Aggregates Resources STP774 PDF

Copyright by ASTM Int'l (all rights reserved); Thu Dec 22 07:41:33 EST 2011
Downloaded/printed by
Indian Institute of Technology Kharagpur pursuant to License Agreement. No further reproductions authorized.
EXTENDING AGGREGATE
RESOURCES
A symposium
sponsored by ASTM
Committee D-4 on
Road and Paving
IVIaterials
Orlando, Fla., 10 Dec. 1980
ASTM SPECIAL TECHNICAL PUBLICATION 774

W. W. Hotallng, Jr., symposium chairman
ASTM Publication Code Number (PCN)

04-774000-08
€1' 1916 Race Street, Philadelphia, Pa. 19103
Copyright © by AMERICAN SOCIETY FOR TESTING AND MATERIALS 1982
Library of Congress Catalog Card Number: 81-70762
NOTE
The Society is not responsible, as a body,
for the statements and opinions
advanced in this publication.
Printed in Baltimore, Md
June 1982
Foreword
The symposium on Extending Aggregate Resources was presented at
Orlando, Fla., 10 Dec. 1980. The symposium was sponsored by ASTM Com-
mittee E-4 on Road and Paving Materials. W. W.Hotaling, Jr., Hotaling
Associates International Inc., presided as symposium chairman.
Related
ASTM Publications
Asphalt Pavement Construction: New Materials and Techniques, STP 724
(1981), 04-724000-08
Quality Assurance in Pavement Construction, STP 709 (1980), 04-709000-08
Cement Standards—Evolution and Trends, STP 663 (1979), 04-663000-07
Recycling of Bituminous Pavements, STP 662 (1978), 04-662000-08
Low-Temperature Properties of Bituminous Materials and Compacted Bitu-

minous Paving Mixtures, STP 628 (1977), 04-628000-08
A Note of Appreciation
to Reviewers
This publication is made possible by the authors and, also, the unheralded
efforts of the reviewers. This body of technical experts whose dedication,
sacrifice of time and effort, and collective wisdom in reviewing the papers
must be acknowledged. The quality level of ASTM publications is a direct
function of their respected opinions. On behalf of ASTM we acknowledge
with appreciation their contribution.
ASTM Committee on Publications
Editorial Staff
Jane B. Wheeler, Managing Editor

Helen M. Hoersch, Senior Associate Editor
Helen P. Mahy, Senior Assistant Editor
Allan S. Kleinberg, Assistant Editor
Virginia M. Barishek, Assistant Editor
Contents
Introduction 1
Aggiegate-Related Applications for Residues from Fluidized Bed
Combustion Boilers—R. J. COLLINS 3
Resource Conservation by Use of Iron and Steel Slags—D. w. LEWIS 31
Economics of Using Incinerator Residue as a Highway Construction

Material—UDAY PATANKAS, M. R. TAYLOR, AND W. C. ORMSBY 43
Beneficiation of Natural Aggregates by Polymer Impregnation—
R. P. WEBSTER AND I. I. FONTANA 64
Beneficiation of Marginal Coastal Oregon Aggregates with Cement—
CHEN CHANG, R. G. HICKS, AND T. S. VINSON 74
Slag Utilization in Pavement Construction—j. J. EMERY 95
Properties of Marginal Aggregates Treated with Asphalt Emulsion—

G. L. EVANS AND R. G. HICKS 119
Potential for Asphalt Stabilized Sand Bases in Texas—D. N. LmxE,
J. W. BUTTON, AND I. A. EPPS 139
Meeting Construction Aggregate Fracture Requirements with Natural

Product-Sized Material—B. H. POLHEMUS 161
Sea Dredged Sand for Concrete—s. D. RAMASWAMY, M. A. AZIZ, AND

C. K. MURTHY 167
Use of Geotextiles to Extend Aggregate Resources—Q. L. ROBNETT

AND I. S. LAI 178
Ceramic Roadway Aggregates—A. V. PETTY, JR. 196
Copyright by ASTM Int'l

Indian Institute of Technolog
STP774-EB/Jun. 1982
Introduction
In its annual practice of developing and presenting a meaningful sympo-

sium during its winter meeting, ASTM Committee D-4 on Road and Paving
Materials looks to its membership for themes that will draw out the best in
prospective authors. The most informative, interesting and above all timely
papers are then compressed between two covers to produce a volume such as
is now in your hands. This is possible because of the efforts of committeemen
and paper reviewers too numerous to list by name here. But it would be
equally impossible without the Editorial Staff, the Society's publications de-
partment, and others at ASTM Headquarters. Because this is so I dedicate
this edition to the real editors—our magnificent, magnanimous, and often
miraculous counterparts in Philadelphia.
That needed saying, and now said I will explain what we are trying to ac-
complish with this 1980 Symposium edition.
Aggregates are all too often thought to be mundane matter of little worth.
Yet they are expected to occupy most to the final volume of an engineered
mass, mixture, etc. Just so much bulk to take up the most space at the least
possible cost? Perhaps that attitude was tolerable some years ago, but today
far too many geographic areas are on the short side when it comes to long-
term reserves of once plentiful "top quality" aggregates.
Users are becoming increasingly aware of this short-fall of premium ag-
gregate in selected areas through the economic facts of life: diminishing sup-
plies + increasing demands = higher prices. This has caused many to reassess
the practice of unrealistically specifying aggregates. Considering that there
are four primary attributes expected of construction aggregates—that they
be:
1. chemically inert,
2. dimensionally stable,
3. physically durable, and
4. structurally strong—
each should be specified to a level not more than is essential to assure reason-
able performance consistent with the useful life of the proposed improve-
ment. Eliminating over-specification can do much to "extend" aggregate re-
sources through better utilization of what is available, and several papers
herein address this.
Another obvious way to extend what exists is to supplement that supply
Copyright by ASTM Int'l (all 1 rights reserved); Thu Dec 22

Copyright® 1982 by A S TM International www.astm.oi-g
Indian Institute of Technology Kharagpur pursuant to License Agreemen
EXTENDING AGGREGATE RESOURCES
with another resource. Often called an "artificial," "ceramic," or "synthetic"

aggregate, such man-made volumes hold a great deal of promise for certain
applications in selected areas. Again, several papers contained in this publi-
cation shed as much light on this aspect as the current state-of-the-art per-
mits. In fact, design engineers, specifiers, builders, contractors, owners, and
users of every description are encouraged to explore the potentials such ag-
gregate materials may afford in economy, performance, supply, and overall
value.
Other papers discuss extending aggregate resources by using less aggre-
gate initially, stabilizing aggregate mixtures, and in general upgrading indi-
vidual aggregate particles. Each approach can contribute to conserving this
valuable natural resource, the humble yet noble and often unappreciated
"naturally occurring" construction aggregate. And looking to the future,
some methods even fall into the "resource recovery" category—a conserva-
tion minded, environmentally desirable solution to matching disposal needs
with re-use potentials. Committees D-34 on Waste Disposal and E-38 on Re-
source Recovery take note!
Finally, no small tribute is due my fellow committeemen of ASTM Com-
mittee D04.92 on Papers and Symposia—the small group that actually plans,
conducts, and otherwise ramrods this work under the aegis of the main com-
mittee. Each has made my year as Chairman of both the subcommittee and
the symposium much easier than it would have been. To these peers I am for-
ever indebted: J. J. Emery, Trow Ltd. (my successor); J. A. Scherocman,
Barber-Greene Company (my predecessor); and F. T. Wagner, North Caro-
lina State Division of Highways.
W. W. Hotaling, Jr.
Hotaling Associates International Inc., Wash-
ington, D.C., fonnerly, Engineering Ser-
vices, National Crushed Stone Association,
Washington, D.C.; symposium chairman.
R. J. Collins1
Aggregate-Related Applications for

Residues from Fluidized Bed
Combustion Boilers
REFERENCE: Collins, R. J., "Aggregate-Related Applications for Residues from

Fluidized Bed Combustion Boilers," Extending Aggregate Resources, ASTM STP 774,
American Society for Testing and Materials, 1982, pp. 3-30.
ABSTRACT: Continuing development of atmospheric fluidized bed (AFB) combustion

systems offers great potential for producing energy from low quality coals and other com-
bustible materials in an environmentally acceptable manner. The process involves burn-
ing the crushed coal in a bed of limestone or dolomite that has been fluidized by jets of
hot air. The resultant by-products are a dry, sand-size spent bed residue and a relatively
high carbon fly ash. Potential aggregate-related applications for these by-products, and
in particular the spent bed residue, are discussed.
This paper is a further progress report for a research program sponsored by the U.S.
Department of Energy to evaluate potential commercial applications for fluidized bed
combustion wastes. Included in this paper are laboratory data related to a number of dif-
ferent AFB by-product sources; a summary of separation efficiencies of several beneficia-
tion techniques; and supporting data related to three of the most promising applications
for spent bed- residues. These applications are stabilized road base compositions,
masonry block, and synthetic aggregate. Data pertaining to laboratory studies and small-
scale field demonstrations for each of these applications is presented and discussed.
KEY WORDS: aggregates, concrete, fluidized bed, solid waste, coal-associated waste,
combustion, waste utilization, limestone, road base, masonry block, synthetic aggregate
Atmospheric fluidized bed (AFB) combustion is among the most promis-

ing of the emerging technologies for producing energy from coal. The AFB
process is effective in burning all types of coal (including high sulfur, high
ash coals), as well as other combustible materials, while still complying with
current environmental standards.
Essentially, AFB boiler systems burn crushed coal in a bed of finely graded
limestone that is fluidized by jets of heated air. Combustion of the coal usu-
ally takes place at temperatures ranging from 815 to 927°C (1500 to 1700°F),
which is significantly below the normal temperature regimes associated with
'Executive vice president, Valley Forge Laboratories, Inc., Devon, Pa. 19333.
Copyright by ASTM Int'l (all rights reserved); Thu Dec 22

Copyright® 1982 b y A S T M International www.astm.org
4 EXTENDING AGGREGATE RESOURCES
conventional stoker-fired coal burning boilers. The principal advantages of

the AFB process are:
1. capture of sulfur dioxide (SO2) emissions by chemical reaction of sulfur
in the coal with calcium oxide in the limestone bed,
2. reduced nitrogen (NOx) emissions due to lowered combustion tempera-
ture,
3. elimination of the need for flue-gas scrubbing systems, and
4. relatively small size of the boiler unit itself.
Use of the AFB system eliminates the need for disposal of by-product
sludge when burning high-sulfur coal. Instead, dry solid waste residuals are
generated by the process in two forms:
1. a fine dust (referred to AFB fly ash or fines) removed by conventional
dust collection systems, and
2. a granular discharge from the fluidized bed (referred to as spent bed
residue or bed drain).
An industrial-size fluidized bed boiler rated at 100 OCX) lb of steam per
hour may produce from 12 to 20 tons per day of fines from 18 to 30 tons per
day of spent bed residue under normal operating conditions. This represents
a total of 10 000 to 20 000 tons of solid waste that is generated annually. It
has been further estimated that a 100-MW AFB utility boiler will generate
approximately 15 tons per hour of by-products, or in excess of 100 000 tons
per year of waste. Therefore, it is evident that disposal of AFB by-products
may represent significant environmental problems [1].^
Some concern has already been expressed by various government agencies
regarding the disposition of solid wastes from fluidized bed combustion. The
U.S. Department of Energy, recognizing the importance of waste disposal to
the future of the AFB program, has funded two programs to investigate
possible uses for AFB by-products. The first deals with use of the spent bed
material in agricultural applications. The second covers a broad spectrum of
other uses, many of which appear to have some commercial potential.
This paper is a progress report that discusses the objectives and current
findings of the commercial applications program, which was instituted in
December 1976. The focus of this report is to describe program progress in
the area of aggregate-related applications for fluidized bed by-products.
Description of Research Program

The program for investigation of possible commercial applications for
AFB residues involves the following:
Product Characterization—Residue samples from a number of AFB boiler
^The italic numbers in brackets refer to the list of references appended to this paper.
COLLINS ON RESIDUES FROM FLUIDIZED BED COMBUSTION BOILERS 5
units were and are being analyzed to determine physical and chemical com-
position and engineering properties.
Beneficiation Techniques—Different systems for separating or benefi-
ciating AFB residue have been studied for the purpose of removing or con-
centrating desirable components in AFB residue.
Application Studies—The majority of effort is being directed toward a
wide-ranging laboratory investigation of possible uses for AFB residue or its
components or both, including small-scale field demonstrations of those ap-
plications considered most promising.
Handling and Disposal—In case commercial applications are slow in
developing or of insufficient magnitude to utilize all of the waste material,
methods for environmentally acceptable handling and disposal of AFB
residues are also being evaluated.
Review of Findings
Laboratory Characterization Studies

To date, by-product samples from eight different fluidized bed sources
have been involved in the program. Most of the analytical work has been
done on spent bed residues. Collection and handling systems for AFB fines
are not yet fully refined; therefore, the availability of the AFB fines has been
somewhat limited.
Samples of AFB by-products that have been analyzed thus far were re-
ceived from the following sources:
Location Boiler Designer or Operator
Alexandria, Va. Pope, Evans, and Robbins

Alliance, Ohio Babcock and Wilcox
Columbus, Ohio Battelle Laboratories
Georgetown University Foster-Wheeler
Linden, N. J. Exxon Research and Development
Minneapolis, Minn. Fluidyne Engineering
Rivesville, W. Va. Pope, Evans, and Robbins
Windsor, Conn. Combustion Engineering
Spent bed residue samples were received from all of the preceding sources.
Fine dust or fly ash samples have also been received from the Alexandria,
Alliance, Georgetown, and Windsor plants. Only one sample of spent bed
residue was received from the Columbus and Windsor plants. Therefore,
these samples may not be representative of spent bed material from these
sources.
With the exception of the Columbus spent bed residue sample, in which
pulverized limestone was used in the bed, the physical appearance and parti-
cle size distribution of these materials is similar. Figure 1 shows the range of
particle size distribution for the spent bed residues. The gradation range of
the samples was determined in accordance with ASTM Standard Method for
Particle-Size Analysis of Soils (D 422-63). Figure 1 also shows the average
gradation of fine aggregate as specified in ASTM Specification for Concrete
Aggregates (C 33-78). From Fig. 1, it can be seen that the average gradation
of ASTM fine aggregate falls within the gradation range of spent bed
residue.
The chemical composition of AFB residues from six sources, representing
both limestone and dolomite sorbents, is reported in Table 1. Analyses of
these materials was performed in accordance with the procedures outlined in
ASTM Standard for the Chemical Analysis of Limestone, Quicklime, and
Hydrated Lime (C 25-72). Spent bed residues from boilers using limestone
sorbents have a high calcium content, while residues from boilers using
dolomite sorbents have less free lime and higher magnesium oxide concentra-
tions [2].
Figure 2 shows the particle size distribution of four different AFB fly ash
sources. As shown in this figure, a wide range of gradation was observed in
these fines. The chemical composition of AFB fly ash fines is quite different
from that of the spent bed residues, as shown in Table 2. These analyses are
indicative of the high carbon content and reduced calcium levels in the AFB
- -- r "" j H
-- • •'
— —_ —
nr —- _ - -1 ' , • ! ' ' ' '
"I' m - —
—= 7 •m
—
/Ff • v^J-\ Gradation Ranoe
î Ll J 1L
—— — — fi-'^
—- ~~ , —_ —
f-i-I l l J
r —
—— • "
/ '
' 1 1.1
• • 1 . i . 1
1 1'
m
"
#
— ———
—— - - - f ' L
. ! 1 : ' :
'^rriT
M —
—— — /-•-nii-N-!:--/
Averaqe ^,ri IdL ion
f o r flSTM
# 1 1 , •, t 1 ! • •
-- f t 1"
W 1 -1 !
\D - J * Fine Aoqre
r ' har - - ——
^ - 1>i »n r 1; - - ' ' f
-- J * " ' • ZXJ ——'
<^ J . . . 1. . —
^ *' ^ , 1T1 f1 1 —".
4.. 1 J
il • J
.
100
1 .60. ! IX
40 20 4 1/4
• 1
3/B V2
l_
3/4 1 11/2
(U.S. sieve No.) '^ (Inch) '
U, S, Standard Sieve Size
FIG. 1—Typical range of gradation for spent bed residues from fluidized bed combustion
systems.
COLLINS ON RESIDUES FROM FLUIDIZED BED COMBUSTION BOILERS
TABLE 1—Chemical analyses of spent bed residues from fluidized bed combustion systems.
Limestone Sorbents" Dolomite Sorbents
Alexandria Alliance Georgetown Rivesville Linden Minneapolis
LCI 3.78 9.89 11.73 6.56 12.63 10.64

SiOj 16.68 4.12 2.70 19.13 6.14 16.56
AI2O3 2.64 1.76 2.91 4.04 0.42 3.06
FejOj 3.66 3.14 4.37 6.05 4.18 3.14
CaO 50.13 58.27 62.48 42.36 35.99 27.91
MgO 1.49 0.87 3.01 1.68 13.71 11.47
SO4 21.61 23.12 14.50 18.39 25.75 21.28
"Analyses for Columbus and Windsor samples not included due to limitations in sample size
and confirming data.
Sle e Opening
001 0.1 I.Omm
100
7 ^ ~ ? ^ '"
HT' > •
>T
90 .. 3
^
\/'^ .'* 2;
h
3
i y
y >^' /
1
1\
Tz
'
r
//
1 ^v. .', •
K
^T
t '
^ 60 1/
f
^ 50 1 —
1
1
£ 40 111 1
1
1 1
30
1
1 .- ALEXAriDRIA FLY ASH
1
1 1
11
ALLIANCE FLV ASH
//
20
1 GEORGETOWN FLY ASH
11 ^' WINDSOR FLY ASH
10 L --'
..*••
7
' '/
0
1 400 200
t_
.
100 60 40
U. S. standard Sieve Size
FIG. 2—Typical range of gradation for fly ash fines from fluidized bed combustion systems.
fines. In comparison, fly ashes from conventional pulverized coal-fired

boilers usually have carbon contents well below 10 percent.
Separation or Beneficiation Techniques

Another area of investigation has been the separation or beneficiation of
the spent bed residue by-product by use of commercial processing technol-
ogy. The following techniques have thus far been investigated:
1. particle size separation and dry tabling,

2. blunging and wet tabling,
3. air classification,
4. magnetic separation,
5. elutriation,
6. centrifugal separation,
7. crushing and grinding, and
8. hydration or separation of lime constituent or both.
At this time, the most practical and effective means of separation or
beneficiation or both appear to be either grinding and fine screening or a wet
blunging hydration, followed by mineral tabling. Table 3 presents a com-
parison of the separation efficiencies of various beneficiation techniques used
for extraction of a lime-rich fraction from the spent bed residue.
Regardless of the separation technique employed, it is essential to hydrate
the spent bed by-product as a preconditioning step prior to use of the
material. The hydration of spent bed residue with its free lime base is ex-
othermic in nature and is accompanied by considerable release of heat upon
exposure of the material to water [3].
Potential Commercial Applications

Because of the greater availability of spent bed residue and its relatively
high lime content, application studies have thus far focused more on in-
vestigation of possible commercial uses for this by-product. Early in the
study, it was discovered that fluidized bed residues had an apparent value as
admixtures in certain commercial products. For example, blending the spent
bed residue with pulverized coal fly ash in varying proportions serves as an
excellent matrix in a variety of construction-related applications, including
cementitious products. In addition, substantial quantities of calcium oxide
in the spent bed residue indicate certain obvious possibilities for use of this
material as a possible substitute for commercial lime products [4].
This report deals with cementitious applications for the spent bed by-
TABLE 2—Chemical analysis of fly ash fines from fluidized bed combustion systems.
Alexandria Alliance Georgetown Windsor
LOI 47.95 38.73 35.76 30.16

Si02 15.43 9.64 20.03 30.18
A1203 5.29 5.28 4.76 5.49
FejOa 7.68 20.92 11.10 12.81
CaO 13.90 16.15 20.10 10.07
MgO 1.87 1.30 2.78 1.81
SO4 5.44 8.66 8.08 4.23
COLLINS ON RESIDUES FROM FLUIDI2ED BED COMBUSTION BOILERS
, o
B S
J) -p
gS5; ^^
I I
c
a
l7. M
21 ^ ^ (N ^ to ^
^ (N 00 (N i-H r-'
z
-s: u.S
'3 I
îPli
M M 3" 2 g _u
•Sin'Eb S
2 2 S
_3 ^
3 IS K
I
a •c -c
ip
lillJ Ifil
•O T !
c s
mfill
X X
<<
3 111
^?^fsi;
.3* 00
a
I JS 1
I 1
c u
o O
a>
H u **-( •a
I
i
M 00 00 S
m
a
§ •^ S .s s § i §§ ii iB
II
V3 VI
1-1
••§
'y
•§ •§ 1 s i
-M +- ^ — -P
Vi
0) a>
I/}
3
M
v
PQ "rt s 3
•a "3 m "ea <f
« « t is = ?fc4
< 1 ^^Q<m >
Q O J >>>:>
S^
* u u z T3 • ^
product from AFB boiler systems that presently offer potential as a supple-
ment or extension of existing aggregate resources. The applications that ap-
pear to offer such potential are: stabilized road base materials, masonry
block units, and synthetic aggregate.
Stabilized Road Base
Laboratory Studies
In combination with fly ash from conventional coal-fired boilers, spent bed
residue from AFB systems offers considerable promise as a stabilizing agent
for use in road base, subbase, and shoulder applications. The strength
development of such compositions is attributable to the cementitious
characteristics of the AFB bed residue and the pozzolanic properties of the
fly ash. An extensive laboratory test program, using a single source of
bituminous coal fly ash, has been conducted to evaluate a large number of
possible road base compositions. Well over a hundred different mix formula-
tions, involving several different AFB residue sources, have been analyzed in
this program.
Table 4 presents a summary of the laboratory compressive strength test
results for a selected number of different road base formulations in which
AFB residue is included [2]. Some of these formulations also incorporate
AFB fines. In all mixes, spent bed residue materials were first hydrated with
approximately 30 percent water, then allowed to reach equilibrium before
blending with other mixture components.
These compressive strength test results were obtained in accordance with
the procedures described in ASTM Test for Compressive Strength of Molded
Soil-Cement Cylinders (D 1633-63). All mixes shown in this table were
molded at maximum dry density and optimum moisture content, in accor-
dance with the sample preparation procedures of ASTM Specification for Fly
Ash and Other Pozzolans for Use with Lime (C 593-76a).
From Table 4, it is evident that substantial compressive strength values are
attainable from combinations of AFB residue and fly ash, regardless of the
percentage of spent bed residue or its source. Even at lower curing
temperatures, significant compressive strength development was observed in
similar compositions.
In addition to strength development, early laboratory studies of stabilized
road base compositions were also concerned with the durability characteris-
tics of these mixtures. Vacuum saturation tests were performed on selected
formulations in accordance with ASTM C 593-76a. The results of these tests
confirmed that formulations containing AFB residue possessed acceptable
durability and residual compressive strength to resist freezing and thawing.
Because of the levels of calcium oxide and anhydrous calcium sulfate
observed in spent bed residue samples, the potential for volume instability
TABLE 4—Summary of compressive strength development of stabilized road base

compositions containing AFB residue.
Compressive Strength, psi

Formulatiuii, yo 4 „* TIOE: irtAor?
Cureti ill ij L^ Cured at xuw r
AFB
Lime Residue Ash Aggregate 7-day 28-day 7-day 28-day
Limestone Aggregate
2 6 15 77 200 2280 2090 2230
2 12 12 74 230 1860 1340 3310
2 16 12 70 260 2020 1010 4000
2 20 12 66 450 1990 2070 3730
2 24 12 52 380 1510 1680 3830
0 6 15 79 230 1920 1450 2030
0 9 22 69 795 2128 NT- NT
0 10 30 60 540 1540 NT NT
0 12 15 73 250 2060 990 1980
0 16 15 69 120 830 1160 1820
0 20 15 65 180 550 670 3290
0 24 15 61 80 450 390 1980
0 30 10 60 235 520 NT NT
0 44 12 44 310 2040 NT NT
Sand and Gravel Aggregate
0 6 15 79 478 2100 1673 2220
0 10 30 60 188 860 NT NT
0 12 15 73 930 2720 1509 2495
0 15 10 75 838 2474 1252 2784
0 20 20 60 892 2283 2250 2864
0 30 30 40 873 1317 1578 1925
"NT denotes no test performed.
was recognized for these materials. Early investigation of the possibility of

volume expansion in road base compositions containing different levels of
AFB residue consisted largely of periodic measurements of cured cylindrical
test specimens using calipers.
Field Installation
In October 1978, approximately 160 tons of experimental road base
material containing spent bed residues from two atmospheric fluidized bed
boilers were placed to a depth of 152.44 mm (6 in.) on 57.9 (190 ft) of a two-
lane highway reconstruction of State Route 149 in eastern Ohio. Spent bed
residues from AFB installations in Alliance and Columbus, Ohio, were used
in the experimental mixes. Three different mix formulations were used. The
mix proportions, quantities, and laboratory strength development character-
istics of these mixes are given in Table 5.
One concern of any road construction project is the stockpiling of an ade-
quate supply of material in accordance with contractor requirements.
Indian Institute of Technology Kharagpur pursuant to License Agreement. No further reproductions authorize
s
r2
00
s
t
j< H
III § s s
»4^
x; .§ II tf g>
to -
^ '•» J3 'S
e e
US s «
1/1 E " C o li<
o Ss S
"^ 5 H H
H "3 (rt 1/2
O i<<
Q 2 J : -c
a .2 M
O S +3 -H
< M'S * *
o * •S "O o u
u.S S S
E -o E J) -a "p
_4*s
a g s
' ~ S -S
° Su a
rt ^ « ,«
U < H H
Preparation of the spent bed residues for this project involved premixing the
residues with aggregate and water in a pugmill mixer several weeks before
placement. The stockpiled material blends were then mixed with optimum
proportions of fly ash and water on the day the products were shipped to the
job site.
The mixing plant used for supplying this project had been used for a
number of years to prepare lime-fly ash-aggregate mixtures for similar types
of road base installations. Unfortunately, the automatic moisture controls
for the plant were not functioning on the day of final mixing. Consequently,
an excess of moisture above optimum was added to Mix A during the initial
mixing operations. The first several truckloads of material were rejected for
being overly wet. Even the material that was placed was on the wet side of op-
timum, making it extremely difficult to compact. Moisture adjustments were
made so that Mixes B and C were blended in the mixing plant by eye at their
appropriate moisture contents. All mixes were sealed and overlaid with a
conventional wearing surface 76.2 mm (3 in.) thick, in keeping with standard
practice in the State of Ohio.
Several weeks after placement of the experimental base materials, one of
the compositions began showing signs of linear expansion, similar to that
observed with steel slag aggregate that is not sufficiently aged prior to use.
The disturbance was most pronounced in Mix A, which contained the
highest percentage of AFB residue and was mixed with excessive moisture.
Evidently, the spent bed residue was not sufficiently hydrated during initial
mixing and, in the presence of excessive moisture in a compacted state, addi-
tional hydration and accompanying volume changes occurred. Although the
problem was most pronounced in Mix A, it was decided to terminate the test
program pending further investigation of other formulations and alternative
means of conditioning and handling the spent bed residue [5].
Preconditioning Experiments
Similar instances of volume instability have been previously noted in struc-
tural compositions using pulverized coal fly ash derived from lignite and sub-
bituminous coals. In order to properly precondition these materials, the fly
ash is fu-st mixed with water, allowed to "cure" in a stockpile for several
weeks, and then crushed to a uniform fineness for use.
To investigate means of alleviating the expansive tendencies of AFB
residue, while at the same time providing a more uniform product for com-
mercial use, arrangements were made to process some 18 tonnes (20 tons) of
spent bed material at the American Fly Ash Company's conditioning plant
near Chicago. This facility is used routinely to process high calcium lignite
and subbituminous fly ash prior to distribution to the Chicago ready-mix
market. The conditioning process involved adding 25 to 32 percent water to
2.25 tonne (l-Vz ton) batches of AFB residue, thoroughly mbcing for several
minutes, discharging the AFB residue onto a stockpile, and then recrushing
several weeks later.
The consistency of each batch in the mixer initially resembled that of a
slurry, but, upon discharge from the plant, the conditioned residue was in
the form of a moist sand. After several weeks of stockpiling, the crusher
reduced the particle size of the residue so that all particles were finer than
2.00 mm (10 mesh sieve).
Besides an investigation into field preconditioning tests, an extensive
laboratory program was also initiated to study the expansive characteristics
of a number of AFB spent bed materials. This program included observation
over a one-year period of expansion bars, in accordance with ASTM Test for
Potential Alkali Reactivity of Cement-Aggregate Combinations (Mortar-Bar
Method) (C 227-71). In addition, dimensional measurements were also made
on test pads, 0.305 by 1.22 by 0.15 m (12 by 48 by 6 in.), and proctor-size test
cylinders representing various mix formulations.
A review of the findings from this test program has led to the following
conclusions:
1. Exposure to free water is necessary in order for excessive expansion to
occur.
2. Compositions with ratios of AFB residue to fly ash of 3:7 or less show no
detrimental expansion, when considering linear expansion in excess of 0.1
percent to be detrimental.
3. Expansion bars maintained free of direct contact with water for a
longer time after molding showed considerably less expansion than their
counterparts.
4. Expansion bars made from AFB residue that was hydrated at the
American Fly Ash Company's conditioning plant generally exhibited reduced
expansion characteristics.
5. Expansion measurements from a number of 0.305 by 1.22 by 0.15 m
(12 by 48 by 6 in.) test pads confirm the previous observation that composi-
tions with ratios of AFB residue to fly ash in excess of 3:7 will expand greater
than 0.1 percent [5].
Further Field Tests

Based on these observations, several larger size test pads were placed in the
field during the late Fall of 1979 in Canton, Ohio. All test pads used AFB
residue from the Alliance, Ohio, boiler that had earlier been preconditioned
at the American Fly Ash plant. Some of the AFB residue had also been
passed through the crusher at the conditioning plant prior to mixing.
Table 6 summarizes the mix proportions and low temperature compressive
strength characteristics of the test pad mixtures placed in Canton. Following
installation of the test pads, a number of control points were placed in the
Q S
Q* t i
'I
i-H O O l/> LO
•Sac ^' o o o^ K
•S o a
ed
i
111 §;:
. ON
1^ vd
TH O
s—'00
p u
• o 2
=".2 S
I &^
- SS
•c
£ .2
I Kl
S
aa
1
1f
11
U
>
4) j j
SS2e 8
0) J5 J3 X !
> &a tS CO
'" '8 g g g
y Y Y
lii
a
cS i3 O. D. O.
< <;
I ^ (N (S r^ Ti-
lii
surface of each mixture for dimensional monitoring. Figure 3 shows the

dimensions and respective locations of these test pads. The surfaces of all test
pads were permitted to remain exposed in order to make periodic measure-
ments of dimensional stability and to assess the penetration resistance of the
test pads.
Periodic monitoring of these test pads has been made in order to evaluate
the physical appearance, strength development, and dimensional stability of
the respective compositions. Table 7 summarizes the dimensional stability
measurements made between adjacent control points since May 1980. Table
8 summarizes the average values of penetration resistance for each of the test
pads. Penetration resistance data was obtained by recording the number of
I- -10'-0"
~B
^1 I I
f^ I , =
MIX 3 ^ A MIX 4-A ol ?
4i
1 lO
C
-±\ ^ LJ_
T^ --t
MIX 2
to '
M- C
• - ^ •
MIX 1 MIX 1
F
C
a.
AFB Residue Percent Percent Percent

Preparati on Aggregate^ AFB Residue' Fly Ash»
1 As Received 50 15 35
2 As Received 70 10 20
2-A Pre-crushed 70 10 20
3 Pre-crushed 75 5 20
4-A Pre-crushed 20
'Sand and gravel, local to Canton.

'From Alliance, Ohio.
A bituminous coal fly ash from western Pennsylvania.
FIG. 3—Dimensions, mix proportions, and locations of control points for Canton, Ohio, road
base demonstration.
TABLE 7—Dimensional stability measurements Canton, Ohio, test strips.
Monitoring Date
5/2/80 7/9/80 8/27/80 11/7/80

Mix Measurement
Original
No. Identification" Length, ft* Change in Original[Length, %''
1 A 6.0 + 0.17 +0.17 + 0.34 + 0.34

B 15.0 -0.14 -0.14 -0.14 -0.14
C 15.0 -0.07 -0.07 -0.07 0
D 6.0 + 0.17 + 0.34 + 0.34 + 0.34
E 15.0 + 0.07 + 0.07 + 0.07 +0.07
F 15.0 -0.07 0 -0.07 -0.07
G 6.0 + 0.52 + 0.69 +0.52 +0.52
2 A 6.0 + 0.52 +0.78 + 0.69 + 0.69
B 15.0 -0.49 -0.42 -0.42 -0.42
C 15.0 -0.83 -0.70 -0.70 -0.70
D 6.0 + 3.13 + 3.30 +3.30 +3.30
2-A A 6.0 + 1.22 + 1.40 + 1.22 + 1.22
B 10,0 + 0.10 + 0.21 + 0.21 +0.21
C 10.0 + 0.31 + 0.42 +0.31 + 0.42
D 6.0 + 1.04 + 1.22 + 1.22 + 1.22
3 A 6.0 -0.17 0 +0.34 + 0.17
B 15.0 -0.21 0 -0.07 -0.07
C 15.0 -0.14 -0.14 Q*** +0.07''
D 6.0 0 + 0.17 Q««« -0.87''
4-A A 5.98 + 0.35 + 0,70 +0.70 + 0.52
B 10.0 + 0.21 + 0.31 + 0.42 + 0.42
C 10.0 0 + 0.21 +0.10 + 0.21
D 6.0 0 + 0.35 + 0.34 + 0.34
"Refer to Fig. 3 for locations of measurements.

''I ft = 0.3048 m.
'Plus ( + ) means increase in length; minus ( —) means decrease.
''Measurements reflect replacement of point on 8/27/80.
hammer blows required to drive a Number 4 reinforcing bar in 2.54 cm (1

in.) increments into the material.
From the standpoint of dimensional stability, Mixes 1 and 3 appear to be
the least expansive. Mix 2-A, containing pre-crushed residue, is generally
more dimensionally stable than Mix 2 of the same composition, in which the
residue was conditioned as received.
As of this writing, three additional test strips have recently been placed in
Canton, Ohio, to provide further documentation of dimensional stability and
strength development for stabilized road base compositions containing
relatively low percentages of AFB residue. The beneficial effects of sealing
the base mixes with an asphalt coating and a bituminous wearing surface will
also be evaluated. This information will then be used to design compositions
for placement of a state highway project during the 1981 construction season,
TABLE 8—Average penelnition resistance measurements Canton. Ohio, test strips.
Monitoring Date
5/2/80 7/9/80 8/27/80 11/7/80
Mix No. Average Number of Blows for 101.6 mm (4 in.) Penetration"
1 23.3 18.7 32.0 22.0

2 24.0 23.0 33.7 30.0
2-A 21.7 29.7 26.0 31.0
3 22.0 18.3 29.3 24.3
4-A 21.3 16.7 19.7 15.3
"The total number of blows required for 101.6 mm (4 in.) penetration were recorded at each of
three locations in each test strip. These totals were averaged.
in cooperation with the Ohio Department of Transportation and the Ap-

palachian Regional Commission.
Production of Masonry Block
Initial Test—Kurtz Brothers

To date, masonry block units containing AFB residue have been manufac-
tured at commercial block plants on three separate occasions. An initial pro-
duction run was made in January 1978 at the Kurtz Brothers plant in
Ephrata, Pa. At that time, approximately 500 blocks were made, of which
300 contained AFB residue from Alexandria, Va. During this initial produc-
tion run, the spent bed residue was evaluated as a possible substitute for the
fine aggregate in the block. The residue was conditioned or hydrated in the
batching unit at the plant just prior to block production.
Several different mix formulations containing AFB residue were used.'but
only two compared favorably with that of the control mix used at the plant.
Selected block from these two mixes and the control mix were tested in the
laboratory for specific gravity and absorption characteristics in accordance
with ASTM Methods of Sampling and Testing Concrete Masonry Units (C
140-75). Masonry units were cured outside for periods of one month and one
year following 24 h of kiln curing. Table 9 presents a comparison of the mix
composition and physical test data for these block. Both of the mixes con-
taining AFB residue satisfy the physical requirements for Type N-I block, for
general use, as given in ASTM Specifications for Hollow Load-Bearing Con-
crete Masonry Units (C 90-78). Mix 1 is classified as medium weight and Mix
2 as normal weight.
The test results from Table 9 confirm that a partial replacement of the fine
aggregate portion of a conventional block mixture with AFB residue results
in a comparable, and in some cases, a superior masonry block product. The
TABLE 9—Comparison of composition unci physical properties of AFB concrete masonry

block from first production run.
Control Mix 1 Mix 2
Aggregate 90.1 55,5 52.4

(% dry weight)
Fly ash 1.8 5.8 11.0
(% dry weight)
AFB residue 24.2 27.6
(% dry weight)
Cement 8.1 9,4 8.9
(% dry weight)
Mix water 6.1 9,7 9.5
(% total dry weight)
Number of block 163 49 48
Dry unit weight (lb/ft-')
1 month 131.6 119,5 131.3
1 year 127.3 121.8 130.4
Compressive strength (psi)"
gross area
1 month 800 1070 1322
1 year 1040 1253 1355
Absorption, 24 h (%)
1 month 9.41 10,79 6.93
1 year 8.38 10,91 7.91
Dimensional stability (average %)
1 year + 0.34 + 0,37 + 0.41
"1 psi = 6.9 kPa.
lighter color and finer surface texture of the AFB block were felt to be
desirable features of these units.
Second Test—Calcite Block Company

A second block production run was made during January 1980 at the
Calcite Block Company in Lebanon, Pa. During this run, a total of 385 block
were made. Approximately 190 of these block contained Alexandria AFB
residue, using five different mix formulations. During this run, the spent bed
residue was also evaluated with a reduced level of cement in the mix. Prior to
production, the residue was conditioned in a ready-mix concrete truck, then
placed into drums and shipped to the block plant.
One of the AFB mixes from this run was comparable to the control mix at
the plant. Several other mixes were able to meet the physical requirements
for hollow non-load bearing block, as given in ASTM Specification for Non-
Load Bearing Concrete Masonry Units (C 129-75). Table 10 presents a com-
parison of the mix composition and physical test data for these block.
The data in Table 10 confirm the findings of the earlier production run.
The data shows that masonry block units meeting Type N-I requirements for
TABLE 10—Comparison of composition and physical properties of AFB concrete

masonry block from second production run.
Control with
Control Fly Ash Mix 2 Mix 5 Mix 7
Aggregate
(% dry weight) 92.6 90.4 56.5 59.4
Fly ash
(% dry weight) 2.5 6.4 6.2 17.1
AFB residue
% dry weight) 28.2 28.5 78.1
Cement
(% dry weight) 7.4 7.1 8.9 5.9 4.7
Mix water
(% total dry weight) 6.1 5.7 8.6 8.3 20.2
Number of block 51 105 39 45 45
Dry unit weight (Ib/ft^) 134.2 134.0 112.5 112.1 86.9
Compressive strength
(psi)" after 30 days
Gross area 1360 1590 1000 590 590
Net area 2420 3110 1760 1120 1110
Absorption, 24 h (%) 6.54 7.05 14.79 16.00 27.70
"1 psi = 6.9 kPa.
general use can be produced by replacing approximately half the aggregate

with AFB residue and maintaining slightly higher cement levels. Mix 2 meets
these requirements as a medium weight block.
These data also show that mixes using reduced cement levels, including
total substitution of aggregate with spent bed residue and fly ash, can meet
the requirements for non-load bearing masonry units. In this case, Mix 5 is a
medium weight block and Mix 7 is a lightweight block.
Third Test—Kurtz Brothers

In September 1980, a total of approximately 700 additional masonry block
were produced at the Kurtz Brothers plant in Ephrata, Pa. Of this total,
there were approximately 510 block containing AFB residue from Alliance,
Ohio, that was conditioned during July 1980 in a ready-mix concrete truck.
The residue was placed in drums after conditioning and shipped to the block
plant. In this production run, none of the block contained fly ash.
Table 11 summarizes the composition and physical test data for the con-
trol mix and the three test mixes containing AFB residue. These data show
that Mixes 1 and 2 meet criteria for medium weight Type S-I block for
limited use and Mix 3 easily satisfies the requirements for non-load bearing
hollow masonry units.
The block from the third production run was used to construct a 18.29 m
(60 ft) long masonry wall, which was erected during October 1980 on the
TABLE 11 —Comparison of composition and physical properties of AFB concrete masonry

block from third production run.
Control Mix 1 Mix 2 Mix 3
Aggregate
(% dry weight) 91.4 55.2 55.2 55.2
AFB residue
(% dry weight) 34.5 36.2 37.9
Cement
(% dry weight) 8.6 10.3 8.6 6.9
Mix water
(% total dry weight) 7.0 9.5 10.5 9.0
Dry unit weight
(Ib/ft^) 129.0 118.0 112.5 119.9
Compressive strength (psi),"
after 40 days
Gross area 1355 886 767 597
Absorption, 24 h (%) 8.6 12.8 14.7 12.4
"Ipsi = 6.9 kPa.
property of Highway Materials, Inc. in Malvern, Pa. A total of 178 commer-

cially available masonry block and 262 experimental AFB residue block were
used. The wall was ten courses high—three below ground and seven above
ground. The two end sections of the wall contained the control block and
each of the three intermediate sections were made from one of the AFB
residue mixes. Figure 4 shows the dimensions of this experimental masonry
wall.
After the wall was erected, various sections were treated with different
coatings (including parging, oil-base paint, and latex paint) to determine
ease of application, porosity, and compatibility. Control spikes were also in-
stalled to monitor the dimensional stability of the wall. Monthly inspections
of this wall are currently in progress. Since this block wall has been installed,
some signs of distress (such as mortar cracking and expansion) have been
observed in sections of the wall containing fluidized bed block that has been
exposed to weathering. No such signs of distress were noted in sections of the
wall that had been sealed to prevent the entry of water into these block, in-
dicating that these block are more suited to interior use.
The results of masonry block investigations to date have shown that high
quality block can be made with replacements of up to 40 percent of the ag-
gregate using conditioned AFB residue and fly ash, while maintaining ap-
proximately the same cement content. Because the density of the individual
blocks containing AFB residue is less than conventional block, the amount of
cement per block is not increased in the AFB blocks. Reduction in the ce-
ment content have thus far resulted in block that meet requirements for
special use hollow load-bearing units or non-load bearing units.
5 9'-4"
16'
Mix 3 Mix 3
2" Exp. - 2 " Exp. 2" Exp. 2" Exp.
Courses
t
CONCRETE FOOTING - 12" THICK X 30" WIDE
FIG. 4—Experimental wall using AFB masonry block.
Manufacture of Synthetic Aggregate
Sintered Aggregate
Efforts to produce synthetic aggregate using AFB residue have come about
in several ways. The first attempts were directed toward production of a
Hghtweight sintered aggregate using four size ranges of spent bed residue,
ranging from "as received" to 100 percent passing the 100 mesh sieve. Fly
ash was also added to AFB residue at the rate of 20 percent fly ash by weight.
Pellets were fired at temperatures ranging from 982 to 1316°C (1800 to
2400°F). Pellet strength and texture generally improved with increasing
temperature and decreasing particle size. The highest pellet strengths were
measured with the - 1 0 0 mesh residue and fly ash, fired at 1316°C (2400°F).
It was felt, at the time, that these pellets did not represent an optimum for-
mulation and the energy input required to produce a sintered aggregate
product was not particularly attractive from an economic standpoint.
Pelletized Aggregate
The use of pelletizing equipment was also investigated for production of
synthetic aggregate from AFB residue. Both roll pelletizing and pan pelletiz-
ing equipment were used in this effort. It was found that pelletizing of AFB
residue was successful only when the material was first ground to —50 mesh
particle size.
When using a roll pelletizer, rectangular shaped pellets were produced at
an optimum moisture content of approximately 25 percent. The pellets were
usually about 6.35 mm ('A in.) on a side and 12.7 to 25.4 mm (V2 to 1 in.) in
length. Although these pellets did not fracture when dropped to the floor
from a height of 1.22 m (4 ft), the surface of the pellets abraded fairly easily.
In the case of the pan pelletizer, a solution of 15 percent sodium silicate
mixed with 85 percent water was used as the binding agent. In general, the
spherical pellets produced from the pan pelletizer ranged from 9.5 to 19.1
mm (% to % in.) in diameter. These pellets also did not fracture when
dropped to the floor and had a more solid surface than the pellets from the
roll pelletizer.
Although pellets were first formed using AFB residue alone, the resultant
product was not considered sufficiently hard to qualify as a commercial
grade aggregate. Additional trials were performed using combinations of
spent bed residue with either fly ash or AFB fines. The best pellets were pro-
duced with an equal weight blend of AFB residue and fly ash. Standard ag-
gregate evaluative tests were performed according to ASTM C 33-78 on both
the roll and pan pellets from these blends. The results of these tests are sum-
marized in Table 12. Despite high porosity and absorption values, the
physical properties of these pellets suggest the possibility of producing
acceptable lightweight synthetic aggregate from either pelletizing process.
Both the roll and pan pellets from the equal weight blending of AFB
residue were also evaluated as possible coarse aggregate in lightweight
Portland cement concrete mixes. A control mix using 19.1 mm (% in.)
limestone served as a point of reference in terms of density, workability, ce-
TABLE 12—Properties of pelletized aggregate from AFB residue.
Roll Pelletized, Pan Pelletized,

% finer % finer
by weight by weight
Sieve size
lin. 100 99.7
% in. 98.1 97.5
1/2 in. 44.0 31.0
% in. 9.2 12.8
No. 4 0.5 1.1
Clay lumps and
friable particles
(ASTM C 142-78)" 1.10 2.68
Specific gravity
(ASTM C 127-77)
Bulk 1.22 1.26
Bulk SSD* 1.48 1.74
Absorption, (24 h) 21.4 38.1
Unit weight
(ASTM C 29-78)*^ 38 Ib/ft^ 45 Ib/ft^
Voids in aggregate
(ASTM C 30)'' 50.1 42.7
Particle shape rectangular ('/2 to 1 in. long) spherical
"ASTM Test for Clay Lumps and Friable Particles in Aggregates (C 142-78).
''SSD refers to saturated surface dry.
''ASTM Test for Unit Weight and Voids in Aggregate (C 29-78).
''ASTM C 30 replaced by ASTM C 29-78.
ment content, and compressive strength. The approach used in proportion-

ing the lightweight mixes was to fix the weight relationship of the fine aggre-
gate, cement, and water, and add the pelletized aggregate to the maximum
extent that permitted acceptable workability.
To prepare the concrete mixes, the pelletized aggregates were soaked for
24 h prior to batching. All mixing was done in a Hobart mixer to achieve a
pumpabie condition (approximately 101.6 mm or 4 in. slump). Test results
of these concrete mixes are shown in Table 13. These data indicate that low
strength lightweight concrete can be designed using pelletized synthetic ag-
gregates composed of AFB residue-fly ash blends.
Briquetted Aggregate
Production of briquettes was also investigated as a possible technique for
converting AFB residue into synthetic aggregate. It was believed that the use
of commercial briquetting techniques offered a practical method of compact-
ing wet materials into shapes with a strong enough compactive effort to
approximate either a primary crusher run product or a coarse aggregate. In
addition to making use of spent bed residue, briquetting also offered an op-
portunity to produce combinations of fly ash and other by-products in a com-
paratively inert, stable form.
Initial briquette experiments were conducted in the laboratory using a
Carver press mold with a 13 789 kPa (2000 psi) compactive effort. Combina-
tions of AFB residue, fly ash, and a gypsum sludge waste were used to mold
TABLE 13—Proportions and properties of lightweight concrete mixes using

pelletized aggregate from AFB residue.
Proportions, Weight per yd-*
Materials Control Mix Mix 1, pan Mix 2, roll
Cement, Portland Type 1 728 581 782

Sand (dry) 1453 1160 1203
Crushed Umestone-y4 in. (dry) 1556 0 0
Pelletized aggregate (saturated surface
dry) 0 1303 1170
Water" 309 247 332
Waterxement ratio (lb/lb) 0.42 0.42 0.42
Consistency (visual) pumpabie* pumpabie* pumpabie*
Unit weight (Ib/ft^) at:
molding 149.9 121.9 129.2
27 days, 22.8°C (73°F), 100% RH 150.8 123.1 130.7
28th day, oven dry 146.9 113.6 121.6
28 day compressive strength (psi) 4280 1970 3020
"Weight of water does not include absorption.

*Pumpable refers to a slump of approximately 101.6 mm (4 in.).
COLLINS ON RESIDUES FROM FLUIDI2ED BED COMBUSTION BOILERS 25
cylindrical test specimens, which were cured under various conditions and
later tested for compressive strength. The information from this initial in-
vestigation was then used to develop target mix formulations for a later
feasibility testing program. This program was performed in March 1980 at
the laboratory of K.R. Komarek, Inc. in Chicago, 111.
A laboratory-scale roll briquetting machine was used to produce synthetic
aggregate pellets with various combinations of AFB residue and fly ash, and,
in some cases, gypsum sludge waste. The objective of the feasibility testing
program was to produce briquettes that attained sufficient "green" strength
to be easily handled and stockpiled after molding. Chemical additives were
also introduced into some of the mixes to react as a binding agent.
Of the ten mix combinations evaluated, virtually all resulted in briquettes
that possessed sufficient "green" strength for normal handling. All bri-
quettes were evaluated to determine their specific gravity, absorption, and
crushing strength in order to select optimum formulations for larger scale
production.
During June 1980, a total of 386.4 kg (850 lb) of briquettes were produced
at K.R. Komarek, Inc. In this phase of the program, 18 different mixes were
used, consisting of combinations of AFB residue, fly ash, sulfate waste from
three sources, and two chemical additives. Table 14 is a summary of the mix
formulas used and an evaluation of the "green" strength of the briquettes
produced from each mix.
These briquettes were then aged in four different environments:
1. moist cured at 22.8°C (73°F);
2. sealed in plastic bags at room temperature;
3. placed outdoors in open, drained containers; and
4. moist cured at 12.8°C (55°F).
Periodic evaluations were made of the bulk specific gravity, absorption,
and scratch hardness of these aggregates. Eighty groups of briquettes,
representing various ages and curing conditions, were tested for specific
gravity and absorption in accordance with ASTM Test for Specific Gravity
and Absorption of Coarse Aggregate (C 127-77). The average bulk specific
gravity was 1.69, ranging from 1.43 to 2.04. The absorption averaged 15.0
percent with a range from 6.5 to 22.1. Briquettes containing the magnesium
carbonate additive performed best according to scratch hardness.
Briquettes representing optimum mix formulations and curing conditions
were further evaluated for their potential as aggregate in portland cement
concrete mixes. Thirty-one briquette samples were crushed in a laboratory-
scale jaw crusher. The gradation of the resulting samples was determined
and compared to that of a typical limestone aggregate that was similarly
crushed. The range of gradations obtained from the crushed briquettes is
shown in Fig. 5. The gradation of the crushed limestone is also shown for
comparison. As seen in this figure, the crushed briquettes are a more finely
TABLE 14—Mix formulas (% by weight) and evaluation of AFB aggregate briquettes.
Sulfate Wastes Visual

% Observation of
Desig- AFB Fly University of Chemical "Green"
nation Residue" Ash* Glidden Tennessee Rockwell Additive Strength'^
1 30 70 easily handled
immediately
after molding
(good)
2 30 70 1% set up within 1 h
ethylsili- (excellent)
cate 40
3 5 45 50 fair to good
4 15 35 50 better than Mix
#3 (good)
5 15 35 50 1 % ethyl- too wet to mold
silicate
40
6 30 20 50 excellent
7 30 20 50 1 % ethyl- good to excellent
silicate
40
8 70 30 good
9 70 30 1 % ethyl- excellent
silicate
40
10 30 70 10% excellent
MgCOj
11 5 95 poor
12 30 70 poor
13 100 poor
14 100 fair
15 5 95 poor to fair
16 30 70 fair to good
17 5 95 fair
18 47.5 47.5 5% good
MgCOj
"Alexandria AFB residue.

*Fly ash from Wagner Station, Baltimore Gas and Electric Co.
'•"Green" strength is the ability of the briquettes to hold together after molding.
Poor = slightest finger pressure will break briquette.
Fair = can handle briquettes without them falling apart.
Good = can be dropped on floor and not fall apart.
Excellent = difficult to break with finger pressure—feels like hard rubber.
graded material than the crushed limestone, but are also a well-graded ag-
gregate.
Los Angeles abrasion tests were performed on selected groups of briquettes
representing mixes with the best apparent physical properties. Due to the
limited supply of briquettes, the test was modified in terms of sample size
and weight of charge from the procedures outlined in the standard test,
Sieve Opening
1 1-!'
- i _
RANGE OF GRADATIONS FOR 1 f-N^li

CRUSHED BRIQUETTES 1 J f! >? >•sKl
>I
v ^
1 S- K v ' ^ : ! !
— .'^fl
^s
I / »^^
1 iJR'N's
—
1 1 / >^ '^^ ->S 1 1
/l '/^: V N N'i
! /i'sjCV
—— — ! irîvvv f t
JW'S'SV^; s
:-i
_«
— [-1 ; ^N'^'^v^J LI CRUSHED LIMESTONE
1 i-1 ir^v^vvif
yS 's'sN'sy
^1
1
JFi \\.K-^
^<:s:'s\J^
J]
ixi
! ! , 1 j ^ ^ ^ ^ " ^
1 ; 11 ^^^VKTiJ^
.. 1 1 j # k S \ S S ^ X 1
; j jjr^^<;<;jpr
OÔJ^ i >
1 T" 1 1
1 1 i-J^
1
J_
200 100 '^^^
60
L l T TTI
40
1 U w
' 11
20
1
4..1M
U~
ri~
3/B Va
1
3/4
il 1
1
i 1
11/2
(U.S. Sieve Ho.) ^^ (Inch) '
U. S. Standard Sieve Size
FIG. 5—Gradation of crushed briquettes and comparison with crushed limestone.
TABLE 15—Abrasion resistance of briquettes as determined by a modified

Los Angeles abrasion test.
Charge Number of Retained on

Mix No./Cure" Weight, g Spheres #12 Sieve, g Wear, %
Limestone 2500 6 2023 19.1

2/A 2185 6 1634 25.2
10/A 2138 6 1474 31.1
4/A 2247 6 1485 34.1
19/A 2145 6 1377 35.8
1/C 3016 7 1883 37.6
8/D 2250 6 1367 39.2
9/B 3498 8 2000 42.8
6/B 3565 8 1815 49.1
9/B 2236 6 1005 55.1
"Curing conditions:
A = in a moist room at 22.8°C (73°F).
B = sealed in plastic bags; at room temperature.
C = outdoors in open; drained containers.
D = in a moist condition at 12.7°C (55°F).
O -rt ir^
gl
T3 1/^ «
• s fO -^
+2 -^ i/i
•o o -< 00 O
m to
•5??
3
o^
-H
rn
rn
P5
r- o
II?
T3 'TH r - <N i n
•« & 5
•w ^ ^
3
o
•5; d ' ^
2 •«• IT) ON O^
•O t^ Tf
•5 9!^
• O^ ON ;«
"^.-3
|:!-t
C^^ C
ON Q,
in J*
OOÔ g
>= 5= „ 5
•cue"'
liil
i3â "s 9*g.>>
rt
a T3
T3
00
•? e « s r-
CO u ? o
ASTM Test for Resistance to Abrasion of Small Size Coarse Aggregate by

Use of the Los Angeles Machine (C 131-76). The results of these modified
Los Angeles abrasion tests are shown in Table 15. These data clearly show
that the moist cured briquettes were the most abrasion resistant. The moist
cured briquettes also most closely resembled the gradation of the crushed
limestone after being passed through the jaw crusher.
The crushed briquettes were selectively screened, weighed, and recom-
bined to produce the gradation of a Pennsylvania 2B aggregate. This ag-
gregate was then used as the coarse aggregate fraction in concrete mixes with
the following general proportions by weight: 1.00 parts cement, 2.81 parts
fine aggregate (dry sand), 2.70 parts coarse aggregate (dry crushed briquet-
tes), and 1.16 parts water.
Concrete cylinders (101.6 mm or 4 in. diameter by 203.2 mm or 8 in, high)
were made from ten different briquette mixes, under various curing condi-
tions. A total of 22 different concrete mixes were made. Table 16 summarizes
the mix proportions and available compressive strengths from selected
mixes. Several of the mixes were able to attain 28-day compressive strengths
in the range of 20 684 kPa (3(XX) psi), indicating potential use in some con-
crete applications.
Additional investigative work will be directed toward the manufacture of
over a hundred tons of large size briquettes at a commercial briquette plant.
The briquettes thus produced will be stockpiled as primary crusher run and
eventually crushed for evaluation and use as aggregate material.
Conclusions
There are a number of applications for fluidized bed combustion by-
products that offer the promise of extending commercial aggregate re-
sources. The most promising of these at this time are stabilized road base
compositions, masonry block, and synthetic aggregate. briquettes. Regard-
less of intended use, AFB spent bed residues, because of their free lime con-
tent, must be thoroughly hydrated prior to utilization. As AFB boilers
become more plentiful, increased opportunities should present themselves
for combining these materials with other by-products.
References
[1] Collins, R. J., Journal of Testing and Evaluation. Vol. 8, No. S, Sept. 1980, pp. 259-264.
[2] Minnick, L., Quarterly Technical Progress Reports, "Development of Potential Uses for the
Residue from Fluidized Bed Combustion Processes," U.S. Department of Energy Reports,
DOE/FE/2549, UC-90e, National Technical Information Service, U.S. Department of
Commerce, Springfield, Va., 1977-1980.
[3] Miller, R. H., "Potential Uses for the Residue from Fluidized Bed Combustion Processes,"
Proceedings, Fifth International Conference on Fluidized Bed Combustion, Washington,
D.C., 12-14 Dec. 1977.
[4] Minnick, L. J. and Miller, R. H., "Utilization of By-Products from Fluidized Bed Combus-
tion Systems," Proceedings, Sixth International Fluidized Bed Conference, Atlanta, Ga.,
9-11 April 1980.
[5] Minnick, L. J., "Potential Uses for Fluidized Bed Combustion Residuals," Proceedings,
Conference on Fluidized Bed Combustion System Design and Operation, U.S: Department
of Energy, Morgantown, W. Va., 28 Oct. 1980.
D. W. Lewis1
Resource Conservation by
Use of Iron and Steel Slags
REFERENCE: Lewis, D. W., "Resource Conservation by Use of Iron and Steel Slags,"
Extending Aggregate Resources, ASTM STP 774, American Society for Testing and
Materials, 1982, pp. 31-42.
ABSTRACT: Use of iron and steel slags as construction materials is believed to be the
oldest and best developed "resource recovery" operation in the world. Iron blast-furnace
slag use in the United States began more than 100 years ago with current use about equal
to production. Steelmaking slags were largely wasted prior to World War II but are now
used extensively in flexible pavement construction and railroad ballast. The 1979 Bureau
of Mines data show use of nearly 36 million tons (32.5 X 106 Mg) of slag in commercial
applications, replacing a similar amount of virgin materials.
Supplies of blast-furnace and steel slags are expected to increase by only a small
amount during the remainder of this century. Use patterns will undergo marked changes,
greatly increasing energy conservation and continuing conservation of virgin materials.
Large amounts of blast-furnace slag will be processed for use as a cementitious material,
saving both the virgin raw materials that would be used to produce a similar tonnage of
Portland cement and the huge amounts of energy used in cement kiln operation. Increas-
ing recognition of the skid resistant characteristics of steelmaking slags will result in
greater usage in asphalt concrete surface courses.
These changes in use patterns will decrease the amounts of slag available for use in
base course and embankment construction. In many areas, it will be possible to use lower
quality natural aggregates, recycled construction materials, or suitably processed wastes
in these applications, thus further extending the supply of high quality virgin aggregates.
KEY WORDS: aggregates, cement, construction materials, energy conservation, recy-

cling, resource recovery, slag, waste materials, concrete
In recent years shortages of high quality natural aggregates have been ex-
perienced in a number of areas of the United States. Depletion of local sup-
plies, environmental regulations, restrictive zoning, and higher land values
for other uses have all contributed to this situation, which is particularly
critical in urban areas. At the same time, disposal of waste materials has
become more difficult and expensive as a result of increasingly stringent en-
'Chief engineer, National Slag Association, Alexandria, Va. 22314.
Copyright by ASTM Int'l (all 31 rights reserved); Thu Dec 22 0

Copyright® 1982 by A S T M International www.astm.org
Indian Institute of Technology Kharagpur pursuant to License Agreement.
vironmental regulations and shortages of suitable, nearby disposal sites. The

Resource Conservation and Recovery Act of 1976 provides requirements for
solid waste disposal and also outlines the Federal Government's responsibil-
ity in encouraging the production and use of recovered resources (including
fuel) from solid waste.
The disposal of waste and industrial by-product materials by utilizing
them in construction applications, and thereby helping to alleviate aggregate
supply problems, is of long-standing interest in some areas and of greatly in-
creased activity in all. This is evident from the recent formation of new
ASTM committees such as E-38 on Resource Recovery and D-34 on Waste
Disposal, and from extensive research studies such as those of the National
Cooperative Highway Research Program [1,2].^ Recycling of construction
materials [3], upgrading of low-quality aggregates [4], and specification of
lower quality materials in appropriate uses [5] are also being given a large
amount of attention.
In any consideration of means for conserving and extending the resources
of high quality natural aggregates, or of conserving energy, the use of in-
dustrial by-products such as the slags from the steel industry immediately
comes to mind. That these materials have long been used in various applica-
tions is common knowledge; the extent and type of use is not so well known.
It is the purpose of this paper to discuss the characteristics and uses of these
slags and the potential for increased or improved applications to further
benefit the conservation of aggregate and energy resources. The paper is con-
cerned only with the iron blast-furnace slags and the steelmaking slags pro-
duced by the steel industry.
Iron Blast-Fumace Slags

ASTM Definitions of Terms Relating to Concrete and Concrete Ag-
gregates (C 125-79a) defines blast-furnace slag as "the nonmetallic product
consisting essentially of silicates and alumino-silicates of calcium and other
bases, that is developed in a molten condition simultaneously with iron in a
blast furnace." The blast furnace is continuously charged at the top with iron
oxide sources (ore, pellets, sinter, etc.), flux stone (limestone and dolomite),
and fuel (coke). Molten iron collects in the bottom of the furnace with the liq-
uid slag floating on it. The iron and slag are periodically tapped from the
furnace. This is the first stage in production of steel followed by final refining
of the iron into the desired end product.
The slag consists primarily of the silica and alumina from the iron ore
combined with the calcium and magnesium oxides from the flux stone. It
comes from the furnace as a liquid at about 1500°C. Dependent upon the
^The italic numbers in brackets refer to tfie list of references appended to this paper.
LEWIS ON IRON AND STEEL SLAGS 33
cooling method used, one of the three different types of product described in
this section is made from the molten slag [6,7].
Air-Cooled Blast-Fumace Slag

The molten slag runs into a pit adjacent to the furnace or is transported in
large ladles and poured into a pit some distance away. It is permitted to
solidify under prevailing atmospheric conditions, after which the cooling may
be accelerated by water sprays on the solidified mass. The slag is
predominantly crystalline, with a cellular or vesicular structure resulting
from bubbles of gases that were dissolved in the molten slag. After cooling
sufficiently to be handled, the slag is dug, crushed, and screened to the
desired sizes.
Air-cooled blast-furnace slag crushes to angular, roughly cubical particles
with pitted surfaces that provide excellent bond with binder materials and
high stability from internal friction when used without cements. Bulk specific
gravity and unit weight are dependent upon grading and particle size (as in
lightweight aggregates) because of the vesicular structure. Slag screenings
(finer than the 4.75-mm sieve) are near natural sand in density values, while
coarse aggregate sizes are as much as 20 percent lighter than natural ag-
gregates having the same gradings. The material is highly resistant to
weathering effects, does not polish to produce slippery surfaces, is not sub-
ject to harmful cement-aggregate reactions, and is hydrophobic in nature,
thus assuring adhesion of bituminous material. Air-cooled slag is included in
most standard specifications for all types of construction use. The difference
in weight compared to natural aggregates is not sufficient to classify it as a
lightweight aggregate; specifications class air-cooled blast-furnace slag with
gravels and crushed stones as "normal weight" aggregates. The weight sav-
ing from slag use is a significant factor in many applications, however.
Granulated Blast-Fumace Slag

The molten slag is quickly quenched to produce a glassy, granular product
called "granulated slag." Quenching with water is the most common pro-
cess, but air or a combination of air and water may be used. This is the most
rapid of the cooling processes, permitting very little mineral crystallization.
Dependent upon the slag composition and its temperature at the time of
quenching, and the speed of quenching (factors such as water volume,
pressure, etc.), the granulated slag may vary from a friable, popcorn-like
structure to small, sand-size grains resembling a dense glass.
The slag glass consists of the same major oxides as does portland cement,
although the relative proportions of lime and silica differ considerably. The
granulated slag has excellent hydraulic properties and, with a suitable
activator such as calcium hydroxide, sets in a manner similar to portland ce-

ment. Granulated slags may be crushed, graded, or ground for specific ap-
plications.
Expanded Blast-Fumace Slag

Expanded slag is produced by treating the molten blast-furnace slag with
controlled quantities of water, less than that necessary for granulation.
Various machine and pit processes may be used to combine the slag and
water, producing a lightweight product with a more vesicular or cellular
nature than the air-cooled slags. Variations in the relative amounts of water
and slag and the process used control the cooling rate and can result in prod-
uct variations from materials resembling air-cooled slags in crystallinity to
highly glassy materials from some machine processes. A pelletizing process,
developed by National Slag Limited in Canada, uses limited amounts of
water followed by chilling of slag droplets thrown through the air by a rapidly
revolving finned drum [8\. This produces spherical pellets of highly glassy
slag.
The expanded slag particles are either angular and roughly cubical in
shape, or spherical, depending upon the processing procedure. They have a
pronounced cellular structure that results in densities in the lightweight ag-
gregate category. The expanded slags also possess cementitious properties in
varying degrees.
Steelmaking Slags
Steelmaking slags are the by-products of the processing of molten iron
(known as hot metal) and scrap (known as cold metal) into the desired type
and grade of steel in "steelmaking" furnaces. Three common types of furn-
aces are the open-hearth furnace, the basic-oxygen furnace, and the electric
furnace. The first two of these are the most common in integrated steel mills
where molten iron from the blast furnaces can be combined with scrap metal.
Open hearths were originally the predominant type but now have been
largely replaced by the more efficient basic-oxygen furnaces. Electric furn-
aces can operate on very high (up to 100 percent) scrap charges and are
largely used in production of special alloy steels and in "mini-mills" where
only cold metal is available for the furnaces. The slags are often referred to as
"steel slags," "open-hearth slags," "BOF (basic oxygen furnace) slags," and
"electric-furnace slags."
Steel making is a "batch" process, where the furnace is charged with the
necessary materials, the composition adjusted and alloying materials added
to produce the desired composition of steel in that particular "heat." Both
slag and molten steel are then emptied from the furnace and the process
repeated. The different types of furnaces, grades of steel produced, and in-
dividual furnace operating practices result in variations in composition and

properties at different mill locations. Specifications covering steelmaking
slags are usually local rather than national in nature, being fitted to the
materials available in the particular area.
Steelmaking slags are composed primarily of calcium silicates, calcium
aluminoferrites, and fused oxides of calcium, iron, magnesium, and manga-
nese. Materials added to the melt just before the end of a heat may result in
incomplete incorporation in the slag, thus some free oxides, including CaO,
may exist in the final slag.
Compared to air-cooled blast-furnace slags, the steelmaking slags are
much heavier, harder, denser, and less vesicular in nature. They have
unusually high resistance to polishing and wear in pavement surfaces.
Slag Uses—Past and Present

Slag has been produced as long as iron has been made, and its use in con-
struction constitutes the world's oldest and best developed "resource
recovery" operation. Slag from the Roman Catalan forges was used in road
building about 2000 years ago. Such use was apparently widespread in the
Roman Empire, with recent excavations in England uncovering many such
roads [9]. Use was quite sporadic until recent years, with early uses even in-
cluding application as cast cannon balls in the 16th century. Blast-furnace
slag's hydraulic properties were recognized in the early 1800's when it was
used in some forms of cements in Germany and use as a concrete aggregate
has been reported to date back to about 1870.
In the United States, early records of the use of blast-furnace slag are far
from complete, but use in road building is known to date back at least to
1830. In the 1860's slag bases were built on many streets in Chicago, and it
was used as railroad ballast in 1875. Use in mass concrete began in the
1880's, and structural concrete use followed by the turn of the century. Use
in concrete paving and in bituminous surfaces began prior to World War I [6].
A large amount of granulated slag was used to produce cement in the early
part of the 1900's. Consisting of ground slag and lime, it was largely dis-
placed by the growing portland cement industry. Portland blast-furnace slag
cements were utilized extensively during periods of cement shortages after
World War II.
Air-cooled blast-furnace slag, used in aggregate applications, has been the
predominant type. Although extensive research and field trials and long ser-
vice records had established the outstanding properties of the slags long
ago—excellent adhesion to cementing materials, lighter weight than natural
aggregates, superior fire resistance as a concrete aggregate, better polish
resistance and skid resistance properties in bituminous surfaces, superior
stability in fills, bases and bituminous paving, and excellent durability under
all weathering conditions—essentially complete use of blast-furnace slags
Indian Institute of Technology Kharagpur pursuant to License Agreement. No further reproductions authorized
was not attained until the late 1950's. Since that time, usage has approx-
imately equalled—and in some years exceeded—production. As a result,
many of the previously existing piles of "waste" slag were "mined" and used.
The use of steelmaking slags developed later than that of the blast-furnace
slags, and it has not progressed quite as far. Prior to 1940, the steelmaking
slags (primarily from open-hearth furnaces) were wasted, often in dumps
that also contained all manner of other waste material and refuse from the
steel mills. Recognition of the value of the steel contained in the slags led to
processing during World War II solely for metallics recovery, with the slag
itself still being wasted. A few of the early attempts at use as aggregates were
unsuccessful due to volume changes from hydration of CaO (and MgO from
burned dolomite from the furnace linings in some open-hearth slags). These
difficulties, usually associated with use of fresh (and often unprocessed)
materials in rigidly confined locations, where even small amounts of expan-
sion would be detrimental, inhibited rapid development of further uses.
Present day steelmaking slags (largely from basic-oxygen furnaces that
represent more than 60 percent of the U.S. steel production) are being suc-
cessfully used in two major types of application: as a source of both iron and
flux materials as part of the blast-furnace charges and as a high quality
mineral aggregate for specific uses. No volume change problems have been
encountered in use for unpaved shoulders or bases, for erosion control, as a
railroad ballast, or in bituminous surfaces where these slags have also
demonstrated outstanding skid resistance properties [10]. Base course and
fill applications, where exceptionally high stabilities are obtained, may re-
quire proper selection, processing and aging (weathering) before use. Al-
though some steelmaking slags have no significant expansion characteris-
tics, they are not used in rigidly confined applications—such as for concrete
aggregate or as fill under building floors and inside foundation walls. In
some areas, steel slags that are not recycled to the furnaces are still being
wasted.
Current Blast-Fumace Slag Uses

Statistics on the use of air-cooled blast-furnace slag for the last three years
are shown in Table 1, taken from U. S. Bureau of Mines data [11]. During
this period, the total tonnages of blast-furnace slag of all types were 25.7
million tons (23.3 X 10^ Mg) in 1977, 28.4 million (25.8 X 10^ Mg) in 1978,
and 27.5 million (24.9 X 10^ Mg) in 1979; closely approximating the
estimated total production in each year. The air-cooled slag constituted
about 88.5 percent of the total during 1977 and 1978, increasing to nearly 91
percent in 1979. During these three years the percentage of the air-cooled
slag going into road base and fill increased from less than 46 to more than 49
percent. Since air-cooled slag is currently less valuable than granulated or ex-
panded, and road base and fill applications are among the least valuable of
TABLE 1—Air-cooled iron blast-furnace slag

sold or used in the United States, by use."
Quantity, thousand short tons
Use 1977 1978 1979
Concrete aggregate 2 553 2 192 2 367

Concrete products 738 764 421
Cement manufacture 146 148 W
Asphaltic concrete aggregate 4 024 3 916 3 421
Roadbases 7 289 8 875 8 452
Fill 3 097 2 677 3 861
Railroad ballast 3400 2 417 2 505
Mineral wool 525 987 826
Roofing, built-up and shingles 255 234 247
Sewage treatment 46 89 W
Soil conditioning 2 ytfd W
Glass manufacture 217 187 21
Ice control 27 W W
Other uses* 433 2 632 2 889
TotaF 22 753 25 119 25 009
"Data from Bureau of Mines [7/].

'includes airport runway base, drainage, miscellaneous, and uses indicated by symbol W.
•^Data may not add to totals shown because of independent rounding.
''W = withheld to avoid disclosing individual company proprietary data; included in other.
NOTE—One short ton (2000 lb) = 0.907 Mg (approximately).
the air-cooled uses, it would seem at first glance that the industry is going the
wrong way. In fact, these changes only reflect the economic situation with
respect to local aggregate supply and demand, effects of economic and en-
vironmental pressure on the steel industry, and increasing concentration of
steel production in a few areas. The slag industry cannot move operations to
new locations, nor alter the amounts of material that must be processed and
marketed.
Table 2 shows similar data for granulated and expanded blast-furnace
slags, indicating decreases in production primarily caused by the shut-down
of specific furnace operations. The major use for granulated slag was not in
cement manufacture, but in road base construction—as it has always been
with the exception of the post World War II cement-shortage years.
Steelmaking Slag Applications

Table 3 presents the Bureau of Mines data on steel slag for the same period
of time. Again, road base and fill applications constitute a major portion of
the tonnages used. The tonnages shown do not include materials recycled to
the furnaces, but do include an estimated one to two million tons (0.9 to
TABLE 2—Granulated and expanded iron blast-furnace

slags sold or used in the United States, by use."
Granulated Expanded
Use 1977 1978 1979 1977 1978 1979
Lightweight concrete aggregate 83 509 1030 931

Concrete products 119 107 916 652 607
Cement manufacture 60 yfd W 50 78 W
Roadbases 1064 936 637
Fill 151 214 W W W
Soil conditioning 45 48 25
Ice control 8 1
Other uses* 40 67 109 153 110
TotaF 1488 1372 855 1475 1914 1648
"Data from Bureau of Mines [11].

*Includes drainage, miscellaneous, and uses indicated by symbol W.
•^Data may not add to totals shown because of independent rounding.
''W = withheld to avoid disclosing individual company proprietary data; included in other.
TABLE 3—Steel slag sold or used in the United States, by use.'
Use 1977 1978 1979
Asphaltic concrete aggregate 3% 607 822

Roadbases 2929 4793 4237
Fill 2330 1784 1882
Railroad ballast 354 549 530
Soil conditioning 44 W
Fire protection 35 8
Other uses"^ 580 723 773
Total 6668 8457'' 8252
"Data from Bureau of Mines [//].

Êxcludes tonnage returned to furnace for charge material.
'Încludes ice control, miscellaneous, and uses indicated by symbol W.
''Data do not add to total shown because of independent rounding.
^W = withheld to avoid disclosing individual company proprietary data; included in other.
1.8 X 10^ Mg) obtained from old bank deposits. The total returned to the
furnaces and used in aggregate applications from current production is
believed to be 75 to 80 percent of the total production. Three to four million
tons (2.7 to 3.6 X 10* Mg) are estimated to be wasted annually.
The steelmaking slag statistics do show one trend that is believed to be

highly significant: the increase (both in tons and percentagewise) of the use for
asphaltic concrete aggregate. This reflects the growing use in pavement sur-
faces to provide the higher friction, safer pavements mentioned previously.
Use in railroad ballast has also increased in the past few years.
Effect of Slag Use on Resource Conservation

The total tonnage of slag used—32.4, 36.9, and 35.8 million tons (29.4,
33.5 and 32.5 X 10* Mg) in 1977, 1978 and 1979, respectively—seems small
in proportion to the total tonnages of construction aggregates used annually
in the United States. It is, in fact, slightly less than 2 percent of the total. It
should be kept in mind that slag is available only in local areas where steel is
produced, often the same areas where natural aggregates are in short supply,
and where it becomes a very important factor.
Basically the current usage of slag accomplishes the following:
1. Conservation of about 40 million tons (36.3 X 10* Mg) of high quality
natural aggregates per year.
2. Conservation of several million tons of iron ore and flux stone by recycl-
ing steel slags to the furnaces.
3. Conservation of a major portion of the energy required to mine and
transport the alternative materials needed if the slag were not available, and
of the energy needed for processing of alternative lightweight aggregates.
4. Separation and recovery of metallics are a part of all slag processing,
making available several million tons of iron and steel for recycling each
year.
5. Elimination of mountains of discarded material in the vicinity of the
steel mills—about 1 Vi billion tons (1.4 X 10^ Mg) over the past 70 years.
Future Prospects
The future of slag usage and its impact upon conservation of aggregate
resources is directly tied to the future of steel industry with respect to
amounts of materials available. The steel industry will also determine the
locations in which slags will be available, dependent upon plant expansions
and closings dictated by the economic situation.
Available Quantities of Slag

A report of the American Iron and Steel Institute (AISI) [12] predicted an
annual increase in steel consumption of IV2 percent per year during this
decade. The problems of the industry—low rates of return, foreign imports,
excessive governmental control and regulation, price controls, and escalation
of costs—are discussed and two possible scenarios for the situation in 1988
are presented. The first, portraying the circumstances if appropriate govern-
ment policies were adopted, indicates imports remaining at about the 1978
level, with predicted increases in actual furnace production of about 10 per-
cent. The optimistic estimate of slag supplies would involve a similar in-
crease—only about 1 percent per year. The second AISI scenario ("A Con-
tinuation of Current Trends") indicates imports increasing from 18 to 40 per-
cent of consumption with huge decreases in domestic production during the
decade 1978 to 1988.
The true situation will probably be somewhere between the most pessi-
mistic and optimistic scenarios. Not only in this decade, but for the re-
mainder of this century, increases in slag production will probably be small.
It is expected that use of steelmaking slags will increase somewhat by use
of materials now being wasted. This would be only partially in the construc-
tion aggregate field and partially in others, including recycling to the fur-
naces. Even complete utilization of all slags—a zero wasting level—would
not result in any large increases in available supplies.
Use Patterns
Future changes in resource conservation from the use of steel industry
slags must come about primarily from modifications in the patterns of use
that currently exist. The precise effects of these are obscured in a rather
cloudy crystal ball, with the number of differing predictions approximately
equal to the number of predictors. Nonetheless, there are several trends,
already under way or obviously in the best interests of both industry and na-
tion, that should be considered.
Slag Cements—The shortages and high costs of energy have stimulated a
great amount of interest in the use of ground granulated blast-furnace slag as
a partial replacement for cement in concrete construction [13]. In the last few
years, numerous studies and conferences on energy saving have pointed out
great benefits to be derived from use of such slags mixed with portland ce-
ment. Each ton of slag so used conserves more than 1.6 tons of natural raw
materials plus the energy required to manufacture the cement clinker, ap-
proximately that of 0.23 tons of coal or 0.9 barrels of oil or 167 m-' (5900 ft^)
of natural gas [14].
Large tonnages of slag are used in cement manufacture in other nations,
and separately ground slag is added to portland cement at the mixer in South
Africa, England, Canada, and recently on a small scale in the United States.
Water-granulated slag is being used in South Africa and England, pelletized
slag in Canada, and a regular machine-expanded slag in the United States.
ASTM Committee C-1 is currently revising the specifications for portland
blast-furnace slag cements, while a specification for separately ground slag is
being prepared by a task group in ASTM Committee E-38.
All of the blast-furnace slags produced in the United States are chemically
suitable for use as cementitious materials. However, locations of slag produc-
tion and cement needs do not coincide well, and the ultimate potential for
such use can only be the subject of a "guess." It has been suggested that
seven to eight million tons (6.3 to 7.2 X 10* Mg) of slag per year might be
used in cement within the next decade or so, and that a reasonable ultimate
goal would be about half of the total blast-furnace slag production. Pelletized
slags are being considered for use in cements at several locations at the pres-
ent time and installation of the first new water granulation facility is under-
way.
Other Blast-Fumace Slag Uses—The special attributes of slag for use in
concrete and bituminous mixtures lead to expectations that use in these
fields will increase, especially in the lightweight aggregate applicatons as use
of the pelletizing process increases. The result of this, and the use in cement,
will mean greatly reduced amounts of blast-furnace slag for use in railroad
ballast, fill, and base course applications.
Changes in Steelmaking Slag Uses—The use of steel slags for their lime
and iron oxide values is expected to increase, both as a blast-furnace charge
and for applications such as neutralization of acidic wastes and as a raw
material for production of portland cement, which are currently being in-
vestigated along with a number of other uses [75]. These slags are ideally
suited for use as railroad ballast and for bituminous surface courses, uses
that should increase markedly in the future. Such changes in use patterns
will be accomplished by a corresponding decrease in other applications,
primarily the base and fill categories.
Summaiy
The use of slags from the steel industry is presently close to the total pro-
duction, a goal that will be attained in the near future. This is the oldest and
probably the most successful "resource recovery" operation in the world.
From the standpoint of environmental effects, conservation of construction
materials resources, and energy saving, it is of great value to the nation. Slag
use has been developed in competition with high quality natural aggregates;
its success is based on economic availability and the ability to perform at
least as well as the competitive materials in the various applications for which
it is marketed. There have been no compromises in construction quality, and
there should be none in the future use of alternative materials.
There will be only small increases in the total amounts of slag available in
the near future. Increased chemical and cementitious applications of the
available supplies will tend to provide even greater savings in natural
resources and energy. This will be accompanied by decreased usage in base
course and fill applications.
It is believed that developments now underway in use of lower quality

natural aggregates, recycled construction materials, or suitably processed ur-
ban and industrial wastes will permit use of these materials in base and fill
applications. Much of the research needed to establish appropriate tests,
specifications, and procedures to ensure maintenance of adequate construc-
tion quality in these and other uses is already underway. Combined with the
shift to higher and better slag uses, the overall effect will be to further extend
the supply of high quality virgin aggregates.
References
[1] Marek, C. R., Herriri, M., Kesler, C. E., and Barenberg, E. J., "Promising Replacements
for Conventional Aggregates for Highway Use," NCHRP Report No. 135, Highway
Research Board, Washington, D.C., 1972.
[2] Miller, R. H. and Collins, R, J., "Waste Materials as Potential Replacements for Highway
Aggregates," NCHRP Report No. 166, Transportation Research Board, Washington, D.C.,
1976.
[3] "Recycling Materials for Highways," NCHRP Synthesis No. 54, Transportation Research
Board, Washington, D.C., 1978.
[4] Cady, P. D., Blankenhom, P. R., Kline, D. E., and Anderson, D. A., "Upgrading of Low-
Quality Aggregates for PCC and Bituminous Pavements," NCHRP Report No. 207,
Transportation Research Board, Washington, D.C., July 1979.
[5] Living With Marginal Aggregates, STP 597, American Society for Testing and Materials,
Philadelphia, Pa., 1976.
[6] Josephson, G. W., Sillers, P., and Runner, D. G., "Iron Blast-Fumace Slag: Production,
Processing, Properties, and Uses," Bureau of Mines Bulletin 479, U.S. Department of In-
terior, Washington, D.C., 1949.
17] "Processed Blast Furnace Slag: the All-Purpose Construction Aggregate," NSA 178-1, Na-
tional Slag Association, Alexandria, Va., 1978.
[5] Margesson, R. D. and England, W. G., assignors to National Slag Limited, U.S. Patent
No. 3,594,142, 20 July 1971.
[9] Lee, A. R., Blastfurnace and Steel Slag, Edward Arnold (Publishers) Ltd., London, 1974,
pp. 6-7.
[10] Ryell, J., Corkhill, J. T., and Musgrove, G. R., "Skid Resistance of Bituminous-Pavement
Test Sections: Toronto By-Pass Project," Transportation Research Record 712, Transpor-
tation Research Board, Washington, D.C., 1979, pp. 51-61.
[;/] Zlobik, A. B., "Slag-Iron and Steel," Minerals Yearbook 1978-79, Vol. I, Bureau of
Mines, U.S. Department of Interior, Washington, D.C., 1980, pp. 821-883.
[12] "Steel at the Crossroads: the American Steel Industry in the 1980's," American Iron and
Steel Institute, Washington, D.C., Jan. 1980.
[13] Price, J. D., Troop, P., and Gershman, H. W., "Potential for Energy Conservation
Through the Use of Slag and Fly Ash in Concrete," Report SAN-1699-T1, U.S. Depart-
ment of Energy, Washington, D.C., 29 Dec. 1978.
[14] Dikeou, J. T., "Cement", Minerals Yearbook 1978-79, Vol. I. Bureau of Mines, U.S.
Department of Interior, Washington, D.C., 1980, pp. 153-191.
[15] Barnes, T. M. and Strong, J. M., "Use of Lime Values in Steelmaking Slag Wastes,"
paper presented at Seventh Mineral Waste Utilization Symposium Chicago, 111., 20-21
Oct. 1980.
Uday Patankar,l M. R. Taylor,1 and W. C. Ormsby2
Economics of Using Incinerator

Residue as a Highway Construction
Material
REFERENCE: Patankar, Uday, Taylor, M. R., and Ormsby, W. C , "Economics of

Using Incinerator Residue as a Highway Construction Material," Extending Aggregate
Resources, ASTM STP 774, American Society for Testing and Materials, 1982, pp.
43-63.
ABSTRACT: The technology for the use of residue from municipal incinerators in high-
way construction has been developed over a number of years in conjunction with studies
performed for the Federal Highway Administration's Materials Division. This study fo-
cuses on an evaluation of the economic factors influencing the use of incinerator residue
in bituminous highway construction. Five Standard Metropolitan Statistical Areas
(SMSA) (Chicago, 111.; Harrisburg, Pa.; Miami, Fla.; New York, N.Y.; and Washing-
ton, D.C.) were chosen to evaluate the economic factors that weigh on the greater use of
residue in highway construction. The residue was evaluated for use as subbase material
and in bituminous base. A cost analysis model was developed to determine economic tea-
sibility that has broad application for other substitute aggregate economic analysis. The
results of this study show that incinerator residue would prove competitive with virgin
aggregate in highway construction in three of the five metropolitan areas studied. Projec-
tions can therefore be made that it is likely that incinerator residue would be economi-
cally feasible in other urbanized SMSAs having municipal incinerators.
KEY WORDS: incinerator residue, economics, solid waste, highway construction, ag-
gregates, cement, concrete
The technology for the use of residue from municipal incinerators in high-
way construction has been developed over a number of years in conjunction
with studies performed for the Federal Highway Administration's (FHWA)
Materials Division. A number of studies have been conducted on the source
and quality of residue, the proper treatment of the material for use in high-
way construction, and the optimum mix of residue and asphalt to ensure a
'Vice president and research associate, respectively, JACA Corp., Fort Washington, Pa.
19034.
2
Chief, Chemistry and Coatings Group, Materials Division, Federal Highway Administration,
U.S. Department of Transportation, Washington, D.C. 20590.
43
Copyright by ASTM Int'l (all rights reserved); Thu Dec 22
Indian Institute of Technology Kharagpur pursuant to License Agreeme
good bond. Test strips have been placed in Houston, Baltimore, Philadel-
phia, Harrisburg, and Washington, D.C.
With the state-of-the-art developed to a highway application stage, an
evaluation of economic feasibility for using the residue became necessary.
The methodology selected to evaluate this feasibility was to analyze the cur-
rent real-world situations at five geographically distributed and carefully se-
lected Standard Metropolitan Statistical Areas (SMSA). Data were gathered
on the cost of processing the residue for use in road construction, transporta-
tion costs for both incinerator residue and quarried material, current prices
of constituent materials such as asphalt cement and hydrated pulverized
lime, and the current cost of aggregate in each locale.
This information provided the basis for a cost analysis model developed to
evaluate the economic feasibility of using residue as opposed to quarried ma-
terial. This model shows areas within each of the SMSAs where the use of in-
cinerator residue as subbase and base course would compete favorably with
quarried material.
While these case studies represent specific situations in selected areas,
conclusions of a broader scope concerning the economic feasibility of using
incinerator residue in highway construction were made.
In order to ensure a broad applicability of the information generated in
this study, certain technological assumptions were necessary. The following
results of previous FHWA studies concerning the processing necessary to im-
prove the incinerator residue to acceptable quality for use as subbase and in
bituminous mix were assumed as given during our research:
1. In order to improve the incinerator residue to acceptable quality for use
as subbase, a certain amount of pulverized lime is added to the mix. The lime
acts as a stabilizer and binder.-' A value of 6 percent was assigned to the hy-
drated pulverized lime added.
2. The density of residue from the incinerators studied was assigned a
value of between 45 and 50 lb/ft-' or approximately 1350 Ib/yd^ (796 kg/m-').
Studies performed in New York City'' and information gathered during site
visitations bear this value out.
3. Pugmill processing of the residue is required in order to screen the ma-
terial to acceptable highway specification requirements. This processing re-
moves the larger size particles and is generally assumed to add $0.50 per
cubic yard ($0.66 per cubic metre) to the price of the residue.
4. As incinerator residue is anticipated to be an aggregate extender or sub-
stitute when used in bituminous mixtures, an additional amount of asphalt
cement needs to be added to the mix due to the greater specific surface of the
^Gnaedinger, J. P., U.S. Patent No. 3,293,999, 27 Dec. 1966.

''"Marketability of Recovered and Clarified Incinerator Residue in the New York Metropoli-
tan Area," Leonard S. Wegman Co., Inc., U.S. EPA Report No. 530/SW-42C-73-008, Wash-
ington, D . C , 1973.
PATANKAR ET AL ON INCINERATOR RESIDUE 45
residue as compared to virgin aggregate. For the purpose of this study, it was
assumed that 2 percent additional asphalt is needed with incinerator residue.
The economic analysis contained in this study addresses two scenarios for
incinerator residue use in highway construction: (1) residue used in subbase,
and (2) residue used in bituminous base.
For Scenario 1 (residue used in subbase), the pugmill processing is as-
sumed to take place at the incinerator site before the residue is shipped di-
rectly to the job site.
For Scenario 2 (which involves application of asphalt), incinerator process-
ing to desired highway-use quality is assumed to take place at the incinerator
site prior to the shipment to the asphalt plant.
Locations and Characteristics of Municipal Incinerator Residues

In order to evaluate the economic aspects of using incinerator residue in
highway construction, a first step is to locate the various incinerators and es-
timate the amount of residue produced and the residue characteristics that
might affect the proposed use. Previous studies in this area were obtained
and, to the extent feasible, utilized as an information base that was then
modified or augmented by work under this project.
Over the past two decades a number of studies have been conducted locat-
ing municipal incinerators operating in the United States. The results of
these surveys have shown a continuous decline in the number of incinerators.
The first of these studies was performed in 1966. It listed comprehensive de-
sign and operational characteristics of 290 incinerators then operating in the
United States and Canada. In 1973, a U.S. Environmental Protection
Agency (EPA) study located 193 plants. In 1975, Richard Fenton, then As-
sistant Administrator of New York City EPA, presented a list of 142 incinera-
tors operating to the Solid Waste Processing Divsion of the American Society
of Mechanical Engineers.
Fenton's list was utilized by Pindzola and Collins^ in an FHWA study that
developed a list of potential residue sources that might provide aggregate for
use in highway construction.
By utilizing the Pindzola and Collins study as a base and searching other
sources such as the U.S. EPA's Compliance Data System (CDS) and Na-
tional Emissions Data System (NEDS), a list of incinerators currently operat-
ing in the United States was developed.
In order to verify the information obtained from the two computerized sys-
tems and collect additional data on the status of incinerators, telephone calls
^Pindzola, D. M., Collins, R. J. et al, "Technology for Use of Incinerator Residue as Highway
Material," Federal Highway Administration, Report No. RD-77-151, Washington, D.C., 1976.
were made to numerous state and local air pollution control and solid waste
management agencies as well as to the incinerator operators themselves.
The data shows 72 municipal incinerators currently operating in the
United States. The vase majority of these plants are located east of the Mis-
sissippi, with the greatest number located in the urbanized areas of the
northeast.
This list was compiled in early 1979 and consequently there are probably
fewer municipal incinerators operating now due to the impact of the Clean
Air Act requirements. Numerous cities have found it uneconomical to up-
grade the air pollution control equipment on their facilities and have closed
them down.
In order to determine the quantity of incinerator residue that would be
available for use in highway construction, the residue output of each plant
was calculated by multiplying the plant refuse design capacity by the number
of operating days per year times the estimated weight fraction of refuse re-
maining after incineration. From these predicted residue outputs, the
amount of incinerator residue for use in highway construction annually on a
national basis was determined to be 2 305 048 tonnes (2 561 164 tons). This
figure is likely to be overstated as the operating hours, solid waste flow, and
degree of burnout in actual municipal incinerators surveyed during the case
studies revealed a lower tonnage of residue available. A ratio of predicted res-
idue quantity against actual output at ten incinerators surveyed during the
case studies projected a realistic range of between 1 262 989 tonnes
(1 403 301 tons) to 2 305 048 tonnes (2 561 164 tons) for 1979.
Aggregate Demand
The demand for aggregate is an important factor that, to a large extent,
will determine whether or not incinerator residue will be used as an aggregate
substitute or extender. Consequently, it is necessary to determine both cur-
rent and future aggregate demand. Moreover, aggregate demand should be
calculated on a local basis since it is unlikely that national or regional de-
mand estimates will accurately reflect local demand. Also, incinerator resi-
due will only be available locally.
Aggregate demand was calculated by using a cross-sectional model to esti-
mate demand. With this method, demand is calculated for one year and cor-
related to parameters that influence it. This model is deterministic in nature
and can theoretically be used to estimate aggregate demand for any area
(that is, nationally, state, county, city, etc.).
With aggregate production for bituminous pavements determined on a
state-by-state basis, factors that would influence demand were selected.
From previous study, it was known that population and aggregate demand
were significantly correlated. Other factors relating to highway construction
that were chosen include: highway expenditures for capital and mainte-
nance; highway mileage; number of registered vehicles; miles of travel; road
density; and land area.
These seven variables were then regressed against state aggregate demand
that resulted in predictive Eq 1.
A D = ($0-3274) (^7.6231) (p0.4212) (j^^-7.4072) (£)7.3560) (yO.2866) (J-0.1752) g-1.9456
(1)
where
AD = aggregate demand, in lOOO's of tons;
$ = expenditures for capital and maintenance, in lOOO's;
A = land area, in square miles;
P = population, lOOO's;
M = mileage;
D = road density, in miles/square miles
V = number of vehicles, in lOOO's;
T = travel miles, in billions; and
e = the natural logarithm base.
This equation was found to have a multiple coefficient of determination of

0.8312 and a standard error of estimate of eO-4292 Q^ + 1350 tonnes (1500
tons).
The single most important variable was expenditures for capital and main-
tenance. By itself, this variable accounts for approximately 75 percent of the
variation in aggregate demand. Overall, Eq 1 explains at least 83 percent of
the variation in aggregate demand. Problems in regard to multicovariant lin-
earity among the independent variables were assumed to be insignificant since
it was desired to obtain as high a value of r^ as possible.
Although Eq 1 would have been used for estimating demand, it was found
that data for some of the variables could not be obtained at the SMSA level.
Consequently, Eq 1 was modified to Eq 2.
A D = (SO'tS^S) (^0.2695) (pO.3463) g-2.2273 (2)
Although somewhat lower than Eq 1, Eq 2 still had a high multiple coeffi-

cient of determination of 0.8127 and a standard error of estimate of ± 1399
metric tonnes (1541 tons). Equation 2 was used to determine both current
and projected aggregate demand.
In order to determine economic competitiveness of incinerator residue with
virgin aggregate, the aggregate demand was determined as part of the five
SMSA case studies using Eq 2. Table 1 presents the findings of this research.
TABLE 1—Incinerator residue as a percent of total aggregate demand.
1975 Annual Incinerator Residue as Part of

Aggregate Demand, Residue Output, Total Aggregate
SMSA Tons" Tons" Demand, %
Chicago, 111. 9 310 000 154 560 1.7

Harrisburg, Pa. 1 134 000 32 760 2.9
Miami, Fla. 2 407 000 17 500 0.7
New York, N.Y.-N.J. 8 147 000 625 000 7.7
(6 incinerators)
Washington, D.C. -Md.-Va. 4 581 000 65 000 1.4
" Metric equivalence: 1.1023 tons = 1 tonne.
Study Methodology
SMSA Case Study Selection Criteria

The selection of Chicago, Miami, Harrisburg, New York, and Washing-
ton, D.C. as the SMSA's to be studied in-depth was done in conjunction with
FHWA personnel and based on the following criteria as well as other factors:
1. availability of incinerator residue,
2. quality of incinerator residue,
3. number of locations of landfills,
4. availability and type of aggregate,
5. number of asphalt plants, and
6. asphalt demand.
By relying on data generated in the development of the incinerator list and
the aggregate demand figures, the most likely candidates for further study
were selected.
Development of the Mathematical Model

In order to evaluate the comparative economic advantage of virgin aggre-
gate and incinerator residue, a cost analysis model was developed. For each
of the five SMSA's, a comparative cost analysis was performed for two
models: (a) Model 1—virgin aggregate and incinerator residue used as sub-
base, and {b) Model 2—virgin aggregate and incinerator residue used as bi-
tuminous aggregate.
Sub-models were developed for Models 1 and 2 based on the amount of
residue that is required to provide equivalent road properties and perfor-
mance to virgin aggregate. For Model 1, the sub-models were designated as
K = \,2, and 3, signifying, when K = 2, for example, that 2 tons of residue
are required to give comparable performance in road use to 1 ton of virgin
aggregate. Only two sub-models were developed for Model 2, A" = 1 and 2.
For each of the five SMSA's the costs of processing incinerator residue to
subbase and bituminous aggregate quality as well as the transportation cost
in each area were gathered. From this data, cost equations were generated.
These equations specified the FOB, job site and FOB, asphalt plant cost of
virgin aggregate, incinerator residue used as subbase, and incinerator resi-
due used as bituminous aggregate. For Model 1 the general form of the cost
equations were
C^ — Pj + Cjg + Cj • djj
where
C^ = the cost per ton of incinerator residue ready to use at the job site,
FOB job site;
Pj = the price per ton of incinerator residue, FOB incinerator;
Cj^ = the cost per ton to improve incinerator residue to acceptable high-
way quality subbase aggregate,
Cj = the cost per ton mile of transporting residue; and
d^ — the distance from incinerator to a hypothetical job site.
and
Qf Q Q ' QJ
where
CQJ = the cost per ton of virgin aggregate ready to use at the job site, FOB
job site;
PQ = the price per ton of virgin aggregate, FOB quarry;
CQ = the cost per ton mile of transporting virgin aggregate; and
CIQJ = the distance from quarry to a hypothetical job site.
For Model 2, the general form of the equation was
Qp = Pi+ CiB + Ci • dip

where
Cjp = the cost per ton of incinerator residue ready to be used at an as-
phalt plant, FOB asphalt plant;
PJ = the same as in Model 1;
Cjg = the cost per ton to improve incinerator residue to acceptable high-
way quality bituminous base aggregate;
CJ = the same as Model 1; and
djp = the distance from incinerator to hypothetical asphalt plants.
and
CQP — PQ + CQ • dQp
where
CQP = the cost per ton of virgin aggregate ready to be used at the asphalt
plant, FOB asphalt plant;
PQ = the same as Model 1;
CQ = the same as Model 1; and
dgp = the distance from quarry to hypothetical asphalt plants.
For each of these models, the relative cost, FOB end-use point, between
residue and virgin aggregate was determined by calculating the expressions
K(C]j) =CQJ for Model 1
K(Cip) I CQP for Model 2
In Model 1, incinerator residue as subbase, C^, is multiplied by 2 and 3 for

sub-models K = 2 and 3, respectively, because the sub-models specify that 2
or 3 tons of residue are required to provide a thickness of subbase having
equivalent properties to one ton of virgin aggregate in subbase use. In Model
2, unfused incinerator residue as bituminous aggregate, Cjp, is multiplied by
2 for sub-model K = 2 because under this sub-model, 2 tons of residue will
be equivalent to one ton of virgin aggregate in bituminous use.
In each of the five SMSA's, cost information was gathered or developed for
each of the variables of the cost equations. For each SMSA, specific equa-
tions were then generated. The C^, cost FOB job site, and Cjp, cost FOB as-
phalt plant, equations were then solved for various distances from the incin-
erator. Once these FOB job site or asphalt plant costs were determined, the
costs were substituted for CQJ or C^p and the CQJ and CQP equations solved
for dQj, distance from quarry to job site, and d-Qp, distance from quarry to as-
phalt plant.
The dQj and dQp show the distances that job sites and asphalt plants must
be from actual quarry locations for the cost of residue, FOB job site or as-
phalt plant, and cost of virgin aggregate, FOB job site or asphalt plant, to be
equal. In other words each djj and djp determines an FOB destination cost
and the cost, in turn, determines the dQj and dQp. When this mathematical
solution is plotted on a map for a particular quarry location, it describes a ge-
ographical boundary where residue and virgin, aggregate costs are equal. In
those areas, on one side of the boundary, residue is cheaper than virgin ag-
gregate, FOB destination, and on the other side of the boundary, virgin ag-
gregate is cheaper than residue, FOB destination. When the analysis is
performed for all actual quarry locations in a 48.28 km (30-mile) radius of an
incinerator, a locus of boundaries is generated where, on one side, residue is
more economically feasible to use, and, on the other side, virgin aggregate is
more economically feasible to use. This comparative cost analysis was per-
formed for each of the sub-models K = 1,2, and 3.
SMSA Case Studies

Using the analytical methodologies discussed in the previous section, a
case study was conducted in each of five selected SMSAs. A capsulatized ver-
sion of each of these case studies follows.
Chicago, III.
Solid Waste Management Practices—The incinerator residue source stud-
ied in the Chicago SMSA was Chicago's Northwest incinerator plant. The fa-
cility is a reciprocating grate, waterwall type unit opened in 1972. It has four
furnace units and a refuse capacity of 1 440 metric tons per day (TPD) (1600
TPD). The plant has a steam generating capacity of 199 580 kg (440 000 lb)
per hour and the city has plans to sell this steam as cooking heat to a nearby
bakery.
The plant produces approximately 360 metric TPD (400 TPD) of incinera-
tor residue. Some post screening of the residue would be required if it were to
be used as aggregate. The residue from the plant is currently being landfilled
at Steam's Quarry, a municipal owned facility 12.8 km (8 miles) from the
northwest plant. The previous experience with the use of incinerator residue
in road construction in the Chicago SMSA involved its use as subbase mate-
rial for a high school parking lot.
Aggregate Sources—There are no quarries currently operating within the
City of Chicago. The closest ones to the incinerator are located in McCook,
111., approximately 19 km (12 miles) to the southwest and Thornton, 111., 38
km (24 miles) due south.
Cost Analysis—The data was developed from current real-world experi-
ence in the Chicago SMSA.
As previously discussed, the equations to determine where incinerator resi-
due is less expensive are as follows:
Subbase
K(Pi + Cjs + C,X djj) < (PQ + CQX dQj)
Bituminous Mix
K{Pj + CiB + Q X dip) < (PQ + CQX dQp)
Where Q is the cost of transportation of incinerator residue in dollars per

ton-mile:
C/ = first 10 miles = $2.06/ton or $0.21 ton/mile every additional 8 km
(5 miles) = $0.38/ton or $0.076/ton mile, or
Cj = $2.06/ton + $0,076 (d - 10) where c? > 16 km (10 miles).
Where CQ is the cost of transportation of quarried material in dollars per

ton-mile:
CQ = 9.6 km (6 miles) = $1.67 to $1.70 ton every additional mile =

$0.20/ton, or
CQ = SI.70 + $0.20 (d - 6) where d > 9.6 km (6 miles).
Where PQ is the price of quarried material per ton at the quarry site:
PQ = $4.00/ton.
Pj, the price of incinerator residue per ton at the incinerator, can be said to
be $0, assuming the municipality's desire to lessen the load on their landfill
facility by giving the residue away free.
Now that the costs of the materials and their transportation to any given
job site have been determined, the cost to improve the incinerator residue to
acceptable highway quality for use as subbase must be determined.
Cj^, the cost of improving incinerator residue to acceptable highway qual-
ity for use as subbase, would include the following elements:
Cj^ = cost of (residue + 6% lime) + pugmill processing cost
= 0 + 0.06 ($30/ton) + ($0.50/yd^)
= 0 + $1.80/ton + residue ($0.50/yd3 ^ 1350 Ib/yd^) X 2000 lb/ton
= 0 + $1.80/ton + ($0.74/ton)
= $2.54/ton
The pugmill processing of the residue is required to bring it up to highway

construction quality and presumes a cost of $0.50 per cubic yard where the
incinerator residue from Chicago's northwest incinerator has an average un-
compacted density of approximately 796 kg/m-' (1350 lb/yd-'). For each ton of
subbase produced, 6% of the mix will be lime and 94 percent will be residue.
The pugmill processing costs are applied to the total weight of the residue-
lime mix. The current cost of hydrated pulverized lime in the Chicago SMSA
is $30.00 per ton.
To determine the cost to improve the incinerator residue to acceptable
highway quality for use as bituminous mix, Cjg, the following elements must
be considered.
Cjg — cost of processed residue + 2% asphalt cement
= $2.54/ton + 0.02 ($90.00/ton)
= $2.54/ton + $1.80/ton
= $3.34/ton
The addition of 2 percent of asphalt cement presumes that the incinerator
residue, due to its greater specific surface, would require more asphalt than
virgin quarried material. Asphalt cement is currently selling for $90.00 per
ton in the Chicago SMSA.
In order to compare the viability of incinerator residue with quarried ma-
terial, an equivalent weight ratio that represents the balance between the two
aggregate materials is required. This ratio, expressed as K, was applied to
three different scenarios based upon the ainbunt of incinerator residue per
unit of quarried material required to provide equivalent road properties and
performance K is equal to 1, 2, and 3 in the scenarios.
Based on the costs just determined, our Model \, K = \ submodel is
C„ = $2.54 + $2.06 + $0.076(d - 10)ford > 10
CQJ = $4.00 + $1.7 + $0.2 ( d ' - 6) for d ' > 6
For Model 1, AT = 2
C^ = 2 [$2.54 + $2.06 + $0,076 (d - 10)] ford > 10
CQJ = $4.00 + $1.7 + $0.2 id' - 6) for d ' > 6
For K = 3, Cjj is multiplied by 3.

Model 2, K = I equations are
Cjp = $3.34 + $2.06 + $0.076{d - 10)ford > 10
CQP = $4.00 + $1.7 + $0.2 id' - 6) for d ' > 6
For K = 2, Cjp is multiplied by 2.

Results of Analysis—The results of this analysis are shown in Table 2.
TABLE 2—Percent of SMSA where incinerator residue is cheaper

than quarried material.
Model 1 Model 2
Residue as Subbase, % Residue as Bituminous Mix, %
K = 1 79.7 70.9
K = 2 4.3 0
K = 3 0 0
Miami, Fla.
Solid Waste Management Practices—Dade County currently operates an
incinerator with a design capacity of 270 metric TPD (300 TPD) located in
North Miami. According to county solid waste management personnel, the
incinerator currently receives approximately 202.5 metric TPD (225 TPD),
an annual volume of approximately 58,5(X) metric tons per year (TPY) (65 OCX)
TPY). It operates 24 h a day, 5 days a week, and trucks the residue to a land-
fill 11.2 km (7 miles) from the site.
Incinerator residue from Miami's plant has been used by county personnel
as road material at their north landfill.
Aggregate Sources—Aggregate is mined at a number of quarry sites in the
area west of the developed section of Dade County. Lignite limerock is dug
for use in both portland cement and bituminous construction. The aggregate
is usually excavated from marshy areas and consequently its wetness can re-
quire more drying than conventional aggregate.
Cost Analysis
C, = $0.40 + S0.i2d,d > 1
Where base price = $0.40, plus $0.07 ton/mile, plus $0.02 per light or $0.06
ton/mile (assuming three traffic lights per mile). The traffic light cost repre-
sents a starting and stopping fee initiated in 1950 in south Florida.
CQ = $0.40 + SO.Ud', d' > 1
Where base price = $0.40, plus $0.07 ton/mile, plus $0.02 per light or $0.06
ton/mile (assuming three traffic lights per mile).
PQ = $2.45/ton
Pi= $0
Cj^, the cost of improving the incinerator residue to acceptable highway

quality for use as subbase is
Cjg = cost of (residue + 6% lime) -f pugmill processing cost
= 0 + 0.06 ($70.00/ton) + ($0.50 yd^)
= 0 + $4.20/ton + ($50/yd3 -i- 1350 Ib/yd^) X 2000 Ib/yd^
= 0 + $4.20/ton + ($0.74/ton)
= 0 + $4.20/ton + ($0.74/ton)
= $4.94/ton
Cig — processed residue + 2% asphalt
= $4.94/ton (with landfill savings) + 0.02 ($87.50/ton)
= $4.94/ton + $1.75/ton
= $6.69/ton
Residue for use as bituminous mix aggregate would not be economically

feasible for any asphalt plants in the area for both sub-models K = \ and
K = 2. The percentage of the Miami, Fla. SMSA where incinerator residue
is cheaper than quarried material is shown in Table 3.
TABLE 3—Percent of SMSA where incinerator residue is cheaper than

quarried material.
Model 1 Model 2
K = 1 <l <]
K = 2 <1 <1
K = 3 0 0
Hatrisburg, Pa.
Solid Waste Management Practices—Harrisburg Incinerator has a design
capacity of 648 metric TPD (720 TPD) with its two 324 metric TPD (360
TPD) boilers. They are currently receiving between 324 metric TPD (360
TPD) and 360 metric TPD (400 TPD), seasonally adjusted. It operates 24 h a
day, 7 days a week. The residue is stored on a four-acre site adjacent to the
incinerator.
The city of Harrisburg had some previous experience with the use of resi-
due from its incinerator. A 54 m (180 ft) long strip of bituminous course con-
taining incinerator residue was laid as part of a resurfacing project on Route
22 in Dauphin County, north of the city of Harrisburg, between Interstate 81
and the Rockville Railroad Bridge. Harrisburg also installed an experimental
wearing surface section using a blend of natural aggregate and incinerator
residue on Wayne Street, a residential street in the southern section of the
city.
Aggregate Sources—There are several vertically-integrated aggregate sup-

pliers located in the city of Harrisburg itself and across the Susquehanna
River in Cumberland County. There are currently 12 asphalt plant opera-
tions in the Harrisburg SMSA.
Cost Analysis
Cj, cost of transportation of incinerator residue in dollars per ton mile =
$0.20 per ton mile.
CQ, cost of transportation of quarried material in dollars per ton mile =
first 8 km (5 miles) = $1.00/ton; after 8 km (5 miles) = $0.35/ton +
$0.07/mile
CQ = $1.00 + [$0.07 id- 5) + 0.35] where c/ > 8 km (5 miles)
PQ = $3.85/ton in Harrisburg SMSA
Pi = $0
Cf^, cost of improving the incinerator residue to acceptable highway qual-

ity for use as subbase is:
Cj^ = cost of (residue + 6% lime) + pugmill processing cost
= 0 + 0.06 ($50.00/ton) + ($0.50/yd3)
= 0 + $3.00/ton + (0.50/yd3 ^ 1350 Ib/yd^) X 2000 Ib/yd^
= 0 + $3.00/ton + ($0.74/ton)
= 0 + $3.00/ton + $0.74/ton
= $3.74/ton
Hydrated pulverized lime FOB at the processing site is $50.00 per ton.
Cjg, the cost of improving the incinerator residue to acceptable quality for
use as bituminous mix is:
Cjg = processed residue + 2% asphalt
= $3.74/ton + 0.02 ($90.00/ton)
= $3.74/ton + $1.80/ton
= $5.54/ton
The cost of asphalt cement is currently $90.00 per ton in the Harrisburg
area.
K, the equivalent weight ratio representing the balance between the two
aggregates, remains 1, 2, and 3.
Based on the costs just determined, our Model 1, if = 1 equations are
Cu = $3.74 + $0.20 5
For K = 2 and 3, C^ is multiplied by 2 and 3, respectively.

Model 2, K = 1 equations are
Cjp = $5.43 + $0.20d
CQP = $3.85 + $1.00 + [$0.07 W - 5) + $0.35] if d ' > 5
For if = 2, CJP is multiplied by 2.


quarried material.
Model 1 Model 2
K = 1 10.7 0
K = 2 0 0
K = 3 0 0
Washington, D.C.-Md.-Va.
Solid Waste Management Practices—Solid Waste Reduction Center #1
(SWRC #1) is a $21 million complex with a design capacity for waste disposal
of 1350 metric TPD (1500 TPD). It operates 24 h a day, 5 days a week. The
plant has six furnaces, each with a waste capacity of 225 metric TPD (250
TPD); however, it currently receives approximately 900 metric TPD (1000
TPD) of waste. The residue from SWRC #1 is currently being used in a land
reclamation project at a Federal health facility approximately 12.8 km (8
miles) from the incinerator.
In June of 1977, a 114 mm (4V2 in.) bituminous pavement base composed
largely of incinerator residue was placed on 14th Street in the southeast sec-
tion of Washington.
Aggregate Sources—Aggregate for highway construction is provided from
a number of quarrying sites located north and west of the District of Colum-
bia in Fairfax and Prince William Counties in Virginia and Montgomery
County in Maryland. Sand and gravel for this type of construction also comes
from north of the District, in Maryland.
Cost Analysis
Ci = $0.12 ton/mile, where cf > 10
CQ = $0.12 ton/mile, where c? > 10
PQ = $4.00/ton
Pj= SO
Cjs, the cost to improve the incinerator residue to acceptable highway

quality for use as subbase, is
CJS = cost of (residue + 6% lime) + pugmill processing
= 0 + 0.06 ($28/ton) + ($0.50/yd3)
= 0 + $1.68/ton + (0.50/yd3 - 1350 Ib/yd^) X 2000 Ib/yd^
= 0 + $1.68/ton + ($0.74/ton)
= 0 + $1.68/ton + $0.74/ton
= $2.42/ton
The price of hydrated pulverized lime in the Washington, D.C. area is cur-
rently approximately 328.00 per ton. A density of 1350 lb/yd'' is assumed for
the uncombined incinerator residue from SWRC #1.
CjB, the cost to improve the incinerator residue to acceptable quality for
use as bituminous mix in highway construction, following the standard meth-
odology, is
CjB = processed residue + 2% asphalt
= $2.42/ton + 0.02 ($100/ton)
= $2.42/ton + $2.00/ton
= $4.42/ton
The cost of asphalt cement, according to the National Asphalt Pavement

Association, in the Washington, D.C. area is currently $100 per ton, FOB at
the asphalt plant.
K, the equivalent weight ratio representing the balance between the two
aggregates, remains 1, 2, and 3.
Based on the costs developed above our Model \, K = \ equations are
Cj7=$2.42 + $0.12i/
CQJ = $4.00 + $0.12i/
For K = 2 and 3, C^ is multiplied by 2 and 3, respectively.

Model 2, K = 1 equations are
Cjp = $4.42 + $0.12^
C Q ^ = $4.00+ $0.12J
For K = 2, CJP is multiplied by 2.

Results of Analysis—When AT = 2 for Model 2, it would not be economi-
cally feasible to use residue in the Washington, D.C. area. The percentage of
area within the Washington, D.C.-Md.-Va. SMSA, where the use of inciner-
ator residue is cheaper, is shown in Table 5.

quarried material.
Model 1 Model 2
K = 1 65.4 45
K = 2 10,7 0
K = 3 0 0
New York, N.Y.-N.J.

Solid Waste Management Practices—There are currently six municipal in-
cinerators available to burn solid waste in the New York SMSA, operated by
the City of New York. Each of the city's six incinerators has four furnaces
and a design solid waste capacity of 900 metric TPD (1000 TPD), however,
due to constraints of the city's air pollution control regulations, only 12 fur-
naces can be operated throughout the system at any one time.
Aggregate Sources—Aggregate used in highway construction in New York
City is quarried north of the city and barged down the Hudson to job sites
and asphalt plants. Large quarry operations that supply the city are located
in Poughkeepsie, Kingston, Haverstraw, and West Nyack in New York; New

Haven in Connecticut; and in the Watchung Mountains in Northwestern
New Jersey. Sand and gravel is usually provided by pits located on Long Is-
land and in New Jersey.
Cost Analysis
Ci = $0.40/ton/mile
CQ = $3.50/ton
PQ = $5.65/ton
Pj= $0
Cj, the cost of improving the incinerator residue to acceptable highway

quality use for subbase is
Cjs = cost of (residue + 6% lime) + pugmill processing
= 0 + 0.06 ($60/ton) + ($0.50 yd^)
= 0 + $3.60/ton + ($0.50/yd3 H- 1350 Ib/yd^) X 2000 lb/yd
= 0 + $3.60/ton + ($0.74/ton)
^ 0 + $3.60/ton + $0.74/ton
= $4.34/ton
The cost of hydrated pulverized lime in the New York SMSA is currently
$60.00 ton FOB the processing site.
C[B, the cost to improve incinerator residue to acceptable quality for use as
bituminous mix in highway construction is
Cjg = processed residue + 2% asphalt
= $4.34/ton + 0.02 ($95/ton)
= $4.34/ton + ($1.90/ton)
= $6.24/ton
A price of $5.65/toii was used as representative of the cost of virgin aggre-

gate, FOB dockside, in the New York area. Based on the costs developed
here, Model 1, /iT = 1 equations are
Cu = $4.34 + $0.4(W
CQJ = $5.65 + $3.50
For K = 2 and 3, Cjj is multiplied by 2 and 3, respectively.

Model 2, K = \ equations are
îp $6.24 + $0.40d
CQP = $5.65 + $3.50
For K — 2, Cjp is multiplied by 2.


quarried material.
Model 1 Model 2
K = 1 100 47
K = 2 0 0
K = 3 0 0
Conclusions from SMSA Studies

A review of the five SMSA's studied shows a wide range in the economic
feasibility of using incinerator residue in highway constriiction. Table 7 pre-
sents a comparison of this economic feasibility for the three K values studied.
The comparison shows that, in the highly urbanized SMSA's studied, the
economic feasibility of incinerator residue when K = 1 \s good. Projections
can, therefore, be made that it is likely that incinerator residue will be eco-
nomically feasible in other urbanized SMSA's having incinerators.
A second conclusion possible from the SMSA studies is that the location of
the incinerator within the SMSA is important to the economic feasibility of
residue use. In the Miami SMSA, the location of the municipal incinerator in
the far northeast comer of the SMSA makes the economic use of the facility's
residue uneconomical in almost all of Dade County. In Chicago the area of
economic feasibility would include a large section of Lake Michigan within
TABLE 7—Comparison of economic feasibility of incinerator residue in the five SMSA's studied.
Percent Area Where Residue is Cheaper
Model 1 Model 2
Area Subbase, % Bituminous Mix,, %
Under Study,
SMSA square miles K = 1 K = 2 ^ = 3 K = 1 K = 2 K = 3
Chicago 1965 79.7 4.3 0 70.9 0 0

Miami 756 <1 0 0 0 0 0
Harrisburg 1398 10.7 0 0 0 0 0
New York 1047 100 0 0 47 0 0
Washington, 2035 65.4 10.7 0 45 0 0
D.C.
Weighted 57 4.2 0 39 0 0
average of all
five SMSA's
the 48.28 km (30 mile) radius. Harrisburg's plant is located less than a
quarter of a mile from one of the SMSA's largest aggregate producers, mak-
ing the competitive value of residue much less in the region.
In general, the studies performed on the five selected SMSA's show that in
highly urbanized areas that have incinerators, future consideration should be
given to the use of incinerator residue in highway construction. It is antici-
pated that the economic feasibility of this technology in this area will be high.
Summary and Conclusions
This study has examined the potential use of incinerator residue in bitumi-
nous highway construction. Findings from this study are:
1. There are currently 72 operating incinerators in the United States that
produce between 1 262 989 tonnes (1 403 301 tons) and 2 305 048 tonnes
(2 561 164 tons) per year of incinerator residue from municipal waste.
2. Incinerator residue supply can meet 1.5 percent of current and less than
1 percent of projected demand for aggregate in bituminous highway con-
struction.
3. The cost analysis model developed to evaluate the economic feasibility of
using incinerator residue as opposed to virgin aggregate has broad applica-
tion as an analytical model for evaluating the feasibility of other alternative
aggregate materials such as fly ash, slag, mine tailings, shredded tires, etc.
4. On a weighted average basis, the detailed five SMSA analysis shows that
residue used as subbase can be economically viable in 57 percent of the geo-
graphical areas for K = I, and 4.2 percent of the geographical areas for
K = 2.
5. Again, on a weighted average basis, for the application of residue as bi-

tuminous base material these values are 39 percent for ^ = 1, and zero for
K = 2.
6. An analysis of five SMSAs reveals that in highly urbanized areas the
economic feasibility of incinerator residue when K = 1 is good. Projections
can, therefore, be made that it is likely that incinerator residue will be eco-
nomically feasible in other urbanized SMSAs having incinerators.
R. P. Webster1 and J. J. Fontana1
Beneficiation of Natural Aggregates

by Polymer Impregnation
REFERENCE: Webster, R. P. and Fontana, J. J., "Beneficiation of Natural Aggregates

by Polymer Impregnation," Extending Aggregate Resources, ASTM STP 774, American
Society for Testing and Materials, 1982, pp. 64-73.
ABSTRACT: The use of polymer impregnation as a means of upgrading the physical and
mechanical properties of natural aggregates was evaluated in a series of tests performed
using four aggregates of varying quality. The strength of concrete cast with polymer im-
pregnated coarse aggregate was also tested. Two monomer systems were used in the in-
vestigation, a methyl methacrylate-based system and a styrene-based system.
In general, significant improvements in the physical and mechanical properties of each
of the four aggregates resulted from polymer impregnation. The strength of concrete cast
with impregnated aggregates varied, being increased in some cases and decreased in
others.
KEY WORDS: aggregates, marginal materials, beneficiation, polymers, polymer im-

pregnation, aggregate durability, concrete, cement
The supply of good quality natural aggregate in the United States is so vast
that it is estimated to be inexhaustible [1,2].2 However, because of geo-
graphic distribution, the quality and quantity of aggregate is not always suf-
ficient to meet the local needs in many areas of the country, requiring large
amounts of aggregate to be transported to these areas. Potential fuel short-
ages and steadily increasing transportation costs have heightened the need
to examine alternatives to shipping aggregates great distances. One alter-
native is to beneficiate locally available aggregate previously rejected because
of marginal or poor quality.
The results of a research program on the use of polymer impregnation as a
means of improving natural aggregates are presented in this paper.
'Materials research engineer and project chemist, respectively, Process Sciences Division,
Department of Energy and Environment, Brookhaven National Laboratory, Upton, N.Y.
11973.
2
The italic numbers in brackets refer to the list of references appended to this paper.
Copyright by ASTM Int'l (all 64 rights reserved); Thu Dec 22 0

WEBSTER AND FONTANA ON POLYMER IMPREGNATION 65
Procedure
Test Aggregates
Four aggregates of varying quality, two carbonate and two siliceous, were
used: Aggregate 09, a highly porous, poorly cemented limestone; Aggregate
23, a highly porous dolomite; Aggregate 43, a fine-grained volcanic lithic
arenite (sandstone); and Aggregate 48, a highly porous, poorly cemented
quartzite. These were selected, from a sampling of 40 aggregates collected
from various state highway departments across the United States, to repre-
sent a broad range of aggregates with widely varying physical and mechanical
properties. A brief petrographic description of each is given in the Appendix.
Impregnants
Test results are presented for aggregates impregnated with one of two
thermoplastic monomer systems, both of relatively low viscosity, ~ 15 mPa • s
(15 cP) at 24°C (75°F):
(a) 83 percent by weight methyl methacrylate (MMA), 5 percent by weight
trimethylolpropane trimethacrylate (TMPTMA), 12 percent by weight
poly(methyl methacrylate) (PMMA).
{b) 57 percent by weight styrene (Sty), 38 percent by weight TMPTMA, 5
percent by weight polystyrene (PSty).
The basic compositional structure of both monomer systems was selected
on the basis of the results of research performed at Brookhaven National
Laboratory in the area of polymer impregnated concrete. This research had
shown significant improvements in the strength and durability properties of
Portland cement concrete when impregnated with MMA-TMPTMA and Sty-
TMPTMA monomer systems [3-7]. The MMA and Sty serve as the base
monomer for their respective systems. The basic material characteristics of
each of the cured impregnant systems are primarily derived from the proper-
ties of the respective base polymers. The TMPTMA is a comonomer added to
each system to cross link the polymer, thus helping to improve the chemical
resistance and thermal stability of the cured impregnant. The TMPTMA
also helps to increase the rate of polymerization of each system. The PMMA
and PSty were added to each of the respective monomer systems to increase
the viscosity of the impregnant, and therefore help to eliminate excess
drainage of monomer from the aggregate prior to polymerization of the
system. In addition, one percent by weight silane coupling agent(7-metha-
cryloxypropyltrimethoxysilane) was added to each monomer system to im-
prove the bond between the polymer and the aggregate by developing a
chemical bond between the polymer and the silica present in the aggregate.
One percent by weight initiator (2-t-butylazo-2-cyanopropane) was added to
each monomer system to initiate the polymerization reaction.
A number of other impregnants have also been used to impregnate ag-

gregate including: epoxies, linseed oil, kraft lignin, sulfur, water-ethylene
glycol solutions, polyester-styrene, and chlorostyrene [8-10].
Impregnation Procedure
The procedure used to impregnate each of the aggregates was as follows.
The aggregate was washed to remove any dirt, clay, or other deleterious
materials from the surface, dried in an oven at 150°C (302°F) to constant
weight, and allowed to cool to room temperature, 24°C (75°F). It was then
put in a wire basket and placed inside the impregnation vessel. The vessel
was sealed and evacuated to ~ 760 mm Hg for a minimum of 60 min. Next,
the impregnant was introduced into the sealed vessel, and an overpressure of
103 kPa (15 psig) was applied for 4 h with nitrogen gas. Then the excess im-
pregnant was drained from the vessel, and the impregnant was cured in situ
by injecting steam, at approximately 112°C (234°F), into the vessel for a
minimum of 2 h.
It should be pointed out that while the impregnation procedure just
described is very suitable for impregnating small quantities of aggregate in
the laboratory, it would be necessary to modify the procedure in order for
large-scale field applications to be feasible. However, it is not the objective of
this paper to present an in-depth discussion of field impregnation pro-
cedures.
Evaluation of Impregnated Aggregates

The effectiveness of polymer impregnation as a means of improving natu-
ral aggregates was determined by evaluating various physical and mechanical
properties of the individual aggregates, and by testing the strength of con-
crete cast with impregnated coarse aggregate.
The following tests were used to evaluate the physical and mechanical
properties of the individual aggregates: ASTM Test for Specific Gravity and
Absorption of Coarse Aggregate (C 127-77), ASTM Test for Soundness of
Aggregates by Use of Sodium Sulfate or Magnesium Sulfate (C 88-76), and
ASTM Test for Resistance to Abrasion of Small Size Coarse Aggregate by
Use of the Los Angeles Machine (C 131-76).
The tests used to evaluate the strength properties of the concrete cast with
impregnated aggregate were as follows: ASTM Test for Compressive
Strength of Cylindrical Concrete Specimens (C 39-72), ASTM Test for Flex-
ural Strength of Concrete (Using Simple Beam with Third-Point Loading) (C
78-75), and ASTM Test for Splitting Tensile Strength of Cylindrical Con-
crete Specimens (C 496-71).
In addition, freeze-thaw tests were done on a limited number of cylinders
in accordance with the cycling procedure specified in ASTM Test for
Resistance of Concrete to Rapid Freezing and Thawing (C 666-77), Pro-

cedure A. Freeze-thaw durability was evaluated by determining the com-
pressive strength at 0, 25, and 50 cycles.
The concrete mix used, had a cement factor of 7.92 sacks/m-' (6.05
sacks/yd^), a water-cement ratio of 0.45, an air content of 5 ± 1 percent,
and a design slump of 7.6 to 10.2 cm (3 to 4 in.). The fine aggregate was a
graded silica sand with a fineness modulus of 2.42. The coarse aggregate was
soaked in tap water for 24 h before incorporation into the concrete mix. All
concrete test specimens were cured in saturated-lime water for 28 days before
being tested.
Compressive strength and splitting tensile strength tests were done on 7.6
by 15.2 cm (3 by 6 in.) cylinders, and flexure strength tests on 10.2 by 7.6 by
40.6 cm (4 by 3 by 16 in.) beams.
Results and Discussion
Physical and Mechanical Properties of Aggregates

The results of the tests done to evaluate the physical and mechanical prop-
erties of the aggregates, summarized in Table 1, generally show significant
improvements as a result of impregnation.
Specific Gravity and Water Absorption—The bulk specific gravity and
water absorption (ASTM C Vll-ll) of each aggregate were determined to
assist in evaluating the physical properties of impregnated aggregates. In
general, the bulk specific gravity increased and the water absorption de-
creased as a result of impregnation. The improvements in the bulk specific
gravity and water absorption within each aggregate group generally varied
directly with increasing polymer loading. In each case, the Sty (that is, Sty-
TMPTMA-PSty) impregnated aggregates exhibited higher polymer loadings
than did the MMA (that is, MMA-TMPTMA-PMMA) impregnated ag-
gregates.
Abrasion Resistance—Each aggregate was tested in the Los Angeles Abra-
sion Machine (ASTM C 131-76) to evaluate the resistance to wear due to
abrasion, crushing, and impact forces. In general, significant reductions in
abrasion loss were obtained for each of the impregnated aggregates, the most
notable being a reduction from 79 to 20 percent for Aggregate 48 im-
pregnated with the MMA system.
With one exception, the STY impregnated aggregates showed greater im-
provements in abrasion resistance than did the MMA impregnated ag-
gregates. The improvement in abrasion resistance within each aggregate
group increased with increasing polymer loading. The improvement noted in
the abrasion resistance is due principally to the formation of an interconnect-
ing network of polymer throughout the pore volume of the aggregate, which
serves to reinforce the aggregate.
|Jf •a
oôôovorjoo^^rsr-ô
1/5 S^
§tf B
S
O Sfi
"§ « "0, S
1
0« < ^^
^
^
g.
•1
•Q
^_,
§ §.'1'
•ts • ^ (U
a>
"cs £,?
1§ ^ 2 •"
^ ^#
la.
.§
ill
-lO
Iff
.1
;2 S? i3P
; rS o ts in o -^ •-
<
e ^' _• ^ ^
l-l
pa
<:
H (A
Sulfate Soundness—Sodium sulfate soundness tests (ASTM C 88-76) were

done to evaluate the resistance of the aggregates to simulated weathering.
Significant reductions in the sulfate soundness loss were obtained for each of
the impregnated aggregates, the most notable being a reduction from 79.2 to
3.7 percent, again for Aggregate 48 impregnated with the MMA system. The
reductions in sulfate soundness loss for each aggregate group tended to vary
directly with the reduction in the water absorption value for the aggregate.
As with abrasion resistance, the greatest reductions in sulfate soundness loss
were obtained with the Sty system.
Strength Properties of Concrete

The results of tests done to evaluate the strength of concrete cast with
unimpregnated and impregnated coarse aggregate are summarized in Table
2. Concrete test specimens were cast with each of the four test aggregates,
unimpregnated and impregnated with the MMA-based system, but with only
one Sty impregnated aggregate (because of supply limitations). The test
results for each aggregate group can be summarized as follows.
Aggregate 09—Two of the three properties measured to evaluate the
strength of concrete specimens were improved by using the MMA im-
pregnated aggregates: the flexure strength was increased by 2 percent and
the splitting tensile strength by 11 percent. The compressive strength,
however, was reduced by 16 percent.
Aggregate 23—Again two of the three properties were improved by using
the MMA impregnated aggregate: the flexure strength increased by 7 per-
cent and the splitting tensile strength increased by 5 percent. The com-
pressive strength decreased by 16 percent.
Aggregate 43—Two of the three properties were reduced by using the
MMA impregnated aggregate: the compressive strength by 19 percent and
the splitting tensile strength by 14 percent. The flexure strength was in-
creased by 4 percent.
Aggregate 48—All three of the properties were improved by using either
the MMA impregnated or the Sty impregnated aggregate. The concrete cast
with the MMA impregnated aggregate exhibited a 9 percent increase in com-
pressive strength, a 6 percent increase in flexure strength, and a 25 percent
increase in splitting tensile strength. For the concrete cast using the Sty impreg-
nated aggregate, these increases were 12, 13, and 27 percent, respectively.
The reductions in strength of three of the five concrete mixes cast with the
impregnated aggregates may be due partly to the presence of residual polymer
on the surface of the aggregate, which could have adversely affected the bond
between the matrix and the aggregate. The strength reductions may also be due
partly to the lack of modification of the concrete mix proportions for each ag-
gregate group to take into account the differences in the specific gravity and
water absorption between unimpregnated and impregnated aggregates; the
.a c a<
(N fS fW ro' (N <N <N o j ro
V5
O O Q O O O O O O
3 t«£'
X c -
(N fO •^' "^ (N (N r^* rr -"tf
SQ O O O O O O Q
0- 5>
e£2i2 (N-nroOrOi-Ô'-HOq
i n i-H CO 00 <N 00 r o i/i i/j
IS? ><
•?^
IP O fO O • s
O (N O ^ --^
tl
m
< < <; < o
I
a
<
u ill
concrete mix proportions were based on the properties of the unimpregnated

aggregates. It is not readily apparent, however, why concrete cast with a par-
ticular impregnated aggregate would exhibit increases in some strength prop-
erties and reductions in others.
The results of the freeze-thaw tests done to evaluate the durability of the con-
crete specimens cast with unimpregnated and impregnated aggregates are
summarized in Table 3. The tests were done on concrete cast with Aggregate 48
impregnated with the MM A system and with the Sty system. In general, the
concrete cast with the unimpregnated aggregate exhibited a 13 percent reduc-
tion in compressive strength after 50 cycles of freeze-thaw. The concretes cast
with the impregnated aggregates exhibited essentially no reductions in
strength. This improved freeze-thaw durability may be due to the improved
strength of the individual aggregates resulting from the interconnecting net-
work of polymer developed throughout the pore system. This internal reinforc-
ing system strengthens the aggregate enough to let it withstand the hydraulic
pressures developed within the pore system by the water being expelled during
freezing. This prevents the aggregate from fracturing and thereby weakening
the concrete.
Conclusions
The test data presented in this paper indicates that polymer impregnation
can be used to improve the physical and mechanical properties of natural ag-
gregates, and that, in some cases, the strength and durability properties of
concrete cast with impregnated aggregate are superior to those of concrete
cast with unimpregnated aggregate. However, the research described here
was of a very limited nature and further research is needed before the poten-
tial benefits of polymer impregnation can be fully understood and utilized.
TABLE 3—Freeze-thaw test results for concrete cast using Aggregate 48.
Compressive Strength, MPa (psi)
Polymer Number of Freeze-Thaw Cycles

Loading,
Impregnant % by weight 0 25 , 50
a 0 23.0 (3340) 20.8 (3020) 19.9 (2890)
MMA' 1.72 25.1 (3650) 24.7 (3580) 25.5 (3700)
Sty"^ 1.73 25.8 (3740) 24.6 (3570) 25.6 (3720)
"Unimpregnated control.
''83% by weight MMA, 5% by weight TMPTMA, 12% by weight PMMA.
"^57% by weight Sty, 38% by weight TMPTMA, 5% by weight PSty.
Acknowledgments
This research was performed as part of a larger study entitled,
"Technology for Improvement of Wear Resistance Properties of Aggregate
by Materials Impregnation," sponsored by the Federal Highway Administra-
tion under auspices of the Department of Energy under Contract No. DE-
AC-02-76CH00016.
APPENDIX
Petrographic Descriptions of Test Aggiregates
Aggregate 09, macroscopically, is a white porous limestone. The bulk of the rock
consists of round (in cross section) and ovoid, polycrystalline particles, typically 0.5
mm in diameter, composed of submicroscopic (<0.01 mm) calcite grains. Lack of
concentric zoning in most particles suggests detrital fecal pellets. Concentric zoning in
some indicates a small proportion of oolites. Fossils and fragments of fossils are also
present. Fossils consisting of single calcite grains up to 0.5 mm in diameter are rare.
The detrital material (pellets, oolites, and fossils) is only partially cemented together
by anhedral calcite grains averaging 0.02 mm in diameter. Much of the pore space is
unfilled.
Aggregate 23 is a homogeneous, reddish brown, porous dolomite. The texture and
mineralogy are uniform throughout the section. Euhedral dolomite rhombs are par-
tially cemented by concentrically layered overgrowths. Cores are cloudy with minute,
reddish brown, opaque inclusions. Rims are clear and in optical continuity with the
cores. Interstices between rhombs are not entirely filled, and there is significant pore
space. Many rhombs show partial to complete replacement of their cores by fine-
grained, polycrystalline aggregates of quartz. The replacement of cores but not
overgrowths suggests that quartz replacement occurred before formation of the
overgrowths.
Aggregate 43 is a fine-grained volcanic lithic arenite (sandstone). Arenite indicates
that it is composed dominantly of sand sized grains; lithic indicates that a significant
proportion of these grains are fragments of other rocks rather than individual
minerals; volcanic indicates that most of the detritus is probably volcanic in origin.
This is indicated by the presence of fine-grained glass, well-developed zoning in some
plagioclase, embayments on some quartz and plagioclase pumice, spherulitic felsic
material, and fragments of nearly aphanitic volcanic rocks. Those characteristics
taken together with the textural immaturity and the mineralogy, indicate that this
rock represents a first cycle reworking of material derived from a high Si02 volcanic
terrain.
Aggregate 48 is a moderately well-sorted, porous quartzite. Detrital grains are
mostly rounded to angular single crystals of quartz with minor fragments of feldspars,
muscovite, biotite, and chlorite. Quartz grains are commonly strained, some showing
advanced stages of polygonization. Optically continuous overgrowths are present on
many quartz grains. Alkali feldspars are highly altered to sericite. Grain boundaries
are mostly tangential or point contacts, though sutured contacts are not uncommon.
The porosity is high but difficult to estimate accurately because grains are plucked
during cutting and polishing of the thin section. Individual grains are extremely hard,
but they are only loosely bound together.
References
[1] Marek, C. R. and Jones, T. R., Jr., "Will There Be a Materials Crisis?," Optimizing the
Uses of Materials and Energy in Transportation Construction, TRB Special Report 166,
Transportation Research Board, 1976, pp. 18-28.
[2] Miller, R. H. and Collins, R. J., "Waste Materials as Potential Replacements for Highway
Aggregates," NCHRP Report 166, National Cooperative Highway Research Program,
1976.
[3] Steinberg, M. et al., "Concrete-Polymer Materials, First Topical Report," BNL 50134
(T-509), Brookhaven National Laboratory, and USER General Report 41, U.S. Bureau of
Reclamation, Dec. 1968.
[4] Steinberg, M. et al, "Concrete-Polymer Materials, Second Topical Report," BNL 50218
(T-560) and USBR Rep. REC-CCE 70-1, Dec. 1969.
[5] Dikeou, J. T. et al., "Concrete-Polymer Materials, Third Topical Report," Bureau of
Reclamation Report REC-ERC-71-6 and BNL 50275 (T-602), Jan. 1971.
[6] Dikeou, J. T. et al, "Concrete-Polymer Materials, Fourth Topical Report," Bureau of
Reclamation Report REC-ERC-72-10 and BNL 50328, Jan. 1972.
[7] Kukacka, L. E. et al, "Concrete-Polymer Materials, Fifth Topical Report," BNL 50390,
Brookhaven National Laboratory, and USBR Rep. REC-ERC-73-12, U.S. Bureau of
Reclamation, Dec. 1973.
[8\ Cady, P. D., Blakenhom, P. R. et al, "Upgrading of Low-Quality Aggregates for PCC and
Bituminous Pavements," NCHRP Report 207, 1979.
[9] Lin, C. H., Walker, R. D., and Payne, W. W. m Living with Marginal Aggregates, ASTM
STP 597, American Society for Testing and Materials, 1976, pp. 76-84.
[10] Steinberg, M. and Colombo, P., "Preliminary Survey of Polymer Impregnated Stone,"
BNL 50255, Sept. 1970.
Chen Chang,1 R. G. Hicks,1 and T. S. Vinson1
Beneficiation of Marginal Coastal

Oregon Aggregates with Cement
REFERENCE: Chang, Chen, Hicks, R. G., and Vinson, T. S., "Beneficiation of Mar-
ginal Coastal Oregon Aggregates with Cement," Extending Aggregate Resources, ASTM
ABSTRACT: Coastal Oregon has a severe shortage of quality construction aggregates.

There is, however, an abundant supply of lower quality, or marginal, aggregates. In an
attempt to use marginal aggregates, the feasibility of beneficiation with cement has been
investigated. Specifically, the optimum cement contents to stabilize three marginal ag-
gregates (Tyee sandstone, marine basalt, dredge spoil sand) were determined. The
resilient modulus and fatigue life characteristics were measured for the aggregates
stabilized at the optimum cement contents. The influence of moist curing time and wet-
ting and drying or freezing and thawing cycling on the properties measured were
evaluated. The results indicated the resilient moduli of all the materials tested increased
with moist curing time. The resilient moduli of the stabilized aggregates were not affected
by freezing and thawing cycling, however, the moduli were reduced after wetting and dry-
ing cycling. The fatigue life characteristics, expressed in terms of tensile stress versus
number of load repetitions to failure, were not significantly different for the moist cured
stabilized materials compared to the stabilized materials subjected to wetting and drying
or freezing and thawing cycling. The resilient moduli and fatigue life properties
measured were used to arrive at typical layer thicknesses and layer equivalencies for road
sections employing the cement stabilized marginal aggregates. The results indicate
beneficiation of marginal aggregates with cement is economically justifiable.
KEY WORDS: aggregates, cement stabilization, durability, fatigue life, layer equiva-
lencies, concrete, cement, fatigue (materials)
Along the Oregon Coast the demand for quality aggregate for road con-
struction far exceeds the available supply. Two primary factors have con-
tributed to this situation, namely, environmental legislation that limits the
development of new aggregate resource areas and rigid specifications for ag-
gregate for use in construction. Owing to the expanding population of
Oregon, the aggregate shortage will become more acute in the future [l].2
'Graduate research assistant, professor, and associate professor, respectively, Department of

Civil Engineering, Oregon State University, Corvallis, Ore. 97331.
2
Copyright by ASTM Int'l (all 74 rights reserved); Thu Dec 22

Indian Institute of Technology Kharagpur pursuant to License Agreement
CHANG ET AL ON MARGINAL COASTAL OREGON AGGREGATES 75
Currently, the solution to supplying Coastal Oregon with aggregate is to

import quality materials from outlying areas as far away as 100 km (60
miles). As the extraction and transportation costs of impo;led quality ag-
gregates increase, it will be necessary to investigate techniques to upgrade the
abundant supply of lower quality or marginal aggregates found near the
coast. In recognition of this situation, a study was conducted to investigate
the feasibility of using cement as a stabilizing agent for marginal coastal ag-
gregates. The results from the study are described herein. The scope of work
includes: (1) the evaluation of resilient properties and fatigue life characteris-
tics for three marginal and one quality aggregate found along the Oregon
coast, (2) the determination of typical layer thicknesses for road sections
employing the cement stabilized marginal aggregates, and (3) the determina-
tion of layer equivalencies, based upon the characteristics of the quality ag-
gregate selected for the study.
Material Chaiacteristics
Aggregate Types
Four aggregates were selected for the study as follows: (1) Eckman Creek
marine basalt, (2) Tyee sandstone, (3) dredge spoil sand, and (4) Oceanlake
marine basalt. Eckman Creek marine basalt is a marginal aggregate owing
primarily to its potential for degradation in the presence of water [2,3]. Tyee
sandstone is an interbedded sandstone and siltstone of the Tyee Formation.
The aggregates derived from this formation have high mechanical degrada-
tion potential. Dredge spoil sand is a clean, uniform fine sand with inter-
mixed shell fragments. Oceanlake marine basalt is a high quality aggregate.
It was chosen for the study for performance comparison with the other
treated materials.
Aggregate Processing and Gradation

The processing operations varied with aggregate type. Block samples of
the Tyee sandstone were obtained from the field. The samples were broken
down with a sledge hammer, then processed through a crusher to obtain
material sizes smaller than 25 mm (1.0 in.). The Eckman Creek and
Oceanlake marine basalt were crushed at a commercial processing plant
before receipt. The dredge spoil sand was used as received. The gradations
for the materials used in the study are presented in Table 1.
Aggregate Properties
Standard indicator and degradation tests were performed on the ag-
gregates. The results of the standard indicator tests, presented in Table 2, in-
TABLE 1—Aggregate gradings.
Percentage Passing
Marine Basalts
Tyee Dredge Spoil
Sieve Size Oceanlake Eckman Creek Sandstone Sand
19.0 mm (^4 in.) 100 100 100 100

12.5 mm (Vl in.) 100 100 93 96
9.5 mm (% in.) 86 86 89 93
6.3 mm (^4 in.) 67 67 81 87
4.75 mm (No. 4) 55 55 75 85
2.00 mm (No. 10) 40 40 60 77
425 Mm (No. 40) 18 18 39 62
75 iim (No. 200) 9 9 15 1
TABLE 2—Aggregate properties—standard indicator test results.
Specific
Gravity, Absorption, Liquid Plastic AASHTO*
Aggregate SSD" % Limit Limit Classification
Oceanlake
marine basalt 2.88 1.0 23 5.3 A-l-a
Eckman Creek
marine basalt 2.68 5.0 20 2.9 A-l-a
Tyee sandstone 2.39 10.0 27 4.2 A-l-b
Dredge spoil sand 2.25 1.4 nonplastic A-3
°SSD = saturated surface dry.

*AASHTO = American Association of State Highway and Transportation Officials.
elude ASTM Test for Specific Gravity and Absorption of Coarse Aggregate
(C 127-77) and ASTM Test for Specific Gravity and Absorption of Fine Ag-
gregate (C 128-73), ASTM Test for Liquid Limit of Soils (D 423-66), and
ASTM Test for Plastic Limit and Plasticity Index of Soils (D 424-59). The
AASHTO classification (AASHTO-M-145-66) is also given in Table 2. The
results of the degradation tests are presented in Table 3. Also presented are
the coastal road construction specifications for aggregates. The ASTM Test
for Resistance of Abrasion of Small Size Coarse Aggregate by Use of the Los
Angeles Machine (C 131-76) (L. A. Abrasion Test) is generally used to deter-
mine mechanical degradation. For chemical degradation, each northwest
state has developed its own test procedure to measure the production of
plastic fines under simulated weathering conditions. Therefore, three tests
were used, namely, the California Durability Test (AASHTO T-210), the
Oregon Air Degradation (OSHD Test Method 208-75), and the Washington
fl
2 >.c
IS '^ 2
etj 3 (8
^Q 8
t SI
3
E
•^ in vo o "> E
^ O •^* (N PS *3
00
I— (N O^
1 5>
fO ON rt s^l
1
a
BO
g^
<
JlHiil:
c« g c4
tu .Molil
Durability (WSHD Test Method No. 113 A). In addition, the Sand Equiva-
lent Test (AASHTO-T-176) was conducted on all four aggregates.
The results of the degradation tests shown in Table 3 indicate Eckman
Creek marine basalt is low quality, failing every test except the L.A. Abra-
sion Test, which was borderline. In particular, it performed very poorly in
the Washington Durability Test.
The Tyee sandstone also performed poorly in most tests. The test results
from both the Oregon Air Degradation and L.A. Abrasion Test indicate the
sandstone has low mechanical strength. The tendency to degrade even under
relatively light physical agitation (mechanical sieving) presented problems
throughout the study.
The dredge spoil sand met all of the specifications for which it could be
tested. Oceanlake marine basalt, the control aggregate selected for the study,
met all of the coastal road construction specifications.
Specimen Preparation
The moisture-density relationships for the aggregates considered in the
study were determined at Standard Proctor Compactive effort using ASTM
Tests for Moisture-Unit Weight Relations of Soils and Soil-Aggregate Mix-
tures, Using 5.5-lb (2.5-kg) Rammer and 12-in. (304.8-mm) Drop (D
698-78). Optimum cement contents for the cement-stabilized aggregates
were determined following the Portland Cement Association (PCA) wetting
and drying (ASTM Method for Wetting-and-Drying Tests of Compacted
Soil-Cement Mixtures (D 559-57)) and freezing and thawing (ASTM Method
for Freezing-and-Thawing Tests of Compacted Soil-Cement Mixtures (D
560-57)) durability test method [4].
Test specimens for the resilient modulus and fatigue life tests were pre-
pared at the optimum moisture contents, maximum dry densities, and opti-
mum cement contents given in Table 4. A kneading compactor was used to
achieve the maximum dry density. Following compaction, the 10 cm (4 in.)
diameter by 6.4 cm (2.5 in.) high specimens were extruded and placed in a
moist room for curing. Specimens to be used in the resilient modulus test
program were cured for 3, 7, 14, 28, and 90 days. Further, another group of
the test specimens used in the resilient modulus program were subjected to
12 wetting and drying or freezing and thawing cycles. Specimens to be used
in the fatigue life test program were cured for seven days and then subjected
to 12 wetting and drying or freezing and thawing cycles or cured for 31 days
with no wetting and drying or freezing and thawing cycling.
Test Method and Procedure

Diametral resilient modulus tests were conducted using the apparatus
shown in Fig. 1. The cyclic vertical load is applied through a pneumatic
TABLE 4—Summary of optimum moisture and cement contents and maximum dry
densities of test specimens.
Cement Content Optimum Maximum Dry

Aggregate Used, % Moisture, % Density, kg/m^
Oceanlake
marine basalt 3 7.0 2243
Eckman Creek
marine basalt 6 13.8 1842
Tyee sandstone 6 14.0 1810
Dredge spoil
sand 9 9.0 1858
FIG. 1—Diametral resilient modulus test apparatus: (a) overview of test system, and (b)
displacement yoke, test specimen, and load cell.
piston. The load is measured with a load cell at the base of the specimen.
Horizontal deformations are measured with a displacement yoke that brings
two linear variable differential transformers (LVDTs) in contact with the
specimen. Cyclic loads of 668 N (150 Ibf), 891 N (200 Ibf), and 1336 N (300
Ibf) were used in all tests and the resultant horizontal deformation recorded.
The load pulse was 0.1-s duration with a 1.9-s dwell time. The resilient
modulus was calculated from the test results employing the following rational
equation
,^ P(0.9976M + 0.2692) ,. ., ,^,

MR = ^^:^ (in psi) (1)
where
Ah — horizontal elastic deformation under one load cycle (in.),
P = vertical cyclic load (Ibf),
h = height of the specimen (in.), and
fi = Poisson's ratio (assumed equal to 0.2).
Fatigue life tests were conducted with the electrohydraulic closed loop test
system shown in Fig. 2. Vertical loads and deformations were measured with
a load cell and LVDT, respectively. A cyclic load pulse of 0.1-s duration with
a 0.9-s dwell time was employed. A 222 N (50 Ibf) seating load was applied to
the specimen to prevent excess vibration during the test. When a fatigue
failure occurred, the electrohydraulic closed-loop test system was program-
med to stop to provide a record of the number of repetitions to failure. The
applied tensile stress was calculated as follows
2P
i£max
:h'D (2)
where
a, = tensile stress (Pa or psi),
^max — maximum vertical cyclic load (N or Ibf),
h — height of specimen (m or in.), and
D — diameter of the specimen (m or in.).
The test system shown in Fig. 2 was also used to determine the tensile
strength of the test specimens. The test specimen was carefully centered on
the lower loading strip and the lower plate was raised until contact was made
between the specimen and the upper loading strip. The specimen was tested
at a loading rate of 5.0 cm/min (2.0 in./min). The tensile strength was
calculated as follows
5, = ^ 5 ^ (3)
irft-D
FIG. 2—Electrohydraulic closed-loop fatigue test system: (a) overview of test system, and (b)
test specimen.
where
tensile stress (Pa or psi),
maximum vertical cyclic load (N or Ibf),
h = height of specimen (m or in.), and
D = diameter of specimen (m or in,).
Resilient Modulus Test Results
Effect of Curing Time

The modulus of the cement stabilized aggregates increased with curing
time for all four test materials. The increase may be approximated by a
straight line on a logarithmic plot as shown in Fig. 3. The test results indicate
that Oceanlake marine basalt has the highest modulus, followed by Eckman
Creek marine basalt, dredge spoil sand, and Tyee sandstone. The values of
modulus for the four materials at 28 days moist cure ranged from 8100 MPa
(1.17 X 10^ psi) to 25 100 MPa (3.64 X lO*" psi). The increase in modulus
with curing time appears to be greatest for the Oceanlake marine basalt com-
pared to the marginal aggregates.
Effect of Wetting and Drying and Freezing and Thawing

The effects of wetting and drying or freezing and thawing cycles on
resilient modulus are varied and depend to a certain extent on the type of
material. Based on the test results shown in Fig. 4, the modulus was only
slightly affected after the freezing and thawing cycles; however, it decreased
significantly after wetting and drying cycles for all four test materials, in par-
ticular the Eckman Creek and Oceanlake marine basalt. The modulus of
5.0
4.0
O 20
i..:.
2 .4 • Ocaan Lake Marina Baaalt
0 Eckman Creak Marine Basalt
• Dredge Spoil Sand

1^ Tyee Sandatone
1 psi = 6.9 kPo
• I I _L
20 30 SO
Moist Curing Time, days
FIG. 3—Resilient modulus versus moist curing time.
_ - 26 day m o i s t curing
3.0
~
•' fter freeze/thaw cycling
2.0
~
H' fter wet/dry
1 psi = 6 . 9 kPa
cycling
CO
a.
ID
o 1.0 1.0
-_
•
0 1
(a) Eckman Creak Marina Basalt
~
(b) Tyae Sandstone

•
»
o
-
c
3.0 — 3.0 -
2.0 2.0
-
1.0
-
•
m/// 1.0 -
1
n 1
(c) Dredge Spoil Sand
0 il 1
(d) Ocean Lake Marine Basalt
FIG. 4—Effect of freeze/thaw and wet/dry cycling in resilient modulus.
Eckman Creek marine basalt decreased to 38 percent of the value for 28-day
moist curing, Oceanlake marine basalt decreased to 52 percent, dredge spoil
sand decreased to 62 percent, and Tyee sandstone decreased to 79 percent.
The reason for this decrease is not entirely clear. However, chemical
degradation potential for the Eckman Creek marine basalt and the grada-
tions used in the study are undoubtedly important factors. The test
specimens of the marine basalts contained relatively larger void spaces com-
pared with the Tyee sandstone or dredge spoil sand. These voids may permit
more reactions between aggregate and water during the wetting and drying
cycles.
Fatigue Life Test Results

The results from fatigue life tests of cement-treated material are expressed
in terms of an S-N curve, that is, the relationship between the number of
Indian Institute of Technology Kharagpur pursuant to License Agreement. No further reproductions author
repetitions to failure (N) and either the ratio of stress to strength or applied
stress (S). Because the results expressed in terms of the ratio of stress to
strength provide a better comparison of fatigue behavior for the different
materials and the results based on stress are convenient for layered elastic
analysis, all test data are expressed in both forms. Test specimens for each
material were tested after 31 days moist curing, after wetting and drying
cycles, and after freezing and thawing cycles. Near the end of each test, a
small crack developed at the center of the specimen. This crack gradually ex-
tended along the vertical diametral direction until failure occurred.
Regression analyses were conducted to obtain the equation that would
satisfactorily predict the relationship between the number of load applica-
tions to failure and either applied stress or the ratio of stress to strength. The
relationship used is as follows
LogAr^=Ci^ + C2 (4)
where
Nf = number of repetitions to failure (fatigue life),
S = stress ratio or applied stress (psi),
Ci = slope of the logarithmic relationship, and
C2 = intercept of the logarithmic relationship.
The regression coefficients are summarized in Table 5. Average tensile
strengths of the stabilized materials for the curing conditions considered are
also given in Table 5. The applied ratio of stress to strength was from 0.55 to
0.90.
Typical results for the fatigue life tests (applied stress versus number of
repetitions) are summarized in Fig. 5. The results indicate that Oceanlake
and Eckman Creek marine basalts generally have steeper S-N curves than the
Tyee sandstone and dredge spoil sand. The results also indicate that fatigue
life is decreased as the applied stress is increased for all materials. Further,
the difference in the fatigue lives of test specimens between the moist curing
condition and after durability tests vary with the different materials. In
general, except for Tyee sandstone, the number of repetitions to failure at a
given applied stress is substantially reduced after wetting and drying or freez-
ing and thawing cycling when compared to the results for moist cured
specimens. For Tyee sandstone, the results indicate there is no significant
difference in the fatigue life for the moist cured specimens and those sub-
jected to wetting and drying or freezing and thawing cycling. As shown in
Fig. 5, the effect of the wetting and drying cycling is slightly greater than the
freezing and thawing cycling for Eckman Creek marine basalt, whereas for
dredge spoil sand the effect of the freezing and thawing test tends to be
larger. For Oceanlake niarine basalt and Tyee sandstone, the effect of both
the wetting and drying and freezing and thawing cycling is the same. If the
o o o o o o o o" o o o o
o ro ^ '-H irj ro Tj- esq o 1-* n r--

ON od 00 '^* CO r--' rS 00 (N r-* oo" (N
'-('-H'-H(N'-HT-((N'-Hrora'-i»-i
mooor5'-if0ôt--(NO'^"g-
oo^<N'rHrtr'>ooir>(Noo
o o o o o o o o o o o o
I M I I I I 1 I I I I
^r^-tNoooOQQp^fSTrr-o
i/ir^fniooooi/)'OfNTrr~-r--
o o o o o o o o o* o o o
aNOI--ro<Ni^roQâNOO
-~{NGi0'<roQ0cPiwr--o
oi 00* 00*rrrr\0\6I^"t^ro 00* Q t^* O i/i
I
o o o o o o o* o" f^ o o o
I I I I I I I I I I I I
m
o
U
00
s
•c
3
u
li I
••a
o
. . . .
I O 3 1 days moist curing Z
250
r 9 after rr«e2e/th«w cyctingZ
: ; V after wet/dry cycling :
200 - specimen did not fail ~
~ - _
! . — • ; 1 p 8 i - 6 . 9 ItPo
150
-^ -i
- -
100 V ^ --V-'
^ ^~—: _ _,^^îÔA_^^__ ^
r
^^'~V "^ -jy*-^#>!:-J-:^fmn \M_. 1
*^ 50 r 0*^ J
b** —
- —
I
il 1 1 1 r • 1 1 1 il 1 Mil 1 . Ill 1 1 III 1 1 III 1 1 1'
,• ,n2 ,„3 ,„< .„5
(>) Eckmtn Crttk Mtrint B t f t i t (b) T y » Sandstont
CO
1_1 ' '*l ' * '*! ' • •! I Lul 1—\ Mi_ 1 aL_i Liii_j uiI ' I I LUl 1 ' "' I
1? i? i?~ IC?
(c) Ori4|( Sfdl Stud (d) Ocitn Lake Marina Basalt
Number of Repetitions to Failure
FIG. 5—Applied tensile stress versus number of load repetitions to failure.
results are expressed in terms of the ratio of stress to strength, the fatigue
curves after the durability tests are generally similar to those for the moist
curing condition.
Development of Layer Equivalencies
Methodology
The approach used to develop layer equivalencies is given in Table 6.
Typical road sections consisting of an asphalt concrete (AC) surface over a
cement-treated base (CTB) and subgrade are assumed. Multilayer elastic
analysis theory is used to calculate the maximum tensile stress in the CTB
and the vertical strain on the subgrade. These values are compared with the
limiting tensile stress determined from the fatigue life test results of the
material used for the CTB, given in Fig. 6, and the limiting subgrade strain
criteria by Santucci [5] given in Fig. 7.
The relationships between the thickness of CTB and associated flexural
stress at the bottom of the CTB and subgrade strain were calculated and are
plotted in a manner similar to Fig. 8. For a given design number of load
O O 3
5)00 "
£<J §
o 5 t>
,
•§ e °
11
3 •'3 •- «
ul 1
•i;3
•o E
Sb a
S)
"3 "O _
&;! = •
^ 'P ^ .9
S u u 'S
<c a B «
280
Ocean Lake Marine Basalt a f t e r w e t / d r y
250 ,' cycling
a.» Eckman Creek Marine Basalt a f t e r w e t / d r y

,• cycling
b 200
/ Tyee Sandstone a f t e r f r e e z e / t h a w cycling
I p8i = 6 . 9 kPo
J 1_1_U _l L-U I I L_lJ • • .1
Number of Repetitions to Failure

FIG. 6—Fatigue life test results used for layer equivalencies.
t 100
'O
o
CO
w 1 KIP = 4.45 kN
10
I I ' • I I II III I
o 1
10" 10- 10^ io'
Number of I 8 KIP Equivalent Axle Load(EAL) Applications
FIG. 7—Subgrade strain criteria [5].
repetitions, one can establish a limiting tensile stress from the previously
derived fatigue curves for the material tested. The required thickness of CTB
to preclude fatigue is then determined using the approach shown in Fig. 8. A
similar technique is used to determine the required thickness of CTB based
Developed from Multilayer Elastic

Developed from Fotigue Life jest Analysis Using Mr ofCTB Materiol
•C
c
Q> quality aggregate
^,/ marginal aggregate \ quality aggregate
w
CO
c v\ marginal
/ aggregate
Design Fatigue Life . D2
Number of Repetitions to Failure CTB Thickness
Refer to Figure 7 Developed from Multilayer Elastic

Analysis Using Mr of CTB Material
V, quality aggregate
\, N, marginal
^ ^ " " ^ - - ^ . ^ ^ \ . \ . / aggregate
Design Fatigue Life , Dp

D|
u
Number of ISkEAL Applications CTB Thickness
FIG. 8—Methodology to establish design CTB thickness: (top) limiting tensile stress criteria,
and (bottom) subgrade strain criteria.
on limiting subgrade strain. The larger of the two derived thicknesses for
CTB represents the design control condition.
Limiting tensile stresses from diametral fatigue tests are not equivalent to
limiting flexural stresses measured in bending. Generally, limiting tensile
stresses from diametral fatigue tests are lower than limiting flexural stresses
measured in bending. Consequently, the methodology employed leads to a
conservative estimate of required CTB thickness for this mode of failure.
Layered Analysis
The computer program ELSYM5 [6] was used in this study to calculate the
critical stresses and strains in the pavement system. The program determines
stresses, strains, or displacements in a three-dimensional ideal elastic layered
system. Each pavement layer is assumed to be of uniform thickness and ex-
tend infinitely in the horizontal direction. The subgrade is infinite in

thickness. The applied loads are described by any two of the following items:
load in pounds, stress in pounds per square inch, or contact radius in inches.
The program determines the missing value. Each layer of the system is
described by its modulus of elasticity, Poisson's ratio, and thickness. The
total layers for a solution are limited to a maximum of five.
The assumed pavement structure that includes the AC concrete, CTB, and
subgrade is shown in Fig. 9. An elastic modulus of 2760 MPa (4 X 10^ psi),
Poisson's ratio of 0.40, and thicknesses of 5 and 10 cm (2 and 4 in.) were
assumed for the AC concrete surface layer. Elastic moduli from 20.7 MPa (3
X 103 psi) to 138 MPa (2 X lO'' psi), Poisson's ratio of 0.4, and a semi-
infinite thickness were assumed for the subgrade layer. Elastic moduli from
690 MPa (105 psi) to 13 800 MPa (4 X W psi), Poisson's ratio of 0.15, and
4 5 0 0 Ibf 4 5 0 0 I bf
12.69 in.
Asphalt
Concrete ( A O - d,= 2 , 4 i n . E , ' 4 0 0 , 0 0 0 psi
Layer I /i»0.4
Cement Treated Eg' 100,000, 1,000,000, 2,000,000

Base (CTB)- d2=3.6,10,14 in. 3 , 0 0 0 , 0 0 0 , 4,000,000 psi
fL2' 0.15
Layer 2
'cTB 'SG
mmmmmmmmmmmmmmmmmmmmmmmmmm
Subgrade dg.OO Ej" 3,000, 1 0 , 0 0 0 , 2 0 , 0 0 0 psi
(SG) ^ 3 = 0.4
I in. = 2 . 5 4 cm
I p s i = 6 . 9 kPo
I lbs = 4 . 4 5 N
I kip = 4 . 4 5 kN
DESIGN CONDITIONS:
(1) Traffic^ 50,000 or 5 0 0 , 0 0 0 - 1 8 kip single axle loads
over the design period
(2) Regional Foctor ( R ) ° 2
(3) Terminal Serviceability Index (Pt)= 2
FIG. 9—Assumed pavement structure to determine layer equivalencies.
Indian Institute of Technology Kharagpur pursuant to License Agreement. No further reproductions authoriz
thicknesses of 3, 6, 10, and 14 in. for CTB were used to derive the relation-
ships between thickness of CTB and the tensile stress or vertical compressive
subgrade strain. An example of the results obtained from these analyses is
given in Fig. 10. Detailed results are given by Chang [7].
The results obtained from the layered analyses for the assumed pavement
structure, given in Fig. 9, may be summarized as follows: (1) using the stress
criteria, the required CTB thickness increases with increasing CTB modulus
or with decreasing allowable tensile stress; (2) using the subgrade vertical
strain criteria, the required CTB thickness increases with decreasing CTB
modulus or with decreasing allowable vertical strain; and (3) the calculated
subgrade strain was always less than the allowable subgrade strains,
therefore, the design CTB thickness was primarily controlled by the stress
criteria.
Layer Equivalencies
Layer equivalencies were determined for two traffic conditions, low volume
(50 000-80 kN axle loads) and moderate volume roads (500 000-80 kN axle
1 1 1 1 1 1 1 1 T—I—1 ' 1 1 1—r—T—r—r~

Assumed A C modulus, E | = 4 0 0 , 0 0 0 psi
AC T h i c k n e s s , ()|=4 in.
600 _r S i bgrade Modulus, £ ^ = 3 0 0 0 psi —
(CBR 2) ;
;-
1 i n . = 2 . 5 4 cm
':
500
1 p8i = 6 . 9 kPo
0> ^ -.
•>
« 400
4->
CO
'— ^ —.
•> CTB Modulus, Eo.psi
•5 300 6
1^
e i~ q • 4 x 10
a>
1- L ^ • 3 X lof J
<d
: O 2 x 10° "
200 ; * • 1 X 10® '_
3 V O . I x 10^
X 2—
100 :— —.
LJ 1 L. • 1 j _ i 1 , 1 , , . ? 1 , • , 1-
2 4 6 8 10 12 14 16 18 20
CTB Thickness, d^, inches

FIG. 10—Flexural tensile stress versus CTB thickness.
Indian Institute of Technology Kharagpur pursuant to License Agreement. No further reproductions
loads). The required CTB thickness for each material was obtained using the
modulus after freezing and thawing cycling, the fatigue curves after the
durability tests (Fig. 6), and the stresses derived from the multilayer elastic
analysis. The thicknesses determined are summarized in Table 7.
Based on the results presented in Table 7, the required CTB thicknesses of
Tyee sandstone and Eckman Creek marine basalt for low volume roads are
very close to that of Oceanlake marine basalt. The required CTB thickness
for the dredge spoil sand tends to be about 1.3 to 1.4 times greater than the
thickness required for the Oceanlake material, depending on the AC layer
thickness and subgrade modulus. The required CTB thickness for the Tyee
sandstone, owing to its low modulus value, is equal to the thickness for the
good quality aggregate.
Table 7 also summarizes similar results for load repetitions of 500 000-80
kN axle loads. The results indicate that the required CTB thickness is only
slightly increased with an increased design traffic load.
Sanunaiy
The suitability of using marginal aggregates treated with cement as the
base course of a road section has been evaluated. Three marginal aggregates,
Eckman Creek marine basalt, Tyee sandstone, dredge spoil sand, and one
quality aggregate, Oceanlake marine basalt, were selected for the study. The
resilient modulus and fatigue life characteristics for the aggregates stabilized
at the optimum cement content were evaluated after moist curing and wet-
ting and drying or freezing and thawing cycling. Multilayer elastic theory
analysis was used to determine the required CTB thickness for the different
materials in an assumed pavement structure. The significant findings of the
study may be summarized as follows:
1. The resilient modulus of the four materials tested increased with curing
period. The mean modulus values for all materials after 28-day moist curing
ranged from 8100 MPa (1.17 X 10^ psi) to 25 100 MPa (3.64 X 10^ psi).
Tyee sandstone had the lowest modulus.
2. There was no significant difference in the moduli for moist-cured
specimens compared to the moduli for specimens subjected to freezing and
thawing cycling.
3. The moduli for specimens subjected to wetting and drying cycling
decreased from 30 to 65 percent compared to the modulus values for 28-day
moist curing. For the Eckman Creek and Oceanlake marine basalts, the
decrease in modulus exceeded 50 percent.
4. In general, there was no significant difference in the slope of the fatigue
life curve expressed in terms of stress ratio or stress versus number of load
applications for 31-day moist-cured specimens and specimens subjected to
wetting and drying or freezing and thawing cycling.
5. The number of repetitions to failure at a given applied stress is substan-
8S88g8S88SSSS
s
oôo^-Ofoooop'-Hoq^S'-Hcq'0ôooo^-;ooo^
•^.s oo6odo^r-'^--'doôdo^<»t-~'daôdod^^''Of^'-Hd'-ôô6
II
1^
2-B
•a
s
I
t
I fSfS(N-^'^TrrstSfSTrTr-«rfSfS(S"<a-"*'*r*(sr^-<i-^^
^
CO th
" II 00
11 II ._
•S„
H
U
•s
ss
g
~S-^—l
«s 4)
— -,
Ji
u
•-;
•s
11
n III
•a
J
1^ 1- 1
II II II
MS 3 »2
B T3
tially reduced after wetting and drying or freezing and thawing cycling when
compared to the results for moist-cured specimens.
6. The results from a multilayer elastic analysis of an assumed pavement
structure indicate that the required thickness of CTB increases with an in-
crease in CTB modulus and decreases with an increased allowable tensile
stress or tensile strength of CTB.
7. The resilient modulus and tensile strength of CTB increase concurrently
with age. The concurrent increase in modulus and strength should be con-
sidered in the determination of the required CTB thickness.
8. Subgrade strains for the assumed pavement structure utilizing the
cement-treated marginal aggregates are relatively small compared to limiting
values of subgrade strain.
9. Based on the test results and multilayer elastic theory analyses, the re-
quired thickness of CTB for the assumed pavement structure was approx-
imately the same for the Tyee sandstone, Eckman Creek marine basalt, and
Oceanlake marine basalt (a quality aggregate); the layer equivalency for the
dredge spoil sand was calculated to be about 1.3 to 1.5.
Acknowledgment
This work is a result of research sponsored by the Oregon State University
Sea Grant College Program, supported by the National Oceanic and At-
mospheric Administration (NOAA) Office of Sea Grant, U.S. Department of
Commerce, under Grant No. NA79AA-D-00106. The U.S. Government is
authorized to produce and distribute reprints for governmental purposes not-
withstanding any copyright notation that may appear hereon.
References
[/] Chintakovid, V., "Evaluation of Aggregate Needs and Problems Along the Oregon Coast,"
a project report submitted to the Civil Engineering Dept., Oregon State University, Cor-
valHs, Ore., April 1979.
[2] Clemmons, G. H., "An Evaluation of Coastal Oregon's Marginal Aggregates," Transpora-
tion Research Report 79-5, Transportation Research Institute, Oregon State University,
Corvallis, Ore., June 1979.
[3] Day, H. L., "A Progress Report on Studies of Degrading Basalt Aggregate Bases," Highway
Research Board Bulletin 344, 1962.
[4] "Soil-Cement Laboratory Handbook," Portland Cement Association, Skokie, 111., 1971.
[5] Santucci, L. E., "Thickness Design Procedure for Asphalt and Emulsified Asphalt Mixes,"
4th International Conference on Structural Design of Asphalt, Vol. 1, 1977, pp. 424-458.
[6] Hicks, R. G., Swait, J. D., Jr., and Chastain, E. O., "Use of Layered Theory in the Design
and Evaluation of Pavement System," Dept. of Civil Engineering, Oregon State University,
Corvallis, Ore., Jan. 1978.
[7] Chang, C , "Evaluation of Selected Marginal Aggregates Stabilized with Cement,"
Transportation Research Report 80-8, Transportation Research Institute, Oregon State
University, Corvallis, Ore., June 1980.
J. J. Emery
Slag Utilization in Pavement

Construction
REFERENCE: Emery, J. J., "Slag Utilization in Pavement Construction," Extending

Aggregate Resources, ASTM STP 774, American Society for Testing and Materials,
1982, pp. 95-118.
ABSTRACT: An overall review of the current and potential utilization of ferrous, nonfer-
rous, and boiler slags in pavement structures is given. It would appear that the optimum
uses for iron blast furnace slag are in cements and base stabilization processes that take
advantage of the energy "invested" in the suitably vitrified slag (granulated or pelletized),
rather than bulk aggregate applications. Steelmaking slag that is not recycled can be
used to advantage in asphaltic concrete (high stability and skid resistance), and may have
applications in base and shoulder stabilization and construction (noting the expansion
potential). Nonferrous slags such as copper and nickel slag can be used in granular base
and structural fill construction, and related applications such as ballast. Boiler slags are
used in a number of applications such as seal coat, asphaltic concrete, and granular base
construction. While most ferrous slag is utilized, there is still much scope to utilize non-
ferrous and boiler slags that are often disposed of with consequent environmental impact.
With the present indications of materials and energy shortages, there is a new concern
that all byproduct slags be utilized effectively.
KEY WORDS: slag, blast furnace slag, steel slag, nonferrous slag, boiler slag, utiliza-
tion, cement, stabilization, concrete, asphalt, base, fill, aggregates
The purpose of this review is to outline the demonstrated and potential

uses of ferrous, nonferrous, and boiler slags in the construction and mainte-
nance of pavements. Slag use, particularly in cementitious applications,
forms part of an overall strategy to optimize the utilization of available ma-
terials—selective and more efficient use, conservation, beneficiation, new
materials and methods, more use of residuals and by-products, etc.—that
must be coupled with environmental and energy considerations [l].2 The an-
nual world production of iron blast furnace slag (approximately 120 million
tonnes based on a generation of 23 percent of iron production) and steelmak-
'Manager, Trow Ltd Consulting Engineers, Hamilton Branch, Hamilton Ontario, Canada.
2
Copyright by ASTM Int'l (all95 rights reserved); Thu Dec 22

ing slag (approximately 53 million tonnes based on a generation of 18 percent

of steel production and 60 percent recycle to blast furnace burden) is very
small when compared to world aggregate needs of about 13 billion tonnes [2].
However, there are many applications other than bulk aggregate uses where
slags already make a significant contribution to our materials requirements,
and can be further exploited. For instance, if all iron blast furnace slags were
produced in a vitrified form (granulated or pelletized) suitable after grinding
for cementitious uses, approximately 15 percent of the world requirement for
cement could be met with significant energy savings [3]. There is little doubt
that the escalating costs of binders such as asphalt and portland cements
(high energy content materials) will foster a dramatic increase in North
American slag cement production.
Although the current utilization of ferrous slags may not always be optimal
in terms of materials and energy considerations, they are being widely and
generally totally used. At present, much nonferrous slag is disposed of rather
than utilized. Throughout this review it will become apparent that much re-
search, development, and demonstration work is still required to ensure the
full exploitation of many slags. It should also be noted that while pavement
uses are emphasized, slags have applications over the full spectrum of con-
struction industry activities from aggregate in foundations to instilation in
walls. The review has been developed mainly from research and development
activities in Ontario (particularly with pelletized blast furnace slag and steel-
making slag) and participation in Organization for Economic Co-operation
and Development (OECD) and Reunion International des Laboratories d'
Essaes et de Recherches sur les Materiaux et les Constructions (RILEM)
studies [1,4,5].
Ferrous Slags
Ferrous slags (iron blast furnace, steelmaking, foundry, ferroalloy) are
probably the industrial byproducts of greatest interest to pavement construc-
tion, given their wide availability and scope of uses. Optimal uses appear to
be slag cement manufacture for vitrified blast furnace slags, and aggregate in
skid resistant asphaltic concrete for steel slags. In considering the use of fer-
rous slags, there are several important factors to note: pavement use is not
new (iron slags were used during Roman road building in England); they are
not wastes, but valuable by-products processed for the construction industry;
there is a continuing trend towards cementitious applications; and blast fur-
nace slags (relatively low bulk density—typically 1300 kg/m^ for air-cooled—
and stable with no expansion tendencies) must not be confused with steel
slags (high bulk density—typically 1900 kg/m-'—and potential expansive na-
ture related to uncombined oxides) [1].
EMERY ON SLAG UTILIZATION IN PAVEMENT CONSTRUCTION 97
Iron Blast Furnace Slag Production and General Uses

Iron blast furnace slag results from the fusion of fluxing stone with coke
ash and the siliceous and aluminous residues remaining after the reduction
and separation of iron from the ore. The blast furnace operation is a continu-
ous process with the carefully controlled raw materials being fed in (burden)
and the uniform products, molten iron and liquid slag (Fig. 1), being drawn
off at regular intervals. Selective cooling of the liquid slag results in four dis-
tinct types of blast furnace slag: (1) air-cooled (solidification under ambient
conditions), which finds extensive use in conventional aggregate applica-
tions; (2) expanded or foamed (solidified with controlled quantities of water,
sometimes with air or steam), which is mainly used as a lightweight aggre-
gate; (3) granulated (solidified by quick water-quenching to a vitrified
(glassy) state), which is mainly used in slag cement manufacture; and (4) pel-
letized (solidified by water and air-quenching in conjunction with a spinning
drum), which is used both as a lightweight aggregate and in slag cement
manufacture. Typical ferrous slag production statistics for several countries
given in Table 1 indicate a trend in Europe towards more granulation (that
is, cementitious applications) than in North America where air-cooling (that
FIG. 1—Molten iron blast furnace slag in runner (Fos, France). A direct runner pelletizer is
operating on the right side.
+•^ + + +
J < as
I 2
x5x
i * X*
1 y
il ±±
00 TT •TH <N
CO O
o
a -ft
I
(N I/) 0^ t-' IQ
CO
11.
jl
J3 C ao
4!
I Jllll
is, aggregate applications) currently predominates, but pelletizing and gran-

ulating are growing as slag cement manufacture develops.
Since the pelletizing process is relatively new (25 machines currently active
in nine countries), it will be described before focusing on specific applica-
tions for blast furnace slags. The basic action of the pelletizer with its spin-
ning drum (Figs. 1 and 2) is to take a pyroplastic expanding slag, and chill it
quickly in the dispersed form with air and water, so that hydrogen sulfide
emissions are limited. To do this, a controlled slag flow (up to 3 tonnes/min)
is expanded under water sprays on a vibrating feed plate, and then passed
over the internally-cooled spinning drum (280 rpm) where the eight concave
throwing vanes (Fig. 2) break-up the now pyroplastic material and "hurl"
the slag into the air for a sufficient time that surface tension forms pellets
(Fig. 3). By controlling the process, a more crystalline pellet can be produced
for aggregate applications, or a more vitrified pellet (increased drum speed
and more cooling water) for cementitious applications [6]. Because of this
rapid cooling, the pelletizer can be considered an "air granulator".
While gaseous emissions control was the early motivation for installing pel-
letizers (hydrogen sulphide reduced from 4000 ppm to 10 ppm in the hood),
the prime reason for most installations now is the versatility of applications
for the slag product. Other benefits of pelletizing as a slag handling process
that have been demonstrated include: immediate pellet removal; minimized
FIG. 2—Schematic of direct runner pelletizer.
FIG. 3—Physical appearance of minus 12.5 mm pelletized slag {\eit) and expanded slag (nght).
space requirements; low installation costs; and water conservation. The op-
timum conditions for pelletizing appear to be a uniform flow of basic slag
that is as hot as possible, which favors a direct runner pelletizer(s) installed
at the furnace (Figs. 1 and 2).
Aggregate Applications for Air-Cooled Blast Furnace Slag [1,4,5J

Granular Base, Ballast, Trench Fill, Backfill, and Engineered Fill—The
general features of air-cooled blast furnace slag that makes it attractive for
such bulk aggregate applications include: low compacted bulk density (typi-
cally 1200 to 1450 kg/m-') that reduces dead load, lateral pressures, and
transportation costs on a volumetric basis; high stability (California Bearing
Ratio (CBR) > 100) and friction angle (approximately 45 deg); ability to
stabilize wet, soft underlying soils at early construction stages; placeable in
almost any weather; very durable with good resistance to weathering and ero-
sion; free draining and non-frost susceptible; almost complete absence of set-
tlement after compaction; and non-corrosive to steel and concrete. Large
vibratory compactors have proven very effective during placement. A partic-
ularly advantageous use is in difficult fill conditions over soft ground. Even
though blast furnace slag may exhibit a sulfur odor and early contacting
water may take on a greenish tinge, an alkaline condition is maintained (pH
8 to 10, that is, steel protected), and the leachate does not contain significant
concentrations of toxic constituents as indicated clearly by the data sum-
marized in Table 2. Most of the iron blast furnace slag produced in North
America currently goes to such bulk applications.
Asphaltic Concrete—Air-cooled blast furnace slag is used to some extent
in both dense graded and open graded asphaltic concretes. However, asphal-
tic concrete applications appear to be limited somewhat by high asphalt ce-
ment content requirements (about 8 percent for dense graded) related to
both the vesicular surface and absorption of the slag. High skid resistance
has been found for these mixes, and also for blended blast furnace-steel slag
mixes. Finer air-cooled blast furnace slag is also used in slurry seals with ex-
cellent results. Further details on slag skid resistance are given later.
Portland Cement Concrete—Air-cooled blast furnace slag is used as the
coarse aggregates in all types of concreting operations associated with road
construction: pavements; precast and prestressed units; foundations, curb,
and gutter; and ready-mix. Concrete mixes incorporating air-cooled blast
furnace slag compare favorably in compressive and flexural strength with or-
dinary mixes, with significant savings due to lower unit weight. Durability
and aggregate-paste bond are excellent, but workability requirements may
involve proportioning to allow for slightly more sand and water. It should be
noted that research and performance records indicate blast furnace slags are
non-corrosive to embedded steel and steel such as piling.
TABLE 2—Typical ferrous slag solubility test data (Calspan Corporation for Environmental
Protection Agency).
Analysis of Filtrate,. mg/1"*
Slag Cr Cu F Mn Pb Zn pH
Blast furnace (iron) <0.01 <0.03 1.9 <0.01 <0.20 <0.01 10.6
Open hearth (steel) 0.01 0.04 3.1 <0.01 0.30 <0.01 12.5
Basic oxygen (steel) 0.03 0.03 4.0 <0.01 0.20 <0.01 12.5
Electric arc (steel) 0.27 0.03 1.5 <0.01 0.44 <0.01 12.4
Iron and steel foundry 0.05 0.25 ND'' 0.06 <0.20 0.12 10.6
Ferrochrome 0.02 0.02 ND 0.30 0.40 0.20 9.9
Ferronickel 0.01 0.74 ND 0.07 1.00 2.00 8.5
Silicomanganese <0.01 0.17 ND 0.10 <0.20 0.05 6.8
Ferromanganese 0.02 0.04 ND 2.10 <0.20 0.03 5.9
"One solubility test for each slag. Two parts distilled water (pH 5.5) to one part slag gently ag-
itated for 72 h. Mixture filtered through 0.45 /^m micropore and filtrate analyzed. Selected sam-
ples probably represent the more comminuted fraction of slag. Ni <0.05 mg/1 for all slags and
Co <0.02 mg/l for ferronickel slag.
*A1I ferrous slags listed considered non-hazardous. (Toxic materials not leached at greater
than 1 mg/1. Greater tolerances for F, Mn, and Zn.)
'•ND= not determined.
Aggregate Applications for Expanded and Pelletized Blast Furnace Slags

/l,4-67
Lightweight and Semi-lightweight Portland Cement Concrete Applica-
tions—Expanded and pelletized slag aggregate use in lightweight structural
concrete can provide a significant overall reduction in total construction costs
through reduced dead loads and attendant savings in concrete quantities,
steel, formwork, foundations, etc. for structures associated with roads.
There are some advantages of pelletized slag over expanded slag in such ap-
plications as its smooth surface (Figs. 3 and 4) reduces water absorption and
cement demand.
Insulating Blankets—While not used in North America for insulating
blankets under pavements to reduce frost susceptibility, this is a major use
for pelletized slag in Sweden and Finland (thermal conductivity of about 0.19
W/m°C at 1185 kg/m^).
Cementitious Applications for Granulated and Pelletized (Vitrified) Blast

Furnace Slags /'l,4-6/
Slag Cements—The utilization of the cementitious properties of vitrified
blast furnace slag takes advantage of the substantial energy invested in the
slag during iron production (Fig. 1). This energy saving can be significant
FIG. 4—Thin section of pelletized slag ('courtesy of H. Wilson, CanmetA
when it is noted that wet process portland cement manufacture requires

about 6700 MJ per tonne and only about 25 percent as much energy is re-
quired to produce ground slag for portland cement replacement [3]. The use
of granulated blast furnace slag in cement manufacture is not that new, and
was probably foreshadowed by Vicat (1778-1861) who noted the similarity
between the chemical composition of blast furnace slag and portland cement
as shown by the typical data in Table 3. It should be noted that pelletizing is
now used to produce vitrified slag for slag cement manufacture in several
countries, but most notably in Hamilton where the process was developed
(about 100 000 million tonnes of separately ground slag cement in 1980).
Vitrification of the slag makes it active as the sudden cooling prevents the
molecules from "organizing" themselves into crystals. The cementitious be-
havior of glassy slags is a complex subject, but if of suitable composition,
glassy slags hydrate readily in the presence of certain activators (Ca(OH)2
from hydration of portland cement, Ca(OH)2, KOH, NaOH, sulfate com-
pounds such as gypsum) that can dissolve the impermeable acidic gels on the
surface of slag grains that inhibit hydration. This surface phenomenon and
the resulting reaction forms hydration products (tobermorite that is primar-
ily responsible for strength development, etc.) that are precipitated from
solution. Thus, ground glassy blast furnace slags are not pozzolans since dis-
tinct hydrated compounds are formed in the presence of low levels of activa-
tors (the optimum level, and type, of activator(s) is the subject of much re-
search). Some vitrified foundry slags are also cementitious.
In addition to glass content and composition, the fineness of ground vitri-
fied slag has a significant influence on strength development. While a high
degree of fineness is obviously desirable, this must be related to the econom-
ics of grinding. A suitable optical method for determining the degree of vitri-
TABLE 3—Chemical composition of iron blast furnace slags, portland cement,

and steel slags flj.
Iron Blast Furnace Slag Steel Slag
Usual Typical Portland Open Basic

Constituent, % Range Pellets Cement Hearth Oxygen
Calcium oxide (Ca) 36-45 37.6 64.1 25.8 41.3

Silicon dioxide (SiOj) 33-42 34.8 22.0 16.4 15.6
Aluminum oxide (AI2O3) 10-16 8.1 5.5 2.4 2.2
Magnesium oxide (MgO) 3-16 15.4 1.4 10.0 6.9
Iron (FeO or FejOs) 0.3-2 0.8 3.0 26.0 20.0
Sulfur (Sr 1-3 1.3 2.1*
Manganese oxide (MnO) 0.2-1.5 0.61 11.2 8.9
Titanium dioxide (Ti02) 0.31 0.8 0.5
Free lime (free CaO) 0.47 2.1 3.3
"Principally as calcium sulfate.
fication has been developed by R. D. Hooton at McMaster University [7\.

With the close interplay of glass content, composition and fineness of ground
slags in determining strength development, it is considered that direct tests
(cube tests for slag cement and standard compressive strength tests for base
stabilization) still give the best indication of potential performance.
The three prime cement applications for vitrified blast furnace slag are:
Portland blast furnace slag cements produced by intergrinding various pro-
portions of glassy slag with portland cement clinker; super-sulfated cement
produced by intergrinding glassy slag, anhydrite and portland cement; and
separate dry ground glassy slag for addition at the mixer as cement replace-
ment. These slag cements are used in a full range of pavement applications
from cement treated base to structural slabs.
While there is some retardation in early strength development with slag ce-
ments, for up to about 50 percent replacement of portland cement, strength
equality with 100 percent portland cement mixes is obtained in 28 days or
less, and ultimate strengths are often superior. Slag cement/portland cement
mixes require a lower water to cement ratio for equivalent workability and
this must be considered during mix design. Appropriate cold weather prac-
tices are adopted when necessary since a low heat type of cement is actually
involved that can be used to advantage in applications where low heat of hy-
dration is required. The main advantages of using ground slag for cement re-
placement appear to be: lower cost (substantial energy for clinkering is
avoided); improved workability; chemical resistance; lower heat of hydra-
tion; and versatility in obtaining desired fineness of ground slag.
Pelletized slag is currently being used in slag cement manufacture as it can
be produced with a high degree of vitrification for cementitious applications.
As noted earlier, the pellet structure is important (Fig. 5) as a high degree of
vitrification is possible. Indeed, some hydration products are evident in Fig.
5. The largest plant doing this (Standard Slag Cement in Hamilton) produces
about 100 000 tonnes/year of separately ground pelletized slag cement meet-
ing Canadian Standards Association (CSA) Preliminary Standard A363 for
Cementitious Hydraulic Slag. Markets for such slag cements are growing
rapidly as the substantial energy-saving possibilities in comparison to port-
land cement manufacture are recognized. The sulfate resistance of pelletized
slag cement (cementitious hydraulic slag)/portland cement blends is cur-
rently being studied at Trow Ltd. It has been found that for the slag cement
produced in Hamilton, a 50 percent cementitious hydraulic slag (CHS)/50
percent Type 10 portland cement (normal portland cement) blend has an
equivalent sulfate resistance to Type 50 portland cement (sulfate resistant ce-
ment) as shown in Fig. 6. Some typical test specimens from this study are
shown in Fig. 7. Such blends are now being used in concrete applications re-
quiring sulfate resistance.
While still at the early stages of development, recent research at McMaster
University has indicated that separately ground slag cement can be used as a
FIG. 5—Structure of pelletized slag at cement fineness.
FAILURE AT 191 DAYS

£ „ . , • 1.39 %
C109 PRISMS
50,000 vmt SO3 (AS N*2SQ,) SOLUTIONS
NORMALIZED FOR £ WATER ' °
o*>^
SO/SO CHS/TYPE 10
7 0 / 3 0 CHS/TYPe 10
TYPE SO ( 2 . 0 % C 3 A )
SO/SO CHS/TYPE 2 0
_] I I L
300 400 SOO 600 800 900
TIME IN SOLUTION (DAYS)
FIG. 6—Sulfate resistance of slag cement (CHS - cementitious hydraulic slag).
FIG. 7—Slag cement sulfate resistance testing, mortar cubes, and bars in 50 000 ppm SOj as
Na2 SO4, 400 days, (heit Xtay) 50 percent cementitious hydraulic slag (CHS)/50% Type 10 (nor-
mal) portland cement (12.2% CjA). No specimens show any deterioration. ^Right tray^ 100per-
cent Type 10Portland cement (12.2% CÂ). All specimens have disintegrated.
filler in asphaltic concrete mixes. This filler appears to be active and can be
used to replace up to 30 percent by volume of the asphalt cement in a typical
mix.
Base Stabilization—While slag cements are used in soil cement, cement
treated granular base, and lean concrete base, it is also possible to use granu-
lated and pelletized slag directly, or partially ground, to produce a stabilized
base. Utilization of granulated blast furnace slag as the binding medium in
stabilized base construction (typically 80 percent gravel or sand and 20 per-
cent granulated slag with 1 percent additional lime activator) is common in
Western Europe, particularly in France where it forms a major class of road
construction. Research by the author at McMaster University, and recent ex-
perience in France, have indicated the suitability of pelletized slag for this
process and the advantages of partially grinding the vitrified slag.
A satisfactory stabilized base was developed using 70 percent minus 19
mm crusher run air-cooled blast furnace slag, various percentages of un-
ground pelletized slag (30 percent less the minus 75 iim ground fines), and
minus 75 ^m ground pelletized slag (fines) to provide the cementitious ma-
terial. It can be seen from Fig. 8 that the required minimum strength is read-
ily achieved, and an aim compressive strength of 6.9 MN/m^ is met at the 4
to 5 percent fines level. Grinding the pelletized slag to get the desired net
fines in the total pelletized slag is generally more efficient from a strength
viewpoint than blending unground pellets to get the desired proportions. It
should be noted that accelerated curing conditions were adopted in the early
stages of this study (38 ± 2°C). However, as indicated by the compressive
strengths in Fig. 8 for curing at 21 ± 2°C (representative of field condition),
the required minimum strength is achieved in approximately 11 days for a 5
percent fines level, and higher fines levels can be used if earlier strength de-
velopment is required.
For actual field applications it is necessary to have an efficient grinding
procedure to give the desired fines level, preferably in the moist condition
from the pelletizer to eliminate the need for either dry or wet grinding with
their associated higher costs, and to avoid potential dust problems. It was
found that an Eirich R7 intensive mixer with a high energy rotor achieves
these requirements and large machines of this type are available.
Based on the laboratory study (including freeze-thaw resistance), it ap-
L E G E N D AND MIX PROPORTIONS
MIX = 5% SEALED CURING AT 38 i

20 MIX = 5%
MIX = 10% SEALED CURING AT 21 t
MIX = 15%
EACH MIX CONTAINS^

70 % CRUSHER RUN AIR COOLED SLAG
X7o TOTALLY GROUND PELLETS - • 75 jim
( 3 0 - X ) % UNGROUND PELLETS
- 4 75 mm to + 75 ^r
8.5%, WATER CONTENT
o
MINIMUM
ACCEPTABLE NSL PELLETS USED
STRENGTH (500psi) GLASS CONTENT 78%
28
CURING TIME, DAYS
FIG. 8—Compressive strengths for pelletized slag base stabilization specimens cured at 21
and 38°C.
pears that the cementitious nature of ground palletized blast furnace slag
makes it a viable and economic alternative to the use of portland and asphal-
tic cements in base stabilization. Future studies on this application for pel-
letized slag must include detailed fieldwork and a consideration of construc-
tion techniques adopted in France. Similar base stabilization processes
incorporating steel slag and iron oxides are described in the next section.
Steel Slag Production and General Uses /1,4,5/

While much slag from the various steelmaking processes (open hearth,
basic oxygen, and electric arc) is suitable for return to blast furnace burden,
there is still a large quantity available that can be effectively utilized in road
construction (Table 1). In considering steel slags, three points must be em-
phasized: there is typically a large variation in steel slags, even for the same
plant and furnace; the economic worth of steel slag for return to blast fur-
nace burden and other applications such as fertilizer must be recognized;
and a management strategy for all by-products must be developed. The vari-
ous steelmaking processes generally introduce restrictions in the uses of steel
slags in comparison to blast furnace slags. Steel slags have a high bulk den-
sity and potential expansive nature (volume changes of up to 10 percent at-
tributable to the hydration of calcium and magnesium oxides), while blast
furnace slags have a relatively low bulk density and are stable. Since serious
damage may result from indiscriminate use of steel slags in confined applica-
tions, potential long-term volume changes must be checked before such
usage. Obviously, steel slags should not be used in portland cement concretes
(unless shown otherwise by a detailed evaluation) since expansion will result
in rapid destruction of the concrete. However, there are many applications
where expansion is tolerable, has been controlled by suitable aging or treat-
ment with spent acids, or where sealing of the steel slag with an asphaltic
coating eliminates any problems. Further, the potential for selective use,
based mainly on the free lime content, has been developed.
Expansive Nature and Processing of Steel Slags—The expansive nature of
steel slags can be traced back to the steelmaking process in which the conver-
sion of pig iron to steel involves the controlled adjustment of various impuri-
ties,, and the addition of small quantities of constituents that give special
properties to the steel. Chemical compositions for typical Canadian open
hearth and basic oxygen furnace steel slags are compared with those of blast
furnace slags in Table 3. While the steel slag constituents are similar to those
of blast furnace slag, the proportions are quite different, with the higher iron
content being reflected in the high specific gravity of 3.2 to 3.5 for steel slag
compared to 2.2 to 2.5 for air-cooled blast furnace slag. The mineral compo-
sition of steel slags is fundamentally different from blast furnace slags where
the calcium and magnesium oxides are always combined in silicate and
aluminosilicate minerals, and consists essentially of calcium silicates, cal-
cium oxide-ferrous oxide solid solutions, oxides, and free lime. Steel slag
phase compositions are similar to that of a weak portland clinker and, while
the dicalcium silicate present is in the potentially metastable jS form, it ap-
pears to be inactive. The calcium and magnesium oxides are not completely
combined in steel slags, and there is general agreement in the literature that
the hydration of unslaked lime (free CaO) and magnesium oxide (MgO) in
contact with moisture is largely responsible for the expansive nature of most
steel slags. The unslaked lime hydrates rapidly and can cause large volume
changes in a few weeks. Magnesium oxide hydrates more slowly and contrib-
utes to the long-term expansion that may take several years to develop in the
field, even when old steel slag dumps are being used.
A separate problem that can occur with steelmaking slags is contamina-
tion with steel works rubbish such as refractories. Inevitably, some refractor-
ies get into the steel slag from furnace linings, but the major problem results
from using ladles and pits as waste receptacles in the furnace areas. Fortu-
nately, the value of steel slags is increasingly being recognized, particularly
for return to blast furnace burden, and contamination is avoided. Small
quantities of refractories do not appear to cause problems in highway appli-
cations.
Processing of steel slags for steel recovery is very important as it results in
an angular, generally well-graded, material that is relatively free of metallics,
and the recovered steel (2 to 4 percent of raw steel production) is a valuable
scrap. It appears that improved, or even new methods for removing the steel
from steelmaking slags would be highly desirable since this is a very space
and time consuming activity. Slag modifications to lower the viscosity in the
molten state, thus allowing the steel to separate, have been developed and
patented, but not used commercially. Similar processes are being developed
to give a steel slag suitable for cement manufacture. Since the steelmaking
slags are reduced in size, and water is involved during processing, it is likely
that hydration of calcium oxide occurs and the aging process is accelerated.
Processed steel slag is still high in iron, lime, and manganese, so that it is
possible to return most steelmaking slag (particularly the large sizes) to blast
furnace burden. However, there are still many operations where most of the
steel slags are not returned to burden in a relatively closed cycle for several
reasons: production of steels involving additions that prevent full or partial
return; equipment not available for return (particularly the fines); economics
not favorable; or the operation is not integrated with blast furnaces.
Expansion Tests—There is a large amount of steel slag used for applica-
tions such as bases, and many old steel slag heaps or fills are being built on,
or used for construction purposes. It is critical that these steel slags be checked
for potential expansion, since even aging for long periods in large dumps
does not guarantee the elimination of expansive behavior (particularly if un-
processed so large lumps are involved). Leachates are not a problem (Table 3).
It was considered desirable to develop a simple, economical, and rapid test
procedure for evaluating the expansion potential of steel slags. The pro-
cedure adopted involves preparing steel slag specimens and a non-expansive
control using the Standard Proctor method and stainless steel molds with
perforated base plates to allow for moisture movement during the immersion
period. The specimens are then totally immersed in a water bath at 82 ± 1°C
and the vertical expansion with time monitored. The 82°C test temperature
was selected on the basis of the initial expansion test series shown in Fig. 9.
Since the expansion levels of 5 to 9 percent observed at 82°C (about three
times that at 60°C) are similar to the levels of long-term expansion often ob-
served in the field, this test temperature was adopted. A short monitoring
period of one to seven days appears adequate to detect potential expansive
behavior. The actual steel slag gradation involved for a project should be
tested since the expansive behavior is clearly a function of gradation. Also,
surcharge weights to simulate overburden conditions can be used. Support
for the adoption of the laboratory accelerated expansion tests at 82°C was
given by the series of long-term expansion tests at 20 ± 1°C shown in Fig. 10
that resulted in expansion at 475 days of about half those observed in seven
days for tests at 82°C.
The test series given in Figs. 9 and 10, and other tests, have indicated that
aging in stockpiles (preferably after processing and in small quantities),
spent acid treatments, and the use of coarser sizes all tend to limit the poten-
tial expansion of steel slags. These results are in qualitative agreement with
STEEL SLAGS
FRESH OH + BOF
FRESH BOF
AGED OH + BOF (LABORATORY)
AGED OH + BOF (STOCKPILE)
OH-OPEN HEARTH
BOF-BASIC OXYGEN FURNACE
CONTROL SHOWED NO
EXPANSION
o
CO
z
X
111
TIME , days
FIG. 9—Expansion tests at 60 and Sl^C for various steelmaking stags.
EXPANSION TEST TEMPERATURE » 20 C

STEEL SLAGS
- • - FRESH O H + B O F - 4 75 mm
— i - FRESH OH + BOF
- • - AGED OH + BOF (LABORATORY) -4.76mm
~»- AGED 0 H + BOF (LABORATORY)
CONTROL SHOWED NO EXPANSION
O
CO
z
X
25 50 75 100 125 150 ' 450
TIME, days
FIG. 10—Long-term expansion tests for various steelmaking slags at 20°C.
field observations. It should be noted that aging steel slag in large heaps or
pieces is not very effective, as steel slag remains expansive for extremely long
periods if not directly exposed to weathering.
The discussion in this section has been concerned with the potential expan-
sion of steel slag that has not been coated with asphalt cement. As indicated
in following sections, the use of steel slag in asphaltic concrete results in an
excellent product, and the asphalt cement film coating the steel slag limits
potential expansion. However, the question still often arises concerning the
need for aging steel slag prior to use in asphaltic concrete. In the Hamilton
area, it has been found that if the finer sizes are used (minus 13 mm), aging
is not critical prior to use in asphaltic concrete since the watering and screen-
ing during processing, travel through the asphalt plant dryer and screens,
etc. allows for any immediate expansion. However, a minimum aging period
of 30 days is still recommended by many authorities, particularly for the
coarser minus 19 mm size asphalt mixes.
Steel Slag Utilization in Asphalt Mixes /1,4,5/

Mix Practices—While the main uses of available steel slags may continue
to be ballast, pavement bases and shoulders, fills, asphalt mixes, ice control
grit, etc., it would appear that the economics of handling and using a heavy
aggregate, the virtual elimination of any expansion related problems by the
asphalt cement coating, and the excellent performance of steel slag asphaltic
concrete mixes detailed in following sections (particularly skid resistance

qualities), favor usage in asphalt mixes (asphaltic concrete, coated chip-
pings, and slurry seals). The use of steel slags in asphalt mixes is highly local-
ized to steel production areas and such utilization is still not permitted by
some agencies, even though the performance record of steel slag asphaltic
concrete mixes is generally excellent.
Steel slag has been used in asphaltic concrete mixes in the Hamilton-To-
ronto area since 1969. There have been no problems with the quality and
durability of steel slag asphaltic concrete pavements, and any surface pop-
outs observed (due to traffic picking out softer inclusions such as refractories
and lime) are minor. Steel slag has also been used in hot mixes for winter
patching with the following advantages being demonstrated: fairly dry so
heating costs were reduced; retains its heat very well; and the high unit
weight and stability tends to hold patches in place.
Mix Characteristics—Typical average Marshall method physical charac-
teristics for open hearth and basic oxygen furnace steel slag mixes are sum-
marized in Table 4. These results, which are typical of current research and
the literature, indicate the following general features of steel slag asphaltic
concrete mixes: they are designed in the same way as any other mixes and no
problems have been experienced in meeting relevant specifications; they
combine very high stabilities (l'/2 to 3 times greater than most other mixes)
with good flow properties, resulting in a mix that resists rutting after cooling,
but is still compactible; they have excellent resistance to stripping of asphalt
cement from the steel slag as anticipated from the free lime contents of Table
3; and they have bulk densities 15 to 25 percent greater than most other
mixes that can result in a cost disadvantage on a volumetric basis.
The high stability of steel slag asphaltic concrete mixes offers a distinct ad-
vantage for applications where rutting resistance is required, such as indus-
trial roads and industrial parking areas subjected to heavy axle loads. Re-
search by the author at McMaster University has shown that the resilient
modulus (Af/j) values of steel slag asphaltic concrete mixes tend to be some 20
to 80 percent higher than those of natural aggregate asphaltic concrete mixes
at typical design temperatures (29°C). Using elastic analysis techniques such
as the Shell BISTRO computer program, this can typically result in a 37 mm
thinner steel slag asphaltic concrete pavement (163 mm) compared to a typi-
cal natural aggregate asphaltic concrete pavement (200 mm) for the same
critical strain levels. Obviously, if thinner steel slag asphaltic concrete pave-
ments can be used, this will help to compensate for any high bulk density dis-
advantages due to asphalt cement and transportation costs.
Skid Resistance—Previous experience with steel slag asphaltic concrete
mixes in several countries has indicated that its skid resistance is very satis-
factory for most applications. However, very little comparative evidence was
available until the Ontario Ministry of Transportation and Communications
(MTC) paved a number of skid resistance test sections on Highway 401, one
TABLE 4—Properties of steel slag asphaltic concrete mixes [1].
Marshall Designs
Property of Laboratory Open Basic Oxygen

Compacted Mix" Hearth* Furnace*
Stability (kN) 15.4 18.0

Flow index (mm) 4.3 5.4
Voids in mineral aggregate (% vol.) 20.5 20.4
Air voids (% vol.) 3.3 3.4
Asphalt cement content (%)'' 6.2 6.2
Density 2.872 2.869
Immersion-Marshall Stripping Test Results''
Stripping Rating Number

Retained
Steel Slag Stability, % (i) Unstripped (ii) Stripped
Open hearth* 73.5 3.0 4.0

Basic oxygen furnace* 82.5 3.0 4.0
"Determined using the standard Marshall method procedures.

*Recent gradations tend to be coarser.
•^Percent by weight of total mix.
''A stripping rating number of 3 indicates no stripping, of 4 indicates slight stripping, etc.
of the busiest freeways north of Toronto. Full details cannot be given here,
but the first year's performance summarized in Fig. 11 is indicative of trends
to date.
The dotted line on Fig. 11 is the minimum acceptable skid number (SN) at
100 km/h (60 mph) of 32. Sections 1 through 6 incorporate various traprock
asphaltic concrete mixes with Section 1 being the standard HLl mix used for
wear and skid resistant surfacings in Ontario. Section 7 is a 100 percent steel
slag mix, while Section 8 is a mix with steel slag as the coarse aggregate and
natural aggregates as fines (performance inferior to Section 7 as anticipated).
Section 9 is a 100 percent blast furnace slag mix, while Section 10 is a mix
with blast furnace slag as the coarse aggregate and natural aggregates as
fines (performance inferior to 100 percent slag mix again). Sections 11 and
12 are sand mixes and Sections 13 through 16 are open graded mixes (non-
slag) that are known to exhibit good skid resistance. It is clear that the all-
steel slag and all-blast furnace slag mixes have retained the highest skid re-
sistance (Sections 7 and 9), while Section 1 (HLl) has fallen well below the
minimum acceptable level. Recent measurements show a decrease in all skid
resistance levels with time, but confirm these trends after five years [8].
Polished stone values (PSV) for typical aggregates available in Southern
Ontario support the preceding general finding that steel slags and blast fur-
nace slags are superior to traprock in skid resistance; steel slag approxi-
INITIAL SN (Aug,. Sepr. 1974 )
LATEST SN I Oct. 1975)
INCREASED SN
MIN. RECOMMENDED SN
10 liN 17 II
SECTION NUMBERS
FIG. 11—Skid resistance of test sections after one year ('courtesy of Ontario MTCA
mately 59; blast furnace slag approximately 53; traprock approximately 46;
and limestone approximately 46. Also, a recent study of skid resistance in the
urban context has shown that steel slag asphaltic concrete mixes appear to
provide superior skid resistance to other mixes. The wet road accident rate
did not significantly exceed the dry road accident rate for steel slag asphaltic
concrete surfacings, while the reverse was generally true of other surfacings.
A Verglimit-steel slag asphaltic concrete mix has been used on a difficult ur-
ban pavement section in Hamilton recently to provide improved year-round
skid resistance.
Based on highway, laboratory, and urban studies, it appears that steel slag
asphaltic concrete mixes of adequately coarse gradation provide satisfactory
skid resistance for most applications. Indeed, for Ontario, they are the most
skid resistant aggregate currently available and are now included in the On-
tario MTC specifications. Highway trials using a blend of air-cooled blast
furnace slag (coarser portion) and steel slag (finer portion) in asphaltic con-
crete surface courses have proven most satisfactory with excellent skid resis-
tance developed. This type of mix would allow a much fuller utilization of the
finer steel slags not suitable for recycling.
From the previous discussion, it is possible to indicate the most promising
applications for steel slag asphalt mixes in order of relative merit: surface
course mixes that take advantage of the excellent skid resistance and good
wear resistance of steel slag asphaltic concrete (they also appear to be good
for reflection crack prevention); industrial paving where the high stability
and consequent rutting resistance is an important asset; winter patching;
slurry seals (design methods have been developed); and binder courses where
some depth reduction related to higher resilient moduli is possible.
Steel Slag Utilization in Stabilized Base

There is currently interest in Europe concerning the utilization of steel slag
in stabilized base, and at least one company, Pelt and Hooykaas B. V. of the
Netherlands, is using such a process commercially. The stabilized base con-
sists of 60 percent blast furnace slag (0 to 60 mm), 25 percent steelmaking
slag (0 to 15 mm), and 15 percent granulated blast furnace slag. This base
material is placed and compacted with standard highway equipment (at ap-
proximately 10 percent water content) and the results appear excellent.
Use of Iron Oxides in Base Stabilization

A somewhat related area is the use of various iron oxides in base stabiliza-
tion processes that also incorporate slags. For example, Merox in Sweden
(Granges Oxelosunds Jamverk AB) has developed Merolit, a stabilized base
of various grades that includes iron oxides. A typical composition is 70 per-
cent air-cooled blast furnace slag, 25 percent ground granulated blast fur-
nace slag, 3 percent iron oxides, and 2 percent hydrated lime plus a suitable
mixing water content. Much Merolit is used in plant areas, parking lots,
farm areas, and local pavements, with a growing demand. Iron oxides are
used in a similar way for base stabilization in steelmaking areas of Australia.
It would appear that the role of the oxides and hydrated lime is similar to the
activation role of steelmaking slags in stabilized base.
FerroaUoy Slags [/,</, 5]

The ferroalloy industry (principally ferrosilicon, ferromanganese, ferro-
chrome, and ferronickel) is the smallest of the ferrous industries and the
available slags tend to be localized. Silicomanganese slag (approximately 1.1
kg slag/kg of silicomanganese) is glassy textured, typically in large pieces,
and can be used for granular bases and engineered fill. Ferrochrome slag
(approximately 1.5 to 2.0 kg slag/kg of ferrochrome) occurs in dense 10 to 15
cm pieces with a vesicular surface and is used in road construction. Ferro-
nickel slag (up to 31 kg slag/kg of ferronickel) is often granulated and used
in bases and fills. For silicomanganese and ferrochrome slags the engineering
properties are similar to air-cooled blast furnace slag except that heavy slags
are involved (1800 to 2000 kg/m^). Leachates from ferroalloy slags do not
contain significant concentrations of toxic constituents (Table 2), but care
should be taken to ensure falling slags are not involved.
Nonferrous Slags [1,4,5]

The nonferrous slags considered—copper, lead, lead-zinc, nickel, phos-
phorus, and zinc—are those that appear to be produced in quantities that
make potential pavement applications attractive.
Copper Slag
Copper slags are usually discharged from the furnace at 1100°C to 1300°C
and when cold have a black color and stoney, or glassy, appearance. As a
rule, the specific gravity will vary with the iron content, from a low of 2.8 to
as high as 3.8. Current research and applications have shown that several
nonferrous slags (copper, lead, and nickel) have great potential as pozzolans.
It has also been suggested that a similar cementitious blend of finely ground
copper or nickel slag and portland cement could be used as the binder for
base stabilization in road construction. A summary of the available informa-
tion on ground nonferrous slag (copper, lead, and nickel)-portland cement
binders indicates the following: must be vitrified; must be finely ground
(about Portland cement consistency); slag content of the binder must not be
too high or reactivity is limited; and the slags act as pozzolans with a rela-
tively slow reactivity.
In areas where large amounts of copper slag are available, use in railway
ballast is common. The high unit weight of the slag means that haulage on a
volume basis is more expensive than for rock ballast. It would appear that
the only significant current use of copper slag in pavement construction is as
granular base and engineered fill. This usually involves crushing and sizing
to ensure adequate compaction. As with most nonferrous slags, potential
leachate problems should be checked for before use. The main reason why
more copper slag is not used as fill and base is the often remote location of
copper production from highway construction, and the high unit weight of
copper slag.
Lead, Lead-Zinc, and Zinc Slags

Since lead and zinc are often related as coproducts in both source and
metallurgical treatments, the various combinations of these nonferrous
slags—lead, lead-zinc, and zinc—will be considered as one group. Slags of
this group are often black or red in color, cohesionless, glassy, and have
sharp angular particles that are cubical in shape. The specific gravity tends
to be as high as 3.6. Slag generation is highly dependent on specific processes
and sources, and generally speaking, the quantities available are relatively
small. Lead, lead-zinc, and zinc slags should generally be usable in fill con-
struction and as granular base materials. However, they are often produced
at some distance from construction projects and their high unit weights in-
crease transportation costs on a volume basis.
Nickel Slag
The use of nickel slags in road and railway construction has been the sub-
ject of a fair amount of research in Canada. The iron-silicate slag is dumped,
and allowed to air cool so that a "clinker" is formed. This material is then
crushed and processed as required, but the quantity of fines produced tends
to be low. The specific gravity tends to be as high as 3.5. Nickel slag from
Sudbury has been used for 40 years as railway ballast and has established it-
self as being the best material used. The processed slag is heavy (specific
gravity of 3.2 to 3.6), tough, hard, angular, and packs well under ties to form
a better support for track than rock ballast. Slag placed in track over 20
years ago is still in good condition. However, the density of the slag means
that transportation on a volume basis is more expensive than rock ballast.
More energy is required to compact the material into the track bed, a prop-
erty that makes the material valuable once in place. This slag meets or ex-
ceeds typical specifications for rock ballast. The other applications for nickel
slag are similar to those discussed for copper slag, and a large amount is used
as fill and base.
Phosphorus Slag
At most elemental phosphorus plants, molten slag is tapped into a pit, air-
cooled or cooled by applying water on the slag surface, and the material is
then crushed for use as ballast or aggregate. However, both granulated and
expanded phosphorus slags are produced, and there have been some experi-
ments with pelletizing in the United States. (Most phosphate rocks contain
between 30 and 200 ppm uranium, and most of this uranium comes out in
the phosphorus furnace slag. However, Tennessee Valley Authority tests
show that radiation is negligible and presents no hazard.) While many of the
pavement applications have been at the experimental stage, the present and
potential uses for phosphorus slag closely parallel those for iron blast furnace
slag: crushed air-cooled phosphorus slag (concrete aggregate, ballast, engi-
neered fill, granular base, asphaltic concrete); granulated phosphorus slag
(cement manufacture); and expanded phosphorus slag (lightweight aggre-
gate).
BoUer Slag
Boiler slag [/] is generally the bottom ash from coal-burning, electric gen-
erating plants that accounts for 25 percent of the total ash production. (Some
is also produced during refuse incineration.) Large quantities are available in
the vicinity of these generating plants, which are more prevalent near large
metropolitan centers. There are many uses being made for this material such
as: seal coat aggregate; sanding and deicing aggregate; asphaltic concrete
aggregate; and fill or base. Economically, boiler slag is very competitive, un-
less it has to be hauled for any great distance or requires extensive process-
ing.
Conclusion
While the major emphasis in the review was cementitious applications for
iron blast furnace slags and the use of steel slags in asphaltic concrete, there
are also considerable amounts of nonferrous slags available for use as pave-
ment construction materials. Applications research on nonferrous slags to
extend their utilization, particularly as cementing agents, should be a prime
topic for construction materials researchers. With the present indications of
pending material and energy shortages, there is a new concern that all bypro-
duct slags be utilized. It is critical that any utilization strategy accounts for
contained material and energy values, and that bulk applications are not
substituted for full resource recovery.
Acknowledgments
The author wishes to thank National Slag, Eirich Machines, Standard
Slag Cement, Stelco, and Dofasco for their generous support of the Con-
struction Materials Laboratory at McMaster University. The work of C.
Kim, R. Gupta, R. D. Hooton, C. Low, M. Lee, and E. Krakowski in the
Laboratory has been of particular significance in the development of a
greater utilization of slags. Their contribution to this review is gratefully ac-
knowledged. This review is based on an updating of a presentation for the
Third Industrial Minerals Congress, 1978 [1].
References
[/] Emery, J. J. in Proceedings, Third Industrial Minerals International Congress, Paris 1978,
Metal Bulletin, London, 1978, pp. 127-142.
[2] Mineral Commodity Summaries—1980, Bureau of Mines, U.S. Department of the Interior,
Washington, Jan. 1980.
[3] Smith, M. A., Resources Policy, Vol. 1, No. 3, 1975, pp. 154-170.
[4] Use of Waste Materials and By-products in Road Construction, Organization for Economic
Co-operation and Development, Paris, 1977.
[5] Gutt, W. and Nixon, P. J., Materiaux et Constructions, Vol. 12, No. 70, 1979, pp. 255-305.
[6] Emery, J. J. in Concrete International 1980—Lightweight Concrete, Construction Press,
London, 1980, pp. 36-46.
[7] Hooton, R. D., Ph.D. thesis, McMaster University, Hamilton, 1981.
[8] Ryell, J., Corkhill, J., and Musgrove, G., "Skid Resistance of Bituminous Pavement Test
Sections: Toronto By-pass Project," Transportation Research Record 712, Transportation
Research Board, Washington, 1979.
G L. Evans1 and R. G. Hicks2
Properties of Marginal Aggregates

Treated with Asphalt Emulsion
REFERENCE: Evans, G. L. and Hicks, R. G., "Properties of Marginal Aggregates

Treated with Asphalt Emulsion," Extending Aggregate Resources, ASTM STP 774,
ABSTRACT: The suitability of asphalt emulsions for upgrading marginal aggregates has
been evaluated. The aggregates tested included a high quality basalt, two low quality
marine basalts, and a fine grained dune sand. Mix properties evaluated included
diametral resilient modulus and diametral fatigue life for both as-compacted specimens
and specimens conditioned by moisture exposure. Layered elastic design principles were
used with the dynamic test results to develop layer equivalencies for emulsion treated
marginal aggregates compared with hot mix asphalt concrete. The results indicate that
beneftciation of marginal aggregate with asphalt emulsions should make acceptable pav-
ing mutes, particularly for low volume roads.
KEY WORDS: aggregates, concrete, beneficiation, asphalt emulsions, mechanical

properties, evaluation
The problem of supplying construction aggregates to Oregon's coastal

areas is evidenced by current practices of importing quality aggregates to ag-
gregate deficient areas. Much of this aggregate comes from as far as 100 km
(60 miles), resulting in poor energy efficiency and high costs from trucking
aggregate through coastal mountains and returning empty. Environmental
restrictions associated with the usage of existing and proposed rock quarries
in this area have compounded the shortage [1-4].3 Faced with continually in-
creasing fuel costs and decreasing resource availability limiting the amount
of quality aggregate that could potentially be exported, alternative methods
must be developed to supply the coastal area with construction materials.
The purpose of this paper is to evaluate the option of treating locally
'Geotechnical Engineer, Williamette National Forest, Eugene, Ore. 97440.

2
Professor, Department of Civil Engineering, and director, Transportation Research Insti-
tute, Oregon State University, Corvallis, Ore. 97331.
3
Copyright by ASTM Int'l (all 119 rights reserved); Thu Dec 22 07:41
available and abundant marginal aggregates with asphalt emulsions. The

scope of work includes: (1) the evaluation of the resilient and fatigue proper-
ties for three marginal and one quality aggregate, and (2) the determination
of layer equivalencies for the aggregates tested compared with conventional
asphaltic concrete.
Materials and Sample Praparation

An outline of the experimental program is given on Table 1. The program
was carried out in four main stages, and included:
1. Selection and determination of aggregates and basic aggregate proper-
ties.
2. Asphalt emulsion mix design.
3. Dynamic test phase.
4. Analysis of the dynamic test results.
Aggregate Types
The aggregates used in this study (judged by availability and abundance),
are marine basalts, which exhibit poor durability characteristics, and sands,
which have low stabilities and excessive amounts of fines. Four aggregates
were selected as follows: (1) Eckman Creek marine basalt, (2) Berry Creek
marine basalt, (3) Oceanlake basalt, and (4) dune sand.
TABLE 1—Asphalt Emulsion Test Program.
Select Mix Experiment Analysis

Aggregate Design
Determine Select Dynamic Modulus (M^)—three samples

Aggregate Emulsion per material. Test after 1, 3, 10, 20, and
Properties. Type. 40 days and ultimately cured. Test
samples again after vacuum saturation
and water soaking.
Trial mixes at various

emulsion and
water contents. Fatigue—12 samples per material.
Test six samples after ultimate cure and
six after vacuum saturation and water
soaking.
Compaction tests at
optimum emulsion
an(} water contents.
EVANS AND HICKS ON ASPHALT EMULSION TREATMENTS 121
Aggregate Properties
Standard indicator and degradation tests were performed on all ag-
gregates. The results, given in Table 2, include specific gravity and absorp-
tion (ASTM Test for Specific Gravity and Absorption of Coarse Aggregate
(C 127-77) and ASTM Test for Specific Gravity and Absorption of Fine Ag-
gregate (C 128-78)) and optimum moisture-density characteristics (ASTM
Tests for Moisture-Unit Weight Relations of Soils and Soil Aggregate Mix-
tures Using 5.5-lb (2.5-kg) Rammer and 12-in. (304.8-mm) Drop (D 698-78)).
Also included are results from tests such as Los Angeles Abrasion (ASTM
Test for Resistance to Abrasion of Small Size Coarse Aggregate by Use of the
Los Angeles Machine (C 131-76)), California Durability (AASHTO T-210)
and Sand Equivalent (AASHTO T-176) as well as typical roadway specifica-
tions. An open gradation was used for the marine basalts while the as-
received gradation was used for the dune sand (Table 3). Only the Oceanlake
basalt met all the specifications required for quality rock.
Emulsions
The emulsions used for this testing include a CMS-2 for the open graded
basalts and a CSS-1 for the sand. The CMS-2 is a medium setting cationic
emulsion, while the CSS-1 is a slow setting cationic emulsion, which allows a
greater mixing time. The properties of the emulsions used comply with
ASTM Specification for Cationic Emulsified Asphalt (D 2397-79).
Emulsion Mix Design

The optimum emulsion and water contents were determined based on tests
to establish the effect of fluid content (water and emulsion) on workability of
a mix, the time to initial break, and the thickness and percent of coating of
asphalt on the aggregate particles [5]. The results given in Table 4 were used
in the dynamic testing program.
Dynamic Test Equipment and Procedures

To simulate the performance of the emulsion mixtures in an actual pave-
ment section, dynamic tests were conducted to evaluate their stress-strain
and durability characteristics. The properties determined included the
dynamic modulus and fatigue and durability characteristics. A diametral
testing apparatus similar to that developed by Schmidt [6,7] was used to
determine modulus and fatigue characteristics of all test specimens. Dura-
bility characteristics were determined by vacuum saturating and water soak-
ing specimens to find the effect of moisture conditioning on modulus and
fatigue life. All specimens were vacuum saturated for a period of 2 h and
soaked in a water bath for seven days.
| . H 1/5
s ^ Z '^•
1^
tl H >/5
ri 3 «J
UQ 8
•O -3 3
\ o r~- t~-
00 ro
^ (N en :
g
l o r-~ Lo
- H TT vO
II
So tu
O O O
-^ rS uS
00 "J OO
r j (N <N
f3^
E
- - V
c
<N 1rt
O OJ
J4 a
1
5§ 1
1^ 1B-l
^ 13
pa u ^ z
VI H
TABLE 3—Aggregate gradation.
Percent Passing
Marine Dune
Sieve Size Basalt Sand
19.0 mm O'l in.) 100

12.5 mm ('/2 in.) 57
4.75 mm (No. 4) 10
2.0 mm (No. 10) 3 100
0.30 mm (No. 50) 1 88
0.074 mm (No. 200) 1 1
TABLE 4—Mean values and coefficients of variation for density of laboratory specimens.
Mean Total Density

Emulsion Added Number
Content, Water, of Mean, Standard
Aggregate % % Specimens pcf° Deviation
Berry Creek
marine basalt 6.0 1.7 31 124.8 3.3
Eckman Creek
marine basalt 6.0 2.7 24 130.4 2.9
Oceanlake basalt 5.0 1.1 36 125.9 2.0
Dune sand 7.0 7.2 28 114.1 2.5
'1 pcf = 16.02 kg/m^.
Modulus Tests
For the diametral modulus, specimen 102 mm (4 in.) diameter by 63 + 8
mm (2.5 ± 0.3 in.) were prepared using a kneading compactor according to
ASTM Method for Compaction of Test Specimens of Bituminous Mixtures
by Means of California Kneading Compactor (D 1561-76). For each ag-
gregate type, three samples were tested after time intervals of about 1, 3, 20,
and 40 days of air curing at 24 ± 3°C (75 ± 5°F). The resilient modulus of
each material type was determined over a range of confining stresses from 0
to 41 kN/m^ (0 to 6 psi) and a dynamic load range from 89 to 334 N (20 to 75
lb). The deviator stress pulse had a duration of 0.10 s and a load frequency of
30 cycles per minute.
Fatigue Tests
For fatigue testing, six samples of each aggregate type were failed over a
range of 500 to 50 000 load repetitions. All samples were air cured for four
days at 24 ± 3°C (75 ± 5°F) and placed in a desiccator at 79.4 kg/m^ (23 in
mercury) vacuum until all of the moisture was removed to obtain the ultimate
cured condition. Tests were also run on specimens after moisture condition-
ing. Loads were applied at 60 cycles per minute and 0.10-s duration until
failure. Failure of the sample was determined by an approximately 12 mm
(V2 in.) wide vertical crack across the diameter of the sample or after ex-
cessive deformation of the sample.
Dynamic Test Results
Mix Properties
Table 4 summarizes the properties of the specimens according to the ag-
gregate type. The highest densities were obtained with the Eckman Creek ag-
gregate, probably attributable to excessive degradation characteristics [3].
The dune sand specimens had the lowest density.
Resilient Modulus
The resilient modulus of each aggregate mix was determined as a function
of curing time and after the water-conditioning procedure.
Modulus Time—A relationship between the modulus and the number of
days cured for a typical mix is shown in Fig. 1. The data points represent the
average of three samples tested. The slopes of the curves for all of the mixes
are quite similar, with the exception of the dune sand. For this mix, the curve
rises much faster in the first ten days of cure, probably caused by the hydra-
tion of the Portland cement additive. This substantiates other findings [7]
that note that the inclusion of this material significantly decreases the time
required for curing emulsion mixes. Also shown is the influence of confining
pressure (a^) on the modulus. Generally, the confining stress has a greater in-
fluence on the specimens in the early stages of cure. As the specimens
become stiffer, the confining stress has less effect.
Table 5 summarizes the ultimate modulus for all materials tested. The
Eckman Creek mix resulted in the highest ultimate modulus values, followed
by Berry Creek, Oceanlake, and dune sand mixes, respectively. The Eckman
Creek aggregate degraded considerably upon compaction, resulting in a
significantly higher density and denser gradation than the other mixes. As
the same compactive effort was used for all of the mixes except the dune
sand, this would explain the magnitude of the modulus values obtained for
this mix. The dune sand, although lacking strength gain from coarse particle
grain interlock, contained a higher asphalt content than the other mixes, and
1.5 percent portland cement. The Oceanlake mix contained less residual as-
phalt than the other mixes. The open-graded mixes, having higher densities,
resulted in higher modular values, which is consistent with results presented
by Hicks et al [8].
Ultimate
'""" "l Vocuum s a t u r a t e d at lOImm
- ^ = 1 g = [3-}(4m)H9,
0 — I Vacuum saturated at
© — f 584mm (23 inJHg.^
0 03= 21 k N / m ^ t3psi)
B 03=41 k N / m ^ (6p5i)
Cure and Test-Temperoture = 2 4 ± 3 ° C (75+5°F)
I . . . . I ' . i I I I . 1 . 1 . 1
10
DAYS CURE
FIG. 1—Variation in modulus with time of curing. Berry Creek marine basalt II psi = 6.59
kN/m^}.
Modulus versus Conditioning—The effect on modulus of conditioning is

also given in Table 5. The greatest loss in modulus resulted from vacuum
saturating and water soaking the Eckman Creek test specimens. This mate-
rial lost approximately 45 percent of its ultimate stiffness after 102 mm (4 in.)
vacuum saturation and about 60 percent after the 584 mm (23 in.) vacuum
saturation-water soak process. The dune sand generally experienced a
smaller reduction in stiffness at both levels of vacuum saturation.
It is evident that both the confining stress and amount of vacuum satura-
tion have a significant effect on the stiffness. Generally, testing at a higher
confining stress and after a lower level of vacuum saturation results in a
higher stiffness value.
Fatigue Characteristics
Diametral fatigue curves were obtained for each of the study aggregates
using unconditioned and moisture-conditioned specimens. One of these
curves is given in Fig. 2. Table 6 summarizes the fatigue equations for each
of the mixes tested.
In comparing the fatigue life characteristics of laboratory-tested open-
graded emulsion mixes with more conventional mixes, the intrinsic nature
(high void content) of these materials must be considered. The extremely
high void content (~ 25 percent) allows for much greater flexibility than con-
ventional mixes, and when used in pavements, they seldom fail in fatigue [9].
i(N^ O a^ -^ •^'^'O ^-HTTO
-H • (N •
III n o ^ r^r4(N
ii
^roro mô^
r-( ^ ^ ^ (N
f
ill o o (N ON O <N
5>l -H (N
O ^
<N <N (N <N
00 O
T T O O i/>»H^ vOfS(N 00(NO
LO vO (S O fO Q 1-1 (N -O
r- ô ON 2;^: ^ r^ <o r-- pn • ^
"1^ ^ fvl <N ^ H <N ^ <N (N
1-1
m
<;
O f O ^ 0 < ^ ^ O r O O O f O ^ I
H 6s
J3
e
1
c 300
f ^^.Unconditioned, 0 seating lood
2
X 200 ""^^ Conditioned, 0 seating load
z
? 0 D
z
^ 100 ^''"'"^tr^ •
E
to^ n ^~~-->^^^
-
" •
lli t
^ ® ^ \ ^
• ^^*^^~^
i 50
40
Unconditioned, 4 4 . 5 N seating lood ^ ^ = 5^s> •
30 - ©
1 1 1 1 1 1 1 1 J 1 1—1 1 i n l 1 LQ 1 1—J 1 • 1
10,000
NUMBER OF REPETITIONS
FIG. 2—Effect of conditioning (and seating load) on diametral fatigue results, Eckman Creek
marine basalt.
TABLE b—Summary of fatigue equations (Nj = k (1/t)"
Aggregate
Berry Creek, unconditioned,

combined seating load 7.67 X 10-'' 1.53 0.55
Berry Creek, conditioned,
0.0 seating load 8.50 X 10-" 1.72 0.78
Eckman Creek, unconditioned,
0.0 seating load 1.78 X 10-* 2.44 0.60
Eckman Creek, unconditioned,
44.5iV seating load 1.61 X 10-* 2.09 0.88
Eckman Creek, conditioned,
0.0 seating load 3.65 X 10-^ 2.61 0.69
Oceanlake, unconditioned,
combined seating load 1.63 X 10-^ 1.38 0.93
Oceanlake, conditioned,
0.0 seating load 6.59 X lo-* 1.63 0.47
Dune sand, unconditioned,
44.5Af seating load 9.13 X 1 0 - " 3.20 0.73
Dune sand, conditioned,
44.5iV seating load 1.56 X 19-26 7.09 0.96
NoTEÂy = number of repetitions to failure.

e = initial tensile strain, in./in.
k,n = constants.
r = correlation coefficient.
In the laboratory testing of the open-graded materials, a majority of the

samples failed primarily from plastic deformation, rather than from a dis-
tinct vertical crack through the center of the specimen.
The effect of water conditioning on fatigue for one of the study aggregate
mixes is also given in Fig. 2. As seen, the conditioned curve is essentially the
same as for unconditioned specimens; however, a considerable shift is ob-
tained due to the effect of the seating load. All of the conditioned specimens
were tested without a seating load, while the unconditioned curves were
developed from samples tested both with and without a seating load.
Recommended Modular Values

As the retained stiffness at 41 kN/m^ (6 psi) confining stress for each of the
aggregate mixes conditioned with the 101 mm (4 in.) vacuum saturation pro-
cess was greater than 1379 MN/m^ (200 000 psi), this value was used in the
development of structural layer equivalencies. A confining stress of 41
kN/m^ (6 psi) was chosen to represent conditions existing in an actual pave-
ment. This is the standard modulus value also recommended for design of
open mixes by Hicks et al [9], indicating that the modular values for all of the
open mixes (including those with marginal aggregates) are at an acceptable
level.
The low levels of stiffness reduction found for the dune sand mix treated
with Portland cement are consistent with results found by Schmidt [7].
Schmidt also found that the modulus of exposed and unexposed cement
treated sand mixes continue to increase for a considerable period of time.
The sand tested by Schmidt was comparable to the dune sand studied here;
however, a harder based asphalt was used in the emulsions in his study
(85/100 penetration compared to 100/250). This resulted in conditioned
modular values greater than the 1379 to 1723 MN/m^ (200 000 to 250 000
psi) found in this study. From this and similar test data [7] indicating that
much higher modular values are obtained from lower penetration base
asphalts, the stiffness of the dune sand is likely to be improved significantly
by using a harder based asphalt emulsion such as CSS-lh. For the develop-
ment of structural layer equivalencies for the dune sand, a value of 2068
MN/m2 (300 000 psi) is assumed.
Recommended Fatigue Relationships

Since little or no information is available relating diametral fatigue curves
of open-graded emulsion mixes to actual performance obtained in field con-
ditions, curves developed by Santucci [10] and confirmed by Hicks [9] and
Kallas [11] are used for further design analysis. Judging from the modular
values obtained for the study aggregate mixes, these curves are felt to provide
a reasonable representation of the fatigue behavior of the study aggregate
mixes. The discrepancies between the fatigue results obtained in this study
and actual field conditions might be remedied by further testing varying test
factors such as the seating load, load duration, and frequency.
Development of Structural Layer Equivalencies
Methodology
Layered system elastic theory principles [12\ are used together with proper-
ties of the study aggregates treated with emulsified asphalt and of hot mix
concrete to determine relative layer equivalencies. These equivalencies com-
pare required pavement thicknesses of the study emulsion mixes with
thicknesses for a high quality hot mix pavement over various base and
subgrade conditions. Pavement failure criteria are defined by establishing
the critical strain levels in a structural section for a given number of load
repetitions. With the aid of a computer, the maximum tensile strain levels in
the pavement mixes are calculated over a range of pavement thicknesses. Us-
ing this information, a pavement thickness is determined to limit strain in
the bottom of the surfacing layer (fatigue) and in the top of the subgrade
(rutting) for various amounts of traffic.
Failure Criteria
The fatigue behavior of open-graded emulsified asphalt pavements devel-
oped by Santucci [10] are given in Fig. 3 for the emulsified asphalt mixes and
Fig. 4 for cement modified asphalt emulsion mixes. These curves were devel-
z
<
10= 10°
FIG. 3—Fatigue criteria Jor asphalt and emulsion mixes (1 psi = 6.89 kN/m^).
•"I- 1 - r T-
2000
1000
X - 40,000 . :
2 500 " 200,000 ^~^_^
' i.soo.ooo ~Z^~~^~~^S:ss~^^!——-—.^

<
\
'.
1 1
-
10'' 10"
FIG. 4—Fatigue criteria for cement-modified emulsified asphalt mixes (1 psi = 6.89 kN/m^).
oped from extensive laboratory and field testing. They have also been proven
to closely simulate actual fatigue behavior of other emulsion mixes [9,11].
The subgrade strain criteria, also given by Santucci, is shown in Fig. 5.
This curve has also been shown to closely simulate actual conditions for
open-graded asphalt emulsion mixes [9] for a limiting rut depth of 9.5 mm
(% in.). The critical strain levels given by these relationships were used for all
of the material types in this study.
Development of Layer Equivalencies

With the failure criteria known, the maximum strain levels in a given pave-
ment system were determined in order to develop equivalent thicknesses of
the marginal aggregate mixes and hot mix asphalt concrete. These strain
levels were determined for the typical pavement cross-sections shown in Fig.
6. ELSYM5, a layered system analysis computer program [12], was used to
make the necessary calculations. Loading was assumed to be an 80 kN
(18 000 lb) single axle applied by two 20 kN (4500 lb) circular loads with ap-
plication pressures of 586 kN/m^ (85 psi) and spaced 213.9 mm (13.32 in.)
apart. Poisson's ratio was assumed to be 0.35 for all layers. Cement-modified
mixes are reported to have slightly lower values {13\; however, this is not ex-
pected to affect the results significantly. The maximum strain was deter-
mined at the bottom of the surfacing layer and at the top of the subgrade for
varying values of surfacing mix and subgrade modulus and for different sur-
facing and base layer thicknesses. With this information, layer equivalencies
can be determined by obtaining the critical strain levels from Figs. 3, 4, and
5 for a given number of 80 kN (18 000 lb) equivalent axle load (EAL) applica-
tions and then entering a surfacing layer thickness versus strain charts to ob-
o —
lO"* 105 ,06
NUMBER OF REPETITONS
FIG. 5—Subgrade strain criteria.
El E,
,"1
E2 = 4 5 , 0 0 0 psi E2 = 15,000 psi
h2 = 12 in. "2 = 12 in.
¥//////////////////,
£3= 30,000 psi
W////////////M,
E 3 = 10,000 psi
(a) "Good Subgrade" (b) "Fair Subgrade"
E2= 4 , 5 0 0 psi E j = 4 , 5 0 0 psi

h5= 12 in. tij, = 2 4 in.
%
E = 3 0 0 0 psi
(c) "Poor Subgrade l"
E = 3000 psi
(d) Poor Subgrade2'
FIG. 6—Pavement cross-section analyzed (I psi = 6.59 kN/m^, 1 in. = 25.4 mm).
tain the required layer thickness. The hot mix modulus was assumed to equal
2758 MN/m^ (400 000 psi), the cement-modified emulsion mix 2068 MN/m^
(300 000 psi), and the marginal emulsified asphalt mixes 1379 MN/m^
(200 000 psi) for the development of layer equivalencies.
The layer equivalencies determined from Santucci's fatigue criteria are
shown in Tables 7 and 8. The layer equivalencies determined from rutting
criteria are given in Table 9. An examination of these tables shows the design
thickness is controlled by fatigue criteria in all cases except for the "Poor
Subgrade 1" condition, where a greater thickness is required to preclude rut-
ting in the pavement. In Tables 7 and 8, no general trends are evident in the
layer equivalencies obtained. The average layer equivalency values in open-
graded emulsion mixes is 1.27 for fatigue and 1.24 for rutting. For the
cement-modified mix the average layer equivalency for rutting is 1.09. This
"layer equivalency" indicates that this multiple of layer thickness is required
for an emulsion mix over that structural thickness required using dense-
graded asphalt concrete. However, in Table 8, it is evident that the cement-
modified layer equivalencies based on fatigue criteria are highly dependent
upon the level of traffic. Significantly greater relative thicknesses are re-
quired at lower traffic repetitions, this difference diminishing with increasing
traffic level. This difference arises because of the significantly lower tensile
strains allowed for the cement-modified mixes in comparison with the hot
mix.
Based on the results of this analysis, the layer equivalencies given in Table
10 are recommended to provide the most conservative design thicknesses.
These results are also compared with layer equivalencies presently specified
by various agencies in the Pacific Northwest [9]. An important consideration
in the use of the layer equivalencies are their variation with the resilient
modulus of the mix. Layer equivalencies determined in this study for open
graded emulsion mixes are given in Fig. 7 over a range of surface moduli and
for both the fatigue and rutting models. As the same allowable subgrade
strain is applicable to both open-graded emulsion and cement-modified
emulsion mixes, the rutting criteria layer equivalencies apply to both types of
mixes. As seen from the curves, the open-graded emulsion layer equivalency
depends significantly upon the subgrade condition and the modulus of the
mix, especially at low modulus and poor subgrade conditions. A similar
variation with modulus exists for the cement-modified material, however, the
difference resulting from the various subgrades is less pronounced and
because of this, level of traffic and modulus of the mix appear to be the most
significant factors affecting the magnitude of these layer equivalencies.
Conclusions
The experimental program presented in this paper provides a valid ap-
proach to evaluate stiffness and durability properties of quality aggregate,
marginal marine basalt, and cement-modified dune sand emulsion mixes.
O ' O <N O 00 <N

Tf •5I-' - 4 00 vd rt
lO 0 0 ( N 00 r-- (N r-- ^^ ( N
X r-' uo ^ o^ r-* '-H ON r~-* •^*
00 O
r- r-- <N r^ r- fS i-H uo <N
-fc TT H ^ r-' iW '-^ t--" i/i i-H
c ^ 0 0 (N (N O ; <N ^ 0 0 (N
o rS (N w >0 -"T '-H -O Tf' ^
•a
I fO (N O (N <N
K
a
.1
II
H
.B5
lil 1^
.a
00 00
1- O —1 o oo ^_
•*• TJ-' rt' OO' '£> ^
i
"h O 00 tN (N r- "H
o K I/:) w a\ r-' •-/) (N O
VO VO l O ^ T-H
^ ^ O <N O
00 r-" t~-' —' rt
I
13
00 (N 00 r j o o r-- fn fN o
irj lO •^' TH 1-H K Ô ' O i-H •p-^"
XI
1
•a
•S
•^_ 00 -rr o4 o
t -^ oo r-- 'd TH \ri ^ -^ ^ -^
J3
!
t-- 00
00 O •'T ^ O t^ O t^ fS O
(N (N (N T-H ^ f^ VO ^ © - ^ i-H •^ TT f^* T-< T-l
\0 ft 'n -' y (S(N<Nrt^

S
•a
a .H .H
Is • ill
.S
ill
s
I J I .§ 11
.2 4) U S e
i!iir'I iSlI
•5.3
i : * ^1 ^ - •«ls B
< UKJ J
Q
i'l
<N ( N <N ( N (N iy-> >
• ^
^ 1
•5 ^
'^
M t H
ft ffsrÔ
:i>2o s
< 1 a: c Q -a
"3 -2 c C
^(2 §>!
0£ D [£ 6^
Copyright by ASTM Int'l (all rights reserved); Thu Dec 22 07:41:33 EST
Indian Institute of Technology Kharagpur pursuant to License Agreement. No furth
50 60 70 80 90 100 200 300 400

MODULUS 10* psi
FIG. 7—Variation of layer equivalency with pavement modulus (1 psi = 6.89 kN/m^).
The test results and discussion substantiate the idea that the marginal ag-
gregates can be treated with emulsions to provide acceptable pavements. The
marine basalts are recommended for construction in open-graded mixes
treated with CMS-2 emulsion, while the dune sand should be treated with a
CSS-lh emulsion and a small amount of portland cement. The stiffness of
these mixtures after moisture exposure does not fall below conventional
modular values, and the fatigue life does not appear to be affected by such
conditioning.
Little or no information is available on laboratory fatigue characteristics of
open-graded mixes. When comparing the results presented with those from
other test methods and materials, the diametral test does not appear to be ac-
ceptable for open graded mixes.
Finally, the layer equivalencies developed here allow for the use of mar-
ginal aggregates with standard design procedures for practically any required
conditions. These should also allow for cost comparisons to be made easily
once the prices of the emulsion and hot mixes are determined for a given
situation. The findings of this study indicate that marginal aggregates can be
used to provide acceptable roadways that may result in significant energy and
construction cost savings.
Acknowledgments
This work is the result of research sponsored by the Oregon State Univer-
sity Sea Grant College, supported by NOAA Office of Sea Grant, U.S. De-
partment of Commerce, under Grant No. NA79AA-D-00106. The U.S. Gov-
ernment is authorized to produce and distribute reprints for governmental
purposes notwithstanding any copyright notation that may appear hereon.
References
[1] Chintakovid, V., "Evaluation of Aggregate Needs and Problems Along the Oregon Coast,"
Master's project, Oregon State University, Corvallis, 1979.
[2] Clemmons, G. H., "An Evaluation of Coastal Oregon's Marginal Aggregates," Master's
thesis, Oregon State University, Corvallis, June 1979.
[3] Evans, G. L., "Properties of Marginal Aggregates Treated with Asphalt Emulsion,"
Master's thesis, Oregon State University, Corvallis, Aug. 1980.
[4] Chang, C , "Evaluation of Marginal Aggregates Treated with Portland Cement," Master's
thesis, Oregon State University, Corvallis, March 1980.
[5] "Mix Design Procedures for Open and Dense Graded Emulsified Asphalt Pavements,"
Federal Highway Administration, Region 10, June 1976.
[6] Schmidt, R. J., "A Practical Method for Measuring the Resilient Modulus of Asphalt
Treated Mixes," Highway Research Record 404, Highway Research Board, 1972.
[7] Schmidt, R. J., Santucci, L. E., and Coyne, L. D., in Proceedings, Association of Asphalt
Paving Technologists, Houston, Tex., Feb. 1973, pp. 300-319.
[8] Hicks, R. G., Williamson, R., and Santucci, L. E., "Effect of Laboratory Curing and
Compaction Methods on the Stress-Strain Behavior of Open Graded Emulsion Mixes,"
Transportation Research Record 712, Transportation Research Board, Washington, D.C.,
Jan. 1979, pp. 37-43.
[9] Hicks, R. G., Hatch, D. R., Williamson, R., and Steward, J., "Open Graded Emulsion
Mixes for Use as Road Surfaces," Transportation Research 702, Transportation Research
Board, 1979, pp. 64-72.
[10] Santucci, L. E., "Thickness Design Procedure for Asphalt and Emulsified Asphalt Mixes,"
Proceedings, Fourth International Conference on the Structural Design of Asphalt Pave-
ments, Vol. 1, 1977.
[//] Kallas, B. F. and Shook, J. F., "San Diego County Experimental Base Project," Final
Report, Research 77-1, The Asphalt Institute, Nov. 1977.
[12] Ahlbom, G., "ELSYM5, Computer Program for Determining Stresses and Deformation in
a Five Layer Elastic System," University of California, Berkeley, 1972.
[13] Yoder, E. J. and Witczak, M. W., Principles of Pavement Design, Second Edition, Wiley,
New York, 1975.
D. N. Little,1 J. W. Button,2 and J. A. Epps1
Potential for Asphalt Stabilized

Sand Bases in Texas
REFERENCE: Little, D. N., Button, J. W., and Epps, J. A., "Potential tor Asphalt
Stabilized Sand Bases in Texas," Extending Aggregate Resources, ASTM STP 774,
ABSTRACT! Depletion of high quality aggregate sources in concert with inflated

transportation costs has created a need for utilization of locally available aggregate
sources as base materials. Many state specifications, developed only for asphalt concrete
surface courses, preclude the use of low quality aggregates that may be suitable for bases
in properly designed pavement systems. Paving mixtures containing poorly graded sands
from Texas and asphalt cement or emulsified asphalt are evaluated using results from
two laboratory tests: (1) resistance to lateral flow (R-value) before and after moisture
treatment, and (2) diametral resilient modulus. Ranges of resilient moduli that have pro-
vided successful pavements in Texas are presented. Moduli determined in the laboratory
have been correlated to moduli collected in the field by in situ deflection testing.
KEY WORDS: aggregates, concrete, asphalt, sand, pavement, cement
The shortage of high quality aggregates together with increased traffic has
created a need for treating local materials for use as base courses. In the
southwest and midwest, sands transported by water or winds or both often
are the only substantial aggregate source available. Asphalt has become a
common base stabilizer for these marginal materials in the last few years.
However, the criteria developed for materials selection and design and con-
struction techniques have been based primarily on requirements developed
for asphalt concrete surface courses. Thus, because of these "strict" re-
quirements, materials evaluation and pavement design techniques are being
used that significantly increase cost and provide a stabilizer material whose
properties are in excess of those required by traffic and the environment.
'Assistant professor and professor, respectively, Texas A&M University, College Station,
Tex. 77843.
2
Assistant research engineer, Texas Transportation Institute, Texas A&M University, College
Station, Tex. 77843.
Copyright by ASTM Int'l (all139 rights reserved); Thu Dec 22

Indian Institute of Technology Kharagpur pursuant to License Agreeme
To provide an economical material to satisfy the particular requirements

of asphalt base courses, current materials selection criteria and pavement
design methods should be investigated and altered, if necessary. A first step
in this direction is to define a number of interacting mixture properties.
These include:
1. rheological characteristics,
2. fracture strength,
3. fatigue resistance, and
4. durability.
This paper specifically treats emulsion stabilized sand bases and hot sand
bases (and full-depth hot sand asphalt) in Texas. Materials selection criteria
and pavement design criteria to provide satisfactory performance in a given
environment will be suggested based on:
1. characterization of laboratory molded mixtures,
2. characterization of field core samples, and
3. in situ structural evaluation of the asphalt stabilized sands.
The purpose of this study is to evaluate the potential of Texas sands
stabilized with asphalt cement or asphalt emulsion to serve as a base course.
Specific objectives are to evaluate the asphalt stabilized sands with respect
to:
1. elastic or resilient deformation potential,
2. resistance to lateral flow,
3. Marshall stability, and
4. moisture susceptibility.
A sizeable research effort has been underway at Texas Transportation In-

stitute to evaluate asphalt stabilized marginal aggregates peculiar to Texas.
This research was jointly sponsored by the Texas State Department of
Highways and Public Transportation and the Federal Highway Administra-
tion.
Data have been collected from laboratory compacted sand asphalt mix-
tures, field cores of sand asphalts and in situ pavements. The following prop-
erties were measured in order to evaluate the potential of these materials as a
base course:
1. resistance value (following exposure to moisture),
2. resilient modulus,
3. air void content, and
4. Marshall stability and flow.
Each phase of the testing program will be discussed later.
LITTLE ET AL ON ASPHALT STABILIZED SAND BASES 141
Materials
The aggregates listed in Table 1 were used in the laboratory molded
specimens. Generally, these aggregates represent either silicious wind blown
sands or silicious river deposited sands. Most of these sands are rather poorly
graded.
Two criteria were used to evaluate the potential of these sands for
bituminous stabilization: the Chevron [1]^ U.S.A. criteria and the Herrin
[2] criteria.
The Chevron criteria judges the sand's acceptability for stabilization based
on gradation, plasticity, sand equivalence, and resistance of the untreated
material to lateral flow. Results of the evaluation with respect to these
criteria is summarized in Table 2.
The Herrin criteria identifies an aggregate as being suitable for stabiliza-
tion with asphalt as either a soil bitumen, sand bitumen, or sand-gravel
bitumen mixture. Expected performance within each category is based on
percent fines, plasticity index, and liquid limit. These results are also sum-
marized in Table 2.
A few sand asphalt cores were tested in the laboratory for data comparison
with results from laboratory compacted specimens. Aggregates used in these
pavements are listed in Table 3. Because these are field cores, virgin ag-
gregates were not available for laboratory testing. However, based on the
locations of the cored specimens and the general aggregate descriptions, it
was assumed that the aggregates comprising the cores are quite similar to the
aggregates used in the laboratory molded specimens.
Laboratory Molded Specimens

Asphalt emulsion stabilized mixtures were tested according to the se-
quence shown in Fig. 1 (top). The Chevron U. S. A. procedure [1] was
followed throughout the testing sequence. Detailed procedures are given in
Appendix A of Ref / . Each aggregate was mixed with a cationic slow setting
emulsion (CSS-1) conforming to criteria in ASTM Specification for Cationic
Emulsified Asphalt (D 2397-79).
Sand asphalt mixtures were tested according to the sequence shown in Fig.
1 (bottom). The asphalt used was an AC-10 supplied by the Exxon refinery in
Baytown, Tex. All asphalt cement conformed to criteria in ASTM Specifica-
tion for Viscosity-Graded Asphalt Cement for Use in Pavement Construction
(D 3381-76).
'The italic numbers in brackets refer to the list of references appended to this paper.
4:?
S -o •o t! •o S?
C —
3 S s M
,g JO
V9 >
'+-
a. S « 3 g S ^
X
»*-( as
X
•a X X
•c H 13 H t3 ^ _ M
,
•« X
1
cu
0
I—1
•o
o 00 m m
•g o o
•
J J
^
(N R
^
3
OJ
.S >
^
s bo
^
i^
s
1
v>
2
'3
U"
a
(U •a u
a
.2 •o •a
•o & & •a
c
ca
•o
ct
q
u
o 1 --s S,
a
s5 o 1/^ --^ r>J ' ^ rM -t-" ( ^ ' N
' " 00 ' _ -—, • ' f^ '^ W9 -—\ • ' 13 ' ^ lO <N "i^ ^ "• OO
?? at
1 a
S •a
5 ! j(5S(5S(5fe
<Q
+-
V
•« Sa ^
f
<
^ 11 (1< p 1/5
LITTLE ET AL ON ASPHALT STABILIZED SAND BASES 1 4 3
o • c8 'a
42 : « : §
a
•o •a •o
U.O
• s •-
1i l l 8
c c
u s 5
B _
8
u
C
 £1 T3 A
= •2 .-3 .•3 73 C
0
e 3E 3E -y3 3e
,^ ."ti
X> .^
Ui•
. 0 •a x>
: 3 :73 a rs
;^ g S
^
s as S S sss s
-a
1 111
Iflllllfllll
I
-as
M M M (A M t/l M C/} M (A (/J
U U U U U U U V 4> V V
a.
^1 >i>-.>.>,>.>>>>0 >>>.>,>,
s 'Sb—i 'S)
e
S S § t 1
0
If
1§ £ 2 1
i i
UH
N '
5BB tL,
C" p C^ H3 JO >-s 00
OfeQ OJfci« Sj
1
w tfl
II .2 rt S
(N (S (N —
' '
* j
-4-»
0 U U U u t3 tS tS ^ « ^ -5
•c •c -c -c •c •c -c -c -c x -c -c
to "K tfl ^ tS tfl tn
"•- Z
(h 3 00 •^ ix
^
^ « J .^ .S
Q Q Q Q O 5 5 S 5 b Q5
TABLE 3—Sand asphalt field cores.
Designation Aggregate
District 11 mixture of East Texas sands (poorly graded)

(SH 103)°
District 11 mixture of East Texas sands (poorly graded)
(Loop 287)
District 11 Gibson sand (poorly graded)
(Loop 287)
District 11 Bradley Pit sand (poorly graded)
(Loop 207)
District 20 poorly graded subgrade sand
(SH 87)
District 20 sand used at hot mix plant site
(SH 96)
District 20 subgrade sand (poorly graded) Layer B
(FM 255)*
District 20 subgrade sand (poorly graded) Layer C
(FM 255)
District 20 poorly graded 1(K)% sand
(FM 255)
"SH indicates state highway.

''FM indicates farm to market road.
Determine Re-
Mix and Compact F u l l y Cure Vacuum Determine
s i l i e n t Modu-
Specimens* a t Optimum Specimens Saturate R-Value
l u s 0 Selectee
Asphalt Content and
Temperatures
According t o Chevron Moisture
Procedure** Pick Up
3 specimens were molded at each of 3 asphalt contents: Optimum, Opt. + ^% and Opt - 1%.
The Chevron U.S.A. procedure was followed throughout this test sequence.
Determine Marshal" Determine

Mix and Compact Air
Specimens*at Optimum
A s p h a l t Content ^r** F u l l y Cure
Determine
Resilient
^ Flow Voids
According t o Test For 24 Hours Modulus fl

Method Tex-126-E ? 73 + 5°F Selected
Temperature
Vacuum
Saturate R- V3lue
Manual of testing procedures, Texas State Department of Highways and Public Transportation,
Volume 1 , 1974.
FIG. 1— Asphalt emulsion mixture test sequence (top) and asphalt cement mixture test se-
quence (bottom).
Field Cores
Field cores were obtained for testing on selected projects. Tests on these
cores were performed primarily to verify or to compare with tests performed
on laboratory molded specimens.
The following tests were performed on field cores:
1. Marshall stability and flow, ASTM Tests for Resistance to Deformation
and Cohesion of Bituminous Mixtures by Means of Hveem Apparatus (D
1560-76),
2. diametral resilient modulus,
3. resistance value (R-value) following vacuum saturation, and
4. air voids content.
The diametral resilient modulus, R-value, and vacuum saturation testing
followed procedures outlined in Appendix A of Ref 1.
Tests Results on Laboratory Molded Specimens
Asphalt Cement Stabilized Sands

Laboratory molded hot sand-asphalt mixtures exhibited comparatively low
stabilities. This is particularly true of the District 5 and District 25 sands.
Gradations of these sands are shown in Table 4. Laboratory test results
reflect the poor gradation of these two blow sands. The District 20 silty sand
is also a poor candidate for a high stability mix due to its poor gradation
(Table 4.)
The Chevron and Asphalt Institute emulsion mix criteria were used to
evaluate the sand-asphalt mixes. Although several states and agencies have
adopted sand-asphalt criteria, these criteria are generally derived from sur-
face hot mix specifications and thus require low air void contents (less than 6
percent). Based on the sand-asphalt air voids criteria, the sands evaluated in
this study would not be suitable.
Table 5 shows that only the District 21 sand meets the Chevron {1\ and
Asphalt Institute [3] criteria established for emulsion stabilized mixtures
that requires a minimum R-value of 78 after vacuum saturation and a max-
imum of 5 percent moisture pick-up. Moisture pick-up is the increase in
weight of the specimen as a percentage of the dry weight of the specimen [/].
The District 20 silty sand is marginal, and the poorly graded blow sands from
Districts 5 and 25 are clearly substandard based on the Chevron or Asphalt
Institute criteria.
Excessive moisture pick-up and low stabilities following vacuum satura-
tion is in part a result of high air void content of the poorly graded sands.
A criteria for mix design based on resilient modulus versus temperature
relationships does not exist. However, researchers at Texas A&M have been
collecting diametral resilient modulus versus temperature data for several
.^-^-OlOf^•vu^'-•oÔ'-'t--l^
SB
. ^ ^ _ fsi o •^ o r-- o t--

• o o o (N* r--' K rH a^' iW -qf
<N • ^ s D r ^ 0 0
Q ^
,'-Hrs'î/)oor--<T'(N
Z
•— ^
1 •n g
.B D
a"
• O O O O --H ^ 0^ H 95 •^' ^ '
&•
o
•t«
5 ^
o
•s ^ ^
•s •S2 i -r-i ^£3 t~-' PO r-' I
•—I t/) v^ t~- r - I
•S
i5£
•n JO
Mb
" "^ d i S s I ^
's>
,1—ir^-q-^^n-ÔôO'-H
1
•f Q!i
'-" • ^ lO \ 0 t-. ^
a -H O z
4W v^ ^ ^
1—^
QQ oooornrorooooNvors<Nr-.oco CU
^H •n iS
,t3 « z
(N so
,>«Trmr--i/)oo\OTriooOQO(N
: o - r4 r4 >H K <^ 0^ p ^ rg ^
•|B
; o -H r n 00 O r~; 00 (N 0.
IS S
55 « z
1 I
=. .3 .s" ^ 00 2 S S ? 55 S S § i ^ 'I1:g ^
.S lo GO Q* ^ Q ri d o' d d 6 o' d § *" .5
- H m m Z Z Z Z Z Z Z Z Z Z Z ! « tt.
:llili
Copyright by ASTM Int'l (all rights reserved); Thu Dec 22 07:4

Indian Institute of Technology Kharagpur pursuant to License A
TABLE 5—Summary of stabilities and moisture pick-up of laboratory

molded asphalt cement stabilized sands.
R-Value Moisture
Optimum" After Vacuum Air Voids, Pick-Up,
Aggregate % AC Used Saturation % %
District 5 6 too weak to test 19 11
(FM 168)
District 20 6 80 15 6
(U.S. 96)
District 21 6 80 3 4
(Beck Pit)
District 25 6 70 18 10
(FM 3182)
"AH tests were performed on mixtures containing 4, 5, and 6 percent AC-10. The op-
timum AC percentage is based on the highest R-value and lowest moisture pick-up.
years on various types of asphalt mixtures (both laboratory molded and field
cores). These data have been compared to field performance data.
Figure 2 illustrates the band of resilient moduli versus temperature
selected as an acceptable range for performance of asphalt treated bases and
full-depth asphalt sections in Texas. Resilient modulus versus temperature
for the asphalt cement stabilized fine sands is show^n by the solid lines. The
length of the arrows represents the magnitude of decrease in resilient
modulus at 20°C (68°F) after vacuum saturation.
Resilient moduli of Districts 20 and 21 asphalt stabilized sands are within
the acceptable band. Those of Districts 5 and 25 mixtures are below this
band.
Based on R-value before and after moisture conditioning and resilient
modulus testing, these previous data indicate that hot asphalt stabilized
sands from Districts 20 and 21 have the potential to perform as satisfactory
base materials. Poorly graded blow sands exhibited very low stabilities, high
air voids, excessive moisture pick-up, and low resilient properties.
Quality of paving mixtures seems to depend on gradation. Poorly graded
sands with low minus 200 sieve fraction should be used with caution.
Emulsion Stabilized Sands

All but one of the emulsion stabilized sands evaluated are poorly graded
(Table 1). The results of the poor gradation is a high air void content
resulting in a relatively high moisture pick-up.
Table 6 summarizes the stability and moisture pick-up data from the sand-
emulsion mixes. The R-values of the wind blown sand (District 5) and the
Padre Island beach sand (District 16) were unacceptable according to
Chevron criteria [/] (lower than 78). Excessive moisture pick-up is evident for
^0 25 40
Temperature, "C
Indicates percentage drop in resilient modulus followinq

vacuum saturation (at 20^C). This is depicted praphicall.y by
the arrow.
FIG. 2—Resilient modulus—temperature relationship for asphalt cement stabilized sands.
the mixtures using the District 5 blow sand, the District 25 blow sand, the
District 16 near beach sand, and the District 11 (Gibson) sand. All other
mixtures recorded moisture increases only slightly greater than the criterion
of 5 percent moisture pick-up by weight of the dry sample.
Generally, the poorly graded East Texas river sands from Districts 11 and
20 are acceptable in terms of R-value and marginal in terms of moisture
pick-up.
Figure 3 is a plot of resilient modulus data at 23°C (73°F). These resilient
moduli compare favorably with the band of resilient moduli values plotted on
Fig. 3. This band represents a range of resilient moduli at 23°C for mixtures
that have performed acceptably in the field as bases or full-depth asphalt
pavements [4-6\.
Based on stability, moisture susceptibility, and resilient modulus test
results, emulsion stabilized sands studied are potentially suitable for base
course layers.
TABLE 6—Summary of stability and moisture pick-up of laboratory

molded emulsion stabilized specimens.
Optimum % R-Value
Asphaltic After Vacuum Air Voids, Moisture
Aggregate Emulsion Saturation % Pick-Up, %
District 5 10 61 25 15
(FM 168)
District 11 11 78 24 16
(Gibson)
District 11 11 80 10 3
(FM 3736)
District 11 11 81 11 5
(Daniels)
District 16 12 68 27 15
(Padre Island)
District 20 12 89 12 7
(FM 255)
District 20 10 86 12 5
(SH 87)
District 25 8 78 19 11
(FM 3182)
Test Results on Field Coras

Field cores from Districts 11 and 20 were tested for resistance to lateral
flow (R-value) and moisture pick-up following vacuum saturation. These
data are summarized in Table 7. Except for those from District 20 (U. S.
Highway 96), all cores meet the minimum requirements of 78 for the R-value
following vacuum saturation. A description of the aggregates in each core is
in Table 3.
The fact that the air void contents for both laboratory specimens and field
cores are of the same order of rriagnitude indicates that the laboratory com-
pactive effort is comparable to that used in the field.
As in the laboratory molded cores, the percent moisture pick-up is above
the recommended maximum of 5 percent. This once again indicates that this
criterion may be too stringent for sands because of their inherent high air
void content. One suggestion for a moisture pick-up criterion for sand mix-
tures is to establish some percentage of the existing air voids in a compacted
mixture.
The resilient moduli of the field cores at 23°C is compared to the resilient
moduli of other base materials typically used in Texas in Fig. 4. All of the
cores exhibited a modulus in the range of other quality base materials. This
indicates their ability to function satisfactorily in a pavement system under
short duration dynamic loading.
10"
Band of acceptable resilient
moduli at 23°C based on Texas
A&M University experience -
laboratory mixed and compacted
FIG. 3—Resilient moduli at 23°C for asphalt emulsion stabilized sands (laboratory mixed
and compacted^.
Structural Evaluation
Laboratory derived resilent modulus versus temperature data were used
together with layered elastic computer modeling to evaluate the structural
potential of selected sand asphalt and asphalt emulsion stabilized sand
bases.
Accurate determination of the resilient modulus is important in evaluating
pavement structures by layered elastic modeling. In addition, temperature
susceptibility of asphalt bound materials makes it necessary to evaluate the
relationship between resilent modulus and temperature.
In order to establish a credible relationship between resilient modulus and
temperature, both laboratory testing and in situ testing were used. Diametral
resilient moduli were determined at 5, 20, 25, and 40°C to establish the
resilient modulus-temperature relationship. In addition, the Falling Weight
Deflectometer and the Dynaflect were used to evaluate the in situ dynamic
modulus of the pavement layer in question. Thus, laboratory and in situ
TABLE 7—Summary of stabilities and moisture pick-up of field cores.
R-Value Moisture
After Vacuum Air Voids, Pick Up,
Aggregate Asphalt Saturation % %
District 11 AC-10 85 16 6
(SH 103)
District 11 AC-10 81 22 11
(LP 287)
(LP 287)
(US 287)
District 20 asphalt 78 16 7
(SH 87) emulsion
(SH 96)
(FM 255) emulsion
(FM 255) emulsion
(FM 255) emulsion
readings were used concurrently to establish a resilient modulus versus

temperature relationship.
In Situ Modulus Determination

Discussion of the in situ resilient modulus determination is divided into
two parts: (1) Falling Weight Deflectometer (FWD) and (2) Dynaflect.
The FWD is a nondestructive pavement tester that simulates the effect of
a fast moving wheel load. It is capable of a load range of 13 to 50 KN, an
associated loading time of 26 ms, and vertical deflections that may be taken
at any desired position from the center of the loading plate outward along the
deflection basin. Several European research projects [7,8] have compared
the deflection, stresses, and strains of the FWD with those caused by moving
wheel loads. Correspondence of the two was remarkably good (within 10 per-
cent) [9].
Assuming that the elastic moduli of materials may be derived from deflec-
tion tests, it is believed that the FWD stands a better chance of representing
actual wheel loads than most other available steady-state loading systems.
Using the concept of establishing pavement layer moduli by deflection basin
matching, the modified linear elastic program, ELSYM 5, was used to match
the FWD basin by an iterative procedure [9].
The ELSYM 5 program was modified so that variable stiffnesses depend-
ing on the state of stress under each corresponding deflection sensor could be
Acceptable r e s i l i e n t moduli bands based on Texas A&M

experience a t 23"C ( f i e l d c o r e s ) :
n ^ ^ ^ Dense graded a s p h a l t concrete bases
lliilllli ^™'^ s t a b i l i z e d bases and subbase
^^^^^ Untreated g r a n u l a r bases
• ^ CC -^03 — 03
FIG. 4—Resilient moduli at 23°C for field cored asphalt stabilized sands.
used for calculating the total deflection. The individual solution is valid for
that particular deflection position only.
The ELSYM 5 program, using principals of the methods of equivalent
thicknesses as well as Bouissinesq's equation, has been streamlined into an
iterative procedure through which unique solutions can be obtained quickly.
The resulting program is called ISSEM 4 [9]. The obvious advantage of the
FWD, together with the analysis package, is that it allows one to vary load or
stress level and to evaluate the in situ elastic response to each layer in the
pavement system as a function of stress level.
The Dynaflect was also used to evaluate the in situ elastic moduli of several
sand asphalt cement and asphaltic emulsion stabilized materials. The
Dynaflect applied a sinusoidal load of 4.45 KN amplitude at a frequency of 8
Hz on two steel wheels 508 mm apart in contact with the pavement. This is
relatively low when compared to the FWD and is a problem in evaluating
layers whose elastic response is highly stress dependent. A second noteworthy
limitation of the Dynaflect is that the magnitude of the load is fixed and
therefore stress sensitivity cannot be suitably evaluated within layers of the

pavement structure.
However, like the FWD, the Dynaflect has the capacity to measure the
deflection basin developed by the 4.45 KN sinusoidal load. This is done by
accelerometers located at 254, 396, 660, 949, and 1245 mm from the two load
points. Accelerometers record acceleration and integrate it twice to deter-
mine surface deflection.
Previous research at Texas Transportation Institute (TTI) [10-12] in-
dicates several ways to use the Dynaflect deflection basin to evaluate in situ
elastic moduli. All of these methods are based on deflection basin matching
by means of layered elastic analysis. The method selected for this analysis has
been used at TTI in comparative evaluations of recycled pavements, sulfur
extended asphalt pavements, and conventional pavements sections. This
method is based on a modification of Vaswani's procedure [13] in which the
dual parameters of maximum Dynaflect deflection, </max. ^^^ spreadability,
S, (the ratio of the average deflection within the basin to if max) are evaluated
to graphically determine an effective thickness of pavement above the
subgrade.
Figure 5 is an example of dual parametric charts developed specifically for
the Dynaflect load configuration using the layered elastic computer program
BISTRO [14]. Once the loci of d„^^ and S are plotted for a given pavement,
the subgrade elastic modulus and effective thickness for a selected composite
modulus of the structural pavement (above the semi-infinite subgrade) can
65 ^
. 245 MPa
Locus of d
max
and S for FM 1632
.06 .04 .02

Maximum Deflection, d Maximum Deflection d , mm
FIG. 5—(a) Dual parametric Dynaflect chart for a composite pavement elastic modulus of
689 MPa. (b) Dual parametric Dynaflect chart for a composite pavement elastic modulus of
1379 MPa. Note that D equals composite pavement (thickness). (Note: Interpretation between
(a) and (b) yields a composite pavement elastic modulus of 965 MPa for FM 1632.)
be determined. If several charts are developed reflecting different composite

pavement moduli, the composite (or weighted average) modulus of the pave-
ment in question may be evaluated by knowning the pavement's cross-
sectional thickness. Such a procedure was first developed by Vaswani [13].
Since the dual parametric procedure gives only a composite or weighted
average modulus of all layers above the subgrade, it may be ineffective in
selecting a specific layer in a multilayered structure. Fortunately, in this
analysis the layer in question was either the only layer of structural
significance or one of only two structurally significant layers. Thus, the in
situ elastic response could be effectively evaluated for the sand-asphalt ce-
ment and asphalt emulsion stabilized sands.
To briefly illustrate the preceding procedure, deflection data from FM
1632, a full-depth (203 mm) emulsion stabilized sand with a surface treat-
ment, will be considered. Plotting the loci of d^^^ and S in Figs 5a and b, the
elastic modulus of the emulsion stabilized sand can be interpolated as 965
MPa at an in situ pavement temperature of 27°C.
Matching Laboratory and In Situ Data

The diametral resilient modulus device developed by Schmidt [15] was
used to evaluate the resilient modulus versus temperature relationship for
laboratory tested field cores. The in situ measured modulus at the pavement
temperature at the time of evaluation was plotted on the same chart, and the
curve was shifted to pass through the in situ measurement. Pavement
temperature at the time of in situ evaluation was computed by The Asphalt
Institute procedure [16]. Figure 6 illustrates this procedure for U. S.
Highway 96. The in situ modulus closely approximated the laboratory curve
in each case.
The procedure just discussed was used for each pavement evaluated in situ
to develop average annual elastic moduli for the respective climatic condi-
tions of each pavement. These moduli were used, as will be explained in the
following section, to develop layer structural coefficients for hot sands and
emulsion stabilized sand bases evaluated.
Stmctural Coefficients
The AASHTO Interim Guide for Flexible Pavement Design is based on the
experience of a factorial road test experiment. The concept of structural layer
coefficients, a,, is familiar to most pavement engineers as a relative measure
of the performance of a given material in a given position within the pave-
ment structure.
Structural coefficients are actually regression coefficients that describe the
contribution of the material and layer in question to the total pavement
structure. As one might expect, these coefficients are highly sensitive to the
Laboratory r e s i l i e n t modulus t u r v e
S h i f t e d r e s i l i e n t modulus curve t o pass
throvKih i n - s i t u determined r.radulus
(U.S. Hiqhwa.y 96)
I n - s i t u determined modulus
Laboratory determined modulus
Pavement temper-
a t u r e at time of
f i e l d testinq
was 33"C
20 25 40
Temperature, "C
FIG. 6—Procedure for shifting the laboratory resilient modulus versus temperature curve to
reflect in situ data.
interactions within the pavement structure. Indeed these coefficients are not
unique material properties but are a function of temperature, pavement
structural geometry, interdependency of structural layers, load intensity, etc.
Clearly, any design parameter that is so sensitive to so many variables is
quite limited as a design parameter. Despite the limitations, the structural
coefficient can be effectively used as a comparative performance index. It is
presented as such in this paper.
Previous research [11,17,18] has shown that the structural coefficient for
base course material is highly correlated to the temperature versus stiffness
relationships of the base material and the design layer thickness of the base.
Furthermore, previous research has also shown that the single most signifi-
cant parameter associated with the fundamental AASHTO flexible pavement
performance equation is subgrade deformation, W^. This was verified by ex-
tensive regression analyses using the original AASHTO data and by
evaluating other mechanistic parameters such as tensile strain at the bottom
of the asphalt bound layers and vertical compressive strains within layers and
at the top of the subgrade [11,17,18].
It is therefore possible to compute an AASHTO structural coefficient for a

given material by:
1. evaluating the stiffness versus temperature relationship of the material,
2. evaluating the performance parameter of subgrade deformation by
layered elastic analyses, and
3. selecting a structural coefficient based on the relationship between per-
formance life and subgrade deformation.
Figure 7 presents a chart by which the structural layer coefficient of a base
course can be determined based on temperature versus elastic modulus,
resilient modulus, or dynamic modulus data. Note that the structural coeffi-
cient, 02, is dependent not only on the modulus (in this case resilient
modulus, MR) but also the base course thickness, /i2.
To develop Fig. 7, the stress sensitive layered elastic computer program
PSAD2A [19] was used to model 27 pavement sections from Loop 4 of the
AASHTO Road Test. Materials comprising these sections had been
previously characterized. Asphalt concrete layers were characterizd in terms
of dynamic modulus versus temperature while the untreated base, subbase,
and subgrade materials were characterized in terms of resilient modulus as a
function of stress. A W^ computed from PSAD2A for each section was then
regressed against the number of design loads (18 kip (40 KN) single axle) to a
selected terminal serviceability {P, = 2.5) for each of the 27 sections.
Once a relationship between pavement life, in terms of design load ap-
plications to a selected serviceability level, and W^ was established, struc-
tural coefficients of bases of different resilient moduli from those used in
Loop 4 of the road test could be determined. This was done by substituting
the sand asphalt moduli versus temperature relationships for those of the un-
bound bases used at the road test and computing W^. A performance life,
W„ in terms of number of design load applications was determined from the
regression equation. Finally, with the VK, known, aj could be calculated from
the fundamental flexible AASHTO design equation.
logW, = log p + G,/^
where
W, — number of design load applications,
p — a function of design load and the structural coefficients,
§ — & function of design load and the structural coefficients, and
G, = a function of the selected terminal serviceability.
This evaluation is based on the potential of the subgrade to distribute
stresses in the elastic range. The asphalt stabilized sands appear to be well
suited to adequately distribute these stresses and thus to protect the
subgrade. However, their resistance to all important permanent deformation
and thermal cracking is not a part of this analysis.
FIG. 7—Relationship between structural coefficient, 82, and average annual resilient
modulus.
Current research at Texas A&M University [20] is geared to investigate the

plastic deformation and thermal cracking potential of the asphalt stabilized
sands. Thus, although the present analysis does not treat permanent defor-
mation and thermal cracking, it is encouraging that these sands possess
suitable resistance to lateral flow after moisture conditioning and a suitable
range of resilient moduli.
Results
Pavements and asphalt stabilized sand bases evaluated by laboratory
resilient modulus testing and in situ deflection testing are given in Table 8.
The FM 842 and FM 2680 contain hot sand bases that were evaluated in
situ by the FWD. The SH 6 and U. S. Highway 84 contain emulsion stabi-
lized limestone bases and were similarly evaluated. The SH 6 and U. S.
Highway 84 are included for comparative purposes as they represent asphalt
stabilized bases of accepted good quality in Texas.
Structural coefficients and elastic moduli from which these coefficients
were evaluated are listed in Table 9. The elastic moduli represent the
weighted annual average value for each respective location based on the pro-
cedure described in the preceeding section.
The sands stabilized with asphalt cement as well as those stabilized with
TABLE 8—Pavement bases tested by in situ deflection and laboratory resilient modulus.
Pavement Type Base Pavement Description
FM842'' AC-10 stabilized 152-178 mm of hot sand base

Lufkin, Tex. sand over 152 mm lime stabilized
subbase
FM2680 AC-10 stabilized seal coat over 165 mm of hot
Lufkin, Tex. sand sand base and 305 mm of select
materials
SH 110* AC-20 stabilized 38 mm HMAC over 203 mm asphalt
Smith Co., Tex. sand stabilized sand base
U.S. %'^ emulsified asphalt seal coat over 152-254 mm
Jasper Co., Tex. stabilized sand emulsion stabilized sand and
(plant mix) 152 mm of lime stabilized
subbase
FM255 emulsified asphalt seal coat over 203 mm of
Jasper Co., Tex. stabilized sand emulsion stabilized sand base
(plant mix) and 305 mm of select material
FM 1632 emulsified asphalt seal coat over 203 mm of
Tyler Co., Tex. (road mix) emulsion stabilized sand base
SH6 emulsified asphalt 38 mm HMAC over 102 mm
Waco, Tex. stabilzed limestone emulsion stabilized limestone
base and 305 mm of select material
U.S. 84 emulsified asphalt 38 mm HMAC over 51 mm emulsion
Waco, Tex. stabilized limestone base and 203 mm of gravel base
"FM indicates farm to market road.

*SH indicates state highway.
"Û.S. indicates U.S. highway.
TABLE 9—Average annual elastic moduli and structural coefficients.
Pavement Stabilizer Aggregate ^•avg. M P a <^2
FM842* AC-10 poorly graded sand 827 0.26

FM2680 AC-10 poorly graded sand 1655 0.33
SH 110-^ AC-20 well graded sand 1103 0.29
U.S. W emulsion silty gravel 1517 0.32
FM255 emulsion silty sand 1241 0.26
FM 1632 emulsion poorly graded sand 965 0.28
SH6 emulsion limestone 1724 0.34
U.S. 84 emulsion limestone 1241 0.30
"Based on 203 to 254 mm base thickness.

*FM indicates farm to market road.
"^SH indicates state highway.
''U.S. indicates U.S. highway.
asphalt emulsions possess structural coefficients, a-i, in the range of 0.26 to

0.33. This compares favorably with structural coefficients established at the
AASHTO Road Test of 0.30 for asphalt stabilized sands. In addition, the
coefficients of the emulsion stabilized limestone for SH 6 and U . S . Highway
84 serve as a comparative standard by which to evaluate the asphalt stabi-
lized sands as well as the analytical procedure.
If one compares the structural coefficients in Table 9 to those developed at
the AASHTO Road Test for structural base layers, the Texas asphalt
stabilized sands appear to be viable alternatives based on the criteria of
subgrade protection. For example, the structural coefficients in Table 9 are
generally significantly superior to those for the lime, cement, and untreated
bases studied at the AASHTO Road Test.
Conclusions and Recommendations

Selected Texas blow sands and river sands are suitable aggregate sources
for asphalt stabilized bases. Many of these sands produce mixtures with high
air void contents and high water absorption; however, suitable resilient
moduli, resistance to flow (stability), and resistance to loss of strength in the
presence of water are obtained.
The aggregate selection criteria established by Chevron U. S. A. [7] and
Herrin [2] are generally suitable. However, these criteria certainly do not
assure a suitable mixture, and thus mixture testing is essential.
The Chevron [/] and Asphalt Institute [J] mixture design methods for
bases stabilized with asphalt emulsions appear to have suitable criteria in
terms of resistance value following vacuum saturation. However, their
criteria for moisture pick-up in asphalt stabilized sands may be undully
restrictive as many of the mixtures evaluated in the study had adequate
resilient moduli, resistance to flow, and water-susceptibility behavior but ex-
cessive moisture pick-up as determined by the established criteria.
The authors suggest that the diametral resilient modulus test be used to
evaluate the structural contribution of the asphalt stabilized sand base to the
pavement system. The structural contribution of asphalt stabilized sands can
be evaluated by layered elastic analysis. Resilient modulus versus
temperature relationships should be developed in the laboratory and used in
combination with the Santucci procedure \22\ to design the pavement struc-
ture.
The in situ derived elastic modulus of asphalt stabilized sands are com-
parable to other stabilized bases. This indicates the ability of these bases to
effectively protect the subgrade from high constant stresses applied at the
surface. The relative contribution of the sand asphalt bases to the structural
function of the pavement system is illustrated by the structural layer coeffi-
cients. The coefficients are competitive with bases stabilized with cement,
lime, and untreated, high-quality aggregates such as crushed stone.
Indian Institute of Technology Kharagpur pursuant to License Agreement. No further reproductions autho
However, plastic flow and thermal cracking potentials must be investigated.

Current research at Texas A&M University includes detailed creep and
thermal cracking analyses.
References
[1] Bitumuls Mix Manual. Chevron U. S, A., Asphalt Division, San Francisco, Calif., Jan.
1977.
[2] Herrin, M., Darter, M. I., and Ishai, Ilan, "Determine Feasible Testing Methods for
Asphalt-Aggregate Cold Mix Bases," Project IHR-505, Illinois Cooperative Highway
Research Program, University of Illinois, Urbana, III., March 1974.
[3] A Basic Asphalt Emulsion Manual. The Asphalt Institute, FHWA-IP-79-1, Jan. 1979.
[4] Epps, J. A., Little, D. N., Holmgreen, R. J., and Terrel, R. L., "Guidelines for Recycling
Pavement Materials," NCHRP Report 224, Transportation Research Board, Sept. 1980.
[5] Epps, J. A., Little, D. N., O'Neal, R. J., and Callaway, B. M., "Mixture Properties of
Recycled Central Plant Materials," Recycling of Bituminous Pavements ASTM STP 662,
Dec. 1977.
[6] Unpublished data from FHWA Research Contract No. DTFH6180C0048, "Design and
Characterization of Paving Mixtures Based on Plasticized Sulphur Binders," Texas
Transportation Institute, 1980.
[7] Bohn, A., Ullidtz, P.. Stubstad, R., and Sorenson, A., "Danish Experiments With The
French Falling Weight Deflectometer," Third International Conference on The Structural
Design of Asphalt Pavements. Vol. 1, London, 1972.
[8] Classen, A. and Ditmarsch, R.. "Pavement Evaluation and Overlay Design—The Shell
Method," Fourth International Conference on The Structural Design of Asphalt
Pavements, University of Michigan, 1977.
[9] Sharma, J. and Stubstad, R., "Pavement Evaluation in Florida Using The Falling Weight
Def lectometer." paper presented at The Annual Meeting of The Transportation Research
Board, Washington, D. C , 1980.
[10] Swift, F., "A Graphical Technique for Determining the Elastic Moduli of a Two-Layered
Structure from Measured Surface Deflections," Research Report 136-3, Texas Transporta-
tion Institute, 1972.
[//] Lytton, R. L. and Michalah, C. H., "Flexible Pavement Deflection Equation Using Elastic
Moduli and Field Measurements," Research Report 207-7F, Texas Transportation In-
stitute, 1979.
[12] Little, D. N., "Structural Evaluation of Recycled Pavement Materials", Ph.D. thesis,
Texas A&M University, 1979.
[13] Vaswani, N. K., "Method for Separately Evaluating Structural Performance of Subgrades
and Overlaying Flexible Pavements," Highway Research Record 362, 1971.
114] Pertz, M. G. F., Jones, A., and Van Kempen, H. P. M., BISTRO. Shell Research Publica-
tion, 1972.
[15] Schmidt, R. J., "A Practical Method for Measuring the Resilient Modulus of Asphalt-
Treated Mikes," Highway Research Record 404, 1972.
[16] "Asphalt Overlays and Pavement Rehabilitation," The Asphalt Institute, MS-17, 1977.
[17] Darter, M. I. and Devos, A. J., "Structural Analysis of Asphaltic Cold Mixtures Used in
Pavement Bases," Research Report 505-4, Transportation Research Laboratory, Univer-
sity of Illinois, 1977.
[18] Jung, F. W. and Phang, W. A., "Elastic Layer Analysis Related to Performance in Flexible
Pavement Design," Ministry of Transportation and Communications, Ontario, Canada,
1974.
[19] "PSAD2A—Stress Sensitive Layered Elastic Computer Program," University of California
at Berkeley, 1978.
[20] Unpublished data from Research Project 2235, "Economic Black Bases," Texas Transpor-
tation Institute, 1980.
[22] Santucci, L. E., "Thickness Design Procedure for Asphalt and Emulsified Asphalt
Mixes," Vol. 1, Fourth International Conference Structural Design of Asphalt Pavements,
University of Michigan, 1977.
B. H. Polhemus1
Meeting Construction Aggregate

Fracture Requirements with Natural
Product-Sized Material
REFERENCE: Polhemus, B. H., "Meeting Construction Aggregate Fracture Require-

ments with Natural Product-Sized Material," Extending Aggregate Resources, ASTM
ABSTRACT: Areas of the country which are not fortunate enough to have ample accessi-
ble supplies of good quarry stone rely on diminishing gravel deposits for the bulk of their
construction aggregates. This paper intends to show how producers in those areas which
have strict minimum fracture requirements for asphalt material can utilize a larger por-
tion of the natural supplies.
Wear course asphalt specifications often require 80 to 100 percent of the particles in
the asphalt mix which are retained on the #4 mesh to have at least one fractured face.
Field experience has shown consistently that centrifugal impact crushing of natural ag-
gregates that are already within this size range achieves fracture counts of 80 to 95 per-
cent, often in a single pass. The operator can then utilize this material and reduce or
eliminate the need to separate, handle, and stockpile a waste material.
A centrifugal impactor can accomplish this because it does not depend on a "closed
side setting" to achieve crushing as do compression units, so virtually every particle enter-
ing the machine has an equal chance of fracturing, even if it is already product size.
On the other hand, the vertical shaft impactor has the ability tp minimize the loss to
fines (minus #10 mesh) often a problem with conventional horizontal shaft impactors.
This is a direct result of the crusher's geometrical design and its control of the trajectory
of the rock. The design minimizes scuffing on the rock and incorporates no grinding
path, consequently allowing the option of reducing production of fines.
The requirement for fractured faces on asphalt aggregate is an important parameter to
assure proper interlocking of particles and stability of the resulting asphalt mix. But with
the proper equipment the operator can meet this requirement, waste less material, and
conserve more of the often scarce high-quality aggregates.
KEY WORDS: aggregates, concrete, cement, fracture (materials), gravel
Many areas of the country do not have ample accessible supplies of good
quarry stone for use in base course and bituminous concrete manufacture.
'Application engineer, Spokane Crusher Mfg. Co., Spokane, Wash. 99220.
Copyright by ASTM Int'l (all161 rights reserved); Thu Dec 22 0

Producers of construction aggregates in these areas must rely on natural

gravel deposits of both river gravel and glacial original. As these sources of
natural gravel are consumed, well graded sources are depleted first, requir-
ing producers to turn to pits with an overall higher percentage of small rock
and sand.
Specifications for aggregate used in asphaltic concrete often require some
percentage of fractured or crushed material. In these cases, natural material
at product size (including pea gravel) cannot be used in large quantity since it
is smaller than the closed side setting of the conventional crushing equipment
employed. The inherent design of conventional crushing equipment requires
discarding natural product size material to achieve the required fracture per-
cent for asphalt aggregate. The net effect is a supply of aggregate that is con-
sidered marginal for asphaltic concrete because of the economics of handling
a large tonnage of material to recover and process a very little amount of coarse
sizes. The percentage of waste is high and the available yield of crushed product
is low.
One solution is the use of a vertical shaft centrifugal impact crusher to put
fractured faces on gravel that is already near the size required for the aggre-
gate product, without excessive reduction. Thus turning a "marginal" re-
source into a profitable aggregate resource.
Background
The advantages in terms of asphaltic concrete stability of using crushed
aggregate as opposed to a rounded natural aggregate have been shown by
many researchers. Field^ indicated in his report:
In a bituminous mix the physical attributes obtained by using "crushed aggre-
gates" instead of "noncrushed" are:
(a) Much more stability.
(b) A slight increase in void content (in bituminous mix).
(c) A slight increase in VMA (voids in mineral aggregate) which improves the
durability of the pavement.
The reasons why "crushed" material provides the physical attributes mentioned
in conclusion number two are:
(a) Edge sharpness improves internal friction.
(fc) Coarse textured faces allow for good asphalt tenacity thus improving co-
hesion.
(c) Coarse textured faces of adjoining particles (regardless of particle size dif-
ference) create more internal friction.
As a general conclusion from this test. Field states "The stability is defi-
nitely affected by the amount of crushed material in the coarse aggregate.
Considerable increases in stability are obtained by increasing the percent
^Field, Frederick, "The Importance of Percent Crushed in Coarse Aggregates as Applied to
Bituminous Pavements," Proceedings, Association of Asphalt Paving Technologists, 1958.
POLHEMUS ON NATURAL AGGREGATES 163
crushed in the coarse aggregate. The 60 percent crushed (minimum) specifi-

cation commonly stipulated should be rigidly enforced. In fact, consideration
might be given to raise this minimum percent crushed specification. (The
stability of 100 percent crushed was . . . 33 percent higher than that at 60
percent crushed.)"
Kalcheff^ has reported the advantages of using a crushed or angular fine
aggregate (minus #4 sieve) in asphalt mixtures compared to a fine aggregate
with rounded particles (see Table 1). The marshall stability values reported
in this table are a common method of measuring strength of asphaltic con-
crete. A higher value indicates a more stable mixture if it is properly placed
and compacted.
It is the added particle interlock due to fractured faces on the aggregate
that tends to increase the structural properties of asphaltic and concrete,
consequently, the marshall stability ratings.
The traditional methods of crushing pit run gravels down to product size—
jaw crushers, cone crushers, and roll crushers—effectively demands that ex-
cess material of approximately product size be screened out and wasted. The
closed side setting in cones and rolls does not allow materials smaller than
that setting to be crushed unless the crushers are fed at optimum rate and
very uniformly. Feeding to these requirements can often be difficult at best,
but the alternatives are an ever-increasing recirculating load or failure to
meet specifications due to insufficient fractured material, or both.
Solution
One solution is in the fundamental design of the vertical shaft impact
crusher. Laboratory testing and field experience have shown that centrifugal
TABLE 1—Summary of asphalt mix stabilities comparing crushed to uncrushed fine aggregates
with crushed coarse aggregates (see Footnote 3).
Mix Identification by Maximum Size
12.5 mm 9.5 mm 4.76 mm

Fine Aggregate, 4.76 mm (Via in.) (V2 in.) ('/8 in.) (3/16 in.)
Type Shape Index Marshall Stabilities at Optimum Asphalt, lb
Uncrushed silica 47 1660 1415 1035

Crushed limestone 49 3050 2520 2200
NOTE—All data from specimens at optimum asphalt content.
^Kalcheff, I. V., Laboratory Activity Report—April to June 1979, prepared for distribution to
the National Crushed Stone Association.
impact crushing of natural aggregates that are already within the desired
product size range achieves fracture counts of 75 to 95 percent, usually in a
single pass. The operator employing a centrifugal impact crusher at some
point in the circuit can then utilize this material and reduce or eliminate the
need to separate and stockpile a waste material.
An analysis of the design concept of a vertical shaft impactor shows why
this can be achieved. The centrifugal impactor does not rely on a "closed side
setting" to achieve crushing as do compression units. It utilizes "free body
crushing." The rock is fed centrally to the impeller and accelerated by the
impeller as it is semi-tangentially thrown against stationary impact blocks.
Every particle entering the impeller assembly is thrown against the impact
block regardless of size. So, every particle has an equal chance to fracture.
On the other hand, the vertical shaft impactor has the ability to minimize
the loss to fines (here considered minus #4 mesh), often a problem with con-
ventional horizontal shaft impactors. This is a direct result of the crusher's
geometric design and its control of the trajectory of the rock.
Feeding at the center of the impeller assembly and accelerating the ma-
terial reduces the scuffing notorious to horizontal shaft units where the feed
first contacts the periphery of the rotor, the point of highest velocity. Then
the placement of the stationary impact blocks assures a right angle of impact
with minimal scuffing, as well as sufficient space to eliminate any grinding
path.
This total design affords the operator the option of minimizing production
of fines (and minimizing wear) while retaining the advantage of impact crush-
ing for maximum fractured faces. The operator must acknowledge that the
probability of any given particle fracturing (and the extent to which it will re-
duce in size) is dependent on two factors:
1. the friability of the feed material, and
2. the "crack speed" or velocity at which the cohesive strength of the rock
is overcome by the rapid deceleration on impact.
Crack speed is partly affected by particle size. Reitter'' states that, "the co-
hesive strength of rock determines the power {E„) we need to crush it. This
cohesion is proportional to the cross-section of the r o c k . . . . " Experimenta-
tion will establish the optimum speed at which the operator produces a maxi-
mum percent of fractured particles and minimum percent of product passing
the #4 mesh.
Data shown in Fig. 1 are for a test performed in a laboratory size vertical
shaft impactor. The feed material, a tough river gravel, was screened at 19.0
mm (% in.) and at #4 mesh on a vibrating screen to provide a consistent 19.0
mm (% in.) by #4 mesh feed for all seven test speeds. The material was crushed
using a single pass through the laboratory size impactor and then screened
''Reitter, Guenther, "An Analysis of Impact Crushing," presented at the Annual Meeting,
National Limestone Institute, Manufacturers Presentations, Jan. 1980.
POLHEMUS ON NATURAL AGGREGATES 165
1 ^ ^
o
z
- i^
g-
0%
PERCENT CF MAXIMUM SPEED, %
FIG. 1—Lab results from crushing 19.0 mm (A in.) by #4 mesh gravel at various impeller
speeds in a vertical shaft impactor, showing effect of impeller speed variation on both fracture
percent and percent passing #4 mesh.
over a #4 sieve to determine the fines in percent by weight. This value is re-
ported on the graph in Fig. 1.
The crushed material retained on the #4 sieve was split for a fractured face
count. Particles having more than 25 percent of surface area as a fresh frac-
ture were considered fractured. The percent by weight of fractured particles
from the material retained on the #4 sieve are also reported on the graph in
Fig. 1.
Five initial tests were performed and the impeller speed was recorded as
percent of maximum speed. Initially speeds of 100, 83.3, 66.7, 50, and 33.3
percent of maximum were tested. A severe drop of fracture count was noticed
between 50 and 33.3 percent of maximum speed. So speeds of 44.4 and 38.9
percent of maximum were tested to further define the curve in this area.
The comparison of the curves on the graph in Fig. 1 indicates that approxi-
mately 50 percent of maximum speed yielded the highest ratio of attained
fracture compared to loss to fines for this particular gravel. At this speed, the
fractured count was 89.9 percent and the percent passing the #4 sieve was
28.6 percent.
One field trial took place in the state of Louisiana under the following con-
ditions:
1. A local supply of sand and gravel that rarely exceeds a 50.8 mm (2 in.)
maximum size.
TABLE 2—Vertical shaft impact crusher output (field results, wet feed).
Feed to Crusher, Crusher Output at 65% of

Sieve Size % Maximum Speed, Percent Passing
25.0 mm 1 in.
19.0 mm 3/4 in. 100 100
12.5 mm '/2 in. 64 87
9.5 mm ^/s in. 28 72
6.3 mm V4 in. 10 51
4.75 mm #4 mesh 6 35
2.0 mm #10 mesh 2 13
425 ^m #40 mesh ... 4
150 Mm #100 mesh ... 2
75 ^m #200 mesh ... 1
NOTE—Single pass fracture count average = 84 percent of particles retained on #4 mesh.
2. State and other asphaltic concrete specifications that require 65 percent

fractured faces for base mixtures and 75 to 80 percent for wearing course
asphaltic concrete.
3. An extremely tough highly abrasive silica gravel.
4. The desire to produce a minimum amount of fines through the crushing
action while achieving the highest possible yield of fractured coarse aggre-
gate.
By top sizing the natural material through screening at virtually the top
product size and feeding it to a vertical shaft impactor, the operator was able
to determine the optimum speed at which he achieved 80 to 85 percent frac-
tured faces using a single pass. The percentage of minus #4 sieve size ma-
terial was held to an acceptable level (see Table 2). Hence, he was able to uti-
lize the large amount of his aggregate supply that fell between 19.0 mm (%
in.) and #4 size.
Summary
Requirement for fractured faces on aggregate for asphaltic concrete is a
common specification strategy to assure proper interlocking of particles and
stability of the resulting asphaltic mixture. However, the producer who has a
scarcity of larger gravel sizes does not want to reject product size natural ma-
terial to meet this requirement. With the proper equipment, the operator can
meet the fracture requirement for mineral aggregate, waste less material,
and turn a resource that was previously considered marginal into a resource
that can be better utilized. The net result is a more efficient use of available
aggregates and a lower overall cost per ton due to reduced crushing costs and
increased product yield.
S. D. Ramaswamy,1 M. A. Aziz,l and C. K. Murthyl
Sea Dredged Sand for Concrete
REFERENCE: Ramaswamy, S. D., Aziz, M. A., and Murthy, C. K., "Sea Dredged
Sand for Concrete," Extending Aggregate Resources, ASTM STP 774, American Society
for Testing and Materials, 1982, pp. 167-177.
ABSTRACT: A study was conducted by the authors to assess the possibility of using sea
dredged sand from Singapore in concrete construction. Various combinations compris-
ing normal sand, washed and unwashed sea dredged sand, potable water and seawater
for mixing were considered. The properties of concrete made with washed and unwashed
sea dredged sand by using potable and seawater for mixing were evaluated and compared
with those determined for concrete made by using normal sand and potable water. The
study has shown very encouraging results. The authors report the details of this study and
offer arguments based on their research in favor of using sea dredged sand as aggregate
in making concrete.
KEY WORDS: sea dredged sand, seawater, concrete, strength, durability, efflorescence,
corrosion, cement, aggregates
Little published information is available on the properties and use of con-

crete made with sea dredged sand. Sea dredged sand differs from river sand
or washed land dug sand mainly because of the presence of sea salts such as
chlorides and sulfates. Because of the contaminated nature of the sea
dredged sand, it is sometimes presumed that it is unsuitable for any type of
concrete construction. Available offshore deposits of sand are sometimes
used for reclamation purposes as a fill material and mostly ignored as of no
value.
At present, opinion differs as to the suitability of sea dredged sand for use
in concrete [l]2 although it is known that such aggregates have been used for
building concrete structures for several decades with few reported difficulties
[2]. Marine aggregate production in the United Kingdom, which was 7.2
million tonnes in 1966, has increased to 16.5 million tonnes in 1979 out of
1
Associate professor, Department of Civil Engineering; associate professor, Department of
Civil Engineering; and associate professor, Department of Building Science; respectively, Na-
tional University of Singapore, Singapore.
2
Copyright by ASTM Int'l (all 167 rights reserved); Thu Dec 22 07:41:33
Copyright® 1982 by ASTM International www.astm.org
which roughly a quarter is exported [3]. This increase is due to the con-
fidence gained in the use of aggregates from deep seawater that were found to
be suitable for making concrete because of lower salt content. If it can be
established that sea dredged sand is not so bad after all and that it can at
least be used for certain types of concrete construction without encountering
short-term or long-term harmful effects, it would be a break-through
especially for fast developing countries like Singapore where the demand for
sand surpasses the supply. With the unprecedented amount of construction
activity taking place in Singapore and with the forecast of a construction
boom in the coming decade, discovering a new source of sand would be very
timely.
The sand needs of the construction industry in Singapore are often fulfilled
by importing sand from neighboring countries. The cost of sand has more
than doubled within a period of two years, the cost at present being $16/m-'.
Judging from the future increase in demand and the expected inflation, fur-
ther increase in unit cost of sand appears imminent. It is interesting to note
that the present cost of dredging sand from the sea-bed deposits lying some 4
to 5 km distance from the shoreline of Singapore is not likely to be more than
SS/m-'. The economic benefits of using sea dredged sand are thus obvious.
Some 40 million m^ of sand was dredged from the sea bed and pumped to
form 640 hectares of reclaimed land to make way for the new Changi Interna-
tional Airport presently under construction in Singapore. Limited investiga-
tions carried out in this connection up to about 5 km offshore from the
coastline at Changi revealed good sand deposits occurring under the sea bed.
Further investigations therefore might reveal more of such deposits.
The effects of salts present in the sea dredged sand on various properties of
concrete have been examined in this paper. For construction onshore and
offshore, it may be economical and sometimes more practical to use seawater
for mixing provided the specifications permit. From this point of view, the
extreme effects of using sea dredged sand and seawater for mixing were also
considered in this study. Steel rods embedded in cylindrical specimens made
with various mixes were subjected to extreme conditions and studied for
long-term corrosion of reinforcing steel. The durability of concrete made
with different mixes was also assessed.
Materials
Freshly placed sea dredged sand from the Changi reclamation site and the
washed sea dredged sand (washed in freshwater) were subjected to chemical
analysis to determine the percentage of main chemical constituents. The
results of chemical analysis are shown in Table 1. It is seen from Table 1 that
the salt content of the sea dredged sand is negligible and after washing, the
salt content is brought to an insignificant level. Although opinion differs as
to the suitability of sea dredged sand for use in concrete, Table 1 confirms
RAMASWAMY ET AL ON SEA DREDGED SAND 169
TABLE 1—Main constituents of washed and unwashed sea dredged

sand as a percentage by weight of cement in 1:2:4 mix.
Percentage by Weight of Cement in 1:2:4 Mix
Sea Dredged Sand

Unwashed Sea Dredged Washed in Freshwater,
Percentage Sand, 100 g of sand per 100 g of sand per litre
Constituents in Seawater litre of distilled water of distilled water
Chloride 1.648 0.2857 0.059

Sulfate 0.270 0.0055 0.002
Sodium 0.987 0.1257 0.025
Magnesium 0.103 0.0040 0.000
Calcium 0.037 0.0240 0.000
Potassium 0.031 0.0226 0.000
the view generally held that sand from the sea bed does not contain harmful
quantities of salts [1,4].
British Standard (BS) CPllO, 1972 stipulates that for concrete containing
embedded metal, anhydrous calcium chloride content should never exceed
1.5 percent by weight of cement. Other stipulations are that calcium chloride
should be preferably completely avoided in pretensioned prestressed concrete
nor in the main concrete of posttensioned prestressed concrete unless there is
an impermeable and durable barrier in additional to any grout, between the
main concrete and tendons. Calcium chloride is also prohibited when high
alumina cement is used. The clause 6.3.8 as added in May 1977 to the BS
CPllO-Part 1, 1972 states that the total chloride content (expressed as
percentage of chloride ion by weight of cement) of the concrete mix arising
from the aggregate together with that from any admixture and any other
source should not exceed 0.06 percent for prestressed concrete, structural
steam cured concrete, and concrete for any use made with cement complying
with BS 4027 or BS 4248 and should not exceed 0.35 percent for 95 percent of
test results with no result greater than 0.5 percent for reinforced concrete
made with cement complying with BS 12 and plain concrete made with ce-
ment complying with BS 12 and containing embedded metal. It is evident
from Table 1 that even the unwashed sea dredged sand (washed in seawater)
is quite acceptable on this basis for structural concrete works.
The American Concrete Institute (ACI) Committee 222 has suggested
(ACI 201 Guide to Durable Concrete-Chapter 4) a chloride content limita-
tion of only 0.06 percent for prestressed concrete. For reinforced concrete in
moist environment it suggests a chloride content limitation of 0.10 percent
for structures exposed to chloride environment and 0.15 percent for struc-
tures not exposed to chloride environment. For above-ground building con-
struction, where the concrete is expected to stay dry, a limit of 2 percent is
generally recommended for reasons other than corrosion.
Land dug sand from a local sand quarry after washing to remove fine
grained soil contamination is widely used in making concrete in Singapore
and is considered as normal sand. The grading of sea dredged sand and nor-
mal sand compared with the grading limits for fine aggregate set by BS 882
and ASTM Specifications for Concrete Aggregates (C 33-80) are shown in
Table 2. From the gradation curves shown in Fig. 1, it is seen that the
grading for both types of sands lie within the BS 882 as well as ASTM C
33-80 grading zones.
The coarse aggregate used in the investigation was from a local granite
quarry, the grading of which lies within the limits for coarse aggregate of 19
mm maximum size as per BS 882.
For mixing, potable water was obtained from the drinking water tap and
seawater was collected from the sea off Singapore coastline. The percentage
of the main constituents of seawater used are as shown in Table 1. Total salt
content of seawater was found to be 3.67 percent by weight of seawater that is
slightly above the value of 3.5 percent at which level it is considered to be
harmful to concrete if used in mixing according to Biczok [4].
ACI 318-77 suggests that where possible water with high concentrations of
dissolved solids be avoided and sets an upper limit for chloride ion concentra-
tion as 500 ppm in mixing water (0.03 percent by weight of cement).
Cement used in this investigation is the normal portland cement manufac-
tured in Singapore and most commonly used in local construction.
Mixes
The casting, curing, and testing procedures were in accordance with the
BS 1881-Parts 3 and 4. A mixer of 0.2 m^ capacity was used for mixing con-
crete (non-air-entrained) and sufficient care was taken to obtain consistent
TABLE 2—Grading of sea dredged sand and normal

sand compared with BS 882 and ASTM C 33-80.
Percentage by Weight Passing
Grading Limits for Fine Aggregate
Grading of Sea Grading of BS 882 1201: Part 2 ASTM C 33-80,

Test Sieve Dredged Sand Normal Sand Grading Zone I Part 14
9.50 mm 97.41 97.49 100 100

4.75 mm 86.05 86.60 90 to 100 95 to 100
2.36 mm 61.25 61.88 60 to 95 80 to 100
1.18 mm 33.87 32.49 30 to 70 50 to 85
600/xm 12.47 8.19 15 to 34 25 to 60
300 ^m 3.71 1.20 5 to 20 10 to 30
150 /im 0.00 0.00 OtolO 2 to 10
fOO
BS St 2 GRADING ZONE-l LIMITS
ASTM~C33 GRADING
ZONE LIMITS
Q:
Uj
0.3 0.6 1.20 2.1,0 t,.7G

APERTURE SIZE, mm
FIG. 1—Grading of normal and sea dredged sand compared with the grading limits of BS 882
and ASTM C 33-80 grading.
mix for casting the specimens. In order to obtain a design cube strength of 30
N/mm^, the mixes were designed according to the procedure given in Road
Note No. 4 [5]. Specimens prepared from various concrete mixes comprised
of 100 and 150 mm cubes and 150 mm diameter and 300 mm long cylinders
with and without embedded steel bars. The details of concrete batches and
mix types in each batch are given in Table 3. Four mixes were used for each
condition.
An admixture of 2 percent by weight of bentonite was used in Mix D2
because the authors felt that it would be interesting to study its effect on
workability, strength, and durability of concrete in view of its ability to
reduce the permeability of concrete.
Workability and Compressive Strength

Concrete mixes using sea dredged sand and seawater for mixing were
generally found to be more workable than the normal mix. The slump and
compaction factor obtained for various mixes are given in Table 4.
TABLE 3—Details of concrete mixes used for making specimens.
Constituents in Addition to Portland Cement

Mix Type and Granitic Coarse Aggregate
Batch 1°
Al Normal sand + potable water
Bl Sea dredged sand + potable water
CI Sea dredged sand + seawater
Dl Washed sea dredged sand + potable water
El Washed sea dredged sand + seawater
Batch 2*
A2 Normal sand + potable water
B2 Sea dredged sand + potable water
C2 Sea dredged sand + seawater
D2 Sea dredged sand + seawater + 2 percent bentonite by weight of water
"Per cubic metre contents: cement 316 kg, sand 632 kg, coarse aggregate 1391 kg, water-
cement ratio = 0.57, slump 30 to 40 mm.
*Per cubic metre contents: cement 338 kg, sand 601.4 kg, coarse aggregate 1223.6 kg, water-
cement ratio = 0.57, slump 50 to 70 mm.
TABLE 4—Stump and compaction factor for various mixes.
Approximate 28
Slump, Compaction Day Strength, Approximate
Mbc Type mm (in.) Factor N/mm^ Percentage
Al (normal sand
-I- tapwater) 30 (1.2) 0.83 30 100
Bl (sea dredged sand
+ tapwater) 35 (1.4) 0.84 29 97
CI (sea dredged sand
-I- seawater) 40 (1.6) 0.86 32 107
Dl (washed sea dredged
sand -1- tapwater) 30 (1.2) 0.84 30.5 102
El (washed sea dredged
sand -I- seawater) 40 (1.6) 0.85 31 103
A2 (normal sand +
tapwater) 50 (2.0) 0.91 30.5 100
B2 (sea dredged sand
-I- tapwater) 55 (2.2) 0.92 29 95
C2 (sea dredged sand
-I- seawater) 65 (2.6) 0.93 33.5 110
D2 (sea dredged sand,
seawater +
bentonite 70 (2.8) 0.95 33.5 110
Concrete mixes of Batches 1 and 2, each designed for a strength of 30

N/mm^ show consistent parallel results (Fig. 2) despite different mix propor-
tions. It is seen that sea dredged sand compares favorably well with normal
sand so far as the compressive strengths are concerned. From Fig. 3, it is
35 -- I 7- DAy STRENGTH
E [ 2& - OAy STRENGTH
I
I
o 30
-1 I I I I
I
T
25 -
1
1
1
r T
i.
1 ; L
•
1
T 1 1
T
\
1
T 1
J.
i 1
o 1
i
X.
20 1 1 1 1 1 1
Bl CI D! El A2 B2 C2 02
BATCH I BATCH 2
MIX Types
FIG. 2—Comparison of compressive strengths of various mixes.
V.
>^ no -
5^
a1
*l
Ul
< 1
<N 5-
Ul
u.
^ 90
o Ul
U) tr
ot
u
n
?• Washed sea sand and sea water
Uj
>.) -r 70 Washed sea sand and potable water
2 Sea sand and sea water
Sea sand and potable Water
Reference mix
50 I _L
1^ 21 26
AGE , days
FIG. 3—Development of strength as a percentage of 28-day compressive strength of reference

mix.
seen that the sea dredged sand and seawater combination gives a rapid
development of strength. The 7-day strength of Mix Type CI (seawater and
seasand) is very near to the 28-day strength of Mix Type Al (potable water
and seasand) and the 14-day strength of Mix Type Bl (potable water and
normal sand). The compressive strengths of all the mixes tend to be in a
similar range after the age of 28 days.
It is interesting to note that bentonite addition (D2) enhances the
workability further (Table 3) and the compressive strength is also increased
by about 10 percent, the water-cement ratio being the same as for the other
mixes in Batch 2.
Efflorescence
Efflorescence tests were conducted for mixes from Batch 1 as per BS
3921-Part 2. The efflorescence observed in all cases could only be described
as "slight" when potable water was used and "medium" when seawater was
used for mixing. There was no significant difference that was attributable to
salt contaminations from sea dredged sand.
Seawater containing large quantities of chlorides is known to cause persis-
tent dampness and surface efflorescence and because of this reason it is not
recommended for mixing where appearance of concrete is of importance or
where plaster finish is applied. Seawater containing a maximum concentra-
tion of 3.5 percent of salts is not supposed to be harmful to concrete from a
strength point of view [6]. Therefore, seawater can still be used for mixing if
appearance of the surface is not important. Unwashed sand from the seabed
with its insignificant salt content when used with potable water appears to be
quite suitable even for exposed concrete surfaces. Washed or unwashed, sea
dredged sand using potable water for mixing can therefore be considered for
concrete work where both strength and outward appearance are important
while unwashed sea dredged sand using seawater for mixing can still be con-
sidered for concrete work where strength is the primary requirement and
where efflorescence and dampness of surface can be tolerated. It is however
necessary to caution that when using unwashed sea sand with potable water
or seawater and washed sea sand with seawater for reinforced concrete and
prestressed concrete works it is important to observe the limitations on
chloride content as per BS CPllO-1972 and ACI Committee 222 as pre-
viously explained to safeguard against corrosion.
Durability
Standard concrete cylinders (150 mm diameter by 300 mm long) made of
Mix Types Al, Bl, CI, D l , E l , and D2 were subjected to adverse conditions
of curing and exposure to assess the long-term behavior. The specimens were
cured in 10 percent solution of sodium sulfate for a period of three months
and then kept submerged in seawater allowing it to ultimately evaporate ex-

posing the specimens to the indoor admosphere that had a relative humidity
of 70 to 80 percent a n d a temperature variation of 28 to 36°C. Although, the
specimens were at first intended to be tested much earlier, they were left in
that condition until two years later when the compressive strength tests were
performed. This is obviously not a standard durability test. However, the
authors thought that the conditions to which the specimens were subjected
represented an abnormally adverse condition so that any loss of strength
observed could be the worse that would be expected in any practical situation
of exposure.
The loss in strength for the specimens was generally in the range of 60 to 70
percent, except in the case of specimens for Mix Type D2 (with bentonite ad-
mixture) that showed a loss of strength of about 40 percent (Table 5). The
durability tests proved that both unwashed and washed sea dredged sand
when used with potable water for mixing could produce a concrete as durable
as that made by using normal sand and potable water. A small percentage of
bentonite seems to make the concrete more durable since the effect of ben-
tonite is to make the concrete more dense and impermeable.
Corrosion of Steel Reinforcement

In the case of reinforced concrete, the presence of sea salts is believed to in-
crease the risk of corrosion of reinforcement although there appears to be no
experimental evidence to the effect that even the use of seawater in mixing
can lead to such an attack [7]. When the reinforced concrete structures are
submerged permanently under water, the use of seawater in mixing seems to
have no ill effects of corrosion [8]. In practice however, it is generally con-
sidered inadvisable to use seawater for mixing unless it is unavoidable.
TABLE 5—Loss in compressive strength caused by curing in seawater.
Average 28-day Compressive Strength,

N/mm^
Mix Type Cured in Potable Water Cured in Seawater
Batch 1
Al (normal sand -1- potable water) 30 15.5
Bl (seasand + potable water) 29 14.8
CI (seasand + seawater) 32 10.5
Dl (washed seasand + potable water) 30.5 15.8
El (washed seasand + seawater) 31 12.4
Batch 2
D2 (seasand 4- seawater + 2 percent
bentonite by weight of water) 33.5 19.8
In order to study the effects of sea salts on steel reinforcement, standard

cylindrical specimens (150 mm diameter by 300 mm long) were cast with Mix
Types Bl (seasand + tapwater), CI (seasand + seawater), and Dl (washed
seasand + tapwater). Four 13 mm (0.5 in.) high tensile reinforcing steel rods
were polished and tied with links and were incorporated in each specimen
with a clear cover of 38 mm (1.5 in.). The specimens were cured in seawater
for three months and exposed to atmosphere for two years. The steel rods
were taken out of the specimens to examine the originally polished surface
for signs of corrosion. A thorough examination of the surface of steel rods did
not reveal any signs of corrosion.
The cement in concrete provides an alkaline environment of between 12.5
and 13.2 pH that protects the embedded steel reinforcement. In order to pre-
vent the steel from corroding, the alkalinity or the pH of the liquid phase in
concrete should not drop below a level of about 11.5 [9]. Tests conducted on
cement pastes prepared by using seawater for mixing showed a drop in pH
value from an initial value of around 13.75 to a final stabilized value of
around 12.50. The results obtained were in agreement with those obtained by
Gjorv and Vennesland [9] who concluded that an increase in salt concentra-
tions of up to 4.3 percent by weight of water can be tolerated before the pH
value is reduced to 12.
Conclusions
The importance of concrete as a building material continues to increase
and along with it increases the demand for sand. If sand from conventional
sources is in short supply, alternative sources must be considered to make up
for the shortage. Offshore sand deposits may come to the rescue of the con-
struction industry if sea dredged sand is accepted without reservations.
From the results of the study reported herein, it would appear that sea
dredged sand (which gets inevitably washed in the seawater during the dredg-
ing process) using potable water for mixing can be safely considered for con-
crete construction. The authors believe that they have found no indication to
the effect that the sea dredged sand should be forbidden for making con-
crete. Excessive chloride, if any, should however be removed from seasand by
washing it with freshwater. The unconditional use of unwashed seasand
should however be discouraged where corrosion of reinforcing steel is an-
ticipated.
By using sea dredged sand in concrete construction, an unutilized offshore
resource can be used to augment if not entirely satisfy the demand for normal
sand. This would no doubt result in savings both by way of material and
transportation costs and prevent the delays in construction operations
created because of difficult sand supply position.
Acknowledgments
The authors wish to thank their colleague, Dr. C. T. Tam, for his valuable
comments, and their students, A. H. Lim, D. M. Tan, and C. Y. Tan for
assisting in carrying out the various tests.
References
[/] Neville, A. M., Properties of Concrete, Pitman Publishing, London, 1977, pp. 135, 294,
353, 400.
[2] Newman, K., "Aspects of Workability, Strength, Shrinkage and Creep," Proceedings, Sym-
posium on Sea Dredged Aggregates for Concrete, Buckinghamshire, Dec. 1968, pp. 27-29.
[3] "UK Contractors Foremost in Aggregate Dredging," International Dredging and Port Con-
struction, July 1980, pp. 13-19.
[4] Biczok, I., Concrete Corrosion—Concrete Protection, Akademiai Kiado, Budapest, 1967,
pp. 91, 353, 475.
[5] Road Research Laboratory—Design of Concrete Mixes, Road Note No. 4, Her Majesty's
Stationery Office, London, 1950.
[6] Title No. 44-197—Water for Making Concrete, Proceedings, American Concrete Institute,
Vol. 44, 1948, p. 416.
[7] Demsay, J. G., Journal, American Concrete Institute, Vol. 48, Oct. 1951, pp. 157-166.
[tS] Shalon, R. and Raphael, M., Journal, American Concrete Institute, Vol. 55, June 1959, pp.
1251-1268.
[9] Gjorv, O. E. and Vennesland, O., Journal, American Concrete Institute, Vol. 9, Sept. 1976,
p. 512.
Q. L. Robnett1 and J. S. Lai1
Use of Geotextiles to Extend

Aggregate Resources
REFERENCE: Robnett, Q. L. and Lai, J. S., "Use of Geotextiles to Extend Aggregate

Resources," Extending Aggregate Resources, ASTM STP 774, American Society for
Testing and Materials, 1982, pp. 178-195.
ABSTRACT: The use of an interlayer of geotextiles or fabrics was found to allow a sub-
stantial reduction in the aggregate layer thickness required to provide a given level of per-
formance for aggregate-surfaced roads over soft soil deposits. Experimental results were
used to derive equations that relate key performance parameters for aggregate-soil sys-
tems containing Typar 3401 fabric. Numerous laboratory and field studies concerning
the use of fabric to reinforce aggregate layers were cited. Various design methods for
specific commercial product lines and two general design methods were used to illustrate
typical ranges in aggregate savings that may be effected by use of fabric.
KEY WORDS: geotextiles, fabrics, aggregates, roads, pavements, thickness design, concrete
Since early civilization, soon after the discovery of the wheel, man has been
building roads to support his transportation activities. The Romans were the
first scientific road builders with the "Appian Way" initiated in 312 B.C.
[I].2 The Appian Way, which was hand placed stone 1 to IV2 m thick, be-
came standard practice for over 2000 years until MacAdam and Tresagnet
developed new concepts for light surfaced roads [/].
The major component of most road structures, whether it be those of early
civilization or today's modern highways is aggregate. Even though many of
today's modern highways contain portland cement concrete or asphaltic con-
crete or both, of the almost 6.4 million kilometres of roads and streets in the
United States, over 2.2 million kilometres are either unsurfaced or surfaced
only with stone, slag, or gravel [2]. Many other areas of the world have an
even greater proportion of unsurfaced or aggregate surfaced roads.
Increasing demand (about 450 billion kilograms estimated for highways in
'Professors, School of Civil Engineering, Georgia Institute of Technology, Atlanta, Ga. 30332.
2
Copyright by ASTM Int'l (all 178 rights reserved); Thu Dec 22 07:4
ROBNETT AND LAI ON GEOTEXTILES 179
the United States for 1980 [3]), escalating costs, rapid depletion of supplies,
and regional shortages of quality aggregate have made it desirable to evaluate
alternate aggregate sources, alternate ways to extend aggregate resources, or
both.
New roads surfaced only with aggregate are constantly being built to pro-
vide access to and around construction sites, logging operations, mining and
quarrying operations, and as planned stage construction for higher type
roads. It is with these types of roads that geotextiles or fabrics used in con-
junction with aggregate may prove cost-effective and may allow reductions in
the quantity and quality of aggregate used. Broad-based experience indicates
that geotextiles are particularly advantageous and beneficial where very soft,
highly saturated, unstable soil conditions are encountered.
The purpose of this paper is to demonstrate that geotextiles or fabrics can
be used to extend aggregate resources. This will be accomplished by first dis-
cussing the general nature and the manner in which fabrics are used in
roads. Then, mechanisms by which fabrics benefit an aggregate road will be
discussed. And finally, experimental results and information in the literature
will be presented and discussed to illustrate the potential benefits derived.
Geotextiles
Definition
Geotextiles or civil engineering fabrics are defined by ASTM as, "Any
permeable textile used with geotechnical materials as an integral part of a
man-made project, structure, or system."
Fabric Types and Characteristics

A large selection of fabric products is available commercially. These syn-
thetic fabrics are commonly categorized based on construction and fiber
composition. Basically, the two categories of construction are woven and
nonwoven, while the fiber composition may be polypropylene, polyester, ny-
lon, or polyethylene.
Woven fabrics consist of filaments (either monofilament, multifilament, or
ribbon of split film) or yams oriented in two mutually perpendicular direc-
tions over-lapped one over the other. The woven fabrics can be constructed to
have the same or different mechanical properties in the two directions.
Nonwoven fabrics consist of a mat of discrete fibers of either continuous or
staple (discrete length) filaments, which are constructed in either a random or
preferentially oriented direction to a predetermined thickness or basis weight.
The fibers within the mat are then entangled or bonded by needle punching,
spunbonding, or resin bonding to produce the finished fabric product.
Critical and optimum properties and characteristics for use in roadways
have not been firmly established. Bell et al [4] suggest that tensile strength,
modulus, friction/adhesion, creep, bond strength, fatigue, failure elonga-
tion, and burst strength are important mechanical properties; Lavin et al
[8] and Robnett et al [9] have shown the importance of fabric modulus on
performance. Bell et al [4] also suggest that chemical stability, durability,
hydraulic conductivity, and constructability considerations are important.
Space limitations do not allow extensive discussion of fabrics but Refs 4 and
5 are excellent sources of information relative to fabric composition and
manufacturing processes, fabric properties, test methods, and end-use re-
quirements.
Mechanisms Responsible for Benefits

Fabrics are used in road construction in conjunction with a locally avail-
able aggregate such as crushed stone, uncrushed quarry rock or "shot rock,"
sand, gravel, sea shells, etc. to develop a structural layer. Figure la depicts
the general geometry of such a system. In this application, the fabric provides
reinforcement and separation benefits to the system [4,6,7].
Reinforcement Function
In the reinforcement function, it is postulated that the fabric serves to im-
prove the performance (often measured by resistance to pavement deforma-
tion or rutting) of the aggregate-fabric-soil (AFS) system under repetitive
vehicular loading due to a number of mechanisms including (a) restraint ef-
fect of the fabric on the aggregate and subgrade layer, (fe) membrane effect,
(c) friction developed at the fabric interfaces creating a boundary layer ef-
fect, and (d) local reinforcement effect.
Restraint Effects—Two types of restraint effects should occur in the AFS
system. The first is related to the reverse curvature of the fabric outside the
wheel path and the resultant downward pressure or apparent "surcharge"
applied to the soil, Fig. lb. Such an effect should increase resistance to shear
or plastic flow of the soil from the wheel path. A second type of restraint ef-
fect occurs when the aggregate at the soil interface tends to move from under
the loaded area but is restrained or given a tensile reinforcement due to the
presence of the fabric. The strength and modulus of aggregate material are
beneficially affected by this increased confinement.
Membrane Effect—As the roadway undergoes large deformation. Fig. lb,
the fabric is deformed and develops in-plane tensile stress, the magnitude of
which depends on fabric strain and fabric modulus. A normal stress perpen-
dicular to the plane of the fabric will be induced, the magnitude of which
equals the in-plane stress divided by the radius of curvature of the fabric.
The net effect is a reduction in the magnitude of stress imposed on the sub-
(fT">i(fTnii [ T n s /fTTTY
- i\uuhui\ " " ' • ' " ' "
\ AGGREGATE
^ . "ô*
infill
SUBGRAOE
/iTn>|
SUBGRADE '^FABRIC-INDUCED
FABRIC NORMAL STRESS
FIG. 1—Schematic of aggregate-fabric-soil suhgrade system; (a) initial state after construc-
tion, and (b) state after substantial heavy traffic.
grade under the wheel load. A reduction in the rate of rut formation in the
subgrade for a given vehicular loading condition will result. In order to de-
velop fabric-induced stress, substantial vertical deformations, proper geom-
etry, and fabric anchorage are required.
Friction and Boundary Layer Effect—Friction developed along the inter-
face between aggregate-fabric and friction/adhesion at the fabric-soil inter-
face create a "boundary-layer" or composite material of aggregate and soil
immediately adjacent to the fabric. The composite material due to the pres-
ence of the fabric should possess more favorable properties of ductility and
tensile strength. The effectiveness of this phenomenon is closely related to the
magnitude of friction/adhesion developed at the interfaces. Fabrics capable
of developing high friction/adhesion are desirable.
Local Reinforcement—Concentrated loading due to the aggregate layer
weight and imposed vehicular loading can cause a punching or local bearing
capacity failure at the points of contact between the aggregate and subgrade.
Use of fabric between the aggregate and soft soil will serve to distribute the
load, reduce localized stresses, and, in general, provide increased resistance
to vertical displacement.
Separation Function
In the separation function, the fabric serves to prevent the fine-grained
subgrade soil from intermixing with the coarse-grained aggregate material
and reducing its shear strength and stability. Depending on aggregate grada-
tion, 10 to 20 percent additional fines can cause a substantial reduction in
shear resistance.
As a result of these various mechanisms, the performance of aggregate lay-
ers has been found to be superior when "reinforced" with fabric that has the
necessary mechanical and chemical properties to resist forces of the environ-
ment into which it is placed. The degree of benefit offered by fabric depends
to a great extent on the physical, mechanical, and chemical properties of the
particular fabric used and specific job requirements.
Experimental Study
To demonstrate quantitatively the benefits derived from the use of fabric,
the results of a laboratory study will be presented. An extensive laboratory
study has been ongoing at the Georgia Institute of Technology to study per-
formance characteristics of fabric-reinforced aggregate layers over soft sub-
grades. A major objective of this study has been to evaluate the relative per-
formance of aggregate-soil (AS) and AFS systems and the benefits accruing
from the use of Typar 3401.^ Another objective that has been reported on
elsewhere [8,9] was to identify fabric properties critical to AFS performance.
General Description of Tests

In order to evaluate performance, a series of repeated load tests were devel-
oped and conducted using cylindrical test pits, 0.9 and 2.44 m in diameter. Fig.
2. Within these pits, a layer of soft silty clay soil was placed and then a layer
of dense graded crushed stone was placed on top. For one part of the series,
Typar 3401, a nonwoven, spunbonded polypropylene fabric, was installed
between the soil and crushed stone, while for the other part of the series, no
fabric was used in the system.
A repeated load was applied on a circular plate at the surface of the crushed
stone and the amount of permanent vertical plate displacement that accumu-
lated during the repeated loading was monitored.
Materials
Crushed Stone—The crushed stone material used for the testing program
was a crushed biotite granitic gneiss obtained from a local quarry in the At-
•'Registered trademark of the duPont Company.
U BELLOFRAM
AIR CYLINDER
LOAD PLATE
AGGREGATE
FRONT VIEW
1 ROLLER ^ - q g
NOTE; NOT TO SCALE
2.44m
NOTE: NOT TO SCALE
FIG. 2~Schematic diagrams o/(a) 0.9-in and (b) 2.44-m test pits.
lanta, Ga., area. The aggregate gradations used for the 0.9-m and 2.44-m pit
tests are shown on Fig. 3. Certain other engineering properties of the aggre-
gate are summarized in Table 1.
5oi/—The fine-grained soil material used in the study was dry-mill fire clay
(ASTM Specification for Ground Fire Clay as a Refractory Mortar for Layer-
ing-Up Fireclay Brick (C 105-47)) purchased from A. P. Green Refractories,
U.S. S T A N D A R D SreVE OPENING IN INCHES U.S. S T A N D A R D SIEVE NUMBERS HYDROMETER

6 4 3 2 TA 1 41 'A am 3 4 6 810 1416 20 30 40 50 70 100 140 200
I I I
- 9n ^
AGGREGATE USED
IN 2.44ni TESTS
5 1 0.5 0.1 0.05

GRAIN SIZE MILLIMETERS
FIG. 3—Grain size distribution of aggregates and subgrade soil.
TABLE 1—Summary of pertinent aggregate and soil properties and characteristics.
Aggregate
Aggregate description Biotite granitic gneiss

Maximum particle size 2.5 to 3.7 cm
Unified classification GW
Maximum dry density (ASTM D 698-78)" 139.1 Ib/lt^
Optimum moisture content (ASTM D 698-78) 6.3%
Angle of internal friction 48°
Soil
Liquid limit (ASTM D 423-66)* 27%

Plastic limit (ASTM D 423-66) 12%
Plasticity index (ASTM D 423-66) 15%
Specific gravity of solids (ASTM D 854-58)'^ 2.67
Unified classification CL
Maximum dry density (ASTM D 698-78) 109.8 Ib/ft^
Optimum moisture content (ASTM D 698-78) 18.5%
"ASTM Tests for Moisture-Unit Weight Relations of Soils and Soil-Aggregate Mixtures, Us-
ing 5.5-lb (2.5-kg) Rammer and 12-in. (304,8-mm) Drop (D 698-78).
*ASTM Test for Liquid Limit of Soils (D 423-66).
'ÂSTM Test for Specific Gravity of Soils (D 854-58).
Morris, 111. Properties and characteristics of the soil material are summarized
in Table 1 and on Fig. 3.
A series of tests were conducted on the soil to establish relationships be-
tween various measures of strength as follows:
1. unsoaked California Bearing Ratio (CBR) (ASTM Test for Bearing

Ratio of Laboratory-Compacted Soil (D 1883-73)),
2. Cone Index (3.2 cm^ cone as per Ref 10),
3. Torvane shear device, and
4. Proctor penetrometer (ASTM Test for Moisture-Penetration Resistance
Relations of Fine-Grained Soils (D 1558-71)), 6.5 cm^ tip.
Correlations between and among the various strength measures were de-
veloped for a range of strengths by compacting a large number of 15-cm-
diameter by 12-cm-high specimens at different moisture contents. All strength
tests were conducted at least once on each specimen. The results were then
analyzed and the various strength measures related to one another by using
linear regression. The resulting relationships are depicted in Fig. 4.
Fabric—A large roll of Typar 3401 was obtained for use in the testing and
the magnitude of various physical and mechanical properties was determined.
These results are summarized in Table 2.
Test Equipment and Test Preparation

0.9-m Test Pit—The 0.9-m test pit consisted of a piece of steel pipe, 75 cm
in length with a nominal inside diameter of 91 cm and a wall thickness of 7.6
mm. The pipe and load frame are depicted schematically in Fig. 2.
Loading system—A plate of Lucite'' acrylic polymer, 15 cm in diameter
and 5 cm thick, was used in conjunction with a size 36 Bellofram air cylinder
0.5 1.0 1.5 2.0 2.5 3.0

1 1 1 11 1 1
1 1 1 1 1
CBR
3 4 5 6 7 8 9 10 11 12
1
1
1 1
1
1
1
1
1
11 1
1
1
1 1
1
1
1
1
TORVANE SHEAR, psi
10 20 30 40 50 60 70 80 90 100 110
1 1 1 1
1 1
1 1
1 1
1 1 1 1
1 1 1 1 1 11 1
CONE INDEX
20 30 40 50 60 70 80
1 1
1 1 1 1
1 1
1
1
1 1
PROCTOR PENETRATION RESISTANCE, psi
NOTE: 1 psi = 6.89 kN/m^
FIG. 4—Strength correlations for subgrade soil.
Registered trademark of the duPont Company.
TABLE 2—Summary of pertinent properties of Typar 3401 fabric

used in laboratory study.
Weight (ASTM D 1910-64)" 136 g/m^

Thickness (ASTM D 1777-64)* 0.38 mm
Grab tensile (ASTM D 1682-64)"^ 0.6kN
Elongation to break (ASTM D 1682-64) 62%
Grab modulus (ASTM D 1682-64) 5.34 kN
Trapezoidal tear (ASTM D 2263)'' 0.33 kN
Puncture strength (ASTM D 751-73)" 0.22 kN
Mullen burst (ASTM D 774-67)^ 1380 kN/m^
Abrasion resistance (ASTM D 1175-71)* 0.19 kN
Specific gravity 0.95
Equivalent opening size 70 to 100 U.S. standard
(ASTM D 442-78)* sieve
"ASTM Tests for Construction Characteristics of Woven Products

(D 1910-64).
*ASTM Method for Measuring Thickness of Textile Materials (D
1777-64).
"ÂSTM Tests for Breaking Load and Elongation of Textile Fabrics
(D 1682-64).
''This standard has been discontinued.
•ASTM Method for Testing Coated Fabrics (D 751-73).
ÂSTM Test for Bursting Strength of Paper (D 774-67).
*ASTM Tests for Abrasion Resistance of Textile Fabrics (D 1175-
71).
*ASTM Specification for Gypsum Backing Board (D 442-78).
to load the surface of the crushed stone. The frequency and duration of air
flow to the Bellofram unit was controlled by an electric solenoid valve and the
pressure was controlled by a regulator. The loading system calibrated such
that a 480 kN/m^ peak pressure was applied by the plate 20 times per minute
for a total load pulse duration of 0.2 s.
Subgrade preparation—The subgrade soil was thoroughly blended with
the desired amount of water in a large mixer. The soil material was placed in
the test pit in small clumps to produce an uncompacted lift of about 5 cm in
thickness. The lift was compacted with two to three coverages of a pneumatic
tamper. The subgrade was constructed to a thickness of 38 cm.
For each compacted lift, numerous penetration resistance (ASTM D
155871) measurements were made. Additionally, for many tests, Torvane
shear strength was determined for each lift. The penetration resistance or shear
strength was used along with the strength correlations to estimate other
strength parameters such as unsoaked CBR, Cone Index, or vane shear
strength (if not determined directly). In general, the subgrade strength used in
the test series ranged in unsoaked CBR from 0.4 to slightly over 2. The sub-
grade strength was closely controlled for a given test by adjusting the moisture
content and degree of compaction.
Fabric—When Typar 3401 was to be utilized between the subgrade and
crushed stone, a 1.2-m diameter piece was placed and excess fabric was turned
up at the pit wall. The fabric was worked to produce a smooth taut surface
across the subgrade.
Crushed stone—The crushed stone was placed in loose lifts of 5 to 7 cm.
The pneumatic tamper was used to compact the material in four to five cov-
erages. Density was determined by using a nuclear gage. A density of about
94 to 96 percent of the maximum dry density (ASTM Tests for Moisture-Unit
Weight Relations of Soils and Soil-Aggregate Mixtures, Using 5.5 lb (2.5-kg)
Rammer and 12-in. (304.8-mm) Drop (D 698-78)) was maintained through-
out the testing program.
Surface instrumentation—With the loading plate in place, a series of
spacers were positioned between the plate surface and the Bellofram load
piston. Three dial gages were cantilevered from an anchored reference datum
and were used to measure the permanent deflections of three equally spaced
points on the plate perimeter.
2.44-m Test Pit—The 2.44-m test pit consisted of a 2.44-m diameter sec-
tion of corrugated metal pipe anchored in the concrete floor of the test lab-
oratory. The pipe had a reinforced concrete bottom. Many of the testing de-
tails were the same for the 2.44-m test pit as for the 0.9-m test pit.
Major differences were as follows:
1. A hybrid compressed air-hydraulic oil loading system was used to load
a 30-cm-diameter plate.
2. A 76-cm-thick subgrade of the same soil was used.
3. A slightly different aggregate gradation was used because of the larger
plate size, Fig. 3.
4. The aggregate layer thickness was about 38 cm.
Test Procedure
Once all soil fabric, aggregate, loading equipment, and instrumentation
were in place, initial dial gage readings were taken and loading of the plate
was begun. To determine the load cycle-cumulative permanent deformation
relationships, the loading was stopped and changes in surface dial gage
readings were determined at various load cycle intervals. Tests were carried
to a maximum of 20 000 load cycles or a minimum surface deflection of 11.5
to 14 cm (about 20 cm for 2.44-m pit), whichever occurred first. When
loading was completed, final dial gage readings were taken.
Test Results
Test results from the 0.9-m and 2.44-m pit tests are summarized in Table
3. Typical plots of vertical deformation or rut depth (RD) of the surface plate
versus the number of applied applications for the 0.9-m pit are presented in
Fig. 5. Similar results for the two 2.44-m pit tests are presented in Fig. 6.
188 EXTENDING A G G R E G A T E RESOURCES
§11 S |§§8§§8l
1—I ' J D <
<
•a
«
> o S g r-" i/^ iS S

O UO <
ro 00 I
I
o
A
I o o o o o '
?8; i (S w O
(S O O
- ^ lO I
-H -H >
-H (N I
;« ^ ^sg^si
« .2
r^ TT r^ n >
t/5 05 >o V <^* o* f*i T-3 >sd •^' fN ^' '^•^r^^m^^ô-rrr^-rt
!l
ce
H IT) O O 00 r^ f*^ 00 O^ T-H (N
n
o '^ o c4 o (S ^d K Q ^' o ^ »-^ ^ Ov <n ^* t--" V in
?: ^ (^ 1^ ^
Tr'<r•âôoo^•^^"vôo
r r <s Q
ON
•I&
O O O O O O ^ ^ - H t N o o o o o o" '-i i-*' -^ fs o
11
<N O ON Tj- 00 TT ^ - ^ fS ^^ <N O ON • ^ 00 r r so -^^ CM ^^
o *^
i/i c^ (N '-H r-' i/j r-" i-H i/J r--* i/i c^ fN r-I r^" i/j r^' 1-i U-) r-*
^ ,^^"
e
? O
?
100 1000
NUMBER OF LOAD APPLICATIONS
FIG. 5—Typical surface rut depth versus number of load application plots of 0.9-m test pit.
200 300 400

NUMBER OF LOAD CYCLES
FIG. 6—Rut depth versus number of toad application plots for 2.44-m test pit.
Analysis and Discussion of Experimental Results

One indication of the relative benefit of the fabric can be found by examin-
ing the rut depth versus number of load applications data presented in Fig.
5. For the three aggregate thicknesses examined, it is evident that inclusion
of fabric provides one of the following: (a) reduces substantially the amount
of rutting for a given number of load applications, (b) allows a substantially
greater number of load applications for the same amount of rutting, or (c)
allows an aggregate thickness reduction to attain the same rutting after a
given A^.
Another indication of the relative benefit of the fabric can be found by ex-
amining the slope of the rut depth versus number of load application plots,
Fig. 6, for the 2.44-m pit tests. It is found that the rate of rutting is 0.23 cm
per load for the system with no fabric and 0.025 cm per cycle for the system
with fabric. The inclusion of fabric reduced the rate of rutting by a factor of
about 9. As a result of known relationships between rutting performance and
aggregate thickness, it is obvious that the use of fabric translates to a reduced
aggregate thickness for the same performance.
In order to examine the fabric benefits derived over a range of soil strength
and aggregate thickness, the data from the 0.9-m pit tests were put into the
form of a "performance equation" wherein aggregate thickness, subgrade
strength, and number of load applications were related to rut depth.
In order to do this and in order to allow interpretation of results for a broader
range of loading and geometry, the loading, thickness, and soil strength pa-
rameters were converted for all tests to a single parameter called stress ratio,
SR. Stress ratio, as used herein, is defined as the ratio of load-induced verti-
cal stress, a^, at the subgrade surface at the beginning of the test divided by
the subgrade shear strength. Eq. 1.
SR = -y- (i>
where
Oy = vertical stress from wheel load and
S = soil shear strength.
The concept of SR has been used by others [6,7,11] in discussions of the

performance of aggregate-fabric soil and other soil systems. In soil mechan-
ics, the bearing capacity of a soil is often expressed in terms of a coefficient
times the soil shear strength where the coefficient, in general, corresponds to
the SR as defined in this paper.
The SR-values for the various 0.9-m and 2.44-m tests were calculated and
are included in Table 3. The magnitude of a^ was calculated by Boussinesq
theory for a uniform contact pressure on a circular plate using the following
equation
/ (r2 + z2)3/2 J ^^^
Copyright by ASTM Int'l

Indian Institute of Technology
that transposes to
r r / r_ \x2
2 ^-1.51
^ -i-5'
Ov= P ; + 1 (3)
where
ff„ = vertical stress under centerline of circular load, kN/m^;
p = uniform contact pressure over circular area, kN/m^;
z = depth below plane of loading, m; and
r — radius of circular loaded area, m.
From this equation, it is seen that for a given contact pressure, the magni-
tude of Oy for a given contact pressure varies only with the dimensionless ratio
of z/r. Thus, for the 2.44-m pit, the calculated vertical stress at a depth of 38
cm below the 30-cm-diameter plate (z/r = 2.5) should be the same as in the
0.9-m pit at a depth of 19 cm below the 15-cm-diameter plate {z/r = 2.5).
The 0.9-m pit test results were statistically analyzed with multiple linear
regression techniques available in the BMD Statistical Package [12]. The re-
sulting regression equations for the aggregate-soil system without and with
fabric are:
Without Typar
RD = [0.309 + 0.243(SR)''-^ logA^]^ (4)
r2 = 0.87
SE,^ = 0.18
n = 197
With Typar
RD = [0.265 + 0.190(SR)0-65 logiV]^ (5)
r2 = 0.83
SEert = 0.21
n = 229
where
SR = stress ratio, aJS;
RD — surface rut depth, in. (Note: 1 in. = 2.54 cm);
N = number of load applications;
SEest = standard error of estimate; and
n = number of observations.
These two equations thus represent a best-fit relationship between the vari-
ous parameters. The benefit of including the fabric in terms of crushed stone
savings can be determined by solution of the equations for a given set of condi-
tions. For example, assume the following: if TV = 1000, soil shear strength =
27.5 kN/m^ (CBR = 0.8), and RD = 7.6 cm (3 in.), an allowable SR of 2.80
and 4.28 is calculated for the system without and with fabric, respectively.
Equations 1 and 2 can then be used to solve for the thickness of crushed stone
or the depth of cover required over the subgrade to satisfy the allowable SR
values. The resulting thickness for the system with fabric is about 25 percent
less than the system without fabric. Extending this relative benefit analysis to
establish actual cost savings can only be done meaningfully on a job-by-job
basis because of the job-dependent nature (size, location, etc.) of costs asso-
ciated with aggregate, transportation, fabric, labor, etc.
Although a scale factor and field verification are needed before these per-
formance equations can be applied directly to full-scale road systems, the
2.44-m pit test results and 0.9-m pit test results show that fabric:
1. Improves drastically the rutting resistance of a system of aggregate over
a soft soil.
2. Allows a substantial reduction in aggregate layer thickness to effect
equal performance.
Other Information and Discussion Concerning Fabric Benefits

Numerous laboratory and field studies concerning the use of fabric to re-
inforce aggregate layers are reported in literature [6,8,9,14-21]. In the re-
cent International Conference on Fabrics in Geotechnics [22], at least five
papers were presented that showed positive structural benefits from the use
of fabrics in conjunction with aggregate support layers over soft soil. Studies
by Webster et al [19,20] have shown the positive performance (as measured
by rutting) benefits of using fabric based on full-scale road and traffic tests.
Steward et al [11] of the U.S. Forest Service have reported on full-scale field
tests that showed the positive benefits of fabric. Bender and Barenberg [6]
and Kinney [18], based on similar scale-model laboratory testing, demon-
strated the benefits of fabric. Lavin et al [8] and Robnett et al [9] have shown
fabric benefits and discussed the significance of certain fabric properties on
performance of aggregate-fabric-soil systems.
In addition to these studies and information, the authors are aware of five
design methods [23-27] for specific commercial fabric product lines and two
general design procedures, one by the U.S. Forest Service [//] that has had
field verification and one by Giroud and Noiray [7] that is based on simpli-
fied theory and limited experimental results. In concept, the latter method
can be used from a broad range of fabrics in that the procedure does require
design input concerning fabric modulus and percentage failure elongation.
The U.S. Forest Service design procedure [11], although implying broad ap-
plicability, probably is limited to use with nonwoven fabrics since it is based

on work by Bender and Barenberg [16].
Although space limitations do not allow a critique and discussion of each
of these design procedures, Fig. 7 summarizes the range of thickness require-
ments for aggregate surfaced roads with and without the inclusion of fabric
for all except the method of Ref 27. A close study of these various methods
reveals the following general conclusions:
1. In all cases, a substantial reduction in aggregate thickness is allowed

when fabric is used. The aggregate savings effected by the use of fabric gen-
erally range from 25 to 40 percent for soft subgrade conditions.
2. The fabric is most beneficial at low subgrade support values (for exam-
ple, CBR < 3).
3. In most cases, it is assumed that high quality aggregate is used although
the method by duPont [23] allows for substitution of lower quality aggregates
(such as sands, gravels, etc.) but with an associated thickness penalty.
DESIGN CONDITIONS
BUT DEPTH = 7 to 10 cm
AXIAL LOAD ~ 8 0 0 0 - 9 0 0 0 Kg
NUMBER OF LOADS • 1000
(FROM REF. 7, 23, 24, 25, 261
S2"
40 -
RANGE IN DESIGN
REQUIREMENT
WITHOUT FABRIC
30 —
10 -
RANGE IN DESIGN REQUIREMENTS
WITH VARIOUS FABRICS
1.0 1.5 2.0 2.5

SUBGRADE CBR, %
FIG. 7—Typical aggregate thickness requirements for haul roads with and without fabric.
Sununaiy and Conclusions

The use of geotextiles to extend aggregate resources has been examined.
Various mechanisms responsible for benefits were presented. Information
available in the literature and laboratory study results presented in this paper
overwhelmingly demonstrate that the use of geotextiles can lead to a sub-
stantial reduction (25 to 40 percent) in the thickness of aggregate layer re-
quired to provide a given level of performance for aggregate surfaced roads
over soft soil deposits.
Critical design and performance parameters include loading conditions,
fabric properties, soil properties, and aggregate properties. This paper ex-
amined the benefits effected by the use of Typar 3401 but did not examine
the relative benefits of other geotextiles available commercially. All geotex-
tiles, however, will not provide the same performance benefits. Geotextiles
can be used to extend existing aggregate resources and may in fact allow for
the use of marginal or lower quality aggregate in certain types of roads.
Acknowledgments
We wish to thank the E. I. duPont Company for providing the financial
support for this study and for granting permission to publish the information.
References
[/] Public Roads of the Past: 3500 B.C.—1800 A.D., American Association of State Higliway
Officials, 1962.
[2] Motor Vehicle Facts and Figures '78, Motor Vehicle Manufacturers Association, Detroit,
Mich., 1978.
[3] Cady, C. R. et al, "Promising Replacements for Conventional Aggregates for Highway
Use," National Highway Research Program, Report 207, Transportation Research Board,
1979.
[4] Bell, J. R. et al "Test Methods and Use Criteria for Filter Fabrics," FHWA-RD-80-021,
Federal Highway Administration, U.S. Department of Transportation, 1980.
[5] Koemer, R. M. and Welsh, J. P., Construction and GeotechnicalEngineering Using Syn-
thetic Fabric, Wiley, New York, 1980.
[6] Bender, D. A. and Barenberg, E. J., "Design and Behavior of Soil-Fabric-Aggregate Sys-
tems," Transportation Research Record No. 671, 1978, pp. 64-75.
[7] Giroud, J. P. and Noiray, L., "Design of Geotextile-Reinforced, Unpaved Roads," Pre-
print 80-177, The Use of Geotextiles for Soil Improvements, Session No. 64, American
Society of Civil Engineers Convention, Portland, Ore., 14-18 April 1980, pp. 143-170.
[8] Lavin, J. G., Murray, C. D., Murch, L. E., Robnett, Q. L., and Lai, J. C., "Prospects of
Spunbonded Fabrics in Civil Engineering," to be published in Proceedings, Non-Woven
Fabrics Conference, University of Manchester Institute of Science and Technology, 24-26
June 1980.
[9] Robnett, Q. L. et al., "Use of Geotextiles in Road Construction: Laboratory Study," to be
published in Proceedings, First Canadian Symposium on Geotextiles, Calgary, Alberta,
Canada, Sept. 1980.
[10] Soil Traffickability Testing, Section 15, Materials Testing Manual, TM-5-530, U.S. Army,
Feb. 1971.
[//] Steward, J., Williamson, R., and Mohney, J., "Guidelines for Use of Fabrics in Construc-
tion and Maintenance of Low Volume Roads," USDA, Forest Service, June 1977.
[12] Biomedical Computer Programs—P Series, Health Sciences Computer Facilities, Depart-
ment of Biomathematics, School of Medicine, University of California at Los Angeles,
University of California Press, 1979.
113] Kinney, T. C. and Barenberg, E. J., "Soil Movement in Geotextile Reinforced Roads,"
Preprint 80-177, The Use of Geotextiles for Soil Improvements, Session No. 64, American
Society of Civil Engineers Convention, Portland, Ore., 14-18 April 1980, pp. 119-142.
[14] Jessberger, H. L. in Proceedings, International Conference on the Use of Fabrics in Geo-
technics, Ecole Nationale Des Fonts et Chaussees, Paris, France, 20-22 April 1977, pp. 9-
14.
[15] Jarrett, P. M., Lee, R. A., and Ridell, D. V. B. in Proceedings, International Conference
on the Use of Fabrics in Geotechnics, Ecole Nationale Des Fonts et Chaussees, Paris,
France, 20-22 April 1977, pp. 19-22. 4
[76] Barvashov, V. A. et al in Proceedings, International Conference on the Use of Fabrics in
Geotechnics, Ecole Nationales Des Fonts et Chaussees, Paris, France, 20-22 April 1977,
pp. 67-69.
[17] McGown, A. and Andrawes, K. Z. in Proceedings, International Conference on the Use of
Fabrics in Geotechnics, Ecole Nationale Des Fonts et Chaussees, Paris, France, 20-22
April 1977, pp. 161-166.
[18] Kinney, T. C , "Fabric Induced Changes in High Deformation Soil-Fabric-Aggregate
Systems," PhD thesis. University of Illinois at Urbana-Champaign, 1979.
[19] Webster, S. L. and Watkins, J. E., "Investigation of Construction Techniques for Tacti-
cal Bridge Approach Roads Across Soft Ground," TRS-77-1, U.S. Army Waterways
Experiment Station, Feb. 1977.
[20] Webster, S. L. and Alford, S. J., "Investigation of Construction Concepts for Pavements
Across Soft Ground," TRS-78-6, U.S. Army Waterways Experiment Station, July 1978.
[21] Bell, J. R., Greenways, D. R., and Vischer, W. in Proceedings, International Conference
on the Use of Fabrics in Geotechnics, Ecole Nationale Des Fonts et Chaussees, Paris,
France, 20-22 April 1977, pp. 71-76.
[22] Proceedings, International Conference on the Use of Fabrics in Geotechnics, Vol. I and II,
Ecole Nationale Des Fonts et Chaussees, Paris, France, 20-22 April 1977.
[23] "A Method for Constructing Aggregate Bases Using TYPAR Spunbonded Polypropylene,"
E-21317-1, DuPont Company, Wilmington, Del., Oct. 1978.
[24] "Constructing Access Roads with Mirafi 140 Fabric," PM-6, Celanese Fibers Marketing
Company, New York, N.Y., Sept. 1975.
[25] "Mirafi 500X Fabric, Design Guidelines: Haul Roads and Area Stabilization," MTB-1,
Celanese Fibers Marketing Company, New York, N.Y., Jan. 1979.
[26] "Bidim Engineering Fabric, Installation Guide for Access Roads, Storage Yards, Parking
Lots, Logging Yards, Truck Terminals, and Board Roads," Monsanto Textiles Company,
St. Louis, Mo., 1979.
[27\ "Designing with 'Terram', A Summary of Techniques and Physical Data Used in the
Design of 'Terram'/Soil Structures," Imperial Chemical Industries Ltd., Pontypool,
Gwent, Great Britian, Oct. 1977.
A. V. Petty, Jr.l
Ceramic Roadway Aggregates
REFERENCE: Petty, A. V., Jr., "Ceramic Roadway Aggregates," Extending Aggregate

Resources, ASTM STP 774, American Society for Testing and Materials, 1982, pp.
196-212.
ABSTRACT: Through a cooperative program between the Bureau of Mines, Tuscaloosa

Research Center, and the Federal Highway Administration, ceramic aggregates having
high wear- and polish-resistance were developed. Three-hundred aggregate composi-
tions, incorporating a variety of low-cost "waste" materials were evaluated over a
30-month period. Aggregates were produced using conventional ceramic processing
techniques and fired at temperatures ranging from 900 to 1500°C. British Wheel and Los
Angeles Abrasion tests were used for initial screening of the aggregates. These data, in
addition to raw materials costs, its availability, and energy requirements, were used to
select nine compositions for circular track tests at Maryland and North Carolina Highway
Department facilities. Economic evaluations showed that present production costs, based
on a 1000 ton-per-day operation ranged from $10 to $120 per ton of material produced.
Guyana bauxite was used as a standard, and several of the selected compositions
developed surpassed the bauxite in performance and were lower in cost.
KEY WORDS: aggregates, roadway aggregates, cement, concrete, ceramic aggregates,

synthetic aggregates, polish resistance, wear resistance
With the objective of conserving mineral resources, the U.S. Bureau of

Mines, Tuscaloosa Research Center, has, for several years, been engaged in
research to use various industrial and mining wastes as potential raw
materials for producing ceramic products. In 1977, a program was under-
taken to develop synthetic aggregates that were resistant to wear but able to
maintain a high level of skid resistance under severe traffic conditions.
Primary emphasis was given to laboratory investigations of raw materials,
and processing parameters and measurements were made to determine the
physical and mechanical properties of the aggregates produced. This
research was a cooperative effort with the Federal Highway Administration
(FHWA) under Task 2 of the Federally Coordinated Program (FCP) Project
'Ceramic engineer, Ceramic Research Group, U.S. Bureau of Mines, Tuscaloosa, Ala. 35486.
Copyright by ASTM Int'l (all rights reserved); Thu Dec 22 07

PETTY ON CERAMIC ROADWAY AGGREGATES 197
4G entitled, "Ceramic Processes for Production of Wear-Resistant and

Polish-Resistant Aggregates for Pavement Surfaces."
Advantages of synthetic aggregate over many types of natural mineral ag-
gregate include the potential for improved skid- and wear-resistance, the
ability to produce aggregate in areas lacking suitable natural materials, and
the utilization of low-cost "waste" materials. For these reasons, a variety of
raw materials from differing geographic locations were considered.
A problem with many naturally occurring aggregates, particularly car-
bonate rocks that are widely used where available, is that they are composed
of minerals fine in grain size, tightly bonded, uniform in hardness, and thus
the rocks wear at an even rate. This eventually results in polishing of exposed
surfaces. Other rocks, like some sandstones, are composed of minerals with
coarse angular grains, variable hardness, and relatively weak bonding. Wear
takes place differentially with the dislodgement of individual crystals before
the exposed surface becomes polished. Abrasion of the surface however may
be excessive. Aggregates such as expanded clays and shales have a vesicular
structure that does not polish but is susceptible to wear.
As described by J. R. Hosking,^ acceptable aggregate materials can be
classified into five categories: (1) very hard materials; (2) conglomerations of
small hard particles; (3) dispersions of hard particles in a softer matrix; (4)
materials that fracture in an irregular, angular manner; and (5) vesicular
materials. Aggregates representing each of these categories were evaluated.
Description of Raw Materials

Calcined Guyana bauxite, although limited by its high cost, has been
evaluated for skid- and wear-resistance in both the United States and Great
Britain. British Wheel test data are available in the literature (see Footnote
2). A sample of kiln-run RASC (Refractory A-Grade Super-Calcined) baux-
ite, the only grade now being imported by the United States, was obtained
and used for comparison throughout the laboratory experimental stage.
Other commercial materials evaluated or used in the development of com-
posite mixtures included calcined domestic high alumina clays (AI2O3 rang-
ing from 45 to 70 percent), ball clays, low PCE (pyrometric cone equivalency)
(ASTM Test for Pyrometric Cone Equivalent (PCE) of Refractory Materials
(C 24-72)) clays and shales used in the manufacture of structural clay prod-
ucts, silicon carbide, fused AI2O3 grain, bubbled AI2O3, a variety of crushed
commercial refractory bricks, tabular AI2O3, and calcined seawater peri-
clase.
Waste materials evaluated, either alone or in combination with other wastes
or commercial materials, included slate mining waste and slate mining over-
^Hosking, J. R., "Aggregates for Skid-Resistant Roads," Report LR 693, Transport and
Road Research Laboratory, Crowthome, England, 1976.
burden, sintered coal refuse, copper mill tailings, serpentine waste (asbestos
mining waste), fly ash, phosphate slime, aluminum processing wastes, waste
glass from municipal incinerators, metallurgical slags, and sands.
Proceduies for Analysis and Characterization

Chemical analyses and characterization studies included wet chemical anal-
ysis. X-ray diffraction, petrographic microscopy, scanning electron micro-
scopy (SEM), and transmission electron microscopy (TEM). Physical testing
included absorption (ASTM Method for Sampling and Testing Brick and
Structural Clay Tile, Part 16 (C 67-78)), unit weight (ASTM Test for Unit
Weight of Aggregate, Part 14 (C 28-78)), bulk density and apparent porosity
(ASTM Tests for Apparent Porosity, Water Absorption, Apparent Specific
Gravity, and Bulk Density of Burned Refractory Brick by Boiling Water,
Part 17 (C 20-74)), British Wheel Test (ASTM Practice for Accelerated
Polishing of Aggregates Using the British Wheel, Part 15 (D 3319-74T)),
with modifications to: (a) time of run, (b) feed rate of SiC, and (c) flow rate of
water, to meet Texas Highway Administration specification TEX-438-A,
British Portable Tester (ASTM Measuring Surface Frictional Properties Us-
ing The British Portable Tester, Part 15 (E 303-74)), Los Angeles Abrasion
Test (ASTM Test for Resistance to Abrasion of Small Size Coarse Aggregate
by Use of the Los Angeles Machine, Part 15 (C 131-76)), and Circular Track
Tests (Maryland [FHWA-MD-R-77-1] and North Carolina [ASTM Practice
for Accelerated Polishing of Aggregates or Pavement Surfaces Using a Small-
Wheel Circular Track, Part 15 (E 660-78)] State Highway Departments).
Description of Aggregate Production

Throughout the program conventional ceramic processing equipment was
used. Raw materials that required mixing were dry mixed and water added
based on the forming method selected. The material was then either pellet-
ized, extruded and chopped to size, or dry pressed. Firing involved periodic
gas-fired furnaces, rotary calciners, or moving-grate type furnaces. The
calcined material was then crushed using a jaw crusher and sized. A typical
process is schematically outlined in Fig. 1.
Exceptions to the preceding procedure are noted in cases where the raw
materials were melted, foamed, and heat-treated to produce crystalline glass-
ceramics and in cases where only calcination of "as-produced" waste mater-
ial and crushing and sizing were required.
In cases involving composite mixtures of calcined clay and other soft bond
phases, various fire clays from Pennsylvania, Missouri, California, Georgia,
and Alabama were added as pre-calcined material or extruded, calcined and
crushed to size prior to mixing with the softer bond phase material.
HARD BOND
GRANULAR PHASE
PHASE
1 »
1
MIXER
'
PUG MILL
'
EXTRUDER
DUST
ROTARY COLLECTOR
KILN AND
SCRUBBER
''
GRATE
COOLER
'
CRUSHER *.
1 * 1/2"
SCREEN
- 1/2
FIG. 1—Typical process flowsheet.
Experimental Results
Approximately 300 aggregate compositions were produced and evaluated.^
British Wheel Tests were used for initial screening, along with the cost and
availability of the raw materials, and energy requirements to produce the ag-
^Petty, A. v., Jr., "Ceramic Roadway Aggregates with Improved Polish- and Wear-
Resistance," Research and Development Report, Federal Highway Administration, to be
published.
gregates. Summarized in Table 1 are those composites rejected because of

the parameters discussed.
The remaining aggregate compositions were further evaluated using the
Los Angeles Abrasion Test (grading "C", 3330 ± 20 grams charge). Since
the primary emphasis was given to aggregates having improved polish- and
wear-resistance, samples having polish values (PVs) below 32 (British polish
numbers [BPNs] below 50) and L. A. Abrasion numbers above 40 were re-
jected. This led to the final selection of 11 potential candidate materials.
Characterization of these aggregates is summarized in Table 2. These
materials fall into three of the five categories of aggregates outlined by Hos-
king—very hard materials (Samples 150/159, 262, and 290); dispersions of
hard particles in a softer matrix (Samples 271, 273, 274, and 289) and
vesicular materials (Samples 79-83, 122A, 203, and 204). Samples 203 and
204, mixtures of aluminum processing wastes and high AI2O3 bauxitic clay
or low AI2O3 fire clay actually combine the properties of being very hard
materials and also possessing vesicular structures.
Experimental results indicated that for composite materials, mixtures of
60 percent by weight hard phase material finer than the 600-pim (30 mesh)
sieve and 40 percent by weight softer bond phase gave optimum results. This
mixture ratio was used in producing Samples 271, 273, 274, 289.
Production processes, along with BPNs and L. A. Abrasion Test results,
TABLE 1—Aggregate compositions rejected because of poor performance, cost or availability or

both, of raw materials, or high energy requirements.
Aggregate Type Raw Material Limitation
Composite mixtures of a tabular alumina" high cost of raw material

hard, angular material bubbled alumina"
and softer bond phase fused alumina grain"
calcined periclase"
silicon carbide"
calcined bauxite" high cost, imported raw
material
brucite" high cost, used in refractory
production
sand"/waste glass poor bond resulting in low
mixtures* polish numbers
phosphate slime*
Crushed commercial all types high cost
refractories
Glass-ceramics high energy requirements, low
polish numbers
Metallurgical slags presently used for roadway
aggregate
"Hard phase material.

*Bond phase.
4} § ^
bis = •5
p. ^5
M ^ e
&
MS o
til
•5 "
^ ^ _ I U'
111
J .2
1
o 0
(/I
& i-H r- -^
si £i
S
a <
oo to
^.
&
Oc 0
§> ^^ 00 • ^ r4 00 rn • ^ ^ c
•9 S.
C3 -ff
=; (N CO ô »/^_ 0 8 00 (
S •d ^ ' ^
ts
S
d^ X
Q. g •S
^ <: g
0-
^
«.
^
3 2 o
Q
-H rt • (S î
a,
•a
00 '^ m in r-)
. o^ 1-H (N r a m •^
"5"
. Ô t~- TT r-l
00 00 a^ 00
00 ~
•^
o S S
^
a
S i/i II '^
S
I M
<N
at
cq
<< T3
III
IT) ,
.a ^Q
It
!II1fill
H
i l o + + 1+1 + •o ii S
111
W U
ill ill*
= •3-^13
= !•, ^ -3 S
o rt g cj O ^ CH
+3 f ^ î-
S! ^^^
(r^
^ tS"
^ "^
.-(
are given in Table 3. The high BPNs and low L. A. Abrasion numbers should
be noted. BPNs ranged from 57.1 to 89.6 with five ranking higher than
average values obtained on Guyana RASC bauxite. The L. A. Abrasion
numbers ranged from a high value of 36.9 to a low of 19.1. Photographs of
the 11 selected aggregates are given in Fig. 2.
Sample 290, calcined serpentine waste, warrants special mention. At pres-
ent the Environmental Protection Agency prohibits the surfacing of any
roadway with asbestos tailings.'' However, these regulations do not consider
the use of thermally altered materials. Mineralogical analysis of the starting
material showed primarily antigorite with minor amounts of carbonate and
opaques, and traces of chrysotile (asbestos) and fibrous serpentine. However,
examination by the U.S. Bureau of Mines, Particulate Mineralogy Unit,
Avondale, Md., using optical microscopy, transmission electron microscopy,
and X-ray diffractometry of the aggregate following calcination (thermal
treatment) at 1350°C, showed the material to be composed of enstatite and
forsterite. No particles were observed that exceeded 5 to 1 in aspect ratio, and
no skeletal remains of chrysotile were observed.
FIG. 2—Macrophotograph showing the morphology of sintered coal refuse (Sample 79-83).
^Federal Register, Vol. 38, No. 66, 6 April 1973, p. 8822.
< < cp 05 <:'<;oa (0 < < <
13
Z
s a
J -
i o vO O ,-. -^ v£.
ON iW
00 r -
8K
.1 i sa|&i
-£; o 'y
-ji
5 53 " h o
£ Is w
"n -^ •« *; •«
"* •§ " -1 "^
s fr
rrll
B' ^ : £•: ^
S
n
|H
Pi It Pi
<
I I
+ + %M
id. MU !!ft.
If Piimvitmî^
5 ^ ~
II
II
FIG. 3—Macrophotograph showing the morphology of waste slate overburden (Sample 122A).
FIG. A—Macrophotograph showing the morphology of calcined high AUO, clav (Samole
150/159). I 3 y f
Copyright by ASTM Int'l (all rights reserved); Thu Dec 22 07

FIG. 5—Macrophotograph showing the morphology of aluminum waste plus refractory fire
clay (Sample 203).
FIG. 6—Macrophotograph showing the morphology of aluminum waste plus high AlJ^^j clay
(Sample 204).
FIG. 7—Macrophotograph showing the morphology of ci'pprr null iiiiliiig\ iSample 262).
FIG. 8—Macrophotograph showing the morphology of calcined clay plus fly ash (Sample 271).
FIG. 9—Macrophotograph showing the morphology of periclase plus waste glass {Sample 273).
f::}%^:
FIG. 10—Macrophotograph showing the morphology of calcined clay plus waste glass (Sam-
ple 274).
FIG. n—Macrophotograph showing the morphology of calcined clay plus low PCE clavs
(Sample 289).
FIG. 12—Macrophotograph showing the morphology of calcined serpentine wastes {Sample
As indicated in Table 3, seven and nine of the selected aggregates were

sent to Maryland and North Carolina Departments of Transportation,
respectively, for circular track testing. The remaining two aggregate com-
positions listed in Table 3, Samples 150/159 and 273, were not included in
the circular track tests due to the high cost of producing these materials. The
Maryland and North Carolina tests differ basically in that the Maryland
method evaluated only the aggregate whereas the North Carolina test eval-
uated an open-graded bituminous aggregate mix. The results of these tests
are summarized in Table 4. As stated in the report submitted by the Mary-
land Department of Transportation (MDOT), polish values for the seven
synthetic aggregates indicate higher frictional resistance than all previously
tested carbonate and serpentinite rock-types. Also, four reported polish
values are higher than any natural aggregate previously tested. The North
Carolina Department of Transportation (NCDOT) reported that problems
were encountered in obtaining an adequate bond and coating with AC 20
asphalt. Based on results of initial trial runs, the asphalt content was increased
and Vi percent by weight anti-strip additive was used to improve the adher-
ence of asphalt to the synthetic materials. The copper mill tailings were
especially difficult to coat relative to unit weight. Based upon the plotted
curves at VSN40 (variable speed number, 40 mph, Skid Resistance of Paved
Surfaces Using the North Carolina State University Variable—Speed Fric-
tion Tester (ASTM E 707-79)), the aggregates could be ranked for 4-h polish
as follows:
Aluminum waste and refractory clay
Calcined clay and low PCE clay
Aluminum waste and high AI2O3 clay
Calcined serpentine waste
Control
Copper mill tailings
Sintered coal refuse
Calcined clay and waste glass
Calcined clay and fly ash
Expanded waste slate overburden
Four of the aggregates ranked equal to or higher than the control aggregate
while five ranked lower.^
An economic and process evaluation was made by the Bureau of Mines,
Process Evaluation Group, Avondale, Md., based on a production rate of
1000 net tons per day of crushed and screened saleable product. A summary
of this evaluation-' is shown in Table 5, where the fixed capital cost (including
all currently required pollution control equipment), production costs per ton
of aggregate product, and the thermal requirements per 1000 tons of ag-
gregate product are given for each of the 11 aggregate compositions. The
'The control aggregate was a medium to low skid-resistant aggregate when polished.
s
"S ' 1 o<
• » TT - ^
^
11°-
J3 . 2
•<1- T T ^
ll
-
&
111
•*-• -4-' C
§ i
II
c e o - o
»s s s :
i
I
C « CS C
III
IS: •fl III
z0.
I •o 5 s
icS >n
oa
I 5"-c
rt w»
^1 _c «3
1 < Z '-' -^ Z - " ^ -
I
CO <
—m
My
a| ++
13 ° S S
l:^+ +I S
il
o
ll 11
o cs oj 03 m o
S "3 13 o o o u o u
II
u it
I
a^ o o o o <N O O O O.O
Sg
C> 2
^ <N
go ( ^ (TJ fO
Q CH
I
I vO lO 00 0 0
« I UO O ^^
in r^
00 o r*
o i/)'
o; _ f*S n' O
^H ^ (^ n • f s TT i n ^
•<,
I A • i n ^ *^
\ \Q O"^ i~i
> -5 ^ o
lO
in rs ^> 00 -^ in
*) i n Q • ^ '-D
( - ^ -O 00 <*)
^ vo in ^ in ^
<Q
i!
iifi
ii
a
PQ .2 S
<
1^ I (N Q r ^
ON <S K O
1 ^ ^ 1-H <S s
total production costs, including fixed capital costs, raw materials, utilities,
direct labor, plant maintenance, overhead, and fixed costs including taxes,
insurance, and depreciation over 20 years, per ton of aggregate produced
ranged from $10.62 to $120.09 per ton. It should be noted that for those nine
materials submitted for circular track tests the cost range is from $10.62 to
$53 per ton as compared to the present cost for Guyana bauxite of $235 per
ton FOB, port of entry.
The apparent high fked capital cost for Sample 79-83, sintered coal
refuse, is due to the cost of pollution control equipment for the sinter grate
furnace. The high fixed capital costs are offset by the reduced energy re-
quirements, and the actual production costs rank this as one of the less ex-
pensive synthetic aggregates.
Conclusions
Synthetic roadway aggregate can be produced using standard ceramic
processing techniques and equipment.
Synthetic aggregate offers improved polish-resistance over natural mineral
aggregate. British Polish Numbers ranged from 57 to 90.
Based on excellent wear- and polish-resistance and the lowest processing
costs, aggregate produced from calcined serpentine waste was rated best of
those materials evaluated.
Other aggregate materials showing excellent polish- and wear-resistance
and relatively low production costs included expanded waste slate over-
burden, calcined copper mill tailings, sintered coal refuse, and mixtures of
aluminum processing waste and high alumina clays or fire clays.
Production costs for the aggregates tested ranged from $10.62 per ton for
calcined serpentine waste to $120.09 per ton for mixtures of periclase and
waste glass. Other materials showing low production costs included calcined
copper mill tailings ($13.58 per ton), expanded waste slate overburden
($13.05 per ton), sintered coal refuse ($18.56 per ton), and mixtures of
aluminum processing wastes and high AI2O3 clays ($34.25 per ton) or refrac-
tory fire clay ($35.98 per ton).

Aggregates Resources STP774 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Aggregates Resources STP774 PDF

Uploaded by

Copyright:

Available Formats

Copyright by ASTM Int'l (all rights reserved); Thu Dec 22 07:41:33 EST 2011

ASTM SPECIAL TECHNICAL PUBLICATION 774

ASTM Publication Code Number (PCN)

€1' 1916 Race Street, Philadelphia, Pa. 19103

Quality Assurance in Pavement Construction, STP 709 (1980), 04-709000-08

Cement Standards—Evolution and Trends, STP 663 (1979), 04-663000-07

Recycling of Bituminous Pavements, STP 662 (1978), 04-662000-08

Low-Temperature Properties of Bituminous Materials and Compacted Bitu-

ASTM Committee on Publications

Jane B. Wheeler, Managing Editor

Aggiegate-Related Applications for Residues from Fluidized Bed

Combustion Boilers—R. J. COLLINS 3

Resource Conservation by Use of Iron and Steel Slags—D. w. LEWIS 31

Economics of Using Incinerator Residue as a Highway Construction

Beneficiation of Marginal Coastal Oregon Aggregates with Cement—

CHEN CHANG, R. G. HICKS, AND T. S. VINSON 74

Slag Utilization in Pavement Construction—j. J. EMERY 95

Properties of Marginal Aggregates Treated with Asphalt Emulsion—

Meeting Construction Aggregate Fracture Requirements with Natural

Sea Dredged Sand for Concrete—s. D. RAMASWAMY, M. A. AZIZ, AND

Use of Geotextiles to Extend Aggregate Resources—Q. L. ROBNETT

Ceramic Roadway Aggregates—A. V. PETTY, JR. 196

Copyright by ASTM Int'l

In its annual practice of developing and presenting a meaningful sympo-

Copyright by ASTM Int'l (all 1 rights reserved); Thu Dec 22

with another resource. Often called an "artificial," "ceramic," or "synthetic"

Aggregate-Related Applications for

REFERENCE: Collins, R. J., "Aggregate-Related Applications for Residues from

ABSTRACT: Continuing development of atmospheric fluidized bed (AFB) combustion

Atmospheric fluidized bed (AFB) combustion is among the most promis-

Copyright by ASTM Int'l (all rights reserved); Thu Dec 22

conventional stoker-fired coal burning boilers. The principal advantages of

Description of Research Program

Laboratory Characterization Studies

Alexandria, Va. Pope, Evans, and Robbins

U, S, Standard Sieve Size

Limestone Sorbents" Dolomite Sorbents

Alexandria Alliance Georgetown Rivesville Linden Minneapolis

LCI 3.78 9.89 11.73 6.56 12.63 10.64

U. S. standard Sieve Size

fines. In comparison, fly ashes from conventional pulverized coal-fired

Separation or Beneficiation Techniques

1. particle size separation and dry tabling,

Potential Commercial Applications

Alexandria Alliance Georgetown Windsor

LOI 47.95 38.73 35.76 30.16

Stabilized Road Base

TABLE 4—Summary of compressive strength development of stabilized road base

Compressive Strength, psi

"NT denotes no test performed.

was recognized for these materials. Early investigation of the possibility of

Further Field Tests

surface of each mixture for dimensional monitoring. Figure 3 shows the

AFB Residue Percent Percent Percent

'Sand and gravel, local to Canton.

TABLE 7—Dimensional stability measurements Canton, Ohio, test strips.

5/2/80 7/9/80 8/27/80 11/7/80

1 A 6.0 + 0.17 +0.17 + 0.34 + 0.34

"Refer to Fig. 3 for locations of measurements.

hammer blows required to drive a Number 4 reinforcing bar in 2.54 cm (1

TABLE 8—Average penelnition resistance measurements Canton. Ohio, test strips.

5/2/80 7/9/80 8/27/80 11/7/80

Mix No. Average Number of Blows for 101.6 mm (4 in.) Penetration"