SHRP S 329

SHRP-S/FR-92-109
Condition Evaluation of Concrete

Bridges Relative to
Reinforcement Corrosion
Volume 7: Method for Field

Measurement of Concrete
Permeability
David Whiting
Construction Technology Laboratories, Inc.
Skokie, IL
Philip D. Cady
Pennsylvania Transportation Institute
The Pennsylvania State University
University Park, PA
Strategic Highway Research Program

National Research Council
Washington, DC 1992
SHRP-S/FR-92-109
ISBN: 3(Z)-05263-7
Contract C-IOI
Program Manager: Don M. Harriott

Project Manager: Joseph F. Lamond
Copy Editor: Katharyn Bine Brosseau
PTI Technical Editor: Joanne M. Fox
Project Area Secretary. Ann Saccomano
September 1992
Key words:
air flow
bridges
bridge decks
chlorides
concrete
in situ testing
nondestructive testing
permeability
vacuum
Strategic Highway Research Program

2101 Constitution Avenue N.W.
Washington, DC 20418
(202) 334-3774
The publication of this report does not necessarily indicate approval or endorsement of the findings, opinions,
condusions, or recommendations either inferred or specifically expressed herein by the National Academy of
Sciences, the United States Government, or the American Association of State Highway and Transportation
Officials or its member states.
Acknowledgments
The research described herein was supported by the Strategic Highway Research Program
(SHRP). SHRP is a unit of the National Research Council that was authorized by Section
128 of the Surface Transportation and Uniform Relocation Assistance Act of 1987.
We wish to acknowledge the cooperation of the Texas Department of Transportation and the
Pennsylvania Department of Transportation in making structures available for field testing.
In particular, the author would like to thank the following individuals: Melvin Stephens and
Jeff Howell of the Texas Department of Transportation, Austin, Texas, and Thomas Bell and
Gilbert Almalaz of the Texas Department of Transportation, Corpus Christi, Texas, for their
assistance in various phases of the field trials. The author would also like to thank Scott
Sabol and Edward J. Gannon of the Pennsylvania State University for their assistance in
arranging for the field trials in Centre County, Pennsylvania.
Finally, the author would like to acknowledge the contribution of Texas Research Institute
(TRI), Austin, for their participation in the construction and testing of the field prototype
developed under this research program. In particular, the efforts of George A. Matzkanin,
Greg McCord, and Paul Dostal are gratefully appreciated.
..°
111
Contents
Abstract • • • • ° • * " • ° * • ° ° " • • ° • * * * ° ° * ° * • ° • • * ° ° ° ° * ° • * * • ° * • • ° * • ° 1
Executive Summary ............................................ 3
1. Introduction ........................................... 5
Needs 5
Objectives 6
Scope 6
2. Literature Survey ........................................ 7

Basic Mechanisms 7
In Situ Test Methods 10
Ranking of Teclmiques 12
3. Method Development ..................................... 15

Basis and Bench Prototype 15
Design and Preparation of Test Specimens 17
Preliminary Testing of Bench Prototype 18
Moist-Cycled Series Test Slabs 23
Correlation Studies 30
Comparisons of Results 32
4. Design and Construction of a Field Device ....................... 35

Design Criteria 35
Construction of Field Prototype Device 36
Overall Features 36
Internal Components and Electronic Circuitry 38
Operation of the Field SAF Device 42
5. Preliminary Testing of Field Device ........................... 43
Comparisons with Bench Prototype 43
Temperature Effects 47
Testing of Outdoor Slabs 48
6. Field Trials ........................................... 51

Field Trial No. 1--Austin, Texas 51
Field Trial No. 2--Corpus Christi, Texas 56
Field Trial No. 3--Centre County, Pennsylvania 61
7. Summary and Recommendations ............................. 69

Summary of Test Method 69
Applicability of Test Method 70
Limitations of Test Method 71
Recommendations 71
Appendix A ............................................... 73
Appendix B ............................................... 81
References ................................................ 85
vi
List of Figures
3-1. Schematic drawing of the laboratory SAF device ..................... 16
3-2. Preliminary classification of permeability based on SAF measurements ....... 26
3-3. Comparison between temperature prof'des generated using 1500 W heat gun
and IR heater .......................................... 28
3-4. Relationship between SAF, chloride diffusion coefficients, and air and water
permeabilities .......................................... 34
4-1. Outline drawing of the permeability indicator ....................... 37
4-2. Overall view of SAF device configured for use on horizontal surfaces ....... 39
4-3. Remote vacuum head for testing vertical and overhead surfaces ............ 40
4-4. SAF device positioned on tripod support for testing of vertical surfaces ....... 41
5-1. Comparison between SAF results obtained using laboratory bench prototype
and field device ......................................... 46
5-2. Relationship between SAF readings and actual air permeabilities for cores
taken from PCA outdoor slabs ................................ 50
6-1. SAF device in operation on Westover Road Bridge, on Route 1, Austin,

Texas ............................................... 52
6-2. Operation of remote vacuum plate on prestressed beams of bridge over

Westover Road, Austin, Texas ................................ 55
vii
6-3. Comparison of in situ SAF and laboratory permeability testing for
Austin, Texas, field sites ................................... 57

Corpus Christi, Texas, field sites .............................. 62

Centre County, Pennsylvania, field sites ........................... 68
A-1. Instrument module removed from housing of SAF device ............... 74
A-2. General electronic component diagram ........................... 77
A-3. Instrument panel showing digital readout gages for vacuum, flow, and
elapsed time ........................................... 78
o..
Vlll
List of Tables
2-1. Ranking of in situ permeability methods .......................... 14
3-1. Concrete test mixtures ..................................... 18
3-2. SAF measured on air-dried slabs .............................. 19
3-3. Measurement of SAF test repeatability ........................... 21
3-4. Effective depth of vacuum penetration ........................... 22
3-5. Effects of steel placement and concrete cover ....................... 23
3-6. Comparison of SAF test results on air-dried and moist-cycled slabs ......... 24
3-7. Effects of surface drying using 1500 W electric heat gun ................ 27
3-8. Effects of surface drying using IR heater--5 minutes of heating ............ 27
3-9. Effects of surface drying using IR heater--3 minutes of heating ............ 29
3-10. Effects of "natural" drying on SAF ............................. 30
3-11. Chloride diffusion constants (De) for test slabs ...................... 31
3-12. Air permeability as determined from pulse-decay measurements ........... 33
3-13. Water permeability as determined from pulse-decay measurements .......... 33
5-1. Characteristics of concrete mixes used for comparison studies ............. 44
ix
5-2. Comparison of SAF readings for bench prototype and field unit ........... 45
5-3. Effects of temperature on readings taken with SAF field unit ............. 48
5-4. Results of comparative tests on field slabs .......................... 49
6-1. SAF readings taken on deck of bridge over Westover Road, Austin,
Texas ............................................... 53
6-2. SAF readings taken on prestressed support beams of bridge over

Westover Road, Austin, Texas ................................ 54
6-3. Results of air permeability tests carded out on cores removed from
bridge over Westover Road, Austin, Texas ........................ 58
6-4. SAF readings obtained on deck of Nueces Causeway Bridge in Corpus

Christi, Texas .......................................... 59
6-5. SAF readings taken on pier cap of J.F.K. Intra-coastal (;anal Bridge,
Corpus Christi, Texas ..................................... 60
6-6. SAF readings taken on prestressed beams on J.F.K. Intra-coastal

Canal Bridge .......................................... 61
6-7. Results of air permeability tests carded out on cores removed

from bridges at Corpus Christi, Texas ........................... 63
6-8. SAF readings taken on deck of bridge on Route 322, Port Matilda,
Pennsylvania ........................................... 64
6-9. SAF readings taken on deck of bridge on Route 220, Julian, Pennsylvania ..... 66
6-10. Results of air permeability tests carried out on cores removed from
bridges in Centre County, Pennsylvania .......................... 67
X
Abstract
A prototype surface air flow (SAF) device for the estimation of concrete surface permeability
has been developed. The method measures the rate of air flow through a vacuum plate
placed on a concrete surface under a vacuum of approximately 25 in. of mercury (16.6 kPa
absolute pressure). The effective depth of measurement is approximately 0.5 in. (12 mm)
below the surface. For calibration of the method, concretes were cast using a variety of
water-to-cement ratios as well as admixtures such as latex and silica-fume. Results were
found to correlate well with chloride diffusion constants derived from 90-day ponding tests,
as well as with true air permeabilities measured using a pulse decay technique.
A field prototype based on the SAF method was also developed. This portable field device
operates off rechargeable nickel-cadmium (Ni-Cad) batteries. Reusable closed cell foam
gaskets are used to ensure a tight seal to rough surfaces. A detachable vacuum plate can be
used to obtain measurements on vertical or overhead surfaces. Readings can be obtained at
the rate of approximately one per minute, allowing a large amount of information to be
developed at close intervals across a given concrete member.
Executive Summary
While much of the interest in concrete permeability dates back to initial construction of
hydroelectric structures in the 1930s, there is a renewed awareness of the role that
permeability plays in the ultimate durability of concrete structures. Keeping moisture and
aggressive substances from penetrating the concrete can minimize the effects of such distress
mechanisms as freeze-thaw damage, sulfate attack, and corrosion of reinforcing steel.
While various methods axe available for determination of concrete permeability in the
laboratory, these are mostly time-consuming and require the removal of cores from the
structure in question. Since coring is a destructive process, this limits the amount of
information that can be obtained from a given structure. A number of more rapid, in situ
techniques have been developed in recent years; however, many are at least semi-destructive,
and the testing time may be slow on particular types of concrete.
The flow of air through a concrete surface can be used to characterize the permeability. The
surface layer is especially important with regard to corrosion of reinforcing steel, since the
temperature and shrinkage steel is often located less than 2 in. (50 mm) below the concrete
surface. An indication of concrete surface permeability may be obtained by applying a
temporary vacuum seal to the concrete surface and measuring the air flow (in ml/minute)
under a vacuum of approximately 25 in. of mercury (16.6 kPa absolute pressure). That is,
higher air flows would indicate a more permeable concrete, and conversely, lower air flows
a less permeable concrete.
A laboratory device incorporating these principles was developed and applied to a series of
concrete specimens coveting a range of w/c ratios and associated permeabilities. Companion
slabs were ponded with sodium chloride solutions in order to obtain chloride ion diffusion
data. Cores were also obtained for measurement of air permeability using standard
laboratory techniques. Correlation between the air-flow device and standard techniques was
quite good, indicating that a rapid indication of concrete permeability could be obtained with
the proposed technique.
3
By inserti_m of pressure wells into the test specimens, the effective range of the vacuum was
found to be approximately 1/2 in. (12 mm) below the surface of the concrete. Therefore, the
technique can be viewed as being most sensitive to the near-surface region. Moisture content
of concrete was found to exert an influence on the test, but this can be minimized by surface
drying prior to testing using a hand-held infrared heater.
Encouraged by the laboratory testing results, a portable field device that operates off
rechargeable Ni-Cad batteries was developed. Reusable closed-cell foam gaskets are used to
ensure a tight seal to rough surfaces. A detachable probe was developed to obtain
measurements on vertical surfaces. Readings can be obtained at the rate of one per minute,
allowing a large amount of information to be developed at close intervals across a given
concrete member.
Field trials of the device were carried out during the first 6 months of 1991. Trials took
place on bridges in the states of Pennsylvania and Texas. Bridge elements tested included
decks, support beams, and pier caps. While some problems were encountered, mostly
dealing with rough or cracked surfaces, the equipment in general performed satisfactorily. It
was possible to test a 40-ft (12-m) long span in one lane over a period of about 2 hr using a
test grid established on 3-ft (l-m) centers. Support elements can also be tested, although in
many cases scaffolding, man lifts, or other devices may be needed to access elements beyond
reach. In addition to the nondestructive testing, cores were obtained from a number of
locations on each test structure and subsequently tested in the laboratory for air permeability.
There was a general relationship between field measurements mad actual air permeabilities.
However, the relationship was not sufficiently quantitative to allow for prediction of actual
permeabilities from the results of the field tests. The test must therefore be regarded as an
indicator of relative permeability, and if more accurate values are needed, cores should be
obtained and tested using more standardized laboratory techniques.
4
1
Introduction
Needs
Long-term durability of concrete structures may be directly related to the permeability of the
concrete. If an aggressive substance can be kept out of concrete by virtue of low
permeability, then associated problems, such as freeze-thaw deterioration, corrosion of
reinforcement, and formation of expansive components, might be mitigated.
Highway industry interest in producing durable concrete has created a demand for rapid
techniques capable of measuring the permeability of concrete to water and chloride ions.
Measurement of the chloride ion permeability is by far the most important concern, as it is
the chloride ion that promotes depassivation of the reinforcing steel, resulting in initiation of
corrosion. While moisture is also needed for corrosion to be initiated, in most
atmospherically exposed structures sufficient moisture is available from direct rainfall, snow
melt, and runoff so that moisture contents at the level of the steel are usuaUy sufficient for
initiation of corrosion. Current tests for measurement of chloride ion permeability require
either taking cores from the structure (1) or obtaining powder samples for subsequent
chemical analysis (2). The latter is a viable option only if the structure has previously been
exposed to chloride ions. Rapid, nondestructive tests are needed that could afford agencies
with a profded permeability across the structure without resorting to an unduly high number
of cores or drill samples being taken from the structure in question.
5
Objectives
The objective of the research described here was to develop a rapid field test for concrete
permeability that would correlate reasonably well with existing laboratory techniques. An
emphasis was placed on correlation with permeability to chloride ions. It was recognized
that any rapid, nondestructive field test would most likely not yield a true permeability (or
chloride diffusion) value but would afford an empirical result that could be taken as a
measure of relative permeability. The field test method was designed to be centered around
a portable and self-contained device that does not require any disturbance of the concrete in
order to obtain a test result.
Scope
The projex:t's scope was:
1. demonstration of the principles of the device in a laboratory situation;
2. study of the influences of materials and environmental variables on test results;
3. correlation with other measures of permeability;
4. construction of a field device based on the laboratory prototype; and
5. evaluation of the device under actual field conditions.
6
2
Literature Survey
A literature survey of rapid permeability devices provided guidance from the experiences of
others and ensured that devices meeting the criteria of the project had not already been
developed.
As a first step, a position was adopted to view "permeability" in a broader context than
simply that associated with saturated flow under a hydraulic gradient. Permeability may be
viewed as the ability of a given concrete to resist penetration of a particular substance, be it
liquid, gas, or ions. This ability may be expressed in absolute units of mass flow, by the
amount of substance deposited in a given time, or as a relative ranking, depending on the
particular test procedure employed.
Basic Mechanisms
The need for data on concrete permeability dates from the early 1930s, when designers of
large hydraulic structures required information on rates of passage of water through concrete
under the influence of relatively high hydraulic heads. In such instances, flow of water
through concrete can be adequately described by D'Arcy's Law (3) as:
7
Q = _k.A •dp (2-i)
ds
where
Q = volume outflow (cm3/sec)

A = area (cm2)
= viscosity (centipoise)
dp/ds = pressure gradient (atmos/cm)
k = permeability constant CD'Arcy's)
If the International System of Units (SI) is used throughout, then the permeability constant is
expressed in terms of m2. Primary factors influencing permeability were found to be water-
to-cement (w/c) ratio and age of concrete. Glanville (4) noted a twentyfold decrease in
permeability between 7 days and 12 months of age for a concrete with w/c ratio of 0.60.
Recent studies by Gaerty and Freeman (5) and Whiting (6) have confirmed the importance of
age of concrete on permeability, especially as regards the benefits of early moist curing on
subsequent reductions in permeability. Permeabilities to such fluids as air (7), liquid
nitrogen (8), methane (9), and oil (10) have also been measured. In general, permeabilities
to gases are from one to two orders of magnitude greater than those for water, which is to be
expected from comparative viscosities of the respective fluids.
Under conditions other than those of saturated fluid flow, transport of substances through
concrete can occur by a variety of different mechanisms. These may include capillary
attraction, vapor transmission, or ionic diffusion. Data developed by Wing (11) and
Dunagan (12) indicate that movement of water into dry concrete via capillary attraction can
be quite rapid and that flow of water into a surface under low heads follows an inverse
power law:
dr
- at-n (2-2)
dt
where
dr
dt - flow rate (ml/s)
t = time (in seconds)
n = a dimensionless constant
a = the volume of flow that has occurred at time t (ml)
An initial surface absorption test, ISAT (13), has been developed using these principles.
Concern with corrosion of reinforcing steel promoted by chloride ions that penetrate through
the concrete cover and eventually reach the reinforcement has resulted in research that has
demonstrated that such migration of chloride ions occurs primarily through diffusion
processes. Various studies (14,15) indicate the coefficient of diffusion to be on the order of
10Tto 10-s cm2/seconds. Diffusion will generally follow Fick's Second Law as:
0c 02c
- D "_ (2-3)
8t OX2
where
C = concentration of distance x (in meters) from a boundary

t = time (sec)
D = effective diffusion coefficient (m2/seconds)
Procedures used to construct chloride profiles and determine diffusion coefficients for actual
structures are discussed by Weyers and Smith (16). Improved laboratory techniques that can
rapidly assess the permeability of concrete to water, gases, ions, or other substances have
been developed. These include techniques based on high pressures (17), transient gas flow
(18), or ionic conductance (19). While these methods offer a marked improvement in testing
speed over previous steady-state techniques, they still require somewhat time-consuming
specimen preparation and the removal of a core or powder sample from the structure, which
makes testing a destructive process.
In Situ 'rest Methods
One of the first methods developed for field indication of concrete permeability was the
ISAT (131).
The test apparatus consists of a gasketed cap, which is either clamped or affixed with sealing
putty to the concrete test surface. Water is poured into the inlet until the outlet runs clear.
A capillary tube is then affixed to the outlet tube, an initial reading is taken, and subsequent
readings are obtained at 10 minutes, 30 minutes, 1 hour, and 2 hours. To date, the
apparatus has been used on reinforced concrete, paving, and architectural concrete with good
results.
The method has been standardi_ed in BS 1.881 Part 5-1970; recent revisions may eliminate
the 2-hour measurement. In most on-site applications, measurements are often limited to the
10-minute reading. Problems encountered with the technique in the field include difficulties
in achieving a watertight seal, securing the rig to the concrete (which may require the
drilling of anchor-bolt holes in some instances), and inconsistent results during cold weather
or when the surface was damp.
Since the initial development of the ISAT test, there has been an increased interest in
development of test procedures for determination of in situ permeability. A number of
non-steady state, or indirect, methods have been developed by a variety of researchers. Figg
(20), Cather, Figg, et al. (21), Schonlin and Hilsdorf (22), Kasai et al. (23), and Hansen
et al. (24) have concentrated on techniques based on non-steady state measurement of water
absorption or air permeability.
Figg's (20) procedure consists of drilling a hole 1.2 in. deep x 0.2 in. (30 mm x 5.5 mm) in
diameter into the concrete, sealing the hole with a silicone rubber plug, ensuring an air-tight
seal with a hypodermic needle placed through the silicone plug, and then monitoring the rate
of fall of water in a capillary after injecting water with the hypodermic syringe into the small
cavity in the concrete.
The method has also been adapted to measure air permeabilities. The hole and sealant are
the same as previously described for the water injection method. The run is started by
turning the three-way stopcock to allow air to be withdrawn from the concrete until a
vacuum of 112 mm I-Ig (14.8 kPa) is reached. The pump is then isolated and the time
required for the pressure to rise to 150 mm Hg (19.9 kPa) is recorded. This value is taken
as a relative measure of the air permeability of the concrete.
10
The results obtained were found to be a strong function of the moisture content of the
concrete. Variations in moisture content from 1.0 to 1.8 percent led to a maximum
difference in At of 20 seconds. Typical test times run from 100 to 500 seconds, indicating
that errors of up to 20 percent are to be expected if the concrete has not been previously
calibrated with respect to moisture content versus At. Although it is possible in the
laboratory to determine moisture content by subsequent oven-drying (and thereby develop
correction factors to the air permeability data), a satisfactory field technique is not as yet
generally available (23).
Cather et al. recently (21) proposed modifications to the original Figg technique, which
apparently result in an improved test procedure. A hole diameter of 0.4 in. (10 mm) and
plug depth of 1.6 in. (40 mm) were selected. The pressure is reduced to 13 in. (330 mm) of
Hg (43.9 kPa) and the time required for the pressure to rise 14.6 in. (370 ram) (49.2 kPa) is
recorded on a digital manometer. Cather et al. concluded that within the period of the test,
most of the pressure increase is due to air being drawn from a local area near the test hole
rather than from the exterior surface of the test slab.
Scholin's test (22) consists of a 2.0-in. (50-mm)-diameter vacuum chamber (volume = Vs),
which is placed on a smooth concrete test surface previously heated to about 122°F (50°C)
for 5 minutes. Pressure is reduced to 0.02 to 0.20 in. (0.5 to 5 ram) Hg (.06 to .66 kPa)
and then the vacuum pump is shut off. The time at which the pressure reaches a value of
1.5 in. (37 mm) Hg (4.9 kPa) is recorded as to. After 2 minutes, the pressure is recorded as
/1. When the pressure reaches 230 mm Hg (30.6 kPa), a final time (tl) is recorded. The
permeability index is then given as:
M = (PI - Po) "V

P1 (2-4)
(t1 - to) • P. "('P. + to)
2
where
Pa = atmospheric pressure, and other quantities are as previously noted
A final category of in situ permeability techniques meriting discussion is the pressure-applied

surface test developed by Montgomery and Adams (25)--the CLAM test. In this technique, a
pressure vessel containing water is affixed to the concrete surface by means of a rapid-setting
11
epoxy adhesive. As water flows into the concrete, pressure is maintained by means of a
vernier-controlled piston. Vernier readings before and after the test can then be used to
indicate the amount of water flowing into the sample. This test is apparently simpler to
apply than the ISAT or Figg methods; however, the need to glue the test rig to the concrete
surface may result in some damage to the surface during removal.
R2nking of Techniques
The need for a rapid nondestructive technique led to the development of a number of criteria
that should be met by any candidate technique and that could be used to rartk existing and
proposed methods. These include the following criteria, which were used to rank the
previously cited methods as well as others uncovered during the literature search. A scale of
-3 to + 3 was used, with positive values indicating benefits of the technique in terms of the
particular criteria and negative values indicating disadvantages.
• Correlation--The method should correlate either with long-term chloride ingress or

with concrete characteristics known to be determinants of chloride ingress, such as
w/c ratio, extent of curing, and use of permeability-reducing additives.
• Interferences--An ideal method would have no interferences; however, this may be

unrealistic. If interferences are present, they may be manageable if they can be
compensated for. The method will suffer otherwise.
• Invasive--If the method requires removal of concrete, or if substances will be

injected into the concrete, the method may be considered invasive. An ideal
method would rely on surface measurements with no substances injected into the
concrete. On the other end of the scale, methods that rely on cores taken from the
concrete would not be promising.
• Mounting--Methods that require mounting on a concrete surface will damage the

surface and may be objectionable, especially on riding surfaces. Temporary types
of bonding such as resilient adhesive or caulk may be less objectionable.
• Rapidity--Rough calculations indicate that an "instantaneous" method (i.e., one

where each reading takes less than 10 see) would enable one to prof'tle two lanes of
a typical bridge deck 100 ft (30 m) long at 3 ft (1 m) centers in 1 hour. One could
do the same at 5-ft (1.5-m) centers if readings took less than 1 minute each. If
readings took 5 minutes each, then a deck profile at 5-ft (1.5-m) centers would still
be possible within 4 hours (1/2 day) of testing, which is not an unreasonable rate.
12
At over 10 minutes per reading, however, only 24 total readings could be obtained
in the same time period. Obviously, slower tests reduce the amount of information
obtainable.
• Orientation--Since substructures were included within the scope of the project,

methods capable of being used in any orientation (horizontal, vertical, or overhead)
are desirable. Methods relying on force of gravity or on containment of liquids
(which is difficult in vertical or overhead configurations) suffer in this regard.
• Weight--A lightweight instrument is desirable, especially for substructure work.
• Number of Components--Reduction in the number of components results in reduced

time needed for set up and reduced complexity. A self-contained unit is the ideal
case, though this may be impractical if locations with difficult access are to be
tested.
• Power--Units not requiring electrical power or capable of operation on standard

low-voltage batteries are desirable. AC power requires gas generators and makes
testing more cumbersome.
• Number of Operators--The ultimate cost of the test will increase with the number
of operators required. A small unit that can be operated by one person is the
ideal.
• Operator Skill--Highly skilled personnel are not always available. Simple units
that can be operated by relatively unskilled personnel would be ideal.
Test methods were scored using these criteria, and the results are presented in Table 2-1.
The air permeability methods offer considerable advantages in terms of many of the criteria
deemed important for an on-site method. On the average, they scored fairly high on
correlation with such parameters as w/c ratio, strength, and extent of curing. Interferences
are something of a problem, the predominant one being moisture content prior to testing.
This is more of a problem for methods relying on drilled holes than for surface methods,
since the surface layer can be dried prior to testing in a fairly short time period. The
methods are invasive when drilled holes are used, but not for the surface types. Mounting is
generally not required. The methods are not overly rapid, but one might expect to make
some improvements in this area. Air permeability methods score fairly high on orientation
and weight, but many are "first generation" systems utilizing components rather than finished
13
Table2-i. Rankingof in situ
permeability
methods.
: Ranking
Method (Point Score) Reference
Figg Rir :permeabUity and modifications (+.3) to (+6) 20, 21, 23
Hansen surface air pressure (+2): 24

?
Schonlin surface vacuum (+ 11) 22
Mmi-permeameter for rock. cores (+ 21). 26

ISAT (+3) 13
Figgwater absorption .. (0).. 13
Pressure injection.of water (+ 4) 25
packages. The general lack of need for AC power is favorable, and the operator skill
required would generally be at the technician level.
Methods based on water permeability are somewhat less promising. Correlations are not as
good, and moisture appears to have a more serious interference effect. Because most of the
water methods operate from the surface, they are, on the average, somewhat less invasive
than the air techniques. However, most do require mounting and are fairly slow. Limits on
orientation are a problem. In terms of overall weight, number of components, and power
requirements, they are similar to the air methods. Operator skills are expected to be similar.
In summary, the most promising avenue of approach for development of a permeability

technique appeared to be a technique based on air permeability, which can be operated from
the surface and requires no mounting. A rapid flow-measurement type technique similar to
the "mini-permeameter" was very appealing, as wide-range electronic flow meters are
capable of spanning the entire range of flows to be expected, reducing the need to allow for
very long decay times, as appears to be the case for the vacuum decay techniques. The
remainder of this report describes development and testing of a device based on these
findings.
14
3
Method Development
Basis and Bench Prototype
The SAF technique is based on the rate of air flow through a concrete surface, more
permeable concretes being characterized by higher air flow rates. Rather than forcing air
into the concrete, which would require a direct pressure seal to be created at the surface, the
technique relies on the creation of a vacuum at the surface. The vacuum causes air to flow
out of the concrete along the pressure gradient. SAF uses the rate of flow of air for a rapid
measurement once the desired pressure differential has been achieved. The technique was
designed to give an indication of permeability for a wide variety of concrete materials.
These include very low permeability concretes such as those containing silica fume or
polymer modifiers as well as very high permeability concretes that result when high w/c ratio
and/or poor curing practices are employed. The effective test area was chosen as a circle
2.75 in. (70 mm) in diameter, offering a representative test on a small area of most structural
concretes. The test was designed to be applicable to support and substructural elements as
well, necessitating that the device also be operable in vertical and overhead modes.
To evaluate the basic concept of the device, it was deemed prudent to construct a laboratory
bench prototype prior to assembly of an actual field test instrument. A schematic diagram of
the laboratory bench prototype is shown in Figure 3-1. The 4-in. (102-mm) diameter steel
vacuum plate is fitted with a 0.125-in. (3-mm) thick soft rubber ring allowing for a circular
flow area 2.75 in. (70 mm) in diameter. Flow is initially directed through Valve A until the
15
IIIIIIIIIIIIIIIIIIIIIIi11!
II!11111111111111
_
E
E_
oE
',, o
I d -
=_ -_ _
o
0
\_.. _
r-
/ • • L'_
• N
)_ • *
°
¢,)
.W . e_
16 _
system reaches the operating pressure of 120 to 140 mm Hg (15.9 to 18.6 kPa) as measured
on the vacuum gage (a U-tube Hg manometer was used in the bench version). Valve A is
then closed and the air steam is directed through Valve B from the rotameter (flowmeter),
which is calibrated to read in units of ml/minute over a range of 1 to 280 ml/minute. The
total test time is approximately 1 minute, at which time the pump is turned off and air is
allowed to reenter the system through Valve C.
Design and Preparation of Test Specimens
After initial system checkouts were complete, a series of concrete test specimens designed to
offer a range of permeabilities were designed and constructed. The majority of specimens
consisted of plain (unreinforced) slabs 12 x 24 in. (305 x 610 mm) in width and length and
6-in. (152-mm) thick. A few reinforced slabs were constructed in order to determine the
effects of reinforcement cover on the test procedure. These slabs contained No. 5 (12-ram)
rebar mats positioned 0.5 and 1.5 in. (12 and 37 mm) below the top surface of the slab.
Additionally, in a few of the slabs, brass tubes 0.156 in. (4 mm) in diameter were positioned
0.25, 0.5, and 1 in. (6, 12, and 25 mm) from the surface to be used to measure extent of
depth of influence of the vacuum. Thermocouples were embedded 0.25, 0.5, 1, and 2 in. (6,
12, 25, and 50 mm) below the top surface.
The concrete mixtures used are shown .in Table 3-1. The mixtures were proportioned so as
to obtain a relatively wide range of permeabilities, while still remaining within the spectrum
of mixtures that could reasonably be expected to be encountered in actual field practice.
Slabs from Mixtures A through D were maintained under wet burlap/polyethylene overnight,
then transferred to heavy plastic bags for 28 days of curing. This was done in an attempt to
simulate actual field conditions. Although curing membranes may be used to retain moisture
to aid in curing, additional moisture is generally not deliberately provided to the slabs over
long periods of time. In contrast to Mixtures A through D, Mixture E was fully moist-cured
for 28 days to ensure that concrete having a high level of impermeability would be included
in the study. Following recommended practice for latex modified concretes (LMC),
Mixture F was exposed to air drying immediately after demolding, in order to develop
maximum curing of the latex admixture.
17
Table 3-1. Concrete test mixtures.
• Quantities-lb/yd 3 Water/ •
Cement Slump Air
Mix Cement Sand Gravel Water Ratio Admixture (in.) (%)
A 659 1263 1772 252 0.38 --_ 2.6 5.0
B •550 1375 1776 237 0.43 ---_ 3.2 5.3
C 453 1477 1760 227 0.50 ---_ 3.0 5.8
D 374 1563 1720 224 0.60 ---_ 3.8 6.4
E 665 1300 1680 239 0.36 Silica Fume- 5.0 5.0

67 lb
F 685 1579 13•11 185 0.27 Latex-25 gal 6.9 5.6

I
aNo admixtures other than Mr-entraining agents used.

Note: 1 lb/yd 3 = 0.59 kg/m3; 1 in. = 25.4 ram; 1 lb = 0.454 kg; 1 ga] = 3.785 L.
After the conclusion of initial curing, the slabs were coated on the four edges with moisture
barrier grade epoxy and placed in the following environments:
* Continuous air drying at 73 +__3°F (23 4- 1.7°C) and 50 4- 5-percent RI-I (air-dry
series); and
• Cycling of 1 day wet burlap applied to top surface at 75 to 80°F (24 to 270C) and
6 days drying (moist cycle series).
Slabs containing reinforcement was exposed to continuous air drying only.
Preliminary Testing of Bench Prototype
Air-Dried Test Series
The first _ries of tests was performed on the slabs that had been placed in an air-drying
environment for approximately 3 months after initial curing. A series of 15 individual flow
readings was obtained across both the cast and finished faces of each slab. Readings were
18
taken on a rectangular grid of 3- x 4-in. (75- x 100-mm) ceils avoiding the outermost cells so
as to minimize possible edge effects. Results are given in Table 3-2.
Table 3-2. SAF measured on air-dried slabs.
Air Flow (ml/minme)
Slab Readings Face Me,an Minimum Maximum Std. Dov. % C.V.
A-1 15 Cast 67 55 85 8.38 12.5
A-1 I5 Finished 50 40 65 6.97 13.7
B-1 15 Cast 68 55 90 8.80 12.9
B-1 14 Finished 87' 60 155 26.29 30.2
C-1 15 Cast 64 50 95 11.01 17.3
C-1 ---* Finished ---_ --) _b ___b _b
D-1 15 Cast 152 125 175 13.66 8:9
D- 1 15 Finished 128 85 175 29.56 23.1
E-1 I5 Cast 19 14 22 2.23 11_7
E-1 I0 Finished 16 12 23 3.14 19.4
F-1 15 Cast 22 17 30 3.54 15.8
F-1 10 Fini._hed 34 22 65 12:54 36.9
_Slab not tested due to surface crazing.

bNo reading.
The lowest readings were obtained for the silica fume concrete (E), and the highest for the
concrete with w/c = 0.6 (13). Readings for LMC (F) were somewhat higher than for silica
fume, as would be expected. Conventional concretes exhibited an increasing surface air flow
with increase in w/c, at least on the finished face. There was little difference in air flow
between concrete Mixtures A, B, and C on the cast face, perhaps representing an increased
densification at the bottom of the slab due to some absorption of the mix water by the
formwork. On the finished face of Slab C-1 it was not possible to achieve satisfactory
readings. Most readings were offscale (i.e., >280 ml/minute), and those that fell onscale
were unstable. Since it was believed that this might have been due to leakage around the
19
gasket, a silicone vacuum grease was applied to the gasket and some readings were repeated.
This stabilized some readings but most were still offscale. Upon inspection of the surface
that had been coated with grease, a fine network of surface microcracks in the thin surface
paste layer became visible. It is likely that this microcracking led to leakage under the
gasket and high measured air flows. This surface was then given a light sandblast to remove
the surface paste. Repeat testing, however, still indicated microcracks, which were now
even more visible on the sandblasted surface. It is possible that this slab may have been
prematurely finished, or suffered from some plastic shrinkage cracking.
The statistics reported in Table 3-2 were obtained from readings taken at 15 different
locations on each slab. Therefore, the variance estimates represent a combination of
variability in concrete properties from point-to-point on the surface as well as inherent test
variability. Variability is greater for the finished surfaces, perhaps reflecting differences due
to hand finishing across the slab. To obta_ a better estimate of test reproducibility, a series
of measurements consisting of five replicates on each of three positions on cast and finished
surfaces of each slab was performed. Summary statistics are given in Table 3-3 for the cast
faces.
The variability of successive measurements at one position is much less than the variability
across the slab. This indicates that the method actually detects differences in permeability
across the surface of the slab, most likely due to small differences in consolidation and
finishing across the surface. Reproducibility statistics are not shown for Slab F-1 (latex
concrete). This is because each reading appeared to be successively lower, indicating that air
was being evacuated faster than it was being replaced in the time interval between each
measurement.
The slabs cast with embedded brass tubes were used to measure the effective depth of
penetration of the vacuum into the slabs. During casting, a solid brass rod was extended
through the tube such that when it was removed, a 1-in. (25-ram) long by 0.156-in. (4-ram)
deep cavity was created at the 0.5- and 1-in. (12- and 25-mm) depths. A microflowmeter
and a manometer were attached to each well in order to measure air flow and vacuum level
at each depth. Results are given in Table 3-4.
20
Table 3-3. Measurement of SAF test repeatability.
Air Flow (ml/mlnute)
Slab Face Position Mean Std. Dev. % C.V.
A-1 Cast 1 52 2.09 4.0
6 75 2.74 3.6
15 54 2.19 4.1
B-1 Cast 1 58 2.50 4.3
5 68 2.73 4.0
11 50 2.74 5.4
C-1 Cast 1 50 3.06 6.1
7 53 2.33 6.1
15 89 2.24 2.5
D-1 Cast 6 121 6.52 5.4
7 153 2.50 1.6
11 161 4.54 2.8
E-1 Cast 1 16 1.79 11.0
6 21 1.12 5.5
11 23 1.04 4.5
21
For low w/c concrete (Slab A-2), only very small flow rates could be measured, even 0.25
in. (6 mm) below the surface. For a slightly higher w/c ratio (Slab B-2), flow at 0.25 in.
(6 mm) below the surface was approximately 20 percent of the surface reading. For high
w/c ratio concrete, flow was much greater at 0.25 and 0.50 in. (6 and 12 ram) and was even
detectable 1 in. (25 ram) below the surface. These results indicate that while the effective
test depth is somewhat dependent on the characteristics of the particular concrete under test,
the test is predominantly influenced by the near-surface (less than 0.5 in. [12 ram]) region of
the concrete cover.
Table 3-4. Effective depth of vaeuum penetration.
Slab Well Depth Air .Flow Vacuum

(in.) (ml/mlnute) (r.m Hg)
A-2 Surface 40.0 -635
0.25 1.0 ---"
0.5 0.5
I 0.0 ---"
B-2 Surface 65.0 -635
0.25 12.6 -180
0.5 2.1 ----_
1 0.0 --_
D_2 Surface 130.0 -635
0.25 t00.0 -180
0.5 13.5 -160
1 0.4 ----"
"No vaeuum detected.

Note: 1 in. = 25.4 turn; 1 mrn Hg = 0.0133 kPa.
22
Effects of steel placement and cover on air flow permeability readings were also investigated.
Readings were taken directly above and between No. 5 (16-ram) bars having 0.5 and 1.5 in.
(12 and 30 mm) of cover and spaced at 5-in. (127-mm) centers. Results are compared in
Table 3-5. There is little significant difference between readings taken above or between
reinforcing bars, indicating that reinforcing steel (and presumably other embedments) has
little effect on the test, provided cover depth is 0.5 in. (12 mm) or greater.
Table 3-5. Effects of steel placement and concrete cover.
Air Flow (rnl/minute)
0.5-1n. Cover 1.5-in. Cover
Slab Above Bars Between Bars Above Bars Between Bars
A-2 50 50 40 40
B-2 ---_ ----_ 60 65
D-2 155 145 135 130
aNo stable readings obtainable (crack in concrete above bar).

Note: 1 in. = 25.4 ram.
Moist-Cycled Series Test Slabs
Testing was performed on the slabs aged 28 days then subjected to a weekly wetting cycle
(1-day soak, 6-day air dry). The top (f'mished) surface of each slab was subjected to the
soaking, and the bottom (cast surface) was protected from direct wetting during the soak, but
was in a damp environment. Surface air flow tests were performed on both cast and finished
surfaces approximately 1 hour after the removal of the soaking blankets. Test results are
presented in Table 3-6.
For the finished surface, lower readings were encountered for the moist-cycled slabs. The
difference between dry and moist cycle mean readings ranged from about 15 ml/minute for
slabs prepared from Batch A to 30 ml/minute for slabs prepared from Batch D. Differences
for cast surfaces were less, as the cast surfaces were not directly exposed to liquid water.
To examine the effects of wetting on cast surfaces, the cast surfaces were placed upright, put
through three weekly cycles of wetting, then retested. As with the finished surfaces, there
23
Table 3-6. Comparison of SAF test results on Mr-dried and moist-cycled slabs.
Air Flow (ml/minute)
Batch Slab Face Condition _ Mean Minimum Maximum
A A-1 Cast Dry 67 55 85
Finished Dry 50 40 65
A-4 Cast Damp 60 45 70
Finished Moist cycled 45 25 75
B B-1 Cast Dry 68 55 90
B-4 Cast Damp 65 48 95
Finished Moist eycled 60 40 80
C C-1 Cast Dry 64 50 95
C-4 Cast Damp 71 55 115
D D-1 Cast Dry 152 125 175
D-4 Cast Damp 154 85 215
Finished Moist cycled 96 65 155
E E-1 Cast Dry 19 14 22
E-2 Cast Damp 22 15 27
F F-1 Cast Dry 22 17 30
F-2 Cast Damp 18 9 32
"Slabs surface-dry at start of test.
24
were considerable differences due to wetting of the test surfaces. These differences were
greatest for the high w/c concrete (D-4), and were insignificant for very low permeability
concretes (E-2 and F-2).
A summary of air flow measurements on the various sets of test slabs under differing
conditions of moisture is given in Figure 3-2. In spite of the differences induced by varying
surface conditions and moisture levels, a general trend of increasing air flow with decreasing
"quality" of concrete was seen. The concretes designed to have low permeabilities (Batches
E--silica fume, and F--LMC) do have the lowest air flows. Conventional concretes typical of
those specified for structural and paving application (Batches A, B, and C) showed a slight
upward trend for air flow as the w/c ratio is increased going from Batch A to Batch C. The
poorest quality concrete produced in this test series (Batch D) exhibited the highest air flow
reading.
A preliminary classification of permeability based on air flow measurements is suggested by

Figure 3-2. Readings less than 30 ml/minute are associated with low permeability concretes.
Readings between 30 and 80 ml/minute are associated with moderately permeable concrete
(w/c from about 0.40 to 0.5). Readings above 80 ml/minute are associated with high
permeability concretes. Readings in "borderline" regions may be ambiguous, for instance,
very wet high w/c concretes may on occasion give readings closer to the "moderate" zone.
In an attempt to devise a practical means of reducing the effect of near-surface moisture

content on test results, various drying strategies were carried out. A 12- x 2- x 4-in.
(304- x 305- x 102-mm) louvered sheet metal box was mounted on a slab so that an
embedded thermocouple "tree" (1/4-, 1/2-, 1-, and 2-in. [6-, 12-, 25-, and 50-mm]
thermocouple depths) was at the approximate center of the area. An electric heat gun (1500
W capacity) was then placed into a tube mounted on the top center of the box and activated.
Temperature profiles for Slab D-2 (w/c = 0.6) are shown in Figure 3-3. Profiles for other
slabs tested were quite similar. While surface temperatures reached close to 300°F (149°C),
temperatures at 0.25, 0.5, and 1 in. (6, 12, and 25 mm) peaked at approximately 140, 110,
and 95°F (60, 43, and 35°C) after 5 minutes of heating. The slab was then allowed to cool
for an additional 10 minutes until all temperatures fell below 90°F (32°C). At this time, air
flow tests were repeated. Results are given in Table 3-7.
25
Table 3-7. Effects of surface drying using 1500 W electric heat gun.
Initial :Flow FlowAfter Flow After

Air Dry _ Soaking b Heat Drying _
Slab w/c (ml/minute) (ml/minute) (ml/minute)
A-2 0.38 45/50 30135 40/45
B2 0.43 60/70 •45/50 65/70
D-2 0.60 130/140 95/105 125/135
aStored in air at 73 _ 3°F (23 __+1.7"C) and 50 + 5-percent RH for 10 months after curing.
bSoaked once weekly for 7 weeks, then continuously for 3 days.
CDried 5 minutes using 1500 W heat gun 4 in. (100 ram) above surface.
These results indicate that the surface heating was able to restore flow values close to those
obtained on initially dry surfaces. This approach, at least in principle, could offset the
interference that moisture seems to have on the test. The 1500 W heat gun requires a high
energy expenditure, and would necessitate that a relatively large generator be transported to
the test site. In order to provide for more compact equipment, the use of a small hand-held
liquid propane-fired infrared (IR) heater (such as is used for stripping paint) was
investigated. The heater was held 4 in. (100 mm) over the desired test areas for periods of 5
minutes each. The heating prof'lles, also shown in Figure 3-3, compare favorably with the
prof'lles obtained using the heat gun. After cooling, SAF measurements were made on the
heated areas. Results are presented in Table 3-8.
Table 3-8. Effects of surface drying using IR heater after 5 minutes of heating.
Initial :Flow •Flow After Flow After

Air Dry_ Soaking _ Heat Drying ¢
Slab w/c (ml/minu_) (rnl/minute) (ml/mlnute)
A-:2 0.38 45/50 30135 40145
B-2 0.43 60170 45/50 65/70
D42 0;60 130/140 110/t20 t-30/140
aStored in air at 73 + 3*F (23 + 1.7°C) and 50 4- 5-percent RH for 10 months after curing.
*Dried 5 minutes using IR heater 4 in. (100 ram) above surface.
27
Heat Gun - 1500W
300 -
surface
.......... 1/4-inch
200 ........... 1/2-inch

o ........... 1-inch
¢D
"- ...... 2-inch
'-¢ .f"_,--'"'_.,.. 1ram= 0,0394 in.

100 _...----'-_
E f °'" -'_'- - ..... _ °C = 5/9 ( °F 32 )
cU
P
0 , I v I
0 10 20
Minutes
I.R. Heater
300"
_ 1/4-inch
u. 200
o. ........... 1/2-inch
........... surface
1-inch
0- 100 .,,,....'_.'-.,.,-'-
l-• lmm = 0.0394 in.
°C = 5/9 ( °F - 32)
0 , I i ,,I
0 10 20
Minutes
Figure 3-3. Comparison betweentemperature prof'des generated using the heat gun and IR
heater.
28
The data demonstrate that the IR method is capable of restoring the test area to essentially
the initial air dry condition after 5 minutes of heating.
Examination of Figure 3-3 indicates that essentially the same temperatures at any given depth
are reached after 3 minutes using the IR heater, compared to 5 minutes using the heat gun.
To investigate the possibility of using shorter heating times, the slabs were re-wetted and the
tests carded out using 3 minutes of heating with the IR heater. Results, shown in Table 3-9,
were again very favorable, indicating that a 3-minute heating time could be used. Further
tests at shorter heating times were not as promising; therefore, a 3-minute heating time using
the IR heater 4 in. (100 mm) above the test area is recommended in cases where testing is to
be carded out on a wet surface.
Table 3-9. Effects of surface drying using IR heater after 3 minutes of heating.
Initial Flow Flow After Flow After:

Air Dry _ Soaking b Heat Drying*
Slab w/e (ml/mlnute) (ml/minute) (ml/minute)
A-2 0.38 45L50 30/35 45/50
B-2 0.43 60/70 45/50 65/75
D-2 0.60 130/140 70/80 1201130
aStored in air at 73 ___3°F (23 + 1.7°C) and 50 + 5-percent RH for 10 months after curing.
CDried 3 minutes using IR heater 4 in. (100 ram) above surface.
While the surface-heating technique is workable, it does dramatically increase the time
needed to carry out a single test. It was recognized that when testing is being carried out by
local forces, it is often possible to schedule around periods of inclement weather, thus
providing a window of drying time prior to initiation of testing. To investigate the minimum
"natural" drying times needed subsequent to cessation of a wet period, a series of
moist-cycled slabs were selected that had previously been placed at 100*F (38 °) under forced
air flow for a period of 2 weeks, then stored at 73 + 3°F (23 + 1.7°C) and 50 +__5-percent
RH for a period of approximately 8 months. SAF tests were carded out on these slabs, and
the slabs then re-introduced into the moist cycle for a period of 6 weeks. The slabs were
then subjected to 24 hours of drying at 75°F (24°C) with an RH of 25 percent. A low-speed
fan was used to gently circulate air over the specimens during the drying period.
29
Results of SAF tests carried out immediately prior to moist cycling (initially dry condition)
and following the 24-hour dry period after moist cycling are shown in Table 3-10.
Table 3-10. Effects of "natural" drying on SAF.
Flow After Drying for 24

Initial Flow Prior to Moist Hours Subsequent to Moist
Slab w/e Cycling (ml/minute) Cycling (mi/rnlnute)
Mean* Max Min Mean* Max Min
B-4 0.43 102 150 65 100 155 60
D-4 0.60 142 195 115 130 200 105
*Average of 15 readings across surface of slab.
Results indicated that flows could be restored essentially to the original air-dried values
through gentle drying at 750F (240C) for 24 hours. It is recognized that the indoor relative
humidity where these tests were carried out was relatively low and that this would tend to
accelerate drying as compared to conditions normally encountered in temperate climates
during summer months, where average relative daytime humidity would typically range from
50 to 70 percent. However, the indoor tests did not take into account the radiant effects of
direct sunlight, which can raise surface temperatures by significant amounts and thereby
greatly increase rapidity of drying. Since all potential field conditions cannot possibly be
investigated, our recommendations are that SAY tests can be carried out under temperate
conditions (i.e., late spring to early fall) provided the concrete is exposed to drying
conditions for at least 24 and preferably 48 hours subsequent to any rainfall that wets the
concrete surface to a saturated condition.
Correlation Studies
Specimen Preparation
In order to develop correlations between SAY and other methods, companion test specimens
were prepared. The first set of plain air-dried slabs was used to prepare these specimens.
30
Each slab was dry-cut into two 12- x 12- x 6-in. (305- x 305- x 15-mm) sections. One of
each of these half-slab sections was used to obtain a series of 4-in. (102-mm) diameter
full-depth cores. On the second section, a 0.5-in (12-mm) wide foamed polystyrene dike was
affixed to the top (f'mished) surface using silicone caulk. These specimens were then
subjected to the test procedures described below.
Test Procedures
A 15-percent solution of sodium chloride was prepared and poured into the diked section of
each half-slab. The slabs were placed in an environment maintained at 70 ___3°F (23 +
1.7°C) for a period of 90 days. The specimens were covered with polyethylene sheeting to
reduce evaporation of the solution, which was brought to its initial level at weekly intervals.
After 90 days of ponding, the solution was removed from the slabs, a brief rinse of tap water
was used to remove residual salt from the surface, and the slabs were allowed to air-dry.
Two 1.125-in. (28-mm) diameter holes were then drilled into the surface of each slab.
Powder was removed at the following depths: 0 to 0.375 in. (0 to 9 mm), 0.375 to 0.625
in. (9 to 16 mm), 0.625 to 0.875 in. (16 to 22 mm), 0.875 to 1.125 in. (22 to 28 mm), and
1.125 to 1.625 in. (28 to 41 mm). Samples were then analyzed for total chloride ion using
procedures described in AASI-ITO Specification T 260 (2). Results were then averaged for
each set of duplicate samples. Data were then transmitted to Professor Richard Weyers, who
has developed (16) an analytical procedure for obtaining chloride diffusion coefficients (De)
by fitting the experimental data to Fick's law of diffusion. Results are presented in
Table 3-11.
Table 3-11. Chloride diffusion constants (D_) for test slabs.
Slab :w/e Ratio _ Admixture : :i!Dc:(in.:/yr)

A •0.38 ____ : :i: :0.93
• ..:.... '.
B 0,43 --" .....

i .
2.27
.
..::: '. :: ••3.01

D ' 0.60 . =". . i.. . • " " " : i:::: .i
E 0:36 Silica :::fume : 0.19
:F 0.2.7 Latex : : :: 0.18
"No admixtures other than air-entraining agents used.

Note: 1 in.2/yr = 4.88 x 107 mm2/seeond.
31
Three of the 4-in (102-mm) cores from each of the second half-slabs for each concrete were
sliced in order to remove the top 2-in. (50-ram) portion. These 4-in. (102-mm) diameter x
2-in (50-ram) thick slices were placed in a forced draft oven maintained at 140 +_ 5 OF(60 +__
2.8°C) and dried to constant weight. Drying time was approximately 20 days. Each slice
was then wrapped carefully in moisture-resistant packaging and shipped to Core
Laboratories, Inc. in Dallas, Texas. Slices were then tested for air permeability using a
pulse decay permeameter. Details of the theory, test equipment, and procedures are given by
Bourbie and Walls (27). With this equipment, permeability can be measured from 1
millidarcy (10-3_m2) to 10 nanodarcy (lO s/zm2) with an accuracy of + 5 percent. Results
are given in Table 3-12.
After air permeability testing had been completed, the cores were packaged and shipped to
the University of Toronto for water permeability testing. The cores were cut to a length of
1.5 in. (40 ram) so they could be accommodated into the test fixture. They were then
vacuum saturated with water just prior to testing. Testing was carried out using a
pulse-decay water permeameter developed at the university. Results are presented in
Table 3-13.
Comparisons of Results
Relationships between SAF readings, chloride diffusion constants, and air and water
permeab'tlities axe shown in Figure 3-4. Correlations are good and demonstrate that SAF can
be used as a qualitative measure of chloride diffusion coefficients or permeabilities of the
surface layer of concrete. One should be aware that the relationships shown in Figure 3-4
are not meant to imply that SAF testing can be used as a substitute for actual permeability
measurements. If actual permeability values axe desired, then the corresponding physical
tests must be carried out.
32
Table 3-12. Air permeability as determined from pulse-decay
measurements.
.Air Permeability
Slab w/e :Ratio Admixture (mierodareies)
:A 0..38 ---: 134
B O.43 --_ 147
D 0.60 .-_ . ... " .286.
E 0.36 Silica fume . • 11
F 0..27 Latex I
"No admixtures other than air--entraining agents used.

Note: 1 mierodarey = 10_/_m 2.
Table 3-13. Water permeability as determined from pulse-decay

m_urements.
"_¢ater Permeability
Slab wle Ratio Admixture ..(pm/sec)
A 0.38 _ . 3.00.
.. . :
B 0.43 --_ "" 7.30 :
•---_ 6A0
•...D • 0.60.:. :..... • ...) . . .. .. .
E. 0.36 : .Silicafume .... :::: 0.75.:
. ... . • .
:F _0.27 • • ...Latex : '. 0.36
*No admixtures other than air-entraining agents used.
33
,_. 4
.,=_
00 "
•_ _; ur e = FI
o_ o t
300
200
o
¢J
EI=- 0f " "' !

¢D
O. 0 20 40 60 80 100 120 140
;_ Surface Air Flow--ml/minute
i 'F o 7
._ 6 ' 0
E
o =o 4o 6o ao loo 72014o
Surface Air Flow--ml/minute
Figure 3-4. Relationship between SAF, chloride diffusion coefficiems, and air and water
permeabilities.
34
4
Design and Construction of a Field Device
Design Criteria
In order to obtain a workable field technique, it was necessary to package the components
into a fieldworthy prototype instrument. This field prototype required certain characteristics
that would not necessarily have been part of the laboratory bench version of the device. The
characteristics believed to be most important for successful development of a field device
included:
• Ruggedness. The instrument must be water resistant, resistant to accidental impact,

and capable of transport in vehicles typical of those used on construction sites.
• Portability. The instrument should be able to be carried by one person, with total
weight preferably less than 15 lb (33 kg), and capable of operation in horizontal,
vertical, and overhead modes.
• Power Sources. The instrument should be capable of operation from on-board

power sources. A rechargeable battery source capable of powering the device for
one working day and able to recharge overnight was deemed highly desirable.
• Rapidity of Test. While total test time for the laboratory device was less than
1 minute, the sequence of opening and closing of valves during extensive field
testing might become quite tedious. Automation of these functions was deemed to
be an important feature.
35
• Safety of Test. The device should not use any dangerous voltages, hazardous
chemicals, or unshielded moving parts. The basic principles of the SAF method
made for an inherently safe procedure, and no problems were expected in this
regard.
Construction of Field Prototype Device
The aforementioned criteria and a general description of the principles and desirable
configuration of a field device for carrying out the SAF test procedure on concrete structures
were submitted to a number of firms having expertise in development of test equipment for
construction applications. Based on a review of designs submitted and other selection
criteria, Texas Research Institute (TRI), Austin, Texas was selected to construct the device.
Construction took place over a period of 18 months, during which first-generation
instruments were tested and modified so as to meet the intended goals. In order to obtain the
desired characteristics, some sacrifices in the portability of the device had to be made.
The final weight and configuration of the unit made it impractical for one-man operations on
vertical or overhead surfaces. For this reason, an auxiliary lightweight vacuum plate was
developed, which was capable of remote operation from the unit using a connecting vacuum
hose.
Descriptions of overall features, major internal components, overall electronic circuitry, and
operations of the device are given in this chapter and in the appendices to this report.
Further details, including detailed circuit diagrams and a troubleshooting guide, are given in
the operations manual supplied with the instrument.
Overall Features
Figure 4-1 is an outline drawing showing the dimensions of the device and major external
features. There are two folding handles on the device to hold it in a vertical position while
testing horizontal surfaces such as bridge decks. These two handles also contain switches
used to operate the unit. The right handle contains two miniature switches. One switch
activates a solenoid valve, which directs vacuum to the plate at the foot of the device. The
other switch is strictly electronic and allows the operator to "hold" the readings, which are
36
37
displayed on the panel meters at completion of the test. Both handles fold out of the way for
compact storage. The vacuum plate is mounted so that it may swivel or rotate in relation to
the body. This enables the operator to hold the body of the device in a convenient position
while testing and still maintain a vacuum seal at the foot of the device. The three centering
springs return the foot to its original position upon removal from the test location. Two foot
pads axe also threaded into the vacuum plate. The operator can thereby apply pressure to
effect a better seal by standing on the foot pads while testing on horizontal (top) surfaces. A
foam rubber gasket with a 4-in (100-mm) outside diameter and a 2.75-in (70-mm) inside
diameter is placed under the foot immediately prior to testing to effect a vacuum seal. Total
weight of the unit as shown is 33 lb (15 kg). A photograph of the entire device is shown in
Figure 4-2. The valve on the side of the unit allows the vacuum to be directed to the remote
plate when testing vertical or overhead surfaces.
Figure 4-3 shows the remote vacuum head used for testing vertical and overhead surfaces.
The external plate is connected to the main unit by a 0.219-in. (6-mm) inner diameter, 10-ft
(3-m) long vacuum hose. The hose is connected to the plate by means of a quick-disconnect
fitting. A toggle valve allows the operator to apply vacuum to the plate after the hose is
pumped down. A similar fitting is mounted on the main unit. On the face of the plate is a
groove into which a closed-cell foam rubber gasket can be inserted. Pressure is then applied
to seal the gasket onto the test surface by pushing on the handles of the plate. The main unit
is placed on a tripod support (Figure 4-4) when the remote head is being used.
Internal Components and Electronic Circuitry
The major internal components of the SAF device consist of:
• flowmeter;
• vacuum pump;
• solenoid valve;
• vacuum transducer;
• battery pack;
• elapsed time indicator; and
• digital panel meters.
38
Figure 4-2. Overall view of SAF device configured for use on horizontal surfaces.
39
Figure 4-3. Remote vacuum head for testing vertical and overhead surfaces.
40
....:...: _, ...:.
Figure 4-4. SAF device positioned on tripod support for testing of vertical surfaces.
41
The flowmeter is of the thermal mass flow type. This type of flowmeter is not affected by
ambient conditions, can be mounted in any orientation, is small, lightweight, and highly
accurate, and provides digital output. The vacuum pump is a small, lightweight, diaphragm-
type pump, which operates on 12 V DC power.
The solenoid valve is a three-way valve that directs the vacuum to the directional flow valve,
where the operator can select the mode of operation. The vacuum transducer verifies that
the pump is working properly and that the flow readings are taken at the proper vacuum.
The battery pack consists of ten 1.2 V Ni-Cad batteries in series to provide the required
12 V, 7 Ah power source. This allows about 4 hours of continuous operation under field
conditions. The recharge time is 15 to 18 hours.
The elapsed time indicator keeps track of the time that the pump runs during operation of the
SAF. The digital panel meters display the flow rate and vacuum values.
Additional details on the internal components and on the electronic circuitry for the SAF are
presented in Appendix A.
Operation of the Field SAF Device
The SAF device is designed to permit taking permeability readings in all orientations--
horizontal surfaces (topside and overhead) as well as vertical surfaces. For overhead and
vertical surfaces, the external vacuum head is used, connected to the SAF device by means
of a vacuum hose. Step-by-step operational procedures for the use of the device in the field
are presented in Appendix B.
42
5
Preliminary Testing of Field Device
Comparisons with Bench Prototype
After assembly of the field device at TRI, a series of comparison tests was undertaken on
slabs previously tested for SAF using the Construction Technologies Laboratories (CTL)
bench prototype. These slabs were cast from mixtures similar to those used for the
developmental studies and were in the form of 12- x 12- x 3-in. (305- x 305- x 75-mm) thick
slabs. Concrete mixes used to prepare the specimens are given in Table 5-1.
Concrete Mixture E was modified from the previous design to reflect very high strength (and
low permeability) concretes being developed under a separate research program (28).
Mixture G was added to the study to be representative of very high w/c ratio concrete, which
would be expected to have a high permeability and, in effect, establish the upper limit of
permeability that would be evaluated by the technique.
Slabs from Mixtures A, D, and G were maintained under wet burlap/polyethylene overnight,
then transferred to heavy plastic bags for 28 days of curing. Mixture E was fully moist
cured for 28 days to ensure a high level of impermeability. Mixture F was exposed to air
drying immediately after demolding, in order to develop maximum curing of the latex
admixture.
43
Table 5-1. Characteristics of concrete mixes used for comparison studies.
Quantities-lb/yd 3
Water/
Cement Slump Air
Mix Cement Sand Gravel Water Ratio Admixture (in) (%)
A 659 1263 1772 252 0.38 ---_ 2.3 5.0
D 374 1563 I720 224 0.60 --_ 2.4 6.3
Silica fume-
E 800 1000 1800 254 0.23 125 lb 9.2 1.4
F 685 1579 1311 185 0.27 Latex 25 gal 6.3 5.0
G 313 1579 1765 234 0.75 ---_ 1.6 5.0
*No admixtures other than air-entralning agents used.

Note: 1 lb/yd 3 = 0.59 kg/m3; 1 in. = 25.4 ram; 1 lb = 0.454 kg; 1 gal = 3.785 L.
After the conclusion of initial curing, the slabs were coated on the four edges with moisture
barrier grade epoxy and placed in an environment maintained at 73 4- 3°F (23 4- 1.7°C) and
50 _ 5-percent RH for a period of 90 days. The slabs were then tested for SAF at nine
positions on the face of each slab on both top (finished) and bottom (cast) surfaces using the
bench prototype. The top surface of the slab cast from Mixture F could not be tested due to
a very rough finish. After testing, the slabs were carefully packaged and shipped to TRI for
evaluation of the field prototype. Tests were carried out approximately 4 weeks after testing
using the bench prototype. Since the slabs had been in essentially an air-dry condition for
many weeks prior to the first tests, one can assume that the slabs were in essentially the
same state for both of the test series. Tests at TRI were carried out in an interior room
maintained at 70 to 75°F (21 to 24°C). Comparative data are included in Table 5-2.
Immediately apparent are the large differences in results between the two procedures on
identical slabs. There was no reason to believe, as noted above, that the permeabilities of
the slabs had changed significantly between the two sets of tests. After elimination of the
effects of other variables, the differences were traced to the basic operating principles of the
flowmeters used. The flowmeter incorporated into the field device is an electronic mass
flowmeter, which measures absolute mass flow. The flowmeter used in the bench version is
a rotameter, which measures volumetric flow. The electronic flowmeter is calibrated to read
in volumetric units under standard conditions and should be equivalent to the rotameter when
44
Table 5-2. Comparison of SAF tc.adings for bench prototype
and field unit.
SAF (ml/mlnute)-Top Surfaces
B_neh Field
Slab Mean Std. Dev. Mean Std. Dev.
A2 48.1 4.6 I 1.5 3.1
D3 143.0 10.6 35.6 3.2
E5 13.4 1.9 0.5 0.2
F2 --_ .--_ --_ --"
G1 142.0 34.1 26.5 2.2
SAF (mi/minute)-Bottom Surfaces
Bench Field
Slab Mean Std. Dov. Mean Std. Dev.
A2 62.0 6.0 11.9 1.3
D3 53.0 16.1 11:6 4.7
E5 8.5 2.3 0.5 0.2
F2 20.6 3.9 0.4 0.2
GI 154.0 22.6 31.8 3.5
"No reading.
both are used under standard conditions. However, all developmental testing using the
rotameter had been done with the meter between the vacuum pump and the concrete test
specimens. Therefore, the rotameter was operating in a partial vacuum, and not under
standard conditions. However, as the test results are relative in any case, and an absolute
measure of flow or permeability is not required, it was only necessary to establish that a
linear relationship existed between results obtained on the same specimens using the bench
and field versions of the methods. Such a relationship is shown in Figure 5-1. The
45
40-
•-_ 30
G)
CD 0 0
.__
I,/.
2O
_c
E
oo
E 10
I I I
0 50 100 150 200
SAF (ml/minute), Bench Prototype
Figure 5-1. Comparison between SAF results obtained using laboratory bench prototype and
field device.
46
correlation coefficient for the regression of the bench values on the field values was
determined to be 0.963, indicating a fairly linear relationship between the two data sets. The
equation of the line allowed the previous categories of SAF developed using the bench
prototype to be transformed into a new set of recommendations, as follows: readings less
than 5 ml/minute represent low permeability concretes, readings between 5 and 16 ml/minute
represent moderately permeable concretes, and readings higher than 16 ml/minute represent
highly permeable concretes.
Temperature Effects
After comparison testing had been completed at TRI, the unit was shipped to CTL for further
evaluation. In order to be viable as a field device, it was deemed necessary to determine the
effects of temperature, ff any, on the response of the unit. For this purpose, large
environmental rooms available at CTL were used. Rooms were brought to temperatures of
either 38°F (3°C) or 98°F (37°C) prior to carrying out the evaluations. Slabs cast earlier in
the investigation were used for these measurements. A high w/c ratio (0.6) slab from
Mixture D and an LMC slab from Mixture F were used to represent a wide range of
permeabilities. Slabs had been stored in the laboratory at 73 +_ 3°F (23 +__1.7°C) and 50 -4-
5-percent RH for a period of approximately 22 months prior to these evaluations. Tests were
first carried out at standard conditions of 73 + 3°F (23 _ 1.7°C) at 15 positions on the cast
face of each slab. After these test were completed, the slabs were moved to a room
maintained at 98°F (37°C) for a period of 7 hours, and tests were repeated atthis
temperature. The SAF device was allowed to come to the temperature of the room for a
period of 30 minutes prior to testing. After the high temperature tests were completed, the
slabs were moved to a room maintained at 38°F (3°C) and allowed to equilibrate overnight.
The unit was moved into the room in the morning for a period of 30 minutes, then tests were
conducted.
Results are presented in Table 5-3. For these slabs, readings can be seen to be essentially
independent of temperature over the range of investigation. The higher coefficients of
variation associated with Slab F are reflective of the very low mean air flows encountered,
and have no detrimental effect on interpretation of results. Performance of the instrument
was excellent at all temperatures, and no problems were encountered.
47
Table 5-3. Effects of temperature on readings taken with SAF field unit.
' s_. (mlt,_.u_)

Temp.
Slab No. Readings (°F) Mean Minutes Max. Std. Dev. % C.V.
D-2 15 38 16.2 t3.7 18.9 1.3 8.1
I5 73 15.3 t2.7 17.8 1.5 10.1
15 98 15.2 I2.5 18.0 1.5 10.1
I_-2 t5 38 1.3 0.8 1.9 0.3 24.2
I5 73 i.1 0.8 1.3 0.2 17.7
15 98 1.0 0.7 1.3 0.2 17.0
Note: °F= 1.8x°C + 32
Testing of Outdoor Slabs
A series of outdoor test slabs cast by the Portland Cement Association 0aCA) in 1979 were
used for a preliminary evaluation of the unit on field-placed concrete. The slabs were
originally designed to evaluate the effects of such parameters as w/c ratio, slump, air
content, finishing, curing, and consolidation on deicer scaling resistance, and have been
subsequently used to study the process of carbonation in concrete (29). Three slabs with a
range of w/c ratios, air contents, and consolidation were selected. Three SAF tests were
carded out on each slab, and 4-in. (100-ram) diameter cores were extracted from the test
positions immediately after the test. The cores were then sliced to 2-in. (50-ram) thickness
and dried to constant weight. The slices were then tested for air permeability using the
pulse-decay technique (27). A comparison of results is shown in Table 5-4.
The tests appear to reflect the characteristics of the mixes used to cast the slabs. Lowest
average results were obtained on Slab 1, which exhibited the lowest slumps and air contents.
Slab 9 was intermediate, exhibiting somewhat higher slump, perhaps indicating a somewhat
higher water content in spite of the reported equality in design w/c with Slab 1. Slab 11
exhibited the highest permeabilities, reflecting the higher w/c and slumps for this mix.
48
Table 5-4. Results of comparative tests on field slabs.
Air Permeability
Slab Description Test Location SAF (ml/mlnute) (mierodarcies)
1 wle= 0.43 A 6.0 45.2
2.6-in. slump B 3.3 19.1
2.8% air C 6.3 13.8
No vibration
9 w/e = 0.43 A 8.8 57.9
3.8-in. slump B 4.8 38.1
2.8% air C 7.5 53.2
Internal vibration
11 wte = 0.50 A 13.3 91.9
5.3-in. slump B 13.5 118.6

5.5% air C I0_6 85.2
Internal vibration
Note: 1 in. = 25.4 ram; 1 microdarey = 10_s pm 2.
The relationship between SAF readings and actual air permeabilities for these cores is shown
in Figure 5-2. There is a very strong relationship between the two test procedures. These
data indicated that the SAF method could be used on actual field-cast slabs and that test
results could be viewed as a good indicator of the actual permeability of the test areas.
Based on these promising preliminary test results, field tests on actual structures were carried
out. The field trials axe described in the next chapter of this report.
49
120 - o
100 "
a,1
"_ O
&..
"_
O 80
E
I
>_ 60 O o
*_
,.O
_E 40
a.
oL.
< 2O
O ' I ' • • . . •
0 2 4 6 8 10 12 14
Surface Air Permeability-ml/minute
Figure 5-2 Relationship between SAF readings and actual air permeabilities for cores taken
from PCA oudoor slabs.
50
6
Field Trials
A series of field trials was carried out to evaluate the equipment under actual test conditions.
The primary purpose of these field trials was to "shakedown" the device under a variety of
conditions, make modifications (if necessary), and thereby refine the method, making it more
practical and reliable. Due to the absence of convenient and standardized field techniques for
assessing in situ permeability, it was not possible to make an absolute assessment of the
accuracy of the procedure in the field. However, by extraction of cores from test locations
and subsequent testing of these cores using standard laboratory permeability procedures, it
was possible to gain some indications of the ability of the device to determine relative
rankings of the permeability of field concretes. Field trials were carried out in environments
representative of temperate, marine, and deicer exposures.
Field Trial No. 1-Austin, Texas
The first set of field trials was carried out on March 5 and 6, 1991, in Austin, Texas. The
weather was clear and hot, with high temperatures near 95°F (35°C). The site chosen for
the testing was a bridge on Route 1 (also known as Mopac highway) over Westover Road in
central Austin. Testing was carried out on the 10-ft (3-m) shoulder area of the southbound
passing lanes and on three of the prestressed concrete support beams in positions close to the
bridge abutment. Operation of the device on the bridge deck is shown in Figure 6-1. Test
results over a 36-ft (ll-m) length of shoulder are given in Table 6-1.
51
_•ii_••
! •••••••_,•i!_i
••_ : i
•I;•!:
,,,
•
:•••••i
• •_•••_
,,:, ,:
_i_ _
Figure 6-1. SAF device in operation on Westover Road Bridge, on Route 1, Austin, Texas.
52
Table 6-1. SAF readings taken on deck of bridge over Westover Road,
Austin, Texas.
. : SAF (nil/minute)
•. Offset(ft)
Station (ft) 0 3 6 9
:3 21.2 _ 41.3 50.4 82.1
6 56.8 :77.9 --_ 40.1
9 35.3 40.3 __b 44.8
12 36.1 53.4 ___b 10.0 •
15 32.3 57.2 79.7 59.8
18 47.3 __b __ 44.1
2t 18.3 :._b 37.2 20.1 _
24 17.2 32.8 41.1 16.9 _
27 23.1 38.2 ___b 30.1
30 26.4 25.8 --) 32.1
33 26.2 36.9 ___b 34.8
36 26.8 24:0 _ ---_ 30.3
39 35.2 42.1 _b :29.3
"Cores removed for subsequent laboratory testing.

bNo reading.
Note: 1 ft = 0.305 m.
Many of the results showed much higher values for flow than previously encountered in the
laboratory phase, even for concretes having very high permeabilities. Much of the surface of
the deck was very rough and uneven on a localized basis, which may have led to poor
sealing and leaks around the gaskets. Additionally, it was noted that the rough surface
caused premature wear of the gaskets. When the original gasket was replaced, much lower
test readings were obtained. The need to remove traffic control, however, prevented further
examination of the effects of gasket replacement on the results. Cores were taken from five
locations for measurement of air permeability using the pulse decay technique.
53
The following day, tests were conducted on three of the prestressed support beams on the
same bridge. Tests were confined to a 6-ft (3-m) section using the beams near the bridge
abutment so as to allow for access to test locations from the slope without the use of special
equipment. Test were carried out on both the webs and lower flanges of the beams.
Operation using the remote vacuum plate is shown in Figure 6-2. Results axe given in
Table 6-2.
Table 6-2. SAF readings taken on prestressed support beams of

bridge over Westover Road, Austin, Texas.
" Distance from End ofBeam

(at abutment) (ft)
Position •on Beam 2 4 6
Outside Beam,East Face

l
Web 19.2 12.1 t8_8
'..
"Flange . t4.1 8.0 _ 7.7
Outside Beam--West Face
Web 15.9 I8.31 17,2
Flange • . . 11.3 7.0 7.9

Interior Beam 1-2East Face
Web : 22.1 24.7 16.2"

.i
" ..i. • '. • ...
Flange i :: _: ;10.3 •••8.3 9.2

.... :. ...
Interior Beam 1-West Face .... :

• . .:. : ".:.
Web " : : 27.2 120.7 .31.1 a
::Flange : :: 10.9 12.2 8.3

• . ..: ..
Interior Beam 2-:East .Face

• " • : ":..i': :: • : " " " "
Web ... :.i• .!:i:::

.. . .19.1 "• 17.8 11.2'
...... . ...: . :.
!Flange: : : .... :7.3 8_1 8.8

Note: 1 ft = 0.305 m.
54
i'.'"
Figure 6-2. Operation of remote vacuum plate on prestressed beams of bridge over
Westover Road, Austin, Texas.
55
Readings were much lower than encountered the previous day on the bridge deck. It is
believed that this is due to the much smoother surface on the beams, which allowed for a
better seal and did not do as much damage to the gaskets. Cores were taken from three
locations for subsequent laboratory testing. It should be noted that at one location on the
web of interior Beam 1, the core was taken completely through the beam, allowing two test
samples to be obtained from the one core. At this point in the testing it was noted that a
much lower vacuum level was being achieved, even against the standard metal plate. This
was indicative of battery drain, and testing was concluded at this point.
Cores were packaged and shipped back to CTL immediately after testing was concluded.
Upon receipt, the top 2 in. (50 mm) of each core were removed using a water-cooled
diamond saw. As noted above, for the core removed from interior Beam 1, two test samples
were obtained, one from each end of the core. After cutting, the samples were placed into a
forced-draft oven maintained at 140 + 5°F (60 + 3°c). The samples were allowed to dry
to constant weight, which was achieved after approximately 3 weeks of drying. They were
then placed into moisture-resistant sample containers and shipped to Core Laboratories,
Dallas, Texas, where they were tested for air permeability using the pulse-decay
permeameter. Results are presented in Table 6-3.
Although a general relationship between in situ and lab tests is demonstrated, as shown in
Figure 6-3, the correlation is not as quantitative as in previous laboratory series.
Additionally, large changes in SAF readings were not always associated with corresponding
changes in laboratory values. Problems encountered during field testing, as noted above,
may have been responsible for some of these discrepancies.
Field Trial No. 2--Corpus Christi, Texas
A second field trial was carried out May 7 and 8, 1991, in Corpus Christi, Texas. Weather
was warm, with a high temperature of about 800F (27°C). Intermittent rainfall was
prevalent during the period. Sites included the deck of the Nueces Bay Causeway Bridge
carrying Rt. 181 North from Corpus Christi and the J.F.K. Intra-Coastal Canal Bridge
carrying Park Road 22 onto North Padre Island. The Nueces Bay Causeway Bridge had a
recently widened deck with a 10-ft (3-m) concrete shoulder, which was available for testing.
The J.F.K. Bridge had an asphalt-covered deck; however, the prestressed concrete support
beams and reinforced concrete piers were available for testing. Results obtained on the
Nueces Causeway deck are presented in Table 6-4.
56
3OO
0
250
rJ
_= 200
"10
[] []
o
In
O O
o_
E 150 I O DeckCores
'_-
'R [] I [] BeamCores
=_ 100
E
O.
•- 50
<C
I I I I
0 10 20 30 40
SAF (ml/minute)
Figure 6-3. Comparison of in situ SAF and laboratory permeability testing for Austin,
Texas, field site.
57
Table 6-3. Results of air permeability tests carried out on cores
removed from bridge over Westover Road, Austin, Texas.
Air
SAF _ Permeability _
Core Loca_on" (ml/minute) (rolcrodarcies)
1-1 Deck (3-0) 21.2 186
1-2 Deck (I2-9) 10.0 181
1-3 Deck (36-3) 24.0 275
1-4 Outside beam-flange 8.0 116
1-5E Interior beam 1-web 16.2 196
1-5W Interior beam 1-web 31.1 199
1-6E Interior beam 2-web 1.1.2 108

I
aLocations on deck refer to station (along length) and offset along

width (in feet).
bin situ measurements using SAF device.
°Laboratory measurements on cores taken from test locations.
Note: 1 ft = 0.305 m; 1 microdarcy = 10_/_m 6.
The readings appeared reasonable, and no problems were encountered with the equipment.
Prior to initiation of this set of field trials, an alternative gasket material, which appeared to
be more durable, had been obtained. In addition, it was decided to replace the gasket after
every fourth reading, regardless of whether the gasket appeared worn or not. The locations
closest to the curb (offset=0), however, exhibited rather high readings. This was attributed
to some rough surfaces in this area, along with considerable surface cracking. After
approximately 2 hours of testing, a heavy thundershower forced cessation of activities for a
period of time. Immediately after the shower, cores were obtained from a number of
locations on the deck that had been tested using the SAF. After coring was completed, it
was necessary to remove traffic control to allow for rush-hour traffic to utilize the entire
deck. No further testing was carried out at this site.
58
Table 6-4. SAF readings obtained on deck of Nueces Causeway Bridge in
Corpus Christi, Texas.
•SAF (ml/minute)
Offset_ft)
0 57.2 I5.0 6.5 8.4
3 37.7 14_2 18.5 25.2
6 40.6 _ I3.8 10.5 17;0
9 35.4 7.9 _ •10.•4 13.3
12 23.0 12.3 7.5 5;lY
15 20.4 8:5 16.6 a 13.5
18 65.2 5.3 4.8 7.7
21 20.5 6.1 5.2 6.0
24 23.4 11.1 9.4 7.1

Note: 1 ft = 0.305 m
The following day, testing was carried out on one pier cap and two support beams on the
J.F.K. Causeway bridge. Access to the pier cap was provided by Texas Department of
Transportation personnel through a scaffolding arrangement. Results for the pier cap tests
are shown in Table 6-5.
The reading at the location 6 ft (1.8 m) in from the edge and 28 in. (635 mm) up from the
bottom of the cap appeared rather high. A core was taken from this location and one other
on the cap for lab verification. Readings were then taken on two prestressed support beams
at the north facing side of the bridge at the mainland side abutment. The beams were
accessed from the abutment slope. Therefore, the greatest distance from the edge that could
be reached was approximately 8 ft (2.4 m). Results are presented in Table 6-6. All readings
were fairly low, indicative of good quality concrete.
59
Table 6-5. SAF readings taken on pier cap of J.F.K. Intra-eoastal Canal Bridge, Corpus
Clu'isti, Texas.
SAF (ml/minute)
Distance from Edge of Cap (ft)
Height from bottom of cap

(in.) 0 2 4 6 8
I2 ___b 8.9" 15.2 9.5 14.7
28 __b 10.$ 11.2 30.P 12.0
"Core removed for subsequent laboratory testing.

bNo reading.
Note: 1 ft = 0.305 m; 1 in. = 25.4 ram.
These exact same beams had been previously tested by CTL 4 years earlier as part of a
Federal Highway Administration (FHWA) sponsored study on the corrosion of prestressing
steel in prestressed concrete bridges. Results from that study indicated that the concrete had
high quality and low permeability measured via the AASHTO T 277 technique. Chloride
contents were quite low in spite of nearly 20 years of exposure to the gulf marine
environment. The low SAF readings corroborate the previous observations. Cores were
obtained from two positions on the beam. Since the cores were taken through the beam, a
total of four test slices were obtained for subsequent laboratory air permeability testing (one
slice from each end). Slices were shipped back to CTL, dried to constant weight, and
shipped to Core Laboratories for air permeability testing.
Results of air permeability testing on the cores obtained from the Corpus Christi structures
are given in Table 6-7. Also included am companion SAF readings obtained at core
locations.
Again, as with the results for the Austin cores, there is a very general relationship between
SAF reading and air permeability determined on the cores. For very low SAF values, there
is a surprising degree of disparity in the measured permeabilities. SAF readings less than
5ml/minute yield air permeabilities that range from about 50 to 100 microdarcies (5.0 x 104
to 1.0 x 10-4/zm2). The relationship is shown graphically in Figure 6-4. As before, the
results indicate that the SAF readings can be used as an indicator of relative permeability,
but if quantitative data is needed, cores should be obtained and laboratory testing carried out.
60
Table 6-6. SAF readings taken on prestressed beams on J.F.K. Intra-
coastal Canal Bridge.
Distance from End of Beam at Abutment

(ft)
Height from Bottom of Beam
(in.) 2 4 : 6 8
Outside Beam-North Face ....
• . '..
112 4.1 6.1 •:•5.4 6.3
24 3.4 3.3 3.6 a 3.9
Outside Beam-South Face
12 4.2 2.8 •5.4 6.6
24 3.3 _ 3.6 a 3.4
Interior Beam-North Face
12 2.8 3.4 3.5 5.1
24 28 2.7 2.9 _ 3.6
Interior Beam-South Face
12 •3.6 •7.9 3.4 4.3
24 4.6 •4.8 2.3 a 5.1
aCore removed for subsequent laboratory testing.

bNot tested.
Note: 1 ft = 0.305 m; 1 in. = 25.4 ram.
Held Trial No. 3--Centre County, Pennsylvania
The final set of field trials was carried out on two bridge decks located in Centre County,
Pennsylvania. The first bridge was located on Route 322 westbound out of Port Matilda,
Pennsylvania, over Reese Hollow Road with bridge No. BMS 14 0322 0221 1292. The
second was located on Route 220 in Julian, Pennsylvania,over Bald Eagle Creek, with
bridgeNo. BMS 14 0220 0280 1591. All tests were carried out on the deck surfaces.
Weatherduring the testing was cool and overcast, with high temperaturesnear 75°F (24°C).
61
150 -
125
=a o o
=m
u
==
"0 100 []
o
,.- [] []
0
E 75 [] O [] []O Beam Cores

Deck Cores I
"_ o
® 50
E []
D.
._ 25
I I | I I
0 10 20 30 40 50
SAF (ml/minute)
Figure 6-4. Comparison of in situ SAF and laboratory permeability testing for Corpus
Christi, Texas, field sites.
62
removed from bridges at Corpus Christi, Texas.
Air
SAF Permeability d
Core Location _'b (mI/rrfinute)c (mierodarcies)
2-1 :Deck (6-0) 40.6 117
2-2 Deck (9-3) 7:9 58
2-3 Deck (12-9) 5.0 62
2-4 Deck (15-6) 16.6 116
3-1 Pier Cap (6-28) 30.1 • 87
3-2 Pier Cap (2-12) 8.9 78
3-3 Outside Beam-North Face 3.6 90
3-4 Outside Beam-South Face 3.6 99
3-5 Interior Beam--North Face 2.9 65
3-6 Interior Beam-South Face 2.3 46
•Locations on deck refer to station (along length) and offset (along

width), in feet.
bLoeations on cap refer to distance from edge (in feet) and height from
bottom (in inches).
tin situ measurements using SAF device.
dLaboratory measurements on cores taken from test locations.
Note: 1 ft = 0.305 m; 1 in. = 25.4 ram; 1 microdarcy = 10_/_m 2.
Testing on the Route 322 bridge deck was carded out on June 18, 1991. Initially, traffic
control was set up to test along a 10-ft (3-m) shoulder on the westbound lane. However,
after testing commenced, it was noted that flow readings were extremely high (over
80 ml/minute). After a brief wetting of a test area, an extensive system of microcracks in
the top surface layer was detected. It was observed that the driving lane of the deck had
been worn to the point that the paste-rich surface layer was no longer present, and the
microcracking was much less extensive. The test area, therefore, was moved over to the
driving lane in order to obtain more representative readings. Results are given in Table 6-8.
63
Table 6-8. SAF readings taken on deck of bridge on Route 322, Port
Matilda, Pennsylvania.
SAF (ml/minute)
Offsetfit)
0 9.6 22.0 10.1 2.9
3 31.3 10.8 5.3 8.0
6 13.1 9.7 12.2 32.6
9 22.2 9.7 4.0 9.3
12 38.1 13.6 a 9.5 9.3
15 23.3 11.7 30.0 14.0
18 33.9 3.5 _ 6.6 26.9
21 26.6 6.7 3.2 9.4
24 13.1 14.1 3.7 26.P
27 13.9 7.(Y 4.2 29.1
30 10.9 8.7 16.7 22.2
33 56.0 9.8 11.7 ___b
36 14.5 9.3 10.8 6.6
39 18.6 28.0 16.8 4.6
42 12.4 7.9 12.6 __b
45 21.3 8.8 ___b 23.1
48 60.2 4.2 22.2 16.8

I I
_Cores removed for subsequent laboratory testing.

bNoreading.
Note: 1 ft = 0.305 m.
64
A considerable range of values was obtained. Some of the higher readings, for instance at
locations (33-0) and (48-0), appeared to be associated with surface cracks at these test
locations. A number of locations could not be tested due to surface cracking or voids in the
surface at these places. All testing proceeded well for a period of about 2 hour, at which
point a brief shower wet the deck and caused an interruption of testing. Since drying was
slow, and more showers appeared imminent, testing was suspended, and cores were obtained
for subsequent laboratory testing.
Testing on the Route 220 bridge deck was carried out the following day. Tests were carded
out in the curb area of the southbound lanes, since the driving lanes had been overlaid with
asphaltic concrete some years ago. For this narrow curb area the test pattern was reduced to
2 ft (0.6 m) along the length of the deck and 6 and 30 in. (152 and 762 mm) across the
width of the curb. Results of the testing are shown in Table 6-9. On the whole, readings
were higher than those at the Route 322 bridge deck. Four cores were selected to cover a
range of values. Although some values higher than 30 ml/minute were encountered, these
were, for the most part, associated with rough or cracked areas and therefore not chosen for
comparison testing. After testing was completed, cores were shipped to Core Laboratories
for air permeability testing.
Results of air permeability testing on the cores obtained from the Centre County structures
are given in Table 6-10. Also included are companion SAF readings obtained at core
locations.
Again, as in the results for the other sites, there is a general relationship between SAF
readings and air permeability (see Figure 6-5). In fact, correlation for the Centre County
bridges is significantly better than for the first two field sites. This may be attributable to
improvements in gaskets and other aspects of the instrumentation prior to testing at the f'mal
site. Deck No. 4 is the Port Matilda Bridge and Deck No. 5 is the Julian Bridge in
Figure 6-5.
65
Table 6-9. SAF readings taken on deck of bridge on
Route 220, Julian, Pennsylvania.
SA.F (mllminute)
Offset (in.)
Station (ft) 6 30
0 39.0 15.2
2 32.7 32.1
4 8.5 29.5
6 37.3 37.9
8 46.2 17.4
10 21.5 30.8
12 21.0 23.5
14 28.0 53.0
16 8.4 _ 16.6
18 18.9 29.4
20 9.1 26.0
22 14.5 a 19.4
24 8.9 13.1
26 29.2 a 34.0
28 23.8 51.4
30 19.5 62.3
32 7.6 _ 49.2
34 10.5 31.5
36 10.8 19.1
38 15.0 22.0
40 33.2 .... b
*Cores removed for subsequent laboratory testing.

bNo reading.
Note: 1 ft = 0.305 m; 1 m. = 25.4 ram.
66
removed from bridges in Centre County, Pennsylvania.
Air
SAF _ Permeability a
Core Location a.b (ml/minute) (microdarcies)
4-1 Deck (27-3) 7.0 52
4-2 Deck (18-3) 3.5 71
4-3 Deck (12-3) 13.6 113
4-4 Deck (24-9) 26.1 233
5-1 Curb (16-6) 8.4 109
5-2 Curb (22_6) 14.5 159
5-3 Curb (26-6) 29.2 204
5-4 Curb (32-6) 7.6 109
aLocations on deck refer to station (along length) and offset (along

width), in feet.
bLocations on curb refer to station (along length), in feet, and offset
(along width), in inches.
5_a situ measurements using SAF device.
aLaboratory measurements on cores taken from test locations.
Note: 1 ft = 0.305 m; 1 in. = 25.4 ram; 1 microdarcy = 10_s/Lm2.
67
300
•-" o
o_
o
" 200 []
0
>,
[] Deck No. 5
E 0 Deck No. 4 I
m 100
E
•- o
a. o
e_
O I I |
0 10 20 30
SAF (ml/minute)
Figure @5. Comparison of in situ SAF and laboratory permeability testing for Centre
County, Pennsylvania, field sites.
68
7
Summary and Recommendations
Summary of Test Method
The method developed under this research program can be used to gain an indication of the
relative permeability of concretes both in the laboratory and on actual field structures. The
method is based on measurement of the rate of air flow through a vacuum plate placed on a
concrete surface under a vacuum of approximately 25 in. of mercury (16.6 kPa absolute
pressure). The effective depth of measurement is demonstrated to be approximately 0.5 in.
(12 mm) below the surface. Two versions of the technique, a laboratory prototype and a
portable field device, were developed. The laboratory device uses bench components and a
rotameter for measurement of air flow. Flows lower than about 30 ml/minute are generally
associated with low permeability concretes. Flows exceeding 80 ml/minute are indicative of
high permeability concretes. For calibration of the method, concretes were cast using a
variety of w/c ratios as well as admixtures such as latex and silica fume. Results were found
to correlate well with chloride diffusion constants derived from 90-day ponding tests, as well
as with true air permeabilities measured using a pulse decay technique.
The field device is self-contained within a 33-1b(15-kg) package. The unit is powered by
Ni-Cad re,chargeable batteries and can operate continuously for approximately 4 hours on a
single charge. Measurements on horizontal surfaces are carried out by placing the unit on a
closed-cell foam gasket placed on the desired test location. The operator stands on two
"feet" attached to the unit while running the test. On vertical surfaces, a remote vacuum
plate containing an affixed closed-cell foam gasket is placed against the surface to be tested
and is linked to the unit through a 10-ft (3-m) vacuum hose. Since the flow path and type of
69
flowmeter utilized in the field device differs from that used in the bench version,
interpretation of results is somewhat different. Values less than 5 ml/minute axe associated
with low permeability concretes. Highly permeable concretes generally exhibit flow values
greater than 16 ml/minute when tested using the field device.
A recommended test procedure, in American Society of Testing and Materials (ASTM)

format, has been prepared for the SAF method of evaluating the permeability of concrete. It
is presented in Volume 8, "Procedure Manual," of this report.
Applicability of Test Method
The method has been shown to be applicable under both laboratory and field conditions. It
has shown good correlation with standard techniques when applied to carefully prepared
concrete specimens cast under controlled laboratory conditions. In the laboratory, the
method may be used to gain an indication of the relative permeability of various concrete
mixtures and therefore can be used for such activities as materials development, research
activities, and product screening.
The major advantage of the new method is that it can be used under field conditions on
in-place concrete structures. The equipment is portable and fairly simple to operate. A large
number of readings can be obtained in a short period of time. Potential applications of the
device could include such activities as:
• Determination of relative permeability of concrete during rehabilitation surveys;
• Verification that low permeability has been established for special concrete overlays;
and
• Establishing general quality of concrete in routine construction or in cases where

questions have arisen.
While there are some limitations (see next section), the method has been used on a variety of
structures under differing climatic conditions with overall good success.
70
Limitations of Test Method
Field testing introduces a variety of largely uncontrollable variables. The method requires an
initially dry surface. Therefore the testing cannot be carried out under wet conditions.
Additionally, the technique will be influenced by internal concrete moisture content, even if
the surface is dry. If testing is to be carried out on a surface that has recently been wetted,
it may be necessary to pre-dry the test areas prior to initiation of testing. If this is not
practical, then one should wait at least 24 hours after cessation of rainfall before conducting
the tests.
Surface features apparently have an influence on the test results. However, it was not
possible to study the myriad of concrete surface characteristics within the scope of this
development project. Some general guidelines would include the avoidance of rough or
grooved surfaces or surfaces that are extremely weathered and weak. Additionally, surface
microcracking can lead to anomalously high results, and such surfaces should also be
avoided. In many cases, it is possible to ascertain whether a test result is valid by comparing
the vacuum obtained during the test to that obtained on the standard metal plate. If the two
readings differ by more than 5 mm of Hg (0.6 kPa), then an inappropriate test surface is
indicated.
Other variables, such as the presence of reinforcing steel or ambient temperature, do not
appear to have a significant influence on the test results. It is wise to avoid areas directly
above steel with cover less than 0.5 in. (12 ram). Additionally, while temperature, per se,
does not influence the test, the electronic components are not designed for continuous
operations below 32"F (0°C) or above about 120*F (50"C).
Recommendations
It is recommended that further evaluations of the method be carried out to more precisely
determine the effects of concrete surface characteristics on the test. This could include
surfaces having various degrees of roughness as well as scaled and microcracked surfaces.
Gasket materials that conform more closely and also adhere to the surface being tested,
thereby allowing testing of rougher surfaces, would also be desirable.
71
Appendix A
Internal Components and Electronic

Circuitry
Internal Components
The major internal components consist of a thermal mass flowmeter, a 3-way solenoid valve,
a DC-operated miniature vacuum pump, a vacuum transducer, a Ni-Cad rechargeable battery
pack, an elapsed time indicator, and two digital panel meters. Also included are associated
electronics for powering the instrument and scaling the output signals from the transducer
and flowmeter. All of the components are mounted on a module that can be easily installed
or removed from the main housing (see Figure A-l). A brief description of each major
component follows.
Flowmeter
A Teledyne-Hastings-Raydist thermal mass flowmeter Model HF -200L was used. This

instrument measures the flow rate of air, in standard cubic centimeters per minute (sccm),
from the concrete. For the purposes of this method, this is equivalent to ml/minute. The
mass flowmeter contains a heated tube instrumented with thermal sensors on each end. As a
gas flows through the tube, it extracts heat and creates a temperature difference along the
length of the tube. This temperature difference is proportional to the time rate of mass flow
and is scaled to a linear output at standard conditions of 0*C and 760 mm Hg (101 kPa)
absolute pressure by electronics within the meter. This type of flowmeter has several
73
74
I_igure A-l, Instrument module removed from housing of SAF device.
advantages over other types of meters (which measure volumetric flow directly). Deviations
from standard conditions will not affect the output of the reading. Additionally, an electrical
analog signal of 0- to 5-V DC can be displayed on a digital panel meter. The flowmeter can
be mounted in any orientation and weighs only 1.8 lb (0.8 kg). Accuracy is _ I-percent
full-scale and repeatability is ___0.2-percent full-scale.
Vacuum/h_mp
A K-N-F Neuberger, Inc. Model UN79 MNI diaphragm vacuum pump was chosen since it
duplicated the pump used in the laboratory development phase of the project, yet is capable
of operation under DC battery power. It is a 12-V DC pump that is capable of pulling down
to 635 mm Hg vacuum (16.6 kPa absolute.pressure) or better at zero throughput. It is
relatively small and fight (2.0 lb [0.9 kg]) and draws 1.5 A.
Solenoid Valve
A Humphrey Products, Mini-Mizer Model 3El three-way, 12-V DC solenoid valve is used.
When energized, the valve directs the vacuum to the directional flow valve, where the
operator can then select the mode of operation (i.e., horizontal or vertical). The flow
reading can then be obtained. When non-energized, the vacuum is vented to the atmosphere
so that the unit may be moved to another location.
Vacuum Transducer
The transducer is a SENSOTEC Model V/6689-01, and it verifies that the pump is working
properly and that flow readings are taken at the proper vacuum. It has good accuracy
(+ 0.25-percent full-scale) and a linear 0- to 5-V DC output that is easily scaled and
displayed.
Battery Pack
Ten "F" size Ni-Cad cells are connected in series to provide a 12-V, 7-Ah battery pack.
Ni-Cad batteries were chosen because of their high energy-to-weight ratio. Considering the
drain of the vacuum pump, solenoid, and associated electronics, this allows for
approximately 4 hours of continuous operation under field conditions. Longer working times
75
could have been achieved by using larger batteries or a remote battery pack, but both of
these were felt to be undesirable in light of the criteria for portability and light weight
discussed above. The batteries must be charged overnight on the day prior to testing. A
recharge time of 15 to 18 hours is recommended.
Elapsed Time Indicator
A SYR_J._C Model 6108-4S elapsed time indicator indicates elapsed time in seconds. It is
activated when the vacuum pump switch is turned on and stops timing when the switch is
turned off. A separate switch resets the indicator to zero. It is powered by an internal
lithium cell.
Digital Panel Meters
Two ACCU'LEX Model DP-2000 digital voltmeters were chosen to display flow and vacuum
values. These are miniature 4-1/2 digit meters with liquid crystal display for viewing in
direct sunlight. They have an accuracy of + 0.06-percent full-scale, a movable decimal
point, and a "hold reading" capability.
Electronic Circuitry
The general electronic component diagram of the device is shown in Figure A-2. The
circuitry may be viewed as two connected modules, the vacuum pump module and the
readout module. These are both housed in the main unit.
The pump module is powered by the rechargeable Ni-Cad battery pack. The battery pack
consists of ten 1.2-V Ni-Cad batteries in series. A total of 12 V are required to power the
vacuum pump and the timer. The pack is located in the instrument module between the
electronic circuitry and the vacuum pump. It is separated from the electronics by a
0.125-in. (3-ram) neoprene gasket. The battery pack may be recharged through the battery
charger jack located on the panel. The pump switch is located on the instrument panel and
activates both the vacuum pump and the timer. With power to the pump, the pressure is
sensed by the in-line vacuum transducer. This pressure is then converted to an electrical
signal, which is then output to the voltage divider and vacuum readout, respectively. At the
same time, the timer clocks the elapsed time of the pump run. The elapsed time of the
pump, while running, is displayed by a digital readout on the panel (see Figure A-3). It is
76
_o II
jl
_=
< _ --
I,
_ :
l <-
_0
-
f °°i
_'.._11 I _,_ _
0
"____ _ -_
_ "/7
Figure A-3. Instrument panel showing digital readout gages for vacuum, flow, and elapsed
time.
78
accurate to within 0.01 seconds and may be reset at any time by flipping the reset switch on
the panel.
The readout module is powered by the output of the vacuum transducer, which drives four
devices: a 5-V regulator for the use of the readout circuit, the solenoid valve, a DC to DC
converter, which powers the flowmeter, and the vacuum pressure readout. Both the
flowmeter and the pressure readings can be captured by pushing the "hold" switch located on
the right handle. The power switch, located on the panel, turns the readout module on and
off. With power on and the vacuum pump running, the vacuum transducer senses pressure
and converts that pressure to an electrical signal. The signal is between 0 and 5 V. The
output is positive for vacuum and negative for pressure above zero gage.
The DC to DC converter provides power to the flowmeter. The converter accepts the
voltage output of the vacuum transducer and creates two low-current voltage references,
+ 15 V and -15 V, required for operation of the flowmeter. With the power from the DC to
DC converter the flowmeter can measure up to 300 ml/minute of air flow. An output signal
of 0 to 5 V can be delivered at approximately 5 ma to the voltage divider.
The voltage divider accepts the outputs of the flowmeter and the vacuum transducer and
transforms their outputs to 0 to 0.3 V and 0 to 0.76 V, respectively. The panel meters
receive the output voltages from the dividers and a 5-V regulator. The regulator accepts the
voltage output of the transducer and produces a low current 5-V reference for the meter
readouts.
79
Appendix B
Operational Procedures for the Use of the

SAF Device in the Field
Horizontal Flat Surfaces
1. Remove instrument from its case and install the two foot pads. The foot pads
should be screwed all the way into the tapped holes on the suction foot base and
then backed out until the aluminum checkered plates are pointed to the top of the
machine.
2. Unfold the two handles by pushing the buttons on the end of the "T" handle lock
pins and removing them. When the handles are horizontal, the lock pins need to
be reinserted through other holes in the handle brackets to lock the handles in the
extended position.
3. Make sure that both switches in the right handle axe in the "off" position (toward
the end of the handle). If the elapsed time indicator is not set to zero, do so at this
time with the SEC. RESET switch. Return the SEC. RF_ET switch to the "off"
position. Ensure that the directional valve is in the "down" position pointing
towards the feet of the instrument.
4. Turn on the POWER switch and observe that the digital panel meters are
activated. Wait 10 minutes for the instrumentation to warm up and stabilize.
There axe three digits to the left-hand side of the decimal point on the vacuum
81
meter and two digits on the left-hand side of the decimal point on the flowmeter.
These should read zero. The digits to the fight-hand side of the decimal point on
both meters are insignificant and should be disregarded. A minus sign also may be
seen to flicker off and on at the left-hand side of the meters. Pay no attention to
this sign. If the flow and vacuum meters seem to be displaying correctly, the next
step is to check operation on a reference plate. This should be done at the
beginning of the working day or whenever readings are questionable.
5. Place a pre-formed closed-cell foam rubber gasket on an impermeable metznic

reference plate. Center the suction foot over the gasket.
6. Stand on the foot pads with the balls of the feet. About one-half body weight
should be placed on the foot pads and the other half should be supported by the
heels. This action will compress the rubber seal against the test spot.
7. Turn on the PUMP with the switch on the panel. You will notice that both the
flow and vacuum gages will display values, and the elapsed time indicator will
start. The vacuum should quickly stabiliTe between 615 and 635 mm Hg, vacuum.
The flow will have a high initial value due to air in the lines, but will start to
stabilize after about 15 to 20 seconds.
8. Next, push the uppermost switch on the right handle toward the main body with
the thumb. This opens the solenoid valve to the suction foot.
9. When the elapsed time meter reads 45 seconds, slide the lower switch on the right
handle toward the main body. This will "hold" the readings on the digital meters.
Record the flow reading at this point. The vacuum should read between 615 and
635 mm Hg, and flow should be less than 1 seem (1 ml/minute).
10. Turn off the vacuum PUMP and the solenoid valve. Reset the hold reading switch
on the right handle and the SEC. RESET switch on the panel. The device is now
ready to be moved to the test spot.
11. Tests on actual concrete surfaces are performed in a manner identical to the initial
check test. In some cases, however, it may take longer than 45 seconds for the
readings to stabilize. Surfaces should be dry, free of dirt or debris, and not
cracked, grooved, or textured.
82
Vertical and Overhead Surfaces
1. Operation of the permeameter on vertical and overhead surfaces is accomplished

with the use of the remote vacuum plate.
2. To support the system during testing, place the main unit in the tripod.
3. Turn the directional valve to the "up" position pointing in the direction of the
panel. This will direct the vacuum to the remote plate.
4. Make sure that both switches in the fight handle are in the "off" and "release"
positions (toward the end of the handle). If the elapsed time indicator is not set to
zero, do so at this time with the SEC. RESET switch. Return the SEC. RESET
switch to the "off" position.
5. Turn on the POWER switch and observe that the digital panel meters are
activated. Wait 10 minutes for the instrumentation to warm up and stabilize.
There are three digits to the left-hand side of the decimal point on the vacuum
meter and two digits on the left-hand side of the decimal point on the flowmeter.
These should read zero. The digits to the fight-hand side of the decimal point on
both meters are insignificant and should be disregarded. A minus sign may also be
seen to flicker off and on at the left-hand side of the meters. Pay no attention to
this sign. If the flow and vacuum meters seem to be displaying correctly, the next
step is to check operation on a reference plate. This should be done at the
beginning of the working day or whenever readings are questionable.
6. Place the remote vacuum plate on the test location by pushing against the handles
with both hands. This will insure a good seal on the foam rubber gasket. Open
the toggle valve by using the thumb to flip the toggle into the "up" position.
7. Turn on the PUMP with the switch on the panel. You will notice that both the
flow and vacuum gages will display values, and the elapsed time indicator will
start. The vacuum should quickly stabilize between 615 and 635 mm I-Ig, vacuum.
The flow will have a high initial value due to air in the lines, but will start to
stabilize after about 15 to 20 seconds.
8. Next, push the uppermost switch on the fight handle toward the main body with
the thumb. This opens the solenoid valve to the remote plate.
83
9. The flow reading will take somewhat longer to stabilize than when the unit is being
used on horizontal surfaces. As a guide, allow at least 1 minute for stabiliTation of
the reading.
10. Next, slide the lower switch on the right handle toward the main body. This will
"hold" the readings on the digital meters. Record the flow reading at this point.
11. Turn off the vacuum PUMP and the solenoid valve. Reset the hold reading switch
and the SEC. RESET switch. The remote head is now ready to be released from
the concrete surface.
12. Remove the remote head by disconnecting the quick-discoimect fitting on the
permeameter. This will release the vacuum and enable the removal of the remote
head.
84
References
1. American Association of State Highway and Transportation Officials. "Standard Method of Test for
Rapid Determination of the Chloride Permeability of Concrete." AASHTO Designation "1"277-83.
Standard Specifications for Transportation - Materials and Methods of Sampling and Testing, Part 11.
Fourteenth edition, August 1986.
2. American Association of State Highway and Transportation Officials. "Standard Method of Sampling
and Testing for Total Chloride Ion in Concrete and Concrete Raw Materials." AASHTO Designation
"1260-84. Standard Specifications for Transportation - Materials and Methods of Sampling and Testing,
Part H. Fourteenth edition, August 1986.
3. D'Arey, H. Les Fontaines Publiques de la Ville de Dijon. 1856.
4. Glanville, W.H. "The Permeability of Portland Cement Concrete." Building Research Technical Paper
No. 3. London: Building Research Station, 1926.
5. Gaerty, L. and R.J. Freeman. "Permeability of Concrete." New Zealand Concrete Construction, pp. 3-6
(March 1986).
6. Whiting, D. "Permeability of Selected Concrete." In ACI Special Publication SP-108 Permeability of

Concrete, edited by D. Whiting and A. Walitt. Detroit, MI: ACI, 1987, pp. 195-222.
7. Mantialay, R.M. Air Permeability of Concrete. Monografia No. 332. Madrid: Instimto Eduardo
Torroja de la Construeeion y del Cemento, Dec.ember 1975.
8. Hanaor, A. and P.J.E. Sullivan. _Factors Affecting Concrete Permeability to Cryogenic Fluids."
Magazine of Concrete Research 35 No. 124 (Sept. 1983): 142-150.
9. Chou Chen, L. and D. Katz. "Diffusion of Methane Through Concrete." J. American Concrete
Institute. Proceedings 75 No. 12 (Dee. 1978): 673-679.
10. Watson, A.J. and C.C. Oyeka. _Oil Permeability of Hardened Cement Pastes and Concrete." Magazine
of Concrete Research 33 No. 15 (June 1981): 85-95.
85
11. Wing, S.P. Discussion of a paper by Wiley, G. and Coulson, D.C., "A Simple Test for Water
Permeability of Concrete." Proceedings of the American Concrete Institute, Vol. 34, (January-February
193811:76-1 to 76-4.
12. Dunagan, W.M. "Methods for Measuring the Passage of Water through Concrete." Proceedings,
ASTM, Vol. 39, pp. 866-880 (1939).
13. Levitt, M. "The ISAT - A Non-Destructive Test for the Durability of Concrete." British Journal of
Non-Destructive Testing 13 No. 4 (July 1971): 106-112.
14. Broglae, R.D. "Design Prediction of the Life for Reinforced Concrete in Marine and Other Chloride
Environments." Durability of Building Materials I (1982): 113-125.
15. Nagaro, H. and T. Naito. *Application of Diffusion Theory to Chloride Penetration into Concrete
Located in Splashing Zones." Transactions of the Japan Concrete Institute 7 (1985): 157.
16. Weyers, R.E. and D.G. Smith. "Chloride Diffusion Constant for Concretes." Proceedings of Structures
Congress "89, Pe_orrnance of Structural Materials. San Francisco: ASCE, May 1989, pp. 106-116.
17. Bisaillon, A. and V.M. Malhotra. "Permeability of Concrete Using a Uniaxial Water-Flow Method."
Permeability of Concrete - ACI SP-82, edited by D. Whiting and A. Walitt. Detroit,/vii: ACI 1988
pp. 175-194.
18. Bourbie, T. and J. Walls. "Pulse Decay Permeability: Analytical Solution and Experimental Test."
Journal of the Society of Petroleum Engineers (December 1982): 719-721.
19. Mobasher, B. and T.M. Mitchell. "Laboratory Experience with the Rapid Chloride Permeability Test."
Permeability of Concrete - ACI - SP-82, edited by D. Whiting and A. Walitt, 1988 pp. 117-144.
20. Figg, J.W. "Methods of Measuring the Air and Water Permeability of Concrete." Magazine of
Concrete Research 25 No. 85 (December 1973): 213-19.
21. Cather, R., J.W. Figg, A.F. Matsen, and T.P. O'Brien. "Improvements to the Figg Method for
Determining the Air Permeability of Concrete." Magazine of Concrete Research 36 No. 129 (December
1984): 241-45.
22. Sehonlin, K. and A. Hilsdorf. "Evaluation of the Effectiveness of Curing of Concrete Structures." AC1
SP-IO0, Concrete Durability, Katharine and Bryant Mather International Conference, edited by J.M.
Seanlon. Detroit, Michigan: American Concrete Institute, 1987 pp. 207-226.
23. Kasai, Y., I. Matsui, and M. Nagano. non-site Rapid Air Permeability Test for Concrete." AC1 SP-82,
In Situ/Nondestructive Testing of Concrete. Detroit, Michigan: American Concrete Institute, 1984,
pp. 525-42.
86
24. Hansen, A.J., N.S. Ottoscn, and C.C. Peterson. "Gas-Permeability of Concrete In-situ, Theory and
Practice." ACI SP-82. 1984 pp. 343-56.
25. Montgomery, F.R. and A. Adams. "Early Experience with a New Concrete Permeability Apparatus."
Proceedings of Second International Conference on Structural Faults and Repair. Edinburg, London:
Engineering TechnicsPress,1985,pp.359-63.
26. Bradley,
J.S.,R.W. Duschatko,
andH.H. Hinch. "PocketPermeameter:Hand-HeldDeviceforRapid
Measurementof Permeability."
Geological
Notes(1972):568-571.
27. Bourbie,T.,and J.D.Wails. "PulseDecay Permeability:

Analytical
Solution
and Experimental
Test."
Journal
of theSocietyof Petroleum
Engineers22 No. 5 (October,
1982):719-721.
28. Burg,R.G., and B.W. Ost. "BasicEngineering

Properties
of CommerciallyAvailable
High-Strength
Concretes."Researchand DevelopmentBulletin
RD IO4T. Skokie,Illinois:
Portland
Cement
Association,
1992.
29. Campbell,D.H., R.D. Sturm,and S.H. Kosmatka. "Detecting

Carbonation."ConcreteTechnology
Today 12 No. 1 (March 1991):1-5.
87

SHRP S 329

Uploaded by

Copyright:

Available Formats

SHRP S 329

Uploaded by

Document Information

Original Description:

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

SHRP S 329

Uploaded by

Copyright:

Available Formats

SHRP-S/FR-92-109

Condition Evaluation of Concrete

Volume 7: Method for Field

Strategic Highway Research Program

Program Manager: Don M. Harriott

Strategic Highway Research Program

Abstract • • • • ° • * " • ° * • ° ° " • • ° • * * * ° ° * ° * • ° • • * ° ° ° ° * ° • * * • ° * • • ° * • ° 1

Executive Summary ............................................ 3

2. Literature Survey ........................................ 7

3. Method Development ..................................... 15

4. Design and Construction of a Field Device ....................... 35

6. Field Trials ........................................... 51

7. Summary and Recommendations ............................. 69

3-1. Schematic drawing of the laboratory SAF device ..................... 16

3-2. Preliminary classification of permeability based on SAF measurements ....... 26

4-1. Outline drawing of the permeability indicator ....................... 37

6-1. SAF device in operation on Westover Road Bridge, on Route 1, Austin,

6-2. Operation of remote vacuum plate on prestressed beams of bridge over

6-4. Comparison of in situ SAF and laboratory permeability testing for

6-5. Comparison of in situ SAF and laboratory permeability testing for

A-1. Instrument module removed from housing of SAF device ............... 74

A-2. General electronic component diagram ........................... 77

2-1. Ranking of in situ permeability methods .......................... 14

3-1. Concrete test mixtures ..................................... 18

3-2. SAF measured on air-dried slabs .............................. 19

3-3. Measurement of SAF test repeatability ........................... 21

3-4. Effective depth of vacuum penetration ........................... 22

3-5. Effects of steel placement and concrete cover ....................... 23

3-8. Effects of surface drying using IR heater--5 minutes of heating ............ 27

3-9. Effects of surface drying using IR heater--3 minutes of heating ............ 29

3-10. Effects of "natural" drying on SAF ............................. 30

3-11. Chloride diffusion constants (De) for test slabs ...................... 31

3-12. Air permeability as determined from pulse-decay measurements ........... 33

3-13. Water permeability as determined from pulse-decay measurements .......... 33

5-1. Characteristics of concrete mixes used for comparison studies ............. 44

5-4. Results of comparative tests on field slabs .......................... 49

6-2. SAF readings taken on prestressed support beams of bridge over

6-4. SAF readings obtained on deck of Nueces Causeway Bridge in Corpus

6-6. SAF readings taken on prestressed beams on J.F.K. Intra-coastal

6-7. Results of air permeability tests carded out on cores removed

The projex:t's scope was:

1. demonstration of the principles of the device in a laboratory situation;

2. study of the influences of materials and environmental variables on test results;

3. correlation with other measures of permeability;

4. construction of a field device based on the laboratory prototype; and

5. evaluation of the device under actual field conditions.

Q = volume outflow (cm3/sec)

C = concentration of distance x (in meters) from a boundary

M = (PI - Po) "V

Pa = atmospheric pressure, and other quantities are as previously noted

A final category of in situ permeability techniques meriting discussion is the pressure-applied

• Correlation--The method should correlate either with long-term chloride ingress or

• Interferences--An ideal method would have no interferences; however, this may be

• Invasive--If the method requires removal of concrete, or if substances will be

• Mounting--Methods that require mounting on a concrete surface will damage the

• Rapidity--Rough calculations indicate that an "instantaneous" method (i.e., one

• Orientation--Since substructures were included within the scope of the project,

• Weight--A lightweight instrument is desirable, especially for substructure work.