PORTLAND CEMENT ASSOCIATION

RESEARCH AND DEVELOPMENT LABORATORIES

FACTORS INFLUENCING
PHYSICAL PROPERTIES OF
SOIL-CEMENT MIXTURES

5
!

-5

By Earl J. Felt

Authorized Reprint from Bulletin 108

of the Higfiway Research Board
1955. p. I38

Factors Influencing Physical Properties of

Soil-Cement Mixtures
EARL J. FELT, Manager
Transportation Development Section, Portland Cement Association
Introductory Remarks by the Chairman: Felt has presented important results of
the well-conceived and thorough investigations on soil-cement accomplished by
the Portland Cement Association. It i s especially gratifying to notice how in the
development of soil-cement general pedological principles and specific physicochemical concepts have been utilized a t an early stage, together with the moredirect approach irom soils and materials engineering.
The importance of this work for a better understanding and utilization of soilcement i s obvious. Less obvious (but perhaps of even greater significance) is
the fact that the basic principles developed a r e applicable not only to systems
composed of soil and portland cement but to soil stabilization with practically all
types of inorganic cementing materials.
This paper may fill another need: Some have suspected a negative attitude in
the emphasis given in the papers by our nonengineering contributors to the complexity of the science of soil stabilization and to the many problems that still r e main to be solved. It is well for the scientist topoint out these problems and for
the engineer to listen. But it is also the glory of theengineer to view and measure these problems and to solve them with hiscreative genius. How this is done
i s well demonstrated in Felt's paper.
*THE properties of compacted and hyganic matter may have on some soil-cement
drated soil-cement mixtures a r e dependent
mixtures, special laboratory studies (3)
and field experiments have been conducted.
to a great extent upon the types of soil
involved. Certain differences in propThis work has shown that sandy soils of
e r t i e s and in cement reaction a r e due to
this type may be satisfactorily treated.
variations in the chemical composition of
Soil identification and careful sampling
the soils. F o r instance, lateritic clays
a r e , of course, prerequisites to effective
with a relatively low silica-sesquioxide
soil-cement testing and construction. In
ratio (colloids low in silica and high in the
this connection, the Pedological system of
sesquioxides of iron and aluminum) pulwhich includes insoil identification
verize more readily and react more favformation on both the physical and chemiorably with cement than do northern podcal properties of soils i s recommende
zolic o r chernozem clay soils which have
for soil-cement work. Addtional informarelatively high silica-sesquioxide ratios.
tion of value in this field i s available
Laboratory studies (A,?) and field ex(5,9.
perlence have shown the differences that
Other factors which have pronounced
may occur in thepropertiesof soil-cement
influence on the physical properties of
mixtures made with different clay soils.
soil-cement mixtures include the quantity
Sandy soils, too, may react differently
of cement and water added; the density to
with cement dependingupon their chemical
which the mixtureis compacted; the length
makeup and surface chemical properties.
of time the soil, cement, and water a r e
For instance, a great variance exists bemlxed prior to compaction; and the degree
tween the relatively poor effectiveness of
of pulverization of the so11 if it i s a clay.
cement with the A horizon (topsoil) of many
The purpose of this report i s to show
podzol o r podzolic soils and the much
how these last-named factors influence the
greater effectiveness of cement with the
strength and the relative durability of soilC horizon (parent material) from the s a m e
cement specimens a s indicated by their
soil profile. In these soils the organic
resistance to loss in weight resulting from
matter in the A horizon isconsidered to be
alternate cycles of wetting and drying and
the chief offender, a s it may retard o r
of freezing and thawing in the laboratory.
almost completely a r r e s t the cement
Although soil-cement may be used for a
reaction.
number of purposes
this report conBecause of the tremendous effect o r cerns c o m ~ a c t e d soil-cement mixtures

(3

Q!

(z),

(and cement-modified soils, ' to a l e s s e r

extent) a s used principally for base and
subbase courses in highway and runway
construction. It does not include data on
"plastic" soil-cement, which is made with
much higher water contents than the compacted type.

cement mixtures. F o r convenience in

this report these test s e r i e s were identif ied a s follows: Series 1, effect of density;
Series 2, effect of molding moisture content; Series 3, effect of length of mifing
time; Series 4, effect of degree of pulverization; Series 5, effect of air-entraining cement; Series 6, effect of the quantity
of cement; and Series 7, effect of high
SCOPE OF PROGRAM
early strength cement.
~ a t aar e also reported f o r three fineToobtain representative data, indicative
of the performance of soil-cement mixgrain soils treated with relatively low
TABLE 1
GRADATION AND PHYSICAL TEST CON ST ANTS^ OF RESEARCH SOILS

,tal
Soil
No.

Clay
0.005
to
0.000
mm

Used in Test
s e r i e s No,

2a
2a-2
2a-3
2a-4
2a-6

1,2
3
1,2
1,2,3,5
6,7

4b-3
4b-4
4b-5
4b-6
4d

1,2,3,
3
1,2,3,5
4
1,2,6,7

2
3
2
2
1

17
12

1,2

2
2
1

!::,3,5

Physical
Test
constantsb

1,2

Group

NP

12
26
13
17
14

11 18
NP 21
1 17
NP20

7
8

24
26
25
24
24

38
35
34
37
34

13
12
10
12
14

25
26
24
20
16

10
9
11
9

35
40
51
36

53
49
37
47

49
47
51
37

26
26
28
18

17
18
17
16

14

18

68

118

- -

7d

57
59
64
67
67

--

83

Fine Sand
Sandy Loam
Loamy Sand
Sandy Loam
Loamy Sand
Silty Clay
Silty Clay
Silty Clay
Silty Clay
Silty Clay

A-2
A-2
A-2
A-2
A-2

Loam- A-4
Loam A-4
Loam A-4
Loam A-4
Loam A-4

Clay
Clay
Silty Clay
clay

A-6-7
A-6-7
A-6-7
A-7

14

Clay

A-7

a ~ b t a i n e dusing Standard AASHO and ASTM Procedures

b ~ L.
. = Liquid Limit
P. I. = Plasticity Index
S. L. = Shrinkage Limit
N. P. = Not Plastic
' ~ u r e a uof Public Roads
tures made of different types of soil,
percentages of cement (cement-modified
tests were made on mixtures containing
soils) and for three granular soils treated
sandy, silty, and clayey soils. Seven
both with lowpercentagesof cement a s r e series of tests were made to determine
quired for cement-modified soils and with
the influence of various factors upon the
higher percentages a s 'required for soilcompressive strength and resistance to
cement. These test s e r i e s were desigwetting and drying and to freezing and
nated: Series 8, properties of cementthawing of compacted, hydrated soilmodified fine-grain soils, and Series 9.
'Cement-mcdlfled sol, m m u r e s a r e those whlch contam l e s s
cement than the quantltles requ~redto defmtely harcen them
lo proQler roll-cement

Of cement-moddied
soils compared with properties of g r a m iar soll-cement mixtures.

The greatest use for soll-cement mixtures 1s In the construction of pavement

base courses. F o r this reason, the cement
contents used with the various soils when
conducting Series 1, 2, 3, 4, and 5 were
In the range that would normally be used
In building pavement bases. These cement
contents were 6 percent and 8 percent for
the sandy soil, 12 percent and 14 percent
for two silty soils, 12 percent for two
clay soils, and 20 percent for one verytough clay soil. Preliminary tests using
A.'rSHO Methods T135 (wet-dry test) and
T-36' (freeze-thaw test) were made to dete. mine the cement requirements.
In
co,?ducting Series 6, 7, 8, and 9, cement
contents ranging from 1)'~ percent to 34
percent were included in the tests.
MATERIALS AND TEST METHODS
The physical properties of the soils
used in the f i r s t seven test s e r i e s a r e
listed in Table 1. Similar data for soils
used in Series 8 and 9 a r e given in discussion of those series. The first group
in Table 1 is composed of five sandy soils
all identified as Soib 2a. (The numbers
after the 2a, i. e. , 2a-2, 2a-4, etc., indicate different samples from approximately
the same area. ) Second i s a group of five
silty soils; third is a group of four clayey
soils; and fourth i s a single clay soil,
very-heavy textured andextremely plastic.
The numbers used in identifying these
eoils a r e also the AASHO soil-group numbers, thus Soil 4b and Soil 4d a r e A-4 silty
soils, Soil 2a is an A-2 sandy soil, etc.
Soil 2a isa brown, fine, sandy soil from
South Carolina. It is a mixture of the
lower A, the B, and the upper C horizons
of the soil profile and contains some organic contamination from the A horizon,
This contamination varied in the different
samples, depending upon the percentage of
A horizon soil included. As previously
mentioned, organic matter has a deleterious effect upon proper cement hydration, and thus some of the soil-cement
made with samples of Soil 2a hardened a t
a slower rate and developed l e s s strength
than did soil-cement mixtures made with
other samples of soil from approximately
the same location.
Soil 4b (four samples) is a dark-gray,
silty soil from a well-drained a r e a in
Illinois. It is from the lower A horizon of
'ASTY Method. D559 d DSO. respectively.

brown clay soils from Illinois, from the

B and upper C horizons of the soil profile. These soils react with cement in a
normal manner.
Soil 7d is a light-brown, heavy clay
soil from Mississippi. It is from the B
and C horizons of the soil profile and is an
unusually tough clay.
Type I portland cement, used in most
of these tests, was a blend of four brands
purchased on the open market. The soilcement mixture was proportioned in the
laboratory on a dry-weight basis, with
cement contents selected to yield certain
predetermined percentages expressed on a
volume basis. Cement contents a r e r e ported, unless otherwise noted, in t e r m s
of volume of loose cement (94 Ib. per cu.
ft. ) per unit volume of compacted soilcement mixture. Thus, a cement content
of 8 percent by volume indicates that a
cubic foot of compacted mixture contains
0.08 sack (7.52 lb. ) of cement.
Test Methods

In most of the work, three different

tests were used in evaluating the influence
of various factors upon the quality of th
mixture. These were the wet-dry test,
the freeze-thaw test, and the compressivestrength test. Specimens f o r the first two
of these were 4 inches in diameter and 4.6
inches high. The testing methods were in
methods TI35 and
accordance with -0
T136, Wetting and Drying Test aml Freezing and Thawing Test, respectively, except that the specimens were not in all
c a s e s compacted according to the -0
standard procedure (T134) specified i n
these methods, and in some c a s e s the
tests were continued f o r more than the 12
cycles specified in the AASElO procedures.
The compaction procedure was varied in
several instances in order that the density
and moisture content of the specimens
could be varied and is described under
each s e r i e s of tests.
Specimens for the compressive strength
tests were 2 inches in diameter and 2

tons. When the pistons were separated by

2 inches, the specimen was a t the designed
density and contained the designed cement
content by volume. After removing the
specimens from the mold, they were stored
for cement hydration in an atmosphere a t
73 F. and 100 percent relative humidity.
At the selected age for testing, the specimens were removed from storage, immersed in water for an hour, and then
broken in a compression-testing machine.
In Series 8 and 9, liquid-limit and
plastic-limit tests, bearing-ratio tests,
and soniscope tests (Series 9) were made
to aid in evaluating the influence of cement
in modifyingthe soil. Test procedures for
these s e r i e s a r e described later in the
report.
Base L ~ n eDensity

Ficure 1. Moisture v e r s u s d e n s i t y of s o i l cement mixtures used i n S e r i e s 1 and 2.

inches in height, Specimens of this relatively small size were suitable a s the soils
used in the compressive-strength test were
relatively fine textured and did not contain
material retainedon the No. 4 sieve. There
i s no- AASHO standard method for making
compression-test specimens of soil-cement
mixtures.
Specimens were compacted from both.
ends using the double-piston method. A
predetermined weight of mixture containing
e proper moisture and cement content
was placed in the cylindrical mold and
compactive force applied through the pis-

.I"

Soil 4 b - 3

',

D m i t y pcf.
20-3
4b-3

-10

0
*$
-5
Wriation in h s i t y -pet.

96.5

I
*I0

Fippre 3. E f f e c t o f d e n s i t y oa soil-cement
loss from freeze-thaw t e a t .

TEST RESULTS
Selected AASHO
Metkod T I 3 4

\'

Sotl 76
20% Cement

\,

Soil

No.
-

20-3
4b-3
6

Soil-4b-3'

\\a

lot ,.poll
- 1

.-

Bosr Line
dens^ t y p d
122
96.5
98

Most soil-cement mixtures, when compacted in accordance with M H O test T134,

develop parabolic-shaped moisture-density curves, a s shown for Soil 2a and Soil
4b in Figure I, and a maximum density is
indicated a t an optimum moisture content
at the peak of the curve. These values,
obtained usingthe standard test, a r e called
"base-line" values in Series 1 and 2, which
were designed to study theeffect of density
and of moisture content on the properties
of soil-cement mixtures.
The clay soil-cement mixtures used in
the tests in Series 1 and 2 did not have
parabolic-shaped moisture-density curves.
These soils, because of their swelling
characteristics a s they become wet, tend
to develop irregular moisture-density

v.

'\

::--.-\.,.
.,:,
\

6e
12% Cement

'-A

Variation In Dmsity

8% Cement
20-3

-.

- pcf.

F i g u r e 2 . E f f e c t o t d e n s i t y on s o i l - c e ment l o s s from wet-dry t e s t .

curves of ski-slide shape, a s shown for

Soil 7d in Figure 1. Experience (1,2) has
shown that, when a curve of this shape i s
obtained, the selection of a base-line
density and base-line moisturecontent can
be done most accurately after special
tests have been made. The tests which
will be discussed in Series 2 a r e of particular value with soils of this type, a s
they oermit the engineer to select the
moist I r e content which produces maximum dfectiveness from the ce,ment. The
.ase. .me optimum moisture content of
4 1 - ement mixtures which have a secondary m p in the curve, a s for instance
Soil .-t in Figure 1, i s generally taken
about L percentage points above the water
content a t the second hump; the base-line
density is that obtained a t this particular
moisture content.
h e Lme D m s ~ f y
S.l.st@4 AASHO
Method TI34

ard procedure, except that the number of

blows of the compacting rammer per layer
was changed from 25, and was s e t a t
9, 15, 20 and 50. This procedure p r o w
duced specimens having a density variation
of about 10 to 15 ucf. (wunds per cubic
foot). Compressive-strength specimens
2 by 2 inches were then molded a t densities
and moisture contents equal to those of the
wet-dry and freeze-thaw specimens.
Sandy Soil 2a-3, silty Soils 4b-3 and
4b-5, and clayey Soils 6e and 7d wereused
in the tests. Figure 1 shows the baseline densities and base-line optimum moisture contents selected for study. Figures
2, 3, and 4 show the base-line densities
and also the variation in density from the
base.
Results from the wet-dry test a r e shown
in Figure 2 and from the freeze-thaw test
in Figure 3. The ordinates show the weight
loss from the specimens in 12 cycles of
test. The effect of density is readily apparent; specimens had increasingly higher
losses a s the density decreased. Although
all the different types of soil-cement mixture were benefited by increased density,
the silty and clayey soil-cement mixtures
were benefited the most. F o r the sand
Soil 2a-3, each 1-pcf. increase in density
reduced soil-cement losses about 1 percentage point. F o r silty Soil 4b-3 and
clayey Soil 7d, however, the corresponding
reduction in soil-cement losses varied
from approximately 1. 5 to 3. 5 percentage
points for each 1-pcf. increase in densi
in the low density range. (Soil 4b soil
cement contained 8 percent of cement in the
wet-dry test, but 14 percent in the freezethaw test, which is critical for this soil. )
The compressive strengths of specimens
molded a t different densities a r e shown in
Figure 4. Again, the value of high density
i s apparent. Inthe caseof sandy Soil 2a-3,
an increase in density of 1 pcf. resulted
in an increase in compressive strength of
approximately 20 psi. The compressive
strengths of the silty and clayey soilcements practically doubled a s a result of
increasing the density of these materials
10 to 15 pcf. F o r instance, with Soil 4b-3,
at a density 5 pcf. below base-line density,
the compressive strength was 235 psi.,
whereas at adensity 5 pcf. above base-line
density, the compressive strength was
400 psi. Soil-cement mixtures of clay
Soils 6e and 7d showed similar performance. In general, with these soils, a n

'Y

F i g u r e 4.

Effect o f d e n s i t v on 28-day canpressive stren~th.

Series 1: Effect of Density

Experience has shown that soil-cement
mixture of good quality can be made by
compacting the mixture to a density equal
to that obtained by the AASHO compaction
procedure. In Series 1an investigation was
made of the quality of soil-cement mixture
compacted to lower and higher densities.
In these tests the moisture content for each
of the soil-cement mixtures was maintained constant and equal to the selected
base-line moisture content. T o vary the
density at this moisture content, specimens f o r the wet-dry and freeze-thaw
tests were molded according to the stand-

ure 6 a r e considered first; it i s apparent

from these data that soil-cement mixtures
made of silty Soil 4b-3 and 4b-5 andclayey
Soil 7d had much-less resistance to alternate freezing and thawing whe'n they were
c o m ~ a c t e dat moisture contents less than
the base-line moisture contents. As seen
in Figure 5, this was true in the wet-dry
test for soil-cements of clay Soils 6e and
7d, but the soil-cement mixture made of
Soil 4b was not affected significantly.
(Soil 4b mixture contained only 8 percent
of cement in the wet-dry test, but 14 percent of cement in the freeze-thaw test,
which i s critical for this soil. )
It i s apparent from these data that to
obtain high-quality mixtures from siltv and
clayey soils, the mixtures must b e g i a x e d a t o r above, never below, the AASHO
TI34 optimum moisture content when they
a r e compacted to AASHO density. These
data have been corroborated by other tests
at the Portland Cement Associationandare
in agreement with results obtained by other
investigators (2).
The effect of molding moisture content
Base-Line Ma~sture
Cmtent SeIecfed
AASHO TI34

-10
-5
0
15
110
V o r l a t m on Maslure Content, Pacmfaqe Pants

Figure 5. Effect of moisture content on

soil-cement loss from wet-dry t e s t .
increase in density of 1 pcf. in the low
density range resulted in an increase in

To study the effect of molding moisture

content on the quality of soil-cement mixtures, specimens for wet-dry and freeze. thaw tests were molded at moisture contents
below and above the selected base-line
moisture content, using the standard AASHO
compaction
procedure.
Compressive
strength specimens were molded at the
same densities and moisture contents.
Since the compactive effort was constant
and the moisture content varied, the density of the specimens also varied, but the
data indicate that the effect of moisture
content overshadows the effect of the differences in density.
The data from the AASHO wet-dry and
freeze-thaw tests a r e shown in Figures 5
and 6. For convenience, the data in Fig-

6% Cement

Sod 6e
12% Cement

'..

So114 b - 3

20% cemrn

-5
0
.5
.I0
-10
vornotm an Mo,rture Contmt, Pacmtaqe Pomts

.I

Figure 6. Effectof moisture contenton 9011cement loss from freeze- thaw test.

on soil-cement mixtures made of sandy

Soils 2a and 2a-3 wag not the same as that
found f o r the silty and clayey soils. In
this case, moisture contents slightly on
the dry side of base-line moisture content
were favorable, suggesting that sand mixtures may follow to some degree
- : -&
cement-ratio relationships fo~-conc-rsd:e.
The compressive-strength data for the
various soil-cement mixtures a r e shown in
Figure 7. Here it i s seen that specimens

optimum moisture content for mixture of

Soil 7d (36 percent) i s practically equal to
the plastic limit (35 percent) of the soil,
a s indicated by the numerical difference
between the P. I. and L. L. data in Table
1. It i s of interest also to note that the
plastic limit of Soil 6e (23 percent) is
practically equal to the base-line optimum
moisture of 23. 5 percent, which was selected f o r Soil 6e soil-cement mixtures.
Other tests on soils having irregular moisture-density curves confirm this generality.
The wet-dry, freeze-thaw, and compressive strength data, when considered
together, indicate that for maximum effectiveness from the cement, sand mixtures should b e compacted a t optimum
moisture content o r slightly d r i e r , whereas
silty and clayey mixtures should be compacted a t moisture contents 1 o r 2 percentage points above optimum moisture.
Special Series 1 and 2
The test data f o r Series 1and 2 give rise
to questions a s to the relative performance
of soil-cement mixtures compacted to
much-higher densities than those obtained
in Series 1. In following the procedure d
Series 1, i. e., molding specimens at different densities a t the' selected base-line!

Fiaurea 7 .

E f f e c t o f moisture c o n t e n t on 28dev compressive s t r e n g t h .

molded at the base-line moisture content,

o r slightly below, except for Soil 7d, have
maximum strength. The data for mixtures
of Soil 7d indicate that the base-line moisture content selected for this soil was
about 5 percentage points below the moisture content which produced maximum
effectiveness from the cement.
Based upon the wet-dry, freeze-thaw,
and compressive-strength data combined,
it appears that the best moisture content
for Soil 7d soil-cement mixtures would be
about 36 percent, o r 5 percentage points
above the selected base-line moisture
content of 3 1percent. As previously mentioned, the so-called optimum moisture
content (that producing maximum effectiveness from the cement) for soil-cement

Soils used in these tests to determine

value of high densities included sandy
1 23-4, silty Soil 4 4 and clay Soil
-4. The cement contents used were
those that would commonly be used with
these soils in pavement base construction.
As shown by Figure 8, the maximum
densitiesof the soil-cement mixtures cornpacted by the AASHO Modified procedure
were considerably greater than the rnaxlmum densities obtained with the standard
method. For sandy Soil 2a-4 the AASHO
Modified density was 7. 5 pcf. higher, and
for the silty andclayey soils, approximately
13 pcf. higher. The optimum moisture
contents were correspondingly lower with
the AASHO Moddied procedure, being
about 2 percentage pomts lower for the
sand mix and approximately 5 percentage
points lower for, the silt and clay mixtures.
Data in Table 2 indicate the relative
quality of mixture produced using basic
molding data obtained with the two methods. In this table, data a r e presented f o r
specimens compacted a t AASHO standard
maximum dqnsity and AASHO standard
optimum moisture content (A); AASHO
Modified maximum density and AASHO
Modified optimum 3oisture content (B);
and AASHO standard maximum density
and AASHO Modified optimum moisture
content (C). The data obtained show that
compressive strengths obtained using
System B were considerably greater than
those obtained using System A. This we
ight expect, a s according to Series 1 the
high densities a r e beneficial. Apparently
at these higher densities l e s s water is
required in the mixture to effectively
utilize the cement, and the acceptable
moisture content i s thus lower for the
densities achieved by System B than for
the lower densities of Series 2. As there
is little difference between the wet-dry and
freeze-thaw data for specimens compacted
by Systems A o r B and a s the compressive
strengths a r e much higher using System
B, it appears that the soil-cement mixture
compacted by System F i s of the higher
quality.
It i s important to note that the density
indicated by the AASHO Modified method
a s maximum for the silty and the clayey
material i s s o great that it is difficult to
obtain in construction practice. Methods
of compaction which will produce this density a r e not generally available, and the
density most likely to be obtained in con-

struction is about that indicated a s maximum by the ASTM method.

It i s apparent that if only this latter
density were obtained with a mixture compacted a t the relatively low AASHO Modified
optimum moisture content, the resulting
quality of the mixture would be impaired.
This i s indicated in Table 2 by the compressive-strength data for the mixtures
compacted using System C. For example,
for the mixture using Soil 4d, the 28-day

Y c n h n - O m ~ l yR H a t a

20-4
46
60-4

AASW S t a b (

Yddl(nd AASW

mzDrr(blMt
pd
2

blmDn W Y C .
PC(
Z
128 5
9
121 5
12 5
,1155
(38
15
20

121

,
5

I
1025

I0 1
11
,185
10

umtwr -,-a

25

Figure 8. histure-denaity telmtions of

soil-cement mixtures, Special Series
1 and 2.

compressive-strength test f o r System C

shows only 293 psi., as compared with
1,202 psi. for System B and 668 psi. for
System A. System C also yielded inferior
soil-cement mixtures with the clay Soil
6e-4. However, the sandy mixture compacted using System C appears to be satisfactory.
These data correlate well with the data
in Series 1 and 2, where i t was shown that
for AASHO standard density the molding
moisture content of. the silty and clayey
soil-cement mixtures must be equal to o r
slightly greater than the base-line AASHO
standard optimum moisture content but for
sandy soils a somewhat lower moisture
content was satisfactory.
It appears, therefore, that if densi-

I
Compaction
Systema

1
1

Density.
pcf

Cy;hl:;;%

11

Compressive
Strength, ps1
A E in
~ Davs

1 Wet-Drv / ~reeie- haw

LOSS, %

LOSS, %

No. of Test Cvcles

Soil 2a-4, 8'3 Cement by Volume

Soil 4d, 12 q Cement by Volume

A
B
C

108
121.5
108

17
12. 5
12.5

352
787
229

596
933
277

668
1202
293

3
3

7
27

3
3

6
6

31
28

86

97

35
6

Soil 6e-4, 12 LR Cement By Volume

A

B
C

102.5
115.5
102.5

18. 5
13. 8
13. 8

203
326
70

417
427
138

486

709
149

a~ - AASHO Standard maximum density, AASHO Standard optimum moisture content

B - AASHO Modified maximum density, AASHO Modified optimum moisture content
C AASHO Standard maximum density, AASHO Modified optimum moisture content

ties a s great a s those obtained by the

AASHO Modified compaction procedure
could be assured, in combination with
their respective optimum mcisture contents, soil-cement mixtures of superior
quality would be produced. However, a s
these densities a r e not readily attained in
practice, it usually i s desirable to construct a soil-cement mlxture near the
AASHO standard optimum moisture content and to obtain a s high a density a s
possible, preferably equal to AASHO
standard density o r greater. This procedure will provide sufficient water in the
mixture to effectively utilize the cement
at the particular density obtained.
Series 3: Effect of Prolonged Mixing
Soil-cement pavement bases a r c f r e quently constructed using mixed-in-place
procedures in which the soil i s pulverized
in place, cement i s added, and the d r y
mix I S completed; water 1s then added and
the damp mlx 1s started. All of the r e quired water is not added at one time, and

several p a s s e s of the water equipment and

laboratory study of the effect of thii prolonged damp- mixing period was undertaken
in the next series.
With the thought of simulatingfield conditions, soil-cement was damp mixed f o r
periods of 2, 4, and 6 hours in the laboratory and then molded into test specimens.
During the mixing period, water,was added
to the d r y mix in equal increments a t approximately 20-minute intervals. After
each addition of water, the mixture was
mixed by s t i r r i n g for about 2 minutes,
The water added in each increment was
proportioned s o that a t the end of the specified time the mlxture was at optimum moisture content. The optimum moisture
content increases a s the length of mixing
time increases, so preliminary tests were
made f i r s t to determine the altered optimum moisture content of the mixtures.

Under unusual field conditions, mixture

that is damp may remain undisturbed f o r
ong p e r ~ o d swthout the intermittent mixing
reviously described. To obtain information on !he effect of such treatment, test
specimens were molded of mixtures that
had been brought to optimum moisture
content and had then been left in the loose
condition undisturbed (without intermittent
mixing) for periods of 2 , 4, and 6 hours.
Water lost by evaporation was replaced
just prior to molding the specimens.
Wet-drv and freeze-thaw specimens
were molded using the standard AASHO
compaction method, and compressive-test
specimens were molded a t the s a m e density and moisture content. Soils used in
these tests were sandy Soils 2a-2 and 2a4 , siltv Soils 4b-3 and 4b-5, and clay Soils
6e-2 and 6e-4. The cement contents used
were those which commonly would be used
with these soils in pavement base construction.
The AASHO optimum moisture contents
and maximum densities of the various
soil-cement mixtures subjected to intermittent mixing a r e shown in F i y r e 9.
These data show that the optimum moisture content increased and the maximum
density decreased a s the length of mixing
time increased. These effects suggest
that a loose, damp mixture, probably
through base exchange, develops a new
structure and texture a s it ages during
the prolonged mixing periods. That this
change s t a r t s soon a f t e r the soil, cement,
nd water a r e mixed i s evidenced by data
hich show that the resistance to penetration of a s m a l l piston forced into a freshly
compacted soil-cement specimen i s nearly
always greater than the penetration r e s i s t ance offered by the compacted raw soil a t
the s a m e moisture content.
As a result of 4 to 6 hours of intermittent mixing, the optimum moisture f o r the
sand soil-cement mixtures increased 0.6
percentage point; that f o r the s i l t soilcement mixtures increased about 1.3 p e r centage points; and that for the clay soilcement increased 1 to 2 percentage points.
The corresponding decreases in density
ranged from 2 pcf. for the sand and s i l t
mixtures to about 0 . 5 pcf. for the clay
rn ixtures.
Experience has shown that corresponding differences in optimum moisture content and maximum density may be considerably more in the field during mixed-

l&

in-place construction. Thus, the data

presented here, plus field experience,
emphasize the necessity in field construc tion of obtaining moisture-density relations
of soil-cement mixtures near the end of
the moist-mixing period, just before compaction begins. These field optimum moistures and field maximum densities should
then be used a s the c r i t e r i a f o r controlling
the compaction of the mixtures, rather
than laboratory o r field t e s t s made before
the damp mixing has been completed.
This is particularly important f o r mixt u r e s made of silty and clayey s o i l s , since
it i s imperative that they be compacted a t

5 1051

-- ,
d loo c
95

Sod 4 b - 4 B 4b-5,

Figure 9. Moisture-density relations of

soil-cement mixtures used i n Series 3.
moisture-contents at o r slightly g r e a t e r
than the new optimum.
Wet-dry and freeze-thaw t e s t data f o r
these soil-cement specimens, plotted in
Figures 10, 11, and 12, show that the soilcement losses increased a s the length of
the damp-mixing period increased. This
was most pronounced when the mixtures
were not intermittently mixed during the
period. F o r instance, f o r Soil 4b-4 (Figure 11) the losses after two hours were 13
percent for the mixture that was intermittently mixed but 40 percent for that
which was undisturbed during the standing
period. The data indicate that the least
possible time should be consumed in damp
mixing p r i o r to compaction and that inter mittent damp mixing of shor! duration,

mixing, but for Soil 2a-4 it increased. (A

s i m i l a r observation will be noted in other
test data involving prolonged mixing r e ported in Table 7 . ) Soil-cement mixtures
made with Soil 2a-4 were stronger than
those made uith Soil 2a-2; although, a s
shown by Table 1, these two soils had
practically the s a m e gradation. However,

the questionof whether the qualiiy of the

mixture i s adversely affected if l e s s than
80 percent of the clay soil i s pulverized.
Wet-dry and freeze-thaw soil-cement
specimens were molded using Soils 4b-6
and 7h, each containing 0 percent, 20 p e r
cent, and 40 percent of lumps retained on
a No. 4 sieve but passing a 1 -inch sieve.

Freeze- Thow Test

Wet- D r y Test

- 13% Moldlng

Molsture Content
No Interrnlttent Mlaing

o - 13% Moldlnq Mo~slureContent

Intermittent Mirlng

M ~ r ~ nTtrneg
Hours
F i r m r e 10.

Effect of l e n ~ t ho f m i x i n g t i m e upon soll-cemnt

losses: Soil 2a-2 plus 8 percent cement.

Soil 2a-2 contained considerably m o r e

organic m a t t e r , and thus its reaction with
cement was l e s s effective than the reaction
of Soil 2a-4. The data in Figures 14 and
15 for mix made of silty Soil 4b and clayey
Soil 6e show that the strengths decreased
with time of mixing. As in the case of the
wet-dry and freeze-thaw tests, prolonged
intermittent mixing was l e s s harmful than
an undisturbed delay.
Additional data showmg the effect of
prolonged mixing a r e presented in S e r i e s
5, Table 7 , and in Series 8, Figures 20
and 21.
Series 4: Effect of Degree of Pulverization
Specifications for the construction of

The molding data f o r the specimens a r e

given in Tables 3 and 4. In one s e t of
specimens (A), a i r - d r y clay lumps w e r e
added to the minus-No. -4 mixture which
was a t AASHO optimum moisture content..
Specimens were molded immediately.
In the second s e t (B), a i r - d r y clay
lumps were added to the minus-No. -4
mixture, which was a l s o air-dry. Water
was then added to the total mix to bring it
to optimum moisture content.
Thus, in Set A the clay lumps had l e s s
opportunity to absorb moisture during the
mixing period than in Set B. In the l a t t e r
case, s o m e of the clay lumps were unintentionally pulverized during mixing. Immediately p r i o r to molding specimens,

TABLE 3
MOLDING DATA FOR SOIL 4b-6, PLUS 144 CEMENT BY VOLUME
AASHO OPTIMUM MOISTURE CONTENT 19. @ , MAXIMUM DENSITY 102.5 PCF.
(SPECIMENS MOLDED ONLY FOR FREEZE-THAW TEST)
I

1
Set i
NO. i Included

B1

Data on Clay Lumps

20
4 0 '

Molsture Content,
When Added
to M u

3 to 4
3 to 4

After
Mmng

13 to l g D (

Molsture Content
of Specimen, %

Minus
No. 4 M u

'

13 to 18al

D e n s ~ t y I Loss
of
1 Due to
Total
Specmen. / 12 cyC.
M u t v r e I pc*.
! F-T.8

:
1

20.1
19.7

I
1

101

101

6
8

a 14% clay L m p s (unpluver~zedsoil retained on No. 4 sieve) after damp m u completed.

22% clay lumps funpulverlzed soil retained on No. 4 s i e v e ) after damp m u completed.
TABLE 4
MOLDING DATA FOR SOIL 7h, PLUS 129 CEMENT BY VOLUME
AASHO OPTIMUM MOISTURE CONTENT 16.88, MAXIMUM DENSITY 108.4 PCF.
(SPECIMENS MOLDED FOR WET-DRY AND FREEZE-THAW l'ESTS)

Set
No.

Data on Clay Lumps

i Moisture Content, %
'
Included / When Added After
1 to Mix
Mixing

___I__-

- Moisture Content

Density

Loss

.J

a 2% clay lumps (unpulverized soils retained on No. 4 s i e v e ) after damp mlxcompleted.

30% clay lumps (unpulverized soils retained on No. 4 sieve) after damp mixcompleted.

moisture content t e s t s were made of the

clay lumps and of the soil-cement mixtures.
As would be expected, the data in Tables
3 and 4 show that the clay lumps in Set A
had gained little moisture by the time they
were compacted into the test specimens.
but the clay lumps in Set B hadgainedconsiderable moisture. The importance of
this moisture factor i s seen in the f r e e z e thaw and wet-dry test data, a l s o presented
in Tables 3 and 4. Here ~t will be noted
that specimens of Set A had l e s s resistance
to alternate freezing and thawing and wet-

ting and drying than Set B, and in some

c a s e s complete failure occurred by disruption of the specimens a s the dry clay
lumps absorbed water and
t h ~ a ~ ~ e A ~ ~ p 6 r When
i 6 dthes .
clay lumps w e r % 7 7 i E p m - ~ \and thus in
a swelled condition a t the timeof inclusion
in the test specimens, the unpulverized
soil had little harmful effect.
These data show that the inclusion of
damp clay lumps i s not particularly harmful and may be permitted in soil-cement
construction in accordance with properly
written specifications. The inclusion of

~~~~~~~g

drv clay lumps, however, is harmful and

should not be permitted. To eliminate the
o s s l b i l i t y of d r v c l a v lumps, clayey soils.
when necessary, can be pre-wetted a short
trme prior to construction.
S e r ~ e s5: Effect of Alr-Entrain~ngCement
In S e n e s 5 a study was conducted to
derermine the comparative performance of
mi u r e s made wlth am-entra~nlng and
no: -air-entra~ningcements. Both cements

ioil
Yo.

Cement , Tvpc
b
by 1 Cementl Mmmg r i m e I Mu!% Time
0-hr
4-hr
0-hr
4-hr
bni.. ?
I

I Cont.

'22.4

i
LA

Lb-5

:4

i
LA

6,-4

12

I
LA

5
4

:O
!O

r e s t s not made

1718)~
JOf16)

26118)
25(16)

:6
13

:.

28

:i

3
-

' F.yures m 0 a r e s o ~ l - c e m e n t l o s s e s f o r

molded
i' a molsrure content 4 percentage polnts wetter than ASTM
~ d l r a t e doptlmum moisture content.

/ 2 0 4 %M o l d ~ n o

501

% 401
I

%-

F r e e z e - Thaw T e s t

'
/

22.5% M o l d i n q
b o ~ s t u r eC o n t e n t
l n t e r m ~ t t e n tM ~ x i n q

I
0

M o s t u r e Content
N o l n t e r m ~ t t e n tM i x m q

I
2

Mixinq T ~ m e Hours

Fieure 11. Effect o f length o f r n i x i q time

upon soll-cement losses: Soil 4h-4 plus 14
percent cement.

were from the s a m e mill. but one (Type IA)

was with vinsol resin interground. Three
soils were testedwith eachof the cements:
Soil 2a-4 plus 8percent of cement, Soil 4b5 plus 14 percent of cement, and Soil 6e-4
plus 12 percent of cement. Test data were
obtained for mixtures that were intermittentlv m x e d for periods of 0, 2, 4, and 6
hours. For the sake of brevity, only the
data for the 0-hour-and 4-hour soil-cements
a r e presented.
Moisture-density data, presented in
Table 5. were obtained for the various m u tures usmg the standard AASHO method.
The maxlmum densities andoptimum moisture contents for each group were p r a c t ~ c a l l vthe same regardless of the type of

specimens

cement, and the maximum densities generally decreased and the optimum moisture
contents increased a s the length of mixing
time increased, which was aiso observed
previously in S e r i e s 3.
Data from wet-dry and freeze-thaw t e s t s
a r e given in Table 6. Here it will be noted
that there i s relatively little difference in
test data for the two cements: if there is
any small advantage, it is in favor of Type
IA.
Compressive strength results a r e given
in Table 7 . These data show only minor
differences between the strengths obtained
with Type I and Type IA cements.
Although these limited tests show little
difference in the performance of mixtures
made with Type I and Type IA cements,
unusual effects have occasionally been
noted in the laboratory with different s o i l s
and different cements. On this basis, it
appears advisable to conduct the l a b
oratory tests with the same type of cement
that will be used in construction.
S e r ~ e s6: Effect of Cement Content
The tests in S e r i e s 6 were made to investigate the effect of cement content on
compressive strength and wet-dry and
TABLE 7
COMPRESSWE STRENGTHS OF SOIL-CEMLNTS
CCNTAININC T Y P E I AND T Y P E LA CEMENTS
Approx.
SOI~

YO.

Cement
C,,"t. by

'

Type.

C u m p r e s s w e S t r e n e h s , PSI.
Mumg T m e . Hours
Zero
Four

.-- -

freeze-thaw r e s i s t a n c e of soil-cement.
Moisture-density relations w e r e e s t a b ished f o r m i x t u r e s of Soils 2a-6, 4d, and
h w ~ t hcement contents of approximately
8, 12, 16, 22, and 28 percent. F r o m these
c u r v e s , shown in p a r t in Figure 16, the
optimum m o i s t u r e content and maximum
density w e r e obtained by interpolation and
extrapolation f o r molding wet-dry, f r e e z e thaw, and compressive-strength s p e c i m e n s

and drying and f r e e z i n g and thawing, a n d

a s shown in Table 8 , even the I - p e r c e n t cement s p e c i m e n s gave f a i r performance
f o r 96 cycles. All specimens containing
12 percent o r m o r e of c e m e n t w e r e e x cellent, showing only a s m a l l l o s s a f t e r
96 c y c l e s of test.
T h e data in Table 9 show that m i x t u r e s
made of s i l t y Soil 4d improved in quality
as the c e m e n t content i n c r e a s e d t o 3 0 p e r -

Freaze- Thou Test

-19%

Wo

- Dry Tart

Molding Moist. Cont.

NO lntermtttent Mixing

0-21%

Molding Moiat. Cont.

Intermittent MI XI^)

M i r ~ n qTime

F i r v r e 12.

- Hwn

E f f e c t o f l e n g t h o f mixing time upon s o i l - c e m e n t

iosses: Soil 6e-2 p l u s 12 p e r c e n t cement.

containing 6 to 34 percent of c e m e n t , a s
indicated in the tables and c h a r t s .
In this s e r l e s , the wet-dry and f r e e z e thaw s p e c i m e n s w e r e tested through 96
cvcles. T h e data in Tables 8. 9 . and 10
show that m i x t u r e s containmg relatively
high c e m e n t contents had g r e a t r e s i s t a n c e
t o a l t e r n a t e wetting and drying and f r e e z ing and thawing. After 96 cycles many of
the t e s t s p e c i m e n s had practically no l o s s
of material.
Soil-cement m i x t u r e s made of sand Soil
2a-6 w e r e particularly resistant to wetting

cent. Specimens containing 18 to 22 p e r cent o r m o r e of cement showed good

performance in the freeze-thaw and wetd r y t e s t s f o r the full 96 cycles. In the
wet-dry t e s t s e v e r a l s p e c i m e n s split a t
the compaction planes (specimens w e r e
molded in t h r e e l a y e r s ) and then c r a c k e d
vertically into pieces which w e r e hard and
durable. P i e c e s l e s s than '4 inch in s i z e
w e r e d i s c a r d e d and included In the m a t e r i a l loss. The significance of this c r a c k ing h a s not been established; however,
field p e r f o r m a n c e of m i x t u r e s that c r a c k

TABLE 8
Soll 21-6
SOIL-CEMENT WSSES % OF ORIGINAL WEIGHT

Soil 2a-6 showed a rather c o n s ~ s t e n t

increase in strength a t all ages a s the
cement content increased froin 6 percent
to 30 percent. The other two soils showed
a s i m i l a r relationship to age 120 days,
with some inconsistency In strength gain
from 120 to 365 days. Maximum strengths
obtained were 4,700 psi. for Soil 2a-6,
3,100 psi. for 4d, and 2,300 psi. for 7h.
Series 7: Effect of High-Early-Strength
m y p e Im Cement.
Compressive strength tests were made
to study the effect of Type 111 cement in
soil-cement mixtures. Specimens of sandy
Soil 2a-6 and silty Soil 4d, each with 6 ,
10, and 14 percent of cement, were molded a t AASHO optimum moisture content and

F i p u r e 1 3 . E f f e c t o f l e n p t h of m i x i n a t i m e
upon 2 8 - d a v c o m p r e s s i v e s t r e n p t h :
Soil 2a
p l u s 8 p e r c e n t cement.

in the laboratory wet-dry test has been

satisfactory.
The data in Table 10 show that mixtures
made of Soil 7h improved in quality a s the
cement increased up to 30 percent. Specimens containing 18 percent cement o r
more were good up to the 84th cycles.
With this soil there was a sharp increase
in quality a s the cement content increased
from 14 percent to 18 percent, particularly
in the freeze-thaw test. As in the c a s e of
wet-dry specimens made of Soil 4d. Soil
7h specimens cracked in the wet-dry test
but to a l e s s e r extent.
The compressive-strength data for the
various mixes a r e shown in Figures 17.
1 8 . and 19. As might be expected, the
specimens containing sandy Soil 2a-6 were
considerably stronger than those containing silty Soil 4d, which were stronger than
specimens containing clayey Soil 7h.

2 200
I

No Intermittent Mixing

M ~ x ~ nT q
~ m e -Hours
Flpure 1 4 .
E f f e c t 0 1 l e n g t h o f rnixinp t i m e
Sol1 4b
dpon 3 - d a v c o m p r e s s ~ v d s t r e n u t h .
p l u s 1 4 p e r c e n t cement.

TABLE 9
S o d 4d
SOIL-CEMENT M S S E S % O F ORIGINAL WEIGHT

'

Wet-Dry T e s t

o.

Freeze-Thaw T e s t

Cement Contenl b V o l . ,
8 10 12 14 I 8 2 2 26 30 1 8 TO 12 I? I 8 2
12
24
36
48
60
72
M
96

29
53
67
72
75
80
87

25
43
54
63

21
33
48
56
70 62
76 67

8J 78

2
3
27
35
46
41
52
60

2
4

11
18
22
38
47
56

1
2
7
8
9
12
17
20

21
58
70
82
92
5 4 100
7 4 100
9 5 1M)
1 1
1 1
1 1
3 2
4 3

7 3 2 2
24 6 4 2
33 9 9 2
53 16 14 3
6 6 21 18 5
74 4 0 2 0 5
aa 4 s 21 s
9 5 55 24 5

m
2 1 1
2 1 1
2 2 2
3 3 2
5 4 3
5 4 4
3 4 4
5 4 4

maximum density. To determine whether

prolonged intermittent m ~ x i n gwould have
an unusual effect on soil-cement mixtures
containing Type In cement, one set of
specimens was molded after a 4-hour
mixing time. (The same procedures were
used a s inseries 3.) The moisture-density
relations in Table 11 show that the optimum moisture content and the maximum
density for the mixtures containing Type I
o r Type In cement a r e practically the
same. As was established in Series 3, a
prolonged mixing period of 4 hours r e sulted in a general decrease in maximum
density and an increase in optimum moisture content.
Compressive strength data f o r specimens broken a t ages of 1, 2, 3, 4, 6 , 7 ,

22% Moldinq Moist. Con).

Intermittent Miring
Soil 60- 4

~ $ 4 0 0 1

5-

,"

F i g u r e 16. Moisture-densltv r e l a t i o n s of
soil-cen.ent mixtures used i n S e r i e s 6.

10, 14, 28, and 60 days a r e shown in Figures 20 and 21. F o r both soil types and
both mixing procedures the early-age
strengths were consistently greater for
Type Ill than for Type I cement, and in
nearly all cases the 60-day strengths were
also greater for Type IlI.
The prolonged intermittent mixing was
not seriously detrimental with either of the
cement types, although it resulted in some
loss of strength.

21% Molding Moist. Cont.

Intermittent Mixing

2 200
0

Figure 15.

Soil 60 2

E f f e c t or lengrn o r mlxlng time

Series 8 and 9: Cement-Modified Soils

Most clayey soils a r e volumetrically
unstable, for they shrink when dried and
expand when wetted; furthermore, their
strength characteristics a r e unusually
sensitive to changes in moisture content.
Stabilization of these soils i s an important field, and portland cement in quantities less than required for regular soilcement mixtures has been used to reduce
the extent to which the soils shrink, swell,
and lose strength. The material thus produced is referred to a s cement-modified
soil. This type of soil stabilization i s

Cycles 1

Cement Content

8 10 12 14 18 22 26 3 0

Vol. , S
10 12 14 18 22 26 30

a r e associated with s o i l s that show high

shrinkage, a n d v i c e v e r s a . T h u s by d e t e r mining these s o i l constants for c e m e n t modified s o i l s , s o m e m e a s u r e of the
effectiveness of the c e m e n t in modifying
the undesirable plasticity c h a r a c t e r i s t i c s
of the soil mav be obtained. With s o m e
s o i l s , bearing-ratio t e s t s w e r e m a d e to
evaluate f u r t h e r the effect of the cement.
S e r i e s 8: Cement-Modified Fine-Grain
Soils
In this s e r i e s t h r e e cement-modified
clavev s o i l s w e r e studied to d e t e r m i n e the
influence of various c e m e n t contents in
a l t e r i n g the p r o p e r t i e s of the soils. S a m p l e s to d e t e r m i n e test constants a n d g r a i n
s i z e w e r e p r e p a r e d by compacting, a t optimum m o i s t u r e content, sufficient s o i l
and cement mixture to fill a t h i r d of a
s t a n d a r d 4-by-4.6-inch specimen mold.
ma 2

- 7 . .

. .. ..
. .

day comor dr

3 28-

c31m-.
300C

354%.

= zsoc

Camont Content by VOI

,Z

F ~ p r e18. E f f e c t o f cement c o n t e n t ana

a g e on c o m p r e s s i v e s t r e n g t h : S o i l 4d.

rABLE 11
MAXlMl M DEYSITY AND OPTIMUM MOISTURE COYTENTS
OR SOIL-CEMENTS O F SClLS 2d 6 AND 4d

--

jell

-all

uo.

with approx. 1
OY < emeot

23-b

aotl 4d wrth approx.

.45 cement

T h e s e specimens w e r e permitted to hyd r a t e a t 100 percent relative humidity f o r

seven davs: then p a r t of the m a t e r i a l was
p u l v e r ~ z e d to p a s s a No. 10 s i e v e f o r a
h..rdrometer analvsis. and p a r t was pulverized to p a s s a No. 40 s i e v e f o r determming
:est constants.
In one c a s e . ro s e e if repeated freezing
and thawingaffected the permanence of the
cement influence, a set of s p e c i m e n s w a s
subjected to 60 cvcles of a l t e r n a t e freezing
and thawing (followingAASH0 T 1361. T h e
m a t e r i a l s w e r e then pulverized to p a s s the
s l e v e s noted above.
T h e data in Tables 12 and 13 show that
cement effectivelv r e d u c e s the plasticitv
index and i n c r e a s e s the shrinkage limit of
clavev soils. The s m a l l differences b e :ween the t e s t constants before and a f t e r
60 cvcles of freezing and thawing indicate
that the effectiveness of the cement i s not
readilv destroyed. T h e gradation a n a l y s i s
f the cement-modified s o i l s shows that
he percentage of c l a y - s i z e p a r t i c l e s i s
educed by the cement action.
TABLE 12
TEST COSSTANTS AEiD GRADATION C F
CEMENT-MODIFIED SClL
Sot1 b q . Clay Sall from r e x a s

7 Cement by Volume
-1 Cement hv Y e q h t

0
i)

2
2. 16

4
4. 37

6
5.61

8 10
8. 90 11. 24

Liquld L:mnl

54

53

52

50

48

47

i J i 3 s t ~ cLlmll

21

21

24

31

33

35

P l a s t ~ c ~ lindex
v

33

34

:9

!5

I2

s h r ~ n k ~ qLlmtt
p

18

23

2b

29

31

33

5111.
10.05-d. 005rnm.,

Additional data showing the effect of

c e m e n t in mcdifving a s i l t y clay loam s o i l
a r e given i n T a b l e 14. In t h i s c a s e a b e a r ing-ratio t e s t was used a s a n a i d in m e a s uring the beneficial effect of the cement.
T h i s t e s t is the s a m e a s the California
Bearing Ratio (CBR) test3, but h a s been
designated simply a s a bearing-ratio t e s t
because a t the high b e a r i n g values of s o m e
of the t r e a t e d s o i l s the r a t i o d o e s not have
quite the s a m e significance a s the CBR.
T h e t e s t i s made by forcing a piston having
a n end a r e a of 3 sq. inches into the top
s u r f a c e of a specimen confined in a s t e e l
mold 6 inches in d i a m e t e r and 4.6 inches
high. The cement-modified s o i l i s c o m pacted in the mold in t h r e e l a y e r s with the
s a m e compactive effort used f o r compacting soil-cement m i x t u r e s according to
AASHO s t a n d a r d T 134. T h i s r e q u i r e s 56
blows of the 5.5-lb. r a m m e r on e a c h of the
t h r e e layers. In t h e s e t e s t s , the s o i l and
c e m e n t m i x t u r e s w e r e hvdrated in the
molds f o r seven d a v s a t 100 percent r e l a tive humiditv and then i m m e r s e d in w a t e r
f o r four d a v s b e f o r e testing. The load in
pounds p e r s q u a r e inch to f o r c e the piston
into the specimen to a penetration of 0. 1
'The Calllornla B e a n n g Ratlo Test, reported by 0 J P o r t e r
rs based O M
191 and by W H J e r v l s and Joseph B E u s t ~ s
the observatron that a f o r c e of 1, 000 p s ~IS requlred to push
a I 95-mch-diameter plston mto a crusher-run, hngh-quallty
base-course materml lo a depth of 0 1 ~ n c h The relatlve
strength or C a I J o r n ~ a Bearmg Ratlo of other base course
rnaterlals IS obtained bv determlnrng the penctratlon r e s l s t ance in pounds per s q u a r e anch a t 0 1 mch and bv dlvidlng
ihle umt load by I . 000 and mulctplymg by 100 to obrarn a perc e n u g e value Thus a m a t e r n l having a p e n e t n l l o n resistance c4 MO psl h a s a CBR of 50

1c),

40

37

33

:9

31

24

48

46

36

29

20

14

Cia"
less than 0.005mm. 1

G.m*nt Conl~nlby Vol

Figure 19. E f f e c t o f cement c o n t e n t and

age on compressive s t r - n g t h :
S o i l 7h.

S e c l m e n s compacted to appraxlrnatelv 92 pel. a t 26 percent

water, and hydrated 7 days, then pulverlzedlor testconsfanls
and gratn stre.

10

20

3 0 x 0

10

20

30

A -

'60

Age Ooyr
Fipure 20. h p a r i s o n o f compressive s t r e n g t h s obtalned with normai (Tvpe 1) cement and hi&-earlv-strength (Tvpe 3) cement: Soil
2a-6.

inch, divided by 10, i s referred to a s the

comparative-bearing ratio.
The data in Table 14 indicate that the
addition of cement (up to 9 percent) to this
fine-grain soil changed its plasticitycharacteristics by reducing the P. I. from 14
to 0 and greatly increased its strength a s
indicated by an increase in bearing-ratio
values from 2 to 138.
Series 9: Cement-Modified Granular Soils
Relatively small quantities of cement
added to granular soils not naturally suitable for base construction will increase
their all-round stability and strength.
Series 9 i s concerned with this field of
cement stabilization.
From the data previously discussed
on cement-modified fine-grain soils, it
appeared that the bearing-ratio test was
fairly effective in showing the beneficial
influence of the addition of cement to plastic soils which soften when they become
wet. Thus this test was used in studying
the effect of additions of relatively small
quantities of cement to substandard granular materials. In addition, tests on the
specimens a t different a g e s and after v a r ious cycles of alternate freezing and thawing were made with the soniscope, a
device (11) which measures the velocity
with which shock pulses travel through

the material. Normally, the velocity inc r e a s e s a s the strength of the material
increases and decreases if deterioration
occurs due to frost action o r some other
cause.
Three granular soils, containing various percentages of s i l t and clay a s listed
in Table 15, were used in these tests. Soil
2b was a pit-run sand and gravel contain-

Soils 2c and 2dwere gravellv plastic soils,

a i r l v good base materials except that they
ontainedcons~derableplasticciav (partlcularlv Soil 2d) and were subject to softening when thev became wet.
In a first group of tests, relatively low
percentages of cement were added to Soils

2c and 2d, and test constants were deterrnlned for the modified materials after a
hvdration period of 2 davs. The data p r e sented in Table 15 show that the addition
of cement greatly reduced the plasticity of
the solls. T e s t s of this type were not
made on Soil 2b. since it was nonplastic in
its natural state.
More important, however. than the r e duction in plasticity of the soils, is the
effect of the cement in increasing their
strength. This was investigated by the
bearing-ratio test on specimens containing
various percentages of cement. Moisturedensity t e s t s were made using both the 4inch-diameter mold (according to AASHO
T 134 except that the material retained on
a No. 4 sieve was included in the sample)
and the 6-inch-diameter mold using equivalent compactive effort. Data obtained
with both molds a r e given in Table 16.
Specimens 6 inches in diameter and 4 . 6
inches high f o r the bearing-ratio test were
molded of mixtures of each of the three
s o i l s containing l x , 3, 4%, 6 , and 10
percent of cement by weight. These s a m e
specimens were used for soniscope tests.
In addition, standard AASHO specimens
( 4 inches in diameter) were molded for the
AASHO wet-dry and freeze-thaw tests.
The Specimens for the bearing-ratio

TABLE I4

rFsT C O N S ~ A N T S *A N D BEARING R A T I O S ~OF

ZEMENT-MODIFIED SOIL
- o $ l 4e

Cement bv Volume

7 Cement bv Welpht

L I ~ U ~
L ~.

m ~ t

3
2.96

5 . 97

3.02

:B

3b

34

34

29

:3

Plast~c~ , m , t

~ ~ a s t r c l tIndex
y

14

Shrlnuaqe ~~~~t

20

21

26

Coarse Sand. ?
2.0-0. 25mm. I

Fme Sand. 4
0.25-0.05mm. 1

is

15

26

28

Silt. 7
0.05-0.005mm.,

57

z:av.

''

jqd

yoa

j6d

66

less than 0. OOSmm. 1

Bearing Ratlo

N. P.=
%. p.
29

42

Bearlnq Ratlo Speclmens

wo~stureContent. 7

16

16

16

Dens~ty,pcf

38

95

,, ,,-

16

'Test constants and gram slze studles made on cemenl~ o d l I l e ds o ~ khydrated seven days, then pulverized
B e a r ~ n qratlo tests made after seven days hydnltan and
four days lrnmerslon In water.
C
~plastic
~
t
d ~ ~and
l t clav combmed

TABLE 15
TEST CONSTANTS AND GRADATION OF GRANULAR SOILS AND TEST CONSTANTS
OF CEMENT-MODIFIED SOILS

1I

o i l No.

-7Cem.

em.

by Vol.
by WI.

Liquid Limit

II

N . P . ~ 23
N. P.

Plastic Limit
Plasticity Index
Plus No. 4 Gravel

25

Coarse Sand, %
(No. 4-0.25mm. )

67

Fine Sand, %
(0.25-0.05mm. )
Sllt, q
(0.05-0.OOSmm.)

Clay,
( l e s s than 0.005)

54

II

'

I -

29

29

28

34

34

32

33

22

24

28

28

29

32

13
I
1 / 1 5

25

5.8

14.5

111

1.9 3.9 5. 5 7 . 3
1.513
4.516

7.7
6

1.9 1 3 . 9
1.5
3
30
29

/ 15

1
i

a Not Plastic

2d

2c

2b

a
I

- Type I Cement
---

Type

ID

Cement

Age D O Y S
Fieure 21. Comparison o f compressive s t r e n g t h s o b t a l n e d with normal (Tvpe 1 ) cement and h l g h - e a r l y - s t r e n e t h (Tvpe 3 ) cement: S o i l
4d.

and soniscope tests were retained in the

molds and placed in the moist room.
Several specimens of each cement content
were molded s o that they could be tested
after various ages in the moist room and
r A U L E 18
UOISTURE-DENSITY RELATIONSHIPS

126 I I

:d

'D
l'o

123 1 2

I 122

12.5 i127 1 0 . 5 1126 1 1 1124

Urxfimum Denslty. pcf.

o p t m u m Molsture Content.

'i

11.5

also after various cycles of alternate

freezing and thawing. Thus, the t e s t s
permitted a study of the effect of cement
content and time of curing on the bearing
ratio and pulse velocity and, also, a study
of the deteriorating effects of freezing and
thawing. The bearing-ratio and soniscope
specimens for alternate freezingand thaw-

a t 73 F. During the test, the specimens

were placed on their s ~ d e s .
The cement contents used in this s e r i e s

TABLE 17
BEARING RATIOS AT DIFFERENT AGES AND AFTER FREEZING AND THA.WNG
Soil 2b
-

% Cement
by Wt.

As Molded
Unsoaked

After Moist Room Cure, Days

7
21
37
87
161

After F-T, Cycles

12

48

100

TABLE 16
overlap those required to produce soilPL'LSE VELOCITIES A T DIFFERENT ACES AND
cement mlx according to the c r i t e r i a a c AFTER FREEZING AND THAWING
5011 2b
epted for use with AASHO methods T 135
nd T 136. thewet-dry test and the freezeVeloclt Throu h S e c r m e n - 1 0 0 It/sec
; ~ ~ ~ n ' i ~ Mf ots te Rr d m Cur: D G s \ A l t e r F - T . C y c l e s
thaw test.
a
7 21
37
53
i l 87 1 12 24
?b
48
Based upon these tests, the following
minimum cement quantities a r e reqtiired
with the soils to produce "soil-cement":
4 1 2
96 105 115 115 120 118 I M l M 105 104
Soil 2b, 3 percent by weight (4.2 percent
6
107 116 120 124 1?5 125 113 I l b 114 113
by volumei; Soil 2c, 4.5 percent by we~ght
10
119 I27 129 130 136 135 123 128 129 I28
(5.8 percent by volume); and Soil 2d, 6
percent bv weight (7.3 percent by volume).
These cement factors may be kept in mind
161 days. Other mixtures had ratios conto differentiate between so-called cementsiderably over 1,000. In making some of
t h e w tests, the loading piston was forced
modified soil and soil-cement.
into the specimen to a depth of only 0.05
The bearing-ratio and pulse-velocity
o r 0.075 inch. Values in this high range
data a r e presented in Tables 17 to 22. Inhave little significance, other than a s
cluded a r e the bearing-ratio values for
indicators of relative hardness and r e s i s t specimens after periods of 7, 21, 37, 87,
ance to penetration. Freezing and thawand 161 days in the moist room and after
ing reduced the bearing ratios f o r the
12. 48. and 100 cycles of alternate freezemixture containing 1% percent of cement;
ing and thawing. The moist room specbut mixtures containing 3 percent o r more
imens at ages 37, 87, and 161 days a r e
of cement showed no deterioration during
the same age a s the freeze-thaw specthe freeze-thaw t e s t , although their rateof
imens after 12, 48. and 100 cycles of test.
strength gain was l e s s than that of specSoniscope (pulse-velocity) values were
imens continously moist cured.
obtained for specimens a f t e r curing for
The soniscope data for Soil 2b a r e p r e periods up to 87 days in the moist room
sented in Table 18. The pulse velocities
and after a s many a s 48 cycles of freezing
did not d e c r e a s e significantly during the
and thawing. Some of the specimens with
freeze-thaw test, again showing good r e low cement contents deteriorated fairly
sistance to deterioration. As might be
rapidlv, and soniscope readings could not
expected, the pulse velocities f o r soilbe made during the freeze-thaw test: in
cement mixtures of Soil 2b increased with
several c a s e s the 100-cycle bearing-ratio
increased cement content and with time of
tests could not be made.
moist curing. Velocities of more than
Table 17 gives the bearing-ratio data
10,000 ft. p e r sec. were common.
Soil 2b. Here it will be seen that the
Bearing-ratio and soniscope data for
ent-treated soil mixtures, even with
Soil 2c a r e presented in Tables 19 and 20.
tively low cement contents, had very
This soil required higher cement contents
high values. For instance, the mixture
than Soil 2b to achieve the s a m e degree of
containing 3 percent of cement had a bearhardness a s indicated by bearing-ratio
ing ratio of 471 a t age 7 days, and 806 a t
TABLE 19
BEARING RATIOS AT DIFFERENT AGES AND AFTER FREEZING AND THAWING
Soil 2c

5 Cement
by Wt.
0

As Molded
Unsoaked
11

After Moist Room Cure, Days

7
21
37
87
161

After F-T,Cycles
12
48
100
3

TABLE 20
PULSE VELOCITIES AT DIFFERENT AGES AND
AFTER FREEZING AND THAWING
Soil 2c

'svwr

Veloclt Throu h eclmen 100 11 s e c

C e m e n t ! ~ l t e r Moist l i c k Cu:.
Z y s iAfter F-T./Cycles
b

21

31

53

I1

87

12

24

36

48

and soniscope tests. The data show, however, that cement quantities of 4% percent
by werght or more produced soil-cement
mixtures that had good resistance to
freezing and thawing. For instance, the
4%-percentmixture showed a bearing-ratio
of 323 a f t e r 12 cycles of test and 358 after
100 cycles, and pulse velocities for this
s a m e mixture were 7,400 it. p e r sec.

f r o m the slightly g r e a t e r cement contents

required to produce soil-cement mixtures.
Both the bearing-ratio and the soniscope
t e s t s appeared to be effective in measuring
the relative deteriorating effects of a l t e r nate freezing and thawing.
SUMMARY AND CONCLUSIONS
The data presented in this report a r e
intended to be helpful to engineers in
developing an understanding of many of
the physical properties of soil-cement
mixtures. In some instances, more mixt u r e s have been tested under each s e r i e s
than were reported in this paper. The

TABLE 21
BEARING RATIOS AT DIFFERENT AGES AND AFTER FREEZING AND THAWING
Soil 2d
I

by Wt.
Cement

As Molded
Unsoaked

After Moist Room Cure, Days

7
21
37
87
161

1
1

After F-T, Cycles

12
48
100

m a t e r ~ a l sreduced t h e ~ pr l a s t l c ~ t yand inc r e a s e d their strength. The addition of

slightlv g r e a t e r q u a n t ~ t ~ eofs cement ( p r o d u c ~ n gs o ~ l - c e m e n t )materlallv ~ n c r e a s e d
d r ~ n t r t hahnvc. those nf the cement-modlfied s b i ~ sand resulted in m a t e r i a l s having

considerable r e s ~ s t a n c e to d e t e r ~ o r a t ~ o n
from a l t e r n a t e freezlng and thawing. In
t h ~ sstudv. b e a r ~ n g - r a t l o and sonlscop
t e s t s w e r e eifectlve In measuring the p r o
trresslve d e t e r ~ o r a t ~ oofn sDeclmens In the
freezing-and-thawing test.

The data reported w e r e accumulated a t

the Portland Cement Association Laborat j r i e s o v e r a p e r i o d of m o r e than 1 5 y e a r s .
Consequently, the author i s indebted to a
number of engineers who a s s i s t e d in conducting the test. Much of the work was done
by E. G. Robbins and M. S. Abrams. Miles

D. Catton. now a s s i s t a n t to the vice p r e s i dent f o r r e s e a r c h and development. w a s in

administrative c h a r g e of most of t h e r e s e a r c h and development work on s o i k e m e n t
mix. The paper w a s reviewed by Douglas
McHenry, d i r e c t o r of development, whose
comments a r e gratefully acknowledged.

-- 1

References
1. Catton, Miles D., "Research on the
Physical Relations of Soil and Soil-Cement
Mixtures," Highway R e s e a r c h Board P r o ceedings, Twentieth Annual Meeting, 1940.
2. Winterkorn. Hans F., Gibbs, Harold
J., and Fehrman, Rollie G., "Surface C h e m ical Factorsof Importance in the Hardening
of Soils by Means of .Portland Cement, "
Highway R e s e a r c h Board Proceedings,
Twenty -Second Annual Meeting, 1942.
3. Catton, Miles D., and Felt, E a r l J.
"Effect of Soil and Calcium Chloride Adm i x t u r e s onsoil-Cement Mixtures, " Highway R e s e a r c h Board Proceedings, Twenty
T h i r d Annual Meeting, 1943.
4. P o r t l a n d c e m e n t Association, "SoilCement Mixtures - Laboratory Handbook. "
5. Hicks, L. D., " P r o g r e s s in SoilCement Construction," Highway R e s e a r c h
Board Proceedings, Nineteenth Annual
Meeting, 1939; a l s o , "Soil-Cement Design in
North Carolina. " Highway R e s e a r c h B o a r d
Proceedings, Twenty -Second Annual Meet
ing 1942.
6. Felt. E a r l J . , "Soil S e r i e s N a m e s
a s a B a s i s f o r Interpretive SoilClassifications f o r Engineering Purposes," A m e r i -

can Society for Testing Materials, Symposium on the Identification and Classification of Soils, Special Technical Publication
No. 113, 1951.
7. Catton, Miles D. , "Soil-Cement: A
Construction Material, " Proceedings of
Conference on Soil Stabilization, Mass a c h u s e t t s Institute of Technology, J u n e
1952.
8. P o r t l a n d c e m e n t Association, "SoilCement Roads, Construction Handbook. "
9. P o r t e r , 0. J . , "The P r e p a r a t i o n of
Subgrades, " Highway R e s e a r c h Board P r o ceedings, Eighteenth AnnualMeeting, 1938;
and "Development of Original Method f o r
J a n u a r y 1949.
10. J e r v i s , W. H. and E u s t i s , J. B.,
"Accepted P r o c e d u r e f o r the CBR Test.,"