1339 PDF

TRANSPORTATION RESEARCH
RECORD
No. 1339
Materials and Construction
Recycled
Tire Rubber in
Asphalt Pavements
A peer-reviewed publication of the Transportation Research Board
TRANSPORTATION RESEARCH BOARD

NATIONAL RESEARCH COUNCIL
WASHINGTON, D.C. 1992
Transportation Research Record 1339 Sponsorship of Transportation Research Record 1339
Price: $17.00
GROUP 2-DESIGN AND CONSTRUCTION OF
TRANSPORTATION FACILITIES
Subscriber Category Chairman: Charles T. Edson, New Jersey Department of
IIIB materials and construction Transportation
Bituminous Section
TRB Publications Staff Chairman: Leonard E Wood, Purdue University
Director of Publications: Nancy A. Ackerman
Senior Editor: Naomi C. Kassabian Committee on Characteristics of Nonbituminous Components of
Associate Editor: Alison G. Tobias Bituminous Paving Mixtures
Assistant Editors: Luanne Crayton, Norman Solomon Chairman: N. Paul Khosla, North Carolina State University
Production Coordinator: Karen W. McClain Secretary: John E. Huffman, U.S. Oil & Refining Company
Office Manager: Phyllis D. Barber Joe W. Button, Douglas M. Co/will, Ervin L. Dukatz, Jr., Frank
Production Assistant: Betty L. Hawkins Fee, Ilan Ishai, Prithvi S. Kandhal, Kang-Won Wayne Lee, Kamyar
Mahboub, Roderick W. Monroe, John W. H. Oliver, Roger C.
Olson, G. C. Page, J. Claine Petersen, Michael W. Rouse, Russell
H. Schnormeier, Scott Schuler, H. Barry Takallou, Ronald L.
Terrel
Printed in the United States of America
Library of Congress Cataloging-in-Publication Data Frederick D. Hejl, Transportation Research Board staff
National Research Council. Transportation Research Board.
Sponsorship is indicated by a footnote at the end of each paper.
Recycled tire rubber in asphalt pavements. The organizational units, officers, and members are as of
p. cm.-(Transportation research record ISSN 0361-1981 ; December 31, 1991.
no. 1339)
"A peer-reviewed publication of the Transportation Research
Board."
ISBN 0-309-05203-3
1. Pavements, Asphalt. 2. Asphalt-rubber. 3. Tires,
Rubber-Recycling. I. National Research Council
(U.S.). Transportation Research Board. II. Series:
Transportation research record ; 1339.
TE7.H5 no. 1339
[TE275]
388 s-dc20
[625.8'5] 92-22590
CIP
Transportation Research Record 1339
Contents
Foreword v
Design and Construction of Asphalt Paving Materials with Crumb 1

Rubber Modifier
Michael Heitzman
JVirginia's Experimentation With Asphalt Rubber Concrete 9

G. W. Maupin, Jr.
)Florida's Approach Using Ground Tire Rubber in Asphalt 16

Concrete Mixtures
Gale C. Page, Byron E. Ruth, and Randy C. West
Advances in Technology of Asphalt Paving Materials Containing 23

Used Tire Rubber
H. Barry Takallou and Alain Sainton
/Use, Availability, and Cost-Effectiveness of Asphalt Rubber in Texas 30

Cindy K. Estakhri, Joe W. Button, and Emmanuel G. Fernando
q'>ennanent Deformation Characteristics of Recycled Tire Rubber- 38

Modified and Unmodified Asphalt Concrete Mixtures
Neil C. Krutz and Mary Stroup-Gardiner
TRANSPORTATION RESEARCH RECORD 1339
Design and Construction of Asphalt

Paving Materials with Crumb Rubber
Modifier
MICHAEL HEITZMAN
This document is a concise overview of the terminology, proc- or illegal dumps across the country. The Environmental Pro-
esses, products, and applications of c.rumb rubber modifier (CRM) tection Agency estimates that the present size of the scrap
technology. Thls technology includes any use of scrap tire rubber tire problem is 2 to 3 billion tires (1).
in asphalt paving materials . In general, CRM technology can be
Of the available expanding markets for scrap tires, only
divided into two categories, the wet process and dry proces..
When CRM i incorporated into an a phatt paving material , it two have shown the potential to use a significant number.
will modify the propertie of tbe binder (asphall rubber) and act They are fuel for combustion and crumb rubber modifier
as a rubber aggregate (rubber-m dified hot-mix asphalt). The use (CRM) for asphalt paving. Combustion already plays a major
of asphalt rubber crack and joint sealant is common aC(OSS the role, consuming 26 million tires annually. Combustion facil-
country and is routinely u ed by many srate maintenance crews. ities have the potential to use 0.5 to 10 million scrap tires/
A surface treatment using an a phalt rubber spray application is facility/yr. In comparison, the second potential new market,
called a stress-absorbing membrane (SAM) . The use of CRM in
hot-mix asphall paving material has broad variability and po- CRM, presently consumes 1 to 2 million tires/yr. The CRM
tential. Compo ite designs with RM paving materials ace a two- technology can incorporate the rubber from 2 to 6 tires into
layer system and a three-layer system. The growing nationwide a metric ton of hot-mix asphalt (HMA) paving material. To
interest in alternative use for scrap tires has provided CRM recycle 10 million scrap tires annually as CRM, 2 to 5 million
technologist with the catalyst to develop new concept for ap- metric tons of HMA material would require modification.
plying RM . The major interest has been to develop generic dry There are other alternative highway uses for scrap tires.
process mixes and a continuous blending wet procc s. There are
two principal unresolved i sues related to the use of RM in The Transportation Research Board (TRB) initiated a syn-
asphalt paving materials. On !he national level, the ability to thesis in 1989 to document these alternative uses. Scrap tires,
recycle asphalt paving mixes containing CRM has not been dem- or rubber processed for scrap tires, have been examined by
onstrated. Al the state and local level , these modified asphalt a number of highway agencies for use in light-weight em-
mixes must be field evaluated to establish expected levels of per- bankments, retaining walls, safety hardware, and pavement
formance. subbase. Details on these potential uses will be documented
in the TRB synthesis.
The use of scrap tire rubber as a modifier for asphalt cement The environmental risks linked to the presence of scrap tire
has been developing for more than 25 years. However, since stockpiles and a number of recent, well-publicized tire stock-
the late 1980s, the emphasis for this engineering technology pile fires initiated legislative action at the state and national
began to focus on its potential as a solution to an environ- level. At the beginning of 1991, 44 states had drafted, intro-
mental solid waste problem: scrap tires. Pavement perfor- duced, regulated or enacted laws to control the scrap tire
mance is a key component in determining if the use of scrap problem (2). Typical provisions of the states' legislation in-
tire rubber is cost-effective. Because of the variable conditions clude regulations to control the processing, hauling, and stor-
that affect pavement performance, it is probable that some age (stockpiles) of scrap tires; restrictions on scrap tires in
areas of the country will not benefit from this technology. landfills; provisions for funding, normally a tire disposal fee;
and in a number of states, incentives for developing new
alternative use markets.
BACKGROUND Legislation is being considered to consolidate the regula-
tions and stimulate alternative use technology. The Tire Re-
Environment and Legislation cycling Incentives Act (H.R.871 and S.396) addresses both
the regulation and technology issues. Section 1038 of the In-
Each year the United States discards approximately 285 mil- termodal Surface Transportation Efficiency Act of 1991 ad-
lion tires, more than 1 tire/person/yr. Of that figure, 33 million dresses the study and use of CRM by highway agencies.
tires are retreaded and 22 million are reused (resold). Another
42 million are diverted to various other alternative uses. The
remaining 188 million tires are added to stockpiles, landfills, Terminology
Federal Highway Administration, Office of Engineering and Office CRM technology is a general term to identify a group of
of Technology Applications, Room 3118, HNG-42, 400 Seventh Street, concepts that incorporate scrap tire rubber into asphalt paving
S.W., Washington, D.C. 20590. materials. CRM is identified as a modifier because the intro-
2 TRANSPORTATION RESEARCH RECORD 1339
duction of the scrap tire rubber modifies conventional asphalt maintenance surface patch using CRM. In 1968, the Arizona
paving products. Department of Transportation (ADOT) placed its first stress
Publications during the last 20 yr used a variety of terms absorbing membrane (SAM), a surface treatment using an
to define different processes and products as the technology asphalt rubber binder (4). ADOT placed their first stress-
evolved. Conflicting terminology has made it difficult for many absorbing membrane interlayer (SAMI) in 1972 and used the
user agencies to understand it. It is important that this doc- asphalt rubber binder in HMA open-graded friction course
ument standardize the terminology , identify the processes and in 1975.
products, and distinguish between the various concepts as they As the Sahuaro technology continued to expand, the Ari-
are introduced. A diagram of this relationship is presented in zona Refinery Company (ARCO) developed a similar wet
Figure 1. process technology that added a blend of CRM and devul-
In general, CRM technology can be divided into two cat- canized CRM to the asphalt cement. The Sahuaro and ARCO
egories. These categories define the basic process used to add technologies merged and are presently controlled by the pat-
the crumb rubber to an asphalt paving material. They are the ents' co-owners. In this paper, the wet process developed in
wet process and the dry process. The term wet process defines Arizona, is referred to as the McDonald technology.
any method that blends the crumb rubber with the asphalt The dry process was developed in the late 1960s in Sweden.
cement before incorporating the binder into the project. The The European trade name for this HMA mixture with CRM
term dry process defines those methods that mix the crumb as a rubber aggregate was Rubit. The Swedish technology was
rubber with the aggregate before the mixture is charged with patented for use in the United States in 1978 under the trade
asphalt binder. The dry process is limited to HMA applica- name PlusRide. The Alaska Department of Transportation
tions, whereas the wet process has been applied to crack began working with PlusRide in 1976 and is still the principal
sealants, surface treatments, and HMA mixtures. state highway agency developing this technology. Four cor-
It is also important to distinguish between the processes, porations have marketed the PlusRide technology since it was
as defined, and the goods that can be produced. When CRM introduced in the United States; currently it is marketed by
is incorporated into an asphalt paving material, the CRM will EnvirOtire Inc.
modify the properties of the binder and act as a rubber ag-
gregate. The modified binder is commonly called asphalt rub-
ber. When CRM is used as a rubber aggregate, the HMA is CRUMB RUBBER MODIFIER AND MODIFIED
called rubber-modified hot-mix asphalt. Understanding the PROPERTIES
process-product relationship is the key to developing the de-
sign for a specific project. Crumb Rubber Modifier
Tire rubber is the principal component in CRM. Tire rubber

History in the United States is primarily a composite of a number of blends of natural and
synthetic rubbers and carbon black. Although variations in
The use of CRA in asphalt paving did not develop as a solution tire rubber exist, both between tires and within the tire struc-
to an environmental problem. In fact, the development of ture, the rubber composition of a bulk sample of CRM is
natural rubber in bitumen was introduced in the early 1840s reasonably uniform .
(3). For years, engineers and chemists have worked to blend The principal source of raw material for producing CRM
natural rubber (latex) and synthetic rubber (polymers) in as- is scrap tire rubber. Scrap tire rubber can be delivered to the
phalt cements to enhance the elastic properties of the binder. processing plant as whole tires, cut tire, shredded tire, or
Tire rubber, a compound of natural and synthetic rubber, is retread buffing waste. Shredded tire rubber is the preferred
an available raw material that has been included in this effort. and logical alternative as a raw material for producing CRM.
In the early 1960s, Charles McDonald , materials engineer The type of scrap tire raw material and the quality of that
for the city of Phoenix, began working with a local asphalt material are generally the responsibility of the CRM pro-
company , Sahuaro Petroleum, to develop a highly elastic cessors. The capabilities of the processing plant and the buy-
er's specified CRM properties will direct the processor's
operation.
M_
_ ATE
_ R_
IA_
L_.,_
P_RO
_C_E_
s_s __,~ _
TE_C~H_
N~O_
LO_G
_Y_ _.. PRODUCT
There are three methods currently used to process scrap
tire rubber intcf CRM. The crackermill process is the most
Wet < McDonald -----.....
continu~us /
Modified Binder
{asphalt rubber)
common method . The crackermill process tears apart scrap
tire rubber, reducing the size of the rubber by passing the
/ blending i
·''
binder material between rotating corrugated steel drums. The gran-
CRM ulator process shears apart the scrap tire rubber, cutting the
-·:
PlusRide rubber with revolving steel plates that pass at close tolerance.
The micro-mill process further reduces a crumb rubber to a
Dry generic '- - - - Rubber Aggregate very fine ground particle.
chunk rubber/ {rubber modified) As the scrap tire rubber is processed, reducing its size, the
hot mix asphalt steel belting and fiber reinforcing are separated and removed
from the rubber. Talc, or other inert mineral powder , is added
FIGURE 1 Relationship of crumb rubber modifier to the CRM to reduce the rubber particles' tendency to stick
terminology. together. Typically, the amount of talc required should not
Foreword
Ground tire rubber has been used as an additive in asphalt for highway pavement construction
since the mid-1960s. Early use of particulate tires in asphalt was a means to improve per-
formance of the asphalt while simultaneously eliminating a waste product. Although the
technology for tire-rubber-modified asphalt has been available for more than 20 years, only
recently has the waste tire problem become so acute that the Intermodal Surface Transpor-
tation Efficiency Act of 1991 (ISTEA) includes a schedule of minimum requirements for
using recycled tire rubber in asphalt pavements starting in 1994.
Ground tire rubber can be blended with hot liquid asphalt, resulting in a new product
known as "asphalt-rubber." This modified asphalt can be used to make any of the various
types of asphalt pavements produced using conventional asphalts. Ground tire rubber can
also be added to the asphalt concrete mixture as a dry particulate or aggregate. In this case,
the properties of the resulting paving mixture containing the tire rubber are affected by the
presence of the rubber.
At one of the sessions of the 71st Annual Meeting of the Transportation Research Board,
a panel of experts discussed the current technology and practice for using ground tire rubber
and asphalt-rubber in asphalt pavements. Six of the papers presented by these experts are
included in this Record. These should be of interest to state and local design, materials,
construction, maintenance, and research engineers, as well as contractors and material pro-
ducers.
Heitzman presents a Federal Highway Administration (FHWA) overview of the experience
to date with recycled rubber in asphalt pavements. Much of the early research and demon-
stration work with ground recycled tire rubber was conducted by and for the FHW A during
the past two decades. Maupin of Virginia and Page et al. of Florida relate practical experiences
with each of the types of tire-rubber-modified pavement processes and describe some of the
difficulties associated with putting tires into pavements.
Takallou and Sainton report on some new advances in this emerging technology that have
been made recently in France. Estakhri et al. report on the cost-effectiveness of using asphalt
pavements as the repository for recycled tire rubber in Texas. Krutz and Stroup-Gardiner
present the results of a comparative study of the rutting performance of asphalt concrete
containing recycled tire rubber versus that of unmodified asphalt concrete mixtures.
v
Heitzman 3
exceed 4 percent by weight of the rubber. In general, a scrap The surface area can be increased by reducing the size and
tire weighing approximately 9 kg (20 lb) will produce 4.5 to specifying a crackermill process. The rate of reaction is also
5.5 kg (10 to 12 lb) of CRM. The remainder of the tire is influenced by the temperature at which the blend is reacted.
fiber, steel, and any rubber removed with the fiber and steel. The specified reaction time should be the minimum time
Each method of producing CRM generates a unique par- (at a preset temperature) required to stabilize the binder
ticle with specific characteristics. The cracker mill process viscosity.
produces an irregularly shaped torn particle with a large sur- This modified binder, asphalt rubber, exhibits enhanced
face area. The particles can be produced over a range of sizes binder properties when compared with conventional asphalt
from 4.75 mm to 425 µm (No. 4 to No. 40) sieve. These cement in laboratory tests. Changes in the viscosity of the
particles are commonly described as a ground CRM. The binder over the normal range of operating and mixing tem-
granulator produces a cubical, uniformly shaped cut particle peratures indicates that the addition of CRM flattens the
with a low surface area. The particles can be produced over temperature-viscosity curve, reducing the binder's tempera-
a range of sizes, typically 9.5 mm down to 2.00 mm (3/8 in. ture sensitivity.
to No. 10) sieve. This material is called a granulated CRM. A majority of the standard binder tests used to measure
The micro-mill process produces a very fine ground CRM. The the properties of asphalt cement can be applied to asphalt
particles can be reduced to a range of sizes from 425 µm down rubber binder. Only the conventional capillary-type viscom-
to 75 µm (No. 40 to No. 200) sieve. eter tests are known to be ineffective (6). The method used
The project specifications should establish the required gra- to measure the viscosity of these modified binders is rotational
dation of the CRM and the type of particle, ground or gran- shear resistance using the Brookfield Viscometer, ASTM D
ulated. Every CRM producer should have a quality control 2994. Portable versions of this viscometer are commonly used
program to continually monitor the uniformity of the product to monitor the binder during the reaction phase and as a
for both its chemical composition and gradation. Processing production control. Several binder tests may show an increase
scrap tires into CRM is not a mobile operation. A small in- in standard deviation caused by the non uniformity of the mod-
dustrial facility with moderate capital investment in equip- ified binder. Because the crumb rubber does not dissolve into
ment is necessary to produce a quality product. Most CRM the asphalt cement, the swollen rubber particles in the binder
is shipped in 22.7- or 27.2-kg (50- or 60-lb) bags, which are can affect the consistency of the binder during a particular
paper or plastic. The average cost of CRM from the producer test.
ranges from 20 to 35 cents/kg (10 to 15 cents/lb) for coarse The enhancements in the binder properties measured in the
and medium crumb (above 425 µm) and up to 55 cents/km laboratory can be an indication of better performance of the
(25 cents/lb) for fine ground crumb. paving material in the field. However, there are numerous
variables, beyond the properties of the binder, that also affect
the overall performance of the pavement. Setting these other
Modified Properties variables aside, the modified binder properties may influence
the pavement's performance related to thermal cracking, rut-
There are two basic products that can be achieved by adding ting, reflective cracking, aging, and chip retention.
CRM to asphalt paving applications. They are modified binder There is also the potential for certain undesirable side ef-
and rubber aggregate. The size, shape, and texture of CRM fects. To develop enhanced pavement performance charac-
required to achieve these end products varies with the pro- teristics, the mix design will generally require the modified
posed application. binder to increase its role in the paving material. In simple
terms, modifying the asphalt binder with CRM will require
an increase in the binder content. This affects the paving
Modified Binder material's cost, potential to flush-bleed, and may cause
tracking.
When asphalt cement and CRM are blended together, there The ability of CRM to enhance the properties of the binder
is an interaction between the materials (5). This interaction, hinges on the compatibility between the asphalt cement and
defined as asphalt rubber reaction, is affected by a number the CRM. For all practical purposes the ability to change the
of variables. Specifically, the reaction is influenced by the overall blended composition of CRM is limited. The type and
temperature at which the blending-reaction occurs, the length amount of aromatic oil in the asphalt cement plays a major
of time the temperature remains elevated, the type and amount role in determining the compatibility.
of mechanical mixing energy, the size and texture of the CRM,
and the aromatic component of the asphalt cement. The re-
action itself is the absorption of aromatic oils from the asphalt Rubber Aggregate
cement into the polymer chains that are the key component
of the natural and synthetic rubber in CRM. As CRM reacts The other product achieved by adding CRM is rubber aggre-
with the asphalt cement, it also swells and softens. The vis- gate. By limiting the time that the asphalt cement and CRM
cosity of the asphalt-CRM blend is used to monitor the re- are maintained at mixing (reaction) temperatures and by spec-
action. An asphalt cement modified with 15 percent CRM ifying a coarse granulated CRM, the CRM can retain its phys-
can increase the binder's high temperature viscosity by a factor ical shape and rigidity. By specifying a granulated CRM, the
of 10 or more. smooth sheared surfaces of the particle are less reactive (lower
The rate of reaction between CRM and the asphalt cement surface area than ground CRM) and its cubical shape can be
can be increased by enlarging the surface area of the CRM. factored into the combined gradation of CRM and aggregate.
This rubber aggregate product is only applied to hot-mixed point plastic material that allow the operator to charge the
asphalt designs. mixing chamber with the entire bag of CRM. The batch size
Putting aside the effect of any binder modification that may usually corresponds to a whole number of bags per batch.
accompany a design with rubber aggregate, rubber aggregate The dry process can be used with drum mix facilities similar
may influence the pavement performance related to reflective to producing mixes with recycled asphalt product (RAP). The
cracking and ice disbonding. process and equipment required for introducing RAP into a
There are potential disadvantages associated with CRM drum mixer can also be used to introduce CRM into a drum
used as a rubber aggregate. Similar to a CRM-modified binder, mixer. Similar to batch mixing, the dry process at a drum mix
for rubber aggregate to achieve the benefits of delayed re- facility has been labor intensive.
flective cracking and ice disbonding requires a minimum CRM
content in the mix. This affects the cost of the paving material
and may cause raveling. DESIGN AND CONSTRUCTION PRACTICES
Compatibility is not as critical with CRM as it is with a
rubber aggregate. The reaction between the CRM and the Crack and Joint Sealants
asphalt cement does not play a significant role in developing
the performance enhancements of rubber aggregate. The use of asphalt rubber crack and joint (C/J) sealant is
common across the country and is routinely used by many
state maintenance crews. The choice of a sealant for a given
CONSTRUCTION PROCESS FOR CRM location should take into consideration the type of pavement,
type of crack or joint, shape and size of the crack or joint,
Wet Process time before next scheduled major rehabilitation, traffic vol-
ume, degree of pavement distress, maximum and minimum
The wet process defines any method that adds the CRM to temperatures, available equipment and work crews, and traffic
the asphalt cement before incorporating the binder into the control.
asphalt paving project. This process is used to produce a The manufacturers of asphalt rubber C/J sealant provide a
modified binder product. There are three elements to the variety of sealants to meet different climate and pavement
equipment necessary to achieve the wet process. They include conditions. These sealants are normally designed to meet var-
blending the CRM and asphalt cement, reacting the two ma- ious ASTM specifications. Asphalt rubber C/J sealant is typ-
terials, and transferring the modified binder product to the ically preblended and packaged in 22.7-kg (50-lb) blocks. These
desired project application. Two limiting factors to the process blocks must be remelted and reacted before the sealant can
are having sufficient storage area for the shipment of CRM be applied.
and the manual effort required to add CRM to the blending Approximately 80 percent of the states use some amount
unit's hopper. of asphalt rubber C/J sealant. States that apply a large amount
The blending unit must be capable of properly metering of this sealant include: Arizona, California, Georgia, Ne-
the CRM (a dry ground and granulated material, or both) braska, Nevada, New Mexico, New York, Pennsylvania, Texas,
into the asphalt cement (a hot viscous liquid) at the required and Wisconsin. The cost of the material generally ranges from
proportion established by the mix design. The reaction tank 45 to 65 cents/kg (20 to 30 cents/lb). This is the material cost
must be capable of maintaining a uniform blend and a uniform and does not include shipping or the cost of installation (labor,
constant temperature. Most applications require some method equipment or traffic control).
of controlled metering for the modified binder. Special pumps
and frequent calibration are essential to ensure that a uniform
accurate application of the modified binder is achieved . Surface Treatments
A surface treatment using an asphalt rubber spray application

Dry Process is called a Stress Absorbing Membrane (SAM) (7) . The design
of a SAM should examine both the binder and cover aggre-
The dry process defines any method of adding CRM directly gate. The binder is asphalt rubber that may be thinned with
into the HMA mix process, typically pre-blending the CRM a diluent to improve distributor-spray flow. The amount of
with the heated aggregate before charging the mix with as- CRM in the binder is typically 20 to 30 percent by weight of
phalt. This process is normally used when a rubber aggregate asphalt cement. Cover aggregate is generally a uniform size
product is desired. The dry process will generate some re- (9 .5 mm to 6.3 mm sieve, 3/8 in. to 1/4 in .) and preferably
action between the CRM and asphalt cement. The amount hot precoated with 0.3 to 0.5 percent (by weight) asphalt
of fine CRM introduced into the mix will determine the degree cement. The compatibility of the binder and aggregate is a
of modification to the asphalt binder. No special equipment part of the design process.
is needed for this process; however, a calibrated proportioning Once the materials have been selected, the designer must
feed system is necessary when drum plants are used. The dry determine the appropriate spray application rate and aggre-
process has only been applied to hot-mixed paving projects. gate spread rate to achieve proper coverage and embedment.
It does not lend itself to other asphalt paving applications like Typical values that have been used successfully are 2. 7 litre/
surface treatments. m 2 (0.6 gall/yd2 ) of diluted asphalt rubber binder and 19
Incorporating the dry process at a batch mix facility is sim- kg/m 2 (35 lb/yd 2 ) of precoated cover aggregate. The construc-
ple but labor intensive. The bags are made of a low melting tion of a SAM is similar to any surface treatment. The use
Heitzman 5
of an asphalt rubber binder in a surface treatment has par- asphalt rubber. The desired drain-down characteristics do not
ticular benefits for the performance of the pavement. Tem- change. The amount of CRM in asphalt rubber binder for
perature susceptibility and elasticity influence the binder's HMA applications generally ranges from 15 to 25 percent by
ability to resist the stresses induced by climate and traffic, weight of asphalt cement.
thus the name stress-absorbing membrane (SAM). The construction of HMA with asphalt rubber binder is
The engineering properties of a SAM can resist and delay similar to constructing conventional mixes. The target tem-
the development of reflective cracks when the cracks are gen- peratures for mixing, Iaydown, and compaction are typically
erally inactive, like alligator fatigue cracking and closely spaced higher. Release agents for the equipment, particularly
random or block cracking. A SAM cannot resist the amount the truck beds and steel-wheel roller drums, must not be
of strain that is typical of transverse thermal cracks in asphalt petroleum-based products. Pneumatic tire rollers are gener-
concrete pavements or transverse contraction joints in port- ally not permitted because the asphalt rubber binder tends to
Jand cement concrete pavements. build up on rubber materials.
Several states routinely design and apply SAMs to their The field inspection of HMA with asphalt rubber is similar
pavement network. Arizona, California, and Texas are pre- to conventional mixes. The use of extraction test procedures
dominant states involved with using this product. The present is not practical with these modified mixes. The reacted (and
cost of a SAM in place is generally 100 percent higher than unreacted) CRM is not completely soluble in the extraction
a conventional surface treatment. The cost increase is prin- solvents. Nuclear asphalt content gauges can be used to mea-
cipally caused by the asphalt rubber binder. In Arizona and sure the modified binder content of a mix applying normal
California during the late 1980s, the in-place cost for a SAM calibration procedures for each mix.
generally ranged from $1.90 to $2.30/m 2 ($1.60 to $1.90/yd2 ). There are no state highway agencies that routinely use HMA
This section has focused on the application of asphalt rubber with asphalt rubber binder for any particular application. The
as a spray application for surface treatments. Other thin majority of documented field-test sections with appropriate
asphalt surfacing techniques may also benefit from CRM evaluation programs were placed during and after the mid-
technology. 1980s. California has performed the most extensive amount
of field performance research and has not yet resolved all the
issues (9). The cost of asphalt rubber HMA mixtures (in-
Hot-Mix Asphalt place) has ranged from 50 to 100 percent higher than the
conventional mix. The projected future cost of HMAs with
The use of CRM in HMA paving materials has broad variabil- asphalt rubber could reduce to between 20 and 30 percent
ity and potential. Two of the principal variables are the type above conventional HMA if the mix is routinely applied.
of CRM process (wet or dry) and the type of HMA (dense,
gap, or open-graded). For clarity, this section on HMA paving
applications is divided into two parts, HMA applications using PlusRide Technology
the wet process (McDonald) and those using the dry process
(PlusRide). PlusRide is a modified gap-graded mix and the mix design
does not follow normal Marshall or Hveem procedures (JO).
The PlusRide HMA is designed to modify the stability of a
McDonald Technology gap-graded aggregate matrix with the elastic properties of
CRM and a certain amount of binder modification (reaction).
Conventional Marshall and Hveem mix design procedures Conventional specimen preparation equipment and proce-
have been used successfully for dense-graded mixes using dures are performed with some modifications, but the spec-
McDonald's asphalt rubber (8). The characteristics of the imens are not tested for stability. The only measured specimen
modified binder alter the laboratory measured properties of property used to establish the mix design asphalt content is
the mix and should be understood when designing these dense percent air voids. The target air void content is 2 to 4 percent.
mixes. As a general rule, the increase in the designed binder The aggregate gradation and CRM content and gradation are
content will be proportional to the amount of CRM in the relatively fixed by the patent description. The CRM is pre-
binder. dominantly a granulated crumb passing the 6.3-mm (1/4-in.)
The present design concept being developed for modified sieve with the fraction passing the 2.00-mm (No. 10) sieve
gap-graded mixes is to maximize the asphalt rubber content supplemented with buffings or ground CRM. As specified in
of the mix. This design is intended to combine the stability the design, the CRM content is 3 percent by weight of the
of coarse aggregate contact with the elastic properties of as- total mix. The asphalt binder content will generally range from
phalt rubber. Typical asphalt rubber binder contents for gap- 7.5 to 9.0 percent.
graded mixes developed in Phoenix range from 8 to 9 percent. There are only a few modifications to the construction prac-
The design of open-graded mixes (OGFC) with asphalt tices for PlusRide HMA. Compaction concerns are similar to
rubber binder requires two revisions to the procedure. To asphalt rubber HMA. In addition to these modifications, the
determine the binder content of OGFC with asphalt rubber finish roller must continue to compact the PlusRide mat until
will require a revision of the formula to account for the thicker it cools below 60°C (140°F).
binder film associated with asphalt rubber; in essence, com- Poor production, placement, or compaction control will
pute a higher binder content. The procedure for establishing lead to premature failure of the pavement. Inspectors should
the optimum mixing temperature will require a change in the be knowledgeable about the required construction practices.
target binder viscosity to better reflect the high viscosity of Extraction methods will not provide accurate means of mon-
itoring CRM content nor binder content. Similarly, asphalt nor examine the association of the new concepts to the pat-
content gauges will measure all the CRM in the sample as a ented products. State and local highway agencies should be
part of the binder content. aware of the known patented products and make their own
Experimental applications of PlusRide began in 1979. Alaska determination of any conflict between a proposed new con-
is the only state DOT with a substantial background in de- cept and its comparable patented product.
veloping PlusRide in the United States. The cost of PlusRide
HMA (in-place) has ranged from 50 to 100 percent higher
than conventional HMA. The projected future cost of this Generic Dry Technology
rubber-modified HMA could reduce to between 20 and 40
percent above conventional HMA if these mixes are routinely A major interest has been to develop generic dry process
applied. mixes. The concept was originated by Barry Takallou as a
result of his research and practical experience with PlusRide
(11). The principal focus of this concept in CRM technology
Composite Designs-SAM! is to incorporate CRM into conventional dense and gap-graded
HMA mixes using the dry process. Unlike PlusRide, which
Composite designs with CRM paving materials offer similar specifies a particular gap gradation for the aggregate, the
theoretical benefits as the use of paving fabrics (7). The prin- proposed technology considers the available generic grada-
cipal theory of the design is to place a membrane beneath the tions for the locality; hence the name, generic dry technology.
overlay that can resist the stress-strain of reflective cracks and Much of the theory and understanding needed to design
delay the propagation of the crack through the new overlay. and construct PlusRide mixes also applies to generic dry mixes .
Similar to a SAM, the asphalt rubber membrane is called a Although the theory is similar, the variability of application
SAMI. is much greater . There are a number of factors that must be
There are two composite design systems, a two-layer SAMI considered in the design, particularly how the CRM is to
and a three-layer SAMI. A two-layer SAMI places the SAMI modify the mix. The PlusRide concept modified the HMA
on the existing pavement and overlays the SAMI with 25 to primarily through aggregate substitution. It is possible with
75 mm (1 to 3 in.) of HMA . A three-layer SAMI begins with generic dry process mixes to achieve a greater degree of binder
the placement of a leveling course of HMA. This initial over- modification. By specifying a smaller particle size ground crumb
lay provides an acceptable uniform surface for placing the rubber, the dry process combined with the HMA production
SAMI. The SAMI is followed by an additional 25 to 75 mm sequence may be sufficient to permit the CRM and asphalt
(1 to 3 in.) of HMA overlay. This system applies when there binder to achieve a substantial degree of reaction before
is deterioration of the existing pavement cracks and joints. If placement and compaction of the mix.
a two-layer SAMI were used, the deteriorated cracks would Because the intent of generic dry process technology is to
create a discontinuity in the membrane at the location where use conventional aggregate gradations, the design process must
the membrane will be subjected to the highest levels of stress determine the appropriate CRM gradation for the proposed
and the performance of the SAMI would be diminished. mix properties. The designer must take into consideration the
Construction and inspection of a SAMI is the same as a capabilities of the CRM manufacturer. Present grading flex-
SAM. The only additional consideration is to assure that the ibility in most CRM plants is limited. If unusual CRM grading
diluent added to the asphalt rubber binder prior to the spray requirements are specified, the cost of the CRM will increase
application has adequately evaporated from the membrane or the gradation may not be attainable.
before the HMA overlay is placed. The generic concept has been successfully constructed in
The delay of reflective cracking is the principal benefit of experimental field applications in New York and Florida. The
a composite design. This potential performance benefit is sim- New York projects included three generic dry process designs
ilar to the benefit of paving fabrics . The cost of a SAMI is (12). The designs varied the amount of CRM added to the
slightly higher than the cost of the fabric. mix from 1 to 3 percent by weight of total mix. The aggregate
gradation was a standard 12.5-mm (1/2-in.) nominal maximum
dense-graded surface mix. The CRM gradation had a 2.0-mm
NEW CONCEPTS (No.10) sieve nominal maximum size. The asphalt cement
content increased to 7.2 percent for the 3.0 percent CRM mix
The combination of an existing exclusive, proprietary CRM compared with the 6.0 percent asphalt cement for the con-
paving market and growing nationwide interest in alternative ventional design. All the mixes used the same grade of asphalt
uses for scrap tires provided CRM technologists the catalyst cement. In Florida, a June 1989 experimental project using
to develop new concepts for applying CRM. Initial laboratory CRM included one section of generic dry process. The ag-
work in this area did not begin until the mid- to late 1980s. gregate was a standard 9.5-mm (3/8-in.) nominal maximum
The first experimental field applications were placed in 1989. open-graded friction course. Along with four sections of as-
Two conditions should be noted regarding these new con- phalt rubber, the section of generic dry process contained
cepts. First, the design and construction practices are still 180 µm (No. 80) sieve CRM at 10 percent by weight of binder.
being developed and there is no field performance record to The binder content of the control mix was 6.3 percent com-
demonstrate that they can provide an acceptable level of ser- pared with the design binder content of the generic CRM mix
vice over a normal performance period. The second condition at 7.0 percent, both using the same grade of asphalt cement.
is that McDonald's asphalt rubber and PlusRide are patented Additional experimental sections were constructed in a
products. This report did not review the extent of the patents number of states in 1991. Illinois, Iowa, Kansas, and Oregon
Heitzman 7
have constructed projects to evaluate the generic dry In addition to the uniformity of the binder, the actual en-
technology. gineering characteristics of the binder may behave differently
from the commonly known McDonald asphalt rubber. The
very fine gradation of 180 µm (No.80) sieve CRM substan-
Chunk Rubber Asphalt Concrete
tially increases the dispersion of the CRM throughout the
asphalt cement. It has not been determined if this additional
As a part of the Strategic Highway Research Program, the
dispersion will improve or reduce the performance of the
Cold Regions Research and Engineering Laboratory (CRREL)
modified binder. It is possible that the optimum CRM content
of the U.S. Army Corps of Engineers was contracted to eval-
will not be the same as that for other asphalt rubber binders
uate the ice-disbonding characteristics of several asphalt pav-
with coarser CRM.
ing materials. One of those materials was PlusRide (13). In
addition to this research effort, CRREL began modifying the
design to determine if the use of CRM could further modify
FURTHER RESEARCH
the properties of the paving material. Their work focused on
increasing both the maximum size of the crumb and the per-
There are two principal issues related to the use of CRM in
cent CRM in the HMA.
asphalt paving materials that need to be evaluated (15). On
The CRREL concept revised the aggregate gradation from
the national level, the ability to recycle asphalt paving mixes
the gap-graded PlusRide design to a dense-graded aggregate,
containing CRM has not been demonstrated. At the state and
while maintaining the same nominal maximum aggregate size.
local level, these modified HMA mixtures must be field eval-
The CRM gradation was revised to a narrow grading band
uated to establish expected levels of performance. There are
(12.5-mm to 4.75-mm sieve, 1/2 in. to No. 4), with a larger
other areas, as well, that are unclear or need further devel-
maximum crumb size. This revision of the gradations applies
opment but are not as critical for acceptance of this technology
to mixes with CRM contents similar to PlusRide, namely 3
by the highway community. This section divides the further
percent CRM by weight of mix. As the CRREL research
research areas into three categories: national, state, and in-
increased the percent of CRM, adjustments were made in the
dustry issues.
aggregate gradation to provide space in the aggregate matrix
for the substitute rubber aggregate. This research examined
chunk rubber asphalt concrete mixes with 3, 6, 12, 25, 57,
National Issues
and 100 percent crumb rubber by weight of aggregate. As
expected, the optimum asphalt cement content (based pri-
National issues are areas of concern that can be resolved and
marily on air voids) increased as the percent CRM increased.
addressed and applied on a national basis. As noted, the
Actual Marshall mix designs produced asphalt cement con-
ability to recycle asphalt pavements containing CRM is a prin-
tents ranging from 6.5 percent for 3.0 percent CRM to 9.5
cipal research issue. This aspect of the technology is critical
percent for 12 percent CRM.
to its long-term application. If these modified paving materials
This research initiative has been confined to laboratory
cannot be recycled, their benefit is substantially reduced and
testing. There are no scheduled experimental field applica-
a new waste problem is created. Two questions need to be
tions established for this concept. CRREL is currently seeking
addressed in the recycling area. First, can materials containing
sources of research funding to continue the development of
CRM be successfully processed as recycled asphalt pavement?
these unique mixes. Until the material is subjected to actual
and second, how does the recycled paving material containing
field conditions, it is impossible to estimate its performance
CRM perform?
or practical application.
Another area of national concern is the development of
standards, particularly for material testing and the environ-
Continuous Blending Asphalt Rubber ment. It is important for the highway community to develop
and adopt standard material testing methods so that data
One concern regarding the McDonald wet process is the re- collected from various sources across the country can be shared
quired batching and reaction time associated with blending on a relatively equal basis.
CRM and asphalt cement to produce asphalt rubber. As pre- The area of establishing environmental standards for emis-
viously discussed, the time required to react these materials sions and fumes is not the responsibility of the highway
is dependent on a number of factors, including the size of the agencies. Their role is secondary. Highway agencies are re-
CRM. Rouse Rubber Industries applied wet process tech- sponsible for enforcing those standards developed by the en-
nology, blended their 180 µm (No.80) sieve CRM with an vironmental agencies, who coordinate with the industry as the
asphalt cement, and developed a continuous blending standards are developed.
procedure. The prototype blending equipment is still being
evaluated.
The first experimental field application of the concept was State Issues
achieved with the cooperation of the Florida DOT in 1990
(14). Results on the performance of this continuous blended State issues are areas of CRM technology that can only be
asphalt rubber will not be known for several years. Particular resolved through proper evaluation at the state and local level.
attention will be given to the uniformity of the binder prop- The principal research issue at this level is the field evaluation
erties as they relate to the uniformity of the blending oper- of these CRM-modified materials to establish their expected
ation. performance. The performance of CRM has been docu-
mented in some applications. In other applications, particu- 4. L. A. Scofield. The History, Development, and Performance of
larly as an HMA mixture, the use of CRM is still being eval- Asphalt Rubber at ADOT. Report AZ-SP-8902. Arizona De-
partment of Transportation, Phoenix, Dec. 1989.
uated. As applications vary, so do the performance criteria 5. B. H. Huff. ARCO Concept ofAsphalt-Rubber Binders. Asphalt-
and the cost-effectiveness. Therefore, the performance data Rubber User-Producer Workshop, Scottsdale, Ariz., May 1980.
must be application specific. Studies have concluded that lab- 6. R. A. Jiminez. Viscosity Measurements of Asphalt-Rubber Bind-
oratory tests using CRM modified mixes do not correlate with ers. National Seminar on Asphalt-Rubber, Kansas City, Mo.,
Oct. 1989.
measured field performance (16). Therefore, laboratory re- 7. T. S. Shuler, R. D. Pavlovich, J. A. Epps, and C. K. Adams.
sults used to predict field performance may not be accurate Investigation of Materials and Structural Properties of Asphalt-
and may not accurately reflect the cost-effectiveness of the Rubber Paving Mixtures. Report FHWA/RD-86/027. FHWA,
material. U.S. Department of Transportation, Sept. 1986.
8. M. Stroup-Gardiner, N. Kruptz, and J. Epps. Comparison of
Mix Design Methods for Rubberized Asphalt Concrete Mixtures.
In National Seminar on Asphalt-Rubber, Kansas City, Mo., Oct.
1989.
Industry Issues 9. J. L. Van Kirk. Ca/trans Experience With Rubberized Asphalt
Concrete. Technology Transfer Session, An Introduction to Rub-
Industry issues are areas of CRM technology that can be berized Asphalt Concrete, Topeka, Kans., Jan. 1991.
addressed to improve the material, process, and technology. 10. H. B. Takallou and R. G. Hicks. Development of Improved Mix
and Construction Guidelines for Rubber-Modified Asphalt Pave-
The material, process, and technology are continually evolv- ments. In Transportation Research Record 117I, TRB, National
ing. An improvement in one area becomes a catalyst for im- Research Council, Washington, D.C., Jan. 1988, pp. 113-120.
provements in other areas. 11. H. B. Takallou. Evaluation of Mix Ingredients on the Perfor-
As the technology expands, specification parameters will mance of Rubber-Modified Asphalt Mixtures. Oregon State Uni-
versity, Corvallis, June 12, 1988.
require the crumb rubber producers to add flexibility and 12. J. F. Shook. Experimental Construction of Rubber-Modified As-
quality control to their production. This may be particularly phalt Mixtures for Asphalt Pavements in New York State. ARE
true for the gradation control. Larger production rates will Inc., New York Stute Deportment of Trnnsportution, Albany,
require larger material-handling systems equipped with me- N.Y., May 16, 1990.
chanical feed systems. Methods and equipment to handle larger 13. R. A. Eaton, R. J. Roberts, and R.R. Blackburn. Use of Scrap
Rubber in Asphalt Pavement Surfaces. ARA International Tire
containers of crumb rubber could improve the overall effi- Recycling Conference, San Jose, Calif., Jan. 20-31, 1991.
ciency of the wet and dry processes. In addition to the crumb 14. G. C. Page. Florida's Experience Utilizing Ground Tire Rubber
rubber handling system, the blending and metering systems in Asphalt Concrete Mixes. Research Report FL/DOT/M0-89-
for adding CRM (dry process) and asphalt rubber (wet proc- 3666. Florida Department of Transportation, Tallahassee, Sept.
1989.
ess) into an HMA facility will need to be interlocked with the 15. D. Bernard. Hearing Before the Subcommittee on Energy Reg-
other material feed systems to eliminate manual adjustment ulation and Conservation Committee on Energy and Natural Re-
of feed rates during HMA production. sources United States Senate. FHWA, U.S. Department of Trans-
portation, Aug. 2, 1990.
16. F. L. Roberts, P. S. Kandhal, E. R. Brown, and R. L. Dunning.
Investigation and Evaluation of Ground Tire Rubber in Hot Mix
Asphalt. The National Center for Asphalt Technology, Auburn
REFERENCES University, Ala., Aug. 1989.
1. Franklin Associates, Ltd., and R. L. Hershey. Markets for Scrap

Tires. Report EPA/530-SW-90-074A. Office of Solid Waste, En-
vironmental Protection Agency, Oct. 1991. The contents of this paper reflect the views of the author, who is
2. M. Sikora. Third Annual Legislative Update. Scrap Tire News, responsible for the facts and the accuracy of the data presented herein.
Vol. 5, No. 1, Jan. 1991. The contents do not necessarily reflect the views or policies of the U.S.
3. K. Allison. Those Amazing Rubber Roads, Part II. Rubber World, Department of Transportation or the Federal Highway Administration.
March, April 1967. This paper does not constitute a standard, specification, or regulation.
TRANSPORTATION RESEARCH RECORD 1339 9
Virginia's Experimentation With Asphalt

Rubber Concrete
G. W. MAUPIN, JR.
The installation of test sections containing asphalt rubber concrete dealing with the comparison of the properties of asphalt con-
using the wet process and the testing of samples of the materials cretes using coarse rubber or fine rubber.
in the laboratory are described . Also, some preliminary test data
from a laboratory study using both coarse and fine crumb tire
rubber with the wet process are included. The installation of the
test sections went smoothly with only minor problems. The fol- FIELD TESTS
lowing laboratory tests were used to evaluate the mixtures: Mar-
shall, gyratory shear, creep, resilient modulus, indirect tensile Installation
strength, and stripping. The test results indicated that the asphalt
rubber mixtures were more susceptible to permanent deformation In August 1990, two sections of control mixture and two sec-
than the same mixtures without asphalt rubber; however, the tions of experimental asphalt rubber mixture were placed as
temperature at which the mixtures were compacted and tested
may have affected the results, which may not therefore reflect an overlay on Rte. 1 in Fairfax County in an urban area in
how the asphalt rubber will perform in the field. After being which slow-moving traffic often causes excessive permanent
subjected to the indirect tensile stripping test, the mixtures with deformation. These four-lane sections carry 30,000 to 40,000
asphalt rubber displayed less stripping than the mixtures without vehicles/day (5 to 15 percent trucks).
asphalt rubber. The laboratory study indicated that asphalt rubber The asphalt contractor hired a subcontractor to supply the
binder containing either coarse or fine rubber can be stored for crumb rubber and to blend it with asphalt cement and an
a reasonable period (at least 24 hr) with no breakdown of the
rubber. The coarse and fine rubber reacted with the asphalt in a extender oil at the hot-mix plant concurrently with the hot-
similar way, but the mixture containing coarse rubber required mix operation. The AC-30 asphalt cement containing exten-
slightly more asphalt cement than the mixture with no rubber. der oil was heated to approximately 420°F before being mixed
The fine rubber mixtures displayed an optimum rubber content with the crumb rubber, and the resultant binder was stored
in those cases in which the maximum resilient modulus and in- at 360°F before being mixed with the aggregate in the hot-
direct tensile strength were achieved. mix drum plant. The construction operation went smoothly
except for a minor problem with pickup on the leading roller
Attempts to use old tire rubber in asphalt for highways date drum of the breakdown roller.
back to the 1920s (1). Rubber may be used in asphalt concrete
as an aggregate, or it may be reacted with the asphalt cement
to yield asphalt rubber. Most of the rubber that has been used Materials
has been asphalt rubber. The success of this product has been
somewhat mixed: some agencies have expressed overwhelm- The 1/2-in. surface mixtures (Table 1) were devised with a
ing satisfaction with it, and others have expressed doubt about 75-blow Marshall design . The control mixture contained 4.5
the economic advantage of adding rubber to asphalt. Logi- percent of AC-30 asphalt cement, and the experimental mix-
cally, the rubber should impart desirable characteristics that ture contained 6. 75 percent of asphalt rubber composed of
improve the life of the pavement. The use of rubber has AC-30 with 6 percent extender oil and 17 percent crumb
changed since its inception; therefore, there is still a need to rubber (14 percent car tire rubber and 3 percent tennis-ball
experiment using the recent changes to determine how it per- rubber by weight of asphalt cement). The crumb rubber (Table 2)
forms. Recently there has been much emphasis on recycling was required to produce an asphalt rubber binder with the
tires to prevent filling landfills with them, and many state characteristics shown in Table 3. Crafco, Inc., which has had
legislatures have mandated that the use of tires in highway considerable experience with asphalt rubber, designed the
construction be studied. In 1990, Virginia Senate Bill No. asphalt rubber mixture.
287, which encouraged demonstration projects using ground
rubber from used tires in road surfacing, was passed, and an
experimental field project using asphalt rubber concrete was Testing
installed in Fairfax County.
The purposes of this paper are to summarize the results of Marshall, gyratory testing machine (GTM), creep, resilient
the field project, which are primarily laboratory results, and modulus, indirect tensile, and stripping tests were conducted
to report on the initial data from a follow-up laboratory study on mixtures sampled during construction.
Marshall tests were conducted according to ASTM D1559
Virginia Transportation Research Council, Box 3817, University Sta- (2) using the 75-blow compactive effort. Properties that were
tion, Charlottesville, Va. 22903-0817. evaluated were voids in the total mixture (VTM), voids filled
TABLE 1 GRADATIONS OF MIXTURES TABLE 3 PHYSICAL PROPERTIES OF

ASPHALT RUBBER BINDER
%Passing
Minimum viocooity, 360°F 1500 cp
Sieve Test Sections Lab Study Minimum cone penetration, 77'F (ASTM 01191) 20
Minimum ooftening point (ASTM 036) 126'F
314 100
Minimum reoilience, 77'F (ASTM 03407) 16%
112 99 100
318 91 96
4 66 68
8 40 42
30 22 26 Stripping tests performed were the indirect tensile strength
60 13 17 test (ASTM D4867)(2) and the Virginia boiling test for field
100 11
200 5.4 6.2 mixtures (3). The tensile strength test determined suscepti-
bility to moisture damage by testing two sets of specimens:
one set was conditioned to simulate potential moisture dam-
age and tested, and one set was tested in an unconditioned
TABLE2 SPECIFIED
GRADATION OF CRUMB
dry state. The tensile strength ratio, which is the ratio of the
RUBBER conditioned strength to the dry strength, was used to predict
moisture damage. The boiling test was performed by boiling
Sieve % Passing
a sample of mixture in water for 10 min. The sample had to
#10 100 be as well coated as an unboiled dry sample in order to pass
#16 96-100 the test.
#30 70-100
#80 0-20
#200 0-6
Note: No gnidation tests were performed. Results
The design criteria and Marshall properties for the mixtures

with asphalt (VFA), voids in the mineral aggregate (VMA), are listed in Table 4. The VTM was lower than desirable, and
and stability. the VFA was higher than desirable for the southern section
Under the oil-filled mode of operation, specimens were of asphalt rubber mixture. The VTM was at the upper limit
tested with a GTM using initial gyratory angles of 1 and 0.75 for the northern control mixture and slightly greater than the
degrees with a vertical pressure of 120 psi. The specimens upper limit for the southern control mixture. If the Marshall
were compacted until the rate of compaction decreased to void properties are indicative of pavement performance, then
1 pcf/100 revolutions, which simulates the level of compaction it is to be expected that the southern asphalt rubber section
after traffic. The angles of gyration are thought to produce may tend to bleed or become overly dense, and the control
compaction that duplicates the compaction of moderate to sections may tend to be open and age prematurely.
heavy traffic. Properties used to characterize the mixtures The GTM results-predicted voids after traffic, shear
were final voids, shear strength, and gyratory stability index strength, and gyratory stability index-are displayed in Fig-
(GSI). ures 1, 2, and 3, respectively. Both asphalt rubber mixtures
Compression creep tests were performed at 104°F on 2.5- failed all of the tests, and the southern control mixture failed
in. by 4-in.-diam specimens that were compacted on the GTM shear strength at both pressure levels and GSI at 120 psi.
to the predicted void content of pavement after traffic. The These test results indicate that the asphalt rubber mixtures
specimens were preloaded for 2 min at 30 psi, unloaded and and southern control mixture may deform if the traffic is
allowed to rest for 5 min, and then loaded for 60 min at 30 severe enough. Because the properties of the binders may be
psi. Axial deformation was recorded at set intervals when the similar at high temperatures (as used in this test) but very
load was being applied and again for 60 min after the load different at lower summer pavement temperatures, perfor-
was released. The stiffness modulus at 60 min and the un- mance of the asphalt rubber mixtures may have been pre-
recovered strain after the load had been released for 60 min dicted better by performing the traffic compaction simulation
were analyzed. at lower temperatures.
The indirect resilient modulus and indirect tensile strength A summary of the results of the creep test are listed in
at 104°F were measured according to ASTM D4123 (2) with Table 5. It was expected that the addition. of rubber would
the Retsina device using the same specimens that were used increase the modulus and decrease the unrecovered strain;
in the creep tests. Because vertical deformation could not be however, the asphalt rubber mixtures had a lower modulus
measured, Poisson's ratio was assumed to be 0.35. and a higher unrecovered strain compared with the same
TABLE 4 MARSHALL DESIGN CRITERIA AND TEST RESULTS

Southern Northern Design Criteria
Property Rubber Control Rubber Control Rubber Control
V'l'M(%) 2.4 6.1 3.8 6.0 3-4 4-6

VFA(%) 87 66 80 67 60-76
VMA(%) 17.7 17.6 18.9 18.3 <!16 <!15
Stability Obol 2680 2960 2660 3120
Maupin 11
4.11 l.S . - - - - - - - - - -- - -- - - - - - - - - - - - - ,
1.4
4
1.3
3.11 1.2
a I.I
3 l!i
i! 2..11
5 0.9
= D.8
E 2 .. 0.7
D.B
u 0.5
D.4
0.3
D.2
D.11
0.1
D D
SOUTH CONT NORTH CONT SOUTH RUB NORTH RUB SOUTH CONT NORTH CONT SOUTH RUB NORTH RUB
TYJIE or MIXTURE TYPE or MIXTURI!
- D.79 DEGREES GJ 1.0 DEGREES - 0.71 DEGREES 0 l.D DEGREES
FIGURE 1 Pavement voids predicted by GTM. FIGURE 3 Gyratory stability index.
properties of the control mixtures. The average differences resistance of the asphalt rubber is minimal. Also, the high
were significant at a 95 percent confidence level. Although temperature during the GTM traffic-compaction simulation
some research has shown that asphalt rubber increased stiff- allowed the specimens containing asphalt rubber to densify
ness and strength (4,5), other research has shown that asphalt more than is normally observed in pavement, thereby result-
rubber decreased the stiffness of mixtures that were already ing in low air voids and low values for strength and modulus,
performing well with a normal binder (6). which may be misleading.
The resilient modulus of the southern control mixture was sig- The tensile strength stripping test indicated that the control
nificantly higher than the resilient modulus of the northern con- mixtures may be susceptible to stripping. The southern control
trol mixture at a 95 percent confidence level (see Figure 4). mixture had a TSR less than the acceptable 0.75, and both
It was anticipated that rubber mixtures would follow the trend control mixtures had considerable visual stripping. None of
observed with the creep test, in which the modulus of the the mixtures failed the boiling test.
rubber mixtures was lower than that for the control mixtures;
however, the reason for the difference between the control
mixtures could not be explained. LABORATORY STUDY
The indirect tensile strengths of the rubber mixtures (see
Figure 5) were significantly lower than the control mixtures Several gradations of crumb rubber that are used in other
at a 95 percent confidence level, which was the same trend applications are available for use in asphalt rubber. A follow-
previously observed. up laboratory study was conducted to determine how the
In those cases in which the GTM was used for testing or properties of asphalt rubber concrete containing coarse crumb
compacting the specimens, the results are probably biased rubber compared with the properties of asphalt rubber con-
favorably toward mixtures without rubber. The GTM test was crete containing fine crumb rubber. When rubber is mixed
performed at high temperatures at which the deformation with asphalt cement, the rubber particles swell (react) causing
the viscosity to increase, and if heat is maintained for a pro-
longed time, the rubber may melt and break down, resulting
ID
in an undesirable decrease in viscosity. Reaction curves were
developed for asphalt rubber binders containing various per-
5D centages of fine and coarse rubber to determine whether the
~ fine rubber should be used differently in the field. It is de-
E 40 sirable to use the asphalt rubber binder after it has reached
1:1 its maximum viscosity but before the rubber breaks down;
E 30 however, rubber breakdown is not the only concern. If a short
..•B
.. 20
time is required to achieve maximum viscosity, the fine rubber
will require less elaborate blending equipment than the coarse
rubber. Perhaps the fine rubber could be added during the
I ID
TABLE 5 CREEP TEST RESULTS
0 Unrecovered Strain(%)
SOUTH CONT NORTH CONT SOUTH RUB NORTH RUB Sections Modulus, (psi)
Average Std. Dev. Average Std. Dev.
TYJIE OF MIXTURE
- 0.79 DEGREES 0 1.0 DEGREES Control 7,900 900 0.080 0.027

Asphalt Rubber 5,500 560 0.200 0.074
FIGURE 2 Gyratory shear strength.
Marshall designs were performed on each binder at each

rubber content to determine the optimum asphalt content at
90 4.5 percent VTM . The remaining tests-gyratory shear, re-
80
silient modulus, and indirect tensile-were performed at the
optimum asphalt content.
Although the GTM was used, the procedure was changed
11- :: from the procedure used in the study of the field mixtures.
i
SE. so Only one initial gyratory angle of 1.0 degree and a vertical
pressure of 120 psi were used . Also , the specimen was com-
a.. 30
pacted to the expected density of pavement after rolling, re-
moved from the machine, cooled to 140°F, and compacted
20
until the rate of compaction decreased to 1 pcf/100 revolu-
10
tions. Compacting at 140°F to simulate the effect of traffic
0
SOUTH CONT NORTH CONT SOUTH RUB NORTH RUB
was done in an attempt to show the potential advantage of
TYPE OF MIXTURE rubber in resisting permanent deformation at normal pave-
0. 7S DEGREES c:::::J 1.0 DEGREES
ment temperatures.
Resilient modulus and indirect tensile tests were performed
FIGURE 4 Resilient modulus at 104°F. on the specimens made on the GTM by the same procedures
that were used in the study of field mixtures.
mixing of the binder and aggregate without blending equip-
ment.
Results
Materials The reaction curves for coarse and fine rubber are shown in
Figures 6 and 7, respectively . It is obvious that there is con-
A 1/2-in. surface mixture using an AC-20 asphalt cement and siderable variation in the test results because the curves were
a 75-blow Marshall design was used (Table 1). The gradation not as smooth as expected. The variation is caused by diffi-
of the coarse rubber was the same as that used in the field culty in maintaining a constant temperature when removing
experiment, and the fine rubber had 100 percent passing the a binder from the oven for viscosity testing, and the accuracy
No. 80 sieve and a mean particle size of No. 200 (0.074 mm). of the Haake viscometer is limited. It appears that the vis-
cosity reached a maximum after approximately 1 hr with all
of the binders except the one containing 15 percent coarse
Testing rubber, which appeared to still be gaining viscosity after 20
hr. For a rubber content of 10 percent or less, the reaction
Reaction (viscosity-time) curves for the asphalt rubber bind- appears to be almost instantaneous. The rubber did not ap-
ers were developed using rubber contents of 5, 10, and 15 pear to break down after 20 hr, which would have resulted
percent for both the coarse and fine crumb rubber . A portable in a decrease in viscosity; therefore, these binders could be
Haake viscometer was used to measure viscosity. The rubber safely stored at 350°F for this time period without deleterious
was mixed with the asphalt cement and maintained at 350°F, effects. It is evident from both sets of curves that viscosity
and viscosity measurements were taken at regular intervals increases significantly between 10 percent and 15 percent rub-
for approximately 24 hr. ber content. The engineering properties would also change
130 24
120 22
110 20
~ 100
18
~
90
E
80 .
t1
5
16
14
70 ~
..~~ 80
50
t
8..
12
10
5 8
~ 40
i 30
20
6
10 2
0 0
SOUTH CONT NORTH CONT SOUTH RUD NORTH RUB D 2 4 6 e w ra ~ m m ~
TYPE OF MIXTURE TIME (HRS)
- 0.75 DEGREES c:::::J 1.0 DEGREES D '!i'/, CR + Ill"/, CR o 15'1. CR
FIGURE 5 Indirect tensile strength at 104°F. FIGURE 6 Reaction curve: coarse crumb rubber.
Maupin 13
24 21.5
22
0 21 -
(
20
20.5
18
18
~ 20
i:l!!! :ii
"e •• ~
:E
19.5
..
12
E ~ Ill
..
"u
5
10
8
"c
!1l
18 .5
6
1.8
• 17,5
2
0 17
0 2 ID 12 14 16 18 20 0 2 4 6 8 ID 12 14
TIME, HRS RUBBER CONTENT ('!.)
5'!. FR x ID'!. FR v 15'!. FR c nNE CRUMB RUBBER + COARSE CRUMB RUBBER

"
FIGURE 7 Reaction curve: fine crumb rubber. FIGURE 9 Voids in mineral aggregate at optimum asphalt
content versus rubber content.
dramatically in this range. There does not appear to be a more flexible than the fine rubber mixtures with lower asphalt
significant difference in the viscosity of the coarse and fine content.
rubber at a similar rubber content, and both would be ex- The predicted voids after traffic, shear strength, and sta-
pected to behave in a similar way in blending operations in bility index are shown in Figures 10, 11, and 12, respectively.
the field. The mixture with 15 percent coarse rubber had a significantly
The "optimum" asphalt content, which was selected at 4.5 higher percentage of voids (i.e., less densification) than the
percent VTM, and the associated VMA from the Marshall other mixtures, which indicates a possible greater resistance
designs are shown in Figures 8 and 9, respectively. The change to permanent deformation. The GSI of the mixtures with 10
in asphalt content with change in rubber content (i.e., slope percent and 15 percent rubber was greater than 1.1, which
of the curve) was consistent between the coarse and fine rub- indicates that they may have contained too much binder and
ber mixtures. The mixtures with coarse crumb rubber required would be prone to instability. These results are contradictory;
approximately 0.6 percent more binder than similar mixtures therefore, the method of testing and analyzing asphalt rubber
with fine crumb rubber. Because the rubber particles do not mixtures with the GTM needs further study. The shear strengths
melt completely, they tend to push the aggregate particles showed no trends for mixtures with different rubber content;
apart, as indicated by an increase in VMA in all mixtures that however, all of the strength values were very low.
have a higher rubber content. As expected, the coarse rubber The bar graphs of resilient moduli and indirect tensile
particles created higher VMA than the fine rubber particles strengths (Figures 13 and 14) of the specimens made on the
because the larger particles forced the aggregate particles fur- GTM revealed some trends. There appeared to be an opti-
ther apart. If they are stable, it would be anticipated that the mum rubber content at which the maximum value of resilient
coarse rubber mixtures with higher asphalt content would be modulus and indirect tensile strength was achieved for the
7.8 4 .5
7.5
7.4 4
7.3
7.2
t 7.1
3 ,5
~
7
8.9
3
B 6 .8
6.7
. 2.S
.....
~ 8.8
8.5 ~ 2
c 6.4
Ii 8.3 l.S
"
I! 6 .2
~ 8.1
" 6
5.9
0.5
5 .8
5 .7
5 .8 0
0 2 4 6 8 ID 12 14 0 s 10 18
RUBBER CONTENT ('!.) '!.RUBBER
o nNE CRUMB RUBBER + COARSE CRUMB RUBBER - COARSE RUBBER 0 FINE RUBBER
FIGURE 8 Optimum asphalt content versus rubber content. FIGURE IO Pavement voids predicted by GTM.
40
45
~
E ~ 35
i:i .,_
Ill
30
E
a. 20
I] 25
Ill
=
15
1- 20
E 10 ..
1:1 15
~ 10
0 0
0 5 10 IS 0 5 10 15
'/, RUBBER '/, RUBBER
- COARSE RUBBER c:J nNE RUBBER - COARSE RUBBER c:J FINE RUBBER
FIGURE 11 Gyratory shear strength. FIGURE 13 Resilient modulus at 104°F: laboratory study.
fine rubber mixtures for both resilient modulus and indirect using the wet process may be less resistant to permanent
tensile strength. There was a significant difference at a 95 deformation than mixtures without rubber; however, labo-
percent confidence level between average values at 0 and 5 ratory tests may not be able to simulate pavement behavior
percent, 5 and 15 percent, and 10 and 15 percent, but not for these types of mixtures.
between 5 and 10 percent, which confirms the optimum value 2. The mixtures with asphalt rubber displayed less stripping
observation. This optimum condition was not apparent for than the mixtures without asphalt rubber when tested with
the coarse rubber mixture. The magnitude of the resilient the indirect tensile stripping test.
modulus for the coarse rubber appeared to increase as the
rubber content was increased. When the average values were
tested, there was no significant difference between any of the Laboratory Study
average values; therefore, this apparent trend could not be
confi,rmed. 1. Coarse and fine rubber reacted with the asphalt cement
similarly.
2. The rubber did not break down over a 24-hr time period.
CONCLUSIONS 3. Approximately 0.6 percent more asphalt cement was re-
quired for the coarse rubber mixtures than for the fine rubber
Field Study mixtures.
4. An optimum rubber content of 5 to 10 percent yielded
1. Laboratory test results on the mixtures sampled during the maximum resilient modulus and indirect tensile strength
contruction indicate that mixtures containing asphalt rubber for mixtures containing fine rubber.
70
1.2
60
I.I
~
a 0.9 ~
I O.B
~ 40
I.. 0.7
0.8
o.s
~
~
~
30
I
20
0.4
I
!
0.3 10
0.2
0.1 0
0 5 10 15
0 '/,RUBBER
0 5 10 15
'I• RUBBER - COARSE RUBBER c:J nNE RUBIER
- COARSE RUBBER c:J FINE RUBBER

FIGURE 14 Indirect tensile strength at 104°F: laboratory
FIGURE 12 Gyratory stability index. study.
Maupin 15
REFERENCES in Airport Pavements. In Proc., National Seminar on Asphalt Rub-

ber, Phoenix, Ariz., 1989, pp. 191-245.
l. W. J . Halstead. Potential for Use oj' Discm·ded Tires in Highway 5. M. Stroup-Gardiner, N. Krutz, and J. Epps. Comparison of Mix
Construc1io11 a11d Mai111e11ance. VTRC 90-TAR6, Virginia Trans- Design Methods for Rubberized Asphalt Concrete Mixtures. In
portation Research Council, Charlottesville, 1990. National Seminar on Asphalt Rubber Proceedings, Phoenix, Ariz.,
2. Annual Book of ASTM Standards: Vol. 04.03: American Society 1989, pp. 82-117.
for Testing and Materials, Philadelphia, Pa., 1990. 6. T. S. Schuler, R. D. Pavlovich, J. A. Epps, and C. K. Adams.
3. Virginia Test Methods Manual. Virginia Department of Trans- Investigation of Materials and Structural Properties of Asphalt-
portation, Richmond, 1984. Rubber Paving Mixtures. FHWA/RD-86/027, Texas Transporta-
4. D . M. Hoyt and R. L. Lytton. Laboratory Behavior, Performance tion Institute, College Station, Texas, 1986.
Prediction, and Cost-Effectiveness of Asphalt-Rubber Concrete
Florida's Approach Using Ground Tire

Rubber in Asphalt Concrete Mixtures
GALE c. p AGE, BYRON E. RUTH, AND RANDY C. WEST
In 19 8 under a legislative mandate, the Florid a Department of material. The term asphalt-rubber in this report is defined as
Tra nsportation (FDOT) began a concentrated effort to evaluate a binder with GTR blended in a paving-grade asphalt cement.
the potential u es for reclaimed tire rubber in asphalt pavement
con truction . FOOT indica ted that the most advantageous use of
rubber would be as a binder modifie r to improve the performance
of friction course mixture.. T hree demonstra tion projects were BACKGROUND INFORMATION
constructed . T he field construction opei:aiions with the rubber-
modified mixtures were essentially the same a those with con- The first investigation conducted by the FDOT in the use of
ventional friction course mixtures. Curre nrl y all of the rest asphalt-rubber for highway construction was performed nearly
sections are performing well. The optimum rubber content for 10 years before the passage of Senate Bill 1192. That project
den e-graded frict.io11 course mixtures has been identified as 5
percent (by weight of a phalt cement) using a maximum nominal was to evaluate asphalt-rubber as a stress-absorbing interlayer
80-mesh ground tire ruhher. Tt is helieved that the rubbe r will and a binder for seal coat construction. A demonstrntion proj-
provide improved elasticity to the binder and therefore greater ect constructed on SR 60, Hillsborough County, was used to
resilience for the e mixtures in recovery from high strains at in- evaluate the performance of asphalt-rubber in these appli-
tersections. T he optimum rubber contelll fo r open-graded friction cations. The results of this study are documented in an August
cow·se mixtures wa de termined to be .12 percent (by weight of 1980 report prepared by the FDOT for the U.S . Department
asphalt cement) using a maximum nominal 40-mesh ground tire of Transportation (J) . As a result of this demonstration proj-
n_1bber. ~n op e n-~rad e d mixture , the rubber has allowed a sig-
mf1cant increase. ia the rota.I binder conrent, which increa ·ed in ect, the FDOT has permitted the use of GTR in selected
the film thickness on the aggregate particle resulting in improved surface treatment and interlayer construction. In addition ,
~ur a.b.ility . On the basis of lhe-se de monstration project , speci- FDOT currently permits the use of GTR in certain joint seal-
flcauon have been developed for using ground ti re rubber in ers and in railroad crossing pads.
friction course mixtures as a standard practice. Upon passage of enate Bill 1192 by the 1988 Florida Leg-
islature, FDOT personnel in cooperation with University of
The provisions of Section 336.044(3) of the Florida statutes Florida researchers, established and implemented a detailed
created by Senate Bill 1192 in 1988 directed the Florida De- plan to address the legislative mandate. The relatively short
partment of Transportation (FDOT) to expand , where fea- time period allocated for this investigation required concur-
sible , its use of recovered (waste) materials for highway conrent activities . One primary activity was to document perti-
struction. Specifically, the bill directed that an investigation nent information from technical literature on asphalt-rubber
be conducted to determine how ground tire rubber (GTR) and its application in asphalt concrete mixtures. The National
from recycled waste tires could be used in quality asphalt Center for Asphalt Technology (NCAT) at Auburn Univer-
concrete mixtures for highway construction by undertaking sity was selected to conduct this investigation because of the
demonstration projects as part of currently scheduled con- knowledge and experience of the investigators with asphalt-
struction projects. It further stipulated that within 1 year after rubber, paving mixtures, and construction processes. Their
the conclusion of the demonstration projects the FDOT should report, dated August 1989, provided a comprehensive doc-
report to the governor and the legislature on the maximum umentation of material properties , benefits , limitations, and
percentage of GTR that can be effectively used in road con- recommendations for the use of GTR and asphalt-rubber
struction projects. Concurrently with this report, the FDOT binders for asphalt concrete mixtures (2). This state-of-the-
should review and modify its standard road and bridge con- art overview of asphalt-rubber in an asphalt concrete appli-
struction specifications to allow and encourage the use of GTR cation confirmed and validated the direction of FDOT in the
consistent with the findings of the demonstration projects. development of the subsequently constructed demonstration
The purpose of this report is to provide a concise overview projects.
of all FDOT and University of Florida activities pertaining to
the development of the use of GTR in asphalt-rubber binders
for specific asphalt concrete mixtures and other highway con- DEMONSTRATION PROJECT DEVELOPMENT
struction applications, and to document the steps taken by
the FDOT to facilitate the use and quality control of this The purpose of the demonstration projects is to evaluate the
constructibility and short-term field performance of different
G. C. Page and R . C. West , Florida Department of Transportation amounts and sizes of GTR in a number of plant-produced
P.O. Box 1029, Gainesville, Fla. 32602. B. E . Ruth, Department of FDOT asphalt concrete mixtures in order to develop speci-
Civil Engineering, University of Florida, Gainesville, Fla . 32611 . fications and procedures for its use .
Page et al. 17
Current standard specifications for gradation and mixture TABLE 1 SUMMARY OF ASPHALT-RUBBER
properties would continue to be used to determine acceptable DEMONSTRATION PROJECTS
characteristics. This is a conservative approach believed to be ht fwi"t ~D~ f[!,2i"t J[~ f[gj~t
warranted at this time, and is consistent with the legislative
Date: March 1989 June 1989 September J990
requirements for this investigation. Location: N.E. 23 Ave. State Road 16 1-95
Gainesville, FL Starke, FL St. Johns Cty
A number of decisions were made about the demonstration
projects that were based on the relatively short time frame Mix Type: Dense-Graded Open-Graded Open-Graded
(FC-4) (FC-2) (FC-2)
(approximately 2 years) for the development of specifications
and procedures imposed by the legislation. The FDOT mix- Test Section (1) 80 mesh/3.1 % (I) 80 mesh/5 .3% (1) 80 mesh/10%
tures for the demonstration projects were limited to the fric- GTR Size/% (a) (2) 80 mesh/5. 3% (2) 80 mesh/ LL.L % (2) 80 mesh/10%
(3) 40 mesh/I I.I% (3) 80 mesh/17.7% (3) 80 mesh/10%
tion course mixtures both dense-graded (FC-1 and FC-4) and (4) control/0% (4) 24 mesh/20.5% (4) 80 mesh/10%
open-graded (FC-2). This was based on two considerations: (5) control/0% (5) Control/ 0%
(6) 80 mesh/I I.I %(b) (6) Control/ 0%
(7) Control/ 0%
(8) Control/ 0%
1. Improvement in the properties of these mixtures had
previously been identified as desirable: improved durability Tota] Binder (1) 7. 1% (I) 8.0% (L) 7.781'i
Content (2) 7.3% (2) 8.4% (2) 7.78~
and resistance to shoving at intersections for the dense-graded (3) 8.2% (3) 11.4% (3) 7.7811.
mixtures, and increased binder film thickness for improved (4) 7.0% (4) 10.3% (4) 7.78%
(5) 6.3% (5) 6.30%
durability and aggregate retention, with improved resistance (6) 6.9% (6) 6.30%
(7) 6.30%
to binder drainage for the open-graded mixtures. (8) 6.30%
2. The compatibility of the GTR with the efficacy of a
Test Section (I) 3520 (1) 2100 (I) 5260
recycling agent (soft asphalt) to rejuvenate the existing asphalt Length - ft. (2) 3656 (2) 2532 (2) 5655
cement in reclaimed asphalt pavement (RAP) was an un- (3) 2460 (3) 1818 (3) 5513
(4) 2640 (4) 2880 (4) 5937
known. Therefore only mixtures using virgin components would (5) 1761 (5) 5280
(6) 263 (6) 5280
be included. (7) 5280
(8) 5280
It was decided to concentrate on the use of fine ( - 80 mesh)

GTR at relatively low percentages in the demonstration proj- (a) By weight of asphalt cement. GTR contents originally specified as a percent of total binder.
(b) Not preblended - mixed in pugmill
ects based on previous experience in the laboratory in ob-
taining acceptable mix properties. These initial laboratory
results are not included in this report. The first demonstration project was constructed in Gaines-
It was also decided to concentrate on the process under ville during March 1989 using a dense-graded friction course
which the GTR was preblended or dispersed in the asphalt (FC-4) containing 3.1, 5.3, and 11.1 percent GTR by weight
cement before mixing with the aggregates. This was done on of asphalt cement. (Note: Rubber contents were originally
the basis of a "common sense" approach assuming that if the specified as a percentage of total binder. As such, these sec-
GTR "reacts" or "swells" in the asphalt cement, then that tions contain 3, 5, and 10 percent GTR by weight of total
process should take place under controlled conditions before binder. A decision was made later to specify GTR as a per-
mixing with aggregates in the asphalt plant. centage of asphalt cement to simplify calculations. All amounts
of GTR in this report are shown as a percentage of asphalt
cement.) Dense-graded friction course mixtures were found
to be generally more susceptible to change in binder content
DEMONSTRATION PROJECT CONSTRUCTION and particle size of GTR than open-graded mixtures. Tests
conducted on the hot-mix samples with different levels of
From 1989 through 1990, three demonstration projects were GTR indicated that the mix from Test Section 2 with 5.3
constructed to evaluate the use of GTR in asphalt concrete percent GTR appeared to be the mix for which the standard
friction course mixtures. A summary of key information for
the three demonstration projects is presented in Table 1. The
specification requirements for both the dense-graded friction TABLE 2 FRICTION COURSE MIX
course (FC-4) and open-graded friction course (FC-2) are REQUIREMENTS
shown in Table 2. Each project required a substantial prelim- Mix Type
RMuiumcn'I D«mn!'Dded CFC--41 Opcp K!lllosl IFC-21
inary effort to ensure the best possible operational conditions
for production, construction, and testing of materials evalu- Gradation (% passing)
1/2 100 100
ation. This involved development of work plans, special pro- 3/8 85-100
No. 4 1040
visions, mix designs, laboratory testing, and considerable No. 10 75-90 4-12
interaction with the prime asphalt contractor and the sub- No. 200 2-8 2-8
contractor providing the blending of GTR with the asphalt Marshall Properties
cement. During construction, extra sampling and specialized Min. Slability 500 Lbs .
Max. Flow 8-16
tests were performed in addition to the standard quality-control
Void Criteria
and quality-assurance tests. A concentrated effort was re- Min. VMA 15%
quired to furnish a sufficient number of qualified personnel Air Voids 12-16
to conduct these tests and to observe construction procedures Asphalt Content

for assessment of any problems or deficiencies. Min. Effective 5 .0%
specification requirements were met, and in addition had use GTR in a modified binder for friction course construction
increa ed re istance to shear as measured in the Corps of without any major change in construction operations. These
Engineers Gyratory Test Machine. Although all of the a phalt- projects also verified that current standard specified criteria
rubber mixtures exhibited some degree of sticking to the pav- for friction course mixtures (as shown in Table 2) could be
ers' screed, it was only considered excessive during paving of met at design and during production for mixtures with an
Test Section 3, which had 11.l percent GTR. Otherwise, no asphalt-rubber binder. In addition , current standard accep-
major problems were encountered during construction of these tance tests and criteria could be applied and met with the
asphalt-rubber friction courses. The data, discussion, and con- exception of modifying the method of measurement for the
clusions for this first demonstration project are contained in asphalt-rubber binder.
a separate report (3). Although the long-term performance of these pavements
The second demonstration project was constructed on SR cannot be evaluated until some time in the future, sufficient
16 near Starke in June 1989 with 4 sections using 5.3 to 20.5 test data and corroborating information suggest that asphalt-
percent GTR in an open-graded friction course (FC-2). Con- rubber friction courses, particularly open-graded, will have
struction was accomplished without any significant difficulty improved durability over conventional friction course mix-
or observable problems. Test Sections 3 and 4 with 17.7 and tures. This improvement is related to (a) reduced age hard-
20.5 percent GTR, respectively, had high total binder con- ening because of anti-oxidants in the rubber and increased
tents that could result in long-term performance and hydro- film thickness, and (b) improved retention of aggregate be-
planing problems. The results obtained from construction of cause of increased film thicknesses and greater resiliency of
this demonstration project indicated that about 10 to 15 per- the binder. Greater binder contents and the retention of thicker
cent GTR can effectively be used in open-graded friction binder films on the aggregate are possible because of the
course mixtures, but the total binder content for mixtures increase in viscosity produced by the addition of GTR.
with this rubber content should probably be less than that
used in mixtures on this project. The evaluation of binder
content relied to a large extent on subjective visual deter-
minations in the field. The data, discussion, and conclusions TYPE AND AMOUNT OF GTR IN FRICTION
for this second demonstration project are contained in a sep- COURSE MIXTURES
arate report (4).
The University of Florida provided technical assistance and The type of rubber currently determined to be satisfactory
documentation of these demonstration projects (3,4). A re- for use in asphalt-rubber friction course mixtures is that pro-
port prepared by the FDOT Materials Office (5) also provides duced by ambiently grinding tires to a suitable fineness (2).
a general overview and summary of FDOT involvement through Cryogenically produced rubber is not currently acceptable
the construction of the first two demonstration projects. Of because the effect of its smooth-faced particles on reaction
primary importance to the development of draft specifications time and the material properties of the modified binder has
were the preliminary laboratory investigations for each of the not been evaluated.
demonstration projects conducted by the FDOT to establish The amount and fineness (gradation) of the GTR to be
asphalt-rubber blends, verify blend times, and develop mix used in asphalt-rubber blends is based on the application. In
designs. Other special studies were conducted to evaluate dense-graded friction course mixtures, 5 percent of GTR pass-
asphalt-rubber blending requirements and the effectiveness ing the No. 50 sieve (e.g., a maximum nominal 80 mesh) is
of extraction testing (6). recommended. In open-graded friction courses, 12 percent of
The third and last demonstration project was constructed GTR passing the No. 30 sieve (e.g., a maximum nominal 40
on Interstate 95 during September 1990 using 10 percent GTR. mesh) is recommended to be blended with the asphalt cement.
The purpose of this project was to determine whether asphalt- Open-graded mixtures are more tolerant of larger rubber par-
rubber could be blended and incorporated into an open-graded ticulate size and greater GTR contents. From experience of
friction course mixture using a prototype continuous produc- these demonstration projects, it was found that the calcula-
tion blending unit on a conventional construction project with- tions for blending are simplified if the amount of GTR is
out encountering any problems that would contribute to con- specified as a percentage of the asphalt cement rather than
struction defects or delays. The information collected on this of the total binder.
demonstration project is documented in a technical report Another application of GTR is in the asphalt-rubber binder
from the University of Florida (7). This demonstration project for an asphalt-rubber membrane interlayer. In this case about
was constructed without any major technical problems. How- 0.6 gal/yd 2 of asphalt-rubber binder is sprayed over the pre-
ever, the blending time required to provide adequate reaction pared pavement surface and uniformly sized aggregates are
of GTR with the asphalt cement had to be increased with this spread and rolled into the membrane before placement of the
prototype blending unit because of the lower-than-anticipated asphalt concrete structural layers. This asphalt-rubber blend
temperature (275°F instead of 310°F) of the asphalt cement. uses 20 percent of GTR passing the No. 10 sieve (e.g., a
This indicated the need either to increase the blending unit maximum nominal 20 mesh). This provides a membrane that
capacity or provide additional heating for the unit to assure should seal the pavement from intrusion of moisture and re-
adequate blending to maintain hot-mix production rate at the tard reflective cracks, particularly for asphalt overlays of port-
desired 100 tons/hr. land cement concrete pavements.
The constructibility and short-term performance of these Requirements for the GTR and asphalt-rubber binder for
asphalt-rubber test pavements indicates that it is feasible to each application are presented in subsequent sections.
Page et al. 19
BLENDING REQUIREMENTS binder contents. Increased binder in combination with the

improved resilient properties of asphalt-rubber should pro-
GTR must be blended with asphalt cement for a sufficient vide improved aggregate retention and improved durability
period of time to achieve a uniform product with fairly stable and life. Limited performance measurements of pavement
consistency (usually determined by viscosity measurements). friction (ASTM E 274) and rut depth have been made on the
This "reaction" time is significantly reduced when using finer three demonstration projects and no differences have been
GTR, softer asphalt cements, and higher blending tempera- identified attributable to the use of GTR.
tures. This was identified in FDOT laboratory blending stud- The recycling of asphalt concrete pavements with asphalt-
ies as part of the demonstration projects. Another advantage rubber friction course surfaces is not anticipated to be a prob-
of fine GTR is that the resulting asphalt-rubber blend is more lem because of the low rubber content present in the total
homogeneous and is better suited for viscosity testing and amount of reclaimed asphalt pavement (RAP) for normal
other quality control tests than blends containing coarser GTR milling depths (2), and it is thought that the rubber has
(particle sizes retained on the No. 30 sieve). Although "re- "absorbed" all the asphalt it can and asphalt demand has
action" time is reduced at higher blending temperatures, hold- stabilized.
ing the blended asphalt-rubber at elevated temperatures for Issues of air quality and toxic fumes were not specifically
long periods will degrade the quality of asphalt-rubber binder addressed by FDOT in this evaluation, but data and reports
because of volatile loss and accelerated hardening. Field and from the Asphalt-Rubber Producers Group indicate that these
laboratory studies by FDOT have shown that holding the issues are of no more concern than those for asphalt cement.
blended asphalt-rubber binder at normal asphalt cement stor-
age temperatures (300° to 350°F) does not degrade the quality
ESTIMATED GROUND-TIRE RUBBER USAGE
of the binder for typical storage periods. These recent data,
AND COST
which are to be published, show that viscosity and softening
point increased slightly and resilience increased four-fold "dur-
The estimated annual use of GTR is based on the total ton-
ing storage. It is necessary, however, to provide periodic ag-
nage of open and dense-graded friction course mixtures nor-
itation of the blended binder to present separation of the
mally used during one construction year. In addition, the
GTR.
asphalt-rubber membrane interlayer is included in the GTR
Conventionally, GTR is packaged in plastic bags that are
use calculations, based on the estimated number of lane miles
opened and dumped into the hopper of a feeding unit. The
per year . These calculations and the yearly GTR usage pro-
feed of GTR and asphalt cement into a blending unit is
jections are as follows:
adjusted to achieve the desired percent GTR in the asphalt-
rubber binder. The size and operation of the blending unit Tons Rubber/Year
may differ according to the approach selected by the asphalt- Open-Graded Friction Course (FC-2)
rubber blending contractor. Blending at the asphalt cement Tons Tons Asphalt
640,000 Ye ar x 6.8% Asphalt = 43,520
terminal and shipping to the project site appear technically Year
and economically practical. The blending temperature and 43,520 Ton~Asphah x 12% Rubber 5,222
reaction time requirements are given in the developmental ear
specification for asphalt-rubber binder presented in subse- Dense-Graded Friction Course (FC-1 & 4)
quent sections. 160,000 Tyons x 7.0% Asphalt = 11,200 Tons Asphall
ear Year
EFFECTS ON CONSTRUCTION AND PAVEMENT 11,200 Ton~Asphalt x 5% Rubber 560
PERFORMANCE car
Asphalt-Rubber Membrane Interlayer
Properly proportioned asphalt-rubber binders can be used in L1nc Miles Gal
600 Year x 0.6 S.Y. x 20% Rubber 2,160
dense or open-graded friction course mixtures without any
significant effect on conventional mix production operations. Pavement Marker Adhesive, Joint Filler,
Railroad Crossing Pads, Guardrail Spacers 1,200
However, standard asphalt metering pumps on asphalt hot-
Total Estimated FOOT Usage Per Year 9,142
mix plants may not be adequate to handle the higher viscosity
binders . Plants with asphalt weigh buckets will generally op- The total yearly generation of waste tires in Florida was
erate without any problems provided that the spray bar ori- estimated based on 0. 75 tire/yr/capita and the 1990 population
fices do not restrict flow. in Florida of 13,000,000. Approximately 10 lb of GTR is
Conventional paving operations for friction course mixtures recovered from each tire, which results in 48,750 tons of
can be used for the paving of asphalt-rubber mixtures. Long- rubber/yr .
term performance data do not exist for asphalt-rubber mix- On this basis , the FDOT can consume about 20 percent of
tures, but the following performance effects are inferred. Dense- the generated waste tire rubber in highway construction ap-
graded friction course mixtures with asphalt-rubber should plications. Because the amount of road construction activity
tend to reduce pavement distortions at intersections in urban done by cities, counties, and developers exceeds that used by
areas because of the improved resilient properties of the FDOT on an annual basis , it is assumed that their use of GTR
asphalt-rubber. Open-graded friction course mixtures would equal or exceed that of the FDOT. Therefore , the
with asphalt-rubber will tend to reduce or eliminate binder projected highway construction usage of GTR from waste
drainage from the aggregate in trucks even with increased tires in Florida is estimated at less than one half of the total
generated per year. However, the amount of waste tires avail- Florida Department of Transportation Standard Specifications
able is questionable because at the present time two major for Road and Bridge Construction, which pertains to Asphaltic
national suppliers of GTR already obtain some of their waste Concrete Friction Courses, was revised to require the use of
tire supply from Florida. Also, the roofing and tire manufac- asphalt-rubber binder in friction course construction.
turing industry incorporate GTR in some of their products. It should be recognized that these developmental specifi-
Consequently, the exact status of usage cannot be determined cations may be further revised before actual implementation
unless a very detailed and comprehensive study and inventory with input of additional data and review. Furthermore, as
is undertaken. experience is gained on asphalt-rubber construction projects,
Cost estimates performed by the FDOT State Materials it is probable that some modifications to the specification will
Office indicate that an optimistic increase in cost of $4.80/ton be needed to improve their effectiveness.
of mix, or about a 15 percent increase in cost, would occur These specifications are not meant to be the only approach
when using GTR in the binder (assuming $32.00/ton of con- to incorporating GTR into asphalt concrete mixtures. It is a
ventional hot mix). This additional cost translates into an documentation of the approach taken by FDOT in using GTR
increase in binder cost of about 70 percent. This cost estimate incorporated into asphalt cement as a modified binder to im-
is based on using asphalt rubber binders for all FDOT friction prove specific asphalt concrete mixtures currently used by
course mixtures on a continuing, not an experimental, basis. FDOT.
It includes reasonable costs for materials and processing. It
should be noted that the third demonstration project (I-95)
went through the normal bid process. The bid price (yd 2 ) for FLORIDA DEVELOPMENTAL SPECIFICATION
mix with GTR was 31 percent higher than mix without. This FOR GROUND TIRE RUBBER
project contained 4 lane miles with GTR in the open-graded
friction course (FC-2) compared with the remainder of the Scope
project with more than 30 lane miles of FC-2 without GTR.
Others have experienced substantially higher costs for specific The specification controls GTR for use in asphalt-rubber bind-
limited experimental construction (2). How, or whether, this ers for use in a variety of road and paving applications. The
increase in cost is funded is beyond the scope of this engi- specification does not address any safety or environmental
neering investigation but is a definite area of concern. In concerns associated with its use.
addition, there may be other asphalt additives that can have
the same effect in these mixtures. There is a concern that General Requirements
they should be able to compete economically, but the direc-
tion is to specify GTR exclusively. The GTR should be produced by ambient grinding methods.
It should be sufficiently dry so that it is free flowing and
foaming is prevented when it is mixed with asphalt cement.
APPROACH TO THE DEVELOPMENT OF The rubber should be substantially free from contaminants
SPECIFICATIONS including fabric, metal, mineral, and other nonrubber sub-
stances. Up to 4 percent (by weight of rubber) of talc (such
It was necessary to develop new specifications and to revise as magnesium silicate or calcium carbonate) may be added to
existing specifications before attempting to use asphalt-rubber prevent sticking and caking of the particles.
friction course mixtures in construction on a conventional
production basis. Therefore, the FDOT State Materials Office
Physical Requirements
prepared tentative or developmental specifications for use on
these construction projects. The current draft of these spec-
•Gradation: when tested in accordance with ASTM C-136
ifications was prepared using the compilation of information
using a 50-g sample, the resulting rubber gradation should
generated during the asphalt-rubber friction course demon-
meet the gradation limits shown in Table 3 for the type of
stration projects.
rubber specified.
The developmental specifications are presented in the fol-
• Specific gravity of the rubber as determined by ASTM
lowing sections. The specification for GTR for use in asphalt-
D-297, pycnometer method, should be 1.15 ± 0.05.
rubber binder was developed with input from tire recyclers
•Moisture content: maximum 0.75 percent by weight as
and the ASTM specification being developed on this subject.
determined by AASHTO T 255 using a controlled oven tem-
This includes physical, chemical, packaging, and certification
perature of 140°F and a 50-g sample.
requirements for GTR use in dense-graded and open-graded
•Mineral contaminants: maximum 0.25 percent by weight
friction courses and for asphalt-rubber membrane interlayers.
(test method to be developed).
A requirement that GTR be produced from Florida tires was
•Metal contaminants: none (test method to be developed).
dropped on the advice of legal counsel as being restraint of
trade. The specification for the asphalt-rubber binder mate-
rials, blending requirements (temperature and time), and the Chemical Requirements
method of measurement was developed for GTR and the
asphalt-rubber blend based on the laboratory and field testing • Acetone extract: maximum 25 percent
as a part of the demonstration projects in Florida. A speci- • Rubber hydrocarbon content: 40 to 55 percent
fication was developed for an asphalt rubber membrane in- • Ash content: maximum 10 percent
terlayer, but is not included in this report. Section 337 of the • Carbon black content: 20 to 40 percent
Page et al. 21
TABLE 3 GRADATIONS OF GROUND TIRE RUBBER

SillVB SIZB TYPE TYPE TYPE
% PASSING I II III
10 -- -- 100
20 -- 100 85-100
30 -- 95-100 40- 65
40 100 85-100 20- 45
60 98-100 30- 60 --
80 90-100 15- 40 5- 20
100 70- 90 5- 25 --
200 35- 60 -- --
Packaging and Identification Requirements unit may be batch or continuous type and shall provide for
sampling the blended and reacted asphalt-rubber binder ma-
The GTR shall be supplied in moisture-resistant packaging terial during normal production .
such as disposable bags or other appropriate containers. Each
container or bag of GTR shall be labeled with the manufac-
Equipment
turer designation for the rubber and the specific type, max-
imum nominal size, weight, and manufacturer batch or lot
The meter for the asphalt rubber binder shall meet the re-
designation .
quirements for accuracy, condition , and so on , of the Bureau
of Weights and measures of the Florida Department of Ag-
riculture and such fact shall be recertified every 6 months
Certification Requirements
either by the Bureau of Weights and measures or by a reg-
istered scale technician.
The manufacturer of the ground rubber shall furnish the en-
gineer with certified test results covering each shipment of
material to each project. These reports shall indicate the re- Method of Measurement
sults of tests required by this specification. They shall include
a certification that the material conforms with the specifica- The GTR content in the asphalt-rubber binder shall be mon-
tion and be identified by project number and manufacturer's itored by the department on the basis of the weight of ground
batch or lot number. rubber used versus the gallons of asphalt-rubber binder used.
The weight/gal for the various types of asphalt-rubber binders
included in Table 4 are to be used for these calculations.
FLORIDA DEVELOPMENTAL SPECIFICATION The quantity of asphalt-rubber binder material used shall
FOR ASPHALT-RUBBER BINDER be determined by a certified meter meeting requirements as
previously specified.
Scope
This specification controls the production of asphalt-rubber OTHER METHODS FOR USED TIRE RECYCLING
binder for use in asphaltic concrete friction courses and
asphalt-rubber membrane interlayers. This specification does The use of GTR in asphalt concrete and other highway ap-
not address any safety or environmental concerns associated plications previously discussed will not solve the waste tire
with its use. problem. Other uses for recycled tires have to be developed .
A variety of products exist that can be constructed from whole
tires . The U.S. Forest Service has used tire-faced retaining
Materials walls for construction of narrow mountain roads (8) . How-
ever, this is not practical for major highway construction be-
Asphalt cement: The particular grade of asphalt cement as cause of aesthetics and safety for off-road vehicular accidents.
specified in Table 4 for the respective uses shall meet the Tires have been used to control erosion along drainage ch~n
requirements of the standard specifications. The asphalt ce- nels and to stabilize highway slopes (9) . Malaysia is currently
ment shall be fully compatible with the proposed GTR as seeking 35 million tires to use as a barrier reef (10). Other
determined by the State Materials Office. products such as crash barriers , playground equipment,,
Ground tire rubber: The type of GTR as specified in Table breakwater, and installations to control soil and beach erosion
3 shall meet the requirement of Developmental Specification can be constructed from whole used tires.
on Ground Tire Rubber. Whole tires are being used as the fuel in a power plant
Asphalt-rubber binder: The asphalt cement and ground tire in Modesto, California (11) . Tires are burned at a rate of
rubber shall be thoroughly mixed and reacted in accordance 700/hr or about 4.5 million tires/yr to produce electrical en-
with the requirements of Table 4. The rubber type shall be ergy. No preprocessing of the tire is apparently necessary in
in accordance with the approved design mix. The blending this operation.
TABLE 4 ASPHALT-RUBBER BINDER
USES Den•e-gradad Open-graded A•pbalt-R.W>ber

PC PC Membrane
Interlayer
Rubber Type Type I Type II Type III
(or I)''' (or II or I)'''
\ GTR (by wt. of 5 15 25
AC)
AC Grade AC-30 AC-30 AC-20
Minimum Temp. , o F 300 300 335
Maximum Temp. , o F 335 350 375
Minimum 10 min. 15 min. 30 min.
Reaction Time (for Type II) (for Type III)
Unit Weight 8.6 lbs 8 . 7 lbs 8.8 lbs
lb/gal. '(b)
<•l Use of finer rubber could result in the reduction of the

minimum reaction time.
tb> Conversions to standard 60° F are not necessary.
NOTE: The mi n i mum r eact i on t im e. ma y be ad j usted if appr oved by the

state Materi als Off ice depending upon the t e mperature o f the
blend, size of the ground tire rubber and vi s cosity
measurement determined f rom the asphalt-rubber b i nder
mater i al prior t o or during production. Hol d - over ti me of
the asphalt-rubber binder material in excess of six hours
will not be allowed. Any corrective action in hold over
situations in excess of six hours will require the approval
of the State Materials Office.
Research being conducted by the University of Wisconsin 2. F. L. Roberts , P. S. Kandhal, E. R. Brown, andR. L. Dunning .
is directed toward the use of shredded tires to replace sand Investigation and Evaluation of Ground Tire Rubber in Hot-Mix
A sphalt. National Center for Asphalt Technology , Auburn Uni-
and gravel fills (12) . Potential benefits include reduced weight versity, Al abama Aug. 1989, pp. 1- 172.
of fill constructed with rubber chips and soil, conservation of 3. B. E. Ruth , S. Sigurjonsson, and C. L. Wu. Evalu(IJio11 of Ex·
mineral aggregates , and elimination of some of the 20 million p erimemal A sphalt-Rubber, Den e-Graded, Friction Course Mix-
discarded tires in the state of Wisconsin . Small quantities of tures: Materials and Construction of Test Pavem ents on N .E . 23
metals in the leachate from these fills apparently are too small Avenue, Gainesville, Florida. Techni cal Report, U. F. Project
49104504269 12, Department of Civil Engine ring, University of
to affect the groundwater. Florida, May 1989, pp . 1- 161.
Shredded tires have been successfully burned as a fuel in 4. B. E . .Ruth . Evaluation of Experimental Asphalt-Rubber, Ope11-
power plants, in cement kilns, in pulp and paper production , Gmded, Frictio11 Course M/:J:ture : Materials am/ Con truction of
and by tire manufacturing facilities (10) . Generally a "flui- Test Pa vem ems 011 rate Road 76. Technical Report, U.F. Project
4910450426912 Department or ivil Engineering, University of
dized bed" burning system is required to achieve sufficiently Florida, Nov. 1989, pp. 1- 56.
high temperatures for combustion of the rubber . Often a com- 5. G . C. Page. Florida's Initial E.rperie11ce Utilizing Gro1111d Tire
bination of fuels is used that promotes efficient burning and Rubber in A splwli Concrete Mixes . .Research R eport FUDOT/
reduced emissions. Although this is technically feasible, mod- M089-366, Florida Department of Transportation, Materials Of-
ern scrubber systems are necessary to remove particulate and fice, Tallahassee, Sept. 1989, pp. 1-31.
6. R . . West and J . A. Musselman. Extraction Testing of Asphalt
undesirable emission such as sulfur dioxides and nitrous Concrete Mixtures Containing Ground Tire Rubber. Bituminous
oxides. Materials Study 89-4, Florida Department of Transportation , Ma-
Crumb rubber can be mixed with other materials and proc- terials Office, June 16, 1989, pp. 1-7.
essed to make mud guards , floor mats, carpet padding, ad- 7. B. E. Ruth . Documentation of Open-Graded, Asphalt-Rubber
Friction Course Demonstration Project on Interstate 95, St. Johns
hesives , new tires, or other rubber products (10) . However, County. Technical Report, U. F . Project 4910450429812, De-
a Minnesota company established recently to produce rubber partment of Civil Engineering, University of Florida, Gainesville,
products from crumb rubber could not achieve the quality Dec. 1990, pp . 1-28.
desired by its customers. Their inability to meet the pur- 8. G . Keller. Retaining Forest Roads. Civil Engineering , American
chaser's specifications apparently. led to bankruptcy. Society of Civil Engineers , Dec. 1990, pp. 50-53.
9. G . Berthelsen . Erosion Control. AASHTO Quarterly, April 1989,
In summary, the solution to the waste tire problem needs pp. 6- 7.
to he a comprehensive one. 10. R. A . Snpek. National and 111tema1io11al /11d11strie Using f~ec;y
c/ed Paper, Plastics, Used Oil, anti Used Tires. Report prepared
fo r the Floridn Department of Commerce Department o( Public
Administration, University of Central Florida, Orlando, June 30,
1990, pp. 219- 234.
11 . Promotional Literature, The Oxford Energy Company, New York,
REFERENCES N.Y., 1990.
12. T. B. ,Edil , P. J. B os.~c h e r , and N. N. Eldin. D evelopment of
1. K. H . Murphy and C. F. Potts. Evaluation of Asphalt-Rubber as Engineering Criteria fo r hredded or Whole Tire in Flighway
a Stress-Absorbing Interlayer and as a Binder for Seal Coat Con- Applicc1tio11s . Interi m Report 10 the WisconS'in Departme nt of
struction (SR-60 Hillsborough County) . Demonstration Project Transportation, The University of Wisconsin, Madison , June 1990,
37 , FHWA-DP-27-14 , June 1980, pp . 1-28. pp . 1-19.
TRANSPORTATION RESEARCH RECORD 1339 23
Advances in Technology of Asphalt

Paving Materials Containing Used
Tire Rubber
H. BARRY TAKAL LOU AND ALAIN SAINTON
The blending of crumb rubber with asphalt cement has been in The addition of crumb rubber to an asphalt cement binder
practice for year and a number of manufacturing processe bave was successfully accomplished in the United States in the early
been developed in Euwpe as well as in the United States. How- 1960s (1). This process of heating and reacting crumb rubber
ever, all these processes have a major drawback: the asphalt
with asphalt cement has continued to be developed and used
rubber binder must be used within a few hours after being man-
ufactured. 1ll 1987, Beugnet a road contractor in France, de- since those early applications. Success of initial patching
signed a new rubbe(ized asphalt formula . This new formula im- placements using the rubberized asphalt led to its use as a
proved rheological properties of the asphalt rubber binder, including "stress-absorbing membrane" (SAM) in the late 1960s. Con-
increasing the shelf life of the binder for up to 8 days. Therefore, tinued research and development led by 1972 to the devel-
this new asphalt rubber binder could be u ed ju t like conven- opment of a stress-absorbing membrane interlayer (SAMI).
tional asphalts or polymer-modified asphalts. The dry system of By 1975, the next milestone in the development of the rub-
rubber-modified aspbalt concrete has historically been limited to
only a patented proces . Major drawbacks of this system include berized binder was to be seen in its use in an open-graded
the addition of crumb rubber to a unique "gap-graded" aggregate friction course (1).
gradation, and nonconventional design criteria. These factors In these processes, crumb rubber was added to the asphalt
contribute to the high cost of using the material when compared binder, heated and interacted, to form a rubberized binder.
with conventional asphalt concrete. The dry system of rubber- In an effort to standardize a description of the rubber asphalt
modified asphalt concrete has also experienced a major advance. technologies, this process is now widely referred to as the wet
A rubber-modified a. phalt concrete system has been developed
system.
by H. Barry Takallou. This system relates to a proce ·s for pro-
ducing an asphalt concrete composition made up of coarse crumb A different asphalt rubber technology was also developed
rubber and .fine crumb rubber incorporated into a standard den e- in the early 1960s in Europe. It was first used in the United
graded aggregate mixture. This proccs is characterized by the States in a late 1970s demonstration project in Alaska. In this
various constitueuts of the asphaltic binder and fine crumb rub- process, the design required a unique mineral aggregate gra-
ber, mixed in6ma.tely by a physical reaction. This will result in a dation. This unique gradation was required to allow a gap in
higher viscosity binder in which the optimum reaction is achieved the aggregate gradation curve. This gap, in the range of 1/4
when the fine crumb rubber particles reach optimum swelling. A
pre-reaction or pre-treatment of crumb rubber with a catalyst in. to sieve size No. 10, was to be filled by the addition of
may be needed to achieve the optimum crumb rubber particle coarse crumb rubber. The majority of the crumb rubber is
swelling. This ystem can be designed using conventional testi11g larger than sieve size No. 10. The addition of the crumb rubber
procedures and complies with conventional design criteria. The is either at the pugmill with the hot aggregate before the
use of this system is in the public domain. addition of the asphalt cement, or at the recycle fit opening
in a continuous, or drum drier operation. The crumb rubber
Research to improve and enhance the performance of asphalt is added as a separate constituent to the mixture, and there-
concrete by the addition of natural and synthetic rubber to fore referred to as the dry system.
an asphalt concrete binder has continued worldwide for more Both wet and dry systems have been used in demonstration
than a century (1). Historically, the objective of the research projects throughout the world. The systems have continued
of the addition of rubber-natural, synthetic, and combina- to increase in use with each passing year. The engineering
tions of both-to asphalt binder, was to develop a mixture community has recognized an improved performance when
to improve the physical characteristics of asphalt concrete by using crumb rubber in asphalt concrete mixtures. Also, the
improving elasticity. Results of several demonstration proj- addition of crumb rubber from whole tire recycling is a re-
ects using rubberized asphalt indicate increased fatigue resis- cycling option to the used tire disposal problem.
tance, retardation of reflective cracking, improved skid resis- Each year in the United States, approximately 240,000,000
tance, and increased durability (2). However, current interest used tires need disposal. Landfills are reluctant to accept used
in rubberized asphalt has been heightened by its potential as tires for disposal because they resist compaction, take up a
a recycling option for used tires. disproportionate amount of space, and may also become
buoyant, penetrating the covering membrane of the landfill
(3). Through state and local action, many landfills across the
H.B . Takallou, BAS Engineering Consultants, Inc., 1920 Main Street, United States no longer accept used tires . This lack of a prac-
Suite 610, Irvine, Calif. 92714. A. Sainton, Beugnet Group, 66 tical or efficient means of used tire disposal has led to the
Chamos-Elysees, 75008 Paris, France. nationwide practice of used tire stockpiling. In the early 1980s,
recognition of the dangers associated with used tire stockpiling Advances in Technology of Rubber Asphalt Binder
came the public's attention. Used tire stockpiles were an ex- (Wet System)
cellent breeding ground for mosquitoes, ideal habitat for ver-
min and, as evidenced too often, a potential fire hazard, ca- The incorporation of crumb rubber into an asphalt cement
pable of great environmental damage. binder bas been researched in Europe since the 1960s. In
The need for used tire disposal and recycling options ca- France, the incorporation and reaction of crumb rubber par-
pable of eliminating large quantities of waste tires is needed. ticles of particular sizes into the an asphalt matrix was initiated
One solution to the used tire problem is the widespread use in 1981 by Beugnet Company. In this process, finely ground
of asphalt concrete containing crumb rubber from whole tire crumb rubber is mixed directly with asphalt cement at an
recycling. Currently, there are major barriers to widespread elevated temperature (200°C) by means of an oil extender.
use of asphalt rubber. However, the technological advances The proportions of crumb rubber varied from 10 to 30 percent
in both the wet and dry systems described in this paper can and of the oil extender from 3 to 15 percent by total weight
reduce those technological barriers. of the binder. This rubberized asphalt binder is marketed
Passage of the Intermodal Surface Transportation Effi- under the trade name Flexochape.
ciency Act of 1991 contained a section directing the use of The reaction of the rubber with the asphalt binder, creating
recycled paving materials. This act requires that states use an asphalt rubber binder, displayed several improved
crumb rubber from whole tire recycling in 5 percent of their properties:
asphalt concrete usage beginning in 1994. This usage increased
by 5 percent/yr to a maximum use of 20 percent in 1997. • High viscosity (8 poises at 200°C),
Mandated use of crumb rubber in asphalt concrete is required • Ball and ring softening point greater than 60°C,
by Congress as a means to alleviate the used tire problem by • High elasticity and high resilience at low temperatures,
establishing a market for crumb rubber produced from whole and
tire recycling. • Cohesiveness 10 times greater than for asphalt alone at
20°c.
In 1985 it was discovered that the reaction processes could

MAJOR BARRIERS TO WIDESPREAD USE OF be improved by incorporating a catalyst into the mix. The
RUBBER ASPHALT improvements on the original binder were
There are several barriers to the widespread use of both the • Greater viscosity stability (i.e., better ductility for field
wet and dry systems of asphalt rubber, including use of spe- use);
cialized equipment, unique aggregate gradations, specialized • Increase in the softening point temperature (as much as
mix designs, lack of standard design criteria, cost of crumb 15 to 20 percent depending on the amount of catalyst). This
rubber, and use of patented processes. In both the wet and suggests that the binder remains less sensitive to temperature;
dry systems these combined barriers result in the major barrier • Longer preservation of the original elastic properties of
to the use of asphalt rubber: the high cost of asphalt rubber the binder; and
when compared with the cost of using conventional asphalt • Better adhesion.
concrete.
Although high cost is a barrier to both systems, the factors
causing the increase are different in both. Production Barriers
In the wet system the increases in cost are attributable to
the following factors: Even with the improved properties achieved by adding crumb
rubber to asphalt cement binder at evaluated temperatures
• Rubberized binder must be used within hours of its pro- with a catalyst, the contractor faced the same problems in
duction; therefore the high cost of mobilizing the specialized France that were seen in the United States. A major disad-
equipment (blending unit, metering unit, storage tanks, etc.) vantage of the use of the rubberized asphalt binder was that
at the production facility must be recovered in the cost per it had to be used within hours of production. Mobile units
ton of rubberized binder used on that project. were employed to produce the asphalt rubber mixture at the
• License fee for using the patented process. project production site. All costs associated with the trans-
portation, setup, and removal of the mobile equipment had
In the dry system the increases in cost are attributable to to be recovered in the unit cost of the tonnage produced.
the following factors:
• Unique aggregate gradation,

• Introduction of crumb rubber to asphalt plants, Technical Barriers
• Higher asphalt and filler content design requirements,
and Technical questions remained on the quality and consistency
• License fee for using the patented process. of the binder produced because the product is the reaction of
several constituents. As in any chemical reaction, the kinetics
The following sections describe advances in technology that are determined by reaction temperature and reactant pro-
address the barriers to widespread use of asphalt rubber. portions; these also affect storage time.
Takai/au and Sainton 25
Viscosity testing shows that the viscosity of asphalt rubber Viscosity (c:p)
binder at a digestion temperature of 200°C reaches its peak
after 45 min; then remains constant for 1or2 hr. Afterwards, 4 to 8 days
viscosity declines steadily and the quality of the binder is
diminished, as demonstrated in Figure 1. 160'C
,__ woE?n~ ~ature
800
Softening Point Temperature follows the viscosity curve,
1ao c0
falling after a certain reaction time, indicating that the binder

is degrading. Therefore, it must be recognized that there are
certain practical considerations in the production of asphalt
rubber binder: (a) the binder must be mixed on-site and (b)
the binder must be used no later than 6 hr after mixing. . t= tirre
Cooling area
As a result, production of an asphalt rubber concrete mix- FIGURE 2 Desired viscosity characteristics.
ture is limited to the amount of asphalt rubber binder that
can be produced by the mobile equipment at the asphalt plant
site. It seemed obvious that a major technological barrier to
the widespread use of asphalt rubber binder is faced without
the product being storable. Therefore, a major study was Effect of Crumb Rubber
undertaken in France in 1987 to formulate an asphalt rubber
binder that could be stored in vats for several days after mixing The effect of three different crumb rubber contents (8 percent,
without appreciable changes in viscosity or other properties. 9 percent and 10 percent rubber) on the viscosity, softening
The asphalt rubber binder could thus be produced in central point, resilience, and penetrability of asphalt rubber binder
terminal locations and shipped to different asphalt paving was evaluated. The results are presented in Figures 3, 4, and
production plant locations, similar to conventional asphalt 5. The results indicate that the optimum properties can be
binder. achieved at 10 percent crumb rubber content.
Evaluation of Storable Asphalt Rubber Binder Effect of Oil Extender
A laboratory study was performed to evaluate the effect of The effect of four different content levels of oil extender was
mix variations on properties of asphalt rubber mixtures. The evaluated (3 percent, 4 percent, 5 percent, and 6 percent oil
purpose of this study was to develop an asphalt rubber binder extender). The results indicate that an increase in the per-
(Figure 2) that had a storage time of up to 8 days and would centage of oil extender will decrease ring and ball softening
meet the following criteria: (a) a ball and ring softening point point, increase penetrability, increase the capacity for stretch-
temperature greater than 65°C and (b) a penetrability of 70 ing and tension, and reduce resistance. A maximum of
to 100/10 mm at 25°C. 6 percent of oil extender provides optimum asphalt rubber
The effect of crumb rubber, oil extender, and catalyst con- binder properties. The results of this testing are presented in
tent on the properties of asphalt rubber binder are presented Figure 6.
in the following sections.
Vlacoslty (cp) REACTION AT 180 'C

' ~ Vl1co1lty Alpllllt 80l100
'} (pol••) IU>bar powdllr
al axllndef 5"'
Calllyat 2%
I
I
'\
I I 1200 -
\
I 1100 -
I 1000
\
900
10 - '
--' 800 -
I ' 700
I "" '
I ' ' .. 600
I
I
-. 500 ·-
400
300 -·
SWB.LING - . . - DEVU.CAN ZATION
Time
3 7 8 I 10
lh 3h ,... T1llle (hOUl'I)
FIGURE 1 Viscosity evolution through rubber FIGURE 3 Change in viscosity for different
digestion. rubber percentages.
TBA ("C) TBA 'C REACTION AT 180 'C

Aaphalt 80/100
!-''"'~'
RldJer powder 8"'
72 a1 extender
70
Celllyal 2%
68 REACTION ~T 180. 'C

Asphalt 80/100
66 RIAAieJ 116 el
64 Oii eX1ender _ 6%_ - t- - - - -1-..,r -
Celelysl 2%
82 - 90 ·-
60 70
58
65
58 80
54 60
52 -
50 - - - - ·-----t--- -~- 10
50
% ol rub er powder
8% 3% 5% 9"
% ol oll extender
FIGURE 4 Change In ball ring softening point as
a function of the percentage of rubber. FIGURE 6 Changes in penetrability and ball
and ring softening point properties as a function
of the percentage of oil extender.
Effect of Catalyst
Two levels of catalyst, 2 percent versus 3 percent, were eval- The asphalt rubber binder was produced using the opti-
uated . The results indicate a maximum of 2Yz percent catalyst mized storable asphalt rubber binder formulation. The binder
provides the optimum asphalt rubber binder properties. This was produced at 180°C (the production process takes about
catalyst can be selected from ethylenically unsaturated poly- 2 hr), then stored in tanks at 160°C in a hermetically sealed
mers or copolymers that historically have been recommended vessel without agitation. Samplings were taken at regular in-
for incorporation into sulphur-vulcanized bitumen intended tervals in order to monitor the change in the product. Prop-
for road surfacings or similar products. The results of this erties of the asphalt rubber binder after production were as
testing are presented in Figure 7. follows:
A formulation was developed to provide a storable asphalt Ball and ring softening point 65°C
rubber binder composition based on the test results previou ly Penetrability in 1/10 mm at 25°C 70
pre ented. Viscosity (Rheomat) 640 cp at 180°C
1070 cp at 160°C
Storable Asphalt Rubber Binder Composition (percent) Stretching under tension to rupture
Asphalt 80/100 81.5 point at -10°C 330 percent
Rubber powder 10.0
Oil cxlt:mkr 6.U
Catalyst 2.5
REACTION AT 180 'C
REACTION AT 180 'C AlplWI ecwtoo
Alphalt 80/100 111.tJber powder 11%
RublJer powder al extender 5%
a1 extender 5% Clllllyat
Celllyat 2%
(1110 mm)
BAU. 00 RING
SOFTENING POINT
70
65 80
20 so.·
80
55 70
10 -70
50
10%
% ol rubber powder
"or m111tv1t
FIGURE 7 Change in penetration and TBA
FIGURE 5 Change In other properties as a properties as a function of the amount of
function of the percentage of rubber. catalyst.
Takai/au and Sainton 27
Results of the evaluation of the properties of the storable The patented specification requires a unique gap-graded ag-
asphalt rubber binder after 6 days of storage provided the gregate gradation ( 4).
following results:
Ball and ring softening point 72.5°C
Penetration at 25°C in 1/10 mm 85 Production Barriers
IP Pfeiffer +1.34
Viscosity (Rheomat) 1040 cp at 160°C In the PlusRide rubber-modified asphalt concrete system, the
1070 cp at 180°C major components that increase cost are as follows (5):
Rupture point stretching under tension
at -l0°C 350 percent
• Specialized aggregate crushing to obtain the unique gap-
Conclusions of the laboratory study show that the formulation graded aggregate gradation,
developed was consistent with the predetermined specifica- •High mineral filler content (8 to 12 percent),
tion. After a week of storage at 160°C, there was no noticeable •Increase in optimum asphalt cement content (7.5 to 9.5
degradation of the asphalt rubber binder. percent),
• Increase in mixing temperature (300 to 350°F discharge
Advantages of Storable Asphalt Rubber Binder for temperature), and
Industrial Applications •License fee for using the patented process.
Until development of the storable asphalt rubber binder, it

was necessary to prepare the asphalt rubber binder just before Technical Barriers
use in plant production. Hence, it was mandatory to have a
mobile asphalt rubber binder blending unit at the site of the The technical barriers encountered with the use of the PlusRide
asphalt paving plant. Without storable asphalt rubber binder system are a combination of a lack of design criteria and
the following production disadvantages are encountered: nonconventional testing methods (6).
1. Inability to supply high-capacity hot-mix plants with

enough asphalt rubber binder to meet their design production Evaluation of Rubber-Modified Asphalt Concrete
rates. (TAK System)
2. If there is a failure, either at the hot-mix plant or in the
paving operation, the storage of the asphalt rubber binder at A new asphalt rubber system was developed in 1986 by H.
high temperature is impossible (because of the loss of quality). Barry Takallou, referred to as the TAK System (7-9). This
3. For small projects, it is cost-prohibitive to use the mobile system relates to a process for producing an asphalt concrete
unit considering the cost of mobilization. In most cases, it composition made up of a coarse crumb rubber, a fine crumb
was impossible to compete with other techniques using, for rubber, asphaltic binder, and mineral aggregate. A catalyst
example, polymer-modified bitumen that was prepared in fixed to improve the reaction between fine crumb rubber and as-
units and could be stored for several days. phaltic binder, and anti-stripping agents may be added as
determined by asphaltic binder and crumb rubber compati-
However, with storable asphalt rubber binder, it is possible bility tests.
to prepared in advance the asphalt rubber binder to meet the This process is characterized by the various constituents of
high production rates of asphalt plants. Moreover, climate the asphaltic binder and fine crumb rubber, mixed intimately
conditions, plant operations, or paving-equipment failures no by a physical reaction. This will result in a higher viscosity
longer have an influence on the production or quality of the binder in which the optimum reaction is achieved when the
asphalt rubber binder. fine crumb rubber particles reach optimum swelling. A pre-
reaction or pre-treatment of crumb rubber with a catalyst may
be needed to achieve the optimum crumb rubber particle
Advances in Technology of Rubber-Modified Asphalt swelling. The coarse crumb rubber will be added as part of
Concrete (Dry System) the composition to act as an elastic aggregate, to improve
elastic properties of asphalt concrete pavement, and to reduce
The addition of crumb rubber to replace some of the mineral temperature susceptibility.
aggregate in asphalt concrete mixtures was developed in Eu- The aim of this system is to remedy the drawbacks found
rope at approximately the same time that the asphalt rubber in other systems. These include the use of specialized mixing
binder process was gaining notoriety in the United States. equipment, specialized mineral aggregate gradations, and
The system of adding crumb rubber directly to the mixture, specialized design criteria. The TAK System combines coarse
either at the recycle fit opening at a drum plant or with the and fine crumb rubber to produce a binder that has asphalt
dry aggregate at the pugmill, is referred to as the dry system rubber binder quality, and provides superior elastic properties
of rubber-modified asphalt concrete. for asphalt concrete pavement. Therefore, there is no need
This system, developed in Sweden, was patented in North for specialized mixing equipment. In this system the mineral
America and marketed under the trade name PlusRide. This aggregate is constant while the crumb rubber gradation is
process used relatively large pieces of crumb rubber (1/4 in. variable in the composition. The crumb rubber gradation is
minus) produced from used tires. The crumb rubber is then optimized per each mix design. However, the size of a ma-
added at the rate of 3 percent of total weight of the mixture. jority of the crumb rubber should be less than U.S. sieve size
No. 10. Also, the TAK System can be designed with con- ber asphalt pavement was introduced to produce a recycled
ventional testing equipment and conventional design criteria. rubber asphalt concrete mixture. The rubber-modified asphalt
The asphaltic composition of the asphalt binder can be any concrete material proved to be recyclable.
of a variety of conventionally available materials; this in-
cludes, but is not limited to, any polymer-modified asphalt
binder, and any modified asphaltic binder material as long as CONCLUSION
the asphaltic binder and crumb rubber are compatible.
In this system, the rubber content in the composition should The major barrier to the widespread use of rubber-modified
not exceed 2 percent for open-graded or dense-graded wear- asphalt concrete has been the increase in cost of using the
ing course, and 3 percent for binder course, by total weight material compared with conventional asphalt concrete. In the
of the mixture. wet system, specialized equipment that had to be mobilized
The crumb rubber used should be processed from whole with each project led to a high cost for asphalt rubber binder.
passenger and semi and truck tires. Heavy equipment tires The mobile equipment was required because the binder had
should not be used. The crumb rubber larger than 16 mesh to be used within hours of production. With the development
size should be processed by ambient granulation. The crumb of storable asphalt rubber binder (Flexochape), it can be pro-
rubber smaller than 16 mesh size may be produced from either duced in central terminal locations and shipped like conven-
granulation or grinding. Uncured or devulcanized rubber is tional asphalt cement to the production facility for use. This
not acceptable. Rubber tire buffings from either recapping or development will lower the cost of asphalt rubber binder,
manufacturing processes may not be used as supplement to provide uniform and consistent binder qualities, and generally
the crumb rubber. The crumb rubber provided should not be have a positive impact on a more widespread use of the ma-
elongated in shape and should be free of contaminants in- terial.
cluding fiber, metal, and mineral matter. Technological advances in the dry system have also reduced
The use of this system is in the public domain; therefore, the barriers to its use on a more widespread basis. The major
there are no license fees required to use it. The system uses barriers to widespread use were the unique aggregate gra-
a standard dense-graded aggregate so that no unique or gap- dation, high asphalt content, high filler content, high mixing
graded aggregate gradation requirement is necessary. temperature, nonuniform rubber gradation, lack of adequate
Demonstration projects using the TAK System have been design criteria, and use of a patented system. These barriers
constructed by several state departments of transportation. have been addressed with the introduction of the TAK System
The New York Department of Transportation has adopted of rubber-modified asphalt concrete. This system allows the
and issued design specifications for the incorporation of crumb roadway engineer to specify a standard dense-graded aggre-
rubber from whole tire recycling using the TAK System gate specification using standard mix designs, test procedures,
concept. and criteria. TAK System uses conventional mixing proce-
The Province of Ontario, Canada, has also investigated the dures, compaction equipment, and quality-control testing
use of crumb rubber incorporated into asphalt concrete mix- procedures. These developments have lowered the production
tures. In 1990, The Ministry of Transport constructed two and construction costs of using the material. The use of the
demonstration projects using the rubber-modified asphalt TAK System is in the public domain.
concrete TAK System. These demonstration projects were Air emission testing of the TAK System was performed in
constructed in an effort to evaluate its constructibility and the Province of Ontario. Emissions tests were taken during
performance, and to monitor air emissions during production production at both a drum drier plant and a batch plant for
and paving operations. TAK System and conventional asphalt concrete. Air emission
The two projects were constructed in separate regions of test samples were also taken during the paving operation,
the province without any significant problems. In one project, gathered from personal air-monitoring devices worn by the
the mixture was produced at a drum drier-type plant and in crew. Analysis of the air samples taken, both in production
the other project a batch plant was used. The rubber-modified and at the job site, indicates no difference in air emissions
hot mix material was laid with a standard paving machine and with rubber-modified asphalt concrete compared with con-
compacted and rolled with standard equipment. Preliminary ventional asphalt concrete. The recyclability of rubber-modified
results show the rubber-modified sections to be performing asphalt concrete, TAK System, was also confirmed in these
as well as the conventional control sections and again, labo- demonstration projects.
ratory testing indicates superior performance is to be expected
with rubber-modified asphalt concrete.
The air emission testing was performed for both rubber-
modified asphalt concrete, TAK System, and conventional REFERENCES
asphalt concrete. Results of the air emissions testing, from
samples taken at the stack in the production of the mixtures 1. J. A. Epps and B. M. Gallaway. Workshop Summary. In Pr~c.,
First Asphalt-Rubber User-Producer Workshop, Scottsdale, Anz.,
and from personal air monitoring devices worn by the paving May 1980.
crew, indicate that the emissions from rubber-modified as- 2. R. N. Doty. Flexible Pavement Rehabilitation Using Asphalt Rub-
phalt concrete Tak System are virtually identical to those of ber Combinations: A Progress Report. California Department of
conventional asphalt concrete. Transportation, Jan. 1988.
3. H. B. Takallou and M. B. Takallou. Benefits of Recycling Waste
These demonstration projects also addressed the recycla- Tires in Rubber Asphalt Paving. In Transportation Research Rec-
bility of rubber-modified asphalt concrete, TAK System, and ord 1310, TRB, National Research Council, Washington, D.C.,
stack emission testing was performed when the recycled rub- Jan. 1991.
Takallou and Sainton 29
4. Technical Data on PlusRide Asphalt. PaveTech Corporation, Se- ments. In Transportation Research Record 1171. TRB, National
attle, Wash., 1987. Research Council , Washington D.C .. Jan . 1988 pp. 113- 120.
S. Guide Specifications for PlusRide Asphalt. PaveTech Corporation, 8. H. B. Takallou. E11aluario11 of Mix J11gredie111s 011 the Performance
Seattle, Wash., 1991. of Rubber-Modified Asphalt Mixtures . Ph.D . Dis ertation , Oregon
6. J. L. McQuillen. Construction Practices Using PlusRideTM and State University, CorvaUis, 1987.
Arm-R-ShieldTM-Modified Asphalt Pavement. Master's Project, 9. H.B. Takallou and M. B. Takallou. Recycling Tires in Rubber
Oregon State University, Corvallis, May 1986. Asphalt Paving Yields Cost, Disposal Benefits. Elastomerics Mag-
7. H.B . Takallou and R. G. Hicks. Development of Improved Mix azine, July 1991.
and Construction Guidelines for Rubber-Modified Asphalt Pave-
Use, Availability, and Cost-Effectiveness

of Asphalt Rubber in Texas
CINDY K. EsTAKHRI, JoE W. BUTTON, AND
EMMANUEL G. FERNANDO
A study was conducted for the Texas Department of Transpor- partment personnel in each district were conducted and other
tation (DOT) by the Texas Transportation Institute (TTI) to individuals were contacted. Applications of asphalt rubber in
address the following issues: (a) the current extent of use of chip seals, sometimes called stress-absorbing membranes
asphalt rubber by the department, (b) the availability of crumb
(SAM), stress-absorbing membrane interlayers (SAMI), crack
rubber produced from scrap tires and the availability of asphalt
rubber in the state of Texas, and (c) the cost-effectiveness of fillers, and hot-mixed asphalt concrete were addressed. For
asphalt rubber compared with conventional paving materials on this study, asphalt rubber is defined as a blend of 17 to 26
the basis of existing information and the experience of department percent ground tire rubber by total weight of the blend. The
personnel. Published information was reviewed, phone interviews blend is typically formulated at elevated temperatures to pro-
with knowledgeable department personnel were conducted, and mote chemical and physical interaction of the two constitu-
existing laboratory information was evaluated. The Texas DOT ents. Various petroleum distillates are sometimes added to
currently uses asphalt rubber in four different applications. Listed
in order of their volume of asphalt rubber consumption, these the blend to reduce viscosity and enhance workability.
are (a) chip seal or stress-absorbing membrane (SAM) construc-
tion, (b) stress-absorbing membrane interlayer (SAM!) construc-
tion (c) crack or joint sealing, and (d) hot-mixed asphalt concrete
pavement construction (on a very limited experimental basis). AVAILABILITY AND USE OF ASPHALT RUBBER
These applications of asphalt rubber are described in detail in
the body of this paper. Results of this study indicated that the Governmental agencies including state highway departments
major obstacle for widespread use of asphalt rubber in Texas is
and municipal street divisions are under public pressure to
its high cost.
use waste materials to the greatest extent possible. Without
question, this is the direction in which our society must move.
Texas Senate Bill 1516 became effective in September 1989 Using waste materials and by-products is logical, sensible, and
and gave the following mandate (among others) to the Texas many times cost-effective. Incentives· are sometimes offered
DOT: by federal and state legislative bodies to promote the use of
waste products.
(1) If the State Department of Highways and Public Trans-
portation uses rubberized asphalt paving, the Department shall
use scrap tires converted to rubberized asphalt paving by a
facility in this state if that paving material is available. Waste Tire Availability
(2) In comparing bids submitted for road construction that
require paving, the Department may give a preference to bids, According to industry figures, there are as many as 2 billion
the paving materials portion of which includes the use of rub- scrap tires currently on the ground in the United States, with
berized asphalt paving made from scrap tires by a facility in
approximately 240 million tires being discarded in the United
this state if the cost of those materials does not exceed by more
than 15 percent the bid cost of alternative paving materials for States each year (2) . Of these, 200,000,000 are passenger car
the same job. tires and 40,000,000 are truck tires (3).
It is estimated that Texas is accumulating scrap tires at a
In order to make rational decisions about materials selec- rate of 18 million annually and that there are approximately
tion based on comparative cost-effectiveness, the department 150 million located at various storage sites around the state.
initiated the study described herein (1). The objective of this These figures are based on the number of passenger cars and
study is to provide the following information to the depart- commercial vehicles registered in the state and an average
ment: (a) the cost-effectiveness of asphalt rubber compared tire life of 4 years.
with more conventional paving materials based on existing A typical worn-out passenger car tire weighs approximately
information and on the experience of department personnel, 20 lb and will provide about 60 percent rubber, 20 percent
(b) the availability of asphalt rubber in Texas, and (c) the steel, and 20 percent fiber and other reusable products. On
current extent of usage of asphalt rubber in Texas. the basis of these estimates, Texas drivers are generating each
To meet these objectives , an extensive review of pertinent year the following potentially reusable materials: 108,000 tons
literature was performed , phone interviews of cognizant de- of rubber, 36,000 tons of steel, and 36,000 tons of fiber. These
estimates are conservative because they were computed using
Texas Transportation Institute, Texas A&M University System, Col- an average weight for passenger car tires, and truck tires are
lege Station, Texas 77843. much heavier.
Estakhri et al. 31
Asphalt Usage the crumb rubber comes from suppliers in California, Indiana,
and Ohio.
Approximately 32 million tons of asphalt were produced in According to asphalt-rubber suppliers and tire-rubber sup-
the United States in 1987. Of this, about 27 million tons were pliers to the asphalt-rubber industry, a continuous supply of
used for paving, 4 million tons for roofing, and fewer than 1 1 to 3 million tires annually and about $1 million in capital
million tons for other purposes. At $100/ton (a reasonable will be required to open and maintain operations of a prof-
average cost), this translates into $2. 7 billion worth of asphalt itable facility for grinding tire rubber for use in asphalt. There
cement per year for paving purposes. Approximately 90 per- is one producer of ground tire rubber at this time in Texas
cent of this was used in hot-mixed asphalt concrete (HMAC) and several reports of others going into this business. It is
and the other 10 percent was used for chip seals and surface anticipated that there will soon be an adequate supply of
treatments. The approximate quantity of HMAC produced crumb rubber produced in Texas to handle the current
in the U .S. was 500 million tons. At an average cost of $30/ton, asphalt-rubber market .
it is estimated that more than $15 billion dollars were spent
on HMAC during 1987. Although these values have varied
somewhat, they are reasonably typical of the last 18 years.
Use of Asphalt Rubber in Texas
In Texas, about 20 million tons (or $0.6 billion worth) of
HMAC was produced in 1989 according to the Texas Hot-
All of the 24 highway districts in Texas have experimented
Mix Association. Just under half of this was purchased by the
with asphalt rubber as a paving material. As stated previously,
Texas Department of Transportation (DOT). The remaining
the Texas DOT currently uses 12,000 to 14,000 tons of asphalt
went to municipalities, airport authorities, and private buyers.
rubber in paving operations annually. Another 1,200 tons are
In Hscal year 1988, the-Texas DOT used 1,100,000 tons of
used as asphalt-rubber crack sealants. The amount of asphalt
asphalt cement, 200,000 tons of emulsified asphalt, and 110,000
rubber used as a paving material is compared with other mod-
tons of cutback asphalt-a total of 1.4 million tons of asphalt
ified binders in Figure 1.
products. These figures were obtained from the Materials and
Tests Division (D-9) of the Texas DOT.
ASPHALT-RUBBER CHIP SEALS (SAMs)

Potential Tire Use in Asphalt Rubber
Research on SAMs in Texas
If 10 percent of the paving asphalt cement used annually by
the State DOT were routinely replaced with asphalt rubber Texas Transportation Institute
(using Texas tires), this would result in partial recycling of
more than one-fifth of all the scrap tires accumulated annually A research study (3) was conducted by Texas Transportation
in the state. Recall that only 60 percent of a tire is used in Institute in 1982 for the Texas DOT on asphalt-rubber mem-
producing asphalt rubber. Therefore, the remaining 40 per- branes. An evaluation of performance was made for 45 sep-
cent must be either disposed of or used in some other recycling arate projects in 13 highway districts. Approximately 850 lane
process. miles of highways were represented by materials constructed
On the basis of information from asphalt rubber suppliers as asphalt-rubber chip seals or SAMs. All projects reviewed
in Texas, it is estimated that the Texas DOT is currently using were constructed between 1976 and 1981. Data on 148 con-
12,000 to 14,000 tons of asphalt rubber/year in paving oper- ventional chip seal projects throughout Texas were reviewed
ations. Another 1,200 tons are used as asphalt-rubber crack and a comparison of performance was made. Some of the
sealant. Assuming that 20 percent tire rubber was used in the more significant conclusions are listed below.
modified binder and that 12 lb of rubber/tire (60 percent of
20 lb) were used, this quantity of asphalt rubber would ac- l. Flushing distress occurs more often with asphalt-rubber
count for approximately 430,000 scrap tires. However, it should chip seals than with conventional seals at a ratio of 99 percent
be pointed out that at the time of this study, more than 85 of all asphalt-rubber projects and 74 percent of conventional
percent of these tires were coming from out of state. Most of projects.
2. Shrinkage cracking appears in both asphalt-rubber and
conventional seals at approximately the same level, occurring
in about 50 percent of all projects.
3. With all other environmental factors being equal, alli-
Asphalt Cement gator cracking appears in conventional seals at approximately
88.9% Asphalt Rubber
1.1% twice the frequency it does in asphalt-rubber chip seals.
4. Shelling of the cover stone appears in approximately 44
~llf!l!!!I!~~/ percent of the conventional seals compared with 17 percent
AC w/Polymer
0.6% of the asphalt-rubber seals.
AC w/Latex 5. The improved resistance to alligator cracking and rav-
9.4%
eling by asphalt-rubber chip seals and poorer flushing perfor-
mance is not surprising because the typical normal application
FIGURE 1 Current unmodified asphalt binder use rate for asphalt rubber is significantly higher than that for
compared with modified binder use. conventional chip seals.
6. The present performance of asphalt rubber suggests that

improved design methods for these new systems may alleviate
the problems described here.
M~ch of the early research shows that asphalt-rubber chip

seals typically exhibit more distress than the conventional
asphalt chip seals; however, this distress is attributed to con-
struction practices rather than to the asphalt-rubber material
itself. The primary type of distress in asphalt-rubber chip seals
is flushing, which is caused by inappropriate quantities of
binder and aggregate. It should be noted, however, that flush-
ing on an asphalt-rubber chip seal is not as critical as it is on
a conventional asphalt chip seal. Experienced department FIGURE 2 Districts with no
interest in using asphalt-rubber
personnel report that although an asphalt-rubber chip se~l
ch.ip seals in near future.
can be flushed on the surface, it will still have adequate skid
resistance to remain serviceable for a number of years, which basis and the use of asphalt rubber in most districts is not
is not true for conventional asphalt chip seals. This may be standard practice. These districts are shown shaded in Figure
because the rubber particle provides increased skid resistance 2. The primary reason cited by most districts for not using
or the asphalt-rubber binder is much stiffer than an asphalt asphalt rubber is that it is too expensive. Some of these dis-
cement. tricts, which have used asphalt rubber in the past but have
no future plans, report that there were some performance
benefits associated with the material, but the benefits do not
District 24: El Paso offset the additional cost. District 21 tried an asphalt-rubber
chip seal 5 yr ago but believes a conventional AC chip seal
District 24 has applied a total of 606 lane miles of asphalt- is just as good.
rubber SAMis and 1,751 lane miles of asphalt-rubber chip During the earlier years of asphalt-rubber technology, many
seals or SAMs. The typical practice of District 24 is to use of the performance problems that emerged were caused
asphalt rubber on their three main highways: I-10, U.S. 90, by poor design and construction techniques. Now, asphalt-
and U.S. 62/180 from El Paso east to New Mexico. Because rubber technology is more advanced and improved. The five
of the costs associated with asphalt rubber, it is considered districts that are beginning to give asphalt-rubber chip seals
cost-effective only when used on the higher traffic-volume another try and those that use asphalt-rubber chip seals on a
roadways, but "Yes, it is cost-effective," states the district somewhat regular basis are shown shaded in Figure 3. District
operations engineer. It is reported as lasting twice as long as 17 uses asphalt rubber regularly. The managing resident en-
a conventional seal. In El Paso, a conventional chip seal is gineer in Brenham states: "When the pavement is badly cracked
reported to last for 7 yr and an asphalt-rubber chip seal is but appears structurally sound, asphalt rubber is the answer."
reported to last for 14 yr. He further stated that he uses asphalt rubber as often as his
Conclusions about the use of asphalt rubber in SAMs after budget will allow.
a number of years of experience in District 24 are as follows:
1. An excellent material for use in a dry, hot area. Some Cost-Effectiveness

have reservations about use in other climates.
2. Should use oqly precoated aggregate. Best results will To determine the cost-effectiveness of an asphalt-rubber chip
be obtained using %-in. maximum size. seal, the life of that seal must be known. There are many
3. Restrict "asphalt (construction) season" to hottest months variables that affect the life of any pavement surface: envi-
of the year (e.g., June, July and August). ronment, traffic, quality of construction and materials, con-
4. Permits seal application on high traffic volume roads. dition of pavement before surfacing, design, and substrate.
5. Most things applicable to conventional seal coats apply
to this m:aterial-this is a very "forgiving" material.
6. General appearance of asphalt-rubber seal is best after
about 3 yr.
7. "This is a significant advancement in asphalt technol-
ogy."
Texas Highway District Survey
As a part of this study, a telephone survey of all the districts

in Texas was conducted. Texas is divided into 24 highway
districts, and personnel in each district were queried about
their experiences with asphalt rubber.
Although there are a significant number of asphalt-rubber FIGURE 3 Districts currently
projects in Texas, many of these were built on an experimental using asphalt-rubber chip seals.
Estakhri et al. 33
Even with construction techniques that are backed by many When comparing a conventional AC chip seal with an
years of experience, such as conventional chip seals, it is asphalt-rubber chip seal, on the basis of this analysis, an
difficult to estimate the serviceable life for a given roadway asphalt-rubber chip seal would have to last three times longer
class and condition. For asphalt-rubber chip seals, this task than a conventional seal to have the same annual cost. Al-
is even more difficult. From Arizona (4) comes the report though this may be possible, there is little information to
that the life of an asphalt-rubber chip seal is 5 yr on the document these service life extensions in the field. As stated
Interstate, 8 yr on U.S. routes, and 10 yr on state routes. earlier, El Paso reports that the asphalt-rubber chip seal lasts
District personnel in El Paso report that, on U.S. highways, twice as long as the conventional seal. Arizona reports a max-
the life of an asphalt-rubber chip seal is 14 yr, and a conven- imum life of 10 years on a state route. It is commonly reported
tional chip seal lasts 7 yr. It must be kept in mind that the that a conventional chip seal will last about 7 yr in Texas.
climate in both El Paso and Arizona is very arid. In an area The asphalt-rubber seal would have to last 21 yr to have an
of low rainfall, a badly cracked pavement may remain struc- equivalent cost. This seems unlikely. Asphalt rubber is usually
turally sound longer than it would in a wet region. If a pave- only placed on high-volume roads where a conventional chip
ment is structurally sound before placement of an asphalt- seal might have a much shorter life of 3 to 4 yr.
rubber chip seal, or any type of chip seal, that seal is likely Originally, it was intended to compare asphalt-rubber chip
to have a relatively long life. seals with polymer-modified chip seals. Most of the districts
Because of the many factors influencing the life of any in Texas, at the present time, use a polymer-modified AC or
pavement surface, it is difficult to assess the cost-effectiveness polymer-modified emulsion for standard chip-seal construc-
of asphalt rubber. Although reports of experience with asphalt tion. The addition of a polymer into the binder does not
rubber in some locations are quite good (4), research results significantly increase the bid price of the chip seal for rela-
from across the United States (3,5) do not indicate that there tively large jobs. In fact, many bids show an equivalent cost/
are significant improvements in performance with asphalt- yard 2 of chip seal. Although there is no doubt that the addition
rubber seals over that of conventional seals. However, it must of a polymer into asphalt increases the cost of the binder, this
also be kept in mind that much of the research involving is not evident in the overall cost of the chip seal examined in
asphalt rubber was done at a time when the technology was this study, as shown in Figure 4. There are several factors
still in an experimental stage. Many reports of negative perfor- that enter into the cost of the chip seal: size and location of
mance were related to improper construction and design prac- job, aggregate, traffic control, and mobilization. For the jobs
tices rather than to the material itself. With the present state examined herein, the polymer-modified chip seals were not
of the art on asphalt rubber, it is not possible to accurately really any more expensive than the conventional AC chip seal.
estimate the life of asphalt-rubber seals under specific Although those districts that use polymer-modified binders
climates, traffic conditions, and underlying pavement con- report that there are benefits associated with the material,
ditions. For the purposes of this study, an annualized cost none are able to identify whether or not there is an increase
evaluation was performed for a range of service lives of an in the service life. Therefore, the polymer-modified chip seals
asphalt-rubber chip seal, a conventional chip seal, and a thin were not included in the cost analysis because they appear to
overlay. To determine the costs of conventional chip seals be similar in cost to a conventional chip seal (on a yard2 basis),
and asphalt-rubber chip seals, actual construction bids from as shown in Figure 4. Furthermore, no information is available
1989 were reviewed. All compared bids were for jobs of more about the life of polymer-modified chip seals.
than 2,000,000 yd 2 . The following are unit costs for the dif- It should be pointed out that an asphalt-rubber chip seal
ferent pavement surfaces used to calculate annualized costs contains more binder than a conventional chip seal. The con-
for different pavement lives: ventional chip seal used in this analysis contains 0.35 gal of
AC/yard 2 , whereas the asphalt-rubber chip seal contains 0.55
Conventional AC chip seal, $0.47/yd 2 .
Asphalt-rubber chip seal, $1.14/yd 2 • gal/yard 2 • Because of this difference, comparisons with con-
Thin overlay, 1-in. $1.60/yd2. ventional chip seals are not completely valid. Engineers in
the department who have experience with asphalt rubber often
The cost of the overlay is based on an in-place cost of $30/ report that they do not use this material in a location at which
ton of HMAC. The formula for equivalent uniform annual a conventional chip seal is a viable option. An asphalt-rubber
cost used in this analysis is
chip seal is typically used as a rehabilitative measure rather
than a preventive measure when a pavement is badly cracked.
A = _P-=-[i_•_,(_1_+__,_
i)'__,_
'] Therefore, a bigger burden is often placed on an asphalt-
[(1 + i)N - 1) rubber chip seal than on a conventional chip seal. Jacobson
and Schnormeier (6) of the Asphalt Rubber Producer's Group
where report that asphalt rubber applications have been most suc-
A = equivalent uniform annual cost cessful when the pavement lost 80 to 90 percent of its quality
P = initial construction cost and funds were not available to reconstruct.
interest rate Perhaps a more valid performance comparison for an
n = pavement life in years asphalt-rubber chip seal would be with a thin overlay. If an
asphalt-rubber chip seal lasted 9 yr, a thin overlay (1-in. thick)
It must be kept in mind that the annualized cost is based on would need to last 14 yr to have an equivalent annual cost.
initial construction cost only with an effective interest rate of Jacobson and Schnormeier (6) stated, "Cost comparisons
4 percent (interest rate with inflation accounted for). It does (of SAMs) are usually based on the direct cost of asphalt
not include any user costs or expected maintenance costs. rubber versus conventional asphalt. This is O.K. if one is
1.4
1.2 1.14
-"'
0
.,_
(.)
]~ 0.8
a.I <T.
- .....
C(/)
Q)ffl
0.6
0.47 0.48 0 .48
0.56
c-
"'
Q;
0.4
~
0.2
0
Aaphall Cement AC wflatex Emulsion Emulalon wfP Aaphalt Rubber
Asphalt Binders
FIGURE 4 Typical in-place costs for chip seals constructed with
different binders in 1989.
concerned only. with initial cost. It becomes very important traffic (ADT) on some of these pavements. El Paso reports
that all costs be included today and tomorrow. Initial asphalt- that major cracks in the old pavement were sealed with asphalt
rubber costs are twice as much as a conventional asphalt. This rubber before application of the SAMis. Cracks reflected
is a disadvantage because the money made available must be through SAMis by the second winter, but these were only
used to cover as much as the public can and will accept." "hairline" cracks and they tended to heal the following sum-
Jacobson and Schnormeier conclude, however, that asphalt mer. All pavements are still in good to excellent condition.
rubber is cost-effective because less maintenance is required A representative of District 24 stated: "This material provides
of asphalt-rubber chip seals than of conventional asphalt chip the best life-cycle cost we have found for rehabilitation of
seals. cracked, weathered asphalt surfaces needing minor leveling
provided by thin HMAC overlays. "
ASPHALT-RUBBER INTERLAYERS (SAMls)

Survey of Texas Highway Districts
SAMI Research in Texas
Personnel in each highway district were contacted to deter-
Texas Transportation Institute mine their experiences with asphalt rubber applied as inter-
layers. The six districts that have constructed asphalt-rubber
The Texas DOT is sponsoring an ongoing research study with interlayers are identified in Figure 5. The opinions of de-
the Texas Transportation Institute (7) to evaluate the perfor- partment personnel on asphalt rubber used as an interlayer
mance of asphalt-rubber interlayers. Three full-scale test roads are much more favorable than they are for asphalt rubber
were constructed in 1983 and 1984 near El Paso, Brownsville, used as chip seals. Although the cost of an asphalt-rubber
and Buffalo, Texas. The Buffalo test road has an overlay interlayer is still at least twice that of a conventional chip seal
thickness of between 4 and 6 in. and is not showing any interlayer, it is only a small portion of the total overlay system
distress. The Brownsville test road was constructed with ex- cost. Most of the districts that have used asphalt rubber as
cessive interlayer binder application rates and all sections are
flushing. However, the El Paso test road has yielded some
useful information. Nine different types of asphalt-rubber in-
terlayers were constructed there using different binder ap-
plication rates, different rubber concentrations , and different
ground tire rubber suppliers. The control section contained
no interlayer. All of the asphalt-rubber sections are perform-
ing better than the control in terms of delaying reflective
cracking, with some sections performing significantly better.
District 24: El Paso
District 24 currently has six asphalt-rubber interlayers under

observation. These range in age from 1 to 12 yr. Overlay
thickness is from 1-1/2 to 3 in. The average binder application FIGURE S Districts currently
rate was 0.55 gal/yard 2 • Traffic exceeds 100,000 average daily using asphalt-rubber interlayers.
Estakhri et al. 35
an interlayer report that it definitely reduces the rate of re- At the current time, more than 95 percent of all asphalt-
flection cracking. rubber crack sealant that is used in Texas is supplied from
Evidence has been seen of cracks in asphalt-rubber chip Crafco, Inc., in Chandler, Arizona. Crafco has done extensive
seals healing in the summer months. Although this phenom- research in asphalt-rubber formulation, production, and ap-
enon can be observed in an asphalt-rubber chip seal, it cannot plication and has helped the state of Texas in its specification
be viewed in an asphalt-rubber interlayer because it is covered guidelines for asphalt rubber crack sealant. In 1989, Crafco
by an overlay. However, if this healing ability exists in an supplied almost 3.5 million lb of material to Texas at an av-
asphalt-rubber interlayer, then the interlayer may function as erage price of 19 cents/lb, translating to a yearly total of
a waterproofing membrane. Once cracks do develop in the $495,041. This material was used to fill approximately 14 mil-
surface layer, the asphalt rubber may prevent, or at least lion linear ft of crack and joints. The price has varied slightly
reduce, any water intrusion into the underlying pavement during the past several years, with costs ranging from 12 to
structure. 15 cents/lb. The department is currently modifying its speci-
fications to accept a slightly wider variety of products that
would allow other suppliers to enter the market.
Cost-Effectiveness
Based on the literature review, research conducted by Tri, Survey of Texas Highway Districts
and the experience of department personnel, an asphalt-
rubber interlayer can produce an improvement in pavement On the basis of a telephone survey of district personnel in
performance. Although it is generally believed that an asphalt- Texas, Crafco asphalt-rubber sealant is the product of choice.
rubber interlayer extends pavement life, it is not accurately Many of the districts have used other products in the past,
known how long. Because the interlayer is not visible on the and on jobs with very small cracks a polymer emulsion product
surface, its effects are difficult to measure. A common method has proven to be more effective; however, according to one
of evaluation is to measure reflective cracking in the surface DOT engineer, asphalt rubber continues to "last longer and
of the overlay. However, there may be other improvements provide less problems" than other types of sealants.
in pavement performance that are not commonly measured In talking to each of the districts with crack sealing pro-
by highway departments, such as roughness. If there are any grams, it was readily apparent that they were pleased with
benefits from "waterproofing" of the underlying structure, the product. Typical comments were that the rubber is very
this is difficult to measure. stable; vehicle tires do not displace it; the rubber provides
A similar cost analysis as shown in the previous chapter good elasticity and strength; and it does not seem to weather
was performed for SAMis. An annualized cost was deter- or oxidize at all.
mined for a 2-in. overlay and compared with the annualized Almost all of the districts agreed on the material's prop-
cost for an asphalt-rubber SAMI with a 2-in. overlay. As in erties and all independently estimated the typical life of the
the previous cost analysis, this is based on initial construction product to be 3 yr.
cost only and does not include any user or maintenance costs.
The following initial construction costs were used for the anal-
ysis: Cost-Effectiveness
2-in. overlay $3.20/yd •
2
2-in. overlay with SAMI $4.25/yd2 • To be consistent with the rest of this paper it would be ben-
eficial to include a cost-effectiveness comparison with other
On the basis of this analysis, a 2-in. overlay with an asphalt- similar products. However, the extensive use of the asphalt
rubber SAMI would need to last approximately 50 percent rubber throughout the districts makes this type of comparison
longer than a 2-in. overlay alone to yield an equivalent annual difficult. Projects are sometimes encountered that require other
cost. For example, if a 2-in. overlay lasted 8 yr, a 2-in. overlay special sealants; however, these projects are usually very small
with SAMI would need to last 12 yr to be equivalent in cost. and a true performance comparison cannot be established.
ASPHALT-RUBBER CRACK SEALANTS RUBBER-MODIFIED ASPHALT CONCRETE

MIXTURES
One of the most troublesome problems the highway depart-
ment faces in its effort to provide quality long-lasting pave- Field Experience in Texas
ments is the presence of pavement cracks. In the past, main-
tenance forces have used many materials as sealants in at- The 1989 hot-mix asphalt concrete usage within the state of
tempts to seal cracks and effectively extend pavement life. Texas is approximately 8.1 million tons, which is down slightly
These materials include asphalt cements, cutbacks, emulsions, from the 5-yr average of 9.4 million tons. These high values
and latex-modified emulsions. However, during the 1970s and indicate excellent opportunities for use of asphalt rubber.
early 1980s, an asphalt-rubber sealing compound containing However, at this time, only two districts in Texas have tried
ground tire rubber emerged as a new and comparatively ef- the product. Ten years ago, District 21 experimented with the
fective means of crack repair. The compound is composed of rubber-modified hot mix but the job was unsuccessful. District
approximately 80 percent asphalt and 20 percent ground tire maintenance forces applied the hot mix along a 1-mi section
rubber. on S.H.336 in McAllen. The mix raveled severely and the
district was forced to place a chip seal over the mix within 3 chip seals claim they will last twice as long as a conventional
months. chip seal.
In 1989, the Tyler district (District 10) placed a dense- There is not enough available information to accurately
graded, asphalt-rubber hot-mix overlay. The project was lo- determine the cost-effectiveness of asphalt-rubber chip seals.
cated at the intersection of FM 14 and Loop 323 just outside However, an annualized cost analysis performed in this study
Tyler. Asphalt rubber was chosen for the site in hopes of revealed that an asphalt-rubber chip seal would have to last
curing a severe rutting problem caused by large trucks turning three times longer than a conventional asphalt chip seal to
onto and off the loop. So far, district personnel are pleased have an equivalent annual cost.
with the project and are interested in using the product again Districts in Texas that are experienced with asphalt-rubber
but on a more standard hot-mix job. The cost of the asphalt chip seals do not usually construct them on a pavement where
rubber for this job was approximately $80/ton. Tyler's district a conventional chip seal is a viable option. Asphalt-rubber
personnel believe that a larger job would help reduce this chip seals are used successfully as a rehabilitative instead of
high material cost. a preventive measure and they are often placed on high-traffic
District 4 in Amarillo constructed 10 lane mi of dense- volume roads. Therefore, perhaps a more valid comparison
graded, asphalt-rubber hot mix in the fall of 1990. Bid prices for asphalt-rubber chip seals might be with a thin overlay or
showed an in-place cost of $52/ton for the asphalt-rubber multiple chip seal, in which case the asphalt rubber is much
paving material, which is substantially less than the $80/ton more likely to be cost-effective.
reported in Tyler but not particularly attractive when com-
pared with the $30 to $35/ton most districts were paying for Asphalt-Rubber lnterlayers (SAMis)
conventional hot-mixed asphalt concrete.
Only six Texas highway districts have built SAMis. Opinions
of department personnel on asphalt-rubber interlayers are
CONCLUSIONS AND RECOMMENDATIONS much more favorable than those on asphalt-rubber chip seals.
Most of the districts that have installed SAMis believe they
Availability and Use are effective in delaying reflective cracking. Some also believe
SAMis will reduce intrusion of surface water and thus pump-
Approximately 150 million scrap tires are currently stored in ing even after cracking occurs in the surface layer.
Texas and another 18 million are being discarded in the state An asphalt-rubber SAMI may provide cost-effective im-
each year. The scrap tires accumulated annually could be used provements in performance of hot-mixed asphalt concrete
to produce 108,000 tons of rubber suitable for use in asphalt- overlays. On the basis of an annualized cost analysis per-
rubber products. The Texas DOT annually uses more than formed in this study, an overlay with an asphalt-rubber in-
1,000,000 tons of asphalt cement. If 10 percent of this paving terlayer would need to last approximately SO percent longer
asphalt cement were routinely replaced with asphalt rubber, than an overlay constructed without an interlayer to be cost-
more than 20 percent of the annual production of waste tires effective.
in Texas would be used. At the present, slightly more than 1
percent of this paving asphalt is asphalt rubber.
Only about 60 weight percent of a tire is consumed in pro- Asphalt-Rubber Crack Sealants
ducing asphalt rubber. Remaining products include primarily
steel, fiber, and additional rubber. Asphalt-rubber crack sealant, which contains 20 percent ground
The Texas DOT is currently using about 13,000 tons/yr of tire rubber, is essentially the only crack sealant used by the
asphalt rubber, which accounts for approximately 430,000 scrap Texas DOT. The Texas DOT uses approximately 3.5 million
tires. However, most of the waste tires used in this material lb of crack sealant annually.
come from other states. The availability of crumb rubber in Asphalt-rubber crack sealant is considered by all personnel
Texas is a rapidly changing issue. Findings indicate that next interviewed in highway districts to be the best product avail-
year 7 ,000,000 to 10,000,000 tires may be recycled in plants able for sealing cracks in asphalt concrete and portland cement
in Texas. concrete pavements.
Asphal~-Rubber Chip Seals Asphalt-Rubber Hot Mix
Asphalt-rubber chip seals have been constructed, at least on Asphalt rubber has been used on a very limited basis in Texas
an experimental basis, in all parts of Texas. However, there for construction of HMAC. The use of crumb rubber in HMAC
are only 5 out of the 24 highway districts currently constructing is gradually gaining popularity across the United States; how-
asphalt-rubber chip seals with some regularity. ever, the technology is still in a somewhat experimental stage
Use of asphalt rubber for chip seals in most highway districts of development.
in Texas has historically not been standard practice, and 13
districts have no plans for increasing their use in the future.
The primary reason cited for this is that asphalt rubber is too General Recommendation
expensive and has not proven to be cost-effective in this ap-
plication. The Texas DOT and the Texas Legislature should not "go
An asphalt-rubber chip seal costs two to three times more overboard" in promoting the use of tire rubber in asphalt
than a conventional chip seal. Proponents of asphalt-rubber because the benefit-cost ratios are not sufficiently high for
Estakhri et al. 37
every application. Offering incentives to use tire rubber (which Research Report 1902-lF, Texas Transportation Institute, Texas
negate fair competition) or mandating the use of tire rubber A&M University, College Station, Sept. 1990.
2. ATR Develops Scrap Tire Recycling Process. Elastomerics, Com-
in asphalt pavements to solve the waste tire problem does not munications Channels, Inc., Atlanta, Ga., July 1989.
appear to be in the best interest of the tax-paying public. 3. T. S. Shuler, B. M. Gallaway, and J. A. Epps. Evaluation of
Sound engineering, not politics, should govern the choice of Asphalt-Rubber Membrane Field Performance. Research Report
paving materials used in highway construction. A practical 287-2, Texas Transportation Institute, Texas A&M University,
solution to the problem will require more research and en- College Station, May 1982.
4. L. A. Schofield. The History, Development, and Performance of
gineering to provide self-supporting, cost-effective uses for Asphalt Rubber at ADOT. Arizona Transportation Research Cen-
scrap tires. There may be more economically efficient ways ter, Phoenix, 1989.
to recycle tires in much greater volumes than in asphalt pave- 5. T. S. Shuler, R. D. Pavlovich, J. A. Epps, and C. K. Adams.
ments. Investigation of Materials and Structural Properties of Asphalt-
Rubber Paving Mixtures. Research Report RF 4811-lF, Texas
Transportation Institute, Texas A&M University, College Station,
Sept. 1985.
6. C. C. Jacobson and R. H. Schnormeier. Cost-Effectiveness of
Asphalt Rubber. Asphalt Rubber Producer's Group, In Proc.,
REFERENCES National Seminar on Asphalt Rubber, Kansas City, Mo. 1989.
7. C. K. Adams and J. Gonzales. Asphalt-Rubber Interlayer Field
1. C. K. Estakhri, E. G. Fernando, J. W. Button, and G. Teetes. Performance. Research Report 449-lF, Texas Transportation In-
Use, Availability, and Cost-Effectiveness of Asphalt Rubber in Texas . stitute, Texas A&M University, College Station, June 1987.
38 TRANSPORTATION RESEARCH R E CORD 1339
Permanent Deformation Characteristics of

Recycled Tire Rubber-Modified and
Unmodified Asphalt Concrete Mixtures
NEIL c. KRUTZ AND MARY STROUP-GARDINER
In recent years, modified asphalt mixtures have become increas- Phase 2: Permanent deformation characteristics of rubber-
ingly popular in the construction of flexible pavements. These ized and unmodified mixtures.
products have gained popularity because of their ability to in- Phase 3: Low-temperature cracking resistance of rubber-
crease resistance to rutting at warm temperatures while reducing ized and unmodified mixtures .
the occurrence of thermal cracking at cold temperatures . This,
coupled with the growing problem of waste rubber tires, has led Phase 4: Fatigue characteristics for rubberized and unmod-
to the reprocessing (grounding) of tire rubber for use in asphalt ified mixtures .
concrete mixtures. In order to investigate the warm temperature
rutting hypothesis, a laboratory research program using both static The laboratory results from Phase 2 only are discussed in this
and repeated load permanent deformation tests, carried out at report. Phase 1 has been completed and reported in "Com-
two temperatures (77°F and 104°F), was designed to assess the parison of Mix Design Methods for Rubberized Asphalt Con-
potential benefits of rubberized asphalt concrete mixtures. Con-
crete Mixtures" (J) . Phases 3 and 4 are currently being com-
clusions from this research indicated that the addition of ground
tire rubber to asphalt concrete mixtures results in mixtures that pleted .
exhibit less permanent deformation at high temperatures com- The scope of this research program includes one aggregate
pared with unmodified mixtures. The research also indicated that source, one gradation, and six binders. The test matrix is
permanent deformation testing should be carried out at high tem- shown in Table 1.
peratures under repeated loading. The relative ranking of strain
changes from 77°F to 104°F for both methods of testing and static
testing indicates the presence of rubber; however, it does not
indicate anything about the base asphalt . The repeated load test- MATERIALS
ing indicates, in a reliable way, the differences that exist between
binders. Aggregates
The aggregates used in this research program were obtained

In recent years, modified asphalt mixtures have become in- from Granite Rock Company, located in Watsonville, Cali-
creasingly popular in the construction of flexible pavements. fornia. This material is a 100 percent crushed granite that has
These products have gained popularity because of their ability no history of stripping problems with in-service pavements.
to increase resistance to rutting at warm temperatures while The physical properties are presented in Table 2.
reducing the occurrence of thermal cracking at cold temper- The gradation used to prepare the mixture samples is shown
atures. This , coupled with the growing problem of waste rub- in Table 3. This gradation was chosen to meet ASTM D3315
ber tires , had led to the reprocessing (grounding) of tire rub- Yi-in . dense mixture, Nevada Type II, and California Yi-in.
ber for use in asphalt concrete mixtures. medium specification (Table 3) . This gradation was opened
In order to investigate this hypothesis, a laboratory research up slightly on the #30 and #50 to accommodate the presence
program was designed to assess the potential benefits of rub- of rubber.
berized asphalt concrete mixtures .
Binders
RESEARCH PROGRAM
The three grades of neat asphalt used in this research program
The extended research program was designed to include four were obtained from a single California Valley crude source.
phases: The binders used were :
Phase 1: The use of conventional mix design methods for Unmodified: AC-5
determining the optimum asphalt content for rubberized mix- AC-20
tures . AC-40
Both the AC-5 and AC-20 were then modified with crumb
Civil Engineering Department, University of Nevada, Reno, Nev. rubber. The AC5 was also modified with rubber and an ex-
89557. tender oil, yielding a very soft third modified binder. The
Krutz and Stroup-Gardiner 39
TABLE 1 TEST MATRIX FOR PERMANENT DEFORMATION OF MODIFIED AND

UNMODIFIED MIXTURES
Binder ACS Ext. Oil ACS AC20 AC40
Modifier Oriq. Rubber Oriq. Rubber Oriq. Rubber Oriq. Rubber

Added Added Added Added
Static 77°F x x x x x x
Load
104'F x x x x x x
Repeat. 77'F x x x x x x
Load
104°F x x x x x x
TOTAL OF 72 SAMPLES (each X denotes 3 samples)
source of crumb rubber was selected by the sponsor with the gradation specification suggested by Crafco, are shown m
rubber being blended with the asphalt cement by Crafco Inc., Table 4. The resulting modified binders were:
located in Chandler, Arizona. The rubber used in this research Modified: AC-5 + 17% Rubber (AC5R)
program was ambient ground rubber having a hydrocarbon AC-5 + 16% Rubber + 5% Extender Oil (AC5RE)
content of approximately 45 percent and a specific gravity AC-20 + 16% Rubber (AC20R)
between 1.100 and 1.200. The particle size, along with the
OPTIMUM BINDER CONTENTS
TABLE 2 PHYSICAL PROPERTIES OF WATSONVILLE In Phase 1 of this research, binder contents to be used in
AGGREGATE phases 2, 3, and 4 were selected by a committee that included
the sponsor and all of the researchers involved. These selec-
Test Fine A&lreg&1e (-#4) Coane Agnpl.o ( +#4)
tions were based on mix designs conducted at both the Uni-
Bulk Specific 2.589 2.682 versity of Nevada, Reno, and the U.S. Army Corps of En-
OnM1y
2.667 2.735
gineers Water Ways Experiment Station (WES). Optimum
Bulk Spedflc
OnMty,SSD binder contents for both unmodified mixtures ACS and AC20
Coodltloo were agreed upon at 5.3 and 5.7 percent by total weight of
Appuad Spedflc 2.806 2.832
OnM1y
mix, respectively. However, there was disagreement about
Ablorptlon 3.0 2.0 the binder content to use for each of the modified mixtures.
Capacity(%) As a result, a compromise was made that was agreeable to
all parties involved in the extended program. The compromise
yielded binder contents that were higher than the University
of Nevada, Reno (UNR)-recommended optimums. The fol-
TABLE 3 COMPARISON BETWEEN LABORATORY lowing table shows the binder contents used and the UNR-
GRADATION USED IN RESEARCH PROGRAM AND
SEVERAL SPECIFICATIONS
recommended binder content for all modified mixtures.
Binder Content Used Jn UNR-Recommended
si-Sla t..bcnrory ASDID331.5 Ncnda Olllfomla
Ondatloa ~·[)D.o Typon W\JOdlum Preparing Samples Binder Content
Type of Binder(%, Total Weight of Mix)(%, Total Weight of Mix)
Oum.ladvo Perccm '-"II
100 100 9G-100 100
ACSR 8.5 7.7
3/4"
ACSRE 8.3 7.7
AC20R 7.9 7.4
1/'l' 98 9G-100 - 89-100
3/8" 85 - 63-35 75-100 TABLE 4 PHYSICAL PROPERTIES OF GROUND TIRE

RUBBER USED IN PREPARING MODIFIED BINDERS
#4 58 44-74 45.fil 51-74
SiOft Size BabrlOR-24 ~
R• * n•..,....,lbll
#8 40 2S-58 - 35-57
Qi.mWatm> Peniont P..u.
•16 28 -- - - •10 100 100
"16 100 100
•30 20 - - 14-35
•30 78 70.100
#SO 14 5-21 - - #40 49 -
150 27 -
•100 9 - - - #80 9 ().20
#200 s 2-10 3-9 0.11

#100 7 -
#200 0.2 ().5
The result of this compromise is a binder-rich mixture . This which samples were allowed to rebound freely . Tests con-
should be remembered when assessing any of the permanent ducted at 77°F used a static stress of SO psi, and tests conducted
deformation data contained in this report. at 104°F used a static stress of 20 psi.
The second test used to assess permanent deformation was
a triaxial, repeated-loading confined test. This test procedure
SAMPLE PREPARATION followed the interim testing guidelines from the Strategic
Highway Research Program (SHRP) A-003A contractor at
Samples were batched by first separating the aggregates into the time this testing was started. The only change imple-
the 11 individual sizes (V2 in.,% in ., % in., #4, #8, #16, #30 , mented by UNR was the shortening of the test time from
#50 , #100 , #200 , fines) needed to prepare samples, and then 36,000 cycles (approximately 8 hr) to 5,200 cycles (approxi-
recombined to meet the desired gradation . Washed sieve anal- mately 1 hr). The test used a 1-min. preconditioning period
yses were performed on complete batches to ensure that the followed immediately by a 60-min. test. The repeated loading
gradation had been met. sequence consisted of 0.1-sec duration haversine pulse fol-
After all aggregate preparation was completed, batches were lowed by a 0.6-sec rest period. This sequence yields a testing
selected at random and mixed with the selected binder. Dif- frequency of 1.43 cycles/sec. All tests used a confining pres-
ferent methods of mixing and compaction were used for the sure of 15 psi. Tests conducted at 77°F used a peak deviator
rubberized and unmodified mixtures. The procedure for each stress of 50 psi, whereas tests conducted at 104°F used a peak
method is described in the following sections. deviator stress of 20 psi.
Unmodified mixtures were blended in accordance with ASTM Deformations were continuously measured for both tests
D 1561 (2). After mixing, samples were placed in a 140°F using two linear variable differential transducers (L VDTs).
forced draft oven for 15 hr before being reheated to 230°F These LVDTs were instrumented 180° apart and measured
for compaction. Specimens 8 in . in height by 4 in. in diam deformations over the total sample height. These deforma-
were compacted in thirds using a kneading compactor. Each tions were electronically averaged and recorded every 60-sec
lift, or third, received 30 blows at 250 psi . Lifts were com- throughout testing.
pacted consecutively on top of each other. After compaction The data were then used to calculate compressive strains
of the third lift, each sample was placed in a 140°F oven for for each test over the sample height using the following equa-
l1/2 hr before the application of a 5,000-lb leveling load. Sam- tion:
ples were allowed to cool before being extruded.
Rubberized mixtures were blended using the recommen- E(l) = [d(t)/H 0 ]
dations of Chehovits (3) . This involves heating the aggregate

where
to 300°F and the rubberized binder , regardless of base asphalt
viscosity , to 350°F before mixing. Once again, after mixing , E(t) strain at time t, in .fin.
=
samples are placed in a 140°F forced draft oven for 15 hr H = original height of sample , in .
0
before reheating the samples for compaction. Samples using d(t) = deformation of sample height at time t, in .
the rubberized AC-20 were reheated to 300°F for compaction
while the other two rubberized mixtures, AC5R and ACSRE,
were reheated to 230°F . The same compaction procedure de- TESTING PROGRAM
scribed previously was used with the rubberized mixtures with
the exception that the 1Yi-hr cure time at 140°F was extended A total of 72 samples, 12 samples from each of the 6 types
to 3 hr for the AC20R. Rubberized samples were allowed to of binder, were prepared. This allowed for 3 replicates to be
cool before being extruded. tested at each testing condition. The testing conditions used
were static load at 77°F, static load at 104°F, repeated load
at 77°F, and repeated load at 104°F. This testing matrix is
TESTING METHODS shown in Table 1. The number of samples tested produced
sufficient data to estimate the mean, standard deviation , and
After compaction, samples were allowed to cool overnight in coefficient of variation for each type of mixture at each testing
a 77°F room before being tested for bulk specific gravity and condition.
height, ASTM D2726 and D3515, respectively (2) . Samples
were placed under a fan , again overnight, to remove any ANALYSIS OF TEST RESULTS
moisture that may have penetrated the sample during testing.
Samples were then placed in an appropriate temperature con- As stated previously, there were two different types of per-
trol chamber to condition them to the testing temperature to manent deformation tests used in this research program. Then
be used, either 77°F or 104°F. After 24 to 36 hr, samples were within each test, samples from each of the six mixtures were
tested for permanent deformation using one of two tests. tested at two different temperatures. For ease of discussion,
These tests are described in detail as follows. the analysis will be presented similarly; first the static test
The first of two tests used was a modified version of the results and then the repeated load test results will be given.
proposed ASTM creep test (4). This test involved a static-
loading, uniaxial unconfined creep test . This test incorporated
a 2-min preconditioning load , using the test load magnitude , Analysis of Static Permanent Deformation Testing
followed by a 5-min. rest period. Immediately following the
rest period, a static load was applied for a period of 60 min ., The average standard deviation and coefficient of variation
followed by a 15-min. unload, or rebound, period, during (CV) for the strain at 60 min (i.e., strain at the end of the
TABLE 5 SIMPLE STATISTICS FOR STRAIN AT END OF LOADING FOR STATIC

PERMANENT DEFORMATION TESTS COMPLETED AT 77°F
Binder Strain at 3600 seconds of Average standard Coefficient cr-p

Type Loading (in/in) strain Deviation of Modulua
variation (pai)
Rep. A Rep. B Rep. c C•l
AC5 F F F -- --- -- 0
AC20 0.0079 0.0072 0.0122 0.0091 0.0027 29.8 5495
AC40 NA NA NA NA NA NA NA
AC5RE 0.0157 0.0165 0.0174 0.0165 0.0009 5.1 3024
AC5R o. 0132 0.0137 NA 0.0135 0.0004 2.6 3717
AC20R 0.0069 0.0090 0.0059 o. 0073 0.0016 21.8 6881
F - indicates sample failure prior to sixty minutes of loading

NA - indicates data not available
loading period) for all tests completed at 77°F using the static The average compressive strain versus time relationship for
testing procedure are shown in Table S. The AC-40 data have the 77°F static test results is shown in Figure 1. Inspection of
been removed from the data base because of sample damage this figure shows that the mixtures behaved as expected. The
before the test. It can be seen from this table that the CV is unmodified mixtures show that the ACS samples fail at about
somewhat higher than desired; however, it is still in the range 10 min into the test and the AC20 samples yield relatively
of acceptable test results. This table also shows an average low strains. The rubberized mixtures show decreasing strain
creep modulus for each of the five remaining mixtures. A with increasing binder viscosity (i.e., ACSR strains more than
creep modulus of zero indicates that the samples failed before AC20R, and ACSRE strains more than ACSR). It can be
the 60 min of loading. concluded from this figure that for this testing procedure con-
ducted at 77°F, the addition of rubber yields mixtures that
exhibit less deformation (i.e., rubberized ACS strains less
Compressive Strain (in/in) than ACS, and rubberized AC20 strains less than AC20).
0.02 ~~~~~~~~~~~~~~~~~~~~~~-
The average standard deviation and CV for the strain at

60 min (i.e., strain at the end of the loading period) for all
0.015 tests completed at 104°F, using the static testing procedure,
are shown in Table 6. Once again, the AC40 data have been
removed from the data base because of sample damage before
0.01
testing. The CV is again higher than desired; however, it is
still in the range of acceptable test results. The average creep
0.005 modulus for each of the five remaining mixtures is also shown
in this table. A creep modulus of zero indicates that the sam-
ples failed before the 60 min of loading.
0 '11-'~~~~-'-'--1--'-·'--'-..L..l-'-'-'--'--l-'-'-'-..L..l-'-L...l...-'--!---'--lLL-L.J._--'--lJ
The average compressive strain versus time for four of the
0 1200 2400 3600
Time (seconds) six mixtures from the static testing at 104°F is shown in Figure
2. The ACS samples failed drastically during the precondi-
-+- AC5R -+-- AC5 --B- AC20R -*- AC20 -+- AC5RE tioning sequence, leaving no data to present for the testing
FIGURE 1 Compressive strain versus time for static loading sequence. This leaves only one unmodified mixture in the
conducted at 77°F. figure, the AC20. All three curves for the rubberized binders
TABLE 6 SIMPLE STA TISTICS FOR STRAIN AT END OF LOADING FOR STA TIC
Binder Strain at 3600 Seconds of Average Standard Co41t'f!ciant en.p

Type Loading (in/in) Strain Deviation of Modulus
Vllrilltion (p•i)
Rep. A Rep. B Rep. C
C'>
AC5 F F F --- --- --- 0
AC20 0.0087 0.0056 NA 0.0072 0.0022 30.7 2797
AC40 NA NA NA NA NA NA NA
AC5RE 0.0045 0.0064 NA 0.0055 0.0013 24.7 3670
AC5R 0.0045 0.0051 0.0042 0.0046 0.0005 10.0 4348
AC20R 0.0037 0.0041 0.0059 0.0046 0.0012 25.7 4380

Compressive Strain (in/in) Average Mix Stiffness at 60 min (psi)

0.02 ~.....:....~~~~~~~~~~~~~~~~~~~~
0.015
0.01
0.005
1200 2400 3600

Time (seconds)
-+- AC5R --e- AC20R --*""" AC20 --- AC5RE

FIGURE 3 Creep modulus from static testing at 7°F and
FIGURE 2 CompressiVe strain versus time for static loading 104°F.
conducted at 104°F.
fell on top of each other, indicating the same response for at 77°F using the repeated loading testing procedure are shown
any mixture incorporating rubber. All rubberized mixtures in Table 7. This table shows data for all six mixtures . It also
exhibited less strain than the AC20. It is hypothesized that shows the average creep modulus for each of the six mixtures .
in this case, the rubber is absorbing the load and the strain Like the static modulus, this modulus was calculated by di-
is therefore independent of the base asphalt cement. It should viding the strain after 60 min of testing into the peak deviator
be remembered that is for a static unconfined test . stress.
The average creep modulus calculated at 60 min of loading The average compressive strain versus time for the six mix-
for the five mixtures for both temperatures of static testing tures from the repeated load testing at 77°F is shown in Figure
is shown in Figure 3. It can be seen that the ACS shows 4. From this figure, is can be seen that both the ACS and
modulus values of zero for both temperatures. This is because ACSRE failed during testing. This was because of the rela-
of sample failure before the 60 min of loading. The AC20 tively low viscosity of the unmodified ACS and rubberized
shows a drop in the modulus of approximately SO percent ACS that incorporates an extender oil, which is also of very
from 77°F to 104°F. All three of the rubberized mixtures low viscosity . The ACSR finished the testing without failure ;
showed a smaller drop in stiffness than the AC20 . In fact, the however it exhibited large strains. The three mixtures that
ACSRE showed an increase in modulus from 77°to104°. This performed best were the AC20, AC20R, and AC40. It is
would indicate that rubberized mixtures will suffer a smaller interesting to note that the AC20R exhibited a higher strain
loss of stiffness with increasing temperature than will un- than the AC20. In this case the AC20 samples exhibited strains
modified mixtures. that grouped the mixtures with the AC40, which yielded very
low strain. This anomaly remains unexplained.
The average standard deviation and CV for the strain at
Analysis of Repeated Load Permanent Deformation 60 min (i.e., strain at the end of the test) for all tests completed
Testing at 104°F using the repeated loading testing procedure are
shown in Table 8. This table shows data for all six mixtures.
The average standard deviation and CV for the strain at 60 It also shows the average creep modulus for each of the six
min (i.e. , strain at the end of the test) for all tests completed mixtures . The table indicates that the ACS and AC20 samples
TABLE 7 SIMPLE STATISTICS FOR STRAIN AT END OF LOADING FOR REPEATED LOAD
-PERMANENT DEFORMATION
- . - -- .. TESTS COMPLETED AT 77°F
Binder strain at 3600 second• of Average s~.ndard CO.ftici-.,t. cr..p
Type Loadin11 (in/in) Strain o.v iation of Kodlllua
va riation (p•J:)
Rep. A Rep. B Rep. c <' >
ACS F F F --- --- --- 0
AC20 0.0056 0.0037 o. 0037 0.0043 0.0011 25.3 11538
AC40 NA 0.0031 0.0034 0.0033 0.0002 6.5 15385
AC5RE F F F --- --- --- 0
AC5R 0.0015 NA 0.0104 0.0110 0.0008 7.1 4566
AC20R 0.0088 0.0053 NA 0.0071 0.0025 35.1 7092

Compressive Strain (in/in) Compressive Strain (in/in)

0 . 02 ..--~~~~~~~~~~~~~~~~~~~~, 0 . 02 ~~~~~~~~~~~~~~~~~~~~~
0.015 0.015
0 1200 2400 3600 0 1200 2400 3600

Time (seconds) Time (seconds)
- AC40 -+- AC6R -+- AC6 - AC40 -+- AC6R -+- AC6
-e- AC20R --*- AC20 ~ AC6RE -e- AC20R ---- AC20 ~ AC6RE
FIGURE 4 Compressive strain versus time for repeated FIGURE 5 Compressive strain versus time for repeated
loading conducted at 77°F. loading conducted at 104°F.
failed before the 60 min of loading. This is shown in Figure concrete mixtures reduces the magnitude of the loss of stiff-
S. It can be seen from this figure that the ACS failed after ness at higher temperatures.
approximately lS min of loading and the AC20 failed after
20 min of loading. This indicates that even though the samples
failed, the AC20 mixtures were stiffer than the ACS mixtures. Comparison of Static to Repeated Load Permanent
The AC40 mixtures performed well, yielding relatively low Deformation Testing
strains. The modified mixtures yielded strains that also follow
the idea of higher viscosity leads to lower strain. The ACSRE The relative ranking of strain changes for both testing con-
produced the highest strains, followed by the ACSR and ditions when 77°F test results are compared with 104°F test
the AC20R. The ACSR acted similarly to the AC40, whereas results. The 77°F test results are useful to assess the loss in
the AC20R exhibited the lowest amount of strain of any of stiffness when compared with testing at 104°F; however, be-
the six types of mixtures. This indicates that for this particular cause of the low testing temperature, they do not appear to
aggregate source and gradation, an ACSR could be expected be appropriate for characterization of permanent deforma-
to behave like an AC40 in warmer temperatures. An AC20R tion .
could be expected to exceed the permanent deformation per- The static test results at 104°F indicate only the presence
formance of an AC40. It can be concluded from this that the of rubber and nothing about the properties of the asphalt
addition of rubber to the mixture produces a stiffer mixture cement rubber blend. The repeated load testing at 104°F in-
at higher temperature. dicates, in a concrete manner, the differences that exist be-
The average creep modulus calculated at 60 min of loading tween the different binders. This is supported by comparing
for the six mixtures for both temperatures of repeated load the static testing at 104°F (Figure 2) to the repeated load
testing is shown in Figure 6. It can be seen that all unmodified testing at 104°F (Figure 5).
mixtures either exhibited very large decreases in stiffness from On the basis of information presented in Tables 5 through
77°F to 104°F or no stiffness at all . On the other hand, the 8 and Figures 1 through 6, two conclusions can be reached.
rubberized mixtures exhibited either very small decreases, or First, permanent deformation testing should be carried out
as in the case of the ACSRE, showed an increase in stiffness. at elevated temperatures. Not only does rutting occur pri-
This again indicates that the addition of rubber to asphalt marily at the elevated temperatures, but the modified mix-
TABLE 8 SIMPLE STATISTICS FOR STRAIN AT END OF LOADING FOR REPEATED LOAD
Binder Strain at 3600 Second• ot Averaqe Standard coettici•nt c:i;..p

Type Loadincr Cin/:l.nl Strain Dev:l.at:l.on of Jk>dulua
Variation (psi)
Rep. A Rep. B Rep. c <'>
AC5 F F F --- --- --- 0
AC20 F F F --- --- --- 0
AC40 0.0091 0.0050 0.0067 0.0069 0.0021 29.7 2885
AC5RE 0.0141 0.0108 o. 0114 0.0121 0.0018 14.5 1653
AC5R 0.0076 0.0097 NA 0.0087 0.0015 17.2 2312
AC20R 0.0033 NA 0.0036 0.0035 0.0002 6.1 5797
F - indicates sample failure prior to sixty minutes· of loading

Average Mix Stiffness at 60 min (psi)
/
,/
15000 ,/
,
//
/
10000
/
/
5000
/
0
77°F
104°F
AC8
0
0
AC20
11838
0
AC40
18188
2888
AC8RE
0
18113
.....
AC8R
2312
AC20R
7082
11487
FIGURE 6 Creep modulus from static testing at 77°F and 104°F.
tures appear to react differently at the lower temperatures. at 104°F. The static test results indicate only the presence of
This conclusion is supported by both the static and repeated rubber and nothing about the properties of the base binder.
load test results. Second, permanent deformation testing should The repeated load testing indicates, in a concrete manner,
be based on repeated loading. Static testing only indicates the the differences that exist between binders.
presence of rubber and nothing about the base asphalt.
ACKNOWLEDGMENT
CONCLUSIONS
The authors would like to thank the Asphalt Rubber Pro-
The following conclusions can be drawn on the basis of the ducers Group, U.S . Army Corps of Engineers, University of
analysis presented in this paper: Arizona, and Crafco Inc., for their support and cooperation
in this research program.
1. The addition of ground tire rubber to asphalt concrete
mixtures results in mixtures that exhibit less permanent de-
formation at high temperatures compared with unmodified REFERENCES
mixtures, remembering that the rubberized mixtures con-
tained higher-than-optimum asphalt contents. This proved to 1. M. Stroup-Gardiner, N. C. Krutz, and J. A. Epps . Comparison
be true for both static and repeated load testing. of Mix Design Methods for Rubberized Asphalt Concrete Mix-
2. Permanent deformation testing should be carried out at tures. In Proc., National Seminar on Asphalt Rubber, Kansas City,
Mo ., Asphalt Rubber Producers Group, 1989.
elevated temperatures. This conclusion is supported by both
2. Road and Paving Materials; Traveled Surface Characteristics. ASTM
the static and repeated load test results. The relative ranking 1990.
of strain changes for both testing conditions when the 77°F 3. J. G. Chehovits. Design Methods for Hot-Mixed Asphalt-Rubber
test results are compared with the 104°F test results. Concrete Paving Materials. In Proc., National Seminar on Asphalt
3. Permanent deformation testing should incorporate re- Rubber, Kansas City, Mo., Asphalt Rubber Producers Group,
1989.
peated loading. This is not only a better model for including 4. Proposed Standard Test Method for Unconfined Static Creep Test
the effects of moving wheel loads, but is supported by com- on Asphalt Mix Specimens. ASTM Sub-committee D04.20, Dec.
paring the static testing at 104°F to the repeated load testing 1988.

1339 PDF

Uploaded by

Copyright:

Available Formats

1339 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

1339 PDF

Uploaded by

Copyright:

Available Formats

TRANSPORTATION RESEARCH

A peer-reviewed publication of the Transportation Research Board

TRANSPORTATION RESEARCH BOARD

Design and Construction of Asphalt Paving Materials with Crumb 1

JVirginia's Experimentation With Asphalt Rubber Concrete 9

)Florida's Approach Using Ground Tire Rubber in Asphalt 16

Advances in Technology of Asphalt Paving Materials Containing 23

/Use, Availability, and Cost-Effectiveness of Asphalt Rubber in Texas 30

q'>ennanent Deformation Characteristics of Recycled Tire Rubber- 38

Design and Construction of Asphalt

Tire rubber is the principal component in CRM. Tire rubber

A surface treatment using an asphalt rubber spray application

1. Franklin Associates, Ltd., and R. L. Hershey. Markets for Scrap

Virginia's Experimentation With Asphalt

TABLE 1 GRADATIONS OF MIXTURES TABLE 3 PHYSICAL PROPERTIES OF

Note: No gnidation tests were performed. Results

The design criteria and Marshall properties for the mixtures

TABLE 4 MARSHALL DESIGN CRITERIA AND TEST RESULTS

V'l'M(%) 2.4 6.1 3.8 6.0 3-4 4-6

FIGURE 1 Pavement voids predicted by GTM. FIGURE 3 Gyratory stability index.

- 0.79 DEGREES 0 1.0 DEGREES Control 7,900 900 0.080 0.027

Marshall designs were performed on each binder at each

TYPE OF MIXTURE TIME (HRS)

- 0.75 DEGREES c:::::J 1.0 DEGREES D '!i'/, CR + Ill"/, CR o 15'1. CR

TIME, HRS RUBBER CONTENT ('!.)

5'!. FR x ID'!. FR v 15'!. FR c nNE CRUMB RUBBER + COARSE CRUMB RUBBER

RUBBER CONTENT ('!.) '!.RUBBER

- COARSE RUBBER c:J FINE RUBBER

REFERENCES in Airport Pavements. In Proc., National Seminar on Asphalt Rub-

Florida's Approach Using Ground Tire

It was decided to concentrate on the use of fine ( - 80 mesh)

to conduct these tests and to observe construction procedures Asphalt Content

BLENDING REQUIREMENTS binder contents. Increased binder in combination with the

TABLE 3 GRADATIONS OF GROUND TIRE RUBBER

TABLE 4 ASPHALT-RUBBER BINDER

USES Den•e-gradad Open-graded A•pbalt-R.W>ber

<•l Use of finer rubber could result in the reduction of the

NOTE: The mi n i mum r eact i on t im e. ma y be ad j usted if appr oved by the

Advances in Technology of Asphalt

In 1985 it was discovered that the reaction processes could

• Unique aggregate gradation,

falling after a certain reaction time, indicating that the binder

Evaluation of Storable Asphalt Rubber Binder Effect of Oil Extender

Vlacoslty (cp) REACTION AT 180 'C

TBA ("C) TBA 'C REACTION AT 180 'C

68 REACTION ~T 180. 'C

Until development of the storable asphalt rubber binder, it

1. Inability to supply high-capacity hot-mix plants with

Use, Availability, and Cost-Effectiveness

ASPHALT-RUBBER CHIP SEALS (SAMs)

6. The present performance of asphalt rubber suggests that

M~ch of the early research shows that asphalt-rubber chip

1. An excellent material for use in a dry, hot area. Some Cost-Effectiveness

Texas Highway District Survey

As a part of this study, a telephone survey of all the districts

ASPHALT-RUBBER INTERLAYERS (SAMls)

District 24: El Paso

District 24 currently has six asphalt-rubber interlayers under

ASPHALT-RUBBER CRACK SEALANTS RUBBER-MODIFIED ASPHALT CONCRETE