this test, five fleets containing a total of 173 cars were lost for the reasons above. Of
these 173 cars, only three cars survived uncontaminated to 60 000 miles (96 558 km).
The fleet test discussed in the paper was headquartered in Martinez, CA. Vehicles were 1986
1.9L Escorts owned and maintained by the municipality of Contra Costa County. This county
is northeast of San Francisco Bay, has an arid climate, and is moderately hilly. Each car is
assigned to one building inspector who may be required to access remote locations within the
county. Consequently, these vehicles undergo severe service. As the vehicles were not new
when the test started, the following new parts were installed in accordance with ASTM Practice
for Testing Engine Coolants in Car and Light Truck Service (D 2847):
9 Cylinder head.
9 Cylinder head gasket and bolts.
9 Water pump.
9 Timing belt.
9 Radiator.
9 Radiator cap.
9 Top radiator hose.
9 Bottom radiator hose.
9 Radiator hose clamps.
Full flow brass pots capable of holding two ASTM metal coupon specimen bundles were
also installed in the top hose of four vehicles: three pots in test coolant vehicles and one in an
original equipment manufacturer (OEM) coolant vehicle.
All cooling systems were flushed with a commercial power flush system before filling with a
50% solution of either test coolant or OEM coolant. Small initial samples were taken only to
measure glycol content as a function of refractive index. The first check point was at one
month for all physical and chemical properties. Thereafter the fleet was visited twice per year
whenever it was convenient or appropriate for the Test Fleet, Ford Motor Company, and Dow
Chemical Company. At those visits the following evaluations were made:
9 Visual check of coolant volume in the radiator.

9 Sampled coolant and topped-up the coolant level.
9 Checked coolant volume stored in vehicle trunk.
9 Pressure tested radiator cap and replaced if necessary.
9 Inspected cooling system for leaks.
9 Changed coupon bundles as appropriate.
This fleet test was terminated after three years service. At that time cylinder heads, radiators,
and water pumps were retrieved from six high mileage vehicles (Table 1, Car Nos. 142, 143,
144, 146, 147, and 152).
The cylinder heads were sectioned to examine the coolant jacket passages. Radiators were
wind tunnel heat rejection tested to determine any significant loss in heat rejection capacity
and then sectioned to examine the tubes, header plates, and gasket troughs for corrosion.
TABLE 1--Fleet study data summary (normalized to % inhibitor level).
Sebacic Dodec
Date Miles DKP (IC) Moly TTZ SI RA pH Acid Acid
CAR 142
870522 0 100.0 100.0 100.0 100.0 11.00 10.50 0.0 0.0
870615 1 176 89.2 75.0 89.2 52.0 6.45 9.60 0.0 0.0
880609 23 223 43.9 43.9 5.09 8.30 0.0 0.0
880928 30 298 87.7 76.7 87.7 19.3 4.88 8.00 0.0 0.0
890221 41 062 93.8 93.8 4.50 7.80 0.0 0.0
890724 53 965 61.5 61.5 4.24 7.60 0.0 0.0
67 410 74.6 74.6 0.0 0.0
900521 72 085 61.5 76.7 61.5 14.0 0.0 0.0
CAR 143
870522 0 0.0 0.0 100.0 100.0 6.00 8.40 100.0 100.0
870615 1 408 0.0 0.0 109.3 45.0 4.68 8.10 140.0 97.9
880610 16 457 0.0 0.0 70.7 3.10 7.30 111.4 89.0
880928 21 216 0.0 0.0 70.0 32.0 3.10 7.10 102.4 91.8
890221 26 168 0.0 0.0 68.0 4.90 7.10 112.9 97.0
890724 33 556 0.0 0.0 62.7 4.40 7.00 109.7 97.0
47 086 0.0 0.0 68.0 113.7 106.6
900521 58 168 0.0 0.0 62.7 18.7 5.40 6.60 105.2 95.6
CAR 144
870522 0 0.0 0.0 100.0 100.0 6.00 8.40 100.0 100.0
870615 1 227 0.0 0.0 103.3 50.7 4.62 8.10 124.3 97.9
880610 19 738 0.0 0.0 58.7 3.00 7.10 100.6 83.6
880928 26 111 0.0 0.0 56.0 22.7 3.07 7.10 101.9 86.3
890221 33 086 0.0 0.0 66.0 4.70 7.00 112.4 98.3
890724 41 397 0.0 0.0 60.0 4.20 7.00 107.6 96.9
52 666 0.0 0.0 64.0 5.30 6.60 126.9 113.6
900130 10.7
CAR 145
870522 0 0.0 0.0 100.0 100.0 6.00 8.40 100.0 100.0
870615 1130 0.0 0.0 104.0 42.7 4.62 8.10 142.6 98.4
880610 9 524 0.0 0.0 74.0 3.08 7.50 132.5 91.5
880928 11 444 0.0 0.0
890221 13 410 0.0 0.0 72.0 4.40 7.30 I 11.5 96.9
890724 15 760 0.0 0.0 72.0 4.30 7.20 113.2 95.8
18 749 0.0 0.0 84.0 13.2 111.5 112.6
900521 25 470 0.0 0.0 78.0 12.0 5.90 6.90 109.0 107.0
NOTE: 1 mile = 1.6 km.
W a t e r p u m p s were e x a m i n e d for corrosion damage using an optical microscope having a mag-

nification o f X30. The coolant was analyzed using instrumental and wet methods well k n o w n
in the industry.
The available analytical data were normalized for inhibitor levels and are presented in
Tables 2, 3, and 4. This n o r m a l i z e d data were also graphed as a function of mileage for all cars
(Figs. 1 to 7).
Discussion: Inhibitor Depletion

I f we first e x a m i n e the two O E M coolant control vehicles, two discrepancies are noted. Data
from Vehicle 142 (Fig. 1) is very erratic for T T Z throughout the test perhaps because o f a n a -
lyrical error. T h e silicate does not deplete as m u c h as expected during the initial stage but does
level o f f a t approximately 20%. The dipotassium phosphate (DKP) depletion is near normal.
There is no accounting for these behaviors except to speculate that perhaps this vehicle
M A E S A N D V A N DE V E N O N F L E E T T E S T E V A L U A T I O N 67
TABLE 2--Fleet study raw data summary (normalized to % inhibitor level).
Sebacic Dodec
Date Miles DKP (IC) Moly TTZ SI RA pH Acid Acid
CAR 146
870522 0 0.0 0,0 100.0 100.0 6.00 8.40 100.0 100.0
870615 1 362 0,0 0.0 97.3 48.0 4.70 8.10 107.4 102.1
880610
880928
14 900
27 993
0.0
0.0
0.0
0.0 5513 2913 316i 716(J
...
97.5 8918
890221 35 516 0.0 0.0 66.0 ... 4.50 7.00 97.5 88.0
890724 43447 0.0 0.0 60.7 ... 4.30 7.00 103.1 95.0
900521 55 009 0.0 0,0 58.7 ... 5,40 6.50 86.7 86.8
CAR 147
870522 0 0.0 0.0 100.0 100.0 6.00 8.40 100.0 100.0
870615 1 063 0.0 0.0 110.0 44.0 4.70 8.10 104.3 101.8
880610 15 589 0.0 0.0 72.4 3.20 7.30 112.9 93.9
880928 19 051 0.0 0.0 71.1 161() 3.20 7.20 106.7 91.1
890221 24 426 0.0 0.0 68.7 ... 4.60 7.10 101.4 97.7
890724 30 924 0.0 0.0 63.0 4.90 7.10 106.4 99.9
900521 35 524 0.0 0.0 74.5 1417 118.6 114.9
... 40 213 0.0 0.0 64.5 ... 4180 617() 127.4 98.1
CAR 148
870522 0 100.0 100.0 100.0 100.0 11.00 10.50 0.0 0.0
870615 1 399 99.2 76.7 46.2 22.7 6.30 9,20 0.0 0.0
880610 12 767 91.9 49.5 5.20 8.70 0.0 0,0
880928 19 094 92.7 7817 41.1 2~114 4.80 8.30 0.0 0.0
890221 26 780 83.1 ... 46.0 ... 4.40 8.10 0.0 0.0
890724 34 796 89.0 32.9 4.00 7.70 0.0 0.0
900521 44 181 88.2 6216 31.4 1714 3.90 7.30 0.0 0.0
CAR 149
870522 0 0.0 0.0 100.0 100.0 6.00 8.40 100.0 100.0
870615 932 0.0 0.0 96.0 36.0 4.90 8.20 99.7 97.1
880610 15 375 0.0 0.0 66.0 2.90 7.10 117.8 93.7
880928 19 085 0.0 0.0 64.2 3417 2.90 7.00 103.5 87.6
890221 23 822 0.0 0.0 80.4 ... 4.10 6.90 100.9 92.6
890724 29 887 0.0 0.0 63.1 3.70 6.80 97.6 92.6
900130 37 346 0.0 0.0 63.1 30[7 104.1 121.6
900516 41 882 0.0 0.0 67.8 28.0 416() 61z~(J 97.0 101.6
NOTE: | mile = 1.6 kin.
lOOt Inhibitor Level (so qL of Initial oono,)
._. DKP
80 xN~Moly
60 ~ ~TTZ
40
2O
--~ 81
0 i i
0 20 40 60 80
Mllee (Thoulands)
FIG. 1--Inhibitor depletion versus mileage for Car 0142.
T A B L E 3--Fleet study raw data summat T (normalized to % inhibitor level).
Sebacic Dodec
Date Miles D K P (IC) Moly TTZ SI RA pH Acid Acid
CAR 150
870522 0 0.0 0.0 100.0 100.0 6,00 8.40 100.0 100.0
870615 715 0.0 0.0 103.2 41.3 4,80 8.10 131.7 113.6
880610 6 352 0.0 0.0 75.8 3,20 7.40 103.7 97.1
880928 11 750 0.0 0.0 60.9 2~)13 3.00 7.10 106.4 97.0
890221 17 989 0.0 0.0 69.4 ... 4.40 7.00 117.3 94.3
890724 26 820 0,0 0,0 57.8 5.00 7.00 1t9.7 96.5
32 572 0.0 0,0 67.5 1817 4.80 6.60 106.9 107.6
CAR 151
870522 0 0.0 0.0 100.0 100.0 6.00 8.40 100.0 100.0
870615 957 0.0 0,0 102.7 49.3 4.80 8.20 133.9 95.1
880610 19 750 0.0 0.0 61.8 ... 3,30 7.20 114.7 90.5
880928 25 152 0.0 0.0 60.2 16.4 3.40 7,10 103.2 97,5
890221 31 294 0,0 0,0 60.6 ... 5.60 7.10 112.5 90.8
890724 39 576 0.0 0.0 55.5 ,.. 4.70 7.00 111.0 88.0
47 611 0.0 0.0 51.6 10.3 5.20 6.60 98.4 94.9
CAR 152
870522 0 0.0 0.0 100.0 100.0 6.00 8.40 100.0 100.0
870615 711 0.0 0.0 103.4 33.3 4.90 8.20 133.8 96.1
880610 18 706 0.0 0.0 58.3 ... 3.10 7.10 105.7 89.6
880928 25 362 0.0 0.0 55.8 26.7 3.00 7.00 102.0 90.0
890221 31 103 0.0 0.0 54.3 ... 4.40 6.90 125.0 90.7
890724 39 551 0.0 0.0 54.3 ... 4.20 6.90 113.3 90.7
... 49 831 0.0 0.0 59.8 19.6 114.9 110.0
... 54 186 0.0 0.0 58.1 ... 41i6 6146 97.3 90.1
CAR 153
870522 0 0.0 0.0 100.0 100.0 6.00 8.40 100.0 100.0
870615 1 496 0.0 0.0 ... 45.3 4.80 8.10 . . . . . .
880610 2 757 0.0 0.0 . . . . . . 3.30 7.40 . . . . . .
880928 24 229 0.0 0.0 14.7 3.10 7.10
890221 31 950 0.0 0.0 5513 .., 4.60 7.10 1121i 9313
890724 34 670 0,0 0,0 55.3 4.80 7.10 112.9 93.3
... 43 301 0.0 0.0 60.8 2()16 ...... 96.7 100.5
NOTE: 1 mile = 1.6 km.
;Ahlbltor LOvel (ae 9 Of I n l t l m l o o n o . )

140
60
40 ~
81
20 -~
0 ~ l I I I
0 10 20 30 40 50 60
Mile ,b (Thouoande)
FIG. 2--1nhibitor depletion versus mileage for Car 0143.
TABLE 4--Corrosion data from coupon bundles.
Miles Copper Solder Brass Steel Cast Iron Aluminum
CAR 142
6889 l 1.0 1.6 9.7 +0.4 + 1.6 3.9
23223 1.0 3.7 2.4 + 1.1 + 1.9 +0.1
30298 3.5 3.3 6.4 + 3.1 + 1.5 2.7
60331 1.4 0.7 1.1 +0.9 + 1.4 +0.1
73234 1.6 6.4 3.0 +2.5 +5.2 1.2
91354 3.9 5.9 4.3 +6.2 +3.8 +0.6
CAR 143
5832 1.1 1.4 0.8 +0.5 +0.5 0.4
16457 +0.4 0.2 0.3 + 1.9 +3.0 + 1.8
21216 4.4 3.4 3.6 0.4 0.0 +2.4
47570 1.3 0.0 1.4 +0.7 1.2 0.7
54958 1.7 8.5 1.9 + 1.7 +4.3 0.2
79570 5.8 4.6 3.6 1.1 2.3 2.2
CAR I45
5386 1.7 1.6 0.6 +0.5 + 1.1 0.3
9524 + 1.0 2.2 + 1.4 + 1.5 +2.3 +2.6
11444 3.4 3.3 2.8 0.6 0.2 3.5
32790 0.8 4.2 0.4 +0.2 1.7 0.6
35114 1.3 6.1 0.3 1.9 +3.4 0.2
44850 3.8 6.8 2.6 2.0 2.9 0.7
CAR 147
5433 +0.7 1.6 1.1 +0.7 +0.9 0.2
15569 +0.9 0.7 + 1.3 + 1.9 +2.6 + 1.9
19051 3.6 2.4 3.1 0.7 1.2 2.9
42897 1.3 1.9 4.1 +2.8 7.0 0.8
49395 1.2 18.5 1.1 1.1 +4.9 1.0
58648 3.3 9.5 2.2 3.2 1.9 1.7
OVER DURATION OF FLEET STUDY
CAR 142
72085 25.6 18.7 12.7 1.0 +0.3 4.2
CAR 143
58168 6.6 5.6 8.9 1.3 1.0 4.3
CAR 145
25470 3.0 2.0 3.4 ! .2 0.0 0.9
CAR 147
40213 3.1 2.5 4.5 1.2 0.0 2.3
r e q u i r e d a n a d d i t i o n o f c o o l a n t d u r i n g t h e initial stage. T h i s was n o t sufficient to disqualify

t h e d a t a for Vehicle 142 as n o p r o b l e m s were e n c o u n t e r e d w i t h it.
D a t a f r o m V e h i c l e 148 (Fig. 6) are a n e a r t e x t b o o k e x a m p l e o f i n h i b i t o r depletion:
9 R a p i d silicate d e p l e t i o n d u r i n g t h e first 2000 miles (3219 km).

9 A 10% d e p l e t i o n o f d i p o t a s s i u m p h o s p h a t e d u r i n g the course o f this test.
9 T T Z levels off a n d r e m a i n s relatively c o n s t a n t .
Let us n o t e here t h a t w h e n t h e silicate depletes in this style o f coolant, it does n o t go away

[3]. Silicate is k n o w n to f o r m a m o l e c u l a r film with a l u m i n u m a n d is r e m o v e d f r o m solution
b e c a u s e it h a s d o n e w h a t it was s u p p o s e d to do. T h e r e m a i n d e r o f the silicate acts to heal a n y
d i s c o n t i n u i t i e s in t h e film t h a t m a y occur.
Inhlbllor Level (ll 9 of Inlllal oono.)

140
8eb1r
120
Dodeoane
100(
80
60 ~" TTZ
2O
I I I l I
0
0 10 20 30 40 50 60
Mileo (Thouoandl)
lahlbllor Level (IS % of Iltltlll oono.)
1004
80 Do.. ::
60
40
20
t I Z I ]
0
0 10 20 30 40 50 60
Miles (Thousands)

MAES AND VAN DE VEN ON FLEET TEST EVALUATION 71
Inhibitor Le/qll (ae % of Initial oono.)

S e b s c i o
120
1001
80
60
40
20
(3 81
0 I ,. I I I
0 10 20 30 40 50
Miles (Thousands)
Inhibitor Level (as qL of Initial aorta.)

100c
~ . L~..~ ~ . ~ DKP
80
I~ Moly
60
TTZ
20 8i
O i j i i _
0 10 20 30 40 50
Mllee (Thousands)
72 ENGINECOOLANTTESTING:THIRDVOLUME
Inh|bllor Level (all % O! In|t|l! oono.)
", o~0 ,
60 ~ , ~ T T Z
4O ] ~__~_._.___
-El 81
I [ I I
0 lo 20 30 40 50 60
Mllell (Thousands)
As to the inhibitor analysis of the test coolant, the author has previously seen examples of a
total acids content higher than 100%. What we can infer from this is that the total acids do not
deplete significantly during vehicle life. Silicate does again deplete rapidly during the 1st 2000
miles (3219 km) to about 40 to 50% and gradually thereafter. TTZ depletes to 60 to 70% and
remains constant thereafter.
Discussion: Metal Coupon Specimen Weight Loss

Let us now judge the metal coupon weight losses against the performance requirements con-
tained in ASTM Specification for Ethylene Glycol Base Engine Coolant, for Automobile and
Light Duty Service (D 3306). According to mileage, there is no practical difference between
FIG. 8--Cylinder head cooling jacket of Vehicle 0142.

FIG. 9--C.vlinder head coolingjacket ~/ Vehicle 0143.
the test coolant and the OEM coolant for a loss of aluminum, cast iron, and steel. For solder
there is a minor issue that the test coolant may not measure up. For the OEM coolant some
slight concern in weight loss occurred during startup for copper and brass but never recurred.
With the trend to aluminum heat exchangers that do not use solder, copper, or brass, there is
no issue. For the long-term duration, metal coupon weight losses were not considered signif-
icant for either coolant.
Cylinder Head Examination

Photographs 8 through 13 show the interior coolant jacket surfaces of the vehicles denoted
therein. In Fig. 8 a tan off-white coating covers most of the surface area; this is believed typical
of high silicate coolants currently in use. Photos 9 through 13 depict a dark blue black film in
the same areas for the test coolant.
FIG. l O--Cylinder head coolingjacket of Vehicle 0144.

FIG. 11--Cylinder head coolingjacket of Vehicle 0146.

TABLE 5--Chemical analysis of the coatings inside two cylinder

head cooling jackets.
Component Vehicle 142 Vehicle 146
Aluminum 20 20
Silicon 5 5
Iron 5 0.2
Zinc 3 0.5
Calcium 2 2
Copper 2 0.5
Phosphorus 5 0.2
Potassium 1 0.1
Magnesium 1 0.1
Sodium 0.5 0.2
Manganese 0.5 0.01
Boron 0.2 0.2
Lead 1 0.5
Tin 0.5 0.1
NOTE: Elemental Composition of Deposits, approximately % by
weight (Emission Spectrographic Analysis of Insolubles).
Emission spectrographic analysis (Table 5) of the coatings from vehicle 0142 and 0146 were
composed primarily of aluminum and silicon. Metallographic analysis indicates that the tan
off-white coating is particulate in form and unevenly deposited on the surface with a maxi-
m u m thickness of 0.006 mm. The blue black coating adheres to the surface like a film and
extends below the surface with a m a x i m u m thickness of 0.002 mm. There is no difference in
the microstructure of both cylinder heads in all regions: surface to core in an area with deposit,
as well as surface to core in an area without deposit.
Radiator Examination
Figures 14 through 19 depict the interior surfaces of the aluminum radiator cross-flow tubes.
All surfaces show a yellow deposit/film, which is likely caused by the presence of TTZ as a
corrosion inhibitor. In no case was there any significant deposit worth measuring. The deposits
from the OEM coolant appeared heavier. No perforation was noted. Figures 20 through 25
depict the inside surfaces of the aluminum radiator header plates. Again the deposit from
OEM coolant appeared heavier. The remaining radiator heat rejection capacity was measured
by Climate Control Division of Ford Motor Company and found to above 90% for all
radiators.
Water Pump Examination

Visual examination using a X30 optical microscope showed no evidence o f erosion or cav-
itation on interior surfaces. The interior surfaces all evidenced protective coatings that
appeared to be of various thickness; pumps from # 142, 147, and 152 were thought be to
thicker.
Summary
The intent of fleet testing is to evaluate the effect of customer usage upon the durability of
a part or a material in comparison to some standard. We believe that this was a good test, well
FIG. 14--Radialor cros.~'llowtubes J?om l "ehicle 0142.
FIG. 15--Radiator crossflow tubes from Vehicle 0143.

FIG. 16--Radiator crossflow tubes from Vehicle 0144.
FIG. 17--Radiator cross,flow tubes lkonz Vehicle 0146.

FIG. 18--Radiator cros,sflowtubes /rom Vehicle 0147.
FIG. 19--Radiator crossfiowtubes from Vehicle 0152.

FIG. 20--Radiator header plates from Vehicle 0142.

FIG. 23--Radiator header platesfrom Vehicle 0146.
controlled over a three-year period. What we learned was not unexpected and confirmed prior
beliefs.
Certainly it showed that low silicate sebacic acid coolants do not suffer as rapid a depletion
rate for chemical inhibitors as do high silicate alkaline phosphate coolants. We do not know
if the test coolant has a longer life because the coolants were not run to failure. We can say that
the test coolant is no worse than high silicate alkaline phosphate coolants. However, this new
style of coolant does not offer any significant advantages for the consumer over current North
American coolants.
Coolant change intervals vary greatly with the needs of the vehicle manufacturer. Mileage
accumulation of the two cars that used standard OEM fill coolant far exceeded our recom-
mended 3 year/30 000 (3 year/48 279 km) mile interval. In a modern car with a well-main-
tained cooling system, current OEM factory fill corrosion protection can be extended far
beyond the current coolant change intervals.
In cars with a l u m i n u m radiators and heater cores, no lead solder contamination of the

engine coolant occurs. Coupled with an extended coolant change interval the environmental
push for recycled coolant may be delayed until proper re-inhibitors can be developed.
One o f the issues we have not addressed in this paper is cost. At the time this fleet test began,
it was estimated that the added cost for OEM sebacic acid inhibitor technology would be 10
to 15%. When the estimate became a reality in 1989, the cost increase for Ford of North Amer-
ica to replace OEM alkaline phosphate coolants with OEM sebacic acid coolants was 26%. If
the actual dollar cost increase due to the 1989 sebacic acid coolant use in Ford of England is
added to today's Ford o f North America OEM coolant price, the cost increase becomes 66%.
In today's economics, such a significant cost increase is not acceptable for a product that offers
no distinct advantage.
Conclusions
1. Low silicate sebacic acid coolants do not offer any significant advantages for the consumer
over current North American coolants.
2. In a modern car with a well-maintained cooling system, current North American and
OEM factory fill coolant corrosion protection can be extended far beyond previous
expectations.
3. In modern cars with aluminum radiators and heater cores, no lead contamination of the
engine coolant occurs, This may reduce the eminent need/cry for recycled coolant until such
time as proper re-inhibitors can be developed.
References
[1] British Specification TS 10177, Antifreeze Inhibited Ethanediol, AL-39, The Director of Materials
Quality Assurance Headquarters Building, Royal Arsenal East, London SEI8 6TD.
[2] Butler, G. and Mercer, A. D., "Inhibitor Formulations for Engine Coolants," British Corrosion Jour-
nal, Vol. 12, No. 3, pp. 171-174.
[3] Beard, L. K., Jr., Conville, J. J., and Wilson, J. C., "A Demonstration of Aluminum Transport Depo-
sition," STP 887, Engine Coolant Testing: Second Symposium, American Society for Testing and
Materials, Philadelphia, 1986, pp. 44-60.
STP1192-EB/May 1993
DISCUSSION
Joseph A. Lima ~(written discussion)--I agree that use of aluminum radiators is preferable
and will reduce the amount of lead in spent coolant, but what about the continued use of cop-
per/brass heater cores? When do you think they will be eliminated?
N. C. Adamowicz and Daniel F. Falla (authors' response)--Ford Motor Company Climate
Control Division planning volumes for both OEM aluminum radiators and heater cores
exceeds 95% by 1998. There will probably always be a small need for copper/brass heat
exchangers.
Houghton Chemical Corporation, Allston, MA.
Copyright9 1993 by ASTM International www.astm.org

Roy E. Beal and ReJki El-Bourini 2
Corrosion Testing for Aluminum Alloy

Selection in Automotive Radiators
REFERENCE: Beal, R. E. and EI-Bourini, R., "Corrosion Testing for Aluminum Alloy Selection
in Automotive Radiators," Engine Coolant Testing." Third Volume, A S T M STP 1192, R. E. Beal,
Ed., American Society for Testing and Materials, Philadelphia, 1993, pp. 83-106.
ABSTRACT: Durability of aluminum alloy automotive radiators in service depends on the alloy
selected and the expected engine coolant environment. The challenge is to find testing methods
that correlate with service experiences that allow for comparative evaluation of competing mate-
rials without resort to long-term vehicle exposure at initial selection stages. Design, manufacture,
and processing affect radiator performance and must be taken into account in corrosion testing.
Aluminum alloys AA3003 and AA7072 are used to illustrate the test program.
Electrochemical corrosion test methods were used for screening purposes in alloy selection
and corrosion test environment studies. A simulated service test method is used with selected
corrosion media to provide accelerated corrosion data in full sized radiators. The importance of
relating corrosion test media to ultimate engine coolant use and abuse is covered. Methods have
been correlated with field tests to prove the approach used.
KEYWORDS: engine coolants, corrosion, aluminum alloys, automotive radiators
The automotive market is very much a leader in globalization of the world's economies.
Unification of design approaches, standards, and servicing requirements is becoming regular
practice. Radiator materials are under constant review and change by all manufacturers with
a view to utilizing a single material worldwide. The engine coolant area is very important to
this goal, and yet very diverse coolant formulations are prevalent on the major continents and
in industrial countries.
Manufacturers of radiators have generally considered a primary market with a particular
coolant and tried to have a universal radiator material for wider markets, where coolant con-
dition and quality vary widely. A three layer a l u m i n u m alloy with AA7072 clad on AA3003
with a 4343 braze filler on the opposite side has been used on a worldwide basis by one man-
ufacturer. This approach has had a good field record in conjunction with an amine based
coolant.
Other manufacturers have preferred to use a two layer AA3003 material as an alternate radi-
ator alloy for the U.S. market. This radiator material has a good performance record with a
North American style coolant.
Radiator materials in the engine circuit are selected for cost, fabrication capability, and
durability in service. Manufacturers have adopted various national and private standards,
such as Japanese Industrial Standard (JIS), British Standards Institute (BSI), Deutsches Insti-
tut fuer N o r m u n g (German Standards Institute [DIN]), American Society for Testing and
President, Amalgamated Technologies, Inc., Suite 208, 13901 N. 73rd St., Scottsdale, AZ 85260.
2 Engineering manager, Calsonic Technical Center America, Arizona Branch, 10232 S. 51st St., Phoe-
nix, AZ 85044.
83
Materials (ASTM), and Society of Automotive Engineers (SAE) for the development and
application of radiator materials and engine coolants.
Durability of aluminum alloy automotive radiators in service depends on the alloy selected
and the expected engine coolant environment. The challenge is to find testing methods that
correlate with service experiences that allow for comparative evaluation of competing mate-
rials without resort to long-term vehicle exposure at initial selection stages.
No single test can be taken as a reliable guide for radiator material and engine coolant selec-
tion. Comparison of material and coolant candidates by electrochemistry and a simulated ser-
vice test was performed in addition to field tests. Various corroding fluids designed to evaluate
complete radiators were included in the test series. Results were correlated with field studies
of aluminum alloy radiators in service by retrieval and examination.
Technical Background
Aluminum Alloys and Coolants
There are many tests that are performed in the evaluation of engine coolant for automotive
use. When considering new alloys for automotive cooling circuit components, similar consid-
erations are necessary. However, it is now important to examine the influence of the materials
in question against the approved engine coolants. An examination of tests available suggested
that the best approach was to use electrochemical techniques for basic information and then
proceed directly to simulated service testing.
Engine coolant tests in the United Slates have used ASTM corrosive water comprising 100
ppm of chlorides, sulfates, and bicarbonates to accelerate corrosion. An extension of this tech-
nique by Alcan to reduce the amount of coolant to very low levels and increase the percentage
of ASTM corrosive water has shown discrimination between different candidate aluminum
radiator alloys on internal corrosion.
Some Japanese companies have used the Oyama water test regimen for determination of
adequate life characteristics of aluminum alloy radiators. Ethylene glycol based engine cool-
ants are likely to maintain importance and be the major coolant component in the foreseeable
future. Accordingly, all efforts at radiator design and material improvements should, it was
felt, relate to ethylene glycol engine coolant. The amine type coolant, largely used in Japan, is
gradually being replaced, because of compatibility problems. In addition, nitrite free formulas
are becoming more popular, and phosphates have generally been eliminated in Europe.
Organic acid type formulas with fairly neutral pH values are finding increased favor.
All the changes in coolant technology are designed to meet modern vehicle specifications
for improved service. However, changes are also taking place in the materials used and man-
ufacturing methods for aluminum radiators. Therefore, developments are involved with a
moving target making it important to develop a testing routine that can allow for the changes
and give valid results.
Radiator Material Galvanic Effects

Corrosion and design cannot be separated. Design affects the type of corrosion that can
occur in the radiator through material and process selection. Technical factors in corrosion
include open circuit potential, polarization, resistivity of the electrolyte, area ratios, and the
direct and indirect galvanic couplings between materials and fluids.
Radiator manufacturers of aluminum alloy product have long recognized the problem of
galvanic relationships and how the effects can be utilized to benefit corrosion resistance. The
7072 clad aluminum alloy is a good example. The clad material is sacrificial to the core, will
corrode preferentially, and for some time protect the underlying alloy. However, if the coolant
BEAL AND EL-BOURINI ON AUTOMOTIVE RADIATORS 85
environment is too aggressive to the clad material itself, then many corrosion sites will be
developed and the degree of galvanic protection effectively reduced. The potentials of the clad
and the underlying core alloy are therefore preferentially balanced to optimize the protection
afforded in a specific environment. This is the main point. Compromise is necessary to try and
cover a wide range of environmental protection , and optimization will only occur with one
radiator and engine coolant combination. Not surprisingly, therefore, by using the sacrificial
layer approach, results will vary between very good, a large middle ground of no problems with
slight corrosion, and occasionally identifiable difficulties where the clad layer is simply too
sacrificial for reasonable protection.
Influence of Fluid Flow

The type of protection is important under fluid flow conditions. If the corrosion resistance
is obtained by creating a film or passivated layer by interaction between the fluid and the mate-
rial, there usually is a critical flow rate, above which the passive film cannot adequately form
or is scrubbed from the surface. Galvanic effects can also be exaggerated under these condi-
tions because the scrubbed area generally becomes anodic to the rest of the exposed surface.
In time corrosion can be accelerated even below known levels of potential erosion effects.
Radiator Alloys
Some years ago, a major program was executed by Ford Motor Company to evaluate alu-
minum alloys for radiators. At the time, Ford was vacuum brazing all their radiators. Alloy
7072 clad tubes with AA 6951 header material was brazed into radiator form and subjected to
field operations. Corrosion of the 7072 clad tubes was found to be a significant problem,
whereas a comparable radiator series made with 3003 unclad tubes revealed negligible corro-
sion after many miles and years of service. Ford was using vacuum brazing. Radiators were
tested in a variety of North American style coolants. Engine coolant composition was also
found important to good durability [1 ].
A widely produced standard production radiator is a 7072 clad on 3003, joined using the
Nocolok ~ process. This radiator has shown very good corrosion resistance and is currently in
large quantity production and used throughout the world.
The 3003 two layer radiator material is less expensive and is used widely in the United
States. There are several alternate three-layer materials used in radiator construction supplied
by Kaiser and Reynolds that are also widely used by some aluminum radiator manufacturers.
The decision regarding aluminum radiator material cannot be made in isolation from
engine coolant technology with any hope of arriving at the correct solution, either economical
or technical.
An example of a new material is a nonvolatile aluminum-tin cladding that was developed
by Kaiser. This material is labelled K805 and is especially made for vacuum brazed radiators
and heater cores. The clad layer proved valuable in providing cathodic protection to a 3003
core in simulated service testing. During this development effort various electrochemical tests
were performed as part of the program. Particularly, galvanic tests, polarization, and poten-
tiodynamic tests were used as part of the screening experiments in alloy development.
In a separate development, a new material designated K319 was found superior to road
splash external corrosion. Again, the importance of galvanic relationships is exemplified by
the development of a subsurface layer that is anodic to the core alloy that is developed during
the vacuum brazing process.
Alean International Ltd., Canada.

Process selection does influence choice of material. The Alcan Nocolok process can be used
with two- or three-layer aluminum radiator alloys. Specific alloys are used for vacuum brazing.
The European chloride brazing process is a very forgiving system.
The lesson from these and other aluminum alloy radiator developments is clear. Investi-
gation of a new alloy or selection of an existing one for radiators involves knowledge of radiator
material, radiator braze processing, radiator design, other materials in proximity, and engine
coolant formulations. The progression of change with alloys and coolants should therefore
involve the most complete knowledge of all these factors.
Corrosion Properties
Both internal and external corrosion of aluminum radiators are important and directly
affected by material selection and brazing process. Kaiser has observed that the SWAAT test
(ASTM Practice for Modified Salt Spray (Fog) Testing (G 85), Method G 43, using synthetic
sea salt water with acetic acid at a pH of 2.8 to 3.0 is best for external corrosion simulation.
This test program concentrated upon internal corrosion evaluation, using 3003 and 7072 clad
materials as examples of the investigative process.
Aluminum alloys are very reactive, especially in aqueous environments. The metal is rea-
sonably resistant to corrosion because of its affinity for oxygen and the protective nature of the
aluminum oxide film that develops on the aluminum alloy surface. The aluminum oxide sur-
face is penetrated by the corroding medium, and because of surrounding protection tends to
fail by pitting action.
Pitting is considered to be the most commonly encountered form of aluminum radiator
alloy corrosion. The problem is accelerated by chlorides, traces of copper forming electrolytic
cells, or very high or low pH levels. Where perforation is the criterion for failure, statistical
analysis may be judiciously applied to the distribution and depth of pits. Maximum pit depth
can be related to practice if the sample size is large enough. However, in practice the best
approach is to use probability paper so that data with random probabilities can be plotted to
give straight line relationships.
A corrosion rate increasing with time is rarely encountered with aluminum [2]. Corrosion
volume is linearly related, and several authors have demonstrated that maximum pitting
depth extends as the cube root of time as illustrated in Fig. 1 by example [3,4].
Examination of the radiator tubes should therefore be handled statistically over a period of
time and a number of tests in order to achieve reliable pitting data and an understanding of
the probable life times of radiators in service.
Engine Coolant and Water Quality in General Usage in the United States
In order to understand the levels of engine coolant percentage, contamination, and water
quality generally faced by the aluminum radiator in service, and to assist in development of
suitable testing media, a survey was carried out.
A total of 130 used coolant samples was obtained from all regions of the country. Exami-
nation of these samples revealed a very wide range of engine coolant quality in actual usage.
The percentage ranges of glycol found in vehicles from all regions is presented in Fig. 2. The
majority of vehicle users had either lower or higher percentages of glycol than that designated
by the automotive manufacturers. Twenty of the samples had less than 20% glycol. This dem-
onstrates that any test program should be carried out with lower percentages and quality of
engine coolant than the 50% coolant and water optimum.
Water quality is also very important. Most vehicle owners use tap water with coolant when
BEAL AND EL-BOUFIINI ON AUTOMOTIVE RADIATORS 87
5.0
S
2.0
1.5 / ~
1.0 ~ 1 1 d = k t ~/I
d = max pit depth
k = constant dependent o n f l u i d Bnd a l l o y
0.5 t = time, years
o.o
Y I I I I I I I
1 4 8 12 16 24 40 60
TIME, y e a r n ( c u b e r o o t )
FIG. 1--Pitting data on aluminum alloy showing pit depth increases with cube root of time [4].
ALL REGIONS % GLYCOL

130 SAMPLES
40
35
35
30
kd
C> 25
5C~
~C
D 20
O
o 16 / 16
15
13
12
11
10
10
0
0-9% 10--9% 2 0 - - 9 Z 3 0 - - 9 ~ 4 0 - - 9 ~ 50-97o 6 0 - 9 % 70-9~ 80--9% 90--9~
PERCENTAGE RANGES
FIG. 2--Percent glycolfound in random sampling of vehicles in the United States.
additions are required to the engine cooling system. A series of water analyses taken from all
regions of the United States revealed a wide range of water quality.
Four waters had high alkaline values outside the Environmental Protection Agency (EPA)
limits for drinking water. The worst condition was in St. Louis, MO, with a pH of 9.7 and high
sulfates. Florida also showed poor water quality with high chlorides. In Houston, TX, and in
Indianapolis, IN, very high chloride levels were recorded. ASTM has defined acceptable water
quality in Manual M N L 6 [4]. Most areas o f the country have satisfactory water. Some prob-
lems could be expected from the areas mentioned if the aluminum alloys used for radiators
are somewhat susceptible to general corrosion.
Evaluation of Suitable Fluids for Accelerated Radiator Tests

Following a review of engine coolants in use and water quality, it was determined that test
fluids for accelerated evaluation of a l u m i n u m alloys for automotive radiators should reflect
used coolant conditions.
The question is to define how much coolant depletion is acceptable to produce accelerated
corrosion effects without compromising the integrity of the corrosion mechanisms involved.
It is recognized that a corrosion medium that is too severe can produce unrealistic results that
would not place materials under test in their order of merit.
A number of coolant and water conditions were examined to investigate a satisfactory cor-
roding fluid for standardized test purposes.
Two types of water condition were selected for the study. ASTM has developed and defined
a corrosive water medium that contains 100 ppm chlorides, 100 ppm sulfates, and 100 ppm
bicarbonates. The ASTM water composition is widely used in radiator and coolant testing and
is the designated accelerant in several A S T M standards. The second type of water used is
known as Oyama water (OY) (Calsonic Company Standard) and has been adopted by some
in Japan as a corroding test fluid for the evaluation of aluminum alloy radiators.
Two types of engine coolant were included in the study. A normal North American style
phosphate-nitrate formula and a Japanese amine based formula for corrosion protection were
used. The North American coolant was used at the 2% and 17% level in conjunction with
ASTM water and at 2% in deionized water. The amine coolant was used at the 17% level in
ASTM water. A separate test was undertaken with Oyama water only.
Electrochemical Evaluation of Radiator Materials

Electrochemical corrosion testing is an approach to understanding the mechanisms and
amount of corrosion that are likely to occur in a coolant system. Several basic techniques can
be adopted for measuring corrosion rates. Selection of the appropriate method is carried out
by the corrosion engineer, depending on the problem involved.
Metals and alloys tend to corrode over the entire surface or in specific locations producing
pits. Corrosion reactions of both anodic and cathodic type are simultaneously occurring on
corroding surfaces. When general corrosion occurs, these areas move quickly across surfaces;
when pitting occurs, they remain localized.
Corrosion potential depends on material composition and the proximity of other materials
of similar or dissimilar compositions. Materials involved can direct the process of corrosion
depending on the relative sizes of cathodic and anodic areas. Smaller anodic areas rapidly
increase corrosion conditions. Increased corrosion per unit area can therefore be developed in
this situation. When two dissimilar corroding metals are involved by direct contact, the prob-
lem is generally referred to as galvanic corrosion.
The principal electrochemical methods used for radiator corrosion studies were as follows:
(1) Tafel,
(2) galvanic, and
(3) potentiodynamic.
Tafel
The Tafel technique is the most widely used electrochemical method for the determination
of average corrosion rates of metals. Corrosion rates are generally expressed as rate of metal
loss per unit area. This can be mils per day or year.
To obtain this information, a combined cathodic and anodic polarization curve is run with
the material and environment of interest. The intersection of the slopes of these curves pro-
duces a corrosion current density figure at the natural corrosion potential of the system. Using
the density of the metal, the corrosion rate can be calculated in any desired units.
Galvanic
Electrochemical techniques for predicting galvanic corrosion are based on three types of
measurement. These are mixed potential, galvanic current, and polarization measurements.
Each of these techniques have advantages and allow the development of useful information.
Usually, more than one technique is necessary to obtain the required confidence in the study.
The use of potential measurements for predicting galvanic corrosion is based on the mea-
surement of Ecorr for the metals involved. Metals arranged into rank order according to their
potential measurement in a given environment are known as the galvanic series for that
system.
There are drawbacks that prevent this simple approach being adequate. Formation of films,
the effect of time, polarization of specific metals, and distance between corroding species are
all important factors.
The measurement of galvanic currents between coupled metals is based on the zero resist-
ance meter. Galvanic current is measured by an ammeter, adjusting voltage, or resistance so
that the potential difference between the two elements of a circuit indicated by an electrometer
is zero. This allows continuous monitoring of galvanic currents with time.
The measured galvanic current is not always the true situation. When cathodic currents are
appreciable at the mixed potential of the galvanic couple, the measured galvanic current is
lower than the true value. Mixed potential theory of corroding systems is important to galvanic
couples. Using polarization measurements, the corrosion potential of the galvanic couple is
determined for each metal individually. When coupled together, the corrosion rate for one
metal can be increased, and for the other, decreased. The use of polarization measurements in
conjunction with mixed potential of a galvanic couple provides precise information regarding
their behavior.
Potentiodynamic
Potentiodynamic polarization experiments are generally used as a qualitative technique to
fingerprint a material in a given medium, to identify tendencies to passivate, isolate any mul-
tiple reactions, and indicate some idea of stability. Several features of the potentiodynamic
curve are important. The curve is obtained by scanning potential values of the specimen over
a wide range and measuring the current densities observed in the fluid of concern.
The potentiodynamic scan is commenced on the cathodic branch of the Tafel slope by
selecting a sufficiently low potential at which to begin the experiment. The scan proceeds
through the change over from cathodic to anodic current. This can be used also for a quick
corrosion rate measurement. The next important area is Ep,, which is the passivation potential
of the material and coolant. At potentials higher than this figure, current density is reduced
until the stable passive region, if present, occurs. The next most important point is the break-
down potential from passive to active corrosion at which current level is measured.
Material and Coolant Evaluations

Different engine coolants are used in North America and Japan. North American coolants
are generally phosphate-nitrate based, and Japanese coolants have been largely based on
amine additives.
An electrochemical study was carried out to compare aluminum alloy 7072 clad material
to aluminum alloy 3003 in five different coolant conditions:
Aluminum Alloy:
(1) AA7072 clad on AA3003 and

(2) AA3003.
Coolant:
(1) 2% North American in ASTM water,

(2) 17% North American in ASTM water,
(3) 17% Amine coolant in ASTM water,
(4) Oyama water, and
(5) 2% North American coolant.
Note: ASTM water was prepared according to details in ASTM Test Method for Corrosion
Test for Engine Coolants in Glassware (D 1384).
FIG. 3--Princeton 2 73 computer operator electrochemical testing apparatus.

TABLE l--Results of TAFEL analysis: AA3003 alloy.
Corrosion Rate,
Solution Type mm/year I-CORR/~A/cm 2
2% North Americana 2.0117 184.00

17% North American a 0.0836 7.66
17% Amine Coolanta 0.0402 3.68
Oyama water 3.4544 318.00
2% North American 0.508 48.00
a Remainder of solution: ASTM D 1384 type corrosive water.
Electrochemical Test Apparatus

The electrochemical test apparatus comprises a Princeton 273 potentiostat/galvanostat, the
experimental test cell, and the computer and program that runs the experiments, as illustrated
in Fig. 3. A working electrode, neutral electrodes, and a standard calomel electrode for refer-
ence are enclosed in a glass container in which the specimen to be tested is exposed. The test
cell is heated by a standard hot plate. Potential or current scans are automatically ramped
according to the required test. A specimen exposure area of 1 square cm was used. A + 250
m V range about E~o, was utilized at a 0.2-mV/s scan rate.
Test Results from Tafel Corrosion Method

Results from the Tafel analysis with both a l u m i n u m alloys and the five corroding fluids are
presented in Tables 1 and 2.
The AA3003 alloy shows relatively slow corrosion in the 17% amine or 17% North Ameri-
can coolant in ASTM corrosive water. The 2% North American style coolant without corro-
sive water showed similarly low corrosion rates. Two percent North American coolant in
ASTM corrosive water results in a considerable increase in material attack as measured by the
corrosion current and the calculated corrosion rates. In rank order, the 17% amine coolant
corroded the least followed by 17% North American coolant both in conjunction with ASTM
corrosive water. The 2% North American coolant was next followed by 2% North American
coolant in ASTM corrosive water. Oyama water caused the most severe corrosion.
The data for AA7072 a l u m i n u m alloy are presented in Table 2. Some differences are noted
from the above results. With AA7072 a l u m i n u m alloy, the 17% North American coolant in
ASTM water is slightly more protective and essentially equivalent to 17% amine coolant in
ASTM water. The Oyama water demonstrates a higher corrosion rate than the above fluids.
TABLE 2--Results of TAFEL analysis: AA7072 clad.
Corrosion Rate,
Solution Type ram/year I-CORR, ~A/cm2
2% North American" 6.5532 601.00

17% North Americana 0.0134 0.33
17% Amine coolant a 0.0198 0.82
Oyama water 0.4343 39.7
2% North American 1.5494 140.00
a Remainder of solution: ASTM D 1384 type corrosive water.
A further increase in corrosion of AA7072 aluminum alloy over Oyama water was found with
2% North American coolant. The most corrosive fluid for this aluminum alloy was the 2%
North American coolant with ASTM corrosive water.
These results show that specific aluminum alloy materials respond differently in selected
corroding fluids. This is very important. When designing an accelerated corrosion test, care
must be taken to select a fluid condition that can be correlated ultimately with service perfor-
mance. Merely adopting an enhanced corroding condition does not ensure that the expected
corrosion can be accelerated without influencing decisions to be made in radiator material
selection.
Results from Galvanic Corrosion Tests

The AA3003 and AA7072 clad materials were coupled together under galvanic conditions
and tested with the five coolants used for the program.
The O y a m a water revealed the greatest galvanic currents in the series at approximately 400
t~A as shown in Fig. 4. The 2% North American coolant in ASTM corrosive water was gen-
erally lower in current differential. A current of < 2 0 0 uA depending on which alloy was used
as the working electrode on the test was measured as presented in Fig. 5. Both the Amine based
coolant and North American coolant at 17% in ASTM corrosive water did not show any effec-
tive galvanic currents as shown in Figs. 6 and 7. These results show that from a galvanic stand-
point the O y a m a water is the most aggressive of the fluids examined.
Results from Potentiodynamic Tests

Potentiodynamic electrochemical tests were carried out with both AA7072 material clad on
AA3003 and on unclad AA3003 alloy. The same five environments used for other tests were
utilized.
4~0 9 ' | " '''I |

' ' : .... I ' u .... I
GALUANIC CORROSION
270
111
ALLOY 3003
~0
WORNING
SOLN:C'YAMA WATER ELECTROOE
-90
J| r-LAD 7072
-~70! L
! I
-4N0 , . u .... t . ,, .... , , I . I.,
1 2 3
10 12, 10 10
TIME SEC
FIG. 4 - - O y a m a water galvanic test.
89
GALVANIC CORROSION
.~00 2 NORTH AMERICAN
1_=.8 UNCLAD WORKING ELErTROnE
- 150 CL.,~D NOF:KIN,3 ELE,LTRO[',E
8 24 48 72 ~D
TIME SEC
FIG. 5--Galvanic corrosion in 2% North American and A S T M corrosive water.
Summary graphs of the AA7072 clad material and the unclad AA3003 are presented in Figs.
8 and 9, directly comparing these materials in four of the environments. In each case, the most
severe corrosive condition is demonstrated by the Oyama water. There is no evidence of any
passivity, and at very low potentials, a very high corrosion current is observed. Corrosion in
the AA7072 is more general with some pitting evident. The unclad AA3003 shows pitting
attack at relatively low potentials above Eco~.
The next most severe condition was obtained with 2% North American coolant in ASTM
' ' I ' ' '

2 ' '' .... I " '' .... I
GALVANIC CORROSION
17 X NORTH AMERICAN
.4
8
7072 CLAD AND :300:3 SPECIMENS
H --, 4 Y I E L D E D ~AHE R E S U L T S : ( 0
-1.2
. ., .... I . .t .... I , , ., ....

-20 1 7" 3
10 10 10 10
T I HE S~'U
FIG. 6--Galvanic corrosion in 17% North American and A S T M corrosive water.
3
GALVANIC CORROSION
1.8 IN AMINE BASE COOLANT
_n o
v
BOTH 3905 ALLOY AND ? 0 ? 2 CLAD
YISLDED SAME RESULT: ( 0 )
-i. 81
.m ,,. I I I I
o 30 ~0 90 12,3
TIME SEC
FIG. 7--Galvanic corrosive in 17% amine base coolant and A S T M corrosive water.
corrosive water. Overall, corrosion rates are less. An approximate 800-mV range is obtained
where passivity is measured, although the corrosion rate is still higher than desirable and does
not show protection of these materials.
The 17% North American and amine coolants in ASTM water environments indicated less
corrosion as measured by the potentiodynamic technique. Both of these coolants lowered the
corrosion rate but only showed slight passivity.
A comparison was made between the North American and Japanese coolants at the 17%
, ,,i,n]
2400 '""'I '""'I '""'I " '"'I '""'I '""'I "'"'I '"""
7872 CLAD P O T E N T I O D Y N A M I C S
I~00 i- 2 NORTH AMERICAN
Z- 17Z'NORTH AMERICAN ~I,~
3- 17ZAMINE COOLANT 'I~
89
- OYAMA HATER II
800
i?
J
Z
i5 o
_~ .............................................
,~:!: /.___--~- - -
-800
1
- I~00 ~..,..I .,.hi .... ,.J ...,..I ...,..J . ~ . , . J ,,,,..
--~ -2 -I 0 1 i 3 89 _m
I0 10 10 10 10 10 10 10 10 10
I(UA/CM.~.2)
FIG. 8--Potentiodynamic data plot with clad AA 7072.
BEAL AND EL-OOURINI ON AUTOMOTIVE RADIATORS 95
2400 """"1 " "'"'7 " ""7 "'"'"1 " '""1 '""7 ' " ""1 "'""'
UNCLAO ALLOY 3803
1~00 :t-zx NORTH AMERICAN
z-17xNORTH AMERICAN ~',,,~,
3 - 17V.a~-tlNI~ C O O L A N T ~i~
4. -Ft"r I.,,IA]ER .......,~"~
800
N .,...,,''J'
.,,p.ri, t. ,.rq" J''
......
,,~,e"~"~'] X/
~, . . , : ~ , , , ~ : : : . . : : : : q _ . , ~
............. 4
-800
._._,.~- -,..
--'-'-""'-~ x
- 16o0 ., ,,.,l ,,.,.,.l., ,,,.J~L..,,J ,. ,,,,,l., ,,.J ,..,,,,J , , +,
-2 -1 0 t 2 3 4 B
10 l0 10 IO J@ 10 10 10 10
I (UA/CI4...~)
FIG. 9--Potenfiodynamic data plot with unclad AA3OO3.
dilution level in ASTM water in Fig. 10. Clad material in North American coolant has slightly
better protection, and unclad was slightly better in amine coolant.
S u m m a r y o f Electrochemical W o r k
The Tafel, galvanic, and potentiodynamic corrosion studies provide very useful laboratory
type corrosion data that can be utilized for malerial or coolant ranking purposes. The tech-
1~00 " " ~ - ! ""l" 7 " " 1 7 ""r 7 '"v 7 ""v-] "",7 " ' T 7 "'4""
~DR3'H AME~ICAX/ VS ~ 2 N E COOLANTS |Q,..
AT ]7> DILUTION "f~'~l~.,
t2~0 IN ASTm O i Z 8 4 WATER I.,,'
i - UMCLAD I N NORTH ~MERIC~N ,,"/I""
2 - CLAD [N NORTH At~ER~CAN ,,,,.,,.W"
3 - UNCLA[) I N AM [NE ,,".... ~x
~OO
4 - CLAD IN AMJ:NE ..e"'
~ , , ",
- Iz~o .,,,,.I ,.,,.l ,,,,..J ,.7~ ...,.J , ..,.J ,,,,,.I ,.,,,J ,.,,..
-3 ---" - t 0 I ,2 3 4 ~
10 10 10 10 10 1~ 10 10 10 10
! (u~/CI',I.~2)
FIG. l O--Potentiodynamic data plot of North American versus amine at 17% dilution in A S T M cor-
rosive water.
niques are not infallible, but when used with supporting data, provide an increased confidence
level that the true corrosion picture is being observed.
The experimental work in the study demonstrates the viewpoint that North American cool-
ants are aggressive towards AA7072 clad AA3003 alloy. This material does not provide any
significant benefit over AA3003 alloy alone that has shown good service experience in U.S.
radiators.
Japanese studies 3 on radiator comparisons revealed more rapid corrosion with the AA3003
aluminum alloy in O y a m a water than with AA7072 clad material. Mainly, it is understood
that this occurs because o f the sacrificial nature of the zinc containing aluminum surface alloy
that provides cathodic protection to the underlying core material. Results of the electrochem-
ical Tafel work confirm this expectation. Electrochemical data also show that aluminum alloy
AA3003 will corrode more in Oyama water than in the Japanese amine type coolant or North
American coolant with ASTM corrosive water mixes.
Simulated Service Testing

Test Equipment
Two radiator test stands were utilized that have a capability of operating at up to 125~ and
at pressures up to 210 k N / m 2. The test stand is built to operate at a range of temperatures and
pressures and to be pressure cycled if necessary.
The thermal cycle is related to the heating system used and normally takes 15 to 20 min to
heat up depending on the temperature and size of radiator. The pressure cycle is operated inde-
pendently by means of an air to fluid pressurization cell that allows a reasonable rise and fall
in the rate of pressure in the stand. A timer is used to operate the pressure cycle. The test stands
are capable of continuous running or timed periods of operation. An adjustable flow of 0 to
80 L/min is available. The radiator capacity is 2 units at up to 0.9 m wide by 0.75 m high.
Safety equipment is included that will operate with loss of fluid, over pressure of the unit, or
over temperature. The unit operates quietly, and easy access is available to the radiators. A
NEMA enclosure is used for all electrical controls.
Experience with operating similar units on a prescribed pressure and temperature cycle of
0 to 140 k N / m 2 and 80 to I 15~ has allowed correlation with on-the-road service: Using an
8-min pressure cycle in conjunction with the above program, 1000 h on the test unit is equiv-
alent to approximately 80 000 km on the highway. A flow rate of 32 L/min was used through-
out this test program, which was related to actual service flow rates for the radiators used.
The simulated service test frames are shown in Figs. 11 and 12 with a schematic diagram of
engine coolant flow depicted in Fig. 13.
Selected Corroding Test Fluids

2% North American Coolant in A S T M Water--The accelerated corrosion rate coolant used
for radiator simulated service tests was 2% North American style coolant with ASTM D 1384
corrosive water. ASTM corrosive water is composed of 100 p p m chloride, 100 ppm sulfate,
and 100 p p m bicarbonate ions. This corrosive water is used to simulate a U.S. tap water. Cool-
ant was changed every 3 months.
OYAMA Water--Work in Japan has resulted in the adoption of OY(Oyama) water as a
testing fluid for the evaluation of aluminum alloys for automotive radiators by some
manufacturers.
3 Private communication, Calsonic, Phoenix, AZ, Oct. 1989.

4 Private communication, Amalgamated Technologies, Scottsdale, AZ, 24 April 1991.
FIG. 1 l--Simulated service test frame,
The engine coolant is replaced by this aggressive water formulation that produces failure of
aluminum alloy radiators in 3 to 6 months. This test system is also used by the Japanese to
evaluate the rate of pitting formation on the internal surfaces of the tubes in the radiator cores.
Radiators were run continuously in this water at 88"C for up to 90 days. The water was
changed every 4 weeks.
Radiators
The radiators used in this particular test were as follows:
1. Calsonic standard production radiators with AA7072 internal cladding, AA3003 base
material, and AA4343 braze alloy.
2. Radiators with no internal cladding using AA3003 base material, and AA4343 braze
alloy.
FIG. 12--blstallation of radiators in radiator testframe.

F'RES~;URE
OAUO~$'*'ITCH V5~ "~G
'
SIMULATED
~R~CE
, ]~ ER
VALVES
V2 - PRESSURE SHUTOFF
/
V2 - DRAIN
V3 - FLOW ADJUST
V2
V5 - FILL & AIR BLEED
DRAIN
V6 - PRESSURE RELIEF
RZGULATOR V7 - SOLENOID VALVE

3-WAY
V7 0 COMPRK$$ED
~Tr,,. F---~-AIRIN
MUFFLER
FIG. 13--Schematicdrawingofcoolantflow.
Simulated Service Test Results in North American Coolant

Simulated service tests were carried out using equipment and conditions already described
with the above radiators and testing fluids. Pitting examinations on this program were carried
out by direct microscope comparator depth measurements and occasionally by metallurgical
cross section to verify direct measurements.
FIG. 14--Surface nodule on unclad alloy.

FIG. 15--5-urn pit on unclad in coolant.
AA3003 Radiators--Monthly examination of the unclad radiators revealed very little pit-
ting ( < 1 p i t / m m 2) throughout the course of the test. A single 15-um pit was found after 9
months of testing; however, no other pits were found with a depth greater than 5 um. Some of
the pits were found underneath nodules that had formed on the tube surfaces. Examples of a
nodule and a cross section of a tube pit are shown in Figs. 14 and 15. The pitting observed in
Fig. 15 was general with several smaller pits than the one measured.
Examination of the header plates after 9 months revealed very little pitting. Maximum pit
depth found was 5 t~m. An example is shown in Fig. 16. There is some residual braze filler
metal material that acted as a cathodic site and tended to promote pitting.
7072- Clad Radiators

Monthly examination of the clad radiators revealed severe pitting (>> 1500 pits/ram z) occur-
ring as early as 4 weeks after test initiation. M a x i m u m pit depth steadily increased until a pla-
FIG. 16--Pits in header plate of clad radiator.

1O0 ENGINECOOLANT TESTING: THIRD VOLUME
FIG. 17--131-~m deep pit in clad radiator tube.
teau of 96 um after 4 months, where the clad material showed protection of the underlying
core alloy AA3003. Maximum pit depth did not rise above this level until after 8 months into
the program, when clad protection finally broke down and a pit was found with a depth of 115
~m. Final maximum pit depth after 9 months was 131 um, well into the AA3003 core material.
Examples of these deep pits are shown in Figs. 17 and 18. Figure 19 is a cross section of a 58-
~m pit that penetrated the clad layer. At this point the clad layer then becomes protective of
the core material slowing down the rate of pitting attack. Additional examples of pitting are
shown in Figs. 20 and 21.
Pitting Summary
The pitting mechanism in the standard AA7072 clad alloy radiator begins as a shallow
depression in the sacrificial clad, which then spreads along the surface of the AA3003 second
FIG. 18--114-~m deep pit in clad radiator tube.

FIG. | 9--58-um deep pit penetrating clad layer.
FIG. 20--71-urn deep pit in clad tube.
FIG. 21--65-~m deep pit in clad tube.

layer in the AA7072 clad material. Galvanic protection takes place. After a period of time,
pitting attack occurs in the second layer, more rapidly penetrating the tube material.
In the unclad radiators, no such pitting attack develops initially in this coolant condition.
Passivation occurs, producing surface protection. After extended exposure time, pitting attack
results in isolated instances. Results are graphically presented in Fig. 22.
No failures of radiators were noted after 9 months of simulated service testing.
Simulated Service Test Results in Oyama Water

AA3003 Radiators--Two unclad AA3003 radiators were used in exposure to Oyama water
at 88~ Very different results were obtained with these two radiators. After 8 weeks of testing,
one of the radiators revealed serious tube pitting with a pit depth of 110 um. This radiator was
removed from the test in order to examine the header and other tubes. The deepest header pit
found was 115 um. N o deeper pit was found than the original tube examined, although several
other pits almost the same depth were identified. An example of tube pitting is shown in Fig.
23. A tube removed from the other radiator after 8 weeks revealed very little pitting (Fig. 24).
A maximum pit depth of 5 um was found in the tubes when the second radiator was
removed from testing after 12 weeks. It is not unusual to have such wide variations in pitting
attack with an aggressive corroding fluid. The header plate of this radiator revealed a maxi-
mum pit depth of 200 ~m. Results of tube pitting for each of the two radiators tested are graph-
ically presented in Fig. 25.
AA 7072-CladRadiators--Tubes removed from the AA7072 clad radiator reveal very small
pits even after 12 weeks of simulated service. The maximum pit depth measured was 5 urn.
CLAD VS UNCLAD I N 2% NORTH AMERICAN/ASTM WATER

130
120
110
100
O ff] U
90
v 80
T
70
w
60
x 50
40
30
20
10
I I I
0 i i I i I I
10 20 3O 40
TIME (WEEKS)
[] CLAD Jr U~LAD
FIG. 22--Pit depth measurements in 2% North American coolant and ASTM corrosive water.
FIG. 23--100-~m pit in unclad tube after 8 weeks in Oyama water.
Some pitting was found in the header plate with a m a x i m u m depth of 60 um. Pits in the header
had formed adjacent to residual surface braze filler material and were obviously galvanic
attack due to braze filler presence. This is illustrated in Fig. 26. No radiator failures were noted
after 3 months of simulated service testing.
Pitting S u m m a r y - - R e s u l t s of simulated service testing in Oyama water clearly show that
AA3003 is subject to pitting attack in this fluid. Under the same conditions, AA7072 showed
much less pitting attack. Both materials would pass the Oyama water criteria for testing.
Correlation o f Lab Test and Field Data--Success of a simulated service testing program
depends upon adequate correlation with actual results obtained in vehicle service. A program
of controlled testing was performed with known radiator alloys and engine coolants. Ten vehi-
cles were involved in the field study. Radiators were removed from service after a number of
years and miles of operation. The test program is continuing with a new series of evaluations
FIG. 24--Shallow pit in unclad tube after 8 weeks in Oyama water.

I I I I I
1 oo
8o Q
N 6o
~ 4o
2O
OvA ~
0 2 4 6 8 10 2
TIME (Weeks)
FIG. 25--Pit depth measurements in unclad radiator material in Oyama water.
and is expected to be a continuing effort for some time as both radiator alloy materials and
engine coolants evolve.
Results for simulated service testing of four different aluminum radiator alloys in a selected
corroding fluid are presented in Fig. 27 as an illustration. None of these radiators leaked during
the tests. Depths of pitting measured were from radiator tubes after removal from the tested
radiators at intervals during the testing program. The maximum pitting depth obtained for
each radiator was used for evaluation. Alloy 3 was tested in both field service and simulated
service.
Controlled field survey data of Alloy 3 are shown in Fig. 27. Again maximum pit depth
found in radiators removed from service are shown. A good correlation is seen between the
maximum pit depth obtained by simulated service testing and pit depths found from vehicle
FIG. 26--Clad header with 56-~m deep pit adjacent to braze filler metal.
BEAL AND EL-BOURINI ON AUTOMOTIVE RADIATORS 1 05
0.3 rnm
Controlled Simulated Service

Field in Selected Fluid
Survey
mm
..................... i
i *i
,t,!
r 0.1 T mm ~li :: Legend #4

Alloy #3
~> Field Survey
9 Alloy #t
9 Alloy ~2
Alloy #3
9 Alloy
II I : { i/ J J ':1 ] P J
tB 15 1~ ~ ~ t ~ ~ 4 ~ e
Years ~ube Root) ivionths ( C u b e R o o t )
FIG. 27--Correlation of pit depth from test and fieM data.
operation with Alloy 3. Further testing will be used for validation, but presently 3 months of
simulated service approximates 8 years vehicle usage. We have noted a direct aluminum radi-
ator alloy dependency for a given corroding fluid as shown by the results of the other three
alloys in simulated service in the same test fluid. No obvious universal relationship between
each alloy and its service durability in engine coolants was observed. Careful data interpreta-
tion is needed in reaching conclusions from accelerated testing with selected coolants.
Discussion
Using the Oyama water test fluid as a measure of acceptability or selection of a radiator
material, the AA7072 alloy would be adopted. A Japanese radiator manufacturer has found
that the Oyama water test does give credible results that can be correlated with service when
using the amine type Japanese engine coolant [6].
Very different results are obtained when a North American coolant is employed. Under
these fluid coolant conditions, the AA3003 aluminum alloy radiator material is found to per-
form better than the AA7072 clad AA3003 core aluminum alloy radiator material.
It can be readily discerned, therefore, that preference will be given to a particular aluminum
alloy radiator material depending upon the corroding test fluid used. The objective of a testing
program is to identify which aluminum alloy material is best suited to service in a particular
country, continent, or even worldwide. Dangers in specific test criteria become clear. It could
be fatal to be adamant about the accelerating corrosion test media that is most applicable to a
development program, unless some understanding of the engine coolant types to be used in
service has been ascertained by a correlative field survey study.
The Oyama water concept does not appear to be applicable to the North American coolant
situation from electrochemical and simulated service radiator tests carried out on this pro-
gram. The work has shown that it is better to use a corrosive water medium with a percentage
of the engine coolant that is appropriate.
This test program did not include any European style formula evaluation, but this will be
the subject of further work based on these findings. The purpose of the paper is to demonstrate
the need for careful consideration where using accelerated testing to determine material
selection.
Conclusions
A study has been performed on the use of accelerating corrosion media to evaluate alumi-
num alloy radiator materials in engine coolant service. The study has used electrochemical
evaluation of materials and simulated service testing o f radiators using alloy AA7072 clad and
AA3003 as examples of widely used aluminum radiator alloys. Five accelerated corrosion test
coolants were used in the program. These were North American and Japanese style coolants
at 17% and 2% levels together with Oyama water.
Results showed that aluminum alloy selection can directly depend upon the test fluid cho-
sen. Data confirmed other studies on aluminum radiator alloy choices and highlighted the
problem of test coolant preference.
Unless related to the particular engine coolant environments common to the country or
continent of concern, serious mistakes can be made by manufacturers if selection of alumi-
num alloy radiator materials is based on some form o f accelerated test without sufficient cor-
roborative evidence.
Unclad alloy AA3003 performed better in simulated service compared to a AA7072 clad
AA3003 material using a diluted North American coolant with ASTM corrosive water. This
result is consistent with field experience with these two materials, and electrochemical tests in
this study.
A successful accelerated test for radiator service life requires removal of tubes during the
test for evaluation of internal pitting depth after certain exposure periods. Reasonable corre-
lation between simulated service and field experience is obtained by these pitting depth
measurements.
At the present time, it does not appear feasible to have an accelerated test that causes radi-
ator failure by leaking in a short time period that can be realistically related to field service
experiences.
Simulated service tests and field survey work continues with efforts to find improved test
correlation for both pitting and radiator failure.
Acknowledgment
The authors wish to acknowledge Mr. Masashi Iwasaki of Calsonic Corporation for his valu-
able guidance throughout this evaluation program.
References
[1] Barkley, J. M., Wiggle, R. R., and Winterbottom, W. L., "The History of Ford Automotive Alumi-
num Radiators," Engine Coolant Testing: Second Symposium, STP 887, American Society for Test-
ing and Materials, Philadelphia, 1986, pp. 11-26.
[2] Shreir, L. L., Corrosion, Vol. 1, John Wiley & Sons, New York.
[3] Rowe, L. C., Galvanic and Pitting Corrosion, STP 576, American Society for Testing and Materials,
Philadelphia, 1974, pp. 203-216.
[4] Pathak, B. R., Handbook on Corrosion Testing and Evaluation, W. H. Ailor, Ed. Wiley, New York,
1971, p. 553.
[5] Lima, J. A. and Otterman G. R., Manual on Selection and Use of Engine Coolants and CoolingSys-
tem Chemicals, MNL 6, 4th ed., American Society for Testing and Materials, Philadelphia, 1989.
[6] Ando Y, et al., "Development of Aluminum Radiators Using the Nocolok Brazing Process--Coro-
sion Resistance of New Aluminum Radiators by Applying a Nocolok Brazing Process," SAE Tech-
nical Paper 870180, Feb. 1987.
R i c h a r d D. H e r c a m p ~
An Overview of Cavitation Corrosion of Diesel

Cylinder Liners
REFERENCE: Hercamp, R. D., "An Overview of Cavitation Corrosion of Diesel Cylinder Lin-
ers," Engine Coolant Testing: Third Volume, ASTM STP 1192, Roy E. Beal, Ed., American
Society for Testing and Materials, Philadelphia, 1993, pp. 107-127.
ABSTRACT: Cavitation corrosion (liner pitting) became a major problem for wet sleeve diesel
engines in the 1940s. For most engines, generous use of supplemental coolant additives (SCAs)
adequately controlled pitting. In the mid 1980s, problems of SCA overtreatment such as silicate
gelation, water pump seal seepage, and solder bloom raised awareness that too generous usage
of SCAs was undesirable. This new awareness caused a general shift toward lower average SCA
concentration.
In addition, engine power densities increased and load factors increased as speed limits
increased. These factors increased the work load of the engine, resulting in a greater potential for
liner pitting.
All these factors have resulted in an increased incidence of liner pitting in recent years. This
paper is an overview to help place all the various factors related to liner pitting in perspective.
The effects of engine design and materials, cooling system hardware, maintenancepractices, and
coolant composition are discussed.
KEYWORDS: engine coolants, heavy duty, cavitation, corrosion, cylinder liner, cavitation cor-
rosion, liner pitting, additive, SCA
The intent of this paper is fourfold:
1. To serve as a quick review for anyone wanting to learn about the various factors that may
affect the degree of cavitation in any specific case. This background should enable one to
do a better job of problem solving and getting at the real causes of failure.
2. To provide appropriate literature references to guide serious researchers to pertinent pre-
vious work.
3. To present some previously unpublished data.
4. To suggest approaches that could reduce the incidence of cavitation corrosion of diesel
engine cylinder liners.
Description of Cavitation Corrosion of Diesel Cylinder Liners

Figure 1 shows a section of an engine block with a replaceable wet cylinder liner. As the
piston moves vertically, it also moves a small a m o u n t horizontally due to inertia and com-
bustion loading and impacts the cylinder liner causing it to vibrate. When the vibration is
severe, the coolant cannot "follow" the rapid motion of the liner wall. Vapor bubbles form
due to creation of an instantaneous low pressure in the coolant as the liner moves away from
the coolant. Immediately after the vapor bubbles form, the liner surface movement reverses
Consultant, Fleetguard, 55E 1100S, Columbus, IN 47201.
107
FIG. 1--Schematic o f cylinder liner cavitation corrosion,
direction creating a localized high pressure in the coolant causing the vapor bubbles to col-
lapse. The resulting forces from the collapsing bubbles act on the liner surface, along with cor-
rosion, to remove metal in a pitting fashion (Fig. 2).
Just as it is important to know what cavitation corrosion is, it is equally important to know
what it is not. Figure 3 shows a liner with some cavitation corrosion, but it failed at 59 715
miles due to another reason. The cavitation corrosion is in the circled center o f the liner, but
the failure occurred lower at the edge of the crevice seal groove. Figure 4 shows a typical deep
pit with rough surfaces from the area of cavitation corrosion. This is in contrast to the smooth
surface of the failure hole shown in Fig. 5. The seal was held away from its groove by steel shot
FIG. 2--Cavitation corrosion on thrust side.

HERCAMP ON CAVITATION CORROSION OF DIESEL CYLINDER LINERS 109
FIG. 3--Failure at 59 715 miles due to erosion behind the crevice seal.
FIG. 4 - - Typical pitted area from the circled area of Fig. 3.

FIG. 5--Smooth surface offailure hole from crevice seal area of Fig. 3.
(Fig. 6) found near the failure. The failure resulted from flow erosion due to coolant pumping
action behind the rubber crevice seal. The smooth surface of the flow-eroded failure hole is in
stark contrast to the rough surface of the cavitation corrosion pit, which is typical of cavitation
corrosion. A similar comparison is shown by Leith and Mcllquham [1] for sand erosion and
for cavitation erosion on a Pelton water turbine. The sand erosion surface is smooth compared
to the cavitated surface.
Casting defects and stray electrical currents are two other examples of what cavitation cor-
rosion is not. Casting porosity does not routinely occur only on the thrust side of the liner and
less frequently elsewhere, as does cavitation corrosion. This rules out casting porosity.
As for stray electrical currents, it is difficult to imagine an external electrical current choos-
ing to flow through the massive cast iron block to only small discrete areas of the liner and
create pitting.
Commercial Consequences of Cavitation Corrosion Failure of Cylinder Liners

The cost to repair cavitation corrosion failure of diesel cylinder liners is high. At the least,
an inframe engine overhaul is required. At the worst, a hydraulic lock-up at engine start-up
due to coolant leaking into and filling the volume above the piston can cause total engine
destruction. An inframe engine overhaul can cost about $5000, whereas a new engine can cost
$15 000 for an on-highway diesel truck. Furthermore, in severe circumstances, failure can
occur in less than 1000 h or 80 000 km of operation.
Cavitation
A paper published in 1982 states, "Since Euler recognized the phenomenon of cavitation in
1754, more than 1000 papers concerning cavitation corrosion have been published and its
FIG. 6--Steel shot that held crevice seal awayfrom its seat.
deleterious effects are often still seen" [2]. A useful general coverage of the subject was pre-
sented in 1970 in a book entitled Cavitation [3]. A possibly more useful (from an engineering
standpoint) and slightly newer summary is provided in the Handbook of Cavitation Erosion,
published in 1974 [4]. Other books containing compilations of related individual papers are
also helpful [5-9]. Reference 9 contains an author and subject index for the complete series
of Cavitation and Multiphase Flow Forums from 1966 through 1984. Reference 4 lists 243
literature references.
One of the reasons for the large amount of technical literature about cavitation is that the
phenomenon is important in many areas. Some of these are liquid film bearings, radial face
seals, and synovial joints [10]; liquid sodium space power conversion systems [11]; liquid
sodium pumps for nuclear power stations [ 12]; rain erosion of high-speed aircraft surfaces and
liquid droplet erosion in low-pressure stages of steam turbines [13]; diesel engine fuel systems
[14]; ship underwater appendages, ship propellers, hydraulic turbines, pumps, valves, regu-
lators, sluice gates, civil engineering hydraulic structures, underwater sound transmission and
detection devices [15]; and the subject of this paper, diesel engine cylinder liners. In most of
these situations, cavitation tends to be detrimental. However, in some applications, it is useful,
such as a cavitatingjet used to excavate material more rapidly [ 16], ultrasonic cleaning baths,
and as an aid in liquid processing for areas such as biotechnology, chemical catalysis, and the
production of colloids and emulsions [17].
Cavitation Bubble Dynamics

It is generally agreed that the material damage due to cavitation is caused by repeated
mechanical forces acting on the material surface as a result of bubble collapse. It is also gen-
erally agreed that some nucleus, either solid or gaseous, is needed to allow vapor bubble growth
to occur in most commercial situations where cavitation is observed [ 18,19]. Most recent work
has been directed toward determining whether the forces impacting the solid material surface
in the area of cavitation is due to shock waves emanating from bubble collapse or from liquid
microjets formed near a surface at the end of bubble collapse [20,21].
Mechanical versus Chemical Damage

Plesset and Ellis concluded in 1955 that " . . . in so far as the basic process of cavitation dam-
age is concerned, chemical effects are not of primary significance" [22]. However, Plesset later
proposed a "'pulsed cavitation" test method to take into account corrosive effects [23]. At
about the same time, other investigators were confirming the chemical contribution [24,25].
Schulmeister and Chance summarized the situation appropriately: "Chemical and electro-
chemical actions always have an influence when the mechanical part o f the complex stress is
relatively low and the attack is directed against materials sensitive to corrosion" [26]. "There
are two components that contribute to cavitation damage; one is the mechanical component
and the other is corrosion. The relationship between the mechanical and corrosion compo-
nents is complex because the processes involved are frequently interrelated and synergistic"
[27]. Kallar gives a thorough review of the influence of chemical processes acting in concert
with mechanical forces to damage materials [28].
Early Work on Cavitation Corrosion of Diesel Engine Cylinder Liners

LaQue and Hergenroether [29] first briefly reported on cylinder liner cavitation corrosion
in 1949, and Speller and LaQue [30] gave a more comprehensive report on the problem in
1950. Several papers were presented at the June 1956 Society of Automotive Engineers (SAE)
meeting [31-35]. Several more papers were presented in the United Kingdom in 1965 at a
symposium on "Cavitation Corrosion and Its Prevention in Diesel Engines" [36]. Weigand
and Schulmeister [37,38] reported in 1967 and 1968 on their work done in Germany at the
Internal Combustion Engine Research Association.
Theories of Cylinder Liner Cavitation Corrosion

In 1956, Trock [31] listed three principal theories o f cavitation erosion: Poulter's, Beech-
ing's, and Crewdson's. (Another reference published by 1955 listed 18 proposed mechanisms
of cavitation erosion [39].)
Trock described the three theories as follows:
Poulter's Theory--Cavitation-erosion is the result of penetration of a liquid into the pores

of metal and its subsequent escape carrying some of the metal with it.
Beeching's Theory--Cavitation-erosion is the mechanical disintegration of surface layers
under repeated, highly localized pressure pulses. These pulses are caused by the sudden
collapse o f cavities in contact with or close to the surface.
Crewdson's Theory--Cavitation-erosion is caused by super-heated steam at extremely high
pressure. This steam is the gas or vapor in the cavitation bubbles.
My view of typical cavitation corrosion o f diesel cylinder liners has been previously
expressed [40]:
It is the authors' opinion that cavitation corrosion could best be described as corrosion accelerated
fatigue. Nonetheless, it is in this region where the chemistry of the coolant is exceedingly important
since the type of oxide formed and its adhesion to the liner surface dictate whether pitting will take
place. This is the vibration range where a well formulated coolant corrosion inhibitor can offer excel-
lent liner pitting protection.
At high levels of vibration, coolant chemistry effects are minor because the damage taking place is
mainly of a mechanical nature. Bubble collapse has become so violent that base metal is rapidly
fatigued making corrosion inhibitors of little, if any, use.
Little has changed since then to alter this view except that coolant passage geometry has been
added as a significant factor [41].
Some Discordant Notes

Although work in the area o f cavitation damage in general and cylinder liner cavitation
damage in particular has supported the basic view that material damage is caused by a c o m -
bination o f mechanical damage due to the forces resulting from vapor bubble collapse and
chemical damage due to corrosion, not all agree. S o m e o f these other views are as follows:
Bubbles do most of their damage not when they burst but when they billow, according to a new
study by You Lung Chen and Jacob Israelachvili at the University of California, Santa Barbara. In
the May 24 Science, they conclude that a bubble's birth involves violent stresses that can deform and
destroy nearby surfaces. This counters the conventional explanation of how bubbles cause pitting,
or cavitation damage, in turbine blades, pumps, propellers and other materials in contact with mov-
ing fluids [42].
Y a k o v l e v [43] felt that cylinder liner cavitation corrosion occurs at areas of high tensile
stress rather than areas o f greatest vibration levels.
Wall [44] felt that cylinder liner pitting could be due to vibration-accelerated corrosion even
if the cylinder liner vibration levels were low enough to preclude formation of vapor bubbles
in the coolant.
The chemical aspect o f cylinder liner cavitation corrosion was discounted by Z h o u [45] as
follows: " I t was concluded that chemical corrosion had little effect on forming the honey-
c o m b e d structure o f the eroded surface."
A recent paper [46] accepts the c h e m i c a l / m e c h a n i c a l nature of cylinder liner cavitation cor-
rosion but adds s o m e additional views on the failure mode.
The failure mechanism can be described in the following form:
9 penetration of the cooling liquid into the bulk of the metal along large graphite inclusions and
surface defects with subsequent oxidation of the base metal;
9 loosening of the base metal as a result of oxidation and the action of hydrolytic impacts which
can cause plastic deformation of microvolumes of the metal;
9 unwedging of the base metal by the impacts and the oxides which leads to its failure;
9 the development of internal stresses in the heterogeneous structure due to thermal and mechan-
ical effects (the microstructural components pose differing mechanical and thermophysical
properties) where the structural components with low strength (graphite, ferrite, and the non-
metallic inclusions) are easily damaged and washed out of the cavitation zone which then per-
mits an even higher level of corrosion-erosion attack.
Current Commercial Situation

Liner pitting (cylinder liner cavitation corrosion) b e c a m e a m a j o r p r o b l e m for wet sleeve
diesel engines in the 1940s. For most engine designs, generous use of supplemental coolant
additives (SCAs) adequately controlled pitting. In the 1980s, problems of overtreatment such
as silicate gelation [47], water pump seal seepage [48], and solder corrosion [49] raised aware-
ness that not only undertreatment but also overtreatment was undesirable. This caused a shift
toward lower average concentration of SCA (Fig. 7).
Decreased heat exchanger design capacity margins, increased complexity of cooling sys-
tems, and less coolant volume led to many instances of overheating and greater coolant loss
rates. In addition to leakage at hose clamps and so forth, overzealous topping up procedures
also increased coolant loss. These increased occurrences of coolant loss further aggravated pit-
ting problems by widening the distribution of SCA concentration (Fig. 8).
In addition to this, engine power densities (horsepower per litre) continued to increase and
speed limits were increased. These factors increase the work load of the engine, resulting in a
greater potential for liner pitting.
Factors Affecting Liner Pitting Severity (see Table 1)

Engine Design and Materials
Liner Vibration Levels--Anything done by design that reduces liner vibration also reduces
liner pitting. When liner vibration is at normal levels, liner pitting can be adequately controlled
by the proper choice of coolant additive. But not when the liner vibration levels are too high.
In the case of one V-12 engine design, Teflon (Teflon is a registered trademark of Dupont)
pads were added to the piston skirts to reduce piston slap (and, therefore, liner vibration levels)
-6
f-
>
..Q
E
",1
z
~.~verconcentration
/Lir~er /
0
0 .5 1.0 1.5 2.0 2.5 3.0 3.5
SCA Concentration (1 is r e c o m m e n d e d level)
FIG. 7--SCA concentration 1985 versus 1990.
r
Low Coolant
ilJ
-6
r
Loss
>
E
"1
z
~ HighCoolant
/ ~'~...~s Over-
Liner / ~ ~ Concentration
Pittingn
0 .5 1 1.5 2 2.5 3 3.5
SCA Concentration (1 is recommended)
FIG. 8--Coolant loss effect on SCA concentration
TABLE 1--Summary chart offactors affecting liner pitting.
A. Engine Design and Material Variables

Liner Vibration Level Coolant Coolant Condition Liner
Passage Material and
Geometry Coatings
Power density Temperature and pressure of
Design of piston coolant in the block
Piston/liner clearance
Timing
Reciprocating mass
Liner stiffness
Liner support
B. Cooling System Hardware
Aeration
Coolant loss
Temperature and pressure of
coolant (design, malfunction,
misapplication)
C. Maintenance Practices
D. Coolant Composition
Chemical
Physical
because pitting could not be controlled by use of coolant additives alone [50]. Some other
engine designs have been relatively resistant to liner pitting, even if the coolant was not well-
maintained. Some of these have used so-called mid stop liner designs wherein the liner is sup-
ported at its mid section [51,52]. This design reduced liner vibration levels. Some aluminum
piston designs have had large piston-to-liner clearances when running at low piston tempera-
ture, so that piston expansion and scuffing would not occur at high piston temperature at high
load. These designs were prone to pitting under cyclic loading because of the large piston-to-
liner clearance at low temperature [52, 53]. This allowed high impact loads due to severe piston
slap, resulting in high liner vibration levels. Conversely, oil-cushioned [54] and some articu-
lated two-piece piston designs have been shown to reduce liner vibration greatly [52,55], as
well as one-piece gallery cooled [56] and "sled runner" type aluminum pistons [57]. Back-to-
back 200-h engine tests with the same marginal coolant in the same engine using non-sled
runner design pistons showed much liner pitting (Fig. 9). In contrast, none of the cylinder
liners in the test run using "sled runner" design pistons was pitted. In fact, the lubrite coating
was not even removed.
The following factors have also been observed to affect the level of liner vibration:
9 In general, as the power density increases and the rated speed decreases, liner thrust loads
increase. This tends to increase liner motion on the major thrust side where liner pitting
is generally most severe.
9 A heavier piston and a thinner, less stiff liner result in greater liner vibration.
9 Piston dynamics are complex. In addition to piston design, other variables such as timing,
rates of pressure rise, engine speed, and load, can all greatly affect the severity with which
the piston impacts the liner [58-60].
FIG. 9--Five liners that pitted in a 200-h test using weak SCA.
FIG. 9--Continued.
Coolant Passage Geometry--Another important design factor is the coolant passage width
in the thrust area o f the liner most prone to pitting. Generally, the less the width of the coolant
passage, the greater the pitting, all other factors being equal [45,55]. However, an increase of
coolant passage width beyond about 0.4 in. yields little, if any, benefit [41].
Coolant Condition--Because bubbles of coolant vapor must form and then collapse to
cause liner pitting, the design pressure and temperature in the coolant passage at the thrust
side of the liner are important factors. Laboratory tests using a vibrating cast iron specimen in
coolant show the nonlinear effects of temperature and pressure (Fig. 10) [61].
When the coolant temperature is low, relative to the boiling point, or when the coolant pres-
sure is high, it is difficult to form a vapor bubble. Therefore, the vapor bubble either does not
form or the collapse of the relatively small bubble that is formed does not generate a mechan-
ically severe environment. Under these conditions, pitting is not severe. At the other extreme,
when the coolant is near the boiling temperature and pressure, vapor bubbles easily form, but
they do not readily collapse. The m a x i m u m damage rate is obtained at an intermediate tem-
perature and pressure where a significant number of bubbles form and collapse, yielding an
intense mechanical attack.
Liner Material and Coatings--Because liner pitting is a physical and also a chemical phe-
nomenon, both the corrosion resistance of the liner and its strength influence its resistance to
pitting. The ultimate resilience of materials has been found in bench tests to correlate best with
cavitation resistance in the absence of corrosion [62-64]. Ni-Resist and other more corrosion-
resistant alloys have been reported to improve [30,33] or to be of no benefit [65] in reducing
cylinder liner cavitation corrosion.
Because the liner material has generally been optimized for inside (oil) surface wear perfor-
mance and producibility, it is difficult to change the liner material to increase its liner pitting
performance. Therefore, outside diameter (OD) coatings have been evaluated. OD coatings,
such as chrome, have sometimes been reported to be helpful [32,53,66], but many times are
m
a
X
i
m
y e
v
0
I
u C
m i
t
y
d d Frequency Amplitude
a e
0
m t
a e
g r
e i a
0
m
a r a
t a g
t e
i
Temperature Static Pressure
0
n
FIG. l O--Relation between velocity of damage and main influences of test rig (schematic).
not [33]. Coating failures probably relate to the difficulty of applying and maintaining a defect-
free surface on the liner before and during assembly. Any coating discontinuity tends to aggra-
vate pitting severity by physically and galvanically focusing the attack.
Cooling System Hardware

Even when the engine design and materials variables described in the previous section are
optimized to prevent liner pitting, cooling system hardware and plumbing problems can cause
liner pitting. If the cooling system does not de-aerate properly, liner pitting can result [37].
The circulating air bubbles serve as nuclei for cavitation vapor bubbles to form when they
otherwise might not. Aeration of the coolant can also occur when the design of the cooling
system flow circuit results in low coolant pressure at the suction side of the pump. In this case,
air will be aspirated past the coolant pump seal faces. Improper installation or design of cooling
system hardware can alter the coolant pressure or temperature in the block coolant passages
and can result in increased liner pitting. Improper installation of cold starting aids, operation
with malfunctioning thermostats, shutter stats, fan clutches, and so forth, or with improper
activation temperatures are specific examples of these types of problems.
Finally, any cooling system hardware problem that results in excessive coolant loss will
result in liner pitting because the chemical balance of the coolant is upset. Water by itself or a
water and antifreeze mix used to replace coolant loss will result in a low SCA concentration
(Fig. 11). Liner pitting is the end result. Some examples of field problems with cooling system
hardware that result in excessive coolant loss are: silicone coolant hoses leaking at low tem-
perature [67]; leaking head gaskets, water pump seals [68], oil coolers, radiators, and after
coolers; and overheating resulting in boil-over.
2.5
"10
a)
"0 .................... .Danger.leve!!or .............. .X~5~.......,. " ....
c
G)
2
E
E
0
o
a~
"-
(/) 1.5
.m
t-
O = 1 gallon per 15,000 miles dilution
I I I I I I I I I I I I I
t-
G)
o
t-
O
. . . . "ve! . . . . . . . . . . . . . .
o0.5
3 gallons per for l i n e r p i t t i n g
o 15,000 miles
dilution
I I I I I I I
0
0 3O 60 90 120 150 180 210 240
Miles on the Vehicle (Thousands)
FIG. l l--Effect of dilution rate on SCA concentration.
Maintenance Practices
Many maintenance personnel are not aware of two facts: The SCA (not generally the anti-
freeze) contains the ingredients that control liner pitting. Second, the initial dose of SCA is
three to four times the maintenance dose. Therefore, when coolant loss is excessive, the cool-
ant composition gets out of balance because the proper amount of SCA is not replaced at the
same time the system is topped up. The following specific practices have been observed to give
low SCA levels in the coolant and result in liner pitting:
1. Excessive topping-up of coolant. Even ifa 50/50 mix of antifreeze and water is used, the
SCA level can be severely reduced if the cooling system is frequently filled to the cap.
About 10% of the coolant is lost due to expansion as soon as the unit reaches operating
temperature. Low SCA levels result.
2. Refilling the cooling system with a 50/50 mix of new antifreeze and water after repairs,
but not adding a precharge of SCA. (The precharge amount of SCA is about three to four
times the service charge.) This results in low SCA levels.
3. Skippingthe SCA service charge at the oil and filter change interval. The amount of SCA
in the service filter is designed for the nominal 16 000 mile (400 h) maintenance interval.
Grossly extending the maintenance interval or failing to replace the coolant service filter
at this maintenance interval can result in dangerously low SCA levels.
Coolant Analys& and Testing--A test kit should measure the amount of the chemicals used
to prevent pitting if the testing results are to be meaningful [69]. In most of the world today,
these chemicals are nitrite alone or a mixture of nitrite and molybdate compounds. In the past,
some test kits have been used that measure pH, mercaptobenzotriazole (MBT), reserve alka-
linity, conductivity, and so forth. These chemical ingredients and coolant properties do not
relate to the ability of the coolant to protect against liner pitting.
Coolant Composition
Finally, if the engine design, cooling system hardware, and coolant maintenance practices
are correct, liner pitting can still occur due to coolant composition. Coolant is composed of
water, antifreeze, and SCA. Each is discussed below.
Water--Water with a high chloride or sulfate level aggravates liner pitting. Original Equip-
ment Manufacturer's (OEM) limits on chloride and sulfate levels in make-up water vary, but
are in the range of 40 to 100 ppm for each.
Antifreeze--Low silicate antifreeze from a reputable manufacturer is recommended by all
heavy-duty engine builders. The purchaser must be careful when selecting an antifreeze
because some blenders use off-grade glycol that is full of impurities. These impurities are often
corrosive and can cause liner pitting. This problem is much more likely to occur when glycol
supplies are tight and antifreeze prices are high [ 70].
SCA--Although few papers have been published on their mechanism of action [ 71], it is
known that supplemental additives provided by reputable suppliers generally prevent liner pit-
ting. Two basic types of SCAs are used in North America for heavy-duty equipment, phos-
phate/molybdate/nitrites and borate/nitrites. The borate nitrites should provide at least 1500
ppm nitrite as NaNO2 in the precharged coolant to protect against liner pitting. The combi-
nation of molybdate and nitrite has been found to be more effective than nitrite alone or
molybdate alone [ 72, 73].
HERCAMP ON CAVITATIONCORROSION OF DIESEL CYLINDER LINERS 121
80
Test Conditions:
80% duty cycle
0.8 mil double amplitude
4 hours
20,000 hertz
no additives
cast iron specimen
60
Courtesy Deere and
Company, 10/5/90
O3
g
{/)
0~
O
..J 40
.E
O)
20
0
0 20 40 60 80
Glycol Concentration (Percent In DI Water)
FIG. 12--Effect of concentration of uninhibited aqueous EG and PG on cavitation corrosion of cast iron.
Just as the phosphate/molybdate/nitrite SCA has been shown to give superior liner pitting
protection, aqueous coolant based on propylene glycol (PG) in place of ethylene glycol (EG)
also has been found to give additional protection against liner pitting [ 74- 76]. Figure 12 shows
the superior performance of PG over EG when no corrosion inhibitor is added.
Although the above coolant variables are the most important, others have been found to
affect liner pitting adversely:
1. Entrained air [37] and

2. Dirt in make-up water/circulating particulate in the coolant [ 77].
A procedure for troubleshooting liner pitting is as follows:
1. Take a random sample of coolant from 10% of the engines in the fleet. Low SCA levels
in many units indicates the potential for liner pitting problems.
2. Check the records for the unit in question for excessive coolant additions, cooling system
repairs, excessive maintenance interval length, skipped coolant filter at the maintenance
interval, use of maintenance filter after repair work when precharge should have been
used.
3. Check for aeration.
4. Check for abnormal block coolant pressure.
5. Check for abnormal block coolant temperature.
Future Work That Can Contribute Toward Reducing the Incidence of Cylinder-Liner
Cavitation Corrosion
Coolant
Standards need to be developed for both SCAs for heavy-duty engines and for antifreeze
coolants for heavy-duty engines that include a precharge dose ofSCA. These standards would
enable the heavy-duty engine and vehicle manufacturers to specify appropriate coolants, and
the coolant suppliers to identify their products as meeting those standards. Use of substandard
products would be minimized. Of equal importance, use of antifreeze coolant that contains a
precharge dose of SCA would dramatically reduce cavitation corrosion caused by topping up
with coolant without SCA.
The increased use of aqueous PG coolants would further reduce cavitation corrosion due to
their proven improved performance in the area. Furthermore, if heavy-duty vehicle manufac-
turers design their product to use nonaqueous PG coolant, cylinder liner cavitation corrosion
can be eliminated [ 78].
Engine Design and Materials

Widespread use of new acoustic and liner motion test methods on early prototypes of new
engine designs holds much promise to measure quickly the propensity of newly designed
engines to pit liners. This quick measurement of cavitation acoustics or liner vibration levels
allows time for design revisions before engine production.
Designing the block casting to accommodate proper coolant passage geometry, particularly
on the thrust side of the liner, avoids what can be difficult and expensive revisions later during
engine development.
Close attention to piston design to minimize liner vibration is required. Articulated piston
designs appear to offer distinct advantages.
The freedom to alter liner material to improve liner cavitation is severely restricted due to
the fact that the material used has generally been optimized for ID performance. Therefore,
liner OD coatings should receive greater attention. Although past experience with coatings has
not generally been favorable, possibly newer processes such as ion nitriding, physical vapor
deposition, ion implantation [ 79], and so forth should be evaluated.
Cooling System Design

The design of the vehicle cooling system must be controlled so that the coolant pressure and
temperature at the cylinder liner are not altered beyond the limits specified by the engine
builder. Also, the system should de-aerate rapidly and in all intended vehicle operating angles.
A positive contribution might be the addition of an overflow tank containing a rubber bladder
plus a coolant level gage. This would minimize coolant composition imbalance due to unnec-
essary topping up and also would minimize the oxygen content of the coolant to reduce
corrosion.
Maintenance Practices
The most important practice to be recommended is use of an antifreeze coolant truly for-
mulated for heavy-duty engines. Preferably the coolant should be PG based. The second is to
use appropriate field test methods to check the coolant condition [69,80].
Summary
The test tools, the design knowledge, the maintenance practices, and the coolants are avail-
able to eliminate cavitation corrosion failures of diesel cylinder liners as outlined above. Con-
tinued efforts by all involved will readily reverse the trend toward an increased incidence of
cylinder liner cavitation corrosion failures.
Acknowledgments
The author is grateful for the support of Fleetguard and C u m m i n s Engine Company in the
preparation and presentation of this paper. The skill and cooperation of Barb Rees who did
all the word processor work are especially appreciated.
References
[1] Leith, W. C. and McIlquham, W. S., "Accelerated Cavitation Erosion and Sand Erosion," Sympo-
sium on Erosion and Cavitation, ASTM STP 307, American Society for Testing and Materials, Phil-
adelphia, 1962, p. 47.
[2] Zhou, Y. K. et al., "'Cavitation Erosion of Diesel Engine Wet Cylinder Liners," Wear, Vol. 76, 1982,
pp. 321-328.
[3] Knapp, R. T., Daily, J. W., and Hammitt, F. G., Cavitation, McGraw-Hill, New York, 1970.
[4] Thiruvengadam A. Handbk f Cavitatin Ersin ice fNava Research AD-78773 United
States Department of Commerce, 1974, p. 290.
[5] Characterization and Determination of Erosion Resistance, ASTMSTP474, American Society for
Testing and Materials, Philadelphia, 1970.
[6] Erosion. Wear, and Interfaces with Corrosion, ASTM STP 567, American Society for Testing and
Materials, Philadelphia, 1974.
[7] Cavitation. Proceedings of the Sept. 1974 Conference of the Institution of Mechanical Engineers,
Mechanical Engineers Publications, Ltd., New York, 1975.
[8] Second International Conference on Cavitation, Institution of Mechanical Engineers Conference
Publications 1983-8, Mechanical Engineers Publication, Ltd., New York, 1983.
[9] Hoyt, J. W. and Furuya, O., Eds., Cavitation and Multiphase Flow Forum--1985, The American
Society of Mechanical Engineers, New York, 1985.
[ 10] Dowson, D. et al, Eds., Cavitation and Related Phenomena in Lubrication, Proceedings of the 1st
Leeds-Lyon Symposium on Tribology, Sept. 1974, Mechanical Engineering Publication Limited,
New York, 1975.
[11] Young, S. G. and Johnston, J. R., "Effect of Temperature and Pressure on Cavitation Damage in
Sodium," Characterization and Determination of Erosion Resistance, ASTM STP 474. American
[12] Fakkel, R. H. et al, "Comparison of Cavitation Tests on the SNR 300 Prototype Sodium Pump,
Carried Out Using Water at Room Temperature and Liquid Sodium at 580~ '' Cavitation, Insti-
tution of Mechanical Engineers, 1974, p. 193.
[13] Hoff, G. et al, "Rain and Sand Erosion, Phenomena of Material Destruction Caused by Repeated
Loads," Characterization and Determination of Erosion Resistance, ASTM STP 474, American
Society for Testing and Materials, Philadelphia, 1970, p. 354.
[14] Kozirev, S. P., "Eroding Capacity of Liquid Fuels and Lubricating Oils," Tractor Institute, City of
Chekhov, District of Moscow, USSR, pp. 30, 31.
[15] Thiruvengadam, A., Handbook of Cavitation Erosion, Office of Naval Research AD-787073, U.S.
Department of Commerce, 1974, pp. 6, 25-29.
[16] Summers D. A. ``The Vume Factrin Cavitatin Ersin Prceedings f 6th nternatinal Cn-
ference on Erosion by Liquid and Solid Impact, University of Cambridge, 1983, pp. 14-1 to 14-7.
[17] "Sonicator," Heat Systems--Ultrasonics Brochure SON-1 12-87, 1938 New Highway, Farming-
dale, NY, 11735, 1987.
[18] Plesset, M. S., "Cavitation and Cavitation Damage," Developments in Mechanics, Vol. 7, Proceed-
ings of the 13th Midwestern Mechanics Conference, University of Pittsburgh, 1973, pp. 55-69.
[19] Morch, K. A., "Cavitation Inception by Solid Particles in Liquid Flows," Cavitation and Multiphase
Flow Forum, American Society of Mechanical Engineers (ASME), 1986, pp. 16-18.
[20] Eisenberg, P.. "'Cavitation and Impact Erosion-Concepts, Correlations, Controversies," Character-
ization and Determination of Erosion Resistance. ASTM STP 474, American Society for Testing
[21] Fujihawa S. and Akamatsu T. ``Sme Recent Aspects fBubbe Dynamics Secnd nternatinal
Conference on Cavitation, Institution of Mechanical Engineers Conference Publications 1983-8,
1983, pp. 31-40.
[22] Plesset, M. S. and Ellis, A. T., "On the Mechanism of Cavitation Damage," Transactions of the
ASME, Oct. 1955, p. 1062.
[23] Plesset, M. S., "Pulsing Technique for Studying Cavitation Erosion of Metals," Corrosion, Vol. 18,
May 1962, pp. 181-188.
[24] Leith, W. C. and Thompson, A. L., "Some Corrosion Effects in Accelerated Cavitation Damage,"
Transactions oftheASME, Dec. 1960, pp. 795-807.
[25] Preiser, H. S. and Tytell, B. H., "The Electrochemical Approach to Cavitation Damage and Its Pre-
vention," Corrosion. Vol. 17, 1961, pp. 535t-549t.
[26] Schulmeister, R., "Vibratory Tests in Water on the Combined Action of Cavitation and Corrosion,"
Characterization and Determination of Erosion Resistance, ASTM STP 474, American Society for
Testing and Materials, Philadelphia, 1970, p. 124.
[27] Chance, R. L e t al, "Evaluation of Engine Coolants by Electrochemical Methods," Engine Coolant
Testing: Second Symposium, ASTM STP 887, Roy E. Beal, Ed., American Society for Testing and
Materials, Philadelphia, 1986, p. 117.
[28] Kallas, D. H., "Corrosion, Wear, and Erosion--An Ancient Art and a Modern Science," Erosion.
Wear, and Interfaces with Corrosion. ASTM STP 56 7, American Society for Testing and Materials,
[29] LaQue, F. L. and Hergenroether, E. J., "Corrosion Problems of the Automotive Engineer," SAE 337
Preprint, Presented at the Society of Automotive Engineers, (SAE) Summer Meeting, June 1949,
pp. 12-14.
[30] Speller, F. N. and LaQue, F. L., "Water Side Deterioration of Diesel Engine Cylinder Liners," Cor-
rosion, Vol. 6, 1950, pp. 209-215.
[31] Trock, B., "A Study of Cavitation Erosion," SAE 758 Preprint, Presented at the Society of Auto-
motive Engineers, (SAE) Summer Meeting, June 1956.
[32] Schrader, A. R., "Investigation of Cavitation Erosion in Diesel Engine Coolant Systems at the U.S.
Naval Engineering Experiment Station," Society of Automotive Engineers, (SAE) 759 Preprint, Pre-
sented at the SAE Summer Meeting, June 1956.
[33] Blackwood, A. K., "Coolant Side Corrosion of Diesel Cylinder Sleeves--Means for Reducing,"
Society of Automotive Engineers, (SAE) 760 Preprint, Presented at the SAE Summer Meeting, June
1956.
[34] Margulis, W. et al, "Cavitation Control Through Diesel-Engine Water Treatment," SAE Transac-
tions, Vol. 65, 1957, pp. 331-336.
[35] Joyner, J. A., "Reduction of Cavitation Pitting of Diesel-Engine Cylinder Liners," SAE Transac-
tions, Vol. 65, 1957, pp. 337-348.
[36] Dugdale, l. andWall, R., Eds., "Cavitation Corrosion and lts Prevention in Diesel Engines," British
Railways Board, London, Nov. 1965.
[37] Wiegand, H. and Schulmeister, R., "Test on the Influence of Gases on the Material Destruction
Through Cavitation and Corrosion in Regard to the Damage in Diesel Motors," Internal Combus-
tion Engine Research Union, April 1967 (translation from German).
[38] Wiegand, H. and Schulmeister, R., "Oscillator Test Relative to the Influence of Frequency, Ampli-
tude, Pressure, and Temperature Upon the Damage of Material Through Cavitation," Progress
Report of the Internal Combustion Engine Research Association, Feb. 1968 (translation from
German).
[39] Thiruvengadam, A., Handbook of Cavitation Erosion, Office of Naval Research AD-787073, 1974,
pp. 224-227.
[40] Hudgens, R. D. et al, "Refinement of the Vibratory Cavitation Erosion Test for the Screening of
Diesel Cooling System Corrosion Inhibitors," Engine Coolant Testing: State of the Art. ASTM STP
705, W. H. Ailor, Ed., American Society for Testing and Materials, Philadelphia, 1980, p. 242.
[41] Zbou, Y. K. and Hammitt, F. G., "Cavitation Damage of Diesel Engine Wet-Cylinder Liners,"
Worldwide Trends in Engine Coolants, Cooling System Materials and Testing. STP 811, Society of
Automotive Engineers, Warrendale, PA, (SAE) 900437, 1990, pp. 119-134.
[42] Pennisi, E., "Troubles with Bubbles Precedes the Popping," Science News, Vol. 139, No. 22, 1991,
p. 342.
[43] Yakovlev, V. V. et al, "Causes and Methods of Avoiding Premature Damage to the Cooling Surfaces
of Diesel Engine Cylinder Liners," Vestnik Mashenostroeniya, Vol. 55, Issue 7, 1975, pp. 26-29.
[44] Wall, R., "The Effect of Vibration on Aqueous Corrosion," Symposium on Cavitation Corrosion
and Its Prevention in Diesel Engines, British Railways Board, London, Nov. 1965, pp. 31-37.
[45] Zhou, Y. K. et al, "Cavitation Erosion of Cast Iron Diesel Engine Liners," Wear. Vol. 76, 1982, pp.
329-335.
[46] Abramenko, Y. E. and Abramenko, V. Y., "Cavitation-Erosion Failure of Cast Iron Components,'"
translated from Metallovedenic i Termicheskaya Obrabotka Metallov, No. 4, April 1989, pp. 20-
24.
[47] Hercamp, R. D. and Hudgens, R. D., "'Silicate Gelation in Heavy Duty Diesel Engine Cooling Sys-
tems," Society of Automotive Engineers, Warrendale, PA, (SAE) 852327, 1985.
[48] Hercamp, R. D., "Recycling Used Coolant from Heavy Duty Engines," Presented at the Forum on
Coolant Disposal and Recycling, National Automotive Radiator Service Association National
Meeting, Atlanta, 30 March 1989.
[49] Mitchell, W. A., "Statistical Treatment of Laboratory Data for ASTM D 1384-70 Using Soft Sol-
der," Engine Coolant Testing: State of the Art, ASTM STP 705, W. H. Ailor, Ed., American Society
for Testing and Materials, Philadelphia, 1980, pp. 220-232.
[50] Hudgens, R. D. et al, "'Refinement of the Vibratory Cavitation Erosion Test for the Screening of
Diesel Cooling System Corrosion Inhibitors," Engine Coolant Testing: State of the Art, ASTM STP
705. W. H. Ailor, Ed., American Society for Testing and Materials, Philadelphia, 1980, pp. 242-
243.
[51] Crosby, J. A. et al, "The Design and Development of the C Series Diesel Engine," Society of Auto-
motive Engineers, Warrandale, PA, (SAE) 870619, 1987, pp. 4, 5.
[52] Nestrides` E J. ``Examps f Cavitatin Crrsin in Engines and Design Cnsideratins Sym-
posium on Cavitation Corrosion and 1is Prevention in Diesel Engines, British Railways Board, Lon-
don, Nov. 1965, pp. 37-43.
[53] Collins, H. H., "Introduction to Cavitation Corrosion," Symposium on Cavitation Corrosion and
Its Prevention in Diesel Engines, British Railways Board, London, Nov. 1965, pp. 5-9.
[54] Fearon, W., "The Effect of Oil Cushioning on Piston Operation," Symposium on Cavitation Cor-
rosion and Its Prevention in Diesel Engines, British Railways Board, London, Nov. 1965, pp. 20-
30.
[55] Pinelli, J., Cummins Internal Report, 1990.
[56] Moore, H. C., "K 2000 Liner Vibration Test in Response to Cavitation Damage," Cummins Report
0771-88024, Cummins Engine Co., Columbus, IN, 1988.
[57] Wells, J. M., "NTC 400 Test A219-28", Cummins Report 0789-89001, Cummins Engine Co.,
Columbus, IN, 1989.
[58] Affenzeller, J. et al, "Some Investigations About the Cavitation on Cylinder Liners of Diesel
Engines," American Society of Mechanical Engineers (ASME) Reprint 80-DGP-27, New York,
1980.
[59] Yonezawa, T. and Kanda, H., "Analysis of Cavitation Erosion on Cylinder Liner and Cylinder
Block," Society of Automotive Engineers (SAE) 850401, Warrendale, PA, 1985.
[60] Lowe, A. S. H., "An Analytical Technique for AssessingCylinder Liner Cavitation Erosion," Society
of Automotive Engineers (SAE) 900134, Warrendale, PA, 1990.
[61] Sehulmeister, R., "Vibrating Tests in Water on the Combined Actin of Cavitation and Corrosion,'"
Characterization and Determination of Erosion Resistance, ASTM STP 474, American Society for
Testing and Materials, Philadelphia, 1970, p. 113.
[62] Hobbs, J. M., "Experience with a 20-KC Cavitation Erosion Test," Erosion by Cavitation on
Impingement, ASTMSTP408, American Society for Testing and Materials, Philadelphia, 1967,pp.
159-179.
[63] Garcia, R. and Hammitt, F. G., "Cavitation Damage and Correlation with Material and Fluid Prop-
erties," Transactions oftheASME, Dec. 1967, pp. 753-763.
[64] Thiruvengadam, A., "High Frequency Fatigue of Metals and Their Cavitation-Damage Resistance,"
Transactions oftheASME, Aug. 1966, pp. 332-340.
[65] McLaughlin, W. G. and Fast, M. L., "Cylinder Liner Corrosion," Cummins Research Laboratory
Summary Report, Cummins Engine Co., Columbus, IN, 21 March 1956.
[66] Retolaza, J., "Further Developments in the Manufacture and Applications of Steel Chrome-Plated
Cylinder Liners with Special Reference to Cavitation and Erosion Problems on the Water Side,"
Society of Automotive Engineers (SAE) 740315, Warrendale, PA, 1974.
[67] Gallant, T., "Coolant Hose for Heavy Duty Engines," Society of Automotive Engineers (SAE)
881272, Warrendale, PA, 1988.
[68] Stafford, R. J., "Investigation of Deposits on Water Pump Seal Faces," Engine Coolants Testing:
Third Volume, ASTMSTP 1192, American Society for Testing and Materials, Philadelphia, 1993,
pp. 205-214 (in this publication).
[69] Hudgens, R. D. and Bugelski, W. G., "Analysis of Coolants from Diesel Engines," SAE 1990 Trans-
actions, Journal of Commercial Vehicles, Section 2, Society of Automotive Engineers (SAE) 900435,
pp. 136-156.
[ 70] Personal communication from R. D. Hudgens, Fleetguard, to Joe Epner, Tradco, 31 May 1989.
[71] Lee, G. and McWhannell, D., "The Effect of Inhibitors on the Mechanism of Cavitation Erosion,"
Second International Conference on Cavitation, Institution of Mechanical Engineers Conference
Publications 1983-8, Mechanical Engineers Publication Ltd., New York, 1983, pp. 259-268.
[ 72] Hudgens, R. D., Engine Coolant Testing. Third Volume, ASTM STP 1192, American Society for
Testing and Materials, Philadelphia, 1993, pp. 128-148 (in this publication).
[ 73] Hercamp, R. D. and Hudgens, R. D., "Cavitation Corrosion Bench Test for Engine Coolants," Soci-
ety of Automotive Engineers (SAE) 881269, Warrendale, PA, 1988.
[74] Hercamp, R. D. et al, "'Aqueous Propylene Glycol Coolant for Heavy Duty Engines," SAE 1990
Transactions. Journal of Commercial Vehicles. Section 2, SAE 900434, pp. 105-134.
[75] "Water Treatment Recommendations for Engine Cooling Systems," Service Bulletin 4-2429A,
Waukesha Dresser, Waukesha, WI, 1 Feb. 1987.
[76] Oakes, B. D., "Cavitation Corrosion," Engine Coolant Testing: State oftheArt, ASTMSTP 705, W.
H. Ailor, Ed., American Society for Testing and Materials, Philadelphia, 1980, pp. 284-294.
[77] Wang, X. and Hammitt, F. G., Cavitation and Multiphase Flow Forum, Vol. 9, 1984, pp. 46-48.
[ 78] Huang, L. K., "Evaluation of Propylene Glycol as a Potential Engine Coolant," Final Report to Arco
Chemical Co. from Cummins Engine Co., 14 Jan. 1991.
[79] Hu, W. W., et al, "Cavitation Erosion of Ion-Implanted 1018 Steel," Materials Science and Engi-
neering, Vol. 45, 1990, pp. 263-268.
[80] Hemmes, P. et al., Engine Coolant Testing: Third Volume, ASTM STP 1192, American Society for
Testing and Materials, Philadelphia, 1993, pp. 165-179 (this publication).
DISCUSSION
Joseph A. L i m a I (written discussion)--First, I am wondering about alternatives for the SCA

approach to prevent liner pitting. Many European coolants contain some nitrite, and appar-
ently some of these are in successful use in heavy-duty equipment in Europe without SCAs. Is
this a successful approach in Europe, and, if so, why not in the United States?
Second, I note that no "leakage" allows an overconcentration to develop, whereas excess
leakage seems to cause a less than optimal SCA level. For a long-term SCA approach, it seems
a controlled removal, almost what is done in cooling tower treatments, of some of the coolant
could help maintain an acceptable level of total dissolved solids (TDS) and permit addition of
SCAs without the attendant overconcentration problem. Comment?
R. Hercamp (author's response)
1. If the coolant contains a m i n i m u m of 1500 p p m nitrite (as NaNO2) after being diluted to
its recommended usage level (generally 50% glycol), it will control cavitation corrosion in most
engines. In this case, an initial charge of SCA would not be required. However, periodic addi-
tion of SCA to maintain the protection would still be required. This is necessary to compensate
for the effects o f additive depletion and dilution with extended use of the coolant, which is
typical with heavy-duty engines.
2. A better approach would seem to be that of using coolant that contains additives for
reducing cavitation corrosion. Use of this type coolant for both initial fill and for "make-up"
would reduce the starting TDS level and would also allow the amount of solids added at the
maintenance interval to be reduced. The SCA maintenance dose could be reduced to com-
pensate for only depletion rather than a larger dose to take care of depletion plus some average
estimated rate of dilution.

DISCUSSION ON CAVITATION CORROSION 127
George R. Otterman 2 (written discussion)

1. Does cavitation corrosion occur on all cycles of the engine stroke?
2. In the real world, will good engine design and/or quality SCAs completely eliminate cav-
itation corrosion, or do you always expect to see and live with the problem?
3. Why phosphate the liner wall?
R. Hercamp (author's response)
1. I don't know, but the normal location of the most severe damage is on the thrust side of
the cylinder liner. This would indicate that the corrosion occurs primarily during the firing
stroke. Liner vibration measurements and liner vibration models reported in the literature
indicate that m a x i m u m liner vibration generally occurs very soon after firing.
2. Engines can certainly be designed to eliminate cavitation corrosion. Whether these
engines would be economical and practical is a much more difficult question to answer. At the
other extreme, some engines have been designed that are so severe that the proper use of qual-
ity SCAs is not sufficient to control cavitation corrosion. If liner vibration is too severe, coolant
additives are not the answer. The design must be modified to reduce vibration.
I expect that we will need to live with the cavitation corrosion problem for many years to
come. Therefore, it is imperative that A S T M D 15 continue its work to develop standards for
both SCAs and antifreeze coolants for heavy-duty engines that incorporate additives to reduce
cavitation corrosion.
3. Some cylinder liners are coated with a phosphate-rich coating as the last step in their man-
ufacture. The coating generally serves two purposes, that of a rust preventative during storage
and as a solid film lubricant for the liner ID during engine operation. The phosphate coating
has no effect on the rate of cavitation corrosion damage.
2 Electro-Motive Division of General Motors, LaGrange, IL.

R. D. Hudgens j
Phosphate-Molybdate Supplement Coolant

Additives for Heavy Duty Diesel Engines
REFERENCE: Hudgens, R. D., "Phosphate-Molybdate Supplement Coolant Additives for
Heavy Duty Diesel Engines," Engine Coolant Testing: Third Volume, ASTM STP 1192, R. E.
Beal, Ed., American Society for Testing and Materials, Philadelphia, 1993, pp. 128-148.
ABSTRACT: Supplemental coolant additives (SCAs) are commonly used in heavy duty diesel
service. The SCA boosts the performance of the typical antifreeze coolant and replaces additives
lost to depletion and coolant dilution.
In the early 1970s chromate based SCA's were phased out because of chromate's toxicity and
adverse environmental impact. Borate-nitrite based SCAs replaced the chromates and are widely
used today.
This paper describes the development of a third generation phosphate-molybdate additive
package. Relative to the borate-nitrite SCAs, the phosphate-molybdate supplemental coolant
additive gives better overall aluminum protection. Furthermore if overtreated, it is less likely to
cause solder bloom, silicate gelation, and water pump leakage.
KEYWORDS: supplemental coolant additives, heavy duty diesel, engine coolants, antifreeze-
coolants, corrosion inhibitors, corrosion, cavitation
Background
Supplemental coolant additives (SCAs) have been used in heavy duty diesel engine coolants
since the mid 1950s. The primary function of the SCA is to stop liner pitting (cavitation-cor-
rosion) as well as prevent deposits and scale buildup in the cooling system. The typical light-
duty automotive antifreeze coolant does not provide this type of protection. An SCA is also
used to replenish the additives consumed in normal service. This is vital for heavy duty vehi-
cles where the service life of the coolant can range from 200 000 to 500 000 miles (320 000 to
800 000 km) or 5000 to 12 000 h.
The typical supplemental coolant additive is formulated to work in water alone or when
used in conjunction with an antifreeze coolant. F o r this reason there is some overlap between
the additive package of an SCA and antifreeze. Both are likely to contain a phosphate and a
borate buffer, or both, as well as some o f the same corrosion inhibitors (nitrate, silicate, and
tolytriazole). Table 1 compares the features or functions of a heavy duty supplemental coolant
additive to those of a light duty antifreeze coolant.
Chromate based SCAs were used in the United States from about 1955 to 1975. Chromates
were largely phased out in the 1972 through 1975 time frame because of environmental con-
cerns, and to a lesser e antifreeze compatibility problems. The chromates were replaced
with borate-nitrite based SCAs. Borate-nitrite supplemental additives specifically formulated
to protect the cooling system of a heavy duty highway diesel engine first appeared in the late
1950s. Those products were very similar to the borate-nitrites used today.
Senior staffchemist, Fleetguard, Inc., Cookeville, TN 38502.
128
HUDGENS ON PHOSPHATE-MOLYBDATE SUPPLEMENT 129
TABLE 1--Comparison of a heavy duty supplemental coolant additive to light duty antifreeze coolant.
Function Heavy Duty SCA Typical Antifreeze/Coolant
Buffer yes yes

General corrosion protection yes yes
Liner pitting protection yes no
Scale/deposit control yes no
Development of a Third Generation Supplemental Coolant Additive

In 1980, a Fleetguard/Cummins Engine Company technical group was formed to consider
the need for an improved supplemental coolant additive. This group consisted of chemists and
technical service engineers from both companies.
The technical group's field survey showed the following problems associated with the use of
borate-nitrite based SCA's:
9 Solder bloom or corrosion

9 Antifreeze compatibility
9 A l u m i n u m corrosion (especially with high mineral content make-up water)
The survey also showed that heavy duty cooling systems were often fouled with oil, which
can cause heat transfer and corrosion problems. This was a problem not being addressed by
the SCAs available at that time.
With our field survey as a guide, we began the development of a new supplemental coolant
additive. The primary objectives were to formulate a product that gave good aluminum and
solder protection under a wide variety of service conditions and minimize the problems that
occur when an SCA is overtreated. For many operators of heavy duty equipment it is difficult
to control the amount of SCA in the engine coolant.
Figure 1 shows the range of SCA concentration in the coolant of vehicles from one on-high-
way fleet. Many of the vehicles have 3 to 4 times the recommended concentration of supple-
mental coolant additive. This overconcentration o f SCA can trigger the following problems:
9 Silicate gelation
9 Solder bloom (corrosion)
9 Water pump leakage
We began our laboratory work by carrying several different additive mixtures through pre-
liminary/screening type tests. In some instances we took the same mix of chemicals and simply
interchanged phosphate for borate to understand which buffer made the better foundation on
which to build. The bulk of the information generated indicated that a phosphate buffered
additive package gave better corrosion protection, especially for aluminum. Figures 2 and 3
compare phosphate and borate as to their aluminum corrosion protection and aluminum cav-
itation protection. It is clear that in both tests phosphate gives significantly better protection.
These results are consistent with those reported by Oakes [1]. In addition, Vukasovich and
Sullivan of Climax Molybdenum [2] and Wiggle et al. of Ford Motor Company [3] reported
borate to be aggressive to aluminum under cavitation and heat transfer conditions.
Another reason for focusing our development effort on a phosphate buffered product was
that it coupled well with molybdate. In the early 1980s, molybdate was being promoted as an
effective all around corrosion inhibitor by Climax, General Motors, and Ford Motor Com-
pany. We did not observe a strong synergism when we tested borate with molybdate. Our find-
12
11
10
o 8
~ 7
~ 6
.Q
!
0 0.5 1 1.5 2 2.5 3 3,5 4 6.5
SCA Concentration
NOTE: Concentration is given relative to the recommended amount.
Therefore 1 is the recommended treat level.
FIG. 1--Variation in SCA concentration from several vehicles in a large on-highwayfleet.
AI 319 corrosion in MPY

120
IOOF " ~ .............................
80
60
40
20
0
1000 ppm 5000 ppm 10000 ppm
9 Borate(Na2B407-5H20) [ ] Phosphate(K2HPO4)
Base s o l u t i o n c o n t a i n s 100 ppm chloride
FIG. 2--Aluminum corrosion with phosphate and borate buffers.
Weight loss in mg
300
250 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
200-
150
100
50 .... ~ ............
0
Borate(Na2B407-5H20) Phosphate(Na2HPO4)
Borate/phosphateconcentration5000ppm
FIG. 3---Aluminum cavitation-corrosion with phosphate and borate buffers.
ings were consistent with data reported by Vukasovich [2] a n d Wiggle et al. [3], that phos-
phate-molybdate blends were synergistic when it came to protecting cast iron, high lead solder,
and a l u m i n u m from corrosion. In addition, our laboratory work showed that molybdate was
synergistic with nitrite in protecting cast iron from cavitation corrosion protection. This find-
ing is also supported by the work of Weber et al. [4] a n d Al-Borno et al. [5].
Laboratory and Dynamometer Test Methods

The test scheme used in this development work is outlined in Table 2 a n d the details of the
various test methods discussed elsewhere in the literature. Rowe [6] in his overview of engine
cooling system inhibitors suggests the same glassware to field test progression.
TABLE 2--Supplemental coolant additive testing scheme.
Level I a Level 2b Level 3 Level 4
Glassware corrosion erosion corrosion dynamometer test field test

Electrochemical German FVV
Bench cavitation corrosion water pump leakage
Aluminum hot surface D 2570 simulated service
corrosion (D 4340)
Antifreeze compatibility D 2809 water pump
cavitation
Detergency/dispersency
o These tests are more "beaker" or "bench top" test methods that use small quantities of coolant. Test
duration is often short, and several tests can be run at the same time so that they make a good screening
tool.
b These methods are more involved and often longer in duration. The test stands are larger, more
complex, and in some ways simulate actual field conditions.
TABLE 3--Supplemental additive comparison.
Function Borate-Nitrite Phosphate-Molybdate
Buffer borate phosphate

Liner pitting protection nitrite molybdate and nitrite
Corrosion inhibitors nitrate and silicate nitrate and silicate
Glassware tests were run in a mixture o f ethylene glycol and corrosive water, corrosive water
only, and hard water. Test methods included both ASTM Test Method for Corrosion Test for
Engine Coolants in Glassware (D 1384) and Case MT-817. Later in the paper as results are
presented, more information will be given on the test methods used.
The dynamometer testing was conducted using the Cummins NTC-400 oil consumption
test. This test, which runs at rated speed and 5% overfuel, was developed to measure lubricat-
ing oil's ability to prevent piston deposits and limit oil consumption. The fact that the engine
is overfueled makes it run hot and pit liners. It is an excellent tool to measure an SCAs ability
to control liner pitting and prevent hot surface scaling. Hercamp [ 7] has thoroughly described
the test method in an earlier paper.
The field test methods and results will be presented in a later section of this paper.
Chemistry Comparison and Summary of Advantages

Table 3 compares the inorganics present in a borate-nitrite versus a phosphate molybdate
SCA. There are several organic components in each (surfactants, scale inhibitors, defoamers,
and corrosion inhibitors), but the inorganics still compose 75 to 90% of the additive package.
Borate is replaced by phosphate and molybdate when the chemistry of the two SCA types
are compared. The phosphate based system exhibits more synergism among its components,
which allows the use of less nitrite and silicate. Lower nitrite levels in the coolant result in less
solder corrosion while lower silicate concentration reduces the chance of silicate gelation.
Table 4 lists the advantages of the phosphate-molybdate additive package.
Phosphate-Molybdate Versus Borate-Nitrite Performance

The advantages listed for the phosphate-molybdate system will be discussed in more detail
in this section. Only data relative to how the two SCAs compare are included. (General test
results on the phosphate-molybdate system are included in Table 5).
Ahtminum Protection
A borate-nitrite SCA depends heavily on silicate to protect aluminum. This presents a prob-
lem when the make-up water contains high levels of hardness ( > 250 ppm). Calcium and mag-
TABLE 4--Phosphate-rnolybdateadvantages.
Better aluminum corrosion protection

Better solder corrosion protection
Improved antifreeze compatibility
Improved liner pitting protection
Lower toxicity
Fewer water pump leaks
More tolerant if overdosed
TABLE 5--Phosphate-molybdate performance in ASTM tests.
Test Results versus ASTM D 3306 Requirements
Test Method and Conditions Results, mg Required, mg
D 1384 glassware at copper 4.0 10

recommended dose in 50/50 solder 0.4 30
(v/v) ethylene glycol-corrosive brass 3.0 10
water steel 0.9 10
cast iron 1.4 10
aluminum 3.3 30
D 4340 aluminum hot surface at 0.5 mg/cm2/week 1.0 mg/cm2/week
recommended dose in 25/75
(v/v) ethylene glycol-chloride
water
D 2809 water pump cavitation 8.0 8.0 min
at ~ recommended dose in
17/83 (v/v) ethylene glycol-
corrosive water
D 2570 simulated service at copper 7.7 20
recommended dose in 44/56 solder 22.0 60
(v/v) ethylene glycol-corrosive brass 6.7 20
water steel 0.3 20
cast iron 2.7 20
cast aluminum 1.0 60
NOTE; Recommended dose of phosphate-molybdate SCA 3% by volume for liquid and 6000 mg/L
for solid.
nesium (water hardness) react with silicate to form a precipitate. The precipitated silicate is no
longer able to maintain a protective film resulting in aluminum corrosion. Figure 4 presents
data obtained from a D 1384 glassware test using hard water.
It is evident that the phosphate-molybdate additive provides improved aluminum protec-
tion in hard water. This occurs because phosphate and molybdate help provide aluminum
protection even when silicate precipitates from solution. Where aluminum cavitation is a
problem and hot surface corrosion protection is needed, phosphate is the buffer of choice (Fig.
3). Wiggle et al. [8] recommended against the use o f high levels o f borate where hot surface
a l u m i n u m corrosion is a problem (engine block and head). This is shown in Fig. 5 which pre-
sents data from the British Standard Recirculating Rig Test (BS 5117), which tests for hot sur-
face aluminum protection. A borate-nitrite SCA fails after only 12 days while the phosphate-
molybdate additive provides hot surface aluminum protection for 34 days.
Solder Protect ion

A borate-nitrite SCA uses relatively high levels of nitrite to protect liners from cavitation
corrosion or pitting. While nitrite effectively reduces cast iron liner pitting it can be aggressive
to solder. Rowe [6] warned of this problem several years ago. Our experience shows that solder
corrosion can be initiated by sodium nitrite levels in excess of 3000 ppm. This concentration
is exceeded with twice the recommended dosage of the typical borate-nitrite SCA; something
that regularly occurs in the field. Table 6 presents dynamometer test data which shows the
effect o f nitrite on solder corrosion based on the soluble lead in the coolant.
ASTM D 1384 glassware corrosion tests were run on a borate-nitrite and phosphate-molyb-
date SCA at the recommended concentration in a 50/50 by volume mix of ethylene glycol and
ASTM corrosive water. The phosphate-molybdate SCA provides superior solder protection as
Aluminum Weight Loss in mg

250
~\\\\\\
200 . . . . . . . . . . . . . . . . . . ~\\\~
,\\\\\\
\\\\\\\
~\\\\\\
\\\\\\\
150 . . . . . . . . . . . . . . . ,~\\\\X
~ \ \ \ \ \ \

\ \ \ \ \ \ \
100 . . . . . . . . . \ \ \ \ \ \ \
, \ \ \ \ \ \
~ \ \ \ \ \ \
\ \ ~ \ \ \ ,
,-~\\\\'-~
0 . . . . . . . . . .
0 k\\~'~\\~
Phosphate-Molybdate Borate-Nitrite
SCA concentration @ 3000 ppm

FIG. 4--Glassware corrosion test o f SCAs run per D 1384 except hard water substituted.for standard
corrosive water.
Days to Failure
35
0 . . . . . . . . . . . . .
25 . . . . . . . . . . . . .
20 . . . . . . . . . . . . .
15 . . . . . . .
10 " ~ . . . . .
. . . . .
0
Borate-Nitrite Phosphate-Molybdate
Loss of t e m p e r a t u r e c o n t r o l e n d s test
FIG. 5--British Simulated Service Test (BS 5117) o f borate nitrite versus phosphate molybdate SCA.
HUDGENSON PHOSPHATE-MOLYBDATESUPPLEMENT 135
TABLE 6--Nitrite concentration versus solder corrosion.
Nitrite Concentration (ppm NaNO2) 3500 690

Dosage versus Recommended 2X lX
Lead in Coolant, ppm 21 l
shown in Fig. 6. This is due in part to the nitrite concentration, which is less than half that
found in borate nitrite SCAs.
Silicate Gelation
On occasion, a problem develops with SCA/antifreeze compatibility. Silicate gelation is the
most recent problem of this nature. It is formed when silicate polymerizes or reacts with itself.
Our investigation of the problem showed that the following coolant conditions promote the
formation of gel:
9 Silicate concentration ( > 1500 ppm)

9 Glycol concentration ( > 60%)
9 Total dissolved solids ( > 3%)
The typical field problem involves a high silicate antifreeze at concentrations of 60 to 70% with
a double or triple dose of SCA.
Solder Weight Loss in mg

160
140
120
100
80
60--
40-
20 ....
0 -i
Phosphate-Molybdate
9 1X Concentration ~ 3X Concentration
50/50 bv ethylene glycol-corrosive water
FIG. 6--Solder weight loss in A S T M D 1384. SCAs run at recommended and 3 times recommended
concentration (for liquid SCAs this equals 3% and 9% by volume).
136 ENGINE C O O L A N T TESTING: THIRD VOLUME
Solids in ml
3
.5 . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . .
1.5 . . . . . . . . . . . . . . . . .
1 "
0.5 . . . . . . .
0
Phosphate-Molybdate Borate-Nitrite
9 50% Antifreeze [ ] 60% Antifreeze

High silicate antifreeze
FIG. 7--SCA compatibility with a high silicate antifreeze. SCA concentration twice that recommended.
Figure 7 compares the two types of SCA as far as their tendency to form silicate gel. The test
was run by heating the mixture of high silicate antifreeze and SCA at 190*F (88~ for 24 h in
a 100-mL centrifuge tube. The tube was removed from the oven, cooled to room temperature,
centrifuged at 500 RCF for 15 min, and the volume ofsolids reported. Because the phosphate-
molybdate SCA contains less silicate, it forms less gel.
Liner Pitting
The most important function of an SCA is to prevent the pitting of wet sleeve liners in a
heavy duty diesel engine. When pitting penetrates the liner wall, coolant leaks into the crank-
case making a rebuild necessary ($4000 to $5000 for a 10- to 14-L engine). Liners can be per-
lorated by pitting (cavitation-corrosion) with a corrosive coolant in as little as 30 000 miles
(48 000 kin).
The nitrite in a borate-nitrite SCA provides good liner pitting proteclion. We wanted to keep
the good performance of a borate-nitrite SCA but limit the amount of nitrite. Lower nitrite
levels reduce the chance of solder bloom and corrosion when the SCA concentration builds
up in the cooling system. Our laboratory evaluation showed that nitrite and molybdate work
synergistically to protect ferrous metals. This is shown in Fig. 8 where glassware type tests were
run with 1010 steel in water [3]. A mixture of 250 ppm of nitrite and molybdate performed
much better than 500 ppm of each inhibitor separately. Other data published in the literature
confirms this synergism [4].
With a mixture of nitrite and molybdate, we found that nitrite could be reduced by 50 to
60% with improved liner pitting protection. Figure 9 shows results from a cavitation-corrosion
bench test in ASTM corrosive water. An ultrasonic welding system is used to vibrate a cast
iron button in a beaker ofcoolant at 20 000 eps and an amplitude of 0.4 mils. After 22 h, the
HUDGENS ON PHOSPHATE-MOLYBDATESUPPLEMENT 137
Corrosion Rate in Mils/Year

14
12 ,~\\\\\"~
\\\\\\\\
\\\\\\\\
\\\\\\\\
10
\\\XX\\\
. .
,'x\\\\\"~
\ \ \ \ \ \ \ \
\ \ \ \ \ \ \ \
,\\~\\\\
\\\N\\\"
Na2MoO4@500 ppm NaNO2@500 ppm Each@250 ppm
REF:Weber, CORROSION, VOL.42, #9, p542

FIG. 8--Glassware corrosion tests of molybdate and nitrite alone compared with a molybdate-nitrite
mixture.
Weight Loss in mg
300
250
2OO
150
/////
/////
100 /////
/////
/////
5O
0
Corrosive water Borate-Nitrite Phosphate-Molybdate
9 1/2 Concentration ~ 1X Concentration [ ] Corrosive water

SCA's run in c o r r o s i v e w a t e r
FIG. 9--Liner pitting (cavitation-corrosion) bench test of SCAs at half and recommended concentration.
Liner Pitting Severity (Area Count)

3w ........................................................................................................................................................................
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
"~ ....................................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
,oof ............................................................... ................................................. .................................................
0 2 4 6 8 10 12 14
SCA Concentration in ppm x 103
-~- Borate - Nitrite I Phosphate- Molybdate

FIG. lO--Liner pitting dynamometer test of SC'4s over a range of concentration.
button is removed from the apparatus, cleaned, and weighed. The details of the test method
are given in the paper by Hercamp and Hudgens [7].
Bench type tests give an indication of how a product will perform, but they cannot guarantee
that the product will work well in the field. For this reason and the critical nature of liner pit-
ring, we ran further comparative tests on the Cummins NTC 400 dynamometer test [ 9]. Figure
10 presents data where both a borate-nitrite and the phosphate-molybdate SCA were tested at
various concentrations in a 50% by volume solution of GM-6038 antifreeze. The liner pitting
count (y axis) is a relative measure of the area of the liner that has cavitation-corrosion dam-
age. In this severe test the phosphate-molybdate SCA required a concentration of 7500 ppm
compared to 12 000 to 13 000 ppm of the borate-nitrite SCA to eliminate liner pitting.
Toxicity
The health affects associated with chemical products are of growing concern. Figure 11 pre-
sents rat acute oral toxicity data for two SCAs. The lethal dose for half the test group of rats
(LDs0) is about 4000 mg/kg of body weight with the borate nitrite SCA. The phosphate-molyb-
date additive is somewhat less toxic with an LDs0 of 6200 mg/kg of body weight. This lower
toxicity is the result of the phosphate buffer and the lower level of the toxic sodium nitrite (LD~0
100 mg/kg). See Figs. 12 and 13.
Water Pump l,eakage

Water p u m p leakage may occur when supplemental coolant additives are overdosed or
slowly built up in the coolant over time. Our investigation of water pump leakage has shown
the following conditions increase leakage:
mg per Kg of body weight (Thousands)
. . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . .
4 . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . .
0
Toxicity tests run on liquid concentrate

FIG. 11--SCA oral toxicity in rats (LDso).
9 High glycol concentration ( > 60%)

9 High total dissolved solids ( > 3% to 4%)
9 Debris in the coolant
9 Excessive silicate concentration
Note that the above conditions are very similar to those that promote silicate gelation.
A simulated service type test was set up at Cummins Engine Company where a water pump
driven by an electric motor circulated coolant through an engine block. Three times the rec-
o m m e n d e d dose of SCA was added to a high- and low-silicate antifreeze-coolant, and the effect
on water p u m p leakage was monitored. The results of this test are presented in Table 7. The
test results show that when SCAs are over concentrated, the phosphate-molybdate SCA was
not as likely to cause water pump leakage as the borate-nitrite additive.
Field Test
Before the release of the phosphate-molybdate additive, it was field tested in three locations
as outlined below. A borate-nitrite SCA was run as a control or reference at each location.
TABLE 7 - - W a t e r pump leakage versus SCA type.
SCA Type 70% High Silicate 50% High Silicate 50% Low Silicate
Borate-nitrite at 3X leakage leakage no leakage

Phosphate-molybdate at 3X no leakage no leakage no leakage
No SCA added no leakage not run not run
MB R E S E A R C H LABORATORIES, INC.
PROJECT NUMBER : MB 91-753 A
TEST ARTICLE : DCA 4 Liquid (Phosphate-Molybdate)
SPONSOR : FLEETGAURD, INC.
'rrrLE : SINGLE DOSE ORAL TOXICITY IN R A T S / L D 50 IN R A T S
PROTOCOL # : 68D R/A
ABSTRACT
Method Synopsis - Five healthy male and five healthy female Wistar Albino rats were dosed orally
with D C A 4 Liquid at 5.0 g/kg of body weight, Although only one anlmRl died at the initial dose
level of 5.0 g/kg, additional groups of 5 males and 5 females were dosed per sponsor request as
indicated below. The rats were observed 1, 2 and 4 hours post dose and once daily for 14 days for
toxicity and pharmacological effects. All animals were observed twice daily for mortality. Body
weights were recorded pretest, weekly, at death and at termination in the survivors. All animals
were examined for gross pathology. Abnormal tissueswere preserved in 10% buffered formalin for
possible future microscopic examination. The L D 50 vras calculated by the method of Litchfield &
Wilcoxon.
S u m m a r y - Mortality response to the four dose levels were:
Dose # Treated # Dead

wk~ MfF lWF
5.00 5/5 0/1

5.60 5/5 2/1
6.25 5/5 2/4
7.00 5/5 4/4
The deaths occurred on the day of dosing and were preceded by physical signs of lethargy, ataxia and
piloerection. Necropsy of the dead anirn~l~ revealed abnormalities of the lungs, liver, kidneys,
spleen, pancreas and gastrointestinal tract. Physical signs noted in survivors included lethargy,
chromorhinorrhea, prostration, ptosis, ataxia, brown staining of the nose/mouth area and wetness of
the anogenital area. In addition, an ocular abnormality was noted in one animal. Body weight
increases of sur~vors were generally normal. Some instances of slight weight loss were noted in
females during the second week of the observation period. Necropsy results of survivors were
normal.
Conclusion - The LDs0'S and 95% Confidence Limits are: males - 6.2 (5.6-6.9) g/kg;, females - 6.0
(5.3-6.7) g/kg and males & females combined - 6.2 (5.7-6.7) g/kg.
FIG. 12--Single dose oral toxicity in 50 rats/LD.
(1) Strip m i n e near Bowling Green, KY: 4 bulldozers with C u m m i n s K-6 engines, 1-year
duration.
(2) O n - h i g h w a y t r u c k fleet near Cookeville, TN: 18 C u m m i n s N T engines, 100 000 miles
(160 000 k m ) d u r a t i o n .
(3) O n highway fleet in O k l a h o m a City, OK: 6 C u m m i n s L-10 engines, 60 000 miles
(96 000 km) d u r a t i o n .
MB R E S E A R C H LABORATORIES, INC.
PROJECT NUMBER : MB 91-752 A
TEST ARTICLE : DCA 2 Iiqmd (Borate-Nitrite)
SPONSOR : FLEETGAURD, INC.
TITLE : SINGLE DOSE ORAL TOXICITY IN RATS/LD 50 IN R A T S
PROTOCOL # : 68D R/A
ABSTRACT
Method Synopsis - Five healthy male and five healthy female Wistar Albino rats were dosed orally
with DCA 2 Liquid at 5.0 g/kg of body weight. Since more than five animals died at the initial dose
level, additional groups of 5 males and 5 females were dosed as indicated below. The rats were
observed 1, 2 and 4 hours post dose and once daily for 14 days for toxicity and pharmacological
effects. All animals were observed twice daily for mortality. Body weights were recorded pretest,
weekly, at death and at termination in the survivors. All animals were examined for gross
pathology. Abnormal tissues were preserved in 10% buffered formalin for possible future microscopic
examination. The LD 50 was calculated by the method of Litehfield & Wileoxan.
Summary - Mortality response to the five dose levels were:
Dose # Treated # Dead

~z M/F M/F
2.6 0/5 -/0

3.2 5/5 0/3
5.0 5/5 2/4
5.6 5/5 3/5
6.0 5/0 5/-
The deaths occurred on the day of dosing and were preceded by physical signs of lethargy, ataxia,
flaccid muscle tone, prostration and ptosis. Necropsy of the dead animals revealed abnormalities of
the lungs, liver and gastrointestinal tract. Physical signs noted in survivors included lethargy,
chromorhinorrhea, ataxia and wetness of the anogenital areB. Body weight increases and necropsy
results of survivors were normal.
Conclusion_ - The LD50's and 95% Confidence Limits are: males - 5.3 (4.9-5.8) g/kg;, females - 3.4
(2.7-4.4) g/kg and males & females combined - 4.2 (3.4-5.2) ghtg.
FIG. 1 3 - - S i n g l e dose oral toxicity in 50 rats/LD.
I n e a c h c a s e t h e field t e s t v e h i c l e s w e r e split b e t w e e n t h e p h o s p h a t e - m o l y b d a t e a n d b o r a t e -
n i t r i t e S C A s . B u n d l e s o f c o r r o s i o n t e s t c o u p o n s w e r e i n s t a l l e d in t h e u p p e r r a d i a t o r h o s e o f
t h e o n - h i g h w a y t r u c k s a n d m o u n t e d in t h e t o r q u e c o n v e n e r h o u s i n g o f t h e dozers. A t t h e
Cookeville location, each SCA had 5 units at the recommended concentration and 4 units
c h a r g e d t o h a l f t h e r e c o m m e n d e d c o n c e n t r a t i o n . T h e c o n c l u s i o n s f r o m t h e field test w e r e as
f o l l o w s [10]:
9 The phosphate molybdate provided superior solder corrosion protection.

9 B o t h S C A s p r o v i d e d e x c e l l e n t l i n e r p i t t i n g p r o t e c t i o n e v e n w h e n d o s e d at h a l f t h e rec-
ommended concentration.
9 The phosphate-molybdate provided slightly better aluminum protection at the reduced

concentration.
Since the release of the phosphate-molybdate additive package, we have continued to field
test at an on-highway fleet near Columbus, IN. The test has involved 30 to 34 units with Cum-
mins 14-L engines that accumulate about 5000 miles (8000 km) per week. The SCA concen-
tration is controlled at 0.5 to 1.5 times that recommended (3000 to 9000 mg/L) with the use
of coolant filters that contain SCA. Engines in this fleet are pulled from service at 500 000 to
650 000 miles (800 000 to 1 050 000 km) and overhauled because o f oil consumption. The
results from this test fleet show are as follows:
9 No liner pitting or crevice seal failure

9 Water p u m p seal life in the range 450 000 to 650 000 miles (725 000 to 1 050 000 km)
Engine Cleanliness Issues

Heavy duty service with its greater power density and load factor requires efficient removal
of large quantities of waste heat from the engine. If an engine is to deliver a half million miles
(approximately 804 650 km) or more of service, the cooling system must he free of deposits
and fouling. Deposits that build up in the cooling system can cause overheating, lube oil dete-
rioration, and excessive oil consumption. This subject is discussed in detail in a previous paper
by the author [11]. Some of the reasons that heat transfer problems develop in the cooling
system are hard water precipitation, oil fouling, and additive solubility problems. These are
discussed in more detail below.
Engine Scaling
Silicate, phosphate, and molybdate can precipitate with water hardness (calcium and mag-
nesium) and form deposits (scale) on hot cooling system surfaces. This usually occurs on the
upper portion o f the liner and in the head. Scaling in an engine is very similar to what occurs
in a tea kettle or hot water heater. The hardness in water is less soluble at high temperatures
and tends to drop out of solution and adhere to hot metal surfaces. If this scale is allowed to
build up in engine cooling systems, oil consumption and head cracking can occur.
In the 1950s German OEMs experienced scaling problems when phosphate inhibited anti-
freeze coolant was used with hard water. Since that time there has been a reluctance by some
to use phosphate in engine coolants for fear of forming hard water deposits. What is often
forgotten is that the carbonate found in make-up water will form scale so that one needs to
deal with the scaling issue regardless of whether phosphate, silicate, or molybdate are used in
the coolant additive package. Certainly there is the potential for aggravating scale formation
when phosphate is used with hard make-up water, but the technology exists to control this
situation. Scale inhibitors (water soluble polymers in the 1000 to 5000 molecular weight range)
have been developed especially to control hot surface deposition problems. It is a c o m m o n
practice to use phosphate along with a scale inhibitor in high pressure boilers [12]. Scale infor-
mation in this application is far more critical than an engine cooling system.
As part of the development of the phosphate-molybdate SCA, the additive package was
checked for its hot surface scaling tendency in hard water. Figure 14 shows results from a scal-
ing bench test rig. In this test a heater is inserted into a stainless steel tube, and coolant is cir-
culated over the surface of the hot tube. The wattage of the heater is chosen to give a heat flux
similar to the surface of a diesel engine liner. An automotive antifreeze coolant was tested and
then compared to the same coolant with the recommended dose of the phosphate-molybdate
FIG. 14--Hot surJace scaling bench test. Upper tube is a commercial antiJkeeze alone while the lower
tube is the same commercial antifreeze combined with the phosphate-molybdate SCA.
SCA. There is a significant amount of scale for the antifreeze alone, but the antifreeze plus
SCA gave almost no scale at all.
The ability of the phosphate-molybdate additive to control scale was further evaluated using
the NTC-400 test. Since the engine is over fueled and runs hot, it is a good tool to evaluate the
scaling protection provided by an SCA. The phosphate-molybdate additive was tested at the
recommended concentration in 50O/oby volume GM-6038 antifreeze. The make-up water was
Columbus, IN, tap water, which has a hardness of approximately 300 ppm. The GM-6038
FIG. 15--Dynamometer scaling test. Liner on left GM-6038 antiJ?eeze alone while the liner on the right
was run with GM-6038 combined with the phosphate-molybdate SCA.
when run by itself allowed appreciable scaling of the liners. However, when the SCA was added
to the antifreeze, very little scaling took place (Fig. 15). This shows that the phosphate-molyb-
date additive can not only prevent scale but make up for the short coming of the antifreeze.
Oil Fouling
It is common to find lubricating oil in the cooling system. We have taken random samples
from the field on several occasions and found about 10% of the units with some degree ofoil
contamination. Oil causes heat transfer problems throughout the cooling system as it breaks
down to a semi-solid sludge and coats liners, heads, and heat exchange tubes (Fig. 16). In addi-
tion an oil fouled surface is more likely to corrode. This is especially true for the cast iron block
where oil fouling is likely to trigger a corrosion problem. Figure 17 shows a coolant filter
plugged with iron hydroxide, which was the result of lube oil in the coolant.
FIG. 16--Coolant side of a cylinder linerfouled with oil sludge.

FIG. 17--Coolantfilterplugged with rust as a result of oil fouling of the cooling system.
A supplemental coolant additive can be formulated to minimize the problems caused when
oil seeps into the cooling system. Surfactants can be incorporated into the additive package
that lifts oil from the metal surface and emulsifies it in the coolant. Figure 18 shows results
from a laboratory test where the phosphate-molybdate SCA with its surfactant package is com-
pared to a conventional SCA. The test is run by placing 5 mL of used-heavily contained lube
oil in a 100-mL glass bottle. The bottle is then rotated so that the interior surfaces are coated
with a thin film of oil. Next, 100 mL of heated (180~ [82~ SCA solution is placed in the
bottle, and the solution is shaken vigorously.The solution is poured from the bottle, and the
amount ofoil film remaining is a measure of the SCA's detergency. As can be seen in Fig. 18,
the phosphate-molybdate SCA leaves behind far less oil and can be expected to improve heat
transfer and reduce corrosion problems in an oil fouled cooling system.
Low- Temperature Solubility

Phosphate is unusual in that its solubility in coolant can be drastically affected by temper-
ature. It is unique in this regard when compared to the other inorganic coolant additives. Fig-
FIG. 18--Detergency test o f SCAs. First bottle shows oil coating sides before testing. Middle bottle is
results obtainedfrom a borate-nitrite SCA while bottle to the right was obtained with the phosphate-molyb-
date SCA.
Solubility in grams/liter
80O
600
400
200
Na2HPO4 K2HPO4
FIG. 19--Sodium versus potassium phosphate solubility in water at Ooc.
TABLE 8--Silicate gelation versus sodium phosphate crystallization.
Sodium Phosphate
Parameters Silicate Gelation Crystalization
Antifreeze concentration Problem above 60% problem below 50%

High silicate antifreeze Aggravates the problem no effect
greatly
SCA concentration becomes a problem above 2x becomes a problem above 2x
Temperature > 160*F (71"C) < 10*F
Nature of solids once formed will not go sodium phosphate crystals melt
back in solution above 20 to 30*F
Frequency common 2 to 3% rare <0.1
ure 19 shows sodium and potassium phosphate solubility as a function o f temperature. The
solubility of sodium phosphate is very low at reduced temperatures, which can cause problems
in cold operating conditions. Sodium phosphate can crystallize, plug the radiator, and cause
overheating. It is interesting that as soon as the coolant warms slightly the sodium phosphate
crystals melt and there is no further overheating. Table 8 gives the conditions under which
phosphate can crystallize in the coolant and contrasts this problem with silicate gelation. The
phosphate crystallization problem is easily solved by using the potassium salt in place o f the
sodium salt because of its far greater low-temperature solubility (Fig. 19).
Summary
A phosphate-molybdate based [13] supplemental coolant additive has been developed and
patented as an improvement over the borate-nitrite SCAs that have been commonly used in
heavy duty diesel engine coolant since the late 1960s. The phosphate buffer gives better solder
and all around aluminum corrosion protection. The addition of molybdate gives better liner
pitting protection and allows the use of lower amounts of silicate and nitrite in the additive
package. This is important as the SCA is often overdosed or builds up to a high level over time.
Lower silicate and nitrite levels in the coolant result in the following:
9 Reduced solder bloom and corrosion

9 Reduced silicate gelation/better antifreeze compatibility
9 Fewer water p u m p leaks
Finally the phosphate-molybdate additive has been formulated to better tolerate poor qual-
ity make-up water, which is often the cause of aluminum corrosion problems.
References
[1] akes B. D. ``bservatins n Auminum Water Pump Cavitatin Tests Engine Cant Testing:
Second Symposium, ASTMSTP887, Roy E. Beal, Ed., American Society for Testing and Materials,
[2] Vukasovich, M. S. and Sullivan F. J., "Evolution of Molybdate as an Inhibitor in Automotive
Engine Coolants," Materials Performance, August 1983, pp. 25-33.
[3] Wiggle, R. R., Hospadaruk, V., and Styloglou, E. A., "The Effectiveness of Automotive Engine Cool-
ant Inhibitors for Aluminum," Materials Performance, June 198 l, pp. 13-18.
[4] Weber, T. R., Stranick, M. A., and Vukasovich, M. S., "Molybdate Corrosion Inhibition in Deaer-
ated and Low - Oxygen Waters," Corrosion, Sept. 1986, pp. 542-545.
[5] A1-Borno, A., Islam, M., and Khraishi, R., "Multicomponent Corrosion Inhibitor System for Recir-
culating Cooling Water Systems," Corrosion, Dec. 1989, pp. 970-975.
[6] Rwe L. C. ``Appicatin fnhibitrs in Autmbies and Their Envirnment Crrsin nhib-
itors, C. C. Nathan, Ed., National Association of Corrosion Engineers, p. 180.
[ 7] Hercamp, R. D. and Hudgens, R. D., "Cavitation Corrosion Bench Test for Engine Coolants," SAE
Technical Paper 881269, Sept. 1988.
[8] Wiggle, R. R., Hospadaruk, V., and Tibaude, F. M., "Corrosion of Cast Aluminum Alloys under
Heat Transfer Conditions," SAE Technical Paper 810038, Feb. 1981.
[9] Hercamp, R. D., "Comparison of Fleetguard DCA and DCA 4 for Prevention of Cavitation Cor-
rosion of Cylinder Liners," Cummins Engine Company Technical Report 0789-86009, Dec. 1986.
[10] Firestone, N. B. and Mullins, G. A., "DCA 4 Field Test Final Report," Cummins Engine Company
Technical Report 0789-87002, 1987.
[ 11 ] Hudgens, R. D. and Hercamp, R. D., "Filtration of Coolants for Heavy Duty Engines," SAE Tech-
nical Paper 881270, Sept. 1988.
[12] Manual on Water, ST,~ 442. American Society for Testing and Materials, Philadelphia, pp. 204-
206.
[13] Hercamp, R. D. and Hudgens, R. D., "Diesel Engine Cooling System Composition," U.S. Patent
No. 4,717,495, Jan. 1988.
DISCUSSION
Steven Woodward I (written discussion)--Ifthe 70% high silicate coolant passed tests with-
out supplemental coolant additive, why were they added? Can you comment on why the coot-
ant failed after the addition of supplemental coolant additive?
R. D. Hudgens (author's response)--Heavy duty diesel engines fail because of cylinder liner
cavitation corrosion when run using coolant without supplemental coolant additive. There-
fore, the water p u m p seal leakage test in Table 7 was run with a variety of coolant mixtures,
most of them containing SCA.
After the addition of borate nitrite supplemental coolant additive to the high silicate anti-
freeze coolant, the water p u m p seal leaked presumably because of either high total dissolved
solids or high silicate. The phosphate-molybdate SCA contains less silicate and less total dis-
solved solids than the borate-nitrite SCA, and the water p u m p seal did not leak using the phos-
phate-molybdate SCA. Also, the low silicate antifreeze coolant with the borate-nitrite SCA
added did not leak. The low silicate antifreeze coolant has lower silicate and lower total dis-
solved solids than the high silicate antifreeze coolant.
First Brands Corporation, Danbury, CT 06810.

R. Douglas Hudgens 1 and R. B. Bustamante 2
Toxicity and Disposal of Engine Coolants

REFERENCE: Hudgen, R. D. and Bustamante, R. B., "Toxicity and Disposal of Engine Cool-
ants," Engine Coolant Testing." Third Volume, ASTM STP 1192, R. E. Beal, Ed., American Soci-
ety for Testing and Materials, Philadelphia, 1993, pp. 149-164.
ABSTRACT: In recent years there has been considerable concern and controversy regarding
both the toxicity and disposal of engine coolants. The toxicity of engine coolant relative to
human or animal ingestion is largely a function of whether ethylene or propylene glycol is used
to provide freeze point depression. The additives used in engine coolants to prevent cooling sys-
tem deterioration have a measurable, but limited effect on toxicity. The general environmental
effects are largely governed by the type of glycol as well as the contaminantsthat may be present
in the coolant.
This paper presents a literature survey on the environmental,disposal, and toxicity issuessur-
rounding engine coolants. In addition, an Acute Lethal Dose (LDs0)study is presented compar-
ing the LDs0 of fully formulated PG antifreeze and supplemental coolant additives. A review of
government laws and regulations covering disposal and management of waste coolant is also
provided.
KEYWORDS: engine coolants, engine antifreeze, coolant toxicity, disposal, biodegradability,

laws and regulations, recycling, environmentalconcerns
In the late 1960s and early 1970s toxicity and environmental concerns resulted in the elim-
ination of the additives, chromate and arsenite, from engine coolants. Since that time for-
mulations have changed little. However, Federal "Right to Know" laws and our continuing
environmental problems have caused renewed concern about the health effects and disposal
problems associated with engine coolants.
In dealing with coolant toxicity and disposal issues it is helpful to break down an engine
coolant into its component parts:
1. Water (the primary heat removal fluid)--The water content of a coolant will be 40 to
70% depending on the severity of the winter climate. In some warm weather areas freez-
ing temperatures are not encountered and water with a corrosion inhibitor package is
used.
2. Freeze Point Depressant, Boil Point Elevation--In most cases ethylene glycol, and
recently but to a lesser extent, propylene glycol is used in a range of 30 to 60% by volume
to prevent freezing of the water during winter.
3. Additive Package--Containing several different chemicals that are initially added to the
glycol to form an antifreeze or concentrate and eventually blended with water to form
the coolant. These additives are designed to prevent corrosion, deposit formation, and
foaming and are in concentrations of 0.5 to 3% by weight of the final coolant.
1Senior staffchemist, Fleetguard, Inc., Cookeville, TN 38501,

2 Professor and chairman, Department of Civil Engineering, Tennessee Technological University,
Campus Box 5015, Cookeville, TN 38505.
149
4. Contaminants--Build up as the engine is used and result from the following: thermal or
oxidative breakdown of glycol; lube oil and fuel accumulation; metals from cooling sys-
tem corrosion; and decomposition or oxidation of coolant additives.
5. Supplemental Coolant Additives--Or SCAs are used in heavy duty service to prevent
cavitation erosion of cylinder liners and to replenish inhibitor chemicals depleted with
service. SCAs are not used or required in passenger cars that have a coolant life of 30 000
to 50 000 miles (48 279 to 80 465 km). Heavy duty service usually demands 200 000 to
300 000 miles (321 860 to 482 790 km). before coolant replacement and hence the need
to periodically replenish inhibitors.
This paper is divided into sections dealing with the toxicity, disposal, and general environ-
mental effects of engine coolants. Each section, in turn, is broken down further to discuss in
detail the effects of the various coolant components.
Toxicity
General
In 1987 the Annual Report of the American Association of Poison Control Centers'
National Data collection System [1] stated that there were over 1.1 million poisonings
reported by 63 poison control centers. These 63 centers serve about half the U.S. population.
About 92% of the reported poisonings occurred in the home and the majority were accidental
(89%). Children under six years of age were involved in 62% of the incidences, and ingestion
accounted for 77% of the poisoning exposures. This same report noted 2451 poisonings related
to glycols with 2372 exposures being accidental, and of those, 765 poisonings were related to
children under six years of age. Consequently, our discussion on toxicity focuses mainly on
the acute (short-term) oral health effects of coolants on mammals. In general, the greatest con-
cern is the short-term health effects of ingestion by a person or a pet. There can be long-term
(chronic) health problems associated with engine coolants, which are primarily related to
heavy metals contamination. Chronic health effects will be discussed briefly in the section on
disposal.
Although more recent data are available from the American Association of Poison Control
Center ( 1991), 1987 represents a year when only ethylene glycol was available. Propylene gly-
col (PG) was not used in the manufacture ofa PG antifreeze concentrate until 1988. The 1991
data from the Poison Control Center does not vary significantly from the 1987 report, and the
conclusion would be the same: ingestion is the greatest concern.
Glycols
Glycols make up 95% by weight of the antifreeze/coolant concentrate, and after blending
with water, about 40 to 60% by volume of the coolant used in the vehicle. Conventional anti-
freeze has for years been formulated with ethylene glycol (EG). EG has provided an efficient
and cost effective means of freeze protection for engine coolants. EG is used in a variety of
other applications including polyethylene terephthalate for use in polyester films, fibers, and
resins.
While EG has served effectively as a freeze point depressant for engine coolants, its one dis-
advantage is its toxicity to humans and other mammals if ingested. Because of this, there is an
increased demand for safer products. Propylene glycol (PG) is a safer alternative for engine
coolant antifreeze. This product has been available for years and is used as an ingredient in
foods, cosmetics, and medicinal products. During the EG shortage of 1988 through 1989,
HUDGENS AND BUSTAMANTE ON TOXICITY AND DISPOSAL 151
many antifreeze producers were using PG as an extender for EG in production of conventional

antifreeze. However, it was not until the last year or two that there was any significant use by
consumers of a PG formulated antifreeze, that is, one without EG.
EG has an acute oral toxicity of 4700 mg/kg (LDs0) in laboratory rats [2]; a rating of 5000
or lower can classify a material as hazardous. (Note the lower the n u m b e r the higher the tox-
icity.) Although marginally hazardous by this rating system, EG is a known toxin to humans
at relatively low levels (reported as low as 398 mg/kg) [2], and consequently is classified by
m a n y regulatory authorities as a dangerous material. When ingested, EG is metabolized to
glycolic and oxalic acids, which causes an acid base disturbance and results in kidney damage
[3]. EG has the added complication of a sweet smell and taste.
No material is absolutely safe. However, PG is a much safer material with an LDs0 in lab-
oratory rats of 28 000 to 30 000 mg/kg [3,4]. PG is metabolized to lactic and pyruvic acids,
normal body constituents [3,5,6].
As previously mentioned, corrosion inhibitors are blended with glycol and sold as formu-
lated antifreeze. These additives are discussed in more detail later. Although there is a signif-
icant body of data showing the lower toxicity of PG versus EG, what happens to that advantage
once the inhibitors are added?
One means accepted by industrial hygienists for estimating formulation toxicity is a calcu-
lation method [ 7]. The LDso value of the individual formulation component is divided into
the component's weight fraction. This "reciprocal" value is added to that of all the other com-
ponents. The sum of this calculation is then divided into 1. The result is an estimate of the
LDs0 of the formulation. The calculation method described above is used in Table 1 to com-
pare LDs0 values of the standard antifreeze formulation, GM-6038, using either EG or PG.
As can be seen ethylene glycol is the largest single component, and its LDs0 value largely
determines the estimated formulation LDs0. Note, we have ignored any contribution of the
pluronic or dye since they are in extremely low concentrations. Also the water present to sol-
ubilize the additives will tend to dilute the toxic effects of the other components and raise the
LDs0 level of the formulation. For this exercise we have chosen to ignore the effect of the small
a m o u n t of water.
This calculation method would therefore predict that not only is a PG antifreeze less toxic
TABLE I--LDso estimate of GM-6038 formulation.
LDsoa, Reciprocal,
Ingredient WT Fraction mg/kg 10 7
EG 0.9565 4700 2035

PG 0.9565 28000
NaNO3 0.002 3750 " "5 5
Na2B407 " 5H20 0.01 2660 38 38
NazSiO3 95H20 0.0015 1280 12 12
Na3PO4 912H20 0.0045 17000 3 3
NaMBT(50% SOLN) 0.0055 3120 18 18
NaOH 0.002 500 40 40
Pluronic L-61 0.0005 b b b
Green dye 0.00005 b b b
Water 0.0175 b b b
EG PG
Summation of Reciprocal 2150 X 10 -7 457 X 10 -7
LDs0 (l/reciprocal) 4650 21 900
See References 2 and 8.
b Omitted in calculation because of the small amount present.
LD 50 MG/KG IN RATS -THOUSANDS

30
25
20
15
10
0
PG ONLY COOLANT 1 COOLANT 2 COOLANT 3 EG ONLY
FIG. l--Oral toxicity of PB coolants.
than an EG antifreeze, but that PG antifreeze, like PG, should be classified as essentially non-
toxic. To confirm this, several commercial PG antifreezes were actually tested using laboratory
rats and the results are presented in Fig. 1. The results compare well with the estimated value
for the GM-6038 formula and confirm that PG can be formulated to provide good freeze and
corrosion protection while remaining essentially nontoxic.
Please note this calculation method is only a predictive tool to estimate the relative toxicity,
and it should be used only for estimating purposes.
Major Additives
The additive package in an antifreeze/coolant and in SCAs (used to replenish inhibitors)
will typically consist of from 5 to 15 different chemicals. These additives are broken down into
major and minor categories depending on the amount used in an engine coolant formulation
(Table 2).
The minor additives should not be of significance to the toxicity of engine coolants because
these materials are usually not highly toxic and are used in small quantities.
Table 3 lists the most common corrosion inhibitors and buffers used in engine coolants.
Lethal dose information is provided as well as the SAX toxicity rating. Arsenite, which has not
been used in engine coolants for 20 years, is shown to give a historical perspective. Common
materials, such as aspirin and table salt, are also listed for comparison purposes.
TABLE 2--Components of antifreezes and SCAs.
Major (0.05 to 3%) Minor (<0.05%)
Buffer defoamer
Corrosion dye
inhibitors
scale inhibitor
surfactant
chelates
TABLE 3--Common coolant additives, acute oral LDso data.
Other
Additive LDs0 in Rats, mg/kg Sax Rating Comments
Adipic acid 3600 1

Benzoate 4070 1
Borate 2660 2
Carbonate 4090 2
Ethylhexanoic Acid 3000[6] 1
Mercaptobenzothiazole 3120 2-3 sensitizer
Molybdate 2810[I3] 2
Nitrate 3750 2 oxidizer
Nitrite 85 3 strong
oxidizer
Phosphate 17000[ 7] 1
Sebasic acid 3400 .
Silicates 1280 i"
Benzotriazole 560114] 2
Tolytriazole 675[14] 2
Triethanolamine 8000 1
Arsenite 41 3 carcinogen
Aspirin 1500 ...
Table salt 3750 ...
NOTE: All LDs0 data taken from Ref 2.
Sax Acute Ingestion Rating System: (1) low-slight or mild toxicity; (2) moderate, and (3) high-severe
toxicity.
Nitrite is clearly the most toxic of the additives still used in engine coolant with an LDs0 for
rats of 85 mg/kg, which is in the range of arsenite. The triazoles are moderately toxic while
most of the other components have an LDs0 value that is the range of table salt and aspirin. It
is worthy to note that some significant variations were found in reported LDs0 data. For exam-
ple, MBT reported in R e f 2 is 3120 mg/kg. Values as low as 100 and as high as 5000 mg/kg
were found in other literature sources.
The toxicity of some coolant additives is effected by their alkalinity. The more alkaline
forms of silicate, phosphate, and borate have lower LDs0 values. Thus, the more alkaline meta-
borate (Na2B204 9 4H20) has an LDs0 value of 1700 compared to the less alkaline tetraborate
(NaEB407 9 5H20) with a value o f 2300 to 3300 mg/kg. This can also be shown for silicates.
Sodium silicate with an SiO2/NazO ratio of I has an LDs0 value of 600 mg/kg as compared to
1600 mg/kg for the less alkaline silicate with an SiO2/Na20 ratio of 2. The toxicity or more
appropriately the skin corrosiveness o f metasilicate (pH 13 at 5%) is greatly neutralized when
blended into a coolant with a pH in the range o f I0. The best example o f this is blending phos-
phoric acid with potassium hydroxide to yield K2HPO4 in an antifreeze coolant or liquid SCA.
The end product is a mildly alkaline salt which exhibits much less toxicity and corrosiveness
than the starting materials. This is shown graphically in Fig. 2.
The discussion to this point has focused on the toxicity o f antifreeze and engine coolants,
however SCAs are sold as both solid and liquid concentrates. Oral LDs0s were determined on
two c o m m o n liquid SCA packages. As can be seen in the Table 4, the calculated values for the
liquid concentrate (80% water and 20O/oadditives) were about half of the actual measured val-
ues. This probably reflects some positive interaction of the various additives as discussed
above, which will tend to lower the overall toxicity of the SCA. When you consider that the
recommended dose of these SCAs is 3% by volume, they should have minimal effect at their
use dosage on the overall toxicity of the coolant.
20
15
10
0~ ~ '
KOH H3PO4 K2HPO4
FIG. 2--Toxicity of reactants versus products (rat Oral LDso, mg/kg hundreds).
Solid SCAs are usually contained inside a coolant filter which makes access to ingestion
unlikely. Consequently, the solid SCAs were not tested.
The chemicals that make a coolant additive package have many c o m m o n uses. Some of
these are listed in Table 5. Adipate, benzoate, carbonate, nitrate, phosphate, and silicate are
used in foods. Even nitrite, which has the highest oral toxicity (lowest LDs0) of any of the com-
mon additives, is used as a food preservative and in medicines. Borate, benzotriazole, carbon-
ate, phosphate, silicate, and triethanolamine are used in soaps and detergents. As with all
chemical products, additive chemicals should be handled with care but in formulated engine
coolant present no extraordinary health risk.
Contaminants
Over the course of its life, coolant will become contaminated. Table 6 lists some of the con-
taminants and the concentrations that might be found in a severely degraded product. The
m a x i m u m level of each contaminant was obtained by reviewing the results of about 12 000
coolant analyses of both passenger cars and heavy duty vehicles, most of which were sent to
the Fleetguard " M o n i t o r " coolant analysis program [9].
In most instances, the level of contamination will be far lower than the listed values, but the
m a x i m u m values listed in Table 6 are useful in dealing with some coolant disposal and toxi-
cology issues. Table 7 presents the actual distribution o f copper and lead in 12 000 heavy duty
diesel engine coolants. As can be seen in comparing Tables 6 and 7, although lead can be found
TABLE 4--Calculated values for liquid concentrate.

Liquid Borate-Nitrite Phosphate-Molybdate
SCA LDso, mg]kg LDs0, mg/kg
Actual 4000 6200
Calculated 1798 2994
TABLE 5--Other common uses of coolant additives.
Additive Use
(1) Adipate Nylon, food additive, plasticizer, insecticide, adhesive, baking

powder
(2) Benzoate food preservative, antiseptic, medicine, tobacco additive,
mildew inhibitor, dye intermediate
(3) Borate glass, starch, and adhesives, detergents, herbicides,
pharmaceuticals, bleach, paint, insecticide, and antiseptic
(4) Carbonate soaps and detergents, glass manufacture, textile processing,
cleaning compounds, food additives, medicine,
photography
(5) Ethylhexanoate plasticizer, paint, and varnish
(6) Mercaptobenzothiazole rubber vulcanization, tire tread manufacturing, fungicide,
(MBT) E.P. additive in greases
(7) Molybdate medicine, paint pigment, fertilizer, and micronutrient for
plants and animals
(8) Nitrate fertilizer, medicine, black powder and dynamite, antiseptic,
pharmaceutical, food preservative, tobacco additive
(9) Nitrite dye, curing accelerator for rubber, pharmaceuticals,
preserving and curing meat, medicine, photographic
reagent, bleach
(10) Phosphate fertilizer, pharmaceuticals, baking powder, coffee creamer,
glazes for paper and ceramics, dietary supplement, and
animal laxative
( 11 ) Sebasate fibers, paint, plasticizers, candles, perfumes, Nylon,
polyurethanes
(12) Silicates catalyst, soap and detergents, adhesives, bleach, sizing of
textiles, egg preservative
(13) Tolytriazole/benzotriazole dishwashing detergents, tarnish removers, corrosion
inhibitors for dry cleaning equipment
(14) Triethanolamine soaps detergents, plasticizer, insecticide, rubber cure
accelerator
TABLE 6--Contaminants in engine coolant.
Maximum Concentration,
Contaminant mg/L Source
(1) Aluminum 50 corrosion
(2) Copper 100 corrosion
(3) Fuel 100 000 injection system failure
(4) Iron 500 corrosion
(5) Lead 200 corrosion
(6) Oil 100 000 oil cooler failure and sloppy maintenance
practices
(7) Organic Acids 5000 glycol oxidation
(8) Sulfate 1000 combustion gas leaking into the coolant
(9) Tin 50 corrosion
(10) Zinc 200 corrosion and hose/gasket additive
TABLE 7--Copper and lead distribution in H. D. coolants.

Concentration, mg/L Copper, % Lead, %
<5 92.6 81.9

5 to 25 6.2 12.4
26 to 50 0.7 2.7
51 to I00 0.4 1.8
>100 0.1 1.1
Average value 2.5 mg/lL 7.0 mg/L
as high as 200 mg/L, the majority of samples analyzed were below 5 mg/L. Consequently,
although Table 6 presents m a x i m u m contaminant levels, typical values are much lower.
The temperatures in engine cooling systems (93 to 99~ [200 to 210~ leads to a slow
decomposition (oxidation) of the glycol: either ethylene or propylene glycol. Almost every pos-
sible decomposition product that can occur probably does. Many of these compounds are
present in minute amounts, however, there are decomposition products that build up to large
enough amounts that examination is required [10]. Table 8 gives rat oral toxicity data for the
major glycol oxidation products [2,6]. The values range from 375 mg/kg for oxalic acid to over
3700 mg/kg for lactic acid. Degradation rates and the amount of each individual degradation
product can vary widely depending upon the type engine, service, and glycol used. Because of
TABLE 8--Toxicity of Coolant Contaminants [2,6].
Contaminant Rat Oral LD5o, mg]kg Starting Glycol

GLYCOLOXIDATIONPRODUCTS
(1) Glyoxal 1100 EG
(2) Oxalic Acid 375 EG
(3) Glycolic Acid 1950 EG
(4) Glyoxylic Acid N/A EG
(5) Formic Acid 1100 EG and PG
(6) Pyruvaldehyde 1170 PG
(7) Pyruvic Acid N/A PG
(8) Acetic Acid 3310 PG
(9) Lactic Acid 3730 PG
CORROSIONPRODUCTS
Metal
aluminum
chloride 3730
nitrate 4280
copper
chloride 140
sulfate 300
iron
chloride 900
sulfate 1480
lead
acetate 900 LDs0
tin
chloride 700
zinc
acetate 2510
sulfate 2950
its toxicity, there is particular concern about the buildup of oxalic acid in engine coolants. Our
analysis of used coolants shows that oxalic acid makes up less than 5% of the total amount of
organic acids present and therefore should not have a significant impact on used coolant
toxicity.
Toxicity information on the various corrosion products is also presented in Table 8. Beyond
the numbers provided, aluminum, iron, tin, and zinc are considered to have low acute toxicity
[11 ]. Even lead with its associated chronic health problems does not have a significant effect
on the acute toxicity of a used engine coolant. From an acute toxicity standpoint, copper pro-
vides the most reason for concern as CuCI2 has an LDs0 of 140 mg/kg.
Table 9 contains estimated LDs0 values of a GM-6038 type antifreeze (at 50/50 blend with
water) with and without the combination of 1000 p p m of copper chloride and 5000 ppm of
formic acid. This example assumes the most toxic form of corrosion and glycol oxidation
products 9 This coupled with the fact that the calculation method tends to give conservative
(low) LD~o values when dealing with engine coolants means that the effect of contaminants on
acute toxicity are, if anything, over estimated.
As can be seen in the table, the contaminants only marginally lower the LDs0 of an EG based
coolant. The effect on the PG based coolant is more pronounced; however, the LDs0 value is
still high and provides no reason for concern. One can conclude that the toxicity of a used-
contaminated coolant is still primarily a function of whether ethylene glycol or propylene gly-
col is used to provide freeze-point depression.
Disposal
Laws and Regulations
At the federal level, the management of hazardous waste is regulated through the Resource
Conservation and Recovery Act (RCRA) enacted in 1976 and subsequently modified a num-
ber of times. RCRA defines hazardous waste as "a solid waste, or combination of solid wastes,
which because o f its quantity, concentration, or physical, or infectious characteristics may (1)
cause or significantly contribute to, an increase in serious irreversible or incapacitating revers-
ible illness or (2) pose a substantial present or potential hazard to human health or environ-
ment when improperly treated, stored, transported or disposed of or otherwise managed." It
is important that hazardous waste must first be classified as a solid waste in order for the mate-
rial to be regulated as a hazardous waste. This does not mean that a solid waste must be solid,
but only that it meets the legal definition of solid waste. RCRA does define solid waste as " . . .
other discarded material, including solid, liquid, semi-solid, or contained gaseous material
9 Hazardous waste, therefore, although defined as solid waste, may actually be a liquid or
a slurry.
The first question to be asked by a solid waste generator is if the material that is being gen-
TABLE 9--Effect of contaminants on coolant toxicity.

Ethylene Glycol Propylene Glycol
Coolant LDs0, mg/kg LDs0, mg]kg
GM-6038 at 50% 9 300* mg/kg 43 700

GM-6038 at 50% with 8 400 29 000
1000 ppm CuC12
5000 ppm Formic acid
NOTE: Values calculated with reciprocal method as in Table 1
[ 7]. Water present was assumed to have no effect on toxicity.
erated is "hazardous" or not. Management of hazardous waste is a far more delicate and
expensive operation than that of a nonhazardous solid waste. The U.S. Environmental Pro-
tection Agency has defined hazardous waste as one that (1) exhibits characteristics of ignita-
bility, reactivity, corrosivity, or toxicity (T. C. Rule), (2) is a listed waste (nonspecific source
waste, specific source waste, specific commercial chemical), (3) is a mixture containing a listed
hazardous waste, or (4) is not excluded by law or regulation. The terminology of characteristic
or listed hazardous waste refers to criteria 1 and 2 above when they are used to classify a dis-
carded material. Details of characteristic waste criteria and hazardous waste lists can be found
in the Code of Federal Regulations (40 C R F 261).
RCRA regulates all aspects of the management of hazardous waste, from generation to
proper disposal. This is what is referred to as "cradle to grave" management. Generators are
required to identify the material deemed to be hazardous waste and to manage that waste as
licensed treatment, storage, or disposal facilities in full accordance with regulatory require-
ments. The law provides for civil and criminal penalties to be imposed upon violators. There-
fore a generator of a hazardous waste who gives, sells, or pays someone to "dispose" of that
waste is responsible to insure that it is properly disposed of in accordance with the law.
Regulation of hazardous waste involves a significant participation by individual states.
RCRA provides for the delegation of all or part of the hazardous waste regulatory program to
a state provided that such a state has a program that is equal to or more stringent than that
prescribed by the EPA. States may have, and sometimes do have, rules and regulations that
are more demanding than those published by the EPA. It would be proper for generators and
managers of hazardous waste or both to consider the federal regulations as a baseline criteria
and to determine if individual states in which they operate have requirements above and
beyond those imposed by the EPA.
In determining if a waste antifreeze is hazardous or not, it must be assessed if the material
is included in one of the hazardous waste lists published by EPA or by the appropriate state.
A second determination must be made to see if the material meets any of the four criteria that
would make it a characteristic hazardous waste. Of particular interest would be the presence
in the material of heavy metals that might be found as contaminants in the waste. Limits for
particular metals, such as lead, are established by the EPA and the individual states (for exam-
ple, lead, 5 ppm). It is of vital importance that all tests, chemical or otherwise, be performed
on waste materials. Tests performed solely on the fresh, unused antifreeze product would not
by any means be sufficient. The materials regulated by RCRA are discarded wastes not
products.
Should recycle of waste coolant be a goal, it is possible that the material being collected for
that purpose would still be regulated under the hazardous waste rules. It would not be wise to
assume a material not to be a hazardous waste simply because it is to be subject to recycle or
reuse. In all likelihood, part or all management rules indicated for hazardous waste will be
active for the management of waste to be recycled, and this might involve operations such as
manifesting, storage, and treatment.
Title III, Section 301, of the Clean Air Act Amendment of 1990, also has application to
engine coolants. The a m e n d m e n t establishes a list of 172 compounds and 17 compound cat-
egories as hazardous air pollutants; this amends Section 112 of Title I of the Clean Air Act
(1970).
The Comprehensive Environmental Response, Compensation and Liability Act
(CERCLA) sometimes referred to as Superfund, defines a hazardous substance to include any
hazardous air pollutant listed under Section 112 of the Clean Air Act [12]. CERCLA (Section
102) requires the EPA to establish reportable quantities (RQ) for hazardous substances. Unless
superseded by regulations establishing a new RQ, the National Response Center must be noti-
fied if there is a release of one pound or more of a hazardous pollutant. Therefore all newly
identified hazardous air pollutants with no prior RQ should be treated as if they have an RQ
of 1 lb (454 g) and any nonpermitted releases in excess of 1 lb (454 g) should be reported imme-
diately to the National Response Center. EG is on the CERCLA list and currently by default
has an RQ of 1 lb (454 g). PG is not on the list. Failure to notify may result in fines or
imprisonment.
The short review of regulatory issues given in this presentation is an indication of the com-
plexity of the problems associated with the proper management of hazardous waste. Regula-
tions are voluminous, complicated, and confusing. It is sometimes not easy to determine what
the true situation is and what to do to properly satisfy regulatory requirements. In some areas,
local regulations (city or state) supercede Federal requirements. Good knowledge of the reg-
ulations and an extensive technical background on the material generated may optimize the
proper management, treatment, and final disposal of the waste material. Later in this paper
we discuss those components/contaminants in used coolant which can make it a hazardous
waste.
Biodegradability
Glycols--PG and EG are both readily biodegradable. A paper entitled "Biological Treat-
ment of Deicing/Antifreeze Chemicals" [4] was presented at the 1991 National Research and
Development Conference on the Control of Hazardous materials. The paper concluded that
P G and EG are both biodegradable while diethylene glycol (DEG), which is frequently blended
into EG antifreeze to reduce costs, biodegrades slower but will ultimately be consumed. The
oxygen requirements for biodegradation of these glycols can be readily supplied by conven-
tional aeration equipment that is available in most waste treatment facilities. This is confirmed
by data from the glycol biotreatment study that is presented in Table 10. Typical values expe-
rienced in publicly owned waste treatment works (POTW) are also shown and confirm that
spent glycols can be readily biotreated in existing facilities.
Acclimation of the microbial population to glycols is required to rapidly biodegrade any of
the three glycols. Biodegradation of EG, PG, or DEG by unacclimated cultures will occur but
TABLE l O--Glycol biodegradation test conditions and results. (Acclimated biomass and unbuffered
pH).
Typical Values
for Public Waste
Glycol EG DEG PG Treatment Plant
Initial glycol 2000 2100 2400 ...

Concentration, mg/L
Average temperature, *C 19.8 19.3 19.3
Detention time, h 27 27 24 6"to" 12
Initial COD, mg/L 3000 3100 2900 400 to 600
Effluent COD, mg/L 230 2300 140 50 to 100
Initial MLVSS
Concentration, mg/L 2260 2700 2860 2000 to 3000
Specific oxygen uptake 22.5 10.3 16.3 15 to 30
Rate, mg O2/h/1000 mg
MLVSS
Sludge yield, mg TSS Produced/ 0.21 0.25 0.33 0.4 to 0.6
mg COD Removed
NOTES: COD is chemical oxygen demand. MLVSS is mixed liquor volatile suspended solids. TSS is
total suspended solids.
at slower reaction rates. As with other chemical and biochemical reactions, the rate of glycol
biodegradation decreases as temperature decreases. However, estimated effluent chemical
oxygen d e m a n d (COD) concentrations for winter conditions (10 and 5~ indicate that under
typical organic loadings, glycols can be biodegraded to meet effluent discharge limitations
within reasonable residence times.
The biodegradation of glycols results in a lowering of p H apparently from the production
o f acids. Data for P G showed less pH depression than for EG or D E G [4]. Buffered systems in
the range of p H 7 to 8 improved glycol biodegradability considerably. As a result biological
treatment systems for glycol containing wastes waters should have provisions for pH adjust-
ment and control.
Major Additives and Contaminants--EG and PG readily biodegrade but what about actual
engine coolant that contains several different additives. To answer this question, four different
types of antifreeze/coolants were tested for their biodegradability. These coolants were chosen
because they contained a range of various additives used as buffers and corrosion inhibitors.
Growth medium was added to a 1% solution of the various antifreeze/coolants. The solutions
were inoculated with a seed culture from a local P O T W that had been acclimated to EG. They
were maintained at 200C with constant shaking, and the onset o f strong culture growth was
noted as a rapid substantial increase in turbidity. The results of this simple test is presented in
Table 11.
A 1% solution of coolant is a much higher concentration than a POTW is ever likely to see,
yet the antifreeze coolants were readily biodegradable. These results are not unexpected as the
major coolant additives are used in a range of household products that are sewered in large
quantities. Further work showed that a 5 to 10% solution of ethylene glycol was lethal to the
culture. The commercial antifreezes we evaluated also proved lethal in the same dose range.
This suggests that the biodegradability o f glycol and commercial antifreeze coolants are similar
and that sewering of used coolants can be acceptable means of disposal.
Minimizing Coolant Disposal Problems

The best way to manage a waste is to prevent or limit its generation. Strategies have emerged
in the past few years to extend the useful life o f a coolant and recycle waste product.
Extending Coolant L / f e h T h e r e are currently three approaches of which the authors are
aware where the useful life of a coolant is extended well beyond the 30 000 to 50 000 miles
(48 279 to 80 465 kin) recommended by most manufacturers.
9 Long Life C o o l a n t - - S o m e antifreeze coolant manufacturers are increasing the expected

life of their products simply by adding more additives or choosing additives or both that
are more stable and therefore deplete at a slower rate.
9 Extending Coolant Life (Light D u t y ) - - T h i s is done by filtering the coolant to remove
debris and replacing additives that have depleted or been lost through leaks or spills. Sev-
TABLE 11--Biodegradability of antifreeze coolants.
Type of Coolant Days to Acclimation and Growth

Benzoate-nitrite 1 to 2
Sebasate 2 to 3
Phosphate-molybdate 2 to 3
Borate-nitrite 5 to 7
HUDGENS AND BUSTAMANTE ON TOXICITY AND DISPOSAL 1 61
eral companies manufacture portable systems that pull the coolant from the vehicle, filter
it, and return the coolant after a booster inhibitor package has been added.
9 Extending Coolant Life (Heavy Duty)--For years the acceptable coolant life for many
heavy duty vehicles has been 2 years, 240 000 miles (386 232 km) or 6000 h. This has
been achieved by using coolant filters on the vehicles as well as adding supplemental cool-
ant additive (SCA) at each oil drain ( 12 000 to 18 000 miles [ 19 312 to 28 967 km]). Many
fleets, in spite of Original Equipment Manufactures (OEM) recommendations are push-
ing out their coolant life to 500 000 miles (804 650 km) or until in frame overhaul. This
can be achieved with limited risk provided close attention is paid to the amount of SCA
added to the cooling system. When too little SCA is added to the system, liner pitting and
corrosion will result. Water pump leakage, cooling system deposits, and over heating are
caused when too much additive builds up in the coolant. Another caution is that excessive
coolant loss, which is the rule rather than the exception in many heavy duty truck fleets,
can make the management of SCA levels very ditficult.
Recycling--A coolant will finally get to the point where filtration and further SCA addition
will not render it fit for further use. The level of dissolved solids and soluble contaminants will
build to the point that they cause poor heat transfer, water p u m p leakage, and corrosion. Recy-
cling, where the old additives and contaminants are removed, is required at this point; other-
wise the coolant should be properly disposed. Distillation and deionization are the only pro-
cesses that currently meet some Heavy Duty OEM requirements for a truly recycled product.
Note, other light and heavy duty OEMs still have not approved the use of recycled coolant.
Companies are now available, especially in the Western United States, which pick up, process,
and return engine coolant. In addition, there are several different portable systems on the mar-
ket that make it possible for a fleet operation or mining site to recycle their own coolant. ASTM
is currently developing a specification for recycled antifreeze/coolant.
Environmental Concerns and Alternatives

Worldwide nearly 400 million gal (1514 million litres) of antifreeze are sold every year. It is
estimated that 25 to 50% of this volume ends up improperly in the environment. Dumping by
consumers is a major cause of this pollution. This can only be corrected by increased consumer
awareness through education.
Another major source is leakage, spills, and overflows in the heavy duty industry. Experi-
ence with heavy duty vehicles shows that it is c o m m o n to lose 10% of the coolant volume every
12 000 to 18 000 miles (19 312 to 28 967 km). This equates to a gallon/month (4 L/month)
for the typical on-highway truck, or a leakage rate of one drop per rain. A coolant leak that
small is likely to go unnoticed but still results in a significant loss. For example, many heavy
duty fleets never change coolant but purchase enough antifreeze every year to replace the cool-
ant in each of their vehicles.
In some heavy duty operations, overflows account for far more coolant loss than leaks at
the water pump, hose clamps, or radiator core. If a heavy duty radiator without a overflow
tank is topped off, it can lose a quart or more of coolant when the engine heats up (expanding
the coolant) to operating temperature.
Small spills and leaks (less than a gallon) of coolant, while undesirable, present little impact
to the environment. Glycols, the main ingredient other than water, may present a toxic danger
to wildlife, but will fully biodegrade. Additives and contaminants are present in small quan-
tities and consequently do not represent a significant hazard. Cleanup, where possible, is
always recommended to prevent animal poisonings.
Larger spills or on-purpose dumping are of more concern. Although glycols will readily bio-
degrade even at concentrations of 50 000 mg/L, high concentrations of contaminants, such as
copper, could negatively affect soil biological activity. Lead, another corrosion product, can
contaminate soil, leach into water tables, move up the food chain, and have serious effects on
higher life forms. While glycols will biodegrade in soil or water, they place an oxygen demand
on lakes, rivers, and streams. Depleted oxygen levels can suffocate aquatic life. For all the
above reasons, generators of spent coolant, both private and commercial, should not dump
the waste product.
We have discussed at some length the problems that can occur when engine coolant gets
into the environment. However, what are the issues when a business or an individual wishes
to responsibly dispose of waste coolant? First, spent coolant should be considered a hazardous
waste unless proven otherwise. Lead content is the primary reason for the hazardous desig-
nation, but some states are also concerned with the EG content as well. In addition, POTWs
limit the oil and grease content of a waste in the range of 100 to 300 rag/L. Experience with
heavy duty vehicles shows that about 10% have some oil contamination of the cooling system.
Therefore, oil levels of waste coolant also need to be considered before it is discharged to a
sewage system. In some areas the P O T W will simply charge a fee if the oil and grease exceeds
their limit. Table 12 summarizes the disposal concerns of waste engine coolant.
For the individual who occasionally must dispose of a few gallons of waste coolant, the
R C R A / E P A guidelines do not apply. The guideline 40 CFR 261.4 simply states that house-
hold waste will not be considered hazardous even though technically it may be. However for
everyone else, the corner service station to the largest fleet or mine, RCRA does apply. Table
13 breaks down the RCRA requirements based on the amount of waste that is produced each
month.
If coolant wastes are collected and properly treated, the resulting environmental concerns
can be minimal. Bulk collection, transport, and disposal in accordance with laws and regula-
tions is the recommended method of waste coolant management.
Recycling is always desirable but not always practical. Several competent and effective
"antifreeze" recycling operations have begun in recent years, but they are primarily focused
on serving commercial generators, which account for only 10 to 20% of spent coolant. Some
communities have started recycling centers for hazardous wastes, but collection, segregation,
and transport to recyclers is difficult, inconvenient, and costly.
Generators should select a recycler cautiously. The primary liability"rests with the generator
to insure proper disposal including residual metals and additives.
What action should be taken to dispose of spent coolant? Large-scale commercial generators
can recycle or biotreat the waste. Note, A S T M has not yet approved specifications for recycled
antifreeze, but going back to the "pure" glycol and then reinhibiting should yield as good a
product as that originally put in the vehicle. Commercial generators should check with their
local waste treatment authority; permitting may be required, which could involve some type
ofpretreatment. This may be as simple as checking the p H or checking the lead content o f the
TABLE 12--What makes coolant a hazardous waste.
Area of Concern Comment
Lead limit 5PPM

Other heavy metals varies
Oil and grease limit 100 to 300 mg/L of concern
Ethylene glycol in some states
TABLE 13--RCRA requirements by size of waste generator:

Waste Produced, Volume of Waste Content,
Size kg/month (L/month) (approx.) Requirements
Large > 1000 1000 Collect, Manage and Report
Medium 100 to 1000 100 to 1000 ""
Small < 100 < 100 Must Manage
No Reporting
waste coolant. Also treatment plants may ask generators to slowly feed spent glycol so as to
avoid shocking the microorganism population. Under no circumstances should any amount
of spent antifreeze be dumped on the ground or into a body of water.
Consumers, whose cooling systems are typically 2 to 5 gal (7.6 to 18.9 L), should follow
similar guidelines. If recycling is not available, speak with your local officials about setting up
a program. If local regulations allow, pour spent coolant down your drain as opposed to dump-
ing it in the back yard. Municipal treatment facilities will easily handle small quantities of
spent coolant. Caution, do not pour large quantities (above 2 gal [7.6 L]) of spent coolant into
a septic system. Too much coolant too quickly can shock the "bugs" and necessitate reseeding
the system.
As a final note, good housekeeping is always recommended. Clean up leaks and spills, and
properly dispose of used coolant.
Summary
The acute oral toxicity of spent antifreeze is largely determined by the glycol used. Additives
and contaminants will have a minimal effect on coolant toxicity.
Spent coolant should not be dumped on the ground or into bodies of water but properly
disposed of by biological treatment or recycling. No spill is insignificant, but small quantities
are not a major threat to the environment. Regardless of size, spills or leaks can pose an acute
oral toxicity danger to wildlife and pets. Consequently all spills and leaks should be cleaned
up.
Heavy metals, such as lead, present a chronic health concern, and along with ethylene glycol
may be one of the primary reasons why a used coolant fails under RCRA guidelines. Improper
dumping can result in heavy metal contamination of soil and water and present a real danger
to the food chain.
Spent coolant can be safely biologically degraded in waste treatment plants. Testing can con-
firm that a waste is not in the hazardous classification. Commercial generators should check
with waste treatment officials for permitting requirements, and the rules and regulations gov-
erning disposal in their communities. Homeowners and other small quantity generators
should also check their POTW and governing bodies before discharging.
Commercial generators can recycle spent coolant. However, the ASTM has not finalized
specifications for recycled antifreeze. Until specifications are available, generators/users would
be wise to only use antifreeze/coolant products that fully meet the performance requirements
of "fresh" antifreeze. Recycling for the average consumer is not readily available. We would
encourage interested parties to work with their local government to set up recycling programs.
Disposal problems can be minimized by extending coolant life. Extended life products and
maintenance methods are under development by many of the antifreeze producers. Consum-
ers, both private and commercial, should support the development and use of such products.
For instance, filtration and SCA addition have long been used to extent the life of heavy duty
engine coolants. Now this concept is beginning to be applied to extend the life of passenger car
coolants.
The authors urge heat exchanger manufacturer's to continue the trend toward air-side sol-
dering. This will minimize the lead (from the solder) in contact with the coolant, which will in
turn reduce the lead levels in engine coolants. Alternative types of solder, which are available,
should be evaluated in an attempt to eliminate a source of lead contamination and the related
serious health concerns.
Heavy duty OEM's should consider installing overflow systems on their vehicles to simplify
maintenance, limit overfilling of the radiator, significantly reduce coolant loss, and thus ben-
efit the environment.
References
[1] Litovitz, Schmitz, Matgunas, and Martin, "1987 Annual Report ofthe Americas Association of Poi-
son Control Centers National Data Collection System," American Journal of Emergency Medicine,
Vol. 6, No. 5, Sept. 1988, p. 479.
[2] Dangerous Properties oflndustrialMaterials, 7th ed., Sax and Lewis, p. 1625.
[3] Casarett and Doull's, Toxicology The Basic Science of Poisons, 4th ed., Amdur, Doull, Klaassen,
pp. 703,704, 705.
[4] Martino and Andrews, Biological Treatment of Deicing~Antifreeze Chemicals, National Research
and Development Conference on the Control of Hazardous Materials, 1991, p. 3.
[5] Ruddick, J. A., "Toxicology, Metabolism and Biochemistry of 1,2 Propanediol," Toxicology and
Applied and Pharmacology. 1972.
[6] Rowe, V. K. and Wolf, M. A., "Glycols," Patty's Industrial Hygiene and Toxicology, 3rd ed, Vol.
2C, p. 3861.
[7] 1990-1991 Threshold Limit Vahws for Chemical Substance and Physical Agents and Biological
Exposure Indices Americas Conference of Government and Industrial Hygenists, pp. 42-45.
[8] Fisher Scientific MSD Sheet.
[9] ASTM Committee D- 15 on Engine Coolants Minutes of Committee D- 15 October 1990 San Diego,
CA., pp. 159-253.
[10] Internal ARCO Chemical Analysis of Thermal Decomposition Products of Glycols.
[11] Handbk fTxic and Hazardus Chemicas and Carcingens 2nd ed Sittig pp. 6 -62 256-258
522, 523, 542-546, 862-865.
[12] DowUSA, Letter to Customers Concerning Regulatory Status ofEthylene Glycol, 26April 1991.
[13] Vukasovich and Van Riper, "Molybdate: A Low Toxicity Corrosion Inhibitor," Proceedings of 7th
European Symposium on Corrosion Inhibitors.
[14] MSD Sheets from PMC Specialties Company.
DISCUSSION
James A. Lima j (written discussion)--Doug, given your data, it seems to me that filtration
to remove most of the lead, followed by POTW treatment should be a viable means of disposal.
Do you agree?
R. D. Hudgens and R. B. Bustamante (written discussion)--In my experience, the lead in
the coolant is largely soluble. I attribute this to the glycol oxidation products that act as a lead
chelate and hold the metal in solution. If filtration is effective at lead removal, then it will have
to be very fine filtration, below a micron.

Paul R. H e m m e s , 1 T h o m a s H. Kreiser, ~ Sarah Valle, ' a n d
Richard D. H e r c a m p 1
Test Strips for Rapid On-Site Analysis of

Engine Coolants
REFERENCE: Hemmes, P, R., Kreiser, T. H., VaUe, S., and Hercamp, R. D., "Test Strips for
Rapid On-Site Analysis of Engine Coolants," Engine Coolant Testing.. Third Volume, ASTM
STP 1192, Roy E. Beal, Ed., American Society for Testing and Materials, Philadelphia, 1993,
pp. 165-179.
ABSTRACT: This paper reviews the use of test strips for measuring the composition of engine
coolants. The tests described include test strips for freezepoint, cylinder liningcavitation corro-
sion protection, reserve alkalinity, pH, chloride, and mercaptobenzothiazole.
Freezepoint strips work with reasonable accuracy with both ethylene glycoland propylene gly-
col based coolants. They will also work with mixtures of the two. The strips are shown to work
with coolants containing phosphates, borates, or benzoates. Data show that good results can be
obtained with new users so that training is minimized.
Strips for liner cavitation protection are described. Good results were obtained with visually
read strips for nitrite in the range of 200 to 4000 ppm and for molybdate in the range of 75 to
1500 ppm. A hand-held reflectance meter to read these strips is described. Precision and accuracy
data are given. The strips are not sensitive to glycol concentration.
Data are presented for the accuracy of test strips for reserve alkalinity and pH. Test strips for
mercaptobenzothiazole are described. A description is also provided of a prototype test for chlo-
ride levels in used coolants where the makeup water may have contained salts. In all cases, test
strips were found to be accurate, easy-to-use, inexpensive devices useful for routine field moni-
toring of coolants.
KEYWORDS: test strips, freezepoint, reserve alkalinity, cylinder lining cavitation protection,
pH of coolant, mercaptobenzothiazole, chloride
Test strips for monitoring the field condition of engine coolants consist of small chemically
treated pads attached to a plastic handle. When immersed in coolant, the pads develop color.
The strip is removed from the sample after a specified time. After another specified time (rang-
ing from 15 s to 1 rain), the pad color is compared to a color chart, which has been calibrated
to read the level of the particular constituent in the coolant.
The history of test strips began with simple laboratory test papers such as litmus paper for
checking the acidity of aqueous solutions. However, the use of test strips for semi-quantitative
measurements was first developed in the area of medical diagnostics. In the middle 1950s, test
strips were introduced for monitoring glucose and protein levels in urine. These tests were
widely used for screening patients to detect the presence of diabetes and urinary tract problems
[1]. In the mid-1960s, blood glucose strips were introduced to allow patients to control dia-
betes. Today, tens of millions of people, worldwide, use blood glucose strips to help control
diabetes. Urine test strips, some with as many as ten analyses on a single strip, are widely used
in doctor's offices and hospitals to screen for possible health problems.
Vice-president of Operations, senior scientist, assistant research scientist, and consultant, respectively,
EnvironmentalTest Systems, Inc., P.O. Box 4659, Elkhart, IN 46514.
165
The first use of test strips for testing engine coolants was by the military [2] and they were
adapted for coolant maintenance in heavy-duty vehicles.
Test strips have several unique advantages over other methods of field analysis of engine
coolants. These include:
I. Testing using strips is more convenient than laboratory testing because the results are
obtained within minutes at the site where the vehicle is located. The vehicle is still in the
maintenance area so that appropriate maintenance can be accomplished in a single visit.
2. The cost for each test performed by test strips is less than the cost for most laboratory
tests.
3. For fleet maintenance, the used strip can be stapled to the vehicle maintenance record to
provide a positive record that the testing has been performed.
4. The initial cost for freezepoint testing is much lower than for a refractometer, and break-
age is not a problem.
5. Cross contamination from vehicle to vehicle caused by poor cleanup procedures is elim-
inated because each strip is used only once.
Poor coolant maintenance is widespread and leads to major cooling system failures [3]. In
spite of this, field analysis of coolants has almost exclusively been restricted to determination
of the freezing point of an ethylene glycol (EG) based coolant by means of a hydrometer. In
recent years, however, many changes have occurred that make this approach unacceptable.
Some of these changes are:
1. Commercial use of propylene glycol (PG) based coolants raises the possibility of finding
vehicles with EG, PG, or mixtures in any given vehicle. Without knowledge of the type
of coolant in a given vehicle, measuring the freezepoint protection is greatly complicated.
2. With extended use of coolants due to supply shortages and problems of disposal of used
coolants, there is a greater chance of coolant problems developing. This is especially true
o f corrosion and cavitation protection additives, which deplete with time.
3. Extended warranties of vehicle manufacturers make the manufacturer liable for repairs
required because of poor coolant maintenance.
4. Changes in the mid 1980s in the inhibitor chemistries used in heavy-duty vehicle cool-
ants, such as the addition of molybdate to nitrite for liner corrosion protection in the
mid-1980s.
This paper discusses test strips for analysis of the following coolant properties:
1. Freezepoint,
2. Cylinder lining cavitation protection,
3. Reserve alkalinity (RA),
4. pH,
5. Chloride, and
6. Mercaptobenzothiazole (MBT).
Test Strip Overview

Before discussing the individual tests, it is useful to consider some general features of test
strips and the factors that influence visual readings. Strips are read by means of optical detec-
tion of color changes, which occur in response to exposure to the sample containing a partic-
HEMMES ET AL. ON TEST STRIPS FOR ANALYSIS OF ENGINE COOLANTS 167
ular component. To describe color, we must consider three things: the detector, the light
source, and the sample.
For visually read strips, the detector is the eye. The human eye responds differently to light
of various wavelengths. In general, the eye is most sensitive to light at about 570 nm wave-
length and rapidly decreases as wavelength either increases or decreases from this maximum.
The response of the eye is, however, quite individual. It is probably true that no two people see
colors exactly the same way. Nevertheless, we are so adapted to color that these differences
cause far fewer problems than might be expected. The only real exceptions come with the con-
dition of color blindness. In many cases, even mild color blindness does not prevent the user
from using test strips. This is due to the fact that a color comparison is actually done against
the color chart. Color-blind people can often match two colors that to them may appear as
gray shades.
The light source is of more concern. It is c o m m o n for one observer to find that two colors
that match under one light source do not match under another. This is called metamerism.
The three most c o m m o n types o f light sources are sunlight, incandescent lighting, and fluo-
rescent lighting. In theory, colors appear somewhat different under each source. Usually, how-
ever, these effects are small enough that they cause few problems. We use special inks for print-
ing our color charts, which minimizes metamerism. In general, we use fluorescent lighting as
the standard light source for coolant strips. It is important to remember the possible presence
o f metamerism if testing of one sample is done under several light sources.
Finally, the physical properties of the sample affect the visual appearance. For example, at
certain viewing angles, a wet strip may appear to have a glossy appearance. This is due to spec-
ular reflectance from the surface. In general, a rough surface provides diffuse reflectance, which
depends less on the viewing angle. Again, these problems are generally of little importance,
especially for experienced users.
During development of a test strip, we use special instrumentation that measures color in
terms of three color coordinates. This gives us a means to describe color and color changes
quantitatively. Many of the various factors that influence the color of the strip can be con-
trolled by this instrument.
Some people have difficulty in judging color matches especially when the levels of material
being tested lie between the levels shown on the color chart. This subjectivity of visual readings
of strip tests for nitrites and molybdates can be avoided by the use of a hand-held reflectance
meter for these strips, which was recently introduced and will be discussed in the appropriate
section o f this paper.
Freezepoint Strips
The use of hydrometers for measuring the freezepoint of coolants has been widespread due
to their relatively low cost compared to refractometers. The fact that the refractometer is more
accurate than the hydrometer [4] has led to the adoption of the refractometer for larger fleets
where the initial cost is less of a factor. The refractometer is strongly recommended over the
hydrometer for heavy-duty fleet work. As PG-based coolants become more prevalent, the
hydrometer becomes inappropriate due to the large differences between the densities of EG
and PG. It requires prior knowledge of the composition of the coolant to use hydrometers at
all. When PG or EG or mixtures of the two can be found in the vehicles being tested, only
refractometers with both a EG and PG scale can be used and then with diminished accuracy
[51.
Test strips have been in commercial use for measuring freezepoints of engine coolants since
1976. The chemistry of this test is detailed in the Appendix. The strip has been patented [6].
In addition to the general advantages test strips have over other methods, the freezepoint test
can be used without prior knowledge of the coolant composition. That is, the test strip is
equally effective with EG or PG or mixtures o f the two. Table 1 shows experimental refrac-
tometer readings on either the EG scale or the PG scale. The column labeled calculated
freezepoint is a listing of the freezepoint estimated from the known concentrations. The data
for pure PG samples show a tendency for overestimating the freezepoint protection at low
levels of PG and underestimating it at high concentrations. In mixtures containing significant
levels of EG, the test strips are accurate. In very high glycol solutions, there is a small tendency
for underestimation of protection. Planned modifications of the color chart will improve the
performance at these levels.
Table 2 shows the effect of additives on freezepoint depression as measured by strips, refrac-
tometers, and by ASTM Test Method for the Freezing Point of Aqueous Engine Coolant Solu-
tion (D 1 177). This study was done to investigate the effect of various salt additives, which are
expected to lower the freezepoint. One of the salts chosen was sodium benzoate. This was cho-
sen because the benzoate ion is expected to have an abnormal effect on the refractive index of
solutions due to its electronic structure. This was a possible problem for the refractometer
readings. In fact, both the refractometer and the strips read these solutions accurately.
TABLE l--Refractorneter and strip readings on EG and PG mixtures.
Volume
Percent Refractometer, OF Strip,
Total Calculated Visual
EG PG Glycol EG Scale PG Scale Freezepoinff Results
0 10 10 24 26 26 +10
0 20 20 13 19 19 +5
0 30 30 -3 10.5 10.5 -5
0 40 40 -24 -6 -6 -20
0 50 50 -42 -27 -27 -30
0 60 60 -68 -65 -65 -45
10 0 10 25 27 25 +15
10 10 20 13 19 16 +5
10 20 30 --2 11 6.7 -5
10 30 40 - 17 0 -4.3 -20
10 40 50 -41 -25 -28.2 -30
10 50 60 -65 -52 --54.2 -45
20 0 20 14 20 14 +5
20 10 30 0 12 4 -5
20 20 40 -18 -1 -9.5 -20
20 30 50 -42 -26 -32.4 -30
20 40 60 -63 -59 -60.3 -45
30 0 30 2 13 2 0
30 10 40 - 16 1 -11.8 -20
30 20 50 -36 - 18 -28.8 -30
30 30 60 -60 -53 -56.5 -45
40 0 40 -14 2 - 14 - 10
40 10 50 --36 - 18 --32.4 --30
40 20 60 --70 -75 --71.7 -60
50 0 50 --35 --18 --35 -30
50 10 60 --57 --49 -55.7 -45
60 0 60 -59 -52 -59 -45
a The proportion of EG or PG in the mixture determined the calculated freezepoint from the EG or
PG scales on the refractometer.
T A B L E 2--Freezepoints in 50% glycol (EG or PG).
ASTM Method
Glycol Solute D 1177-88 Refractometer Strip
EG None -- 36.7 - 35 - 30
EG 2% Sodium benzoate - 42.9 - 44 - 45
EG 1% NaH2PO4 - 39.8 - 39 - 45
EG 2% Na2HPO4 -- 41.8 - 42 -- 45
EG 1% Sodium borate -- 38.7 - 38 -- 30
EG 2% Sodium borate - 39.4 - 41 - 45
EG 1% Na2HPO4/1% S o d i u m borate - 38.6 - 41 - 45
EG 1% Na2HPO4/0.14M F o r m a t e -- 35.4 - 38 -- 30
PG None - 28.1 -- 30 - 30
PG 2% Sodium borate - 33.4 - 37 - 40
T a b l e 3 c o n t a i n s t h e r e s u l t s o f a s t u d y d e s i g n e d to t e s t t h e a b i l i t y o f n e w o p e r a t o r s t o u s e
b o t h r e f r a c t o m e t e r s a n d t e s t strips. T w o d i f f e r e n t r e f r a c t o m e t e r s w e r e c o m p a r e d u s i n g a n
experienced operator. Then an inexperienced operator used one of the refractometers. Finally,
an experienced lab person read strips along with an inexperienced strip reader selected from
t h e office staff. T h e r e w e r e s o m e s u r p r i s i n g l y l a r g e v a r i a t i o n s b e t w e e n r e f r a c t o m e t e r s as s h o w n ,
f o r e x a m p l e , b y t h e r e s u l t s o f s a m p l e s 2 3 3 0 6 a n d 2 3 3 1 0 . E x p e r i e n c e level w a s n o t s i g n i f i c a n t
in r e a d i n g t h e s a m e r e f r a c t o m e t e r . E x p e r i e n c e level also w a s n o t a s i g n i f i c a n t f a c t o r i n r e a d i n g
strips.
T A B L E 3--Results o f study to test new operators.
Refractometer Readings, EG Strip

Scale, *Fa Readings, ~ b
Sample
Number # l-Lab #2-Lab #2-Otfice Lab Office
23292 -42 -46 -48 -45 -45

23293 -15 -24 -24 -30 -30
23294 -23 -28 -29 -20 -30
23295 -37 -47 -48 -60 -60
23296 -34 -33 -33 -30 -45
23297 10 7 6 0 -5
23298 -54 -54 -54 -45 -70
23301 -37 -38 -38 -30 -45
23302 -37 -37 -37 -30 -30
23303 -68 -60 -59 -70 -70
23304 -54 -51 -51 -60 -45
23305 13 7 7 0 5
23306 -19 -30 -30 -30 -20
23307 -6 -13 -12 -20 -20
23308 17 15 14 10 15
23309 18 15 13 10 15
23310 -23 -31 -31 -30 -30
23311 --1 --5 --6 --5 --10
o T w o refractometers were c o m p a r e d by one experienced lab

operator. O n e refractometer was used by an inexperienced office
worker.
b Strips were read by both a lab a n d office person, t*C = (t*F -
32)/1.8.
The strips are designed to be read at room temperature (65 to 85~ [18 to 29~ Reading
hot solutions causes errors because the indicator's degree of ionization is temperature-depen-
dent. The recommended procedure is to withdraw a small sample into a cup. This should
quickly cool to the proper range. A strip can then be dipped and read.
In summary, the use of test strips is a fast, easy, accurate method for monitoring the freeze
protection of cooling systems. The strips work well in EG, PG, and mixtures of the two.
Cylinder Liner Cavitation Protection

The two coolant additives used in most of the world to protect cast iron cylinder liners from
cavitation corrosion are nitrite and molybdate salts. Although chromate was the predominant
additive in the 1950s and 1960s, it has been replaced because of its toxicity. Nitrites were
almost exclusively used to replace chromates until the mid 1980s when molybdate also began
to be used in supplemental coolant additives (SCAs). The so-called mixed system test kit is
shown in Fig. 1. It is designed to measure both nitrite and molybdate in coolants [7]. It has
enjoyed extensive commercial use since 1988. The chemistry of both strips is discussed in
Appendix 1. The test procedure consists of drawing a sample of coolant, adding an equal vol-
ume o f acidic solution, and then dipping two strips, one after the other. The color chart for
this test is a two-dimensional array (Fig. 2). After dipping in the acidified sample, the molyb-
date strip (called strip A) is read on the vertical brownish orange blocks. The nitrite strip (strip
B) is read on the horizontal, pink blocks. A vertical line from the position where the nitrite
strip best matches the color block will intersect a horizontal line from where the molybdate
strip best matches the color blocks. The point o f intersection of the two lines will lie in a zone
corresponding to a specified treatment procedure for the coolant.
Although the visually read test strips have been well-received for the past several years, some
people are not comfortable with color matching. For this reason, we have developed a small,
portable, reflectance meter to read the test strips (Fig. 3). In operation, the test strips are devel-
oped in much the same way as in the visual test. The molybdate strip is first inserted into the
FIG. l--"Mixed system'" test kit.

FIG. 2--Color chart for the mixed system test kit.
FIG. 3--Hand-held reflectance meter for nitrite and molybdate strips.

meter. A light-emitting diode of fixed wavelength momentarily focuses a bright spot of light
on the reacted pad. Some o f this light is reflected from the pad to a detector. This detector
converts the light intensity into an electrical signal, which is processed by the microprocessor
in the meter. The nitrite strip is read in the same manner. The microprocessor uses the cal-
culated levels of nitrite and molybdate to calculate a recommended treatment code for the
operator. The operator does not have to read strips or compute treatment codes. The actual
concentrations can be obtained from the instrument using a special procedure. This allows for
troubleshooting of the system and also provides a means for generating accurate and precise
data on the coolant composition, if this is required.
One important difference between the strips for nitrite and molybdate as read on the meter
is the different instrument responses to concentration. The reflected light intensity is a linear
function o f molybdate concentration and a nonlinear function o f nitrite concentration. This
FIG. 4--(a) Instrument response to molybdate concentrations in coolants. The y-axis represents instru-
ment units that are proportional to reflected light intensity, The response at both 40% and 60% glycol is
shown. (b) Instrument response to nitrite concentrations in coolants at both 40% and 60% glycol Upper
line is 40% antifreeze and lower line is 60% antifreeze,
TABLE 4--1on chromatography and strip results.
Ion Chromatography, Strip Results,

Sample Number ppm ppm
1 42 7
2 190 60
3 230 95
4 710 700
5 1190 990
6 1000 860
7 1060 990
8 1850 1790
9 1040 1000
10 1050 1170
11 1320 1480
12 2890 2370
13 1740 1990
14 2920 3290
15a 1500 2790
Sample 15 contained a great deal of suspended carbon. When
this sample was applied to the strip, the strip was darkened by the
carbon and gave a high result as a consequence.
difference is shown in Fig. 4a and b. This in no way affects the operation of the system in the
hands of the user. We have deliberately designed the system so that low samples (200 ppm or
less) would read even lower than actual. This ensures that the recommended treatment for
such critically low samples calls for supplemental additives. The design philosophy behind this
system has consistently been to give the maintenance person doing the test the proper treat-
ment recommendations under all circumstances. The system does this effectively. The low
bias is shown in the section on accuracy.
The accuracy and precision of the meter-read test strips were checked with field samples.
Nitrite was determined in a number of field samples using ion chromatography [8] and com-
pared with the meter-read test strip results. The results are given in Table 4.
Table 5 contains results of the molybdate strip test compared with an inductively coupled
plasma (ICP) reference method. To investigate the reproducibility of the test strip results, ten
replicates were run on a single meter at four levels o f molybdate and five nitrite levels. Table
6 gives the results of this study in terms of the true concentration, the average of the ten rep-
licates, the standard deviation of the replicates, and the percent coefficient o f variation (%CV).
The latter is 100 times the standard deviation divided by the average.
The influence of the glycol level was investigated. Figure 4a and b also shows the meter
response to molybdate and nitrite concentrations at 40% and 60% EG. In general, the system
is remarkably free from interferences.
We are developing an improvement of the system. We have been able to develop a molyb-
date and nitrite test strip that does not need the acid dilution step. Both visual and instrument
versions of these strips are under development with a planned introduction near the end of
1991.
Reserve Alkalinity Test Strip

A dip-and-read reserve alkalinity (RA) test has been used by the military for nearly 20 years
[9]. In 1984, a two-pad multiple strip consisting o f a freezepoint test and an RA test was intro-
TABLE 5--Molybdate strip results compared with ICP reference

method.
Sample Number Lab Value (ICP) Strip Value
1 1567 1427
2 979 878
3 1447 1137
4 594 557
5 404 487
6 408 481
7 600 916
8 964 824
9 237 305
10 563 519
11 132 198
12 483 473
13 117 145
14 20 23
15 1069 1076
16 3 0
duced commercially under the name Radi-Aider~ (Radi-Aider produced and sold by Envi-
ronmental Test Systems, Inc., Elkhart, IN 46514. The chemistry and mechanism of the RA
strip are described in the Appendix. In an evaluation of the RA test, a major manufacturer of
coolants obtained the test results shown in Table 7. The strip values are compared with those
measured using ASTM Test Method for Reserve Alkalinity of Engine Coolants and Antirusts
(D 1121).
These results indicate that the RA test gives good results compared with ASTM Test Method
D 1121.
As noted in ASTM Test Method D 1121-88, care is required in the interpretation of any RA
measurement. When the chemical nature of the coolant additives changes, the RA changes.
European coolants using organic acid salts always appear to have a low RA, even when fresh.
When the composition of the coolant is well-defined, as, for example, in military service or
large fleets, a precipitous drop in RA can serve to indicate a problem with the coolant. RA is
a measure of buffer capacity. Thus, a low RA value can exist along with a proper pH. However,
TABLE 6 - - T e n repliates run on a single meter at 4 levels o f molybdate and nitrite.
Parameters Level 1 Level 2 Level 3 Level 4
Molybdate
True concentration, ppm 75 200 600 1500
Average, ppm 74.3 201 631 1460
Standard deviation, ppm 8.9 11.2 46.3 78.3
% CV 12 5.6 7.3 5.4
Nitrite
True concentration, ppm 200 700 1500 2500 4000
Average, ppm 178 698 1439 2479 4481
Standard deviation, ppm 22 68 191 247 203
% CV 12.5 9.8 13.3 10 5
TABLE 7--Results of evaluation of reserve alkafinity test.
% Ethylene RA (ASTM Method

Glycol D 1121-88) RA (Strip)
New Antifreeze Solutions

33 5.1 6
50 7.0 8
80 11.2 10+
Used Antifreeze Solutions
33 9 10
50 7.1 8
47 6.9 6+
47 10.5 10+
42 5.7 6-
37 4.9 6-
42 5.8 6
41 5.2 6
34 3.7 6-
the low RA value implies that the useful life of the coolant is short because the basic materials
added to the coolant are nearly exhausted.
Test Strips for Coolant pH

We have developed a test strip for coolant pH that gives visual pH readings that correlate
with those measured by ASTM Test Method for pH of Engine Antifreezes, Antirusts, and
Coolants (D 1287). The chemistry is simply the use of selected acid-base indicators incorpo-
rated into a paper matrix. The pH of a coolant may or may not be a significant indicator of
condition. If the pH is below 7, indicating an acid condition, the coolant should be replaced
at once. On the other hand, a pH of 8 may not be worse than a more alkaline coolant due to
the type of additives used. The performance of these test strips is shown in Table 8.
TABLE 8--Results of a test strip for coolant pH.
% Ethylene pH (ASTM Method

Glycol D 1287-85) pH (Strip)
New A n tffreeze Solutions

33 10.2 9+
50 10.5 9
80 10.8 9+
*50 7.0 7
*50 6.0 6
Used Antifreeze Solutions
33 7.8 8.5
50 7.3 8
47 7.9 8
47 8.2 8.5
42 8.7 9-
37 9.0 9
42 9.7 9
41 8.3 9
34 8.5 8.5
The pH strip does not respond properly in pure glycol. This may account for the low reading
in 80% solutions. The test is more accurate in used samples. This may be related to the high
RA of such solutions. The two samples of new glycol (marked with an asterisk in Table 8) were
acidified in the lab to the pH values determined by pH meter.
Test Strip for Chloride in Coolants

Chloride is a possible serious contaminant of coolant systems. When present in high levels,
it causes corrosion of iron (Fe) and a l u m i n u m (Al) parts. There is no industry-wide agreement
on what constitutes too high a level of chloride, but opinion ranges from 100 to 200 ppm as
chloride. We have developed a prototype of a test strip for chloride in coolants as part of a
program to produce a screening test for reconditioning coolants. Clearly, a heavily contami-
nated coolant is not worth reconditioning. The chloride test we developed is designed to detect
a threshold level of chloride. That is, if the level of chloride is 175 ppm or more, the strip turns
blue. Below this level, it remains colorless. The chemistry is given in the Appendix. In addition,
we have designed the test to eliminate interferences by mercaptobenzothiazole (MBT) up to
about 700 ppm of MBT. ASTM testing requires a special procedure for the laboratory deter-
mination of chloride in the presence of MBT, Test Method for Trace Chloride Ion in Engine
Coolants (D 3634).
In addition, we plan to use our existing test strip for MBT in conjunction with this chloride
test so that fresh coolants with high MBT levels are not confused with heavily salt-contami-
nated, used coolants. The chloride test strips are being patented. They have been evaluated by
a major coolant manufacturer (Table 9). The samples were prepared by dissolving known
amounts (if it was known, it had to have been weighed) of salt into a glycol-water mixture. A
yes means that the strip was blue, positive for chloride. A no means that the strip was negative
for chloride.
Test Strip for MBT

MBT has been used as an additive in coolants to provide some protection for copper (Cu)
components in the cooling system. The level of MBT in a coolant decreases rather quickly due
to adsorption onto the metal parts and chemical degradation. Because of this rapid decay,
MBT has been replaced with longer lasting additives such as tolyltriazole (TTZ). Nevertheless,
MBT is still found in a n u m b e r of antifreeze formulations (particularly for light-duty use) and
TABLE 9--Results of evaluation by coolant manufacturer.
Samples
Readers A B C D E F G H I
Cl, ppm a 150 160 170 180 190 200 220 250b 300
1 No No Yes Yes Yes Yes Yes No Yes
2 No No Yes No Yes Yes Yes Yes Yes
3 nac na na na na na Yes Yes Yes
a The above results indicate that the chloride test strips serve as an adequate screen for high levels of
chloride in coolants. A Yes means that the strip was blue, positive for chloride. A No means that the
strip was negative for chloride.
b Although one reader found a 250-ppm sample to be negative, this abnormal finding could not be
confirmed by an additional reader.
c The symbol na means that a third reading was not made on this sample.
in some supplemental additives. A strip test for MBT was developed and sold some years ago
but was withdrawn from the market as the use of MBT declined. It is being reintroduced only
in conjunction with the chloride test for the reasons stated in that section. The chemistry of
this test strip is given in the Appendix.
Summary
Test strips are easy-to-use, accurate, versatile, and cost-effective devices for the rapid testing
o f engine coolants. They have been used successfully for more than 10 years to measure
freezepoint and are expected to gain increased use as P G use increases. The use of strips for
monitoring the levels of cylinder liner cavitation protecting agents is growing due to the imme-
diate treatment recommendations they provide. Newer tests for testing the suitability of cool-
ants for reconditioning are being developed. The use of test strips in monitoring coolant sys-
tems, both heavy- and light-duty uses, is expected to grow.
APPENDIX
The chemistries used in various test strips are as follows:
Freezepoint
The freezepoint test is based on the fact that certain acid-base indicators change color when
placed in a mixed solvent containing water and an organic liquid. This effect occurs even if
the solution contains material that in water would buffer the solution to constant pH. The
reaction can be written:
HIn + H20 ~ H30 q- It,l
In this reaction, HIn is the acid form of the indicator and In is the ionic form.
For the indicator used in the freezepoint test, HIn is yellow, whereas I n - is blue. The mix-
ture gives shades of yellow greens.
The effect may arise from several causes. Because the quantity of the deprotonated species,
In, increases with increasing organic solvent content, the presence of the organic liquid must
stabilize this form. This must be a type of specific interaction because electrostatics predict that
the ionic forms will be less stable in water-organic mixtures. This specific effect does not differ
much among EG, PG, or simple alcohols such as methanol. The lack of strong dependence
on the nature of the organic substance is the reason this chemistry works to indicate the
freezepoint of aqueous mixtures of either EG or PG.
Reserve Alkalinity
This test consists o f a strip treated with a buffer compound and an acid-base indicator. In
the presence of alkaline materials, some of the buffer acids are converted to the corresponding
salt. This changes the ratio of buffer acid to salt and thereby changes the equilibrium pH. The
indicator provides a visual response to this p H change.
Cavitation Corrosion Protection Strips

The two types of strips now in use involve tests for nitrite and molybdate. The nitrite test
strips use the following chemistry:
nitrite + sulfonamide - - diazo compound
diazo compound + aromatic amine ~ azo compound
A number ofsulfonamides can be used, including sulfanilamide. The formation of the diazo
c o m p o u n d occurs only in acidic solution. The coupling of this material to an aromatic amine
in acid solution produces a highly colored azo compound. This reaction is specific for nitrite.
The molybdate test is based on the formation of a complex between molybdate and a com-
plexing agent. This complex is then reduced chemically to an intensely colored compound.
This test is highly specific for molybdate.
Test Strip for Chloride

This test consists of a pad treated with silver (Ag) salts and a proprietary indicator material.
When chloride is added, it reacts with Ag first. Once the Ag is consumed, any excess chloride
reacts with the indicator to give a blue color. Because MBT reacts with Ag, it is an interfering
substance. We use a material in the pad that preferentially reacts with MBT. Then the MBT
is not free to react with Ag. This is effective with modest levels of MBT, up to about 700 ppm.
Higher levels still interfere because the material we use to minimize interference is not 100%
effective. At high levels of MBT, the small fraction of MBT that reacts with Ag is enough to
develop some color.
MBT Strip
For the above reason, we supplement the chloride test strip with one specific for MBT. This
provides a warning that high levels of MBT are present and that the chloride test may be giving
a false reading but also is useful in checking that reconditioning has taken place, because many
treatments contain some MBT. This test is based on the fact that MBT can be made to develop
a mixed complex in the presence of certain metal ions and amino acids. The orange color of
the complex of the metal, MBT, and amino acid is quite specific. The test measures levels of
MBT up to about 1200 ppm.
References
[1] Free, A. H., Adams, E. C., Kersher, M. L., and Free, H. L., "Simple Specific Test for Glucose in
Urine," ClinicalChemistrv, Vol. 3, 1957, p. 163.
[2] Conley, J. H. and Jamison, R. G., "Research and Development Efforts in Military Antifreeze For-
mulations," Engine Coolant Testing: State of the Art, AST~I STP 705, W. H. Ailor, Ed., American
Car Population Cooling Systems," SAE 831821, Society of Automotive Engineers, Warrendale, PA,
pp. 121-130.
[4] Hudgens, R. D. and Bugelski, W. G., "Analysisof Coolants from Diesel Engines," SP-811 Worldwide
Trends in Engine Coolants, Cooling System Materials and Testing, SAE 900434, Society of Auto-
motive Engineers, Warrendale, PA, 1990, pp. 79-99.
[5] Hercamp, R. D., Hudgens, R. D., and Coughenour, G. E., "Aqueous Propylene Glycol Coolants for
Heavy Duty Engines," SP-811 Worldwide Trends in Engine Coolants, Cooling System Material and
Testing, SAE 900434, Society of Automotive Engineers, Warrendale, PA, 1990, pp. 47-77.
[6] Van der Weft, L. J. and Rapkin, M. C., U.S. Patent 3,973,909, August 10, 1976.
[ 7] "Coolant Test Kit," Heetguard Bulletin No. 3300396, 1988.
[8] Hudgens, R. D. and Hercamp, R. D., "Test Methods for the Development of Supplemental Additives
for Heavy Duty Diesel Engine Coolants," Engine Coolant Testing: Second Symposium, ASTMSTP
887, Roy E. Beal, Ed., American Society for Testing and Materials, Philadelphia, 1986, pp. 189-215.
[9] U.S. Army Technical Bulletin 750-651, 1971.
STP1192-EB/May 1993
DISCUSSION ON TEST STRIPS FOR ANALYSIS OF ENGINE COOLANTS 179
DISCUSSION
Joseph A. Lima' (writtendiscussion)--A very informative paper was presented in Chemtech

recently that describes the complex technology used to produce test strips. My concern is that
many men are color-blind and would have difficulty comparing colors and making the inter-
pretations necessary to read these strips. I believe it to be important to mention because indis-
criminate use of the strips by people who are not aware of such limitations could present an
obvious problem. I would be interested in hearing your comments about this particular prob-
lem and also would like to know the cost of the color comparison device that'you mention,
which I gather is meant to circumvent this color-blindness problem.
Paul R. Hemmes (author's response)--Color-blind persons have been found to be quite
capable of using strips except in a few extreme cases or with a particular color change on a
strip. What is actually being done when a strip is used is to compare the strip with a color chart.
Thus, one is making a color comparison. Color-blindness is less of a problem when matching
two colors held side by side.
In the medical diagnostic field, millions of diabetics worldwide use test strips to measure
blood glucose levels. Many continue to use visual reading with satisfactory results. The use of
a meter is not dictated by the problem o f color blindness. Some people with perfect vision are
uncomfortable making judgments of the match between the strip and the color chart. This is
particularly true when the measured level is between two color blocks. Such people prefer the
objective meter readout to their own color match.
The only meter for engine coolant test strips is available from Fleetguard, Inc. They should
be contacted for retail prices.
t Houghton Chemical Corporation, AUston, MA.
Copyright9 1993 by ASTM International www.astm.org

W a l t e r Z a m e c h e k I a n d D a l e A. M c K e n z i e 2
Application of Inductively Coupled Plasma

(ICP) Emission Spectroscopy and Laser
Ablation-ICP for Problem Solving in Coolant
Systems
REFERENCE: Zamechek, W. and McKenzie, D. A., "Application of Inductively Coupled
Plasma (ICP) Emission Spectroscopy and Laser Ablation-ICP for Problem Solving in Coolant
Systems," Engine Coolant Testing: Third Volume, ASTM STP 1192, Roy E. Beat, Ed., Ameri-
can Society for Testing and Materials, Philadelphia, 1993, pp. 180-189.
ABSTRACT: Since the introduction of commercial instrumentationin 1975, inductively cou-

pled plasma (ICP) emission spectroscopy has become the technique of choice for the determi-
nation of metals for many environmental,clinical, geological, and industrial applications. This
is due to the method's relative freedom from interferences, speed, and flexibility.Described are
some main characteristics of ICP spectroscopy, typical sample preparation approaches, and
results for the analysis of coolants and deposits from different parts of cooling systems. Laser
ablation-lCP (LA-ICP) was used to map the spacial distribution of metals across the surface of
water pump seals. Design of the ablation system and standardization method are described.
These results are compared to the analysis of the same samples by scanningelectron microscopy/
energy dispersive X-ray (SEM/EDX), electron spectroscopy for chemical analysis (ESCA), and
scanningauger microprobe/energy dispersive x-ray (SAM/EDX).
KEYWORDS: inductively coupled plasma (ICP) spectroscopy, coolants, radiator deposits, laser
ablation-ICP
Since the introduction of commercial instrumentation in 1975, inductively coupled plasma

(ICP) emission spectroscopy has become the method of choice for the determination of metals
in a wide variety of materials with applications covering most industries. The Central Scientific
Laboratory (CSL) of U n i o n Carbide Chemicals and Plastics Company, Inc., has been using
ICP spectroscopy since 1976 as a tool for the routine analysis and problem solving related to
cooling systems. Described are some of the methods that have been developed for the bulk
analysis of coolants and deposits from many parts of the systems.
As a part of a problem-solving effort, laser ablation-ICP (LA-ICP) was used for spacial map-
ping of elements on the surface of water pump seal faces. This method proved to be useful and
provided information used to complement data from other surface analysis methods.
Group leader, Union Carbide Chemicals and Plastics Company, Inc., Central Scientific Laboratory,
777 Old Saw Mill River Rd., Tarrytown, NY 10591.
a Manager, OEM & Technology Services, Corporate Fellow, First Brands Corp., Still River Corporate
Park, 55 Federal Rd., Danbury, CT 06810.
180
ZAMECHEK AND McKENZIE ON ICP EMISSION SPECTROSCOPY 181
Method
Detailed descriptions of emission spectroscopy and ICP emission spectroscopy, its main
attributes, analytical considerations and spectrometer designs, have been published elsewhere
[1,2]. However, a few major aspects of the method are worth mentioning here.
Emission spectroscopy is a method based on the detection and quantification of energy,
emitted as photons, by the outer electrons in excited atoms and ions. Emissions from the
excited species occur at the same time, thus, emission spectroscopy is inherently a simulta-
neous analytical method. In ICP emission spectroscopy, an argon (Ar) gas plasma is used as
an excitation source. The high temperatures and electron densities of the plasma allow for the
determination of all metals and some nonmetals, such as phosphorus (P), sulfur (S), and car-
bon (C). The polychromatic emissions from the source are dispersed with a grating and
detected with photomultipliers. Spectrometers are available as sequential, simultaneous, or
combination instruments.
Quantification is achieved by comparison with standards. In aqueous solutions, quantita-
tive detection limits (ten times the standard deviation of the background), assuming no inter-
ferences are present, are in the range of 0.02 to 0.1 ug/mL, with a linear dynamic range of four
to five orders of magnitude.
Spectral overlap is the major source of interferences in emission spectroscopy. It is caused
by the inability of the spectrometers to resolve completely the large number of emission lines
present in typical spectra. The extent of interferences depends on instrument resolution, line
selection, and the elemental composition of the sample. Typical spectrometer software pack-
ages allow for empirical corrections by comparison with standards. This approach is satisfac-
tory in most cases, however, the accuracy and detection limits for the interfered element are
degraded. In cases where the interferences are severe or high accuracy is desired, a secondary
emission line for the element of interest must be used. In some cases, a chemical separation
must be performed before analysis.
Samples are generally introduced into the ICP source as a fine liquid spray. Solid materials,
where bulk composition is of interest, must be dissolved before analysis. In special cases, sam-
ples have been introduced as slurries, powders, or gases [3]. In LA-ICP, a laser is used to ablate/
vaporize small amounts of sample, which are then swept as a gas into an ICP source and,
except for desolvation, undergo the same excitation processes as aqueous samples. Detailed
descriptions of LA-ICP and analysis of selected materials have been published elsewhere
[3,4].
In our laboratory, LA-ICP capability has been developed for the determination of trace
(ppm) metal contaminants in refractory ceramics and electronic materials. This is necessary
because conventional sample dissolution methods of these hard, refractory materials are prone
to contamination from the grinding and the robust decomposition conditions. In addition, it
is often necessary to analyze special "features" or determine elemental distributions on the
surface of these materials, and this is not possible with bulk dissolution methods.
Experiment
All solution analyses were performed using a Jarrell-Ash Model 1100 polychromater also
equipped with a 0.5-m Ebert monochromater, both operated at standard conditions (Table 1).
Table 1 lists the laser operating conditions. All spectral interference corrections and data
reduction were performed with in-house software, which has been described previously [5],
however, it has undergone several revisions; details are available on request.
Developing LA-ICP required solutions to three problem areas: design of the sample cell,
TABLE 1--Operating system.
Operating System Parameters

Polychromater Jarrell-Ash Model 1100
grating 2400 Grooves/mm
entrance slit 25 um
viewing height 16 mm above coil
Monochromater Jarrell-Ash 0.5-m Ebert
grating 1180 Grooves/mm
entrance slit 25 um
exit slit 25 #m
viewing height 16 mm above coil
Forward power 1000 W
Reflected power < 10 W
Torch sample uptake tube Polyethylene tipped with sapphire
nebulizer Jarrell-Ash Cross-flow
sample gas flow 1 L/min
coolant gas 18 L/rain
Laser Control Laser Model 512Q
wavelength 1064 nm
maximum (cw) power 50 W
Q switched mode 10-20-Hz rate
alignment of the target area with the laser, and calibration. The sample introduction system
for conventional and LA-ICP analysis is shown in Fig. 1. The gas inlet provides tangential flow
of Ar. The three-way valve allows for conventional introduction of solutions or gas from the
sample cell. For target alignment, the sample cell can be moved vertically and the sample
spoon can be moved horizontally or rotated.
A target locator was designed for the alignment o f the spot to be sampled (target) with the
laser. Two small telescopes, one with a pinhole and one with a slit image, were attached to the
main laser tube at such angles that the images are coincident when the entire assembly is at
the proper focal distance from the target. The laser focal spot is at the center of the point where
the slit and pinhole images overlap. Alignment by eye with a resolution of approximately 0.1
m m is possible.
In conventional ICP, calibration is achieved by comparing the sample solution emission
intensity to that of the standard solution. The same is true for LA-ICP if homogeneous solid
standards are available, as in the case of glasses and alloys. Such standards are not available
for ceramic materials. Semiquantitative analysis can be achieved by comparing ceramic sam-
ples with standard glasses or alloys. Alternatively, approximate emission intensity per micro-
gram of an element can be calculated from solution analysis and compared to LA-ICP. In
many cases, calibration is not vital because a simple empirical comparison of a "good" and
"bad" sample or feature can also provide valuable information.
Results and Discussion of Solution Analysis

Analysis of Coolants
The sensitivity o f the ICP technique allows direct determination of coolant additives and
solubilized metallic corrosion products. The spectrometer is calibrated with aqueous stan-
dards, then the sample is diluted 50-fold and analyzed. Typical results are shown in Table 2.
Laboratory testing of experimental engine coolant formulations is performed in a con-
trolled, closed-loop system, with specimens (usually aluminum or brass) of interest suspended
SPECTROMETER
I _ PLASMA
ICP TORCH -"!~I~11 ~----~LASER HEAD
HI I.,~ "~J'U~ SAMPLE
CHAMBEB- - ' - - - - ~ ~ --f ~OLDER

NESUL,ZER_._~I I i t [ VALVE .GO.
ARGON
ABLATED SAMPLE SUPPLY
GASES INTO
THE NEBULIZER
FOR LA-ICP
FIG. l --Conventional and LA-ICP sample introduction system.
in the coolant. Performance evaluation is based on elemental analysis of used coolants and
deposits. A mass balance is also obtained, and the surface elemental composition of specimens
is determined by successive rinsing with nitric acid and potassium hydroxide. Typical results
for the analysis of used coolants reflect changes in elemental concentrations as a function of
time in use; these are shown in Table 3. The changes in the concentrations of silicon (Si), alu-
m i n u m (AI), and calcium (Ca) in this coolant, as a function of testing time, are readily appar-
ent. Water is the likely source of Ca. The concentrations of sodium (Na) and boron (B)
remained constant. Typical results for the basic and acidic washing of a specimen are shown
in Table 4. The analysis of deposits is described in the next section.
TABLE 2--Typical coolant concentrate analysis.
Element ~g/mL
A1 <3.00
B 708
Ba < 1.20
Ca <0.60
Cu <3.00
Fe <3.00
Mg <0.60
Mo <3.00
Na 5990
P 1940
Pb <6.00
Si 622
Sn < 12.0
Zn <0.60
K <6.00
TABLE 3--Analysis of an experimental coolantformulation, solublefraction. Concentration changes

as a fimction of time in use.
Elements
Sample Si, tzg/mL Al, t~g/mL Ca, ttg/mL B, mg/mL Na, mg/mL
Initial 132 <3 26.4 1.14 3.06

100 h 105 <3 12.9 1.13 3.03
200 66 <3 3.0 1.13 3.01
300 49 3.9 2.4 1.13 3.00
400 14 14.3 0.7 1.14 3.01
500 12 8.6 1.6 1.14 3.01
600 13 39 <0.6 1.13 3.01
700 12 31 1.3 1.12 2.99
800 12 30 1.3 1.11 3.00
900 I1 20 2.4 1.17 3.15
In special cases, where better detection limits are needed, more concentrated coolant solu-
tions can be analyzed. We have analyzed coolant solutions with ethylene glycol concentrations
ranging from 10% to 100%. For these cases, the standards must be prepared in matching eth-
ylene glycol concentrations. Detection limits on the order of 0.05 to 2 #g/mL were obtained.
An example is the determination of lead (Pb), which is a toxic element, in used and treated
coolants.
Analysis o f Solids
Solid deposits recovered from different parts of cooling systems have been analyzed. The
analysis of these materials is generally related to a problem-solving effort.
Many used coolants contain insoluble materials. These are separated from the cool-
ant by centrifuging. The supernate is decanted, and the unwashed solids are dissolved in a
loosely capped 15-mL polystyrene centrifuge tube. A mixture of 3:1:0.1 hydrochlo-
ric:nitric:hydrofluoric acids is used a t approximately 90~ digestion at higher temperatures
may volatilize Si and B. In general, these solids have been found to contain one or more metals
TABLE 4--Acid and base rinses of treated specimens.
Element KOH Rinse, #g/mL HNO3 Rinse, tzg/mL
AI 85.4 1260
B 71.3 110
Ca 1.40 702
Cd <0.20 2.23
Cu 90.4 975
Fe 18.8 63.7
Mg <0.10 19.2
Mn 0.79 19.1
Na 229 170
Pb 124 138
Si 81.1 10.9
Sn 26.8 31.2
Zn 72.6 312
K ... 1.41
TABLE 5--Unwashed insolubles in a used coolant--typical

analysis.
Element #g/mL
AI 40
B 12
Ca 36
Fe 52
Cu 6.7
Na 120
P 79
Si 82
Zn 11.2
K 62
related to water quality, corrosion problems, solder components, or silicate dropout. Typical
data are shown in Table 5.
Deposits from radiators, thermostats, water p u m p components, and filters have been ana-
lyzed. For bulk compositional analysis, these samples are generally solubilized by attack with
mineral acids. Organic residue, if present, is eliminated by dry ashing at 450~ or by digestion
with nitric-perchloric acids.
Occasionally, only a small a m o u n t of sample is available for analysis; however, good sen-
sitivity is possible. For example, dissolution of a 10-mg sample into 2 mL, the approximate
m i n i m u m solution volume needed for a simultaneous spectrometer, results in detection limits
in the range of 10 to 50 #g/g. Very small quantities of deposits have been removed by ultra-
sonic agitation or wiping with filter paper and subsequent leaching.
Data for the analyses of deposits from three different radiators are shown in Table 6. These
results exhibit wide concentration ranges and elemental compositions typical of deposits. The
concentration differences for such metals as AI, copper (Cu), tin (Sn), Pb, and zinc (Z) are
important in ascertaining causes of system failures.
TABLE 6--Radiator deposit analyses.
Compound Sample 1, % Sample 2, % Sample 3, %
A1203 44.3 14.9 15.6

B203 2.0 0.4 1.5
CaO 4.4 7.5 2.8
CuO 0.5 5.2 11.3
CdO <0.004 <0.0004 0.036
Fe203 2.8 1.3 2.5
K20 0.5 <0.1 0.4
MgO 1.15 0.04 0.60
MnO2 0.05 0.02 0.06
Na20 5.4 1.9 2.5
P20~ 0.4 <0.2 2.6
PbO 2.5 2.3 11.8
SiO2 25.7 10.6 11.8
SnO2 0.9 0.3 23.9
ZnO 0.7 1.3 6.6
% LOI 450"C 44 74 21
Results reported as oxide. Results corrected for loss on ignition
(LOI) at 450"C.
TABLE 7--Analysis o f deposits on water pump seal springs'.

Sample A, Sample B,
Element Total mg Total mg
Al 1.15 0.068
B 0.20 0.010
Ca 0.006 0.001
Cu 0.29 0.064
Fe 3.03 0.116
Mo <0.001 0.092
Na 1.54 0.16
P 0.57 <0.01
Pb 0.15 0.16
Si 2.6 0.33
Sn 0.024 0.023
Zn 7.67 14.9
K O.12 0.023
Data in Table 7 are for the analysis of deposits on springs of two failed water pump seals
(Fig. 2). The deposit was removed by leaching the spring with 10% hydrochloric-hydrofluoric
solutions. In each case, approximately 15 mg of material was leached, however, the elemental
composition is different. The presence of Zn in the cooling system has been shown to be unde-
sirable [6]. The different concentrations of molybdenum (Mo) and P are indicative of expo-
sure to different coolant formulations.
LA-ICP
The a m o u n t of material that is ablated depends on the laser power, operating conditions,
and sample characteristics. In general, conditions are adjusted to achieve a smooth ablation
rate for the duration of lasing. For metals and alloys, approximately I to 10 ~g of sample are
ablated, whereas 200 to 300 #g are ablated from new alumina ceramics used in water pump
seals. Typical ablation craters are shown in Fig. 3. Results for the bulk analysis of the alumina
by conventional ICP and semi-quantitative LA-ICP are shown in Table 8.
During an investigation of premature water p u m p failures [6], alumina faces from good and
failed seals were examined. A s can be seen in Fig. 4, these faces have three distinct zones; the
FIG. 2--Deposits on exposed pump partx.

FIG. 3--Typical ablation craters on alumina ceramic.
coolant side, the sealing zone, and the air side. The bronze in the sealing zone is deposited from
the phenolic-bronze stator face of the seal. Several surface analysis techniques were used to
map the distribution of metals across the width of used ceramic seal faces; these included scan-
ning electron microscopy/energy dispersive X-ray (SEM/EDX), electron spectroscopy for
chemical analysis (ESCA), scanning auger microprobe/energy dispersive X-ray (SAM/EDX),
and LA-ICP.
Data obtained from LA-ICP revealed that every failed ceramic face contained large concen-
trations of Zn and P in the sealing zone. It was ascertained that this relationship was not
observed with other methods because o f different sampling depths and location of the Zn/P
layer.
TABLE 8--Comparison of lCP and LA-1CP analysis of ceramic

seal face material.
Element Oxide ICP, % LA-ICP, %
A1203 85.8 Major

BaO 0.57 1.1
CaO 1.27 0.9
Cr203 0.13 0.03
CuO 0.017 0.017
Fe203 0.23 0.20
MgO 2.14 3.2
Na20 0.14 1.3"
SiO2 7.8 Major
TiO2 0.12 0.16
ZnO 0.01 0.01
* Contamination of real enrichment of sodium on the surface.
FIG. 4--Ceramic seal surfaces. Areas analyzed b.v LA-ICP.
Lasing of bronze deposits on alumina is akin to ablating metals. Small amounts of material
were ablated with a sampling depth of 100 to 500 nm, the apparent depth of the Zn/P layer
on these samples. Due to extreme surface sensitivity, 1 to 5 nm, ESCA detected approximately
the same concentrations of coolant and bronze components on all faces. Although some Zn
was detected by S E M / E D X , its concentration appeared to be minor. It was determined that
the Zn signal was "diluted" by contributions from the ceramic base, which was sampled due
to the 1000 to 5000-nm S E M / E D X sampling depth. From these data, the film thickness is
estimated to be about 100 to 500 nm. The LA-ICP results were instrumental in determining
the cause of premature water pump seal failures. These data are summarized in Table 9.
TABLE 9--LA-1CP analysis of "good" and "'failed" ceramic

seals, sealing zone.
Relative Concentration
Samples P Cu Sn Zn
"Good"
Seal 1 8 68 11 13
Seal 2 2 78 13 7
Seal 3 3 66 25 7
Seal 4 9 66 19 7
Seal 5 8 73 12 8
"Failed"
Seal 1 21 5 7 66
Seal 2 55 3 12 30
Seal 3 9 13 13 65
DISCUSSION OF ICP EMISSION SPECTROSCOPY 189
Other uses of LA-ICP in our laboratory have included the analysis of silicon carbide pump
seals and radiator hoses. The expected coolant components and some Zn were detected on the
carbide faces. However, no correlation with performance could be established. New radiator
hoses were found to be a potential source of Zn contamination.
Conclusions
With appropriate decomposition methods, ICP emission spectroscopy is a tool for the rapid
determination of metals in engine coolants and other components of cooling systems. The
method is especially useful with a limited sample size or as a part of a problem-solving effort.
The use of laser ablation for sample introduction into an ICP source allows for quick empir-
ical or semiquantitative analysis of refractory ceramics and metallic films. This was especially
important here because the 100 to 500-nm sampling depth o f LA-ICP is not readily available
with other surface analysis methods.
Acknowledgments
The authors would like to thank Dr. J. N. Pike for his assistance in setting up the laser abla-
tion system.
References
[1] Dickinson, G. W. and Fassel, V. A., in Emission Spectroscopy, R. M. Barnes, Ed., Dowden, Hutch-
inson & Ross, 1976.
[2] Winefordner, J. D., Ed., Advances in Analytical Chemistry and lnstrumentation, Vol. 9: Spectrochem-
ical Methods of Analysis. Wiley-Interscience, 1971.
[3] Dittrich, K. and Wennrich, R., in Sample Introduction in Atomic Spectroscopy, J. Sneddon, Ed,
Elsevier Science, Amsterdam, 1990, pp. 107-146.
[4] Abercrombie, F. N., Silvester, M. D., and Stoute, G. S., 1CPlnformation Newsletter, Vol. 2, 1977, p.
309.
[5] Zamechek, W. and Pankhurst, R. G., Developments in Atomie Plasma SpectrochemicalAnalysis, R.
M. Barnes, Ed., Myden & Son, 1981.
[6] Kinisky, T. G. and McKenzie, D. A., "Investigation of Premature Water Pump Seal Failures," pre-
sented at the 1990 Society of Automotive Engineers International Congress and Exposition, SAE
Paper 900806.
DISCUSSION
W. D. Berg 1(written discussion)--From your ICP results, how do you go about determining
the oxide form of each element, as your data indicated? More specifically, can ICP provide
you with CuO and CuO2 level from a simple analysis of copper?
Walter Zamechek (author's response)--ICP results reported as oxides, were simply calcu-
lated as oxides and do not suggest that the elements are present as such. Perhaps Table 6 should
have a statement that reads "reported as oxides." Only elemental concentrations can be deter-
mined by ICP. The results for the deposits were reported as oxides in order to facilitate mass
balance calculations. If c o m p o u n d identification is required, X-ray diffraction or classical
chemical methods can be utilized.
Mobay Corp., Mobay Rd., Bldg. 14, Pittsburgh, PA 15205.

Peter M . Woyciesjes t
The Chemistry of Oxalic Acid Cleaning of

Engine Cooling Systems
REFERENCE: Woyciesjes, P. M., "The Chemistry of Oxalic Acid Cleaning of Engine Cooling
Systems," Engine Coolant Testing." Third Volume, ASTM STP 1192, R. E. Beal, Ed., American
ABSTRACT: Engine coolant candidates undergo a progression of tests to evaluate various phys-
ical properties and performance characteristics. Key to this test progression is the evaluation of
the coolant in vehicles. It serves as the final check point, tieing together the results of the other
various tests. In the fleet test, the corrosion protection and performance characteristics of an
engine coolant are evaluated under actual use conditions.
The methodology used in testing a coolant in a passenger car or a light duty truck is critical.
ASTM Practice for Testing In Car and Light Truck Service (D 2847) serves as an industry
accepted standard for evaluating coolants in vehicles. This standard provides a full discussion of
testing protocol.
Sections 9.3 and 9.4 of ASTM D 2847 outline the cooling system cleaning and conditioning
procedures. These sections stipulate the use of oxalic acid as a chemical cleaner followed by a
borate buffer conditioning fluid. The purpose of these two steps are (1) oxalic acid cleaning
removes inhibitors, rust and other deposition from the cooling system, and (2) borate condition-
ing fluid removes excess oxalate to leave a conditioned surface in the cooling system which is
neither detrimental nor beneficial to the performance of the coolant to be evaluated.
The ASTM Committee D- 15 on Engine Coolants is reconsidering the use of oxalic acid cleaner
in the automotive cooling system. Concerns have been raised regarding potential carryover
effects oxalic acid could have on the performance of a coolant.
This paper reviews the chemistry of the oxalic acid cleaner as well as discusses the effects that
oxalic acid cleaning has on the cooling system surfaces and its potential carryover effects to the
test coolant. An understanding of the chemistry of oxalic acid is crucial in judging its efficacy as
a cleaner to bring cooling system surfaces of a test vehicle to a reproducible base condition before
performance testing.
KEYWORDS: oxalic acid, vehicle testing, oxalate, iron oxalate, borate conditioner, antifreeze
testing
Today's engine coolants are formulated to m e e t a variety o f requirements. Although its pri-
m a r y function is to transfer heat from the engine to the radiator for dissipation, there are other
i m p o r t a n t corrosion inhibition and freeze point properties that the coolant must exhibit to be
fully functional. The A S T M Specification for Ethylene Glycol Base Engine Coolant for Auto-
m o t i v e and Light Duty Service (D 3306) outlines certain m i n i m u m physical, chemical, and
performance requirements that all ethylene glycol base engine coolants must meet.
The physical and chemical requirements have been developed around the use of ethylene
glycol as the coolant base. All m a j o r c o m m e r c i a l ethylene glycol base coolants meet these
standards.
Group leader, Coolant R&D, First Brands Automotive R&D Laboratory, 55 Federal Rd., Danbury,
CT 06810.
190
WOYCIESJES ON CHEMISTRY OF OXALIC ACID CLEANING 191
In terms of performance requirements, ASTM D 3306 outlines a regimen of progressively

more strenuous tests, starting with glassware testing and proceeding to simulated service test-
ing. Each test stresses different aspects of the coolant. ASTM D 3306 provides minimum per-
formance standards that the coolant should meet before it is considered for more advanced
testing such as engine dynamometer and vehicle testing.
Key to the development of antifreeze/coolants is the ability to differentiate performance. A
well designed development program structured initially around the tests outlined in ASTM D
3306 provides the coolant scientist with minimum performance requirements to evaluate a
test coolant against and information that could help avoid a catastrophic coolant related fail-
ure in fleet testing.
Ultimately, fleet testing is the best measure to differentiate coolant performance. It provides
a direct measure of the performance and functionality of the coolant under field conditions
and the needed significance to the data obtained from D 3306 testing. Given the time, effort,
and cost that must be expended to run a fleet test, it is imperative that the fleet test be well
designed, maintained, and controlled.
ASTM Practice for Testing Engine Coolants in Car and Light Truck Service (D 2847), pro-
vides a procedure for evaluating the corrosion protection and performance of an engine cool-
ant in passenger cars and light trucks. As part of the pre-test preparation procedure, both the
engine and cooling system are checked to ensure integrity. Before initiating the coolant test, it
is desirable to remove inhibitors and resulting deposits and establish a reproducible, condi-
tioned surface on the cooling system components. The current cleaning (Section 9.3) and con-
ditioning procedure (Section 9.4) recommends the use of oxalic acid (H2C204) followed by a
low pH borate conditioner (Annexes A.3). This paper discusses some concerns that should be
addressed when cleaning with oxalic acid.
Background
The ability to run a well controlled, statistically significant fleet test is, in part, dependent
upon being able to control the variables (the vehicle and cooling system history, operating
conditions, test procedures, and so forth) introduced into the test.
A brief summary of the cleaning and conditioning steps as outlined in ASTM D 2847 is as
follows: First, drain the cooling system of its coolant and flush the system with water. The
system should then be drained of any remaining water and refilled with water. Oxalic acid is
added to the system, and the engine started. The engine is run for 1 h after the thermostat has
opened to allow the hot oxalic acid solution to circulate throughout the cooling system. The
system is drained and water flushed to remove residual oxalic acid and reaction products. The
system is drained and a 50% solution of the low pH borate conditioning coolant is installed.
The conditioning fluid remains in the system for one week or 1125 km (700 miles).
The oxalic acid reacts with the metal surfaces it comes into contact with during the cleaning
procedure, removing residual rust and deposits. The water flushing removes most of the oxalic
acid and its reaction products. Some of the oxalic acid, however, will react with the ferrous
metals to form an insoluble species, which is not always removed completely in the subsequent
conditioning and flushing steps. As a result, it will be shown that at high concentrations the
oxalate species can have a negative effect on the performance of the test coolant which follows.
Chemistry of Oxalic Acid

Oxalic acid is one of the oldest known acids and for many years has been used extensively
as a precipitating and chelating agent. The oxalate ion, C2042, functions as a bidentate ligand
with the formation o f the five-membered chelate ring. Simple or complex oxalate compounds,
or both, are known for most of the metals on the periodic table [ 1 ]. Except for the alkali metals,
most simple oxalates are insoluble in water. However, all oxalates dissolve in acids and most
dissolve in excess oxalate to form complexes. For instance, the solubility of lead oxalate,
PbC204, increases linearly with excess oxalate ion concentration [2], which is indicated by the
formation o f Pb(C204)f 2.
During the cleaning step, the oxalic acid will react with the surfaces with which it comes in
contact, removing inhibitors and deposits. Oxalate has been shown to form a variety of com-
plexes with the metals typically found in an automobile's cooling system (for example, alu-
minum, iron, copper, and lead (Table 1).
Cooling System Chemistry

Most engines today employ a cast iron block, and either a cast iron or aluminum cylinder
head. Depending on the history of the engine, the condition of the cooling system surfaces can
be quite different from one vehicle to another. The reactions that occur on each surface will,
in part, be dependent on its history. This paper discusses the chemistry of the cast iron surfaces
that can result from the oxalic acid cleaning and borate conditioning steps.
It has been shown that during normal use the metals in the cooling system can be corroded
by direct chemical attack or by electrochemical processes, either galvanic when dissimilar met-
als are electrically coupled, or local-cell action between anodes and cathodes on the surface of
a single metal [3].
Local cell action on an iron surface in a neutral or alkaline environment is represented by
the following equations
2Fe ~ ---"2Fe +2 + 4e- (1)

4e + 2H20 02 ~ 4(OH) (2)
2Fe +2 + 4(OH)- ~ 2Fe(OH)2 (3)
2Fe(OH)2 + xAO2+ H20 --- 2Fe(OH)3 (4)
Equations 1 through 4 provide a path by which iron dissolves at the local anode, and electrons,
e - , flow within the metal to a cathodic area of the same metal surface. Oxygen and water accept
TABLE 1--A listing of several common metal oxalate complexes

that couM be forrned as the result of oxalic acid cleaning an
automobile's cooling system [10].
Solubility in~
Oxalate
Compound Color Cold Water Hot Water
A12(C204)3 9 4H20 white insoluble insoluble

CAC204 colorless 0.00067 0.0014
CAC204 - H20 colorless insoluble insoluble
CuCzO4 9 ~H20 bluish white 0.00253
FeC204 - 2H20 pale yellow 0.022 01626
Fe(C204)3 - 5H20 yellow vs vs
PbC204 white 0.00016 ...
ZnC204 . 2H20 white 0.00079 ...
a vs = very soluble. All oxalate complexes dissolve in acids, and
most of them in excess oxalate to form complexes [1].
the electrons to produce hydroxyl ions, O H - . The hydroxyl ions react with the dissolved iron
to produce insoluble ferrous hydroxide, Fe(OH)2, which can be further oxidized to ferric
hydroxide, Fe(OH)3.
The hydrolysis o f ferric solutions with sodium hydroxide solutions develops with the for-
mation of dimer iron complexes, [(H20)4Fe(OH)2Fe(H20)4] followed by a series of poly-
nuclear complexes. When these complexes are sufficiently large in size, coagulation takes place
and an amorphous gel-like precipitate of ferric hydroxide is formed.
There are a number of iron oxides that are important to iron chemistry (Table 2). The iron
oxides formed in a cooling system depends on the local environment. Certain differences
appear to exist between iron oxides formed in water and glycol solutions.
The iron oxides formed in a car with 100% water as the coolant shows Fe304, gamma-Fe203,
and gamma-Fe203 9 H20 on the surfaces. These three oxides and alpha-Fe203. H20 are found
also in solution. The oxides formed, appear to be larger and more crystalline in nature when
compared to those formed in a car containing a 50% ethylene glycol solution [4].
In the earlier study [4], the surface oxide formed in a car containing ethylene glycol coolant
solution consists of gamma-Fe203 and Fe304 with Fe304 and gamma-Fe203 in solution.
In a latter study [5], the iron oxides formed in cars with neglected ethylene glycol antifreezes
were found to be predominately Fe~O3 with small amounts of FeO. In both glycol studies,
the oxide particles formed are smaller than those in cars run with 100% water, and are
often accompanied by amorphous material. The oxalic acid will be added to this environ-
ment.
Titration of an Automobile's Cooling System

In order to develop a better understanding of the effect the oxalic acid cleaner has on the
cooling system a titration of a vehicle's cooling system that had been oxalic acid cleaned was
carried out using measured caustic additions. A Chevrolet Impala with a cast iron engine was
conditioned using the following procedure: (1) drain the cooling system and water flush, (2)
pre-rust for 2 h with a dilute acetic acid solution, (3) water flush, (4) oxalic acid clean for 1 h,
TABLE 2--Listing of some common forms of iron oxide [4,10].

Crystal
Iron Compound Structure Color Comments
FeO cubic blue, black frequency of occurrence not clear
FeO(OH) ... brown when heated to 200~ forms alpha-
Fe203
Fe(OH)2 hexagonal pale green oxidizes and darkens quickly in air
Fe(OH)3 ... reddish-brown not stable, converts to hydrous oxide
gamma-Fe203 9 H20 orthorhombic light brown to abundant constituent of ordinary
yellow rust
alpha-Fe203 9 H20 orthorhombic yellow generally found associated with
gamma-Fe203 9 H20, but in
minor quantities
Fe203 - xH20 amorphous red-brown hydrated Fe203
gamma-Fe203 cubic often reddish diffraction lines similar to those of
magnetic iron oxide
alpha-Fe203 rhombohedral reddish brown frequently produced at elevated
temperatures
Fe304 cubic black magnetic iron oxide
(5) water flush, and (6) drain the system and fill with distilled water and allow the system to
come to equilibrium, approximately 16 h. At this point, the titration of the car's cooling system
was initiated.
The car was started each morning at approximately 8:15 am and allowed to warm up for 30
min. 1 m L of a 50% by weight sodium hydroxide solution (NaOH, 50%) was added directly to
the cooling system every 45 min until approximately 4:00 pm. The car was then turned off
and allowed to sit overnight. Before each addition, a sample was taken from the system for
analysis. Over the course of 4 days the pH, RA, and oxalate levels were monitored. The titra-
tion profile of the oxalic acid cleaned vehicle is shown in Fig. 1.
The initial pH of the water in the system is slightly acidic. It is clear from the titration profile
that the cooling system is behaving in a manner characteristic of a buffered system between
p H 7.0 and 8.0. The addition of base has very little effect on the pH of the system in the pH
region <8.0. However, at a pH > 8 . 0 , only a small amount of additional base is required to
significantly increase the pH of the solution.
In addition to the titration curve, a profile of pH versus oxalate level was plotted (Fig. 2).
The two profiles can almost be superimposed, suggesting that the p H and oxalate levels are
closely related. Furthermore, this chemistry occurs at a p H > 8.0. This can easily be seen from
a plot of oxalate level versus N a O H addition (Fig. 3). The soluble oxalate level starts out at
zero and increases linearly with the addition the NaOH. The increase closely corresponds to a
2:1 molar ratio of N a O H to oxalate. Only at higher levels of NaOH, at a pH > 8.0, is there
any appreciable deviation of the experimental line away from the theoretical 2:1 molar ratio
line.
This 2:1 molar ratio is consistent with the bidentate coordination chemistry of oxalate dis-
cussed earlier. These results support the idea that the oxalic acid reacts with the iron surfaces
to form an insoluble ferrous oxalate (FeC204 9 2H20) film. This species then behaves as a
pseudo buffer between a p H range o f 7.0 and 8.0. The addition of a high pH fluid, engine
coolant, or as in this case NaOH, will react with the residual ferrous oxalate in a 2:1 molar
ratio to free oxalate from the surface and results in the consumption of base and lowering of
12
j
11
lO
7
j-. F"
S
6
5
0 ~'o r 2'o 2'5 ~'o ~'5 4'o 4'5 sb 5~
mls of NaOH,50~ added
FIG. l--Titration of the coo]ing system of a Chevrolet Impala that has been oxalic acid cleaned.
)H
12
11
10
5~0 10'00 15'o0 2000 25b0

OXALATE (ug/ml)
FIG. 2--Results from the titration o f a Chevrolet Impala: oxalate level as a function o f pH.
OXALATE (uO/mL)
2500
500 ""~'/"
0 i i
~ 1'0 15 2'o 25 ~'0 3'5 ~'0 ~;'5 5b 5~
mls of NaOH.50X added
FIG. 3--Plot o f milliliters titrant (NaOH, 50% by wt) versus the appearance o f oxalate in solution.
(--) theoretical line, 2:1 N a O H to oxalate mole ratio; (- - - -) experimental line, N a O H to oxalate.
the pH and RA of the solution. The magnitude of the pH and RA drop will be dependent on
the amount of ferrous oxalate present in the system.
The Formation of Iron Oxalate

To develop a better understanding of the chemistry involved with the oxalic acid cleaning
step, it is necessary to characterize the actual species formed on the iron surface. Of particular
interest are the oxalate complexes that might form in a cast iron system treated with oxalic
acid. Two common forms of iron oxalate have been reported in the literature: ferrous oxalate
(FeC204 9 2 H 2 0 ) and ferric oxalate ( F e 2 ( C 2 0 4 ) 3 9 5H20). They can be formed directly by reac-
tion of oxalate solutions with iron oxide.
! 1
I I
20~ I I
Nylon f l s h llne
suspension
B~
~I r~
Hodtfled
cartridge
--assembly
FIG. 4--Heat flux test assembly.

In automobiles oxalic acid will react with the cast iron and steel surfaces of the cooling sys-
tem as well as with any iron oxide deposits that might have accumulated in the system. It was
shown in earlier studies [4] that cleaning a cast iron automobile engine with oxalic acid can
result in the formation of ferrous oxalate on the walls of the cooling system.
Work was undertaken in the laboratory to confirm this result and to determine the effect of
the conditioning with low pH borate solution on this film. A test [6] utilizing a series of heat
rejecting and accepting gray cast iron specimens was undertaken (Fig. 4). The test was run as
follows: the metal specimens were oxalic acid cleaned for 2 h at 82~ (Fig. 5), conditioned with
borate at 820C for 24 h and exposed to a fresh fill of a standard 50% engine coolant solution at
82"C for 1 week. Samples taken before the oxalic wash and after each subsequent step were
examined by optical microscopy, scanning electron microscopy (SEM), X-ray diffraction
(XRD), auger electron spectroscopy (AES), and energy dispersive X-ray analysis (EDX) [ 7].
Emphasis was placed on characterizing the heat rejecting surfaces.
The most notable result from the X-ray diffraction work is that a heavy layer of ferrous oxa-
late (FeC204 9 2H20) forms on both the heat rejecting and heat accepting cast iron surfaces
washed with oxalic acid. Cross-sectional SEM micrographs show the ferrous oxalate on the
heat rejecting specimen is often in excess of 10 um thick. The surface appears patchy with thin
spots in areas (Fig. 6).
These results are in agreement with work done by Laudere et al. [8]. They investigated the
reaction of a 5% aqueous solution of oxalic acid with products of iron corrosion. Synthetic
Fe304 and gamma FeO(OH) were treated with oxalic acid, and the course of dissolution was
followed by X-ray diffraction. Crystals of oxalic acid and Fe2(CzO4)3 - 5H20 were isolated from
the liquid phase and solid gamma-FeO(OH) from the reaction of gamma FeO(OH) with oxalic
acid (Eq 5).
FeO(OH)(s) + H2C204 --~ Ee2(C204)3 + H2C204 + FeO(OH)(s) (5)

The reaction of Fe304 with oxalic acid yielded Fe304 and FeC204 9 2H20 in the solid phase.
Crystals of oxalic acid, Fe2(C204)3 9 5H,O and FeC204 - 2H20, were isolated from the liquid
phase (Eq 6).
Fe304(s ) -[- H:C20, ~ Fe304(s ) -~ FeC204 9 2H20(s) + H2C204 + Fe2(C_~O4)s 9 5H20 (6)
FIG. 5--The heat-rejecting iron disk and heat-accepting iron coupon after cleaning with oxalic acid.
FIG. 6--SEM micrograph ( 50) of the oxalic acid cleaned surface of the heat-rejecting cast iron disk.
Oxalic acid reacts with the iron surfaces to form the slightly soluble ferrous oxalate dihydrate
and the soluble Fe2(C204)3 9 5H20. In practice, as a result of the cleaning procedure FeC204
9 2H/O will be left behind as a film on the walls of the cooling system while the soluble
Fe2(C204)3 9 5H20 species is washed out of the system by the subsequent water flushing steps.
Following the borate conditioning step, the surface topography is more patchy, indicating
partial removal of the iron oxalate (Fig. 7). XRD shows that the iron oxalate deposits remain-
ing behind on the heat rejecting and accepting specimen appear to have been converted to
calcium oxalate, CAC204 9 H20. Auger analysis of the heat-rejecting surface reveals iron, oxy-
gen, and calcium together with lesser amounts of sodium, boron, and carbon9 This is consis-
FIG9 7 - - S E M micrograph ( 50) of the heat-rejecting cast iron disk surface following the 24-h treat-
ment in the borate conditioner.
tent with the formation of calcium oxalate with some deposits of sodium borate, Na2B407. Iron
oxides are also present. It is apparent from this work that the borate conditioning step does
not remove all of the oxalate from the metal surfaces.
XRD shows that after one week exposure to a 50% solution of fully formulated coolant,
oxalate species are no longer present on the heat rejecting or accepting metal surfaces. The
only major component shown by XRD on the heat-rejecting surface was gamma-Fe203. The
heat-accepting surface has a relatively thin layer of material present. It cannot be identified by
X R D because of its noncrystalline nature. The heat-rejecting surface also contains some non-
crystalline material, which is masked by the dominant gamma-FezO3 X-ray pattern. AES and
EDX spectra suggest that the underlying bulk consists mostly of Fe203 with a thin upper film
containing silicon, magnesium, oxygen, calcium, carbon, and phosphorous. This suggests an
upper film made up of silicate and phosphate, which are the inhibitors present in the coolant.
These results are consistent with the findings reported above, which describe the formation
ofgamma-Fe203 in ethylene glycol coolant solution. The absence of Fe304 could be a result of
differences in the length of exposure to the coolant or the starting condition of the surfaces, or
both. In both examples, the presence of an amorphous or noncrystallinematerial was detected.
Fleet Test Results

A reasonable question to ask at this point is: "Will there be any carry over effect from the
oxalic acid cleaning on the performance testing of the coolant which follows?" In a fleet test
[9] of 33 Ford Econoline 150 vans with 300 CID (4.9-L) six cylinder cast iron engines, which
were oxalic acid cleaned and conditioned according to D 2847, there were pronounced differ-
ences in the performance of the test antifreezes depending on the starting mileage of the test
vehicles. The fluids' performances were seriously impaired when tested in high mileage vehi-
cles. As can be seen in the 24 140-km (15 000-mile) specimen weight loss for fluids A&B (Table
3), the weight losses were greater in the vehicles having the higher starting mileage. The same
trend holds for the other three coolants tested, with emphasis on aluminum, cast iron, and
solder weight losses.
TABLE 3--24 140-km (15 O00-miles) Ford Econoline fleet test results. The average (of2) metal
specimen weight loss data (mg/specimen) for Fluids A and B.
Initial Oxalatea Modine ASTM Cast Cast

Mileage Level,ppm Solder Copper Solder Brass Steel Iron A1
EXPERIMENTALFLUIDA
2 433 230 41 32 17 24 6 1 10
62 835 125 19 283 10 10 24 22
79 585 865 261 13 482 16 54 95 37
82 890 208 12 465 12 95 44 32
99 189 972 160 4 358 7 83 36 36
EXPERIMENTAL FLUID B
2 982 110 67 34 22 25 5 8 6
3 167 287 31 23 18 22 11 1 7
72 850 606 160 16 361 14 16 62 24
101 744 801 186 16 464 22 74 62 28
135 003 950 96 16 306 15 46 27 27
Oxalate levels were measured at 16 093 text km (10 000 test miles).
Analysis of the 16 093-km (10 000-miles) solution data showed significant correlations
between the RA(Linde) 2 drop, starting odometer mileage and soluble oxalate level in the anti-
freeze solution. The correlation between starting mileage and oxalate is fairly good for each
fluid, R 2 = 0.9965, 0.9662, 0.9252, 0.4180, and 0.4263 for fluids A-E, respectively. This sup-
ports the theory that the higher the starting mileage the higher the oxalate level. In Table 3,
the metal specimen weight loss has not been negatively affected for the lower mileage vehicles
( < 4 0 0 0 miles) with low oxalate levels. However, the specimen weight loss data obtained in
the higher mileage vehicles has been negatively biased by the oxalic acid cleaning step. The
composite correlation for fluids A - E of starting mileage versus oxalate level appears to hold,
R 2 = 0.6400, although there is more scatter in the data. This is due, in part, to differences in
the test coolants and the prior history of the cooling system. Figure 8 gives a composite plot of
the fleet test results for five different test fluids, A-E.
It is proposed that oxalate enters the system as a result of the oxalic acid cleaning step. An
insoluble layer of ferrous oxalate is formed in the cooling system, which is not washed out in
the subsequent flushing steps, and is not always completely cleaned out with the low p H borate
conditioning solution. This residue of ferrous oxalate appears to be the reason for the larger
than usual drop in RA and p H and may be responsible for the large drop in some of the
coolant's inhibitors. Regression analyses of the starting miles versus RA(Linde) drop (Fig. 9)
and oxalate versus RA(Linde) drop (Fig. 10) show that there is a stronger correlation between
the oxalate level and RA(Linde) drop than the starting miles versus the RA drop; R 2 = 0.4985
versus 0.2816, respectively. The magnitude of the RA drop appears to be related to oxalate
2 Reserve Alkalinity (Linde) is equivalent to RA(ASTM) for a 33% by volume solution. It is a way of
expressing the RA for fixed EG-water solution. It provides a way of comparing RAs directly. The RA of
any fluid can be converted to RA(Linde) using the following equation: RA(Linde) = RA(ASTM) * 330/
% by volume ethylene glycol in the sample. It is defined as the number of millilitresof 0.1 N HCI required
to titrate 100 mL of a 33% by volume solution of antifreeze to a p H of 5.5.
KILOMETER, Thousands
140
120 . . . . . . . . . . . . . . . . . . . . . i .......... i ..........................
10080_
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9. . . . . . . 9 9 9 9 9
60 . . . . . . . . . . . . . 9 ..................................
40 . . . . . . . . ram--- W . . . . 9
O_o :i .... , . ........ ...... , ...... .........................
25 225 425 625 825 1,025

OXALATE, ppm
FIG. 8--Oxalate level as a function of starting mileage for Ford Econoline fleet test, at 16092 test km
UO 000 test miles).
KILOMETER, Thousands
140
120 - . . . . . . . . . . . . . . . . . li . . . . . . . . m . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
100 -- ..... J . . . . . . . . . . . . . 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 '
80 - J 22 " . . . . . . . . . . . .
60 ,
40 . . . . . . 9
/ 9 9 ~m 9
20-
9 Nil 9 9 9 9
0 2 4 6 8 10 12 14 16 18 20 22
RA(Linde) DROP
FIG. 9--RA(Linde) drop as a function of mileage for Ford Econoline fleet test, at 16 092 test km
(10 000 test miles).
RA(Linde) DROP
25
0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 . . . . . .
15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 ...................
10 . . . . . . "-. .............. ~ ...................
, .... . ~ , .................. , .......... , ....
I 1 I i
25 225 425 625 825 1,025
O X A L A T E , ppm
FIG. l O--Oxalate level as a function of RA(Linde) drop for Ford Econoline fleet test, at 16 092 test km
(10 000 test miles).
concentration (Fig. 10). However, it is apparent from the magnitude ofR 2that there are other
variables that contribute to the RA drop, such as ethylene glycol degradation and corrosion
processes.
Alternative Conditioning Procedure

From the data presented here it is apparent that oxalic acid cleaning of a vehicle's cooling
system can leave behind an insoluble film of ferrous oxalate that is not adequately removed in
subsequent flushing and conditioning steps. Although oxalic acid is an effective cleaner, the
residual ferrous oxalate can bias the subsequent test. It is necessary to find a way to remove
this film with a conditioning step.
As noted above, the current borate conditioning step is not always effective in removing the
oxalate film. The car titration showed the iron oxalate film buffers in the same range as the
conditioner, and that only at higher pH, >8.0, is the oxalate cleaved from the surface. It is
theorized that a higher pH version of the borate conditioner should remove the residual oxa-
late more efficiently.
Table 4 shows the oxalate levels for a fleet of 1989 Grand Am's equipped with 2.3-L Quad-
4 engines. The engine consists of a cast iron block and a l u m i n u m head. All the test vehicles
had accumulated approximately 48 280 km (30 000 miles) on a previous antifreeze/coolant
test. The vehicles were cleaned and conditioned following D 2847 procedures, except that a
pH 10.0 borate solution was used in the conditioning step in place of the low pH borate con-
ditioner specified in the standard. In all cases the oxalate levels in the test vehicles were low,
and the vehicles did not experience a large initial drop in pH or RA. It appears that the higher
pH conditioner was effective in removing the oxalate from the cooling system surfaces. As can
be seen in Table 5, the metal specimen weight loss data for these vehicle have not been affected
by the oxalate cleaning. The values are similar in magnitude to the low mileage vehicles listed
in Table 3.
At the end of the second 48 280 km (30 000 miles) test, six of the vehicles were cleaned and
TABLE 4--Oxalate levels form the Grand A m fleet test as a function o f test mileage.
Fleet Test 1 Fleet Test 2
Oxalate Level, Oxalate Level,

Test ppm ppm
Vehicle Initial, km At x Test km Initial, km At x Test Miles
846 53 440 31 24 140 . . . . . . . . .

36 48 280
856 51 452 15 24 140 100"993 "i9 8"047
15 48 280
858 51 909 14 24 140 102 660 86 8 047
... 48 280
860 51 727 24 140 101 406 134 8 047
43" 48 280
862 51 752 32 24 140
48 280
864 51 850 "15' 24 140 110 320 64 8 047
16 48 280
868 51 787 ... 24 140 109 320 54 8 044
18 48 280
870 52 709 15 24 140 111 631 134 8 047
19 48 280
TABLE 5--24 140-km (15 000 miles) Grand Am fleet test results. The average (of2) metal specimen
weight loss data (rag~specimen)for Vehicles 846, 856, 858, 860, 862.
Test Initial Oxalate, Modine ASTM Cast Cast

Vehicle Mileage Level, ppm Solder Copper Solder Brass Steel Iron AI
846 53 440 31 27 14 3 7 1 I 21
856 51 452 15 10 19 15 16 2 2 30
858 51 909 14 8 4 5 6 2 1 11
860 51 727 43 a 3 7 3 9 1 - 1 18
862 51 752 32 21 4 4 5 1 0 17
864 51 850 15 23 5 3 7 1 1 27
868 51 787 18a 17 5 1 7 1 -3 26
870 52 709 15 19 4 6 5 2 - 1 10
a Oxalate level, ppm, at 48 280 km.
conditioned for use in a 8047 km (5000 miles) test. In these six vehicles the oxalate levels
remained relatively low as compared to the oxalate levels reported in Table 3; however, they
are higher than in the previous test. A rise in oxalate has been seen in other tests at the First
Brands Corporation's Laboratory. It appears that repeated oxalic acid cleaning can increase
the oxalate levels. It is postulated that the oxalic acid is effectively increasing the surface area
and the amount of residual ferrous oxalate in the cooling system with each cleaning.
Discussion and Conclusions

It has been demonstrated that oxalic acid will react with the metal surfaces with which it
comes in contact in the automobile's cooling system. The reaction of oxalic acid with the iron
surfaces present in the cooling system results in the formation of ferric and ferrous oxalates.
The ferric oxalate and some of the ferrous oxalate is removed from the system with the water
flush following the oxalic acid cleaning step.
A residual ferrous oxalate film will remain on the iron surfaces. The amount of iron oxalate
depends on the condition of the cooling system before the cleaning step and the number of
times it has been oxalic acid cleaned. A cooling system with a larger surface area will have
larger deposits of ferrous oxalate. The borate conditioning step removes ferrous oxalate from
the surface of the cooling system. However, in some cases residual ferrous oxalate will remain.
It has been shown that this oxalate can have a detrimental effect on the pH, RA, and inhibitor
levels of the coolant solution added to the vehicle following the conditioning step. This effect
can be minimized by using a high pH borate conditioning fluid and, when necessary, repeating
the conditioning step with the high pH conditioner.
The option selected depends on the purpose o f the fleet test. In all cases, the better the vehicle
histories are matched in a fleet, the better probability that final test results will be statistically
significant. In new vehicles, the oxalic acid effect appears to be minimal. A single conditioning
step with the high pH conditioner should be sufficient. The older the vehicle, the higher the
mileage, and the number o f times the vehicle has been oxalic acid cleaned will determine the
need for repeating the conditioning step.
As a matter of course, the radiator and water pump should be changed following the con-
ditioning step to minimize the effects that the cleaning and conditioning steps will have on the
cooling system.
A vehicle showing extensive signs of corrosion of the cast iron or cast aluminum cylinder
heads and blocks are not suitable for coolant testing. Heavy corrosion deposits will not be
removed by oxalic acid cleaning, or other commercially available in-situ cleaners.
Acknowledgment
The author wishes to acknowledge the Analytical, Coolant R&D and Engineering Groups
of First Brands Automotive R&D Laboratory for their support of the work presented in this
paper.
References
[1] Krishnamutury,K. V. and Harris, G. M., "The Chemistry of the Metal Oxalato Complexes," Chem-
icalReview, Vol. 61, 1961, pp. 213-246.
[2] Kolthoff, I. M., Perlich, R. W., and Weiblen, D., "The Solubility o f Lead Sulfate and of Lead Oxalate
in Various Media," Journal Physical Chemistry, Vol. 46, 1942, p. 561.
[3] Beynon, E., Cooper, N. R., and Hannigan, H. J., "Cooling System Corrosion in Relation to Design
and Material," Society of Automotive Engineers, Paper 780919, Dec. 1978.
[4] Zientek, E. A. and West, C. E., "The Nature of Rust Formed in Automotive Cooling Systems,"
Internal Report, First Brands Corporation, Danbury, CT, 14 June 1955.
[5] Chessic, R., "A Comparison of Rust Produced by a Simulated Depleted Antifreezeand Rust Formed
by Actual Cooling System Neglect," Internal Report, First Brands Corporation, Danbury, CT, 12
March 1981.
[6] Fisher, A. O. and Whitney, F. L, "Laboratory Methods for Determining Corrosion Rates Under
Heat Flux Conditions," Corrosion, Vol. 15, 1959, p. 257t.
[ 7] Psioda, J. A., "Effects of Oxalic Acid and YA-326 ConditioningPre-Treatment on Corrosion Prod-
ucts Formed on Cast Iron in Borate Test Fluids," Internal Report, First Brands Corporation, Dan-
bury, CT, 24 July 1985.
[8] Laudere, A., Konstants, Z., Upite, A. and Vaivads, A., "Reaction of Oxalic Acid With Iron Hydrox-
ide and Iron Oxide. I. X-ray Diffraction Study," Latvijas PSR Zinatnu Akademijas Vestis, Kimijas
Serija. Wol. 1, 1970, pp. 23-26.
[9] Frost, R. A., "1984 Fleet Test Results on Fluids A-E, Internal Report," First Brands Corporation,
Danbury, CT, April 1985.
[10] Weast, R. C., CRC Handbook of Chemistry and Physics, 68th Edition 1987-1988, CRC Press, Inc.,
Boca Raton, FL, 1987.
Randall J. Stafford ~
Investigation of Deposits on Water Pump Seal

Faces
REFERENCE: Stafford, R. J., "Investigation of Deposits on Water Pump Seal Faces," Engine
Coolant Testing: Third Volume, ASTMSTP 1192, R. E. Beal, Ed., American Society for Testing
ABSTRACT: Mechanical water pump face seals from heavy duty diesel on-highway engines
which were reported to leak were analyzed and were compared to nonleaking seals. The silicon-
ized graphite seal face pairs did not exhibit any significant wear of the faces. Coolant leakage was
traced to deposit film build-up, which resulted in separation of the seal faces.
Analysis of the films by wavelength dispersive spectroscopy revealed elements from the cool-
ant and coolant additives (St, C, O, S, P, K, Na, B, Mo), corrosion metals from the engine (Fe,
Cu, Cr) and hard water elements and contaminants (Ca, Mg, Zn, A1, Mn, CI). Calcium from
hard water and iron from corrosion are believed to accelerate the deposit formation.
KEYWORDS: water pump seal, mechanical face seal, face deposit
The water p u m p has historically been a weak link in the cooling system of engines with a
majority of the water p u m p leakage attributed to the mechanical face seal within the water
pump. Trends in face seal leakage have been documented for the automotive industry [1 ]; but
no comparable documentation has been published for the heavy duty diesel industry. Even
though water p u m p and seal leakage have been a major development area for many years, this
lack of documentation for heavy duty diesels has not allowed a complete assessment of failure
modes.
Standard warranty for heavy duty diesel is 100 000 miles (161 000 km) in the U.S. with
extended warranty coverage up to 500 000 miles (804 500 km). Automotive warranty is typ-
ically 10 000 to 15 000 miles (16 100 to 24 100 km) with extended warranty coverage to
50 000 miles (80 450 km). The greater warranty expectations for heavy duty diesel has driven
the engine manufacturers to increase durability, reliability, and reduce warranty expense
through improvements in seal design and materials.
The design parameters for automotive seal components [2] have been used as guidelines for
the design development of diesel water pump seals. Furthermore, one of the primary failure
modes, abrasion and wear of the face in automotive applications [3], has been greatly reduced
through face material development. The current heavy duty diesel seal construction incor-
porates the design enhancement of unitized seals, and the materials advancements of silicon-
ized graphite primary and mating rings.
The reduction of seal wear by use o f a hard abrasion resistant siliconized graphite face pair
significantly reduced seat failures and shifted attention from leakage due to face wear to seal
leakage being attributed to deposit formation on the faces. These deposits have been reported
in the automotive industry [4,5,6]. Many interrelated factors can contribute to the deposit
formation such as coolant type and composition, engine duty cycle, make-up water, face
J Senior Engineer, Ceramic Engineering & Tribology, Cummins Engine Co., Inc., Columbus, IN.
205
design and composition, and coolant temperature. Deposits have been reported in the heavy
duty diesel industry for alumina/carbon face pairs 2 [5, 7].
Heavy duty diesel coolant is usually a low silicate formulation automotive coolant with the
major component, ethylene glycol, and minor components, such as sodium salts of nitrate,
borate, metasilicate, orthophosphate, merceptobenzothiazole and tolytriazole [8]. The greater
engine life expectancy for heavy duty diesels and higher engine load factor require use of sup-
plemental cooling additives (SCAs) to provide adequate corrosion protection, liner pitting pro-
tection, coolant pH control, and hot surface scale inhibition. The SCAs have major compo-
nents that are sodium or potassium salts of borate, nitrate, nitrite, metasilicate, phosphate,
molybdate, adipate, mercaptobenzothiazole and tolytriazole [9].
The intent of this paper is to document analysis of deposits on the siliconized graphite face
materials in current use for many heavy duty diesels in the U.S. The older seal designs and
secondary market replacement seals represented by two piece construction seals and seals with
alumina/carbon face pairs were not evaluated.
Seal Examination
The seals examined in this investigation were limited to a unitized construction having a
nitrile rubber bellows secondary seal and siliconized graphite primary and mating rings.
The seals were removed from water pumps of heavy duty diesel 10 and 14 L displacement
engines (Cummins L 10 and NT). The water pumps were returned directly by Cummins field
service representatives or returned after collection at a rebuilding facility (Diesel Recon).
Water pumps removed from service due to coolant leakage were examined. The indication of
leakage was visible leakage of coolant or dried deposits of coolant on the water pump body.
The pumps were leak tested (5 psig air pressure in the weephole with water around seal) to
limit the seals for examination to those with face leaks or unidentifiable leaks. A schematic of
the water pump assembly is shown in Fig. 1.
Of the ninety water pumps returned, the majority (70%) of these were attributed to face
leakage due to deposits, debris, or indeterminate causes. The remaining 30% of the water
pumps were removed from service due to leakage through faces cracked in manufacturing or
assembly, torn bellows, misassembly in pump, and excessive rotor run-out, or mechanical fail-
ures of bearings, pulley, shaft, casting and mounting.
The lag time between water pump leakage and removal of the seal for examination was
approximately 3 to 6 months. Because there was no control over the removal and handling of
the pump prior to analysis, the seals were examined in the "as-received" condition and all
analysis was performed without any cleaning procedures. Limited information such as mile-
age was provided on some of the pumps by Cummins service representatives. The mileage at
pump removal ranged from 28 000 to 199 000 miles (45 000 to 320 000 km).
In addition to the leaking water pumps, five nonleaking water pumps were obtained. These
pumps provided baseline information on the face appearance and seal condition of an oper-
ating water pump. The nonleaking pumps were subjected to the same teardown and leak test-
ing procedures as the leaking pumps. The information received with the pumps indicated that
these pumps were removed from service for analysis after 146 000 to 243 000 miles (235 000
to 391 000 km).
Experimental
The seal faces of each seal were visually inspected under a Zeiss reflected light stereomicro-
scope at 20X magnification. A random selection of seals, including nonleaking seals, was
-~Rathert, B., Cummins Diesel Australia, personal communication.

STAFFORD ON WATER PUMP SEAL DEPOSITS 207
FAC E S EAL
BEARINGS
IMPELLER
SHAFT
PULLEY
WAT E R
PUMP
BODY
WEEPHOLE
FIG. 1--Schematic drawing of waterpump assembly.
examined further in an Amray Model 1645 scanning electron microscope (SEM) and the
observed deposits were analyzed for elemental composition by wavelength dispersive spec-
troscopy (WDS) using a Microspec Model WDX-3PC spectrometer with a Model WDX-2A
detector. The deposits on these seal faces were also characterized by Fourier transform infrared
spectroscopy (FTIR) using a Nicolet Instrument Model 450 FTIR, with 4 cm-1 resolution, to
determine the molecular bonding characteristics of/he deposits on the surfaces. Deposits were
analyzed using a microscope attachment to the F-FIR in the reflectance mode.
Results
Visual examination of the seal faces from leaking water pumps showed a range of conditions
from full face deposits to no discernable deposit. Examination of the seal faces from nonleak-
ing water pumps revealed the same range of surface conditions (full face deposit to no deposit).
There was no evidence of any significant wear on any of the seal faces, with only a few localized
scratches in the graphite surface of the siliconized graphite material.
Examination of the seal faces in the SEM showed no evidence of wear. A reference photo-
graph of a new seal face is shown in Fig. 2. The surface microstructure consists of a silicon
carbide matrix with islands of graphite and random irregular pores.
Faces from nonleaking seals are shown in Figs. 3 and 4 and represent the range in face
appearance observed for nonleaking seals. The thick full face deposit in Fig. 3 was the sealing
surface during operation. Where the thick deposit is broken away from the surface, an addi-
tional underlying thin film deposit was observed. There was no deposit on the wear face of the
seal in Fig. 4, but crystalline dried coolant deposits were scattered across the face.
Faces from leaking seals are shown in Figs. 5 through 9 and represent the typical face appear-
ances observed for leaking seals. Comparison of the seal in Fig. 5 to the new seal in Fig. 2 shows
that the graphite islands and surface porosity are covered by a full face deposit. Additional
raised deposits, from gel dehydration, cover part of the outer 50% of the surface. Figure 6 shows
FIG. 2--New seal face of siliconized graphite, primary ring.
a deposit band covering the inner 35% of the seal face. The deposit does not fully cover the
surface graphite within the deposit band. Figure 7 shows sporadic deposits on the face with an
underlying deposit band covering the inner 50% of the face. Figure 8 shows a single local
deposit on the seal face. Figure 9 shows a seal face with no apparent deposit and wear scratches
in the graphite phase.
FIG. 3--Nonleaking seal.lklce with thick deposit film, primary ring.

FIG. 4--Nonleaking sealface with dried coolant crystals on surface, primary ring.
The results of the WDS analysis are summarized in Table 1. The elements present in the
coolant are all present in the seal face deposit including corrosion metals from the cooling
system. For the leaking seals, calcium and magnesium are also present in the adherent deposit.
The presence of sodium and magnesium on the new seal face may be due to handling during
seal assembly.
FTIR analysis showed that silicate, phosphate and methyl group bonds were present in the
deposits on all seal faces with carbonyl, amine and hydroxyl group bonds detected on individ-
ual faces. No clear differences in chemical bonding in the deposits on the leaking and non-
leaking seal faces could be determined.
FIG. 5--Leaking sealface with discrete thick film deposits and underlying smeared film overfull face
width, primary ring.
FIG. 6--Leaking sealface with discrete random deposits and smearedfilm over inner 35% ofJhce width,
primary ring.
Discussion
The deposits analyzed on the leaking seal faces were classified as filming failures, with the
film causing separation of the faces and allowing coolant leakage.
Elemental analysis shows the presence of calcium, iron and phosphorus on the seal faces in
Figs. 6 through 8. The presence of these three elements was in agreement with Kiryu's work
on deposit formation [4], where calcium, iron, and phosphorus in solution and nucleate boil-
ing between the faces are required for adherent deposit formation. Absence of calcium or iron
from the system produced nonadherent precipitates.
FIG. 7--Leaking sealJace with discrete random deposits and smearedfilm over inner 50% offace width,
primary ring.
FIG. 8--Leaking sealface with large discrete local deposit, primary ring.
While the seal deposits in Figs. 6 through 8 agree with Kiryu's work, the seal deposits shown
in Fig. 5 are not in agreement and do not contain iron required for deposit formation. How-
ever, boron and molybdenum are present indicating that some of the borate and molybdate
coolant additives were part of the deposit. Hercamp and Hudgens [10] have shown that silicate
gel formation is enhanced by the presence of borate or calcium, or both, and Kiryu [5] has
shown that deposit height is greatly increased by the presence of borate. The relative peak
heights for silicon, oxygen and carbon in the WDS spectrum showed significant increases in
oxygen and carbon compared to the new seal face analysis. We examined the rubber compo-
FIG. 9--Leaking sealface with no visible deposit, wear scratches in graphite phase, primary ring.
TABLE 1--Deposit elemental compositionfrom WDS.

Element
Figure
References Condition SiC Cc Fe P K S Na Zn Cu Ca Mg AI Mn O B Mo C1 Cr
2 New X X X X X X
3 Non Leak X X X X X X X
4 Non Leak X X X X X Xa X Xb X
5 Leak X X X X X X X X X X X
8 Leak X X X X X X X X X X X X X X
a No presence of sulphur in scan of area without dried coolant crystals.
b Peak height reduction of 55% in scan of area without dried coolant crystals.
c Seal face material constituents.
nents of the seal and the graphite in the seal face and they showed no evidence of damage or
degradation, so therefore it is not believed that the increase in carbon peak height can be attrib-
uted to either of these sources. Formation and deposition of a silicate and glycol colloidal gel
is suggested by the silicate and glycol chemistry observed in the FTIR spectrum and this gel is
believed to be the source o f the increased oxygen and carbon in the WDS spectrum.
The elemental analysis for the seal face in Fig. 4 shows the presence of calcium, iron and
phosphorus, however, no deposits are present. This seal did not leak and although the neces-
sary coolant chemical constituents are present for deposit formation (calcium, iron, and phos-
phorus), the operational characteristics (nucleate boiling in the seal face coolant film) may not
have occurred.
Comparison of the seal face in Fig. 4 with the seal face in Fig. 9 shows no deposits adhering
to the seal surface in either case and elemental analysis for both seals reveals similar compo-
sitions of the surfaces. The similarity in appearance and elemental analysis for these two seals,
one nonleaking and one leaking, confirms that seal leakage is a complex issue.
The other nonleaking seal (Fig. 3) did not show the presence of calcium, iron, or phosphorus
(elements believed necessary to form adherent deposits). The smooth surface of the film indi-
cated that it acted as the sealing surface during operation. Comparison of the relative peak
heights for silicon, oxygen and carbon in the WDS spectrum again show significant increases
in oxygen and carbon compared to the new face analysis. Once again, we examined the rubber
components of the seal and the graphite in the seal face and neither showed evidence of dam-
age or degradation; therefore, the increase in carbon peak height could not be attributed to
either o f these sources. The F T I R spectrum of the deposit was indicative of silicate and glycol
chemistry, thus the deposit was most likely a silicate and glycol colloidal gel, which would
account for the increase in the oxygen and carbon concentration in the WDS spectrum. While
calcium and boron are not present, silicate gel formation is possible, especially if the coolant
silicate additions were greater than the silicate solubility. The heavy film of, presumably, sili-
cate and glycol gel on the seal face still does not provide an explanation for the seal not leaking
or suggest the mechanism for deposit formation and retention on the face. Further work using
water p u m p test rigs is necessary to investigate this phenomena.
Water p u m p test rig work on operating conditions, coolant composition and hard water
effects is in progress to investigate the deposit formation and retention mechanisms in a diesel
engine water p u m p face seal.
DISCUSSION ON WATER PUMP SEAL DEPOSITS 213
Conclusions
Deposits present on leaking seal faces have formed by adherence of calcium-iron-phosphate
complexes precipitated during nucleate boiling episodes induced by engine operational con-
ditions. These deposits are sufficient to cause face separation and leakage of coolant.
Deposits on nonleaking seal faces have formed by a mechanism related to silicate gelation
and are not presently understood.
The similar appearance of one leaking and one nonleaking seal face pair with the same ele-
mental analysis illustrates that seal leakage is a complex phenomena involving coolant chem-
istry, face materials and engine duty cycle.
Acknowledgments
The author would like to thank Cummins Engine Co., Inc. for permission to publish this
paper, and T. M. Yonushonis, C. J. Hurt, J. B. Coffin, J. S. Kolhouse, G. A. Totten, T. Rob-
inson, and W. Moenning for their technical or laboratory assistance, or both.
References
[ 1] Symons, J. D., "Factors Contributing to the Water Pump Seal Leakage Problem," General Motors
Research Publication GMR: 1736, Nov. 1974, pp. 1-83.
[2] Matsushima, A., "Guide to Automotive Water Pump Seals," SAE 780404, 1978.
[3] Hirabayashi, H., Ohtaki, M., and Yukimasa, T., "Trouble and Countermeasure for Seals for Auto-
motive Application--Part II," SAE 800530, 1980.
[4] Kiryu, K., Tsuchiya, K., Yonehara, Y., Shimomura, T., and Koga, T., "An Investigation of Deposits
Formation on Sealing Surfaces of Water Pump End Face Seals," Lubrication Engineering, Vol. 45,
No. 1, Jan. 1989, pp. 49-55.
[5] Kiryu, K., Tsuchiya, K., Shimomura, T., Yanai, T., Okada, K., and Hirabayashi, H., "The Effect of
Engines," SAE 890609, 1989.
[6] Kinisky, T. and McKenzie, D., "Investigation of Premature Water Pump Seal Failures," SAE
900806, 1990.
[7] Hudgens, R. D. and Hercamp, R. D., "Filtration of Coolants for Heavy Duty Engines," SAE
881270, 1988.
[8] General Motors Engineering Standard GM6038-M, "Automotive Engine Coolant, Antifreeze Con-
centrate-Ethylene Gycol Type," Jan. 1977 p. J-41.201.
[9] Hercamp, R. D. and Hudgens, R. D., U.S. Patent 4,717,495, 5 Jan. 1988.
[10] Hercamp, R. D. and Hudgens, R. D. "Silicate Gelation in Heavy Duty Diesel Engine Cooling Sys-
tems," SAE 852327, 1985.
DISCUSSION
W. Merce? (written discussion)--Thirty percent of the observed p u m p failures were attrib-

uted solely to quality control problems in assembly of the pumps. Could you comment on the
possibility that the remainder of the failures were also caused by less obvious quality control
problems?
R. J. Stafford (author's response)--The 30% of the seal leaks referenced represent all quality
problems involved in manufacture of the faces, the seal assembly and the water pump as well
as including all leaks caused by bellows or rubber gasketing failures and leaks due to mechan-
ical alignment and runout. The less obvious quality problems, such as face material and rubber
component quality were included in this group of failures.
Since the focus of the paper was on deposits observed on the seal face, all of the above iden-
tifiable causes of leakage were grouped together to present a total picture o f the water p u m p
analysis. The remaining 70% of the leaking seals were subjected to additional analysis for
deposits and are represented by the data presented.
Peter Woyciesjes I (written discussion)--You stated that you do not see silicate deposits on
the seal surfaces and that a low silicate coolant was used with a phosphate/molybdate/nitrite
package. How can you substantiate that pumps are failing as a result of silicate gels?
R. J. Stafford (author's response)--Silicate gelation is not frequently seen in the heavy duty
diesel cooling system given the present state of technology with specifications for low silicate
formulation coolants. However, deposits are observed on some seal faces, which can be attrib-
uted to silicate gel formation on the seal surface.
Hercamp and Hudgens [1] showed that the silicate concentration in coolant of heavy duty
diesel trucks ranges between 200 and 600 ppm. The initial fill of low silicate coolant with the
recommended dosage of supplemental coolant additives (SCAs) provides 1450 to 2050 ppm
of silicate in the cooling system [2,3]. The excess silicate beyond the 200 to 600 ppm in solu-
tion forms protective layers on internal surfaces of the engine and cooling system components.
Addition of SCAs before depletion of the protective layers, use of precharging SCA dosages
instead of maintenance SCA dosages, or development of coolant conditions that reduce the
silicate solubility will result in silicate sols and gels being deposited on the surfaces within the
cooling system.
The seal faces shown in Figs. 3 and 5 of the paper have deposits attributed to silicate gel
formation. In the analysis of these seal faces, detection of silicate on the seal surface by wave-
length dispersive spectroscopy (WDS) was not possible because of the masking effect of the
silicon carbide in the seal ring. However, strong silicate molecular bonding was detected by
Fourier transform infrared (FTIR) spectroscopy for the deposits on the seal surface and after
removal from the surface.
Deposition of silicate on the seal surface as a full face film can create a sealing surface (Fig.
3) and allow the seal to operate without leakage. In contrast, a similar visual seal face appear-
ance (Fig. 5) with similar WDS and F T I R spectra was shown for a seal that had leaked.
As was stated in the paper, we do not presently understand the silicate gelation mechanism
for these heavy duty diesel engine seals and believe that the coolant chemistry, surface inter-
actions, face materials, and engine operating conditions all play roles in how a silicate gel is
generated and deposited on the seal face.
References
[1] Hercamp, R. D. and Hudgens, R. D., "Silicate Gelation in Heavy Duty Diesel Engine Cooling Sys-
tems," SAE 852327, Society of Automotive Engineers, Warrendale, PA, 1985.
[2] General Motors Engineering Standard GM6038-M, "Automotive Engine Coolant, Antifreeze Con-
centrate--Ethylene Glycol Type," p. J-41.201, Jan. 1977.
[3] Hercamp, R. D. and Hudgens, R. D., "'U.S. Patent 4,717,495, 5 Jan. 1988.

Kenji Kiryu, 1 0 s a m u Hirata, i Akira Yoshino, ~Ken Okada, 1 and
Hiroshi Hirabayashi ~
The Relationship Between Sealing

Performance of Mechanical Seals and
Composition of Coolants for Automotive
Engines
REFERENCE: Kiryu, K., Hirata, O., Yoshino, A., Okada, K., and Hirabayashi, H., "The Rela-
tionship Between Sealing Performance of Mechanical Seals and Composition of Coolants for
Automotive Engines,"Engine Coolant Testing: Third Volume, ASTM STP 1192, R. E. Beal, Ed.,
American Society for Testing and Materials, Philadelphia, 1993, pp. 215-233.
ABSTRACT: Because the mechanical face seals in automotive water pumps play an important
role in maintainingthe good performance of engines, they must exhibit reliable sealing perfor-
mance and long operating life. Generally, the sealing performance of mechanical seals is closely
related to the properties of the liquid to be sealed. Automotive engine coolants are generally for-
mulated with a variety of inhibitors to prevent corrosion and erosion &cooling system parts. A
few of these inhibitors have a detrimental effect on the performance of water pump seals.
In this paper, representative types of deposit formation on sealing surfaces are reported. These
seals were taken from pumps that failed due to seal leakage believed to be caused by a combi-
nation of coolant formulations, contaminationsinto coolant, and changes in engine operation
conditions. The failure mechanisms and possible countermeasures involving the sealing mate-
rials, surface quality, and seal structure are discussed.
KEYWORDS: mechanical seal, long life coolant, inhibitors, deposit, wetting, contact angle
End face mechanical seals are generally used for sealing devices in water pumps of auto-
motive engines [1]. Because they play an important role in achieving the optimum operation
of engines, their reliable long lives are required. A variety of failure modes have been experi-
enced in actual service use, such as thermal cracking of the seal ring, excessive wear of the
sealing surfaces caused by abrasive particles, squeaking, and so forth. By applying the appro-
priate countermeasure for each type of failure mode, the probability of these seal failures has
been decreasing [2, 3]. However, another type of failure has also been observed and is caused
by the formation of deposits on seal faces, which results in severe leakage. A solution to this
worldwide problem is required [4-10].
This urgent problem is attributed to a combination of three factors:
1. Qualitative and quantitative changes in long life coolant (LLC) formulations,

2. An increase in severity of engine operation conditions, and
3. Contaminations into LLC from various parts in the cooling system.
Dr. Eng., assistant manager ofR & D #3 Section; chief engineer ofR & D #7 Section; manager of
design #2 Section; Dr. Eng., senior manager ofR & D # 1 Department; and Dr. Eng., executive vice pres-
ident; respectively, Eagle Industry Co., Ltd., 1500 Katayanagi, Sakado-shi, Saitama-ken, 350-02, Japan.
215
This paper deals with the investigation of deposit formation caused by three different for-
mulations of LLC together with their mechanisms and appropriate countermeasures.
Mechanical Seals for Water Pumps

Figure 1 shows the structure and construction materials of the representative mechanical
seal used in field service. The stationary seal ring is loaded by a coil spring. The mating ring is
used on the rotating shaft side. Generally, the seal ring materials are softer (for example, a
variety of carbon materials), and the mating ring materials are much harder (for example, alu-
mina and silicon carbide). All of the metallic parts (parts 5, 6, 7, and 8 in Fig. l) are composed
of stainless steel, which is corrosion resistant to the LLC containing additives. In actual field
service, the failure of most mechanical seals is due to leakage across the sealing surfaces (parts
2 and 3 in Fig. 1).
The mode and probability of occurrence of seal failures are strongly affected by the internal
design of the pump and the layout of the exterior cooling system. In particular, heat and mass
transfer of cooling water including additive are considered to have predominant influences on
the seal failure. In view of these experiences, reproducible tests were carried out under labo-
ratory conditions, so that test results using LLC with different inhibitor formulations could be
compared.
Properties of LLC
For automotive engine coolants, suppression of corrosion and scale deposition on metallic
parts and degradation of rubber parts is required [4]. To satisfy these requirements, a variety
of organic and inorganic inhibitors are prescribed [ 11,12]. Table 1 shows a list of the typical
chemical properties of the three LLC used in the present investigation. The main features of
these three LLC are:
1. L L C - A : This type of LLC is characterized by inorganic inhibitor at a high concentration

of silicate to prevent mainly corrosion of aluminum (A1) parts. For prevention of silicate
gelation, the pH of the LLC is maintained at a high level, approximately 10.
2. LLC-B." This type of LLC is characterized by triethanolamine, an organic corrosion
inhibitor for ferrous and AI parts. An inorganic phosphoric salt is also added at a high
concentration to prevent corrosion of ferrous parts.
3. LLC-C." In this formulation, inorganic inhibitors to prevent corrosion of ferrous and A1
parts (for example, phosphoric and nitric salts) are substituted for the triethanolamine
found in the LLC-B formulation. In addition, mercaptobenzothiazole (MBT) is added
~)- CUP GASKET : NBR

(~)- MATING RING : CERAMICS
(~- SEAL RING : CARBON
(~)- BELLOWS : NBR
(~)- SPRING HOLDER : STAINLESS STEEL
(~- COIL SPRING : STAINLESS STEEL
f .... / (~)- CARTRIDGE : STAINLESS STEEL
(~)- SLEEVE : STAINLESS STEEL
FIG. 1--Mechanical seal for automotive water pump.

KIRYU ET AL. ON MECHANICAL SEALS AND COMPOSITION OF COOLANTS 217
TABLE l--Chemical analysis of LLC.
Parameters LLC-A LLC-B LLC-C

pH 10.1 8.1 8.3
Organic mercaptan benzo triazole 0.28 0.49 0.73
benzoic acid ND 3.72 3.90
triethanolamine ND 4.10 ND
Inorganic phosphate (PO~-3) 1.01 1.14 0.84
nitrate (NO~) ND ND 0.45
silicate (SiO32) 0.15 ND ND
Ion K+ 0.52 0.03 1.06
Na + 0.17 0.80 0.62
at relatively high concentration as organic corrosion inhibitor for copper (Cu) and brass
parts.
Mechanism of Deposit Formation

In this section, deposit formation phenomena are described in relation to the three repre-
sentative LLC formulations. The process of the seal failure due to deposition and leakage is
believed to progress as follows:
1. Deposit formation on seal ring or mating ring surfaces,

2. Growth of deposits,
3. Increase in gap between sealing surfaces, and
4. Occurrence of excessive leakage.
Deposition of Silicate
Typically, two different types of deposit profiles were found on the surfaces of failed seals
taken from actual vehicles that used the LLC-A type coolant. These deposits, which resulted
in leakage across the seals, are shown in Fig. 2. The type 1 deposit is usually observed when a
combination ofgraphitic C as the seal ring material and alumina as the mating ring one is used.
In this case, a black deposit is formed on the surface of the alumina and severe wear of the C
ring occurs. On the other hand, type 2 deposit formation takes place only when the seal ring
material is amorphous C. This failure is characterized by the formation of a white deposit on
the seal ring side and little deposit formation on the mating ring side. Figure 3 shows scanning
electron microscopy (SEM) images of type 1 and type 2 deposits together with element map-
ping by means of electron probe X-ray micro analysis (EPMA). In the case of type 1 deposits,
the elements on the mating ring surface are found to be mainly silicon (Si), oxygen (O), and
C. Based on the chemical analysis of LLC-A shown in Table 1 and the profile curves showing
severe wear of the seal ring, it is considered that the deposit on the mating ring surface is com-
posed o f silicate from the LLC-A and wear particles o f C from the seal ring. In the type 2 depos-
its, only Si and O are detected on the seal ring surface and it is believed that the deposit is
composed mainly of silicate from the LLC-A.
It was reported by Hercamp and Hudgens [5] that silicate gelation can result in deposit for-
mation and subsequent seal leakage. In connection with this problem, the authors carried out
heating tests of LLC-A type formulations in a container with or without a lid in a furnace at
temperatures of 120~ and 150~ for 1 h. Figure 4 shows changes of LLC-A after the heating
I SE&L R I N G
:'n .,:.u '~l DEPOSIT I'~ $~.'~.
~ "..~ ~.~ 9
Type-1 ~ . ~ LATMOSPHERICSIDE~~
I Q UI I D~ S~ l' :D~E ~
" ~ " '"~:'~;':: ":~ ' "~' ~r
"~ ~ Iz 9 0.Sml "~' , ~"
I
~ D~osrr g" SEAL m~G:~l
PII LIQUID SIDE

T y p e - 2
Sa
:::~ I|
z
5a m 0.Smm
t ,
~
-:'~
I
I
:
t;~
t*~
~IATINGRING Z,'
.t
FIG. 2--Two types of silicate deposits infield service.
test. Under the closed condition with the lid, clusters of colloidal particles are found in the
liquid phase. Comparing the change at 150~ with that at 120"C, it is noted that the number
of clusters increased remarkably. The EPMA analysis (energy dispersive X-ray [EDX] spec-
trometry) of the clusters and the filtrate taken from the LLC-A heated at 150~ by means of
filtration was carried out as shown in Fig. 5. The detective intensity of Si in the clusters is much
stronger than that in the filtrate, so it is believed that the clusters are composed of silicate gel.
In contrast to this, under the opened condition without a lid, the liquid phase transforms into
gel, showing increased viscosity.
Sakka and Kamiya investigated the characteristics of silicate for preparation of functional
materials by the sol-gel method [ 13]. According to them, in hydrolysis and condensation reac-
tion, silicate causes gel formation as coagulation of clusters at higher water content. On the
other hand, at lower water content, it forms network gel. The result of the present heating tests
shown in Fig. 4 are verified also by their investigation. Under the opened condition without a
lid at 150~ the LLC-A coolant completely dries up and forms a solid composed of the addi-
tives in LLC-A, mainly sodium phosphate and silicate.
To estimate the temperature rise on rubbing surfaces due to frictional heat, finite element
method (FEM) analysis is attempted as shown in Fig. 6. The calculated result using the input
data simulating practical operating conditions of mechanical seals indicates the elevated tem-
perature on rubbing surfaces from 140~ to 150~ near to the temperature range of silicate
coagulation. Moreover, the following fact should be taken into consideration due to the vapor
pressure of 50% solution of ethylene glycol: On the inner side of rubbing surfaces near to atmo-
sphere, the liquid temperature exceeds the boiling point, whereas on the exterior side near to
pressurized liquid, hardly any boiling occurs. This means that the vapor-liquid interface is
necessarily formed between the inner and outer peripheries of rubbing surfaces. Consequently,
it is easily suggested that the gelation of silicate in LLC-A occurs between sealing surfaces.
Besides these tests carried out under static condition, a number of running tests were carried
FIG. 3--SEM images and X-ray mapping of silicate deposits.
out at various liquid temperatures to relate with the dynamic phenomena in actual sealing of
LLC-A. The observed results of deposit height are shown in Fig. 7. Here, remarkable temper-
ature dependency is confirmed, that is, the deposit height shows a rapid increase with raising
temperature. This fact is considered to verify that the supposition derived from the static test
results emphasizing the significant role of condensation reaction of silicate by dehydration is
also applicable to the dynamic tests.
From the above discussion, the cause of recent increase in the number of the leakage failures
due to the silicate deposition on sealing surfaces can be explained as follows: The increase in
'::3
I
E
O
>
0 5 10
a Energy Value, keV
>0
P
AI
0 l
5 10
I1 Energy Value, keV
FIG. 5--EPMA analysis of clusters (a) and filtrate (b) of LLC-A.
FIG. 6--Temperature distribution of rubbing rings by FEM.

\
E 10-- o~
E
E
O
0 \
~5
1
\
n,-
\
0.1
2.6 2.8 3.0 3.2 3.4
Reciprocal of Liquid Temperature, -1
FIG. 7--Effect of temperature on silicate deposition.
severity of engine operating conditions (for example, increase in rotation speed and temper-
ature rise of coolant) elevates the temperature of rubbing surfaces of mechanical seals, so that
the silicate deposition is apt to occur in recent years.
Deposition of Phosphorus, Iron, and Copper

Figure 8 shows profiles of deposits formed on sealing surfaces in actual service of automotive
water pumps using LLC-B, where excessive leakage was observed. In this case, the deposit
height reaches approximately 20 um on a C ring surface. The EPMA result shown in Fig. 9
indicates that the principal ingredients of the deposits are iron (Fe) and copper (Cu). The result
LIQUID SIDE ATMOSPHERIC SIDE
S At. RING ~i...

:.'. ~:~.
"' "~ ~.~.'2: . ...:'~..., ,;-. #:.,
~~:2, ;" -;."~,.
i:
"~
"."
.x
.'.'"
.:'~., .....
....:
9: 9 .z"
:, .4
: . . ; . ,~.: { l : ";. r
o PO ,t
9 :i- ::" "-""-':: ".'7-r: -'- ;,'Z"~;-t .,r ,-.r ;:.~':~: ,. (~.7~:
:.'~" .r
MATING RING -i! Oft m 0.5 m
FIG. 8--Profiles of P, Fe, Cu deposits in actual use.

FIG. 9 - - S E M image and X-ray mapping of P, Fe, Cu deposits.
of electron spectroscopy for chemical analysis (ESCA) carded out by ion-sputtering the depos-
its and obtaining a concentration profile with depth is shown in Fig. 10. The ESCA results
indicate that Cu concentrates mainly in an extremely thin layer within a few micrometers on
the ferrous deposit.
This type of deposition was frequently observed in heavy-duty diesel engines running a long
distance, for example, exceeding 105km. Once encountered, even immediately after replace-
Surface of Deposit ~ Fe
Surface of Deposit / / ' ~ 4 ~ - Cu20
=-
c)
/~---~-----'---
s ~ ~ - - . . -~- _~- ~. - ~
. . . .
1o 720 716 712 708 704 700 l O / -b - z ; ~

346 342 338 334 330 326
a Binding Energy, eV Binding Energy, eV
FIG. lO--Distribution of Fe (a) and Cu (b) in depth direction.
ment with a new seal, the deposition is apt to be reproduced. It is suggested that deterioration
of LLC-B causes the deposition of this type.
According to Kikuchi and Tajima [12], who observed time changes in concentration of
metal ions in LLC similar to the LLC-B type formulations, which was collected from engines
in field service, Cu ions began to increase after a few months' operation and then Fe ions
showed a rapid increase. Thompson and Lorking [14] considered the process of the change in
these metal ions in connection with the deterioration of LLC.
Based on these viewpoints, the most probable mechanism of the deposit formation in LLC-
B is explained in the following sequence: Formic acid as an oxidation product of ethylene gly-
col is formed and promotes corrosion of Cu parts. Cu ions dissolved in the LLC-B form gal-
vanic cells on surfaces of Fe and AI parts by producing complex salt with amine. These
galvanic cells act as triggers of corrosion of these metallic parts. The Fe ions dissolve into the
LLC. As a result of an increase in concentration of ferrous ions, the self catalytic reaction pro-
motes ferrous deposition on the sealing surfaces in the presence of phosphate as a corrosion
inhibitor. Succeedingiy, on the surface of ferrous deposit, a thin layer of Cu is formed.
Deposition of Dibenzothiazyl Disulfide (DM)

Figure 11 shows a rubbing seal ring surface that caused excessive leakage in actual field ser-
vice using LLC-C. The deposit formation a few micrometers high is observed. EPMA analysis
shows sulfur (S) as a main constituent (Fig. 12). Fourier transform infrared (FT-IR) micros-
copy was used to analyze the deposit. The infrared spectra found that the precipitate is com-
posed of DM (Fig. 13). The molecular structure of DM is characterized by the existence of the
S-S bond (Fig. 14). It is suggested that the deposit on rubbing surfaces is composed of DM.
However, whereas LLC-C coolant contains MBT as a rust preventive for Cu parts, DM is not
typically found in LLC formulations. This supposition is verified also by Kinisky and
FIG. 1 l--Appearance of seal ring surface with DM deposit.
FIG. 1 2 - - S E M image and X-ray mapping of DM deposit.

226 ENGINE C O O L A N T TESTING: T H I R D V O L U M E
100
80
O
Z 60
,<
N 4O
CO
2
<
er
I-
4O3O 40o
WAVE NUMBER, cm -1
FIG, 13--FT-1R analysis of precipitate in LLC-C.
McKenzie [6] who used a similar analytical method. They found the same result in the case
of LLC without MBT. Figure 15 shows the observed result by Yoshikawa et al. [ 15] who found
that DM could be dissolved into LLC from rubber radiator hoses that included DM as a vul-
canizer. In their study, the amount of DM dissolved into LLC from a new hose was approxi-
mately two times that from a used hose, which corresponded to approximately six times of
initially added corrosion inhibitor. Using LLC-C in which a new rubber hose had been
immersed, the result of running test deposit formation on seal ring surfaces was reconfirmed.
This test result also coincides with Kinisky and McKenzie's study [6].
The above discussion can be summarized as follows: The precipitation of DM occurs as a
result of the dissolution from rubber radiator hoses. The DM precipitate is concentrated by
frictional heat and finally adheres on the sealing surfaces, causing seal leakage.
Countermeasures for Failures

The countermeasures for the deposition on sealing surfaces of mechanical seals should be
established by considering suitable rubbing materials, surface roughness, and seal construc-
tion. Here, as a representative of the deposit types, the experimental results of silicate deposi-
tion are explained as follows:
Sealing Materials [ 7]
The authors attempted to reproduce the silicate deposition under laboratory conditions.
Two types of C materials were used--graphitic C S1 causing the type 1 deposit formation and
amorphous C $2 causing the type 2 deposition (Fig. 2). For the mating ring materials, alumina
M 1 with 92% purity and silicon carbide M2 with 97% purity produced by sintering at atmo-
spheric pressure were selected. The mechanical and thermal properties of these four materials
are listed in Table 2.
FIG. 14--Structure of DM.

15
E
_> 10-
o
k3
E
o
9
(O
0 "
N e w LLC + N e w LLC +
Ii
N e w LLC only
N e w Rubber Hose Used Rubber H o s e
(ca. 500h)
FIG. 15--Amount of DM dissolved from radiator rubber hoses.
Figure 16 illustrates the surface profiles after the laboratory tests. The results are indicated
as follows:
S 1 vs. M l - - D e p o s i t i o n on the mating ring M 1 surface

$2 vs. M l - - D e p o s i t i o n on the seal ring $2 surface
S I vs. M 2 - - N o deposition on both surfaces
$2 vs. M2--Deposition on the seal ring $2 surface
These marked differences in the deposit formation will be discussed from the viewpoint of
interfacial chemistry, that is, a change of wetting free energy Wa for intensity of adhesion,
where a contact angle O plays an important role. This relationship is expressed by Eq 1, where
YLvdenotes interfacial free energy between vapor and liquid [16]. Here, the liquid is distilled
water.
Wa --- 3'Lv(1 + cos O) (1)
TABLE 2--Mechanical and thermal properties of rubbing materials.
Seal Ring Mating Ring

Parameters S1 $2 M1 M2
Materials resin bonded carbon alumina silicon carbide

Filler graphite amorphous carbon A1203 SiC
Gravity 1.76 1.73 3.72 3.15
Hardness 54 HS 68 HS 1500 HV 2750 HV
Strength, MPa 45 75 350 490
Coeflicient of linear 1.1 X 10-5 3.1 10 5 6.7 X 10 -6 4.07( 10 -6
expansion, I/K
Heat conductivity, W/m 9K 9.2 2.9 15.9 125.0
Seal Ring Material: S I Seal Ring Material: S 2

~ SEAL ~ ~] I ~" SEALRING '~ IjJ
Mating :~ I'~. . . .
'~':.. ~.'~ ,. .
iRlng
LIQUID SIDE J W~ ATMOSPHERIC
SIDE
Material:
M1 [~.~ NATDIG RING Sd '
li~ I s.. o.~u ~', o ~" J

~ 1 - ~ i ~ I V~ 1 D t ~:
I. SEAL~NG ~,.,.: SEAL RING ~I

Mating
Ring LIQUID SIDE ;5,
~__~... ~ , ~ ,~'
~~- ATMOSPHERIC'SIDE LIQUIDSIDE ~ ATMOSPHERICSIDE
Material: IP" ~ DEPOSIT
M2 li?'F "~TING RING I

I!':., I I , ":~
FIG. 16--Surface profiles after reproducibilitytests.
Because in this test series the same liquid with the same 7Lv was used, the wetting energy W,
of each material can be estimated and compared by measuring the contact angle on each seal-
ing surface (Table 3). The relationship between the deposit formation and the wetting energy
W, is plotted in Fig. 17. Deposits are preferably formed on the rubbing surface with higher
wetting energy. And, in the case of the combination of rubbing materials where both showed
lower wetting energy, no deposit is formed.
As a suitable countermeasure, a selection of both seal materials having lower wetting energy
is required. In fact, graphitic C as the seal ring material and silicon carbide as the mating ring
are the most desirable material combination.
Surface Roughness of Mating Ring

Because the change of surface roughness of mating rings caused by rubbing is low because
of high hardness, the sealing performance of the mechanical seal is considered to be affected
by surface roughness of the mating rings over long durations. Consequently, it is important to
control surface roughness of mating rings within an optimum range.
The laboratory tests were performed with the combination ofS 1 versus M2 in LLC-A, where
surface roughness of the mating ring, M2, was varied within the range from 0.01 to 0.10 #mRa.
Figure 18 shows the relationship between the surface roughness and the rate of silicate depo-
sition. In the roughness range lower than 0.03 #taRa, the deposit is formed on the surface of
the seal ring S 1. In contrast, in the range exceeding 0.06 ~mRa, the deposition occurs on the
TABLE 3--Wetting energy of each rubbing materials.

Seal Ring Mating Ring
Parameters S1 $2 M1 M2
Wetting free energy, 10 .3 N / m 91.1 113.0 106.0 90.3
KIRYU ET AL. ON MECHANICALSEALS AND COMPOSITION OF COOLANTS 229
Mark Seal Ring Surface Mating Ring Surfacel

( No Deposition _.No Deposition
....~ Deposition No Deposition
No Deposition Deposition
E 120
:=" M2 MI
. . . . . . . $2
~110
~
o 100
i
i
,,,
.~ 9 0 ,
/ I
80 90 100 110 120
Wetting Energy of Mating Ring, xlO-3N/m
FIG. 17--Relationship between wetting free energy and deposition.
mating ring surface. In the range from 0.03 to 0.06 umRa, little or no deposition is generated
on either seal surface.
These results can also be discussed from the viewpoint of interracial forces. Figure 19 shows
the relationship between the surface roughness and the wetting free energy of the mating ring
M 1 surface, which is calculated by Eq 1. As the surface becomes smoother, the wetting free
10 0
O o
O.s E 0
~o: E o
o 0 0
0 0 0 o
0
0 0 0 0
0
oO 0
0 0 0 0
0
o~ oOO 0
o o(
o E
O
~E
~.- 9 o
tim b
~m
10
0 0.02 0,O4 0.06 0.08 0.10
Surface Roughness of Mating Ring , ,umRa
FIG. 18--Effect of swface roughness on deposition of rubbing rings.
100
E O O
O
z
~o O O
O
O
~< 95 O
0 O
0
0 O
tr
0
== 0 0
0 0 O0 o
90 !
0 0
0 0
J 0
o
O
Ul
85
o
O
80
0 0.02 O.04 O.06 O.O8 0.10
Surface Roughnessof Mating Ring , JJmRa
FIG. 19--Relationship betweensurface roughness and wettingenergy.
energy of the mating ring becomes lower than that of the seal ring (Fig. 17). In the range of
lower M 1 surface roughness, the deposit is not formed on the mating ring surface but on the
seal ring surface. However, in the range exceeding 0.06 fzmRa, the deposit is not generated on
the seal ring surface but on the mating ring surface, because the wetting energy of the mating
ring M 1 becomes higher than that of the seal ring S1.
It is very important not only to select the suitable combination of seal materials but also to
control surface roughness within the optimum range to prevent the deposit formation.
Construction of Mechanical Seals

As shown in Fig. 7, the silicate deposition is accelerated by raising liquid temperature. Con-
sequently, to suppress deposit formation, one must decrease the temperature of rubbing sur-
faces. In comparison with the conventional seal structure illustrated in Fig. 1, the authors rec-
ommend the improved seal structure shown in Fig. 20 to decrease the temperature of rubbing
surfaces. In the conventional seal, the rubber cup gasket (part 1 in Fig. 1) disturbs transfer of
frictional heat to the coolant. However, the frictional heat transfer in the recommended seal
structure is much improved because the contact area of the mating ring to coolant increases
according to transposition of the cup gasket to the inner side of the mating ring.
~ ~ (~)- CUP GASKET : ~/BR

(~)- MATING RING : CERAMICS
(~)- SEAL RING : CARBON
(~)- BELLOWS : NBR
(~)- SPRING HOLDER : STAINLESS STEEL
.
(~)- COIL SPRING : STAINLESS STEEL
,/' ! (~)- CARTRIDGE : STAINLESS STEEL
\ V | STAINLESSSTEEL
FIG. 20--Improved seal structure.
FIG. 21--Temperature distribution of improved seal by FEM.
FEM analysis was carried out again to confirm the effect of the transposition of the cup gas-
ket on the temperature distribution of the improved seal. In comparison with the temperature
of the sealing surfaces in the conventional type seal shown in Fig. 6, a lower temperature
around 13"C is expected, as shown in Fig. 21. This is also confirmed by measuring the actual
temperature of the rubbing surfaces.
Furthermore, one should also sufficiently consider the environmental condition near the
seal faces, for example, quenching, aeration, and so forth, which can also affect the transfer of
frictional heat.
Durability Tests of the Improved Seal

The durability tests in LLC-A were performed to confirm the effects of the countermeasures
as mentioned above. Figure 22 shows time changes of total leakage including vapor. In Fig.
22, seal-a is the conventional seal and seal-/3 is the improved one, in which three types of coun-
termeasures, that is, rubbing materials, surface roughness and seal construction, are applied.
Seal-a generates abrupt leakage due to the deposit formation in a short period. On the other
hand, seal-~ maintains the satisfactory sealing performance for long duration. The result of
the durability tests confirmed that these countermeasures show the sufficient effects on pre-
venting the deposit formation.
The description in the above concerns the silicate deposition in LLC-A. However, the
authors confirmed experimentally that seal-/3 shows similar good sealing performance also in
LLC-B and LLC-C, that is to say, these countermeasures are also effective to prevent deposi-
tion of phosphorus (P), Fe, Cu, and DM.
Conclusion
The mechanisms and countermeasures of three types of deposit formation on sealing sur-
faces of water pump seals are summarized as follows:
100
MARK
Seal-o( -O-
Seal-~ -{3-
80 T E S T CONDITIONS
Coolant:LLC-A
Coolant Conc.:5Ovol%
~z Shaft Speed:ZOOOrpm
Sys. Pressure:O.15MPa
d Liquid Temp.:121"C
40
o
<E
20 /
~ ~ U ---~
00" 2 4 6 8 10
Running Time, xlO2hours
FIG. 22--Time changeof leakage in durabilitytests.
Mechanisms of Deposit Formation
1. Deposition ofsilicate--Frictionalheat of mechanical seals accelerates the process from

a state of silicate gel into the solidified film as a result of the condensation reaction by
dehydration. The solidified film adheres and grows on the sealing surface side with higher
wetting free energy.
2. Deposition of P, Fe, and Cu--As a result of an increase in concentration of ferrous ions
caused by deterioration of the LLC, the self catalytic reaction promotes ferrous deposi-
tion on rubbing surfaces in the presence of phosphate, and, subsequently, a thin layer of
Cu is formed on the surface of the ferrous deposit.
3. Deposition of D M - - T h e precipitation of DM occurs as a result ofdissolution from rub-
ber radiator hoses. The DM precipitate is concentrated by frictional heat and adheres to
the sealing surfaces.
Countermeasuresfor the Deposition
1. Rubbing materials--Selection of combination of rubbing materials having lower wet-

ting free energy.
2. Surface roughness--Finishing in an appropriate range of mating ring surface roughness
in relation to wetting free energy of the seal ring surface.
3. Seal construction--Improvement of the seal construction to increase transfer of fric-
tional heat.
Acknowledgment
The authors wish to express their sincere appreciation to Dr. F. Hirano, professor emeritus
of Kyushu University, for his valuable advice during this investigation, and to thank the direc-
tors of Eagle Industry Co., Ltd., for permission to present this paper.
References
[1] Matsushima A., "Guide to Automotive Water Pump Seals," SAE 780404, Societyof Automotive
Engineers, Warrendale, PA, 1978.
[2] Hirabayashi, H., Oka, M., and Yukimasa, T., "Trouble and Countermeasure of Seals for Automo-
tive Application--Part II," SAE 800530, Congress and Exposition, Detroit, MI, 1980.
[3] Kiryu, K., Fukahori, K., Yanai, T., Matsumoto, S., Shimomura, T., and Hirabayashi, H., "A Status
of Sealing Performance of End Face Type Seals for Water Pump of Automotive Engines in Japan,"
SAE 880303, InternationalCongress and Exposition, Detroit, MI, 1988.
881270, International Off-Highway and Powerplant Congress and Exposition, Milwaukee, WI,
1988.
tem," SAE 852327, Truck and Bus Meeting Exposition, Chicago, IL, 1985.
[6] Kinisky, T. G. and McKenzie, D. A., "Investigation of Premature Water Pump Seal Failures," SAE
900806, Society of Automotive Engineers, Warrendale, PA, 1990.
[ 7] Komiya, M., Kiryu, K., Masuya, H., Okada, K., and Hirabayashi, H., "Effect of Silicate in Coolant
on Sealing Performance of Automotive Engines," Japan Society of Lubrication Engineers, Proceed-
ings of the Japan International Tribotogy Conference, Nagoya, 1990, pp. 101-106.
[8] Kiryu, K., Tsuchiya, K., Yonehara, Y., Shimomura, T., and Koga, T., "An Investigation of Deposits
Formation on Sealing Surfaces of Water Pump End Face Seals," Society of Tribologists and Lubri-
cation Engineers, Lubrication Engineering, Vol. 45, No. 1, 1989, pp. 49-55.
[9] Kiryu, K., Tsuchiya, K., Shimomura, T., Yanai, T., Okada, K., and Hirabayashi, H., "The Effect of
Engines," SAE 890609, InternationalCongress and Exposition, Detroit, MI, 1988.
[10] Hirabayashi, H., KJryu, K., Okada, K., Yoshino, A., and Koga, T., "Electro-Chemical Investigation
of Deposit Formation on Mechanical Seal Surfaces for Diesel Engine Coolant Pumps,'" Automotive
Lubrication, 7th International Colloquium, Technishe Akademie Esslingen, Osttildem, 1990, pp.
14.4-1 to 14.4-6.
[11] Ohkawa, S., Kawasaki, T., and Kumagae, K., "A New Antifreeze Coolant for Heavy-Duty Diesel
Engines," SAE 900433, Society of Automotive Engineers, Warrendale, PA, 1990.
[12] Kikuchi, M. and Tajima, H., "Development of Coolant for Heavy Duty Vehicle," Japanese Society
of Automotive Engineers, Vol. 42, No. 4, 1988, pp. 471-477.
[13] Sakka, S. and Kamiya, K., Journal of Non-Crystalline Solids, Vol. 48, No. 31, 1982, pp. 1-217.
[14] Thompson, P. F. and Lorking, K. F., "'Some Aspects of the Corrosion Processes of Iron, Copper and
Aluminum in Ethylene Glycol Coolant Fluids," Corrosion, Vol. 13, No. 8, 53 lt, 1957, pp. 59-63.
[15] Yoshikawa, K., Maeyama, A., and Izumi, K., "An Analysis of Leakage of Coolant Water in Auto-
motive Water Pump," Unisia Technical Report, No. 3, 1989, pp. 94-102.
[16] Bikerman, J. J., PhysicalSurfaces, Academic Press, 1970, pp. 239-246.
Stephen M. W o o d w a r d 1 a n d Aleksei V. Gershun ~
Characterization of Used Engine Coolant by

Statistical Analysis
REFERENCE: Woodward, S. M. and Gershun, A. V., "Characterization of Used Engine Cool-
ant by Statistical Analysis," Engine Coolant Testing: Third Volume, ASTM STP 1192, R. E.
Beal, Ed., American Society for Testing and Materials, Philadelphia, 1993, pp. 234-247.
ABSTRACT: Since the 1950s, the most common choice of engine coolant has been a 50:50 mix-
ture of ethylene glycol and water, with the ethylene glycol containing selected corrosion inhibi-
tors and other additives such as nitrates, silicates, borates, phosphates, tolyltriazole (TTZ), mer-
captobenzothiazole (MBT), antifoam, molybdates, silicones, dye, surfactant, and alkalinity
builder. Manufacturers in Europe and Asia use a different corrosion inhibition technology than
those used in North America.
The function of the engine coolant is simple, to remove the heat from the engine and to protect
the engine from corrosion and pitting of the metal surfaces. As the coolant accumulates miles or
hours in a vehicle cooling system, it develops many different types of problems. With time, the
inhibitors deplete, the ethylene glycol degrades, the pH of the coolant solution decreases, and
corrosion starts to take place. Silicates can combine with excessive levels of magnesium or cal-
cium, thus losing their ability to protect aluminum; MBT is easily oxidized and results in loss of
copper protection; and phosphates can combine with calcium and magnesium to form an insol-
uble sediment.
A consequence of this chemistry is that the coolant that comes out of an automobile after time
and use can be significantly different than what went into it. With the increased emphasis placed
on proper waste disposal and recycling of used coolant, it is important to understand what is
coming out of the car.
During 1989 and 1990, a statistically valid survey and characterization of used engine coolant
was taken from commercial automotive service centers. The purpose of the survey was to char-
acterize the condition of a typical used engine coolant. The samples were evaluated for sus-
pended matter, and analyzed for pH, reserve alkalinity (RA), percent ethylene glycol (EG),
diethylene glycol (DEG), propylene glycol (PG); the presence of degradation acids (acetate, gly-
colate, formate, and ethylenediaminetetraacetate [EDTA]); the level of inhibitors (nitrites,
nitrates, TTZ, MBT, benzotriazole [BZT], benzoate), oil, chloride, fluoride, sulfate concentra-
tions, and 14 other elements (A1, B, Ca, Cu, Fe, K, Mg, Mo, Na, P, Pb, Si, Sn, Zn) in their soluble
and insoluble forms. The samples were also tested following U.S. Environmental Protection
Agency (EPA) guidelines to determine if the engine coolant would be considered a hazardous
waste.
This paper presents the findings of this survey and shows the variation in the condition of
engine coolant taken from automobiles at the time the average consumer decides to change the
fluid.
KEYWORDS: automotive, antifreeze, engine coolant, characterization, used antifreeze
In the early 1980s, the H o m e and A u t o m o t i v e Products Division o f U n i o n Carbide, which

later b e c a m e First Brands Corporation, c o n d u c t e d a survey entitled " A One T h o u s a n d Car
Assessment of the U. S. Car Population Cooling System" [1]. T h e objective o f this survey was
Senior development chemist and project scientist, respectively, Coolant Group, First Brands Corpo-
ration, 55 Federal Rd., Danbury, CT 06108-4001.
234
WOODWARD AND GERSHUN ON CHARACTERIZATION OF USED ENGINE COOLANT 235
to determine the quality of engine coolant in cars currently on the road and to compare that
with the recommended procedures as published by the major U. S. auto manufacturers. The
results showed that there was a "'general state of neglect of the engine cooling system and the
lack of proper attention being paid to the service recommendations of the car manufacturers"
[1]. This survey is different in that the evaluation is made at the end of the coolant's life. In
addition, the samples taken were analyzed for their basic chemical components using current
techniques.
As chemical costs and disposal charges escalate, used coolant is slowly becoming a com-
modity as a resource for raw materials such as ethylene glycol (EG) and recycled engine cool-
ant. In addition to the obvious environmental benefits, the magnitude of the potential source
of recycled EG dictates that new technology be developed at a reasonable cost to take advan-
tage of this opportunity.
As a result, it became necessary to identify the ingredients of the coolant at the end of its
useful life defined by the consumer. This information will help develop a database to address
other important questions: What is the legal classification of used coolant? How do analytical
test methods impact hazardous classification? What is the environmental impact of used cool-
ant disposal? What role does variation in used coolant composition have on efforts to recycle
it? All of this leads back to the need for a detailed survey of the composition of typical used
coolant.
First, we must address the question: "What is the definition of a used coolant?" For our
purposes, used coolant is a material that has been functioning in the cooling system of an inter-
nal combustion engine and now has been removed because of a perceived lack of performance.
This survey, entitled "Characterization of Used Engine Coolant by Statistical Analysis,"
includes a variety of automobile makes and model years, and addresses population variation
as well as averages for used coolant.
Survey Procedure
This survey deals with coolant that has been removed from vehicles during routine main-
tenance service. The owner has decided to replace the coolant in the car with a fresh fill, as a
result of a scheduled mileage time table, a visual inspection by the vehicle owner, or an inspec-
tion by a service station attendant during a routine oil change. The engine coolant was
removed and replaced by a service technician. This survey, therefore, does not necessarily rep-
resent do-it-yourselfers' changing habits. This survey is confined to used coolant from auto-
mobiles and light-duty trucks and does not address heavy-duty vehicles.
Interviews were conducted at several of the service centers while the coolant was changed.
Figure 1 shows the breakdown of vehicle types that came in for an engine coolant change as a
function of manufacturers. For comparison, Fig. 1 presents the percentage of the vehicles serv-
iced against the ratio of vehicles sold in the U. S. during the last 6 years [2,3]. The survey test
sample corresponds to the cross section of cars on the road. The average coolant change inter-
val is shown in Fig. 2. It occurred every 2 to 3 years, with the most c o m m o n period being every
other year. It is interesting to note that the coolant age of at least 41.3% o f the vehicles sampled
was 3 years and older.
The coolant was collected over an 18-month period between April 1989 and September
1990. The samples were tested at the time of receipt.
The samples were collected from service stations along the eastern section of the country,
ranging from New England to Georgia, from ten different sources (Table 1). The coolant was
collected and stored as individual vehicle samples, as a combined sample of 15 to 25 vehicles
(in 55-gal drums), or in a bulk truck (5000 gal). The breakdown indicates a solid cross-sectional
FIG. 1--Population of samples by manufacturer share.
FIG. 2--Population of samples by coolant age.

TABLE l--Used engine coolant survey.

Area # Vehicles # Samples Sampled As
New England 2000 1 5000-Gal truck
75 5 55-gal drum
25 25 Individual car
2100
New York 135 9 55-gal drum
33 33 Individual car
t68
Georgia 240 16 55-gal drum
TOTAL 2508 89 6757 Gal
NOTE: 2 gal = 3.8 L.
base to allow generation of variance data, as well as a large sample size, which adds significance
to the average data. The total sample includes the results of over 2500 cars.
Analytical Techniques
It is important to understand the techniques involved in quantifying used coolant. As seen
on Table 2, gas chromatography was used to identify the glycols, high performance liquid chro-
matography was used to identify the organics, ion chromatography was used to identify the
inorganic anions, and standard laboratory techniques were used for pH (ASTM Test Method
for pH of Engine Antifreezes, Antirusts, and Coolants [D- 1287]) and reserve alkalinity (ASTM
Test Method for Reserve Alkalinity of Engine Coolants and Antirusts [D 1121 ]). One of the
more critical tests was the elemental analysis completed by inductively coupled plasma emis-
sion spectroscopy (ICP). This is because it measures the concentration of heavy metals, which
determines ifa material should be classified as a hazardous waste. Recognizing the importance
and sensitivity of this method, special care was taken to avoid errors caused from interferences
by this complex matrix or by different coolant sample viscosities (caused by glycol concentra-
tion differences between the sample and the standards).
The EPA has set strict guidelines for characterization of a material as a hazardous waste
[4]. The guidelines include a list of hazardous materials and upper limits for each listed mate-
rial. In addition, sample preparation plays a key role in the characterization of the potential
hazard level. The EPA has defined sample preparation methods. Table 3 shows the recent his-
tory of EPA test methods for heavy metals such as lead (Pb).
In 1989, one of the definitions for characterizing waste was called the extraction procedure
toxicity test (EP Tox) [5]. For used engine coolant, the primary concern is the presence of
TABLE 2--Lab methods.

Analysis Required Method Used
Physical properties pH Meter, Karl Fisher

Oil Babcock bottle method
Ethylene glycol, diethylene glycol, propylene glycol Gas chromatography
Nitrite, nitrate, chloride, fluoride, sulfate, benzoate Ion chromatography
TTZ, BZT, MBT, degradation Liquid chromatography
Elemental analysis--A1, B, Ca, Cu, Fe, K, Mg, Mo, Na, Inductively coupled plasma emission
P, Pb, Si, Sn, Zn spectroscopy
TABLE 3--EPA test methods, lead contamination.
Contaminent EP Toxicity TCLP
Effective date 1986-Sept. 1990 Oct. 1990-?

Filtration size 0.45 um 0.6-0.8 um
Maximum lead allowed 5 ppm 5 ppm
EP Toxicity = extraction procedure toxicity test; TCLP = tox-
icity characteristic leaching procedure.
heavy metals, specifically Pb. The source of Pb in used coolant is from the solder used in cop-
per/brass radiators and heater cores that have corroded. Sample preparation consisted of fil-
tering the sample through a 0.45-um filter and then conducting an analysis on the filtrate. The
solids were extracted for analysis only if the solids exceeded 0.5% by weight of the original
coolant. This served as the definition for soluble versus insoluble material.
In September 1990, the EPA replaced the EP Tox test with the toxicity characteristic leach-
ing procedure (TCLP) [4]. The major difference is that the filter has been changed from
0.45 um to a filter which has 0.6 to 0.8-um pore size. The net effect of this change is that the
soluble fraction has increased. TCLP was used to define solubles versus insolubles for this
presentation.
Analysis of Components
The analysis o f the used coolant samples has been broken into several groups as shown on
Table 4. These groups include properties, residual inhibitors, corrosive contaminants, EG deg-
radation products, corrosion products, and gravimetric analysis. Gravimetric analysis covers
the characterization o f the insoluble or suspended solids.
Properties
Table 5 presents the averages and standard deviation for pH, reserve alkalinity (RA), and
the percent glycols, including EG, diethylene glycol (DEG), and propylene glycol (PG). The
pH and RA are measures of the buffering capacity of the coolant, which maintains a stable
inhibitor package. The glycols represent the freeze point protection. EG is the primary agent.
DEG, a byproduct of the EG manufacturing process (with less freeze point depression char-
acteristics), is added to extend the EG base. PG has also been used as an extender for EG during
EG shortages and as a direct replacement for EG, although PG is not as effective for freeze
point as EG [6]. The RA shown here is defined as the number of milliliters of one tenth normal
hydrochloric acid required to neutralize 10 ml of the engine coolant. Based on a 50% solution
TABLE 4--Used engine coolant characterization.
9 Properties
9 Inhibitors
9 Corrosive contaminants
9 Degradation products
9 Corrosion products
9 Gravimetric analysis
TABLE 5--Typical used engine coolant, properties.
Standard
Properties Average Deviation
pH 8.48 0.62
Reserve alkalinity 6.10 2.51
Ethylene glycol, wt% 53.72 11.63
Diethylene glycol, wt% 2.76 1.07
Propylene glycol, wt% 0.37 1.02
o f several leading national brands, the RA would be between 6.4 and 7.4 [I ]. For the purposes
o f this paper, we considered a 50% depletion as one definition of an ineffective coolant. Figure
3, which is a summary of RA data, shows that over 13% of the samples tested fall into this
category. Figure 4 represents the percent EG of the samples. On average, the percent glycol
was 53.7% by volume, which is acceptable. However, the range is wide with values as low as
11.5% by volume to as high as 76.6% by volume EG. There was very little evidence o f PG.
Residual Inhibitors
The most commonly used inhibitors are listed in Table 6. The inhibitors listed include those
found in vehicles in this country from North American, European, and Asian formulations.
As you can see, the variation (defined as plus or minus two standard deviations) of each mate-
rial measured is at least equal to the averages. You could predict with 95% confidence that the
FIG. 3--Reserve alkalinity distribution.

FIG. 4--Percentage ethylene glycol distribution.
value of any inhibitor could be between 0 and twice the average listed here. Anyone skilled in
the art recognizes that an effective coolant formulation (either original or recycled) requires
the correct inhibitors, the correct a m o u n t of those inhibitors, and the correct balance of those
inhibitors. The advantages and disadvantages of each inhibitor must be understood and bal-
anced. Let's take a look at some examples of the variation o f some commonly used coolant
additives.
Figure 5 shows the distribution found for nitrate (NO3). It is an inhibitor commonly used
TABLE 6--Typical used engine coolant, inhibitors (sohtble).

Standard
Inhibitors Average Deviation
Nitrite 77.99 214.59

Nitrate 981.94 503.96
Tolyltriazole 395.87 238.94
Benzotriazole 91.29 51.61
Mercaptobenzothiazole 14 86.26
Benzoate 2347.57 3890.95
Borate (as B) 403.04 271.16
Molybdate (as Mo) 44.19 115.22
Phosphate (as P) 656.51 550.98
Silicate (as Si) 48.98 25.94
Potassium 1180.6 1349.94
Sodium 2706.19 1028.4
FIG. 5--Nitrate distribution.
to prevent pitting corrosion on aluminum (A1) [ 7]. The histogram shows a bimodal distribu-
tion with one center around 800 to 900 ppm and the second smaller one around 2000 ppm.
Figure 6 shows the distribution for boron (B). Boron is commonly used in the form of sodium
borate, borax, or boric acid. It buffers the solution at a desired pH and prevents the pH from
dropping to a range where the corrosion of cast iron and steel is unacceptably high [ 7]. There
are wide ranges and non-normal distributions for these various inhibitors found in engine
coolant.
Corrosive Contaminants
Table 7 lists some of the corrosive contaminants found in a coolant system. Chloride, sul-
fate, and fluoride can be introduced into the cooling system by way of the water used to adjust
the freeze point, during cleaning, or during conditioning of the cooling system. Another poten-
tial source of sulfate is the degradation of mercaptobenzothiazole (MBT; if present in signifi-
cant quantities). Because these contaminants are common, the performance test methods for
evaluating engine coolant outlined in ASTM Specification for Ethylene Glycol Base Engine
Coolant for Automobile and Light Duty Service (D 3306) define a corrosive water, which con-
tains set amounts of these anions to dilute the concentrated coolant samples to the proper
concentration before testing. Corrosive water is defined by ASTM as deionized water contain-
ing 100 ppm each of sulfate, chloride, and bicarbonate introduced as sodium salts.
The chloride level found in these vehicles is similar to the level that would be achieved with
corrosive water as used in ASTM D 3306. However, the sulfate level is significantly higher than
in the standard. These water quality levels were consistent in all the samples taken.
FIG. 6--Boron (soluble) distribution.
Degradation Products
Table 8 lists some of the degradation products found in used coolants. This list includes the
soluble salt forms of the carboxylic acids generated by the thermal decomposition of ethylene
glycol. The formation of acetates, glycolates, and formates depletes the reserve alkalinity when
the acids are formed. The level of glycol degradation products varied from sample to sample;
however, the ratio of glycolate to formate to acetate remained relatively stable at 29:11:6,
respectively. They all are soluble at this level in 50% glycol/water mixtures.
Also included are the insoluble or precipitated corrosion and deposition products of the
interaction ofinhibitors with the coolant system and with calcium (Ca) and magnesium (Mg)
found in the dilution water. The concern with insoluble inhibitor products is that, at certain
levels, they could have two negative results. The first is that they reduce the desired effective-
ness of the inhibitors, which can create a system imbalance promoting corrosion and can act
as a physical catalyst for further precipitation. The second result is that these insolubles may
reduce the efficiency of cooling system components (that is, reduce or restrict the flow through
the heater core and radiator, and so forth).
TABLE 7--Typical used engine coolant, corrosive contaminants,

ppm.
Standard
Contaminents Average Deviation
Chloride 54.40 74.86
Sulfate 269.45 199.30
Fluoride 146.00 101.00
TABLE 8--Typical used engipw coolant, degradation products,

ppm.
Standard
Products Average Deviation
Acetate 147.55 77.50

Glycolate 722.17 918.1
Formate 278.37 155.6
Insoluble
B 29.07 12.4
Ca 9.66 58.35
K 19.91 50.61
Mg 4.35 5.69
Mo 1.04 1.58
Na 76.64 35.58
P 28.28 28.64
Si 19.94 94.03
Corrosion Products
Table 9 lists the soluble and insoluble corrosion products found in the used coolants tested.
These materials represent the metals commonly found in cooling systems. As discussed earlier,
the soluble data are identified by measuring the ion level of a coolant that has been filtered
through a 0.7-um filter, and the insoluble measurement was made on the material trapped on
the filter, which is then acid digested [4]. There are two implications from the generation of
soluble and insoluble metals suspended in the coolant. The first concern is that, as the con-
centration of suspended solids increases, this can cause abrasion of surfaces and dramatic-
ally increase the chance of water pump failures, radiator clogging, and insulating barriers.
The second concern is based on the regulatory impact of dissolved heavy metals in used
coolants.
The most c o m m o n way that a used coolant could be classified as a hazardous waste by the
federal government is if it is characterized as containing heavy metals such as Pb. Used engine
TABLE 9--Typical used engine coolant, corrosion products,

pprn.
Standard
Products Average Deviation
Soluble
Aluminum 2.68 2.22
Copper 0.69 7.27
Iron 2.73 17.52
Zinc 4.65 20.34
Lead 2.85 3.6
Insoluble
Aluminum 10.71 15.15
Copper 1.98 7.91
Iron 43.53 29.53
Zinc 6.13 8.58
Lead 15.49 23.78
coolant is not listed by the federal government as a hazardous waste. We have conducted stud-
ies on data from 55-gal drum quantities of used coolant including EP Tox and TCLP methods
[4,5]. Each d r u m typically contained coolant from 15 to 25 different vehicles. Using the sol-
uble portion of the coolant, it is predicted that 20% of used engine coolant will have Pb in
excess of 5 ppm. The impact of this issue on engine coolant recycling has not been fully
addressed, but it is clear that all disposal options are important and that recycling efforts should
not be discouraged.
Gravimetric Analysis
The final means that we have used to characterize used coolant is gravimetric analysis. The
functionality of a fluid can be identified by concentrating on the suspended or insoluble frac-
tion of a used coolant. Usually, a visual inspection of a sample allows conclusions to be drawn
based on its color and turbidity. Determining the total suspended solids analytically provides
a better determination of the deterioration of a coolant's effectiveness. After the solids are col-
lected on a 0.7-#m filter, the filter is dried at 180"C until the weight loss stabilizes. Figure 7
shows that there is a significant range of values in the samples tested. We believe that a high
level of suspended solids indicates a failed coolant. Over 25% of the samples tested contained
more than 500 p p m suspended solids.
The suspended solids were further characterized by particle size (Table 10). The objective
was to determine the theoretical effectiveness of different filtration techniques on the removal
of particulate matter. It was found that over 10% of the suspended matter was less than 5 um
in diameter.
FIG. 7--Total suspended solids.

TABLE 10--Typical used engine coolant, partMe size

distribution.
Particle Size, um Percent
< 0.1 0.1

0.1-.45 0.3
0.45-.7 0.4
0.7-1.0 0.8
1.0-2.5 4.0
2.5-5.0 7.0
5.0-15.0 9.0
15.0-25.0 13.0
25.0-50.0 15.4
50.0-100.0 13.0
> 100.0 37.0
Summary and Conclusions

This 1989/1990 survey has shown that in many cases, the point at which the consumer
decides to change engine coolant is well beyond its useful life. Over 41% of consumers have
the same coolant for more than 2 years in vehicles. The inhibitors have been significantly
depleted. Corrosion is evident to some degree in a majority of vehicles. In addition to the dam-
age to the cooling system, the lack of good maintenance has generated soluble heavy metals
such as Pb, which represents a hazardous waste situation. The survey also shows the contin-
uing problem with consumers' lack of appreciation for having the recommended engine cool-
ant concentration, as pointed out in the earlier survey [1]. This not only increases the risk of
damage from boilovers and freezeups, but it also accelerates the rate of corrosion due to lower
inhibitor levels.
This survey indicates a significant variation among coolants at the perceived end of their
functionality. Treatment of used coolant for the purposes of recycling either on-site or off-site
presents many challenges. The variety of engine coolant formulations on the market as well
as the variation of coolant condition requires careful consideration when evaluating different
recycling techniques.
The engine coolant industry must be environmentally conscious and proactive in address-
ing the issues of used coolant, but, at the same time, they must preserve the integrity and per-
formance standards of engine coolant for automotive and heavy-duty cooling systems.
References
Car Population Cooling Systems,"; SAE Report 831821, Society of Automotive Engineers, Warren-
dale, PA, 5 Dec. 1983.
[2] Ward'sAutomotive Yearbook 1990, Ward's Communications, Detroit, 1990.
[3] Ward's Automotive Yearbook 1987, Ward's Communications, Detroit, 1987.
[4] "Hazardous Waste Management System: Identification and Listing of Hazardous Waste, Environ-
mental Protection Agency," 40 CFR Part 261, Federal Register. Vol. 55, No. 61, March 29, 1990.
[5] "Hazardous Waste Management System: Identification and Listing of Hazardous Waste; Environ-
mental Protection Agency," 40 CFR Part 261, Test Methods for Evaluating Solid Waste, SW-846,
U.S. Environmental Protection Agency, National Technical Information Service, Springfield, VA,
3rd ed., Nov. 1986, pp. 1310-1311.
[6] Curme, G. O., Jr, Ed., "Glycols," American Chemical Society Monograph Series, Reinhold Publish-
ing Co., New York, 1952.
[7] Chiasson, B. A., "An Overview of Antifreeze Chemistry," internal First Brands Corp. memo, May
10, 1988.
STP1192-EB/May 1993
246 ENGINECOOLANTTESTING:THIRD VOLUME
DISCUSSION
P. Dittma? (written discussion)--My questions relate to the issue of proper disposal under
RCRA:
1. What conclusions has First Brands reached regarding whether used antifreeze should
generally be considered TCLP-Hazardous?
2. Of the samples analyzed, how many were taken directly from vehicles? Is there any sta-
tistically significant difference between the sample sets taken from vehicle versus others,
perhaps revealing the effects of cross-contamination with other shop wastes?
3. What fraction of samples showed TCLP soluble lead in excess o f the EPA 5-ppm
threshold?
4. Do the lead data follow a normal definition? If so, what is the upper limit of the data as
defined in EPA Method SW846?
5. To what extent was perchloroethylene observed and to what detectability limit?
Stephen M. Woodward (author's response)-- 1. First Brands has not made any conclusions
about the classification o f used antifreeze as an EPA Hazardous waste. We do recommend that
if any one or more of the following apply, then the material should be treated as if it were
hazardous:
(a) State and/or local regulations require it.

(b) A representative sample o f the material on hand has been analyzed and has been found
to contain lead above the standard.
(c) A history o f representative samples shows that lead levels have exceeded the standards
for that specific location on a regular basis.
(d) No sampling has been done and no recognized form of recycling is being used that
would remove the hazardous material or deactivate it.
2. Fifty-eight of the coolant samples were taken directly from the vehicles. The only cross-
contaminant that was identified in our samples was oil. There was only a slight difference
found, but no test for significance was conducted.
3. We did not conduct TCLP analysis on every sample, however, based on our data, it is
predicted that at least 80% of the vehicles will not exceed 5-ppm soluble lead (after filtration
through a 0.7-t~m paper).
4. The distribution of data for lead was not a normal distribution because the lower limit is
bound by zero. At this time, the upper statistical limit based on EPA Method SW846 is not
known.
5. Perchloroethylene was not part of the initial list of materials to analyze for because it is
not in any known antifreeze formulation and it is not part of any OEM cooling system prep-
aration. As a result o f a Safety-Kleen letter stating that it was used by OEMs in the preparation
of radiators, several samples of used coolant and several samples from vehicles with less than
50 miles (80 km) on the odometer were tested and perchloroethylene was not detected in any
samples by our methods (2-ppm limit of detection).
J. A. Lima 2 (written discussion)--One o f your specimens was from a system that had not
been changed in 12 years, if I understand correctly. I assume that it was a standard North
Safety Kleen Corp., 1255 W. Old Higgins Rd., Elk Grove Village, IL 60007.
2 Houghton Chemical Corp., Allston, MA.
Copyright@1993 by ASTM International www.astm.org

DISCUSSION ON CHARACTERIZATION OF USED ENGINE COOLANT 247
American coolant, and am wondering what you think the practical life of a North American
coolant might be, if proper charging procedures and de-ionized water were used in the initial
fill?
Stephen M. Woodward (author's response)--The intent of this survey was not to determine
the theoretical life of a coolant that was under ideal conditions, but rather, the objective was
to characterize the quality of the used coolant under "real" conditions. The 12-year-old cool-
ant that was collected was mixed in a drum at the service station and could not be segregated
and analyzed.
Richard D. H e r c a m p ~a n d Robert A. R e m i a s z 2
Coolant Maintenance and Extension of

Coolant Life for Light Duty Vehicles
REFERENCE: Hercamp, R. D. and Remiasz, R. A., "Coolant Maintenance and Extension of
Coolant Life for Light Duty Vehicles," Engine Coolant Testing: Third Volume, ASTM STP
1192, R. E. Beal, Ed., American Society for Testing and Materials, Philadelphia, 1993, pp. 248-
257.
ABSTRACT: For many years coolant has been maintained and coolant life extended for heavy
duty vehicles by using periodic additions of supplemental coolant additives (SCAs) and, in many
cases, use of on-board by-pass filtration of the coolant. This concept has now been applied to
light duty vehicles using a system consisting of coolant testing, offboard full flow coolant filtra-
tion, and SCA addition to the coolant. In this manner, coolant life in light duty vehicles may be
safely extended.
The system used to maintain and extend the life of the coolant is described. Glassware, bench,
and simulated service testing data are reported, which support this approach.
KEYWORDS: engines, light duty vehicles, engine coolant, life extension, recycling, additives,
filtration, corrosion, filling and flushing, maintenance
Several years ago the combination of an ethylene glycol shortage and increasingly stringent
environmental laws caused vehicle service organizations to investigate methods of conserving
automotive light duty antifreeze, The ethylene glycol shortage was short-lived, but disposal
regulations continued to become tighter. The interest in reducing the volume of coolant dis-
posal is expected to continue.
The current issue concerns itself with disposal of light duty (LD) antifreeze. The Environ-
mental Protection Agency's (EPA) guidelines for reducing waste disposal volume consists of
a hierarchy of approaches [1]. In descending order of desirability, they are as follows:
1. Design the product to last longer.

2. Extend the useful life of the product.
3. Recycle the product.
4. Properly dispose of the product.
Consistent with the EPA's primary approach, original equipment manufacturers (OEM)
have become more interested in longer life LD coolants. Work has been underway at several
laboratories. Texaco, for example, has reported on a portion of their work [2]. Additionally
A S T M D 15 on Engine Coolants has formed a task force to start work to develop a specifica-
tion for a long-life coolant for light duty engines.
Consistent with the EPA's secondary approach, heavy duty (HD) engine OEMs have used
periodic addition of supplemental coolant additives (SCAs) and on-board coolant filtration
l Consultant, Daubert Industries, 55E 1100S, Columbus, IN 47201.

2 ECP Inc., Daubert Industries, 4700 S. Central Ave., Chicago, IL 60638.
248
HERCAMP AND REMIASZ ON COOLANT MAINTENANCE 249
for over 30 years [3]. The lives of l i D coolants are routinely extended to 290 000 km (180 000
miles) or more [4].
For the last ten years, the U.S. military has successfully used their MIL-A-53009 SCA to
extend coolant life in LD vehicles [5]. More recently, several commercial ventures have
offered LD SCAs for coolant life extension. This paper offers data developed to support one
of these programs.
Coolant or ethylene glycol recycling (EPA's tertiary approach) is often confused with cool-
ant extension. In March 1989, the National Automotive Radiator Service Association spon-
sored an all day Forum on Coolant Recycling and Disposal. Shortly thereafter ASTM formed
a Subcommittee on Recycled Coolant (D 15.15). As would be suggested by the titles, both
groups focused on the recycling o f spent coolant.
It is worth noting that it is not the intention of the authors to suggest that any coolant, regard-
less of its condition, is extendable. A coolant that has, or is about to fail, is not considered
extendable. It should be recycled or disposed of.
A comparison of l i D and LD systems is provided in Table 1.
A System to Maintain and Extend the Useful Life of Coolants for Light Duty Vehicles
A system has been developed to effectively extend the useful life of L D coolants. This system
includes qualification o f the coolant by chemical, physical, and visual means; and extension
o f the coolant life by filtering and the addition o f an SCA. If the coolant does not qualify, it is
removed for recycling or disposal.
The following quick checks are made to determine whether or not the coolant is extendable.
Since the concept of the system is to extend the coolant life and keep it in good condition, it
is vital to perform the maintenance at about the same mileage that light duty vehicle manu-
facturers have recommended that the coolant be replaced with new coolant, generally about
50 000 to 65 000 km. Therefore, the first criteria to determine if the coolant is maintainable
is that the coolant shall not have more than 65 000 km of use since new, or since its last main-
tenance with this system.
Second, if the coolant is oily, murky, or rusty, it indicates that the coolant has been contam-
inated by foreign material or that excessive corrosion has already occurred. In either case the
coolant must be removed, the cooling system cleaned thoroughly, and repairs made to correct
any cooling system problems.
Third, if the glycol level is less than 25% or the reserve alkalinity (RA) less than 3 (for domes-
tic vehicles), it can be concluded that corrosion, scaling, and so forth, has already occurred to
the extent that the coolant should be removed, the cooling system cleaned, and the cause for
the dilution or loss of RA determined and corrected. The glycol level and RA level can be
TABLE 1--A comparison of liD and LD cooling systems.

HD LD
OEM recommends draining and flushing at OEM recommends draining the system at
300 000 km 50 000 to 65 000 km
Typically 90% of the coolant can be drained Typically 50 to 75% of [6] the coolant can be
drained
Typical capacity of cooling system is 32 to 56 L Typical capacity of cooling system is 8 to 14 L
Routinely extended to 290 000 km or more, with Extendable with proper maintenance and
maintenance and SCA addition SCA addition
Life of vehicle typically 1 600 000 kin, or more Life of vehicle typically 200 000 km or more
FIG. 1--Coolant filtration device.
conveniently determined using a 2 pad test strip [ 7]. If these 3 checks are positive, the coolant
is deemed extendable and is maintained. If not, the coolant is removed for recycling or
disposal.
The device described in the Appendix is attached to circulate and filter the coolant (Fig. 1).
(The benefits of filtration for H D engine coolants have been shown [3], and it can be assumed
that filtration is also beneficial for LD engine coolants.) SCA is then added to reinhibit and
extend the life of the coolant, and glycol or water is added to achieve a 50% glycol final mixture.
The SCA is also described in the Appendix.
Performance Testing
Since extension of LD engine coolant life is relatively new and no industry standards of
performance exist, it was decided to run modified performance tests listed in ASTM Specifi-
cation for Ethylene Glycol Base Engine Coolant for Automobile and Light Duty Services (D
3306). These tests were run on used coolant both as received and after treatment. The tests
were modified in that the same amount of corrosive agent added to new coolant concentrate,
for which the tests procedures are written, was added to the sample coolants. A somewhat dif-
ferent approach using simulated depleted coolant has been reported which was used to eval-
uate proposed SCA's for LD engine use [8]. Since the SCA we use had already been developed
with chemistry similar to the additive specified for the GM 6043 [9] coolant concentrate (anti-
freeze), we chose to test used coolant from the field rather than simulated depleted coolants.
A test using six domestic light duty vehicles with mileages exceeding the 65 000-km criteria
mentioned previously was designed to measure the performance in ASTM Test Method for
Corrosion Test for Engine Coolants in Glassware (D 1384) and ASTM Test Method for Cor-
rosion of Cast Aluminum Alloys in Engine Coolants Under Heat-Rejecting Conditions (D
4340) corrosion tests, both before and after coolant maintenance. The higher mileage coolants
were used to make the test more severe. Unfortunately, a mixup occurred, and one set of data
had to be discarded. The performance data for the other 5 are shown in Tables 2 through 6.
Coolant analysis results are shown in Table 7.
TABLE 2--1986 Mustang, 46 959 miles, pump replaced at 9916 miles, I gal antifreeze added, no
other maintenance, 2.3 L.
Results of ASTM D 1384 Glassware Testing on ECP 524122 D 4340
Average of 2
Weight Loss, mg/Sample TeNs
Coupon Type 1 2 3 Average Maximum Allowed* Maximum(l)
As-REcEIVED
Copper 2 2 2 2 10
Solder 3 3 5 4 30
Brass 5 5 2 4 10
Steel 0 3 0 1 10
Cast iron 120 106 115 "114 ^ 10
Cast aluminum 93 75 103 ^90 ^ 30 ^9.99 ^
AFTER TREATMENT
Copper 1 2 2 2 10
Solder 5 6 6 6 30
Brass 2 1 2 2 10
Steel 0 0 2 1 10
Cast iron 5 2 5 4 10
Cast aluminum 8 2 3 4 30 0.028
NOTES: *: ASTM Specification for Ethylene glycol Based Engine Coolants (D 3306). ^: Exceeds ASTM
D 3306 maximum weight loss. 1 mile = 1.609 km.
TABLE 3--1985 Celebrity, 68 281 miles, Car 425.
Average of 2
Weight, mg/Sample Teas
Coupon Type 1 2 3 Average Maximum Allowed* Maximum (l)
As-RECEIVED
Copper 2 2 2 2 10
Solder 6 7 7 7 30
Brass 2 3 2 2 10
Steel 0 2 1 1 10
Cast aluminum 11 6 5 7 30 ^16.13 ^
AFTER TREATMENT
Copper 2 2 2 2 10
Solder 4 4 4 4 30
Brass 2 2 2 2 10
Steel 1 2 1 1 10
Cast iron 5 4 t 3 10
NOTES: *: ASTM Specification for Ethylene Glycol Based Engine Coolants (D 3306). ^: Exceeds
ASTM D 3306 maximum weight loss. 1 mile = 1.609 km.
TABLE 4 - - 1 9 8 5 Celebrity, 57 699 miles, Cat" 426.
Results of ASTM D- 1384 Glassware Testing on ECP 524126 D 4340
Average of 2
Weight, Loss in mg/Sample Tests
Maximum
Coupon Type 1 2 3 Average Allowed* Maximum(1)
As-RECEIVED
Copper 4 4 3 4 10
Solder 7 l0 8 8 30
Brass 3 3 3 3 10
Steel 3 3 3 3 l0
Cast iron 6 11 9 9 l0
Cast aluminum l0 7 8 8 30 ^16.10 ~
AFTER TREATMENT
Copper 3 3 3 3 10
Solder 4 5 5 5 30
Brass 2 3 3 3 10
Steel 2 3 3 3 10
NOTES: *: ASTM Specification for Ethylene Glycol Based Engine Coolants (D 3306). ": Exceeds
TABLE 5--1986 Blazer, 61436 miles, 2.8 L V-6, no maintenance except water addition.
Average of 2
Weight Loss in mg/Sample Tests
Maximum
Coupon Type 1 2 3 Average Allowed* Maximum ( 1)
As-RECEIVED
Copper 3 3 2 3 10
Solder 7 7 7 7 30
Brass 2 3 2 2 10
Steel 3 4 4 4 10
Cast iron 61 79 78 ^73 ^ 10
Cast aluminum 11 11 7 10 30 ^12.48 ^
AFTER TREATMENT
Copper 2 2 2 2 10
Solder 4 5 4 4 30
Brass 2 2 2 2 10
Steel 5 5 5 5 10
Cast iron 11 13 9 ~ 11 ^ 10
Cast aluminum 4 4 5 4 30 "19.58 ^
NOTES: *" ASTM Specification for Ethylene Glycol Based Engine Coolants (D 3306). ^ ": Exceeds
ASTM maximum weight loss (D 3306). I mile = 1.609 km.
TABLE 6--1986 Chrysler Laser, 45 461 miles, 2.2 L Turbo, no known maintenance.
Average of 2
Weight Loss, mg/Sample Te~s
Maximum
Coupon Type 1 2 3 Average Allowed* Maximum (1)
As-REcEIVED
Copper 3 3 3 3 10
Solder 7 7 6 7 30
Brass 3 2 3 3 10
Steel 5 3 4 4 10
Cast iron 50 48 66 ^ 55 ^ 10
Cast aluminum 10 6 5 7 30 "6.12 ^
AFTER TREATMENT
Copper 2 2 2 2 l0
Solder 6 3 5 5 30
Brass 2 2 2 2 10
Steel 3 2 1 2 10
Cast iron 14 12 14 ^ 13 ^ 10
NOTES: *: ASTM Specification for Ethylene Glycol Based Engine Coolants (D 3306). "^: Exceeds
All five u n t r e a t e d c o o l a n t s failed D 4340 b y a wide m a r g i n , while 4 o f t h e 5 m a i n t a i n e d a n d

e x t e n d e d c o o l a n t s readily passed. O f t h e u n t r e a t e d coolants, 3 o f 5 failed D 1384 b y a wide
m a r g i n , while 3 o f t h e 5 m a i n t a i n e d a n d e x t e n d e d c o o l a n t s passed. T h e p e r f o r m a n c e o f t h e
r e m a i n i n g 2 m a i n t a i n e d a n d e x t e n d e d c o o l a n t s was substantially i m p r o v e d , b u t they did n o t
m e e t A S T M standards.
TABLE 7--Coolant analysis results.
1985 Celebrity
Parameter 1986 Mustang 1985 Celebrity #2 1986 Blazer 1986 Laser
Treated no yes no yes no yes no yes no yes

pH 9.3 9.3 10 9.8 9.8 10 9.4 9.4 8.2 8.8
Reserve 5.15 5.1 6.2 7.2 4.2 7.8 4.0 6.3 3.8 5.1
Percent 48% 49% 58% 53% 40% 52% 54% 52% 59% 53%
Freeze --30 -30 --56 --40 -- 12 --37 --44 -37 --60 --40
Gly./Formate 240 208 78 97 300 300 460 520 770 611
Chloride 70 75 74 200 ~ 80 60 80 84 48 50
NO2 0.0 0.0 280 260 350 350 0.0 0.0 0.0 0.0
NO3 1000 830 a 1000 1500 890 1800 580 1100 2700 2700
PO4 500 4150 2200 3400 1300 3500 960 2200 3300 4000
SO4 700 600 340 460 500 610 540 550 1400 1100
SiO3 20 620 200 700 290 1160 120 550 300 800
a These results also are not logical, based on the SCA composition. Nitrate should increase appreciably
while chloride should remain nearly constant.
b When checked by another lab, the phosphate showed an increase of 180% rather than the 17%
decrease shown here. The 17% decrease is not logical, based on the SCA composition.
TABLE 8--ASTM D 4340 test results.
Weight Loss,
Coolant mg/cm2/Week
NEW coolant 0.1

NEW coolant with additive 0.0
USED coolant 2.8
USED coolant with additive 0.06
In addition to the above tests, the effect of adding the SCA, without tiltrahon, was studied
in the following series of tests with new and used coolant: D 4340, ASTM Test Method for
Caviation Erosion--Corrosion Characteristics of Aluminum Pumps with Engine Coolants (D
2809), and a modified 1500 h ASTM Method for Simulated Service Corrosion Testing of
Engine Coolants (D 2570). Table 8 shows the improved performance with the additive in D
4340 tests. Table 9 shows the improved performance with the additive in D 2809 tests. Figures
2 and 3 show the improved performance with the additive in the modified 1500-h D 2570 test.
Taken as a whole, these data indicate that coolant, which meets all the criteria, including
the 65 000-km limit, will generally be maintainable so that its life can be extended.
Commercial Experience
After 4 years of commercial use of this system to extend the life of coolant in 115 000+
vehicles, warranty figures show that only 28 claims have been filed for cooling systems failures.
Conclusions
Both lab tests and field experience show that, with proper maintenance at 50 to 65 000-km
intervals, coolant life in light duty vehicles can be safely extended.
Proper maintenance includes coolant analysis, filtration, and addition of a well formulated
SCA.
Weight Loss (mg)

50
40
30
20
10
Copper 70/30 Solder Brass Steel

1
C a s t Iron Aluminum
I Without Additive [~ With Additive

* NO weight change
FIG. 2--A verage ,,eight loss of couponsfrom 1500-h simulated service tests (new coolant).
Weight L o s s (mg)
50--
40
3O
20
10
0
|
I I ] 2 L . . . . L
-10 _ _
Copper 7 0 / 3 0 Solder Brass Steel Cast Iron Aluminum
W i t h o u t Additive ~ With Additive

* No weight change
FIG. 3--A verage weight loss of couponsfrom 1500-h simulated service tests (used coolant).
APPENDIX
Description of the Circulation and Filtration Device and the SCA
The Circulation and Filtration Device

The device is described in detail in U.S. Patent No. 4,996,874 [10]. The coolant flushing is
performed by a diaphragm p u m p operating at 8 gal/min (30 L/min). The complete cooling
system is flushed in a closed-loop fashion by manipulating hose pliers. The push/pull action
of the p u m p creates less than 15-psi ( 103 kPa) pressure on system components. The system is
further protected by a liquid pressure relief valve. Simultaneous to flushing, the coolant is fil-
tered through a full-flow filter manifold. Circulating the coolant for 15 min removes 95% of
the particles larger than 5 gm.
In addition, the device contains a venturi p u m p that allows the user to vacuum test the cool-
ing system. The vacuum also allows for the elimination of air pockets when the system is
refilled with coolant by vacuum. This eliminates the need to run the engine and make several
coolant additions.
TABLE 9--ASTM D 2809.
Coolant Rating
New coolant 9
New coolant with additive 9
Used coolant 8
Used coolant with additive 9
Supplemental Coolant Additive

The supplemental additive uses chemistry similar to the G M 6043 antifreeze additive. It is
a colorless liquid with sufficient additive in a 12-oz (0.35-L) bottle to treat 12 qt (11.35 L) of
coolant in a typical light duty cooling system.
The additive uses both borate and phosphate to boost and restore reserve alkalinity and to
buffer pH in the preferred 8.5 to 10 range. The stabilized silicate is present in a sufficient
a m o u n t to protect all of the a l u m i n u m components in the system, but the a m o u n t used is not
so much as to cause gel formation [11 ]. Nitrate is included to insure against a l u m i n u m pitting
corrosion and the resulting failure of a l u m i n u m radiators [12]. Azoles are included to give
yellow metal corrosion protection [ 13]. Additionally, molybdate is included, which is an excel-
lent multimetal corrosion inhibitor [14]. Molybdate also shows synergistic effects with other
inhibitors for reducing corrosion of iron [15], a l u m i n u m [14,16] and yellow metals [17]. Fur-
thermore, molybdate depletion rates are low compared to those of many of the corrosion
inhibitors used in engine coolants. Its durability will provide corrosion protection after other
inhibitors have been depleted [18]. This is a very well balanced and complete package which
is designed for all the needs of LD engine coolants.
References
[1] Environmental Protection Agency, The Solid Waste Dilemma: Agenda for Action, EPA/530-SW-
88-052, Washington, DC, 1988.
[2] Darden, J. W., Triebel, C. A., Maes, J. P., and VanNeste, W., "Monoacid/Diacid Combination as
Corrosion Inhibitors in Antifreeze Formulation," Engine Coolants, Cooling System Materials and
Testing. STP811, SAE, 1990, pp. 135-152.
881270, 1988.
[4] Hudgens, R. D. and Hercamp, R. D., "Test Methods for the Development of Supplemental Addi-
tives for Heavy Duty Diesel Engine Coolants," Engine Coolant Testing: Second Symposium, ASTM
STP 887, R. E. Beal, Ed., American Society for Testing and Materials, Philadelphia, 1986, pp. 190-
191.
[5] Conley, J. H. and Jamison, R. G., "AdditivePackageforUsedAntifreeze,"EngineCoolant Testing:
Second Symposium, STP 887, R. E. Beal, Ed., American Society for Testing & Materials, Philadel-
phia, 1986, pp. 78-85.
[6] Filowitz, M., "Oral Presentation on The Wynn Oil Company Mark X System," Forum on Coolant
Disposal and Recycling, National Automotive Radiator Service Association, National Meeting,
Atlanta, 30 March 1989.
[ 7] Hemmes, et al., "Test Strips for Rapid On-Site Analysis of Engine Coolants," Engine Coolant Test-
ing. Third Volume, STP 1192, R. E. Beal, Ed., American Society for Testing and Materials, Phila-
delphia, 1993, pp. 165-179 (this publication).
[8] Richardson, R. C., "Test Methods for Evaluation of Supplemental Coolant Additives in Automobile
Engine Coolants," Worldwide Trends in Engine Coolant, Cooling System Materials and Testing, SP
811, SAE 900805, 1990, pp. 153-158.
[9] "Automotive Engine Coolant Concentrate--EthyleneGlycol Type, GM 6043-M," GeneralMotors
Engineering Standards, Sept. 1988, pp. J-41.401.
[ 10] Colomer, et al., "Method and Apparatus for Treating Coolant for Internal Combustion Engine,"
United States Patent 4,996,874, 5 March 1991.
tems," SAE 852327, 1985.
[12] Wiggle, R. R., et al., "The Effectivenessof Automotive Engine Coolant Inhibitors for Aluminum,"
Materials Performance, Vol. 20, No. 6, 1981, p. 18.
[13] Rowe, L. C., "Application of Inhibitors in Automobiles and Their Environment," Corrosion Inhib-
itors, C. C. Nathan, National Association of Corrosion Engineers, p. 179.
[14] Robitaille, D. R., "Molybdate Inhibitors for Problem Cooling Waters," Chemical Engineering, 4
Oct. 1982, p. 139.
[15] Weber, et al., Corrosion. Vol. 42, No. 9, p. 542.
[16] Vukasovich, M. S. and Sullivan, F. J., "Evaluation of Molybdate as an Inhibitor in Automotive
Engine Coolants," Materials Performance, National Association of Corrosion Engineers, Aug. 1983,
p. 29.
DISCUSSION ON COOLANT MAINTENANCE 257
[17] Robitaille, D. R., "Molybdate Inhibitors for Problem Cooling Waters," Chemical Engineering, 4
Oct. 1982, p. 140.
[18] Rowe, L. C., et al., "Antifreeze Coolant for Aluminum and Cast Iron Engines," Materials Perfor-
mance, National Association of Corrosion Engineers, June 1983, p. 22.
DISCUSSION
S. Woodward I (written discussion)--How can a filtration/reinhibition process that yields

coolant that fails the ASTM D 3306 specification 20% of the time be recommended for use?
R. Remiasz and R. D. Hercamp (authors" response)--We stated that the system is designed
for use at intervals of 65 000 km, maximum. However, the average usage of the coolants in
the 5 vehicles tested was 55 967 miles or 90 050 km. These used coolants at much higher mile-
ages were chosen to make the test more severe. The results demonstrate that the system per-
forms relatively well even when usage recommendations are exceeded. Even though the
treated coolant failed to meet ASTM D 3306 20% of the time for coolants that had been used
38% longer than recommended, it is reasonable to conclude that the system, when used as
recommended at 65 000 km maximum, will provide coolant meeting D 3306 performance.
S. Woodwardz (written discussion)--It was stated that the on-vehicle recycle equipment
attached directly to one of the heater hoses. Can you comment about the future ability of your
system to work because of the increasing difficulty of accessing heater hoses due to underhood
clutter?
R. Remiasz and R. D. Hercamp (authors' response)--This concern exists today. Some vehi-
cles are much more difficult to service. We continually provide customers with new connec-
tion adaptors and procedures to be used as new vehicle configurations are introduced. We are
not aware of any that cannot be serviced, but the more difficult models require special adaptors
and special procedures. They also may take more time to service.

Robert C. Richardson ~
A Multi-Stage Process for Used Antifreeze/

Coolant Purification
REFERENCE: Richardson, R. C., "A Multi-Stage Process for Used Antifreeze/Coolant Puri-
fication," Engine Coolant Testing: Third Volume, ASTM STP 1192, R. E. Beal, Ed., American
ABSTRACT: Today, the subject of recycling used antifreeze/coolant is a popular and very com-
plex issue. Legislation has forced users in some areas of the United States to either recycle or to
pay for proper disposal through a licensed facility. Recycling represents a viable alternative to
disposal but poses technical challenges never before attempted. This paper describes a unique,
multi-stage purification process (MSPP) based on dual-bed deionization. The process can be
effectively utilized in automotive or heavy-duty applications to remove inhibitors and contam-
inants in bot h ionic and particulate form. This paper presents (1) analysis of several used coolants
before and after the process, (2) comparison of other recycling processes, (3) hazardous waste
characterization testing of process components and finished product, and (4) complete corro-
sion/performance testing of the finished coolant.
KEYWORDS: ion, ion-exchange, total dissolved solids, glycolate, formate, recycling, multistage
purification process
Recycling used engine coolant, while providing a solution for hazardous waste disposal,
poses technical challenges never before attempted. Characterization of the constituents pres-
ent in used engine coolant and their concentrations has created more questions than it has
answered thus far. However, this paper attempts to answer some of those questions and will
provide a platform from which future research can grow. The end result of the recycling effort
should be higher quality engine coolants, both virgin and recycled, as well as a better under-
standing o f engine coolants and cooling systems in general.
A unique, multi-stage purification process for used engine coolant has been developed that
is based on dual-bed deionization. This paper examines the efficiency of the process for remov-
ing inhibitors and contaminants in both particulate and ionic form. Data from corrosion and
performance testing of the finished, prediluted engine coolant shows that this recycled product
meets all current ASTM standards and is being successfully evaluated by at least one major
U.S. Original Equipment Manufacturer (OEM). Hazardous waste characterization testing,
according to Environmental Protection Agency (EPA) standards, of process components and
finished product is also presented.
One o f the greatest challenges facing today's automotive and industrial shops is complying
with governmental regulations concerning the proper disposal of hazardous waste. In recent
years, world attention has increasingly focused on the earth's delicate ecology and ways to
restore and preserve the natural environment. This attention has been aimed at reducing
sources of waste, producing more biodegradable materials, and recycling a larger portion of
the materials we use. We have seen massive changes in business and industry to adapt and to
Laboratory manager, BG Products, Inc., 701 South Wichita, Wichita, KS 67213.
258
RICHARDSON ON ANTIFREEZE/COOLANTPURIFICATION 259
provide products designed for these new markets. The transportation industry has offered one
of the largest and most visible of these new markets, one created mainly by more rigorous
environmental legislation. These changes include the phase-out of chlorofluorocarbons
(CFC), such as dichlorodifluoromethane (R- 12), for use in automotive air conditioners, man-
datory use of oxygenates in gasoline to lower exhaust emissions in large urban areas, and
proper disposal of used fluids such as motor oil and engine coolant. The aforementioned envi-
ronmental mandates have culminated in the introduction of many types of recycling that pro-
vide viable alternatives to disposal. This stems the flow of hazardous materials to toxic waste
sites or landfills, or both, and provides a cost-effective, often profitable, solution for end users.
Legislation
On October 17, 1986, the Superfund Amendment Reauthorization Act (SARA) Title III
was signed into federal law by the U.S. Congress. Section 313 of this legislation designated
ethylene glycol, the major component in most engine coolants, as a toxic chemical. Further-
more, the EPA defines ethylene glycol as hazardous waste under 40 CFR 414.60 and 40 CFR
261.24 if it contains any regulated heavy metals, such as lead, in concentrations above 5 mg/
L (Waste D008). Section 311 of the Clean Water Act prohibits the disposal of hazardous waste
in any sewer system. Although some states allow used engine coolant to be discharged to a
sanitary sewer, provided it goes to a treatment-for-disposal plant, others have set more strin-
gent guidelines than the federal laws.
All of this has led to massive confusion and problems for shops who must service vehicle
cooling systems and periodically replace the used or spent engine coolant.
Commercial Technologies
In the three to four years since recycling engine coolant has become commercially available,
as many as seven different methodologies have been offered. A list of these might include fil-
tration, chemical filtration, chemical filtration/oxidation, centrifugation/filtration, vacuum
distillation, reverse osmosis, and ion-exchange/filtration. Table 1 lists each process in general
terms and comments on the ability of the process to separate inhibitors, contaminants, and
degredation products from the reuseable water/glycol mixture.
As can be seen in Table 1, there are large differences in separation efficiencies between the
various processes. When used a/one, conventional filtration ( > 1 #m) is not adequate to pro-
vide complete purification of spent coolant, since it cannot remove contaminants, degradation
products, and dissolved solids and provide a consistent finished product with a known inhib-
itor balance. Dissolved solids, for these purposes, are defined as compounds that will pass
through a 0.7-tzm filter. They include water impurities, such as chlorides and sulfates, inhibi-
tors, such as borate, phosphate, silicate, and some corrosion metals.
TABLE 1--On-site engine coolant recyclingprocess comparison.

Process Purification Efficiency
Filtration very low
Chemical filtration low
Chemical filtration/oxidation low
Centrifugation/filtration low
Vacuum distillation high
Reverse osmosis high
Ion-exchange/filtration high
Background
Chloride and Corrosion Metals
Chloride is known to promote pitting corrosion of aluminum, as does soluble corrosion
metals (for example, copper) albeit by different mechanisms. Chloride promotes pitting by
setting up a localized corrosion cell within a crevice, casting pore, or other microcell [ 1] while
soluble corrosion metals, such as copper (Cu 2+) and zinc (Zn2+), function by coupling with
metals, such as aluminum, to set up active galvanic cells [2].
There has been some disagreement as to whether or not "soluble" corrosion metals (cations)
actually exist in used engine coolant. This is primarily questioned because it is known that
cations will readily form insoluble hydroxides at a pH of 8 to I0. Therefore, the author sam-
pled some representative used automotive coolants and analyzed each before and after a 0.7-
um filtration. Cations remaining in the coolant after filtration would be considered soluble.
Results are shown in Table 2.
The D 15.15 samples represent coolant from approximately 2100 vehicles and were from
bulk collection in California. The three vehicles sampled were from the Midwest region, and
the DI 5.15 used (simulated) was a coolant made in the author's lab in an attempt to simulate
the physical/chemical properties of the average used coolant more fully outlined in Table 3.
As can be seen in Table 2, some soluble corrosion metals, including zinc and copper, do exist
even at elevated pH.
Glycol Degradation
Another ionic species that cannot be removed by conventional filtration is the oxidative/
thermal degradation products of glycol. For the purpose of this paper, this is limited to gly-
colate, formate, and acetate, which are the major by-products of ethylene glycol (EG) break-
down. EG is still the most widely used freezing point depressant in commercial engine coolant
formulations. Propylene glycol, although not inert, is much more resistant to thermal/oxida-
tive stresses and is gaining ground as a replacement freezing point depressant [3]. EG degra-
dation has been well studied in simple systems with relatively few inhibitors [4], but it has not
been well established whether the degradation products pose corrosion problems at concen-
trations found in a typical used engine coolant. In a recent effort by ASTM Subcommittee
D 15.15 on Recycled Coolants to document the condition of the "average" used coolant, the
following properties (Table 3) were found in a used coolant population of 28 to 100 samples.
As shown in Table 3, glycolate/formate levels in used coolant range from 123 to 1172 mg/
L with a mean of 647 mg/L. Since these are average levels only and would not represent a
TABLE 2--Soluble corrosion metal evaluation in used engine coolants before~after 0.7 um glass fiber
filter-metal concentrations.
Sample pH AI Cu Fe Pb Zn
D 15.15 Round Robin # 1 8.61 0/0 0/0 0/0 0/0 18.0/0
DI 5.15 Round Robin #2 8.34 0/0 0/0 17.0/0 8.0/0 10.0/4.0
'89 Olds Calais 8.07 12.3/10.3 0.2/0.2 42.5/1.5 2.4/0.7 17.0/1.0
'83 Honda Accord 10.70 2.7/2.6 4.6/4.1 0.1/0 4.7/0.4 0.8/0
'80 Chevy Malibu 8.5 4.4/2.6 4.9/0.2 1.8/0 5.9/0.9 6.7/0
D15.15 Used (simulated) 8.74 0/0 15.1/0 69.5/1.5 34.3/17.9 31.4/0.2
NOTE: Analysis performed by inductively coupled plasma atomic emission spectroscopy (1CPAES)
using a 1:10 sample prep in a 1% nitric acid aqueous matrix. All metal concentrations in mg/L w/w.
Detection limit 0.1 mg/L.
RICHARDSON ON ANTIFREEZE/COOLANT PURIFICATION 261
TABLE 3 - - A S T M D15.15 summary sheet: used coolants frequency distribution charts.
Number of Standard
General Properties Samples Mean Deviation (_+)
pH 88 8.82 _+0.94
RA 74 5.13 _+2.57
% Glycol 72 41.10 _+17.00
Freeze point, *F 44 --27.4 _+27.30
Total dissolved solids, mg]L 34 15 581.60 + 17 713.20
Inhibitors, mg/L
phosphorus 56 899.98 _+ 1 179.33
nitrate 67 791.94 _+675.43
boron 60 213.83 _+203.74
silicon 34 137.62 _+ 137.00
tolyltriazole 52 258.67 + 118.18
Contaminants (rag/L)
glycolate/formate 58 647.45 _+524.78
Chloride 80 48.01 _+63.11
Corrosion Metals (rag/L)
lead 100 13.95 _+ 15.29
copper 98 9.11 + 11.59
iron 83 48.19 + 84.24
zinc 28 33.69 46.40
aluminum 47 2.26 5.02
worst-case scenario, it was decided to e x a m i n e potential deleterious effects o f these species.

A S T M Corrosion Test F o r Engine Coolants in Glassware (D 1384) was utilized for this pur-
pose. A m o d e r n phosphate/silicate based coolant, k n o w n to perform well in the test, was used
as the control, and a second, triplicate test using the control coolant spiked with 450 m g / L of
glycolic acid (99 + % H O C H z C O 2 H ) and 195 m g / L o f formic acid (95% to 97% HCO2H) served
as the experimental samples. Results are shown in Table 4, and it is obvious that glycolate a n d /
or formate aggravated a l u m i n u m corrosion.
Figure 1 shows the severe pitting induced on the a l u m i n u m coupon. This suggests that alu-
m i n u m m a y be particularly susceptible to pitting corrosion from these species and that heat
exchangers would be o f particular concern.
TABLE 4--Glycolate/formate corrosivity evaluation as

measured by A S T M D 1384.
Specimen Sample # 1a Sample # 2 a D 4656 Limitsb
Copper 2 2 10
Solder 3 4 30
Brass 2 2 10
Steel 1 1 10
Cast iron 2 1 10
Aluminum 1 14c 30
N o r m Sample ~1:50/50 prediluted automotive coolant. Sam-

ple ~2:50/50 prediluted automotive coolant with 450-mg/L gly-
colic acid and 195 mg/L formic acid added.
a Solutions prepared in accordance with ASTM specification for
prediluted automotive antifreeze/coolant D 4656.
b Average (3) weight loss in milligrams.
c Evidence of severe nonlocalized pitting corrosion.
Specimen: D1384 Cast Aluminum Specimen
Description: Magnification of pit formed in casting void and

propagating through intergranular galvanic corrosion.
FIG. 1--Photomicrograph of D 1384 specimen.
Although Hercamp et al. [3] showed in D 1384 tests with uninhibited EG/corrosive water
solutions that EG breakdown products were corrosive toward lead, no significant increase in
solder corrosion was apparent upon the addition of glycolic/formic acid.
To more fully understand the thermal/oxidative degradation ofEG, a test was designed to
measure the separate and combined effects of c o m m o n corrosion inhibitors. The test utilized
ASTM Oxidation Stability of Steam Turbine Oils By Rotating Bomb (D 2272) apparatus with
the following adaptations for consideration of glycol solutions:
9 50/50 v/v glycol solutions in deionized water

9 380 minute test duration
9 101*C test temperature
9 620 kPa oxygen pressure
9 Electrolytic copper wire catalyst
The following properties of the glycol solutions were evaluated before and after each test to
evaluate the extent of oxidation and the success of inhibitors:
9 pH change
9 Reserve alkalinity loss and % additive depletion (calculated)
9 Conductance
9 Degradation product concentration (by ion chromatography exclusion)
9 Dissolved copper concentration (by inductively coupled plasma atomic emission
spectroscopy)
Interestingly, initial runs at the chosen test length of 380 m i n with a known, fully formulated
coolant, produced well over the m a x i m u m glycolate/formate levels found in typical used cool-
ants and additive depletion ranging from 9.28 to 60.0%. Twenty-five percent depletion is the
specified m a x i m u m in the G M 1825M standard after 1000 h in the simulated service test.
Therefore, it was decided that the test apparatus and conditions were right for the desired
screening of corrosion inhibitors and that glycols may provide insight into the comparative
service life of various formulations in service. Results of the test are outlined in Table 5, Fig.
2, and Fig. 3.
It is interesting to note that most inhibited glycol solutions oxidized to a greater extent than
the uninhibited glycol solution. The exceptions were the glycol solutions with copper inhibi-
tors such as various azoles (for example, mercaptobenzothiazole). This could be partially due
to the fact that glycol degradation products can inhibit further oxidation by the common-ion
principle as has been previously demonstrated [4]. The most c o m m o n engine coolant pH buf-
fers, borate and phosphate, seemed to aggravate the oxidation of EG more than any single
inhibitor evaluated, although simple combinations of mono and dicarboxylic acids, which are
becoming of interest as inhibitor systems [5], were worse. However, in the presence of suffi-
cient copper inhibitor, mono/dicarboxylic acid formulations performed far better than con-
ventional borate/phosphate formulas with respect to end of test (EOT) glycolate/formate
levels.
TABLE 5--Glycol oxidation study, sample legend.
Sample ACondition, Formic, Glycolic, Copper,

ID Description ApH AR.A #S/cm mg/L mg/L mg/L
PGI PG/DI H20 50/50 - 1.18 . . . . . . 0.7 <0.31 94.0

EG 1 EG/DI H20 50/50 - 1.18 27.0 73.0 111.0
EG2 PLUS 900 mg/L P as -1.52 -'li(J "34 136.0 1210.0 5.1
K2HPO4
EG3 PLUS 213 mg/L B as - 1.78 - 1.1 76 220.0 1430.0 308.8
Na2B407
EG4 PLUS 750 mg/L NaNO3 - 1.72 -0.1 37 53.0 120.0 105.4
EG5 PLUS 980 mg/L sodium ... -0.2 17 43.0 220.0 38.7
silicate 28.7% SiOf
EG6 PLUS 2500 mg/L Na - 1.44 -0.4 18 10.0 30.0 0.0
mercaptobenzothiazole
50%
EG7 PLUS 1000 mg/L Na -0.69 -0.2 67 14.0 43.0 0.0
tolyltriazole 50%
EG8 PLUS 1000 mg/L Na --2.84 -0.2 3 11.0 <0.5 1.7
dimercaptobenzothiazole
EG9 PLUS 1.772% oclanoic -0.59 -0.8 -370 170.0 2800.0 169.0
acid/0.772% sebacic
acid
EG10 PLUS 1.89% octanoic --0.67 -0.9 --70 180.0 3400.0 99.0
acid/0.61% adipic acid
EGl 1 PLUS phosphate based -2.50 -0.8 ... 41.0 1686.0 8.0
inhibitor system
EG 12 PLUS GM 6043-M - 3.05 -- 3.0 64 130.0 1420.0 2.6
inhibitor system
EGI 3 commercial long life --0.12 --0.4 0 115.0 110.0 0.0
coolant 50/50 DI H20
EGI4 commercial recycled -2.97 - 1.8 -20 520.0 3100.0 62.7
antifreeze/coolant 2
a Lactic acid.
b Produced by a process other than ion-exchange, distillation, or reverse osmosis.
CooPer, mg/L
350 f l
300
250
200
160
100
50
0
P E E E E E E E E E E E E E E
G G G G G G G G G G G G G G G
1 1 2 3 4 6 6 7 8 8 1 1 1 1 1
0 1 2 3 4
~G. 2--EOTsohzbleeopper.
Total Degradation Products, mg/L x 1000

5-
4-
3- t
2 J
i P E E E E E E E E E E E
G G G G G G G G G G G G
E E E
G G G
1 1 2 3 4 6 6 7 8 9 1 1 1 ! 1
0 1 2 3 4
I GLYCOLDEGRADATION
GLYCOLATE ~ FORMATE
FIG. 3--EOT degradation products.

It has been generally established that elevated levels of dissolved solids can be detrimental
to cooling system performance. Hercamp [6] reported that a heavy-duty coolant with a cal-
culated total dissolved solids (TDS) level of 3.4% w/w promoted water pump seal leakage in a
simulated service rig. The 3.4% corresponds to approximately 41 000 mg/L using an accept-
able gravimetric method (EPA 209B). While water pump seal failure is not univariate,
McKenzie and Kinisky [ 7] found that the primary mechanism by which water pump seal leak-
age occurs is by face seal filming. Coolant is subjected to high temperature between the mating
surfaces of the mechanical seals resulting in nucleate boiling, which leaves patchy films, 5 to
10 t~m thick, of solid matter on the seal face. This separates the face seals providing a leak path
for the coolant to escape.
High levels of TDS can also promote reduced heat transfer. When solubility limits of the
various inorganic salts are exceeded or when incompatibilities between these salts occur, the
result is precipitates or gels coming out of solution. These usually collect in low temperature,
low flow areas of the cooling system, or they can deposit on cooling system components, char-
acterized by high heat flux causing an insulating scale such as calcium phosphate. Silicate is
an inhibitor of particular concern with respect to solubility problems because of its known
propensity to gel and cause flow restriction [8].
This data collectively suggest that processes designed to adequately purify spent engine cool-
ant and provide a consistent finished product must remove a majority of the TDS fraction;
this includes inhibitors, glycol degradation products, water impurities, and corrosion metals.
Description of the Multi-Stage Purification Process (MSPP)

The MSPP was designed to purify used antifreeze/coolant by combining successive stages
of suspended solids and dissolved solids removal. Figure 4 is a simplified flow diagram of the
MSPP where the graduated shading depicts the successive stages of purification.
Filtration begins with a basket strainer (1) removing large particulates such as stop leak
materials that are greater than 115 #m in size. The strainer's main purpose is to protect the
system pump (2) and to extend cartridge filter (3, 4) life. Next is a cartridge filter (3), which
FIG. 4--MSPP flow diagram.

removes suspended solids greater than 25 um, followed by a 5-um cartridge filter (4) with a
core employing granular 20 40 mesh activated carbon to adsorb organics. This protects the
ion-exchange resins from organic fouling, which renders the resins ineffective. Since oil can be
a c o m m o n engine coolant contaminant [9], this is an important design concept.
After filtration in Stages 1,2, and 3, the coolant, now free of particulates and heavy organics,
flows into the cation exchange tank (5) through a tube to the bottom of the tank where it flows
evenly upward contacting the ion-exchange resin. All positively charged ions (cations) are
replaced by hydrogen (H +) ions during this stage. In stage 6, all negatively charged ions
(anions) are exchanged for hydroxide ions (OH-). The H + and O H - ions exchanged in Stages
5 and 6 combine to form water (H20) in an amount depending on the number of equivalents
(moles of charge) of ions removed from the used engine coolant. The resulting purified fluid
flows into a final activated carbon filter (7) whose purpose is to remove any entrained gases
from the liquid stream. A more detailed explanation of the justification for this filter follows
later in this paper.
Finally, the completely purified EG/H20 mixture passes through an in-line conductivity
probe (8) which measures the conductivity of the solution as an indication of the degree of
purification by the ion-exchange system in Stages 5 and 6. The probe is designed to show a
green light when the conductivity is below 50 uS/cm. This ensures high-quality effluent and
also gives the operator a visible indication that the process is under control and that the ion-
exchange beds have not exceeded their capacity. When the ion-exchange resins are exhausted,
dissolved solids remain in solution and the conductivity of the effluent rises and quickly
exceeds the set point of the probe assembly at 50 uS/cm. This trips the green light offand lights
a red light, indicating to the operator that the ion-exchange tanks must be removed for regen-
eration and replaced with a fresh set. The 50-uS/cm set point is normally reached within
a few seconds after the beds have been exhausted, and maximum TDS at the 50-#S/cm
point is 53 mg/L (as NaC1). Since glycol suppresses the conductivity of ions in solution,
as can be seen in Fig. 5, this suppression must be taken into consideration in all TDS calcu-
lations.
These data stem from an evaluation in the authors lab to investigate the effect of glycol on
conductivity derived calculation of TDS versus purely aqueous systems. TDS levels were
achieved using a phosphate/silicate based inhibitor package in deionized water and antifreeze
grade EG.
One set of ion-exchange tanks will generally purify 378.5 to 832.8 L ( 100 to 200 gal) of used
engine coolant before regeneration is necessary. At the average of 567.8 L (150 gal) and an
average TDS level of 15 581.6 mg/L, the capacity of the ion-exchange resins can be calculated
at 3.5 mEq/mL.
The efficiency of the ion-exchange resins at removing dissolved solids depends largely on
two factors: flow rate of the engine coolant through the resin and the ionization constant of
the cation or anion to be removed. However, at an adequate flow rate (0.21 to 0.54 L 9 min-J/
m-3), most of the ionic species found in used engine coolant do not exceed the acid strength
limits (Table 6).
Table 7 gives a good theoretical image of what should be possible to remove from the used
engine coolant using ion-exchange technology. As can be seen, borate and silicate, due to their
low acid strength, are the most difficult species to remove from conventional North American
engine coolants and are the first to be detected as the ion-exchange tanks approach their capac-
ities. Also, it suggests that European, Japanese, and new "long-life" coolants utilizing different
inhibitor systems (that is, benzoate-nitrite and sebacic/adipic acid) may be effectively purified
using the type of ion-exchange system used by the MSPP. This is assuming that these acid
strengths, which were calculated in purely aqueous systems, are somewhat similar in water/
glycol solutions.
RICHARDSON ON ANTIFREEZE/COOLANT PURIFICATION 267
CONDUCTIVITY, mlcroslemens/cm
8000
7500
7000
/
6500
/
/
6000 /I"
5500
5000
4500 /I Z "~-
4000
3500
300O
2500 7/ /
2000 - -
1500
1000 ~ j..~. I-~
500,
0! , , -- , , , i
0 5 ~0 15 2o 25 30 35
TOTAL DISSOLVED S O L I D S , mg/L(Thousar, ds)
'[
,i ~ 10 ~
PERCENT E e
3o_ ~ ,~0_ -e- 70
!
9Specific; COndUc;tstlc;e at 25"0
pH 10.8
FIG. 5--Conductivity suppression in glycol-aqueous EG based engine coolant solution.
Complications
W i t h a n y r e s e a r c h a n d d e v e l o p m e n t effort, u n f o r e s e e n c o m p l i c a t i o n s seem to arise. T w o
different u n d e s i r a b l e c h e m i c a l r e a c t i o n s occurring in t h e cation e x c h a n g e b e d were realized
d u r i n g t h e d e v e l o p m e n t o f t h e M S P P . B o t h created p r o d u c t s t h a t were e i t h e r r e m o v e d or dealt
w i t h using m o r e process e q u i p m e n t or c h e m i c a l additives, or b o t h .
TABLE 6--Dissociation constants versus ion-exchange

limitations.
Dissociation Constant
Species, Anions T, *C pKa Resin Limit
Adipic acid 25 4.63 a 9 to 10

Acetic acid 25 4.75 9 to 10
Benzoic acid 25 4.19 9 to 10
Boric acid 50 9.08 9 to 10
Citric acid 20 3.14 9 to 10
Formic acid 25 3.83 9 to 10
Glycol 25 14.22 9 to 10
Glycolic acid 25 3.83 9 to 10
Oxalic acid 50 4.417 9 to 10
Phosphoric acid 50 2.260 9 to 10
m-silicic acid ... 9.70 9 to 10
a pKa for first ionization.

TABLE 7 - - G M evaluation: coolant properties before and after MSPP. a
Chemical Properties Before MSPP After MSPP Reinhibited # I Reinhibited #2

Contaminants
chloride (C1) 64 28 17 28
sulfate (504) 290 6 82 100
formic acid 780 0 0 0
glycolic acid 1 800 28 15 0
acetate 21 0 0 0
calcium, mg/L 0 1 0 1
magnesium, mg/L 0 0 0 0
Corrosion metals
aluminum, mg/L 3 0 3 3
copper, mg/L 7 0 0 0
iron, mg/L 0 0 1 2
lead, mg/L 1 0 0 0
zinc, mg/L 0 0 0 0
Inhibitors
boron, mg/L 516 0 0 0
molybdenum, mg/L 249 1 181 149
phosphorus, mg/L 990 28 2 612 2 421
potassium, mg/L 1 203 52 4 474 5 315
silicon, mg/L 20 7 86 92
sodium, mg/L 3 014 7 318 380
nitrate (NO3), mg/L 760 27 450 450
nitrite (NO2), mg/L 35 14 14 0
General Properties:
pH 8.37 8.46 10.38 10.38
freezing point, *Fb -- 60 -- 36 - 35 - 34
glycol, wt% 58 52 50 50
reserve alkalinity, ml 7.5 0.3 6.4 6.3
conductivity, ~S/cm 1 879 21 2 769 2 810
TDS (calculated), mg/L 7 500 200 8 300 8 000
TDS (209B), mg/L 16 638 110 11 362 10 703
TSS, mg/L 1 318 90 356 419
visual brown-green clear/colorless bright green bright green
a Analysis performed by BG research laboratory.
b t*C = (t*F - 32)/1.8.
The first reaction was the conversion of nitrite (NO2) ions, which are commonly found in
most heavy-duty engine coolants, into nitrous acid (HNO2). This reaction occurs via proton-
ation of the NO2 ions as they pass through the strong acid (sulfonic) cation exchange resin.
HNO2, which is highly unstable, quickly decomposes into nitric oxide [NO(g)] which is oxi-
dized by the oxygen in the aerated coolant solution to form nitrogen dioxide [NO2tg)], a toxic,
reddish-brown gas. The reaction is summarized below
EG/H20 decomp. 02
NOr ~ HNO2 ~ NOtg) + O H - --" NO2(g)
[H +]
Since NO2tg) is toxic, careful work was done using different process techniques and moni-
toring NO.) gases with sensitive gas detection equipment. The problem was eliminated with a
special filter (see # 7 on Fig. 4) employing a special grade of activated carbon that removes most
of the gas from the liquid stream.
The second reaction involved the oxidation of glycolate, formate, and acetate to carbon
dioxide (CO2). This was again found to be occurring in the cation exchanger because of a com-
bination of the strong acid nature of the resin type being used and an unidentified oxidant.
The CO2 produced created some difficulties in obtaining a balanced pH in the finished engine
coolant. The presence of CO2 was usually noticed upon taking a pH measurement of the puri-
fied coolant solution before any inhibitor addition. The pH usually fell between 3.2 and 4.2,
which is precisely the expected pH for a saturated H20]CO2 solution. With no CO2 present,
the solution would exhibit a near neutral pH after the ion-exchange stage of the MSPP. More
careful measurements found the pre-ion-exchange CO2 concentration was typically 50 mg/L
and the post-ion-exchange concentration around 260 mg/L. After the two-part MSPP inhib-
itor package is added to a near neutral solution of glycol/water, the pH will be 10.3 +_ 0.1.
However, with the CO2 present, the pH would typically fall well below this to around 8.9. Our
hypothesis is that CO2 is converted to carbonate ( C O / ) and bicarbonate (HCO3) when the
purified coolant is made alkaline with the two-part additive addition. At a pH of 9 to 11, the
HCO3 would dominate. This sets up a bicarbonate buffering system that shifts the desired pH
from 10.0 to 8.9.
The solution to this problem is to make a simple pH check after additive addition. If nec-
essary, an adjustment to the desired 10.0 pH is made with the alkaline portion of the two-part
MSPP inhibitor package. Any extra buffering capability derived from the C O ] - / H C O ; is
viewed as beneficial and foaming was shown to not pose a problem.
Corrosion Tests and Performance Data

The best purification process in the world would be useless without a well-formulated, bal-
anced inhibitor package to restore the necessary corrosion inhibitors and performance addi-
tives to the engine coolant. A balanced pH; all system metals protection; foam inhibition; scale
dispersants, sludge dispersants and detergents; buffering capability; and elastomer compati-
bility are all important performance targets that must be met.
A novel, two-component, phosphate-silicate-molybdate based inhibitor system was devel-
oped for the MSPP in automotive/light-duty applications. In mixed-fleet and heavy-duty
applications, a similar two-component, low-silicate, phosphate-nitrite inhibitor system is
used. The two-component system was developed out of necessity due to storage stability prob-
lems incurred with a concentrated, one-component package. However, the two-component
approach allows a more foolproof design for recycled coolant re-inhibition. The two separate
inhibitor concentrates are dyed different colors: one yellow and the other blue. Since all color
(dye) is removed by the MSPP in most cases, the desired blue-green coolant coloration is pro-
duced and provides a visual indicator that both components have been added. Furthermore,
the two components allow for adjusting the pH, since one is acidic and the other is basic. In
normal operation, the two components produce a coolant with a balanced pH of 9.5 to 10.5
when used at a treat rate of 1 oz/qt of coolant or cooling system capacity (3.125% v/v) each.
General Motors--Service Technology Group Evaluation

This program was designed to supply service dealerships with accurate information as to
what commercial processes could be safely used to recycle used coolant without voiding war-
ranty. It was developed out of response to a desperate need of recycled coolant standards in
the field. The Service Technology Group (STG) of GM organized and performed the initial
evaluation while the actual testing of the resultant finished engine coolants was performed by
a leading independent laboratory. The GM standard 1825-M was used as the target for passing
performance requirements. However, since most of the coolants evaluated were prediluted,
ASTM Prediluted Aqueous Ethylene Glycol (50 volume percent m i n i m u m ) Base Engine
Coolant For Automobiles and Light Duty Service (D 4656) was used for the standard for phys-
ical and chemical requirements. It also was followed in preparing the coolant solutions for
corrosion tests.
Approximately 300 gal (1336 L) of spent engine coolant was collected by the STG that rep-
resented approximately 230 passenger cars of various years, makes, and models. This coolant
was collected in a c o m m o n tank and represented North American, European, and Japanese
inhibitor technologies and had a mean service life of seven years. The coolant was then agitated
and divided into 5-gal (19-L) containers for processing. Potential suppliers of recycling pro-
cesses were then invited to recycle 10 gal (38 L) of this spent coolant, balance the freezing point,
and re-inhibit such that it would be ready for use at the commercial levels. Furthermore, units
having some internal capacity were precharged with the spent coolant solution before the final
10 gal (38 L) were processed so that the coolant tested would be representative of the starting
material.
The following is a summary of test data obtained by a leading independent laboratory for
TABLE 8--GM evaluation: corrosion and performance data.
Test Results
ASTM Test
Requirements Method GM 1825M Pail # 1 Pail #2
Physical/chemical requirements a
specific gravity, 60/60~ b D 1122 1.065 min 1.079 1.080
freezing point *F D 1177 --34 rain -34.4 -34.3
boiling point, *F D 1120 226 rain 226 226
effect on automotive finish D 1882 no effect none none
ash content, wt% D 1119 2.5 max 1.04 1.04
pH D 1287 7.5 to 11.0 10.2 10.2
Reserve alkalinity, ml D 1I21 report 5.5 5.6
Chlorides, mg/L D 3634 25 max <4 <4
R.A., % loss after 1000 hours D 1121 25% max NDA NDA
Performance requirements
corrosion in glassware D 1384
copper 10 0 0
solder 20 --0.3 0
brass 10 --0.3 -0.3
steel 10 0 0
cast iron 10 0 0
aluminum 20 --0.3 +0.3
simulated service test D 2570
Copper 10 NDA NDA
Solder 40 NDA NDA
Brass 20 NDA NDA
Steel 20 NDA NDA
Cast iron 20 NDA NDA
Aluminum 40 NDA NDA
heat transfer corrosion D 4340 1.0 <0.1 0.3
foaming, volume (mL)/break time(s) D 1881 50/5 47/1 32/1
cavitation erosion D 2809 8 at 300 NDA NDA
glycol composition by GC NA no req
Ethylene glycol, wt% ... 49.0 48.8
Diethylene glycol, wt% ... 2.5 2.6
Propylene glycol, wt% ... <0.5 <0.5
Per D 4656 prediluted specifications.
b t*C --- (tOF- 32)/1.8.
the G M on-site engine coolant recycling evaluation (Tables 7 and 8). The testing was not c o m -
plete at the c o m p l e t i o n o f this paper so the full data were not available for inclusion, although
preliminary results indicate that the finished coolant produced by the M S P P will satisfy the
requirements o f G M .
D 4656/D 4985 Test Data

Since full p e r f o r m a n c e data were n o t available from the G M evaluation, test data obtained
on testing o f the M S P P t w o - c o m p o n e n t additives, both light and heavy duty, are presented
(Tables 9 and 10). A S T M D 4656 was used for evaluation o f the light-duty inhibitor, since a
TABLE 9--Light-duty performance testing of the MSPP inhibitor in virgin EG.
ASTM D 4656
Requirements Specified Test Specifications Results
Physical/chemical requirements
specific gravity, 60/60~ ASTM D 1122 1.065 min 1.0807
freezing point, 50 vol% ASTM D 1177 --34* rain -34"F
boiling point ASTM D 1120 226"F, rain 227"F
effect on automotive finish ASTM D 1882 no effect ...a
ash content, max wt% ASTM D 1119 5.0 1.2
pH, 50 vol% ASTM D 1287 7.5 to 11.0 10.4
reserve alkalinity ASTM D 1121 report 5.7
General requirements
b
color
effect on nonmetals no adverse effect
storage stability ASTM draft 1 year min 1~Ayear rain
Performance Requirements
corrosion in Glassware ASTM D 1384
copper 10 mg/coupon, max. 6
solder (30/70) 30 1
brass 10
steel 10 ()"
cast iron 10 1
aluminum 30 0
high lead solder ... 33
Simulated service test c ASTM D 2570 (79)
at 33g% antifreeze
copper 20 mg/coupon 4.8
solder (30/70) 60
brass 20 :/12
steel 20 0.6
cast iron 20 0.4
aluminum 60 +5.3
high lead solder ... 8.3
corrosion of cast aluminum at
heat rejecting surfaces ASTM D 4340 1.0 mg/cm2/wk, max. 0.2 mg/cm2/wk
foaming, volume (mL)/ ASTM D 1881 150 mL, max./5 s max. 33 mL/1 s
breaktime(s)
cavitation erosion ASTM D 2809 8 rain at 100 h 10 at 1000 h a
NOTE: t*C = (t*F -- 32)/1.8.
a No adverse effect expected.
b Color is entirely up to the manufacturer.
c Modified--test run on coolant as received (50/50) and using Ford parts.
a Used Ford Test Method BL 3-2, a more stringent test, testing the same factors but under harsher
conditions and extended time.
TABLE lO--Heavy-duty performance testing of the HD MSPP inhibitor in virgin EG.
ASTM D 4985(89)
Requirements Specified Test Specifications Results
Physical/chemical requirements
specific gravity, 60/60~ ASTM D 1122 1.110-1.145 1.08074
freezing point, 50 vol% ASTM D 1177 -- 34* max - 34"F b
boiling point, undiluted ASTM D 1120 325"F, min
boiling point, 50 vol% ASTM D 1120 226~ min 227"F ~
effect on automotive finish ASTM D 1882 none ... d
ash content, max wt% ASTM D 1119 5.0 1.6
pH, 50 vol% ASTM D 1287 7.5 to 11.0 10.4
water, wt%, max ASTM D 1123 5.0
chloride, mg/L ASTM D 3634 25 max < 10
silicon, mg/L UC 250 max 166
reserve alkalinity, m L ASTM D 1121 10 min 6.8
General requirements
color
effect on nonmetals no adverse effect
storage stability ASTM draft 1 year min 1~ year min
Performance Requirements
corrosion in glassware ASTM D 1384
copper 10 mg/coupon, max 3.1
solder (30/70) 30 0.7
brass 10 3.5
steel 10 0.3
cast iron 10 0.9
aluminum 30 4.7
simulated service test ASTM D 2570 (79)
at 33~A%antifreeze
copper 20 mg/coupon 8.0
solder (30/70) 60
brass 20 "1o:1
steel 20 2.5
cast iron 20 0.7
aluminum 60 9.1
high lead solder ... 65.4
corrosion o f cast aluminum at
heat rejecting surfaces A S T M D 4340 no req 0.4 mg/cm2/wk
foaming ASTM D 1881 150 mL, max/5 s max NR
liner cavitation UCg mg loss/22 h 49.0
NOTE: t~ = ( t * F - 32)/1.8.
a Test run on coolant as received (50/50); spec on prediluted AF 1.065 min.
b Test run on coolant as received (50/50).
c Test run on coolant as received (50/50).
a No adverse effect expected.
e Color is entirely up to the manufacturer.
f Modified--test run on coolant as received (50/50) and using Ford parts.
g Test b u t t o n s - - A S T M G32; Test s o l u t i o n - - A S T M standard on corrosion o f water + l/3x 570A/B;
amplitude--0.8 mil peak to peak; duty cycle--100 s [80 on/20 off]; temp (initial) 130*F., temp (final)
170*F.
prediluted engine coolant is the finished product from the MSPP. For heavy-duty, D 4985 was
used, realizing, of course, that this is an antifreeze concentrate specification.
Hazardous Waste Characterization

As previously mentioned, used antifreeze/coolant or recycling process components, or both,
such as filters, are regulated as hazardous waste under 40 CFR 261.24 if certain hazardous or
toxic characteristics are present. These characteristics might include corrosivity, reactivity, or
toxicity. Toxicity limits can be violated if regulated heavy metals, such as lead, are present in
concentrations exceeding the allowable limits.
Lead in used coolant is present as a corrosion metal that is extracted from the soldered joints
in the radiator by the coolant while in service. At a pH of 8.5 to 10.0, this lead would primarily
be lead hydroxide Pb(OH)2. Typical lead concentration in used coolants is 2 to 30 mg/L with
an average of 6 mg/L. As could be expected, this lead is readily filterable using simple mechan-
ical filtration such as that employed in the MSPP as described previously. Table 11 shows data
collected by a certified independent environmental test laboratory on a used coolant collected
at a service garage involved mainly in over-the-road, heavy-duty engine repair. The data sup-
ports that: (l) The used coolant itself was hazardous waste due to lead concentration, (2) The
MSPP purified the coolant to acceptable levels, (3) The MSPP filters removed the lead and
would be characteristically hazardous, and (4) The ion-exchange regeneration solution did not
contain appreciable amounts of lead.
Conclusions
1. Recycling used antifreeze/coolant is becoming a reality in the marketplace and can be
cost-effective and show environmental benefit.
2. Data presented show that even average levels of ethylene glycol oxidation products can
be corrosive, especially to aluminum, and that these species must be removed in a viable recy-
cling process.
3. The combination of mechanical filtration and dual-bed deionization represents a sound
technical process to purify used antifreeze/coolant.
TABLE 11--Hazardous waste testing of recycling components.
Concentration and Units of Samples, mg/kg
Total Metals 1 2 3 4 5 6 7 Method
Arsenic ... . . . . . . . . . EPA 7060

Barium ... "i5 "25 . . . . . . . EPA 70S0
Cadmium ... 10 "1' .. . . . . EPA 7130
Chromium 19 " "~ 2 EPA 7190
Lead "i8 6 500 53 620 0167 0142 EPA 7420
Mercury ... . . . . . . . . . . . . . . . EPA 7470
Selenium ... . . . . . . . . . . . EPA 7740
Silver .. 9 '28 ... i ...... EPA 7760
NOTE: Sample 1: used antifreeze before MSPP; Sample 2: used antifreeze after MSPP; Sample 3:25
#m filter; Sample 4: activated carbon filter; Sample 5:5 ttm filter; Sample 6: cation regenerant solution;
and Sample 7: anion regenerant solution9
4. The removal of a large portion of total dissolved solids (TDS) in a recycling process is key,
both to minimize potential water p u m p failures and to ensure proper additive compatibility
and shelf life.
5. Efficient removal of TDS in engine coolant must also be accomplished by a recycling
process in order to provide a consistent finished product.
6. Recycled antifreeze/coolant produced by a viable recycling process can meet and exceed
current performance standards, such as ASTM D 3306, D 4656, and D 4985.
7. Some portion of all recycled antifreeze/coolant will still require proper disposal, such as
filters and distillation bottoms, because of heavy metal concentrations.
Acknowledgments
The author would like to thank the following individuals for their help in this writing. John
Dick and Ren6 Wiebe for their competent analytical work, Dr. David Shubert for his expert
technical consultation, Jack Richardson for the superb graphics, and Jo Dinsmoor for her tire-
less efforts in assembling and typing the paper. The author would also like to thank BG Prod-
ucts, Inc. for their support of this research effort and permission to publish.
References
[1] Wiggle, R. R., Hospadarak, V., and Stylogou, F. A., "The Effectiveness of Engine Coolant Corrosion
Inhibitors For Aluminum," Corrosion 80, National Association of Corrosion Engineers Conference,
Paper 69.
[2] Hirozawa, S. T., "Mechanism for the Inhibition of Localized Corrosion of Aluminum in Antifreeze/
Coolants," Corrosion 82, National Association of Corrosion Engineers Conference, Paper 263.
[3] Coughenour, G. E., Hercamp, R. D., and Hudgens, R. D., "Aqueous Propylene Glycol Coolant for
Heavy Duty Engines," SAE Paper 900434 from SAE SP-811, Feb. 1990.
[4] Brown, P. W., Clifton, J. R., and Rossiter, W. J. Jr., "Degraded Aqueous Glycol Solutions: pH Val-
ues and the Effects of Common Ions on Suppressing pH Decreases," Solar Energy Materials, Vol.
12, 1985, pp. 77-86.
[5] Darden, J. W., Maes, J. P., Triebel, C. A., and Van Neste, W., "Monoacid/Diacid Combinations as
Corrosion Inhibitors in Antifreeze Formulations," SAE Paper 900804 from SAE SP-811, Feb. 1990.
[6] Hercamp, g. D., "Recycling Used Coolant From Heavy Duty Engines," presented at the NARSA
Forum on Coolant Disposal and Recycling, National Automotive Radiator Service Association
National Meeting held in Atlanta, GA, 30 March 1989.
[ 7] Kinisky, T. G. and McKenzie, D. A., "Investigation of Premature Water Pump Seal Failures," SAE
Paper 900806 from SAE SP-811, Feb. 1990.
tems," SAE Paper 852327, SAE, Dec. 1985.
[9] "Forum Report--Cooling Systems, A Closer Look," Equipment Management, Feb. 1986, pp. 43-
49.
[10] General Motors Service Technology Group, Wayne Bradly, Senior Project Engineer, Warren, MI.
DISCUSSION
J. A. L i m d (written discussion)--I am wondering if you have an estimate of total cost to

"process" a typical vehicle, not what the charge would be for the actual service, but what the
average cost is at the present time to process a car and, in addition, what is the cost of the
equipment itself?.

DISCUSSIONON ANTIFREEZE/COOLANTPURIFICATION 275
R. C. Richardson (author's response)--Realizing that costs can vary slightly from location
to location, an average cost to process a passenger car is approximately $7.50 assuming a 10-
qt (9.4 km 3) cooling system capacity. This cost does not include labor and overhead or equip-
ment amortization but does include everything necessary to deliver a fully cleaned, balanced,
and reinhibited cooling system. The current cost of the equipment is approximately $8745,
which includes a fully equipped MSPP unit with two sets of deionization tanks which gives
capacity to process 200 to 300 gal (757 to 1136 L) of used coolant.
P. Woyciesjes 2 (written discussion)--What cleaners are recommended for the cooling sys-
tem? If they are the EDTA type, what effect do they have on your process and on the perfor-
mance testing of the final product?
R. C. Richardson (author's response)--A citric acid/EDTA based cooling system flush, spe-
cifically designed with the MSPP, is recommended. The effect of the flush on the overall pro-
cess has never been specifically evaluated, although much of the performance data, including
the GM-STP evaluation, was done with the flush added to the used coolant before processing.
As the final coolant from these evaluations performed well, certainly up to industry standards
(D 3306/GM 1825-M), the impact would appear to be minimal. Furthermore, the benefit
from chemically cleaning the cooling system has been evaluated and is critical to proper main-
tenance of a modern automotive cooling system.
A. Gershun 2 (written discussion)--Why is the cation exchange column placed before the
anion exchange column in your process? What would be the effect of reversing them?
A. C. Richardson (written discussion)--The sequence of the ion-exchange resins is critical,
cation exchange followed by anion exchange. Reversing the order would result in the forma-
tion of insoluble metal hydroxides in the anion exchange, which would quickly foul the system
possibly even resulting in restriction of liquid flow rate.

Wayne H. Bradley 1
An Evaluation of Engine Coolant Recycling

Processes: Part l
REFERENCE: Bradley, W. H., "An Evaluation of Engine Coolant Recycling Processes: Part I,"
Engine Coolant Testing. Third Volume,ASTM STP 1192, R. E. Beal, Ed., American Society for
Testing and Materials, Philadelphia, 1993, pp. 276-288.
ABSTRACT: The Service Technology Group (STG) is an organization within General Motors
(GM) responsible for providing service equipment and information to GM dealerships. STG is
charged not just with the responsibility for current GM vehicles, but for the cars and trucks of
the future as well. The tools, the procedures and the programs that technicians will use tomorrow
must be created and validated today. Various staffs within STG address advanced serviceability
readiness, special tool development, service research, and similar themes.
As the awareness of environmental issues has heightened, the disposition of spent engine cool-
ant has begun to affect automotive service. One possible approach to this coolant disposition
problem is that of coolant recycling. General Motors Service Technology Group has conducted
a study in which the objective was to determine the efficacy of (state of the art) engine coolant
recycling. Spent engine coolant was collected from vehicles and combined to form a reference
coolant with a mean age of approximately seven years. Various coolant recycling processes were
evaluated as they were required to process the reference coolant. Once the recycled samples were
obtained, they were sent to an independent test facility. The test criteria was GM 1825M, a virgin
coolant corporate specification.
Included in this study was the evaluation of off-site coolant recycling facilities that reformulate
engine coolant for reuse.
In advancing its pro-active environmental policy, STG has initiated this study to evaluate the
efficacy of engine coolant recycling. It is our belief that the results of this undertaking will be
beneficial to our dealers, the environment, and our customers.
KEYWORDS: corrosion, inhibitors, recycling, inhibitors, reference coolant, deionization, fil-

tration, centrifugation, flocculation, coagulation, reverse osmosis, vacuum distillation, coolant
extension
Background
Engine Coolant: Its Function and Properties
Practically every passenger vehicle on the road today shares a c o m m o n ingredient: ethylene
glycol used as an integral c o m p o n e n t in the cooling system. As coolant serves its purpose in
protecting the c o m p o n e n t s o f the a u t o m o t i v e cooling system against corrosive and erosive
attack and providing e n h a n c e d freeze/boil protection, it begins to take on certain negative
characteristics. These characteristics typically manifest themselves by a lowered capacity to
protect the cooling system because o f the formation of ethylene glycol degraded acids (acetic,
glycolic, and formic), accumulated dissolved and suspended solids, principally metal oxides,
and salts (these originate from the corrosion of various metals within the cooling system) and
m a k e up water contaminants. Cooling system metals are corroded by cavitation corrosion,
Senior project engineer, General Motors Service Technology Group, 30501 Van Dyke, Warren, MI
48089.
276
BRADLEY ON ENGINE COOLANT RECYCLING PROCESSES 277
direct chemical attack, or by electrochemical processes, either galvanic when dissimilar metals
are electrically coupled, or local cell action between anodes and cathodes on the surface of one
metal [ 1].
Sodium (alkali metal) silicate is generally acknowledged to be the most effective inhibitor of
aluminum corrosion. For this reason, it is included in almost every (North American) com-
mercial coolant. The silicate, however, can be depleted in customer service, resulting in a loss
of aluminum protection [2]. Other important corrosion inhibitors included in engine coolants
also gradually deplete in service. It is for all of these reasons that General Motors requires spe-
cific change intervals for the engine coolant in its vehicles.
Regulating the Disposition of Used Coolant

Until recently, there were few questions asked regarding the disposition of spent coolant. It
was either hauled away as waste or disposed of in conventional storm or sewer drains at the
rate of an estimated 200 million gallons (757 million litres) per year. As continued emphasis
on the environment grew, the disposition of spent coolant took on a different light. Section
313 of Title III of the Federal Superfund Amendments and the Reautherization Act of 1986
classifies ethylene glycol as a toxic waste. The Environmental Protection Agency (EPA) reg-
ulatory cutoff for lead (the threshold for allowable lead in any waste) is 5 mg of extractable lead
per litre of waste (rag/L), also stated as five parts per million (ppm) of extractable lead. Because
the EPA toxicity characteristic is concerned with the likelihood that a hazardous waste will
release its heavy metals into ground water and thereby pollute the drinking water system, the
EPA extraction procedure defines a waste as hazardous based upon its extractable levels of
heavy metals rather than its total concentration.
State and Local Policies

Individual states and local governments (that is, cities and counties) are not preempted by
federal law from enacting and enforcing their own environmental laws and regulations regard-
ing the disposal of used coolant which are more stringent than the federal laws and regulations.
If a state enacts a more stringent environmental protection provision, then a generator must
comply with those higher standards. With regard to ethylene glycol, only California has deter-
mined it to be hazardous waste if in sufficient quantities. However, several states have listed it
as a "pollutant" or "special," "residual," "industrial," or "commercial" waste, which may not
be discharged onto the ground, into waterways or into the local sewage system without special
permission of the municipality or local publicly owned treatment works (POTW). Addition-
ally, even in the absence of any such state regulation of ethylene glycol, local POTW discharge
ordinances may prohibit or restrict the disposal of ethylene glycol into the local sewer system
[3].
To provide environmentally responsible engine coolant service alternatives for General
Motors (GM) dealers, STG has implemented a coolant recycling program. The objective in
conducting this program is to accrue knowledge of engine coolant recycling technology while
laying a foundation for future projects designed to further provide beneficial information and
data for the promotion of this type of recycling. Obviously, General Motors has a vested inter-
est in the quality of the environment and the performance of the coolants used in its vehicles.
It is for these reasons that this engine coolant recycling study was conducted objectively, yet
circumspectly. This paper is the first part of a two part series on this engine coolant recycling
study and is intended to describe the manner in which the test was conducted. A short discus-
sion is included relative to the results.
Format of the Study

Objective
The objective of this study was to determine whether a commercial engine coolant recycling
process, available at the time of this study, was capable of recycling a qualified spent coolant
to meet General Motors virgin coolant specification GM 1825M. This is a virgin coolant spec-
ification and is the same required for coolant suppliers for initial fill in General Motors vehi-
cles. It parallels ASTM Specification for Ethylene Glycol Base Engine Coolant, for Automo-
bile and Light Duty Service (D 3306) with the exception of some performance requirements
(Table 1). Also, an objective was to provide to General Motors dealers the information and
the means to choose a recycling process that has been predetermined to be capable of produc-
ing an adequate recycled product for G M vehicles. The results of this study will be shared with
the public by means of papers such as this and volunteered interaction with outside organi-
zations. This study has been presented to various governmental agencies, and it is believed that
it will provide crucial information in shaping and implementing subsequent legislation regard-
ing engine coolant recycling.
Process Selection
The selection process for choosing the various recycling techniques to be evaluated was rel-
atively straight forward. An announcement was made to the ASTM Subcommittee D 15.15 on
recycled coolants in November of 1990 describing the program and its objective. There were
no qualifying factors or requirements that any of the recyclers needed to meet with the excep-
tion that their product be ready for release to the public (not in the development stage). Each
recycling unit chosen represented a different process. In an attempt to simplify the study, the
recycling processes were segregated into two categories: on-site and off-site. The recycling pro-
cesses that were categorized as on-site were those that would perform the recycling at the deal-
ership. Those that recycled coolant at a facility other than the service outlet where the coolant
was acquired were addressed under off-site.
On-Site Coolant Recycling

Coolant Collection Program
In researching any coolant recycling process, it is essential to characterize the coolant that
is to be recycled. The spent coolant utilized as the input coolant (designated as the reference
coolant) in this program was obtained from General Motors employee vehicles at the Tech
Center located in Warren, MI. The selection of the vehicles was accomplished by selecting 200
of the vehicles with the oldest coolant from a total of 600 candidates. The mean age of the
coolant from these 200 vehicles was approximately seven years. It is believed that many of
these vehicles contained the original factory fill with the exception of occasional overflow tank
top off. Individual samples were taken from each vehicle for analysis before mixing with the
reference batch. In order to encompass the effects that foreign inhibitor formulations may have
on the recycled products, 30 Japanese vehicles were drained, mixed, and added to the reference
coolant (unknown age). Although the act of combining all these coolants with varying inhib-
itor chemistries may not represent a singular inhibitor technology, it is indicative of all coolant
recycling practices to varying degrees. Each recycling process, whether it is batch or vehicle
oriented, would be expected to encounter mixed coolant chemistries because of the residual
coolant in the equipment (direct vehicle recycling equipment) or batch processing.
TABLE l - - Comparison of coolant specifications, ASTM D 3306 versus GM 1825M.
Chemical Requirements ASTM D 3306 GM 1825M
PHYSICAL REQUIREMENTS
Specific gravity, 60/60-F 1.11 t o l . 1 4 l.ll tol.14
Freezing point ~ 50% vol. --34 min --34 min
Boiling point, ~
undiluted 300 min 300
50% vol. 226 min NR
Effect on automotive finish none none
Ash content, wt% 5 max 5 max
pH, 50% vol. 7.5 t o l l . 0 7 t o 11
Reserve alkalinity, mL 10 rain 10 rain
Water, wt% 5.0 max 5.0 max
Chlorides, ppm NR NR
RA, % loss after 1000 h NR 25% max
GENERAL REQUIREMENTS
Color dist. dist. green
Odor not off. NR
Storage Stability NR 6 months
PERFORMANCE REQUIREMENTS
Corrosion in glassware
copper 10 10
solder 30 20
brass 10 10
steel 10 10
cast iron 10 10
aluminum 30 20
high lead solder NR NR
impingement coupon (C) NR NR
impingement coupon (A) NR NR
Simulated Service Test
copper 20 20
solder 60 40
brass 20 20
steel 20 20
cast iron 20 20
aluminum 60 40
high lead solder NR NR
Heat transfer corrosion 1.0 mg/cm2/wk 1.0
Foaming 150 mL/5s 50/5 scc
Cavitation erosion 8 min at 100 H 8 at 300
Galvanostatic pitting test NR NR
Engine dynamometer NR NR
Circulation test NR NR
Fleet test NR NR
NOTE: NR -- Not Required.
Reference Coolant Dispersion

T h e reference c o o l a n t f r o m t h e 230 vehicles was c o n t a i n e d in its e n t i r e t y in o n e b u l k con-
tainer. W h i l e agitating, t h e reference c o o l a n t was consecutively p o u r e d i n t o individual con-
t a i n e r s for ease o f use. Before use, each c o n t a i n e r was s h a k e n t h o r o u g h l y to s u s p e n d a n y sed-
i m e n t . It was r e q u i r e d o f e a c h recycling process to recycle two separate c o n t a i n e r s o f the
reference c o o l a n t as two s e p a r a t e tests.
Reference Coolant Characterization

A complete characterization of the used coolant used in this study (designated as the "ref-
erence" coolant) is shown in Table 2 (two samples were analyzed and labeled 316 and 2). This
analysis, conducted at an independent laboratory, was accomplished by the following meth-
ods. All metals were determined by inductively coupled plasma (ICP). tolyltriazole (TTZ),
benzotriazole (BZT), and mercaptobenzothiazole (MBT) were determined by high-pressure
liquid chromatography (HPLC). All anions, including glycolate and formate, were determined
by ion chromatography (IC). The reference coolant was filtered at 0.45 um before analysis.
Also shown in Table 2 are the results when the reference coolant (unfiltered) was treated with
hydrofluoric acid to dissolve all the visible insolubles and colloids. These analyses give a clearer
picture of the soluble/insoluble ratio of various constituents, particularly heavy metals.
How the Testing Was Conducted

Each recycling system was tested at the General Motors Service Technology Group Head-
quarters building located in Warren, MI, before the author. Strict adherence to the manufac-
turer's operational instructions was followed to ensure that a similar method would be repli-
TABLE 2--Analysis of reference coolant.
Solubles Acid Treated
Analyses #316 #2 #316 #2
pH 8.3 8.3 N/A N/A

Glycol 59.9 59.9 N/A N/A
ASTM RA 7.16 7.16 N/A N/A
TTZ 212 212 N/A N/A
BZT 23 20 N/A N/A
MBT 23 20 N/A N/A
Chloride 50 50 N/A N/A
Nitrite 19 11 N/A N/A
Benzoate 2220 2248 N/A N/A
Nitrate 838 852 N/A N/A
Sulfate 265 268 N/A N/A
Sebacate N/A N/A N/A N/A
2-Et Hexanoic N/A N/A N/A N/A
Octanoate
Boron 458 460 377 400
Aluminum ND ND 34 27
Copper 6 5 10 11
Iron <2 <2 181 138
Lead <2 <2 28 28
Potassium 1005 970 755 903
Molybdenum 149 147 130 112
Sodium 2512 2469 2088 2184
Phosphorus 641 639 616 597
Silicon 42 41 20 26
Zinc ND ND 5 6
Fomate 556 519 N/A N/A
Glycolate 1387 1347 N/A N/A
NOTE: ND = not detected. N/A = not analyzed. All data in
ppm.
cared in the field. Before testing, each piece of equipment was drained and primed with the
reference coolant. This eliminated dilution of the reference coolant with the coolant contained
in the equipment. Each system was required to recycle two separate 5-gal (19-L) batches of the
reference coolant. This allowed for some measure of consistency while providing a margin of
comfort (two test points as opposed to one) in the testing. The recycled product from each
process was individually labeled and shipped to an independent test facility experienced in the
testing of engine coolants. In accordance with proper testing methods, a blind test was per-
formed at the testing facility where the correlation of the recycled sample and its origin
remained unknown. To contribute to the validity of the test procedure, control samples were
submitted (also blind to the test facility) in the form of virgin General Motors coolant and the
untreated reference coolant. Samples of each recycled product were taken immediately follow-
ing processing. The corresponding samples were sent to a specialty lab for complete analysis.
Recycling Processes Evaluated

The following list includes the names of the various types of coolant recycling processes eval-
uated in this study. Some of the processes included mechanical or chemical variations; these
are noted. A detailed description (partially or fully derived from each equipment manufac-
turer) of each process follows:
1. Filtration, simple.
2. Filtration, flocculation/coagulation.
3. Filtration, flocculation/coagulation, oxidation (with and without deionization).
4. Deionization.
5. Reverse Osmosis.
6. Filtration, Centrifugation (with and without deionization).
7. Vacuum Distillation (two units tested).
8. Coolant Life Extension.
Filtration, Simple--This process treated 25-gal (95 L) batches of automobile coolant, which
is normally a mixture of approximately 50% ethylene glycol and 50% water. When 25 gal (95
L) have been accumulated in the machine's "dirty coolant" reservoir, the machine is turned
on and the solution passes through a 5-t~m pre-filter and then through a final filtration, which
uses an ultra filtration filter removing impurities as small as 89 in size. After the solution
has been filtered and is now contained in the clean recycled coolant reservoir, the inhibitor is
added to restore the proper balance of additives. The recirculating pump can be run to speed
the mixing of the inhibitor with the recycled coolant. Next, the recycled, re-inhibited solution
is tested for freeze point, total dissolved solids, and reserve alkalinity.
Filtration, Flocculation, Coagulation--This particular process is conducted on the vehicle
via hook up to the cooling system at the radiator and at the heater hose. An alkaline chemical
flush is poured into the cooling system, and the vehicle is run at idle for approximately 10 rain.
With the vehicle shut off, the recycling begins, with coolant directed by ordinary shop air. The
coolant is drawn into the holding tank through a coarse filter to catch most suspended solids.
While coolant is flowing into the holding tank, special chemicals are added to the holding tank
to precipitate out some of the dissolved species. The coolant is then directed back into the
vehicle through a fine filter to remove the precipitated solids. This procedure takes about 3
rain and is repeated two more times. A proprietary corrosion inhibitor package is then added
to the coolant, followed by a stop-leak chemical to replace any filtered stop-leaks that are rou-
tinely added to the coolant at the factory on many new cars.
The final steps require addition or removal of antifreeze to restore a 50:50 glycol/water mix
for recommended freeze and boil protection and addition of a coolant sealant that replaces
any factory stop-leak removed in the filtration process.
Filtration, Flocculation, Coagulation, Air Oxidation (with and without deionization)--This
process (directly recycled on vehicle) consists of filtration to remove suspended solids, air oxi-
dation, and hydroxide precipitation to remove dissolved metals with the filters. The pH is set
to 9.8 to 10.0 with an antifreeze "extender," and the freeze point is adjusted with 100% pure
ethylene glycol. The process can be seen diagrammatically in Fig. 1. This process was evaluated
with and without ion-exchange resins.
Deionization--This batch or direct vehicle process was designed to purify used coolant by
combining successive stages of suspended solids and dissolved solids removal. Figure 2 is a
simplified flow diagram where the graduated shading depicts the successive stages of purifi-
cation. Filtration begins with a basket strainer (1) removing large particulates such as stop leak
materials that are greater than 115 um in size.
The strainer's main purpose is to protect the system pump (2) and to extend cartridge filter
(3,4) life. Next is a cartridge filter (3), which removes suspended solids greater than 25 #m
followed by a 5-um cartridge filter (4) with a core employing granular 20 40 mesh activated
carbon to absorb organics. This protects the ion-exchange resins from organic fouling, which
renders the resins ineffective. Since oil can be a common engine coolant contaminant, this is
an important design concept.
After filtration in Stages 1, 2 and 3, the coolant, now free of insoluble particulates and heavy
organics, flows into the cation exchange tank (5) through a tube to the bottom of the tank
where it flows evenly upward contacting the ion-exchange resin. All positively charged ions
(cations) are replaced by hydrogen (H ions during this stage. In Stage 6, all negatively
charged ions (anions) are exchanged for hydroxide ions (OH-). The H and OH- ions
USEDANTIFREEZEI )" 3
Suspended Hatter SUSPENDED~ DISSOLVED]
HATTER HATTER
Iron Oxide
Copper Oxide
Calcium carbonate Iron Oxide Nitrite Nitrate
Magnesium Carbonate Copper Oxide Sulfate
Hagoexium S i l i c a t e Tin Oxide Chiorlde
Calcium Phosphate Lead Oxide HBT
Dirt/Silt Zinc Oxide Phosphates
Lead Oxide
Zinc Oxide 1 4 2--~ 5
) FILTRATION[ AERATION I
D i s s o l v e d Hatter I
8 Removes Converts
Iron Suspended Dissolved
Copper Hatter H e t a l s to
Zinc Oxides as
Tin Suspended
Lead Hatter 6
Nitrites
Phosphates ADDITIVE I SUSPENDEDI
Borates , I HATTER
Raises pH
Sulfates Adds Hydroxides,
Chlorides Antifoam and Iron Hydroxide
MBT Inhlbitors Copper Hydroxide
Zinc Hydroxide
Lead Hydroxide
Tin Hydroxide
Magnesium
Hydroxide
Calclum Hydroxide
FIG. l--Filtration, flocculation, coagulation, and air oxidationprocessflow chart.

FIG. 2--Deionization process flow diagram.
exchanged in Stages 5 and 6 combine to form water (H20) in an amount depending on the
number of equivalents (moles of charge) of ions removed from the used engine coolant. The
resulting purified fluid flows into a final activated carbon filter (7) whose purpose is to remove
any entrained gases from the liquid stream.
Finally, the completely purified EG/H20 mixture passes through an in-line conductivity
probe (8), which measures the conductivity of the solution as an indication of the degree of
purification by the ion-exchange system in Stages 5 and 6. The probe is designed to show a
green light when the conductivity of the solution is below 50 #S/cm. This ensures high-quality
effluent and also gives the operator a visual indication that the process is under control and
that the ion-exchange beds have not exceeded their capacity. When the ion-exchange resins
are exhausted, dissolved solids remain in solution and the conductivity of the effluent rises and
quickly exceeds the set point of the probe assembly at 50 t~S/cm. This triggers the green light
offand the red light on indicating to the operator that the ion-exchange tanks must be removed
for regeneration and replaced with a fresh set. At the completion of this process the coolant is
re-inhibited and checked for freeze point and reserve alkalinity.
Reverse Osmosis--This batch oriented process is based upon a reverse osmosis membrane.
Reverse osmosis uses a membrane to remove dissolved solids from the coolant. Unlike con-
ventional filters, a membrane system separates the feed solution into two streams. The per-
meate is the fluid that passes through the semipermeable membrane, while the concentrate
stream is the part of the feed stream that is used to flush the concentrated solid from the system.
One of the most widely known uses of reverse osmosis technology is turning seawater into
drinkable water.
Referring to Fig. 3, the pump feeds used coolant to the membrane housings. Coolant (nearly
pure ethylene glycol and water) is extracted by the membrane within the housing and leaves
the membrane housing in two streams as permeate and concentrate. Permeate is collected at
the permeate manifold on systems where more than one membrane housing is used. One per-
meate line is contained in each membrane housing. The permeate then flows through two flow
meters mounted in series and to the outlet point of the machine. External plumbing carries
this flow to a permeate storage tank. Concentrate leaves the last membrane and is plumbed to
ReverseOsmosis
Cooler
i PreConditioningUnits (
Concentrate Permeate
Tank Tank
--O JI
LowPressure Pump
FIG. 3--Reverse osmosisflow diagram.
a globe needle valve that controls the pressure within the machine. The concentrate flow con-
tinues through a flowmeter and a permeate outlet port. External plumbing carries the concen-
trate back to the feed tank. The permeate is then adjusted for freeze point and re-inhibited.
Filtration, Centrifugation (with and without deionization)--This process is utilized as a
drive-up service to recycle engine coolant on-site. It utilizes the vehicle's engine for the power
source and is completely self contained. This process combines a sophisticated series of filters,
a thermo-processing operation, a centrifuge system, and a monitored precision additive sys-
tem that ensures the precise quantifies of specially formulated chemical additives are injected.
Referring to Fig. 4, the engine coolant to be processed is evacuated from a storage container
containing the coolant designated to be recycled, by a vacuum system in the vehicle. The cool-
ant to be recycled can be stored in any size container, the system can evacuate the coolant from
as far away as 100 ft (30.5 m). The engine coolant is drawn through a series of filters reducing
the particulates and contaminates. The engine coolant is then heated to reduce its viscosity
thereby allowing for a more thorough centrifuge separation of the contaminates from the cool-
ant. After the centrifuge, the coolant is immediately cooled to facilitate handling. The coolant
is then injected with the specially formulated additives. As the coolant is processed and flows
through the various operations, it is continually monitored, and tests are performed before,
during, and after the service.
Vacuum Distillation--Two systems were utilized to evaluate this process. The mechanics
of vacuum distillation are straightforward. The coolant to be processed is poured into a boiling
chamber. The chamber is then sealed. Heat is then applied to the chamber and, according to
basic physical principles, the liquid with the lower boiling point (water) begins to boil out of
the chamber first. Exiting the chamber as a vapor, the vapor is directly piped into a heat
exchanger where a cooling affect is enacted. The cooling that occurs in the heat exchanger
FIG. 4--bu centrifugation processflow diagram.
condenses the vapors (lowers the temperature below the boiling point) transforming them
back into a liquid. The liquid is then allowed to flow out of the exchanger and into suitable
collection vessels to contain either the recovered water or ethylene glycol.
The differences between the two units evaluated consisted of process rates and usability. One
system processed coolant at a rate of 5 gal/h (19 L/h) while the other was at a rate of 1 gal/h
(4 L/h). One system offered an automatic version, which mixed the water and glycol to a pre-
determined set point and added the inhibitor package as well.
Coolant Life Extension--This system was developed to extend the useful life of automotive
engine coolant. It includes qualification of the coolant by chemical, physical, and visual
means; and extension of the coolant life by filtering and the addition of a supplemental coolant
additive. Because of the criteria of the qualification tests (mileage < 65 000 km, coolant not
to be oily/murky/rusty, glycol level and reserve alkalinity not < 25% or < 3, respectively) the
reference coolant did not qualify for extension, and therefore this process was not evaluated.
Off-Site Coolant Recycling

There are several facilities in the United States that retrieve spent coolant from service facil-
ities and auto-dealerships for the sole purpose of recycling for reuse as engine coolant. These
facilities typically utilize fractional distillation methods although not exclusively. These recy-
cling processes had to be dealt with differently than the smaller on-site systems for reasons such
as the significantly larger amounts of coolant needed for processing in each run.
Off-Site Approval Program

In order to determine and maintain the efficacy of off-site coolant recycling, an approval
program has been established at General Motors. Letters were distributed to all off-site coolant
recycling facilities regardless of the type of process they utilize. The program requires a sub-
mitting facility to provide quality control information to ensure a stable process and trade-
names of the recycled product along with a list of the outlets where the coolant can be pur-
chased. With this information, finished samples of the recycled product can be acquired by
General Motors and sent to a test facility for compliance to GM 1825M, the same criteria used
in the on-site tests. Since this approval program is intended to be an ongoing effort, submitting
facilities are required to cover the expense of testing. Subsequent retesting will be conducted
to ensure continued compliance. Facilities may lose their approved status if they do not con-
tinue to meet the program's requirements.
The objective of this program is to provide to General Motors dealers a list of all the
facilities capable of providing a recycled coolant that meets or exceeds G M 1825M coolant
specifications.
Laboratory Testing of Recycled Coolant

Upon testing the recycled coolants, the ASTM test procedures found within the GM 1825M
test specification were conducted in a pre-arranged order. As described in Society of Auto-
motive Engineers (SAE) Report J814, special tests are required to evaluate the performance
of coolants with regard to specific forms of corrosive attack. These stepwise procedures are
used to avoid the unnecessary expenditure of time and money performing long expensive tests
on obviously poor coolants, and to ensure that better coolants will meet service requirements,
by the use of more rigorous test conditions. Each of the recycled coolants were tested for com-
pliance of ASTM Method for Corrosion Test for Engine Coolants in Glassware (D 1384),
ASTM Test Method for Corrosion of Cast Aluminum Alloys in Engine Coolants Under Heat-
Rejecting Conditions (D 4340), and chemical/physical tests. If however, the recycled coolant
failed to meet the requirements of either D 1384 or D 4340, it was suspended from the test
program, and no further testing was performed. Those recycled coolants that met the require-
ments of both D 1384 and D 4340 were submitted for ASTM Test Method for Caviation Ero-
sion-Corrosion Characteristics of Aluminum Pumps with Engine Coolants (D 2809) and
ASTM Method for Simulated Service Corrosion Testing of Engine Coolants (D 2570). For
additional information, the radiators and post test coolants were sent for analysis at the con-
clusion of the simulated service test.
Product Confirmation
In order to assure that the inhibitor formulation used for each recycling process in this test
program was identical to that currently utilized in the field, samples of the inhibitor packages
were acquired from typical field outlets from each of the processes that had successfully com-
pleted D 1384 and D 4340. The acquired inhibitor packages were analyzed and compared to
the analysis of the respective recycled coolant inhibitor formulation. The analysis of the inhib-
itors acquired over the counter had to agree with the inhibitor used in the test program.
To ensure that the performance of each process which meets the pre-described criteria
remains unchanged in the field after the completion of this test program, periodic analysis of
inhibitor products will be conducted and compared to previous analysis on file. No changes
will be allowed in the mechanics of the process or the inhibitor formulation. Furthermore, in
order to maintain an approved status, there must not be available alternate inhibitor packages
(GM or non-GM facilities).
DISCUSSION ON ENGINE COOLANT RECYCLING PROCESSES 287
Conclusions
The purpose o f this paper has been to describe the coolant recycling study conducted at STG
for the benefit of any who desire to conduct such a test in the future. Also, to provide insight
to the results of this study that were incomplete during the writing of this paper. Upon com-
pletion, the results will be discussed in the second part of this paper.
References
[1] Benyon, E., Cooper, N. R., and Hannigan, H. ]., "Cooling System Corrosion in Relation to Design
and Materials," Engine Coolant Testing: State of the Art, STP 705, W. H. Ailor, Ed., American Soci-
ety for Testing and Materials, Philadelphia, 1980, p. 312.
[2] Vukasovich, M. S. and Sullivan, F. J., Inhibitors and Coolant Corrosivity, Engine Coolant Testing.
STP 887, R. E. Beal, Ed., American Society for Testing and Materials, Philadelphia, 1986, p. 96.
[3] Tighe, K. P., Heavy Metal Concentrations are of Concern, NARSA Forum on Coolant Disposal and
Recycling, March 1989.
DISCUSSION
Frank A. Kelly ~ (written discussion)--Have you considered providing your dealer organi-
zation any economic assessments of the approved recycling processes?
Wayne H. Bradley (author's response)--Some of the recycling equipment that has been
approved is in the General Motors dealer equipment program. Those manufacturers that are
in the program have made available cost comparison data.
It should be noted however that the actual operating cost is variable and largely dependant
upon the equipment's application. The most significant factor being the average condition
(that is, total solids content) of the used coolant typically recycled. The more solids involved
in the recycling process, the quicker depletion occurs.
Paul Dittmar 2 (written discussion)--(1) What recycling processes passed the test? (2) What
is your opinion about the capability of the processes that passed for maintaining product qual-
ity control on an on going basis?
Wayne H. Bradley (author's response)-- 1. The following equipment has been endorsed by
General Motors for coolant recycling: A C T I V - - v a c u u m distillation, Finish Thompson--vac-
uum distillation, BG--deionization, and American Fluid Technology--Centrifugation.
2. There are several post-test measurements incorporated into each recycling procedure to
assure that the process was successful. Also, variability (in the coolant to be recycled) was
accounted for within the diverse composition of the reference coolant used in this evaluation,
This, accompanied with an ongoing educational effort, will help ensure a successful recycling
program.
Gerry Smith 3 (written discussion)--(1) Of the seven types of processes mentioned as being
tested, which were found to be effective? (2) Were any methods found to be effective in extend-
ing coolant life? If so, what chemistry and/or processes worked?
Wayne 11. Bradley (author's response)-- 1. The following processes were endorsed by Gen-
eral Motors for engine coolant recycling: A C T I V - - v a c u u m distillation, Finish T h o m p s o n - -
vacuum distillation, BG--deionization, and American Fluid Technology--centrifugation.
Caterpillar, Inc., Technical Center Building E, P.O. Box 1875, Peoria, IL 61656-1875.
2 Safety-Keen Corp, 12555 W. Old Higgins Rd., Elk Grove Village, IL 60007.
3 Gold Eagle Company, 4400 S. Kildare, Chicago, IL 60632.
2. The concept of coolant life extension was included in this evaluation. However, because
of their coolant qualifications criteria and the characteristics of the reference coolant used in
the study, coolant life extension was not evaluated.
Joseph A. Lima 4 (written discussion)--With the plethora of on-site systems, the attendant
problems of maintenance and additive control and the apparent success of on-board recycling
on heavy duty equipment, I am wondering why the automobile manufacturers are not install-
ing on-board filtration systems and including coolant system evaluations at their programmed
maintenance intervals, such as 30 000, 60 000, and 90 000 miles (48 279, 96 558, and 144 837
km).
IfI may comment, it would seem to me that this would offer the consumer a more effective
and supervised alternative to present on-site capability.
Beyond this, there will be eventual need to recycle extended coolant, and I am wondering
what evaluations have been made of central off-site recovery systems, where large batch sizes
would allow for economically justifiable testing to ensure that the quality desired is met before
the material being offered for re-use in vehicles.
Wayne 11. Bradley (author's response)--On-board coolant treatment similar to heavy duty
applications has been considered for use in automotive applications. The difficulty in imple-
menting this approach in automotive applications is the difference in maintenance techniques
from those practiced in heavy duty fleets. Regular additions of SCAs and coolant change out
can, and is, controlled in heavy duty fleet applications. In automotive applications however
coolant service is not controlled to this degree, and therefore this concept is not feasible.
To address off-site coolant recycling facilities, General Motors has established an approval
program. Within this program, coolant recycling facilities may submit pertinent information
about their process (that is, quality control data) as well as the names of outlets where their
recycled product can be purchased. GM personnel will acquire samples of the product and
submit them for test at the submitters expense. Providing both the performance and process
requirements are met, the facility will be endorsed. Periodic retesting will occur to ensure prod-
uct quality.
4 Houghton Chemical Corporation, Allston, MA.

F. A. Kelley I
Heavy Duty Diesel EngineCoolants"

Technology Update
REFERENCE: Kelley, F. A., "Heavy Duty Diesel Engine Coolants: Technology Update,"
Engine Coolant Testing: Third Volume, ASTM STP 1192, R. E. Beal, Ed., American Society for
Testing and Materials, Philadelphia, 1993, pp. 289-296.
ABSTRACT: A summary of current heavy duty engine coolant technology is provided which
includes a description of heavy duty engines and applications, current problems and concerns
which face the industry, industry standard specifications and test methods which are under devel-
opment to address the concerns, and future directions for the technology. The special needs of
heavy duty engines dictate that engine coolant formulations be developed and made available
that are tailored to these engines. Efforts to improve fuel economy and durability while reducing
emissions and noise will have an important impact on the application of engine coolant tech-
nology to heavy duty diesel engines in the future.
KEYWORDS: heavy duty engine coolants, liner pitting, cavitation corrosion, coolant recycling,
water pump seals, supplemental coolant additives, coolant maintenance, coolant test methods,
water quality
Heavy duty diesel engines have been the work horse of industry for more than 60 years.
Through those years, technology has been developed continually to meet demands for dura-
ble, cost-effective power systems. Current demands of diesel engine technology are somewhat
different than they may have been in the past, however. Today's manufacturers continue to
work on engines that provide better fuel economy, better durability, and better reliability, but
much more emphasis is placed on improving the environmental impact of diesel engines with
significant efforts under way to reduce emissions and reduce noise. Cooling systems of heavy
duty diesel engines, and the coolants used in them, are increasingly becoming critical com-
ponents whose importance cannot be overlooked if all of the demands facing the industry are
to be met.
Traditionally, engine coolants developed primarily to suit the needs of automotive engines
have been used in heavy duty diesel engine applications. However, for many years now, it has
been recognized that significant differences exist between the two types of engine technology,
requiring somewhat different treatment for satisfactory results.
Some facets of heavy duty diesel engine technology will be characterized, specific engine
coolant-related problems and concerns that face the industry will be discussed, efforts within
the ASTM D-15 Engine Coolants Committee to establish suitable standards to address the
specific needs of the industry will be highlighted, and some comments about future trends in
the technology will be provided in this article.
t Project Manager, Advanced Materials Technology Div., Research Dept., Caterpillar Inc., Peoria, IL.
289
290 E N G I N ECOOLANT TESTING: THIRD VOLUME
Heavy Duty EngineApplications

Heavy duty diesel engines are used in a number of demanding applications (Table 1). The
popularity of diesel engines for these applications is due to a reputation for excellent fuel econ-
omy, high reliability, long service life, and high sustained torque when operating at low speeds.
These advantages generally far outweigh the disadvantages of higher initial cost, somewhat
heavier weight, and operational roughness. Diesel engines come in many shapes and sizes.
They can deliver anywhere from a few kilowatts to more than 15 000 kw (20 000 horsepower)
in large marine engines. Since the overwhelming majority of diesel engines produced in the
United States are used in the on-highway truck, construction equipment, and farm machinery
applications, most of this discussion will relate to the use of diesel engines in these applications.
The typical heavy duty diesel engine used in on-highway truck, construction, and farm
machinery applications is in the 110 to 330 kw ( 150 to 450 horsepower) range. Cooling system
sizes typically range from 20 to 80 L (5 to 20 gallons).
A characteristic that is c o m m o n to many of these engines is the use of a wet sleeve cylinder
liner in the design. This, coupled with the high pressures that are characteristic of the diesel
combustion process and the strong tendency for severe piston impact on the liner, leads to the
occurrence of liner pitting. This problem will be discussed in more detail in the next section.
Several characteristics of diesel engine applications are important to consider that have an
effect on the type of engine coolant formulation and maintenance strategy that is best suited
for use. Today's typical diesel engines are primarily cast iron systems with copper and brass
used in radiators and oil coolers. This contrasts with the extensive use of aluminum in many
automotive engines. Heavy duty diesel engines are expected to operate for better than 800 000
km (500 000 miles) before requiring a major overhaul. It is possible, with the use of good qual-
ity engine coolant and oil formulations and adherence to recommended maintenance prac-
tices, to reach 1 600 000 km (1 000 000 miles).
Heavy duty diesel engines are often operated with high load factors and with high rates of
use in comparison to automotive engines. It is not uncommon for a heavy duty on-highway
truck to log more than 240 000 km ( 150 000 miles) in a year. Many heavy duty trucks travel
routes that stretch from coast to coast and include both flat and mountainous terrain. Con-
struction equipment in some mining applications is operated nearly continuously over three
shifts. In most applications, downtime is not appreciated since this amounts to a loss in reve-
nue for the equipment owner.
The environments in which heavy duty diesel engines must operate have an impact on cool-
ing systems. They must be able to function in cold, arctic climates as well as hot, tropical loca-
tions. Care must be taken not to overconcentrate the cooling system with glycol in obtaining
the necessary freeze protection for arctic applications as this will lead to problems with additive
drop-out. A 60 vol.% concentration of ethylene glycol concentrate is considered to be the high-
TABLE 1--Heavy duty engine applications.

9 On-Highway Truck
9 Construction Equipment
9 Farm Machinery
9 Railroad
9 Marine
9 Electric Power Generation
9 Industrial
KELLEY ON HEAVY DUTY DIESEL ENGINE COOLANTS 291
est practical level for use as this level will provide the maximum practical freeze protection for
ethylene glycol-water mixtures. In warm climates, it is still common for water with an appro-
priate concentration of supplemental coolant additive (SCA) to be used in heavy duty engine
cooling systems. This practice will probably become less acceptable as engine operating tem-
peratures increase and aluminum components find their way into applications in heavy duty
diesels in the future.
Current engine coolant technology requires the use of SCAs in combination with a good
quality automotive engine coolant that is low in silicates to achieve optimum results. SCA
formulations provide liner pitting protection, and some improved capability to control scaling
of heat transfer surfaces. True heavy duty engine coolants are beginning to appear in the mar-
ketplace, eliminating the need for the addition of an SCA at initial fill, but still require main-
tenance dosages of SCA at regular intervals. These maintenance dosages are applied through
the addition of specified quantities of liquid SCA or through filter changes containing SCA in
solid form.
The need to regularly replenish inhibitors in heavy duty engine applications can be attrib-
uted to high use rates resulting in depletion of key inhibitors in engine coolant formulations
prior to the exhaustion of the ethylene glycol itself. It would not be economical to replace the
inhibitors by draining the engine coolant and filling with a fresh charge of coolant. The accu-
mulation of spent inhibitors and regular additions of SCA eventually leads to the need to
replace the entire coolant in the system due to the increase in total dissolved solids that results.
Typical heavy duty engine change interval recommendations are in the range from 160 000
to 320 000 km (100 000 to 200 000 miles) or 1 to 2 years of service.
Current Problems and Concerns

Liner Pitting
Perhaps, the most significant coolant-related engine problem experienced by heavy duty
engine manufacturers is liner pitting. Liner pitting has been known to occur for many years
in diesel engines [1 ]. It is the result of cavitation and erosion processes that occur on the water
side of diesel engine cylinder liners. Cylinder liners vibrate from the motion of the piston
within the cylinder. This vibration creates pressure variations in certain regions of the fluid
causing vapor bubbles to form and collapse in concert with the vibrating surface of the liner.
This produces a cavitation effect where collapsing bubbles violently impinge on the liner sur-
face, eroding away material. Cavitation-erosion produces increased corrosion of the cast iron
by removing protective films that leave the exposed surface susceptible to a rapid, localized
corrosion attack.
This corrosion can be arrested by the presence of very rapid film forming agents in the
engine coolant. Initially, additions of relatively high concentrations of sodium dichromate cor-
rosion inhibitor to the cooling system were found to be effective in reducing or eliminating the
damage. Chromates are no longer used and today's formulations for the protection of heavy
duty engines from liner pitting contain sufficient levels of sodium nitrite, or combinations of
sodium nitrite with sodium molybdate, to control the problem. These film formers are not
normally present in most automotive technology engine coolants.
Silicate Drop-out
Another problem of heavy duty diesel engines is that of silicate drop-out. Silicates have
become very popular in automotive engine coolant formulations because of their effectiveness
in controlling aluminum corrosion in high heat transfer areas of automobiles [2]. Silicates can
be relatively unstable in solution and are prone to precipitate in some situations, such as when
the total dissolved solids level is high, high hardness water is added to the mixture, or glycol
concentration is high. This problem is further complicated by the need to use an SCA for heavy
duty engine applications because this adds more solids to the system and, usually, more sili-
cates as well. The tendency for silicate drop-out to occur has led to recommendations in the
heavy duty engine industry to use engine coolant formulations containing low levels of silicate
corrosion inhibitor to produce minimal interference with SCA additions.
Water Pump Seal Leakage

Water pump seal leakage is a c o m m o n problem in heavy duty applications. The seals used
in heavy duty engine water pumps are generally similar in design to those used in automotive
applications. However, they are plagued more often by problems caused by silicate drop-out
and high levels of total dissolved solids in heavy duty engine systems. Water pump seal leakage
problems are usually difficult to resolve because there are so many contributing variables. In
order to minimize the likelihood of water pump seal failure in heavy duty engines, it is nec-
essary to avoid high levels of silicate in the engine coolant, maintain low total dissolved solids
levels, and to keep glycol concentrations below 70%.
Solder Bloom~Solder Corrosion

Radiators and oil coolers on heavy duty engines must withstand higher levels of stress,
shock, and vibration than their automotive counterparts. For this reason, corrosion of solder
can be a bigger problem in heavy duty engine applications. There are two types of solder cor-
rosion problems. Solder bloom is a solder corrosion product that forms as a highly voluminous
deposit. The make-up of the deposit is generally low in lead but contains high levels of other
materials that have been codeposited with the lead from the engine coolant. The solder bloom
process is not thought to produce weakened solder joints in most cases, but does often lead to
overheating problems by closing offthe ends of the tubes to restrict coolant flow. The potential
for solder bloom to occur is reduced by assuring that solder fluxes are thoroughly cleaned from
any new or rebuilt system and that only good quality engine coolant and water is added to the
cooling system. Other forms of solder corrosion more directly attack the solder joint, leading
to failure of the joint and leakage. One aggravator to solder corrosion is the nitrite corrosion
inhibitor used to protect against liner pitting. If the cooling system is overdosed with SCA, the
resulting high nitrite levels can lead to increased solder corrosion. For this reason, it is neces-
sary to maintain SCA levels within the operating range specified by the engine manufacturer
and SCA supplier.
Scaling
Scaling of high heat transfer surfaces is particularly troublesome in heavy duty engines.
Coolant loss is more c o m m o n in heavy duty engine applications. This leads to frequent addi-
tions of water to the cooling system. If high hardness water is added to the system, the potential
for producing scale deposits on the liners is increased. Scale deposits on the coolant side insu-
late the liners and affect the heat transfer from the cylinder. This can lead to higher operating
temperatures in the ring belt zone and an increase in piston deposits which, in turn, can cause
premature wear-out on the oil side of the cylinder kit. To minimize the problem it is recom-
mended that good quality water be used. Most SCA formulations provide some additives
designed to counteract the effects of scaling.
Recycling
Environmental concerns have produced more awareness for the need to recycle used engine
coolants. A significant number of approaches to engine coolant recycling have surfaced in
recent years. Many of the approaches being marketed are not suitable for the purpose of recy-
cling engine coolant to be used in heavy duty engines because they do not remove the dissolved
solids from the system. The primary reason for draining engine coolant from a heavy duty
engine is to remove contaminants that have accumulated in the system during use. These con-
taminants include corrosion products, spent inhibitors, oxidized glycol, hard water salts, oil,
etc. Many of these contaminants are soluble in the coolant and are not removed by filtration
or simple ion exchange. It is absolutely imperative that recycled coolant for heavy duty engine
usage have these contaminants removed prior to reinstallation in the system. One process that
has been evaluated and has been found to be acceptable for the purpose of recycling used
engine coolant for heavy duty engine usage is distillation. It is very important to follow the
engine manufacturer's recommendation for recycled coolants as improper practices will lead
to premature engine failure.
Maintenance
As evidenced, proper maintenance is critical for the successful operation of heavy duty
engines. Either a low silicate engine coolant formulation with the appropriate amount of SCA
or a true heavy duty engine coolant formulation incorporating the initial dosage of liner pilling
and scaling protection should be used. Maintenance dosages of SCA should be added at rec-
ommended intervals to replenish depleted inhibitors. And good quality water should be used.
Suggested limits for water appropriate for use in heavy duty engine cooling systems have been
developed (Table 2).
Standards Under Development

Most automotive engine coolant formulations fall short of providing adequate performance
in the typical heavy duty diesel engine. The ASTM D- 15 Engine Coolants Committee has been
very active in seeking to develop new standards, test methods, and specifications to support
the needs of the heavy duty engine industry. One heavy duty engine specification has been
introduced and several other specifications and test methods are under development.
The one specification that is now in place is ASTM D 4985-89 Standard Specification for
Low Silicate Ethylene Glycol Base Engine Coolant for Heavy Duty Engines Requiring an Ini-
tial Charge of Supplemental Coolant Additive (SCA). It is intended to provide the heavy duty
TABLE 2--Recommended water quality limits, a
Property Specific Values ASTMTest Method

Total Solids, ppm 340 max. D 1888
Total Hardness, ppm 170 max. D 1126
Chlorides, ppm 40 max. D 512, D 4327
Sulfates, ppm 100 max. D 516, D 4327
pH 5.5-9.0 D 1293
a Recommended by Engine Manufacturers' Association cool-
ants subcommitteebased on a survey of service recommendations
of North American heavy duty diesel engine manufacturers.
industry with a standard engine coolant that minimizes the problem of silicate overconcen-
tration and drop-out that began to cause significant difficulties in the early to middle 1980s.
Heading the list of new heavy duty engine coolant specifications under development is the
supplemental coolant additive specification. The heavy duty engine industry has relied on the
use of SCAs. The new specification under development will provide an industry approved
standard for those existing products which conform to the informal standard that has devel-
oped over the years and provide engine manufacturers with a better means for describing to
their customers those products that meet the necessary requirements.
To address the difficulties of obtaining suitable water meeting suggested quality limits, a
specification will be developed for prediluted engine coolants, which are blended with an
appropriate quantity of ethylene glycol for freeze protection and water meeting very stringent
quality standards. The initial version for heavy duty engines will be based on the existing
ASTM D 4985 low silicate engine coolant specification.
The current movement toward increased recycling of used engine coolant will not be effec-
tive until appropriate specifications are developed defining what is expected of the recycled
product. As mentioned previously, it will be very important for recycled coolants intended for
use in heavy duty engines to have sufficient amounts of contaminants removed such that the
resulting product does not interfere with SCA additions or water pump seal performance. Cor-
rosion performance capability will be required to match that of products based on virgin
materials.
There is a need for a specification that is based on formulations tailored specifically to the
needs of heavy duty engines. A "true" heavy duty engine coolant formulation simplifies the
installation of new engine coolant by not requiring the addition of an SCA at initial fill. This
avoids the inefficiency of combining two separate inhibitor systems during the initial fill and
compatibility problems that can result. It provides an engine coolant mixture that has a
reduced total dissolved solids level at initial fill. These formulations currently still require the
addition of maintenance dosages of SCAs at periodic intervals after the initial fill.
It will be important to develop appropriate bench tests for inclusion in each of the specifi-
cations that are under development. Currently, acceptable performance in the critical problem
areas discussed previously is mainly established through full-scale engine testing. Specifica-
tions based on full-scale engine tests would tend to reduce the participation of many potential
suppliers who would have difficulty justifying the cost of demonstrating conformance for their
engine coolant or SCA products. Formulation specifications that define some or all of the com-
position of the conforming products provide only a partial solution because they may preclude
some products that otherwise may be proven acceptable through proper engine dynamometer
or field testing.
Four bench test procedures have been proposed for development within the ASTM D- 15.11
Subcommittee on Heavy Duty Engine Coolants to provide the necessary performance defi-
nitions desired for heavy duty engine applications. A vibratory cavitation bench test based on
the use of an ultrasonic horn has been chosen as the method for demonstrating acceptable liner
pitting protection. Methods based on this approach have been shown to discriminate between
formulations that perform acceptably in field experience to provide adequate liner pitting pro-
tection and those which perform unacceptably in this regard [3]. An industry standard test for
evaluating the tendency of engine coolant formulations or SCAs to produce water pump seal
leakage problems is also under development [4]. The third bench test under development
addresses the issue of scaling of heat transfer surfaces and is expected to determine the scaling
protection provided by formulations intended for heavy duty engine applications. In this test,
the test fluid composition containing hard water salts is flowed past a heated iron surface which
is then weighed to determine the extent of scaling. Lastly, a test that demonstrates acceptable
compatibility between engine coolant and SCA products is being developed. In this case, the
combined mixture is centrifuged and any precipitated material that collects is measured.
Future Directions
In the coming years, diesel engine technology will experience significant changes as engine
manufacturers work to develop engines that meet new regulations for exhaust emissions and
noise, while seeking improvements in fuel economy, service life, and ease of maintenance.
Many of the approaches to future heavy duty engine technology will result in higher operating
temperatures. It will be a major challenge to achieve these goals, and engine cooling systems
will play a key role in determining ultimate success.
Future emphasis on improved fuel economy and reduced costs will most likely bring more
usage of aluminum into the picture. However, increased aluminum usage in heavy duty
engine cooling systems will occur only if cooling system designs and engine coolant formula-
tions that demonstrate acceptable reliability and durability can be devised. Aluminum radia-
tor technology will have to demonstrate significantly improved durability and serviceability in
order to be of use in heavy duty engine cooling systems. Engineers will need to be certain that
the application of any aluminum component in a heavy duty engine cooling system is com-
patible with expected maintenance practices and available engine coolant technology for the
engine.
Solder corrosion problems have led engineers to search for alternative methods of improv-
ing the robustness of heavy duty radiator systems. Also, manufacturers are looking for ways
to reduce the usage of lead in their products for environmental considerations. The two
approaches that are receiving significant attention are the development of joint designs that
greatly reduce the exposure of the solder to the engine coolant and development of alternative
solders. Zinc based solder development is an example of the latter approach.
Alternative engine coolant fluids have received a significant amount of scrutiny. Most sig-
nificantly, propylene glycol-based engine coolants have received much attention. Although
propylene glycol has been used for many years during periods of short ethylene glycol supply,
only recently has significant interest been placed on the prospects for taking advantage of the
specific properties of propylene glycol as an engine coolant fluid of choice. The advantages
most often cited include reduced toxicity in comparison to ethylene glycol and improved cor-
rosion performance, particularly in the case of aluminum. Some investigators have reported
that propylene glycol-based engine coolant formulations can provide improved protection
against liner pitting [5] compared to their ethylene glycol counterparts. The effect of the pro-
pylene glycol is thought to be primarily physical in that the intensity of the bubble collapse
might be reduced.
Although most of the emphasis on propylene glycol is on traditional aqueous-based systems,
much interest has been generated for the development of anhydrous propylene glycol based
systems [6]. This approach takes advantage of the wide fluid temperature range of - 6 0 to
187"C ( - 7 6 to 369~ that propylene glycol offers. This allows fluid systems to be designed
requiring no water for freeze protection and no pressure for boiling point elevation. Interest
in heavy duty engines comes primarily from the ability to operate at higher temperatures or
the possibility for reducing radiator frontal area for streamlining vehicles ira on-highway truck
applications.
The prospect that coolant formulations can be developed for heavy duty engine applications
that have extended protection capability without the need to rely on SCA additions at initial
fill or for inhibitor replenishment has begun to garner a significant amount of attention. Long-
life engine coolant formulations based on organic acid inhibitor chemistry are representative
of one approach to this type of technology [ 7]. Anhydrous fluid technology, such as the pre-
viously mentioned propylene glycol-based technology or oil-based technology, are other pos-
sible ways of achieving this desirable goal. A true heavy duty engine coolant with long-life
capability would have significant value to the user of heavy duty engine technology.
References
[1] Speller, F. N. and LaQue, F. L., "Water Side Deterioration of Diesel Engine Cylinder Liners." Cor-
rosion. Vol. 6, No. 7, 1950, pp. 209-215.
[2] Wiggle, R. R., Hospadaruk, V., and Tibaudo, F. M., "Corrosion of Cast Aluminum Alloys under
Heat-Transfer Conditions," SAE Technical Paper 810038, 1981.
[3] Hudgens, R. D., Carver, D. P., Hercamp, R. D., and Lauterback, J., "Refinement of the Vibratory
Cavitation Erosion Test for the Screening of Diesel Cooling System Corrosion Inhibitors," Engine
Coolant Testing. State Of The Art. ASTM STP 705, W. H. Ailor, Ed., American Society for Testing
[4] Sehnal, J. F., "Development of Test Stand to Measure the Effect of Coolant Composition on Engine
Coolant Pump Seal Leakage," SAE TechnicalPaper 900808. 1990.
[5] Hercamp, R. D., Hudgens, R. D., and Coughenour, G. E., "Aqueous Propylene Glycol Coolant for
Heavy Duty Engines," SAE Technical Paper 900434, 1990.
[6] Evans, J. W. and Light, J. T., "Expanding the Limits on Engines and Vehicles Imposed by Circulating
Liquid Engine Cooling Systems," SAE Technical Paper 861428, 1986.
[7] Darden, J. W., Triebel, C. A., Maes, J. P., and VanNeste, W., "Monoacid/Diacid Combination as
Corrosion Inhibitors in Antifreeze Formulations," SAE TeehnicalPaper 900804, 1990.
AUTHOR INDEX
Index Terms Links
Adamowicz N. C. 63
Beal, R. E., editor 1 83

Bradley, W. H. 276
Burns, J. M. 25
Bustamante, R. B. 149
E1-Bourini, R. 83
Falla, D. F. 63
Gershun, A. V. 234
Hannigan, H. J. 6
Hemmes, P. R. 165
Hercamp, R. D. 107 165 248
This page has been reformatted by Knovel to provide easier navigation.

Index Terms Links
Hirabayashi, H. 215
Hirata, O. 215
Hudgens, R. D. 128 149
Kelley, F. A. 289
Kiryu, K. 215
Kreiser, T. H. 165
Maes, J.-P. 11
McKenzie, D. A. 180
Mercer, W. C. 44
Okada, K. 215
Remiasz, R. A. 248
Richardson, R. C. 258
Stafford, R. J. 205
Valle, S. 165
Van de Ven, P. 11

Index Terms Links
Woodward, S. M. 234
Woyciesjes, P. M. 190
Yoshino, A. 215
Zamachek, W. 180

SUBJECT INDEX
Index Terms Links
AA7072 clad AA3003 alloy 83

Acute Lethal Dose (LD50), PG antifreeze
and SCAs 149
Alkalinity, reserve
depletion by glycol degradation 234
oxalic acid cleaning effects 190
test strips for 165
in used coolant, statistical analysis 234
Aluminum
corrosion protection of heat transfer
surfaces 11
protection with phosphate-molybdate
SCAs 128
radiators, sebacic acid coolant effects 63
Aluminum alloys
carboxylic-acid corrosion inhibitors 44
radiators, corrosion testing 83
Antifreeze, see Coolant/antifreeze
formulations
ASTM standards
D 1121 234
D 1287 234

Index Terms Links
ASTM standards (Cont.)

D 1384 11 44 63
83 128 248
276
D 2272 258
D 2570 248 276
D 2809 44 276
D 2847 63 190
D 3306 25 44 128
190 276
D 4340 11 25 44
63 248 276
D 4985 25 289
G 85 83
standards under development 289
Automotive radiators, see Radiators
Bench tests, for comparative coolant studies 25

Biodegradability, ethylene and propylene
glycols 149
Borate conditioning 190
Calcium-iron-phosphate complexes on seal

faces 205
Carboxylic acid
cooling formulation 25
corrosion inhibitors 44
Index Terms Links
Cavitation corrosion
cylinder liners 165
diesel cylinder liners 107
heavy duty engines 289
history of 6
phosphate-molybdate SCAs for diesel
engines 128
theories of 107
Centrifugation, of coolants 276
CERCLA (Comprehensive Environmental
Response, Compensation, and
Liability Act) 149
Chloride, test strips for 165
Clean Air Act Amendment of 1990 149
Closed-loop coolant flushing system 248
Comprehensive Environmental Response,
Compensation, and Liability Act
(CERCLA) 149
Contaminants
in engine coolants 149
removal with multi-stage purification
process 258
Coolant/antifreeze formulations
carboxylic acid corrosion inhibitors in 44
comparative, fleet test of 25
heavy duty 248 289
ICP and LA-ICP applications 180
LD50 of PG antifreeze and SCAs 149
life extension of 248 276
light duty 248
Index Terms Links
Coolant/antifreeze formulations (Cont.)

SCA effects 128
and sealing performance 215
silicate-based 44
test methods 289
Coolant pumps, see also Water pumps
bench test for 25
failure, fleet tests 25
Coolant toxicity, see Toxicity
Cooling systems
design history 6
oxalic acid cleaning 190
Copper, deposition on sealing surfaces 215
Corrosion inhibitors
carboxylic acid 25 44
and coolant life extension 248
coolant recycling processes evaluation 276
history of requirements for 6
phosphate-molybdate SCAs 128
process 258
and seal performance 215
sebacic acid extended life coolant 63
traditional, pump failure and 25
Corrosion protection, aluminum surfaces 11
Corrosion testing, radiators
aluminum alloys 83
electrochemical methods for 83
Index Terms Links
Cylinder heads, design history 6

Cylinder liners cavitation protection
diesel engines 107
pitting 107
test strips for 165
1, 10-Decanedioic combination coolants 44

Degradation acids, in used coolant 234
Deionization, of coolants 276
Depletion rates, sebacic acid and silicate
phosphate coolants 63
Diacids, monoacid/diacid inhibitors 11
Dibenzothiazyl disulfide deposition 215
Diesel engines
coolant technology update 289
cylinder liners, cavitation corrosion 107 128 289
water pump seal deposits 205
Diethylene glycol, in used coolant, statistical
analysis 234
Disposal, coolants and additives 149
Dual-bed deionization-based multi-stage
coolant purification 258
Dynamometer test method 6
Electrochemical corrosion testing, radiator

materials 83
Index Terms Links
Electron spectroscopy for chemical analysis,

comparison with LA-ICP sample
analysis 180
Environmental concerns, toxicity and
disposal of coolants 149
Environmental Protection Agency,
hazardous waste characterization
guidelines 234
EP Tox (extraction procedure toxicity test) 234
Ethylene glycol
biodegradability 149
toxicity and disposal 149
Ethylene glycol coolants, history of 6
2-Ethylhexanoic acid coolants 44
Extended life coolants
disposal 149
sebacic acid inhibitor technology 63
Extraction procedure toxicity test (EP Tox) 234
Face deposits, water pump 205

Ferrous oxalate, formation chemistry 190
Filling, of coolants 248
Filtration, of coolants 248 276
Filtration-flocculation-coagulation, of
coolants 276

Index Terms Links
Fleet tests
carboxylic acid coolants 44
coolant pump failure 25
oxalic acid cleaning carryover effects 190
sebacic acid coolants 63
silicate phosphate coolants 63
Flushing, of coolants 248
Fourier transform infrared spectrometry,
protective films 11
Freezepoint, test strips for 165
FTIR, see Fourier transform infrared
spectrometry
FVV Test (German) 128
Gas chromatography, carboxylic-acid

German FVV Test 128
GM 1825M coolants 276
GM 6038M coolants 25 149
GM 6043M coolants 44
Hard water, effects on pump seals 205

Hazardous waste
determination in used coolant 234
EPA characterization guidelines 234
management 149
in multi-stage purified coolant 258
Index Terms Links
Heater cores, sebacic acid coolant effects 63

Heat-exchange surfaces
aluminum, corrosion protection 11
scaling 289
Heavy duty diesel engines, see Diesel
engines
Heavy metals
process 258
High lead solder alloys, carboxylic-acid
ICP, see Inductively coupled plasma

emission spectroscopy
Inductively coupled plasma emission
spectroscopy
carboxylic-acid corrosion inhibitors 44
for coolant systems 180
Inhibitor depletion
high silicate alkaline phosphate coolant 63
sebacic acid coolant 63
Ion chromatography, carboxylic-acid
Ion exchange, multi-stage coolant
purification using 258
Iron
deposition on sealing surfaces 215

Index Terms Links
Iron (Cont.)
effects on water pump seals 205
Iron oxalate, formation chemistry 190
K805 cladding 83
LA-ICP, see Laser ablation inductively

coupled plasma emission
spectroscopy
Laser ablation inductively coupled plasma
emission spectroscopy 180
Lead
in multi-stage purified coolant 258
Lead solder alloys, carboxylic-acid corrosion
inhibitors 44
Lethal dose 50, PG antifreeze and SCAs 149
Life extension, of coolants 248 276
Light duty vehicles, coolant life extension 248
Liner pitting
of diesel cylinders 107 128
heavy duty engines 289
test strips for 165
Long life coolants
carboxylic acid 25
disposal 149
and seal performance 215
Index Terms Links
Maintenance, of coolants 248 289

Mercaptobenzothiazole, test strips for 165
Microscopy, protective films 11
Monoacid/diacid inhibitors 11
Multi-stage process, for coolant purification 258
Octanoic acid coolants 44

Off-site coolant recycling 276 289
On-site analyses, test strips for 165
On-site coolant recycling 276
Oxalic acid cleaning, chemistry of 190
Oxygen, and inhibitor solidification 215
PG antifreeze, LD50 149

pH
test strips for 165
Phosphate-molybdate SCAs, for diesel
engines 128
Phosphorus, deposition on sealing surfaces 215
POTWs (publicly owned waste treatment
works) 149

Index Terms Links
Propylene glycol
biodegradability 149
toxicity and disposal 149
Protective films, analysis, microscopy and
FTIR 11
Publicly owned waste treatment works
(POTWs) 149
Radiators
aluminum, sebacic acid coolant effects 63
aluminum alloy, corrosion testing 83
deposits, ICP spectroscopy for 180
Radiator solder corrosion, history of 6
RCRA (Resource Conservation and
Recovery Act) 149
Recycling
coolant life extension and 248
multi-stage process with dual-bed
deionization 258
off-site 276 289
on-site 276
processes, evaluation 276
regulations for coolants 149
technology update 289
Regulatory issues, coolant toxicity and
disposal 149

Index Terms Links
Reserve alkalinity
depletion by glycol degradation 234
test strips for 165
Resource Conservation and Recovery Act
(RCRA) 149
Reverse osmosis recycling processes 276
Right to Know laws, Federal 149
Scaling
heat transfer surfaces 289
Scanning auger microprobe/energy
dispersive
X-ray, comparison with LA-ICP sample
analysis 180
Scanning electron microscopy/energy
dispersive x-ray, comparison with LA-ICP
sample analysis 180
SCAs, see Supplement coolant additives
Sebacic acid coolants 63
Silicate-based coolants 44
Silicate deposits, high-temperature-related 215
Silicate gelation, phosphate-molybdate SCA
effects 128
Silicate phosphate coolants
depletion rates 63
fleet tests 63

Index Terms Links
Sodium tolyltriazole coolants 44

Solder bloom
in heavy duty engines 289
phosphate-molybdate SCA effects 128
Solidified silicates 215
Statistical analyses, used coolant
characterization 234
Superfund Amendment Reauthorization
Act Title III 258
Supplement coolant additives
coolant life extension with 248
in diesel engines 107 128
LD50 data 149
overtreatment effects 107
phosphate-molybdate 128
SWAAT test, method G43 83
Test strips, for on-site analyses 165

Tolyltriazole coolants 44
Toxicity
borate-nitrate SCAs 128
coolants and additives 149
Vacuum distillation recycling processes 276

Index Terms Links
Water pumps, see also Coolant pumps

failure
coolant composition and 215
fleet tests 25
leakage, phosphate-molybdate SCA
effects 128
seals
and coolant composition 215
deposit compositions 205
in heavy duty applications 289
LA-ICP for 180
Wavelength dispersive spectroscopy, pump
face seal deposits 205

Beal Roy E.eds. Engine Coolant Testing Volum PDF

Uploaded by

Copyright:

Available Formats

Beal Roy E.eds. Engine Coolant Testing Volum PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Beal Roy E.eds. Engine Coolant Testing Volum PDF

Uploaded by

Copyright:

Available Formats

What is the document about?

What is the document about?

What are some corrosion protection methods discussed?

What are some corrosion protection methods discussed?

S T P 1192

Engine Coolant Testing: Third

Roy E. Beal, editor

ASTM Publication Code Number (PCN)

Peer Review Policy

Printedin Ann Arbor,MI

A Review of Automotive Engine Coolant Technology--H. J. HANNIGAN

Corrosion Protection of Aluminum Heat-Transfer Surfaces in Engine Coolants

An Investigation of Carboxylic Acids as Corrosion Inhibitors in Engine Coolant--

Fleet Test Evaluation of Engine Coolants Using Sebacic Acid Inhibitor

Corrosion Testing for Aluminum Alloy Selection in Automotive Radiators--

An Overview of Cavitation Corrosion of Diesel Cylinder Liners--

Phosphate-Molybdate Supplement Coolant Additives for Heavy Duty Diesel

Toxicity and Disposal of Engine Coolants--R. DOUGLAS HUDGENS AND

Test Strips for Rapid On-Site Analysis of Engine Coolants--PAUL g. HEMMES,

The Chemistry of Oxalic Acid Cleaning of Engine Cooling Systems--

Investigation of Deposits on Water Pump Seal F a c e S - - R A N D A L L J. STAFFORD 205

The Relationship Between Sealing Performance of Mechanical Seals and

Characterization of Used Engine Coolant by Statistical Analysis--

A Multi-Stage Process for Used Antifreeze/Coolant Purification--

An Evaluation of Engine Coolant Recycling Processes: Part I - -

Heavy Duty Diesel Engine Coolants: Technology U p d a t e - - F . A. KELLEY 289

Copyright9 1993by ASTMInternational www.astm.org

A Review of Automotive Engine Coolant

Consultant, International Copper Assoc., Ltd., Dallas, PA.

Corrosion Protection of Aluminum Heat-

ABSTRACT: A low depletion monoacid/diacid corrosion inhibitor technology that provides

KEYWORDS: engine coolants, corrosion, inhibitors, aliphatic acid, heat-transfer, monoacids,

9 Monoacid-dibasic acid technology (Sample A)

Block temperature 240 ~

PT-100 is a Resistance Temperature Device

FIG. 2--Dynamic heat transfer test.

TABLE 1--Engine coolant inhibitors.

NOTE: X is presence of inhibitor shown by analysis.

Temperature in the 115 ~

Nominal heat input 5000 W

FIG. 3--Aging test.

TABLE 2 - - T h e A S T M D 4340 corrosion test.

Corrosion Results, mg/cm2/week

SAE 329 0.30 0.32 0.39 0.21 0.07 0.00 0.29

Heat- Transfer Properties

TABLE 3--Dynamic heat transfer test.

Corrosion Results, mg/coupon

)~ = 1.09 10 2Jm IK-Js-I

Analyses of Protective Film

9 Microscopic examination: dendrite structure coinciding with discolored areas.

2 First Brands Corporation, Danbury, CT 06810.

Fleet Test Correlations of Original Equipment

Yellow and Factory-Fill (FF)

Purple and Green

Borax yes yes no yes no yes

Original Pumps Removed at

FF, 123 74, 60% 27, 22%, 50 22, 18%

Car Actual Test Coolant Mileage at Date at

Car Actual Test Coolant Mileage at Date at