HIGHWAY R E S E A R C H BOARD

Bulletin 335

Lime Stahilization Mix Design, Properties

and Process

1962

NOV 19 1962

ational Academy of Sciences—

National Research Council publication 101 !i

HIGHWAY RESEARCH BOARD Officers and Members of the Executive Committee

1962

OFFICERS R . R . B A R T E L S M E Y E R , Chairman C. D . CURTISS, First Vice Chairman

W I L B U R S . S M I T H , Second Vice Chairman F R E D BURGGRAF, Director W I L L I A M N . C A R E Y , J R . , Assistant Director

Executive Committee R E X M. WHITTON, Federal Highway Administrator, Bureau of Public Roads (ex ofHcio) A. E . JOHNSON, Executive Secretary, American Association of State Highway Officials

(ex officio) LOUIS JORDAN, Executive Secretary, Division of Engineering and Industrial Research,

National Research Council (ex officio) P Y K E JOHNSON, Retired (ex officio, Past Chairman 1960) W. A. BuGGE, Director of Highways, Washington Department of Highways (ex officio,

Past Chairman 1961) R. R . BAKTELSMEYEM, Chief Highway Engineer, Illinois Division of Highways E . W . BAUMAN, Director, National Slag Association, Washington, D. C. DONALD S. BERRY, Professor of Civil Engineering, Northwestern University MASON A. BUTCHER, County Manager, Montgomery County, Md. 3. DOUGLAS CARROLL, JR., Director, Chicago Area Transportation Study C. D. CURTISS, Special Assistant to the Executive Vice President, American Road

Builders' Association HARMER E . DAVIS, Director, Institute of Transportation and Traffic Engineering, Uni-

versity of California D U K E W . DUNBAR, Attorney General of Colorado MICHAEL FERENCE, JR. , Executive Director, Scientific Laboratory, Ford Motor Company D. C . GREER, Stote Highway Engineer, Texas State Highway Department JOHN T . HOWARD, Head, Department of City and Regional Planning, Massachusetts

Institute of Technology BURTON W . MAKSH, Director, Traffic Engineering and Safety Department, American

Automobile Association OSCAR T . MARZKE, Vice President, Fundamental Research, U. S. Steel Corporation J . B. MCMORRAN, Superintendent of Public Works, New York State Department of

Public Works CLIFFORD F . RASSWEILER, Vice President for Research and Development, Johns-Manville

Corporation G L E N N C . RICHARDS, Commissioner, Detroit Department of Public Works C. H . SCHOLER, Applied Mechanics Department, Kansas State University WILBUR S. SMITH, Wilbur Smith and Associates, New Haven, Conn. K . B. WOODS, Head, School of Civil Engineering, and Director, Joint Highway Research

Project, Purdue University

Editorial Staff <

F R E D BURGGRAF H E R B E R T P . ORLAND 2101 Constitution Avenue Washington 25, D . C.

The opinions and conclusions expressed in this publication are those of the authors and not necessarily those of the Highway Research Board

Lime Stabilization Mix Design, Properties

and Process

1962

/VRe HIGHWAY R E S E A R C H BOARD Bulletin 335

, Lime Stabilization Mix Design, Properties

and Process

1962

Presented at the 41st ANNUAL MEETING

January 8-12, 1962

National Academy of Sciences-1-̂ ^ • ' - l ^ National Research Council

Washington, D.C. 1962

ho 5 3 6 '

Department of Soils, Geology and Foundations Miles S. Kersten, Chairman Professor of Civil Engineering

University of Minnesota, Minneapolis

COMMITTEE ON LIME AND LIME-FLY ASH STABILIZATION

Chester McDowell, Chairman Supervising Soils Engineer

Texas Highway Department, Austin Conard M. Kelley, Highway Engineer, National Lime Association, Garland, Texas James A. Kelley, H^hway Research Engineer, Physical Research Division, U. S.

Bureau of Public Roads, Washington, D.C. O. L. Lund, Assistant Materials and Testing Engineer, Highway Testing Laboratory,

Nebraska Department of Roads, Lincoln R. L. Peyton, Assistant State Highway Engineer, State Highway Commission of

Kansas, Topeka Willis H. Taylor, Jr., Bituminous Construction Engineer, Louisiana Department of

Highways, Baton Rouge Ernest Zube, Supervising Materials and Research Engineer, California Division of

Highways, Sacramento

Contents FATIGUE BEHAVIOR OF A LIME-FLY ASH-AGGREGATE

MIXTURE — Harold L. Ahlberg and WUliam W. McVinnie 1

EFFECTS OF LIME ON PLASTICITY AND COMPRESSIVE STRENGTH OF REPRESENTATIVE IOWA SOILS

Paul E. Pietsch and Donald T. Davidson 11

FORMATION OF NEW MINERALS WITH LIME STABILIZATION AS PROVEN BY FIELD EXPERIMENTS IN VIRGINIA James L. Fades, F. P. Nichols, Jr., and Ralph E. Grim 31

LIME AND FLY ASH PROPORTIONS IN SOIL, LIME AND FLY ASH MIXTURES, AND SOME ASPECTS OF SOIL LIME STABILIZATION

Manuel Mateos and Donald T. Davidson 40

COMPARISON OF VARIOUS COMMERCIAL LIMES FOR SOIL STABILIZATION

Jerry Wen-Hann Wang, Donald T. Davidson, Elmer A. Rosauer, and Manuel Mateos 65

Fatigue Behavior of a Lime-Fly Ash-Aggregate Mixture HAROLD L. AHLBERG and WILLIAM W. McVINNIE, Research Associate and Re-search Assistant, Respectively, Civil Engineering Department, University of Illinois

The use of lime-fly ash-aggregate mixtures for highway base courses is a relatively recent application to the expanding highway program. Rational methods of analysis for the behavior and design of pavements built with lime-fly ash-aggregate base courses are extremely limited. As a prelude to the development of sound rational methods, it is neces-sary to understand the behavior of the material when it is subjected to repeated loading.

This paper reports on the fatigue behavior of a lime-fly ash-aggre-gate mixture. Specimens, 6- by 6- by 38-in., were tested after a 28-day curmg period by subjectmg them to constant, repeated flexural loads. The relationship between the cycles required to produce failure and the ratio of the applied stress to the static strength is shown.

Repeated f lexural loads applied to the specimens tested caused the material to fracture as a rigid-type structure at stress levels con-siderably below the static strength of the specimens. The test results indicate that the concept of fatugue should be considered in develop-ment of a rational method of analysis for the behavior and design of pavements m which lime-fly ash-aggregate mixtures are used.

• A QUALITY lime-fly ash-aggregate mixture is a composite of relatively inert materials cemented together by the combination of lime, fly ash, and water. Fly ash is a pozzolanic material that when combined with lime in the presence of water will form a cementitious material. After mixing the proper proportions, the material is in a moist, nonplastic state, but it can be readily compacted to form a dense mass. Following a curmg period, the compacted material exhibits rigid, concrete-like prop-erties (1).

Lime-fly ash-aggregate mixtures have been investigated by many agencies and have been found by field experience to produce a satisfactory base course material for high-way pavements. However, a rational structural design method for the use of lime-fly ash-aggregate is not available.

A highway pavement is subjected to repeated loads of varying magmtudes and fre-quencies. Recognizmg this type of loading, it seems only reasonable that a knowledge of the "fatigue" behavior of the material should be understood and evaluated in any rational design procedure. The term "fatigue" refers to progressive fracture of a material caused by repeated cycles of stress at amplitudes of stress smaller than the ultimate strength of the material (2).

The primary purpose of this research is to provide a basic concept of the fatigue properties of lime-fly ash-aggregate mixtures. With no previous fatigue data avail-able, the scope of this study is limited to constant, repeated flexural loads. The authors feel this type of test will give a basic concept of the fatigue properties and pave the way for further and more extensive studies in this field.

PROCEDURES The specimens used to evaluate the fatigue behavior of the lime-fly ash-aggregate

mixture were tested to failure by repeated loadings. The magnitude of the applied load.

which was held constant for each test specimen, and the number of cycles of repeated loading required to fracture each specimen were measured.

The most common method of reporting the results of repeated load tests is by a CT-N diagram in which the relationship between the number of cycles required to pro-duce failure and the applied maximum stress is shown. This method is satisfactory for relatively homogeneous materials with little variation in static strength.

However, a lime-fly ash-aggregate mixture is heterogeneous m character and the strength of the material will vary considerably even under controlled laboratory condi-tions. Other Investigators working with concrete have encountered the same problem and have found that the scatter of fatigue test results could be reduced considerably if another method of reporting the data was used. In this method, the relationship be-tween the number of cycles required to produce failure and the ratio of the applied stress to the static strength is shown. In a diagram of this type, the scatter in test results can be attributed to the uncertainty of determining the correct static strength and the probability of a fatigue failure.

Kesler (3) has done considerable work with the fatigue testing of concrete and has developed methods that may be used to reduce the uncertainty of determining the correct static strength. The methods presented by Kesler are used to estimate the static strength of each specimen by tests on the portions of the specimen remaining after fracture of the original specimen.

Static tests were made on lime-fly ash-aggregate specimens similar to those tested by repeated loadmg to develop the relationships required to estimate the static strength of the specimens, which were tested by repeated loading.

All loads carried by specimens were reduced to stresses before evaluation. The stress distribution in the specimens is not known exactly, especially when it is recog-nized that the stress-strain relationship and the distribution of stresses may change during the fatigue test. Because an extensive test program would be necessary to determine this stress distribution, the stresses were computed by assuming a linear stress-strain relationship. At failure this stress value is defined as the modulus of rupture. The stress caused by the dead weight of the specimen is included in all computations as its maximum value.

NOTATIONS The following notations are used:

f' = compressive strength as determmed by modified cube tests; Mr = modulus of rupture; Mj.g = estimated modulus of rupture; Mpei = estimated modulus of rupture by method 1; Mj.g2 = estimated modulus of rupture by method 2; Mres = estimated modulus of rupture by method 3; Mre4 = estimated modulus of rupture by method 4; Mj.' = modulus of rupture of the remaming pieces after the original speci-

men has fractured; N = number of cycles of stress required to cause fracture; ^min ~ minimum applied stress; "̂ max ~ maximum applied stress; and

TABLE 1 GRADATION OF MATERIALS

Percent

MATERIALS The material investigated was composed of a single mixture of lime, fly ash, and

aggregate. The particular materials used were selected as being similar to those in widespread use in several sections of the country.

The aggregate used in the mixture was taken from the subbase material stockpile from the AASHO Road Test at Ottawa, 111. Gradation of the material as determmed by ASTM D 422 (4) is given in Table 1.

The fly ash was obtained from the Public Service Electric and Gas Company, Sea Warren, N.J. The particle-size distribution of the fly ash is given in Table 1. The loss on ignition was 7.2 percent.

The lime used was a monohydrated dolomitic lime supplied by the Marble-head Lime Company, Chicago, 111.

The materials were combmed by dry weight in the following proportions: 82 percent aggregate, 14 percent fly ash, and 4 percent lime. The proportions were established from a previous laboratory mvestigation of these materials (1̂ ). The moisture-density relationships of the mixture were determined by ASTM D 698 (5). The maximum dry density was 135.4"pcf and optimum moisture content was 7.8 percent.

Sieve

PREPARATION OF SPECIMENS

Size Aggregate Fly Ash %in . lOOa ysin. 87 No. 4 73 No. 8 55 No. 16 41 100 No. 30 31 99 No. 50 16 97 No. 100 11 92 No. 200 8 87 0.05 mm 6 79 0.02 mm 4 34 0.005 mm 2 11 0.002 mm 1 10

^All material larger than 3/h i n . carded.

was d i s -Specimens 6- by 6- by 38-in. were

molded in steel forms. The lime-fly ash-aggregate mixture was mixed at or near optimum moisture content in a 2y2-cu f t Lancaster PC mixer with a mixing time of 4 min. Various methods of compaction were investigated to determine a method that would consistently compact the specimens to maximum dry density at optimum moisture content as found by ASTM D 698. The method selected consists of compacting three equal layers with 100 well-distributed blows of a 10-lb rammer dropped 18 in. The rammer had a circular head 3.87 in. in diameter. To insure bond between the layers of the specimens, each layer was scarified before the placing of the next layer.

The specimens were cured for 28 days at ambient temperatures in moist sand. This consisted of covering the molded specimens with a 1-in. thick layer of moist concrete sand and sprinkling the sand daily at a rate of % gal per sq yd.

Before testing, all specimens were rotated so that the load was applied perpendicular to the direction of compaction to minimize any effects of nonuniform layers or poor bond between layers. Full contact between the specimen and the bearing plates was in-sured by using a capping material at all points of contact.

TESTING MACHINES A schematic diagram of the fatigue testmg machine used m this study in shown in

Figure 1. A repeated load of constant magnitude was applied by the loading lever which was connected to an eccentric by an adjustable connecting rod. The eccentricity and the connecting rod could be adjusted to produce the desired magnitude of load. The load was applied to the specimen by means of a loading plate made of cold rolled steel. Several sets of electrical resistance strain gages were mounted on the loading plate which was calibrated and used as a dynamometer to measure the loads. The frequency

Connecting I— rod

Loading lever

Specimen Loading plate

I 1 ^

Support frame J

Motor driven eccentric

FigTire 1. Schematic diagram of fatigue-testing machine.

of load application was approximately 450 cycles per min. A combination of ball-bearings and rollers was used at the sup-ports to insure simple loading.

For the static tests, a Baldwin hy-draulic testmg machine was used.

Estimating Static Strength Twelve static tests were made on

specimens of this material using the same conditions of loading as was used in the repeated load tests. In most cases, the failure occurred near enough the cen-ter of the specimen to permit a modulus of rupture test to be made on each of the two remaining pieces. The conditions of loading for these two tests are shown in Figure 2. Values for the modulus of rupture were calculated for the original specimen and for the remaining pieces after fracture of the original specimen. The remaining pieces of the specimen were then tested as modified 6-in. cubes as shown m Figure 2 and a value for the compressive strength of the material was obtained.

As stated previously, the main pur-pose of the static tests was to find a re-lationship that could be used to predict the static strength of a specimen that had failed under repeated loading. When a specimen has failed by repeated loading, the modulus of rupture (Mr) of the specimen will be unknown. Various methods of estimating the modulus of rupture are shown herein. This estimated modulus of rupture is denoted as Mfe with an additional numerical subscript to denote the method used in estimating Mr. Mj.g can be compared to the known Mj. of the specimens tested statically to de-termine the reliability of the various methods that have been used to determine Mj.g. All comparisons are made on the basis of the ratio Mre/Mj.. The results of the static tests are shown in Table 2.

Because the procedures used in these static tests to predict the modulus of rupture are to be applied to fatigue specimens, the question arises of the validity of the proce-dure when applied to specimens that were originally subjected to repeated loads. No attempt was made to answer this question m this study, but Kesler (3), while working with a much larger number of concrete specimens, shows that confidence in the validity of the procedures used is justified for concrete. Because this material is quite similar to concrete it has been assumed by the authors that the strengths of the fractured speci-mens remains unaffected. Results of the estimated values of the modulus of rupture for the specimens tested by repeated loading tend to justify this assumption because the average for aU. the estimated values is 184 psi as compared with an average value of 178 psi for those specimens tested statically.

The first method to estimate the modulus of rupture was to determine the average Mj. of all of the specimens tested and use this average value as the estimated value. This is denoted as Mp-i and is given in Table 2 along with the values for Mj-e/Mj.. The standard deviation of ttie ratio Mrgi/Mj. was determined to be ±0.19 with 95 percent confidence limits of ±0.11.

The second method related M^ to the compressive strength (f') of the specimens. The compressive strength was determined from the modified 6-in. cubes. The average for the ratio M ^ f ' was determined to be 0.187. Assuming that the flexural strength and the compressive strength of the specimens are linearly related, the second estimate

(a) Modulus of rupture of specimen

- 6 -

16

I (b) Modulus of rupture of broken specimen

All dimensions in inches

7

(c) Compressive strength from modified cubes

Figure 2. Loading detai ls for s ta t ic tes t s .

of the modulus of rupture (M^ea) t^^Y be computed from

Mre2 = 0-187 f The standard deviation of the ratio Mj.g2/ Mj. was determined to be ± 0.08 with 95 percent confidence limits of ± 0.05.

The third method was to assume My equal to the modulus of rupture of the broken pieces of the specimens (M'-). The average value for the ratio Mr/M' was determined to be 1.04 with a stand-ard deviation of ± 0.08 and a 95 percent confidence limit of ± 0,05. The assump-tion of Mp/M'j. equal to 1.00 lies within the 95 percent confidence limit and is therefore statistically justified. Thus, the value of Mre3 = M'^ was taken as the third estimated value of Mj.. The aver-age value for the ratio Mj.e3/M_ was found to be 0.96 with a standard deviation of ±0.08 and a 95 percent confidence limit of ± 0.05.

Three methods of estimating the modu-lus of rupture have been shown. A fourth method was developed as a combin-ation of two of these methods. It is fairly apparent by a comparison of the standard deviations, that tA.xe2 ^re3 are about of equal correctness in pre-dicting Mp. On the other hand, using Mfgi in combination with either or both of these would give a deviation greater than by using Mi'e2 oi* Mre3 separately. At this point Mj.g]̂ was disregarded and Mre2 ^re3 ^^^^ combined, each method receiving equal weight. Thus, a fourth value for M^g was found from Mre4 %(Mj.e2 +Mre3).

TABLE 2 STATIC TEST RESULTS

Test Mr Mr- I' «re . MreJ "re. Mre. Mre. Mre. Mre. Mre< No. (psl) (psl) (pel) (psl) (pel) (psl) (psl) Mr Mr Mr Mr

19 139 131 748 178 140 131 136 1.28 1.01 0.94 0.98 20 165 161 893 178 167 161 164 1.08 1.01 0 98 0.97 21 221 228 1,055 178 197 228 213 0 80 0.89 1.03 0 97 22 ISO 125 748 178 140 125 132 1.37 1.08 0.06 1.01 23 219 222 1,108 178 207 222 215 0 81 0.95 1 01 0.98 24 197 168 965 178 180 168 174 0.90 0.91 0.85 0.89 25 206 183 1,138 178 213 183 198 0.86 1.03 0 89 0.96 26 186 174 931 178 174 174 174 0.96 0 94 0.93 0.94 27 142 141 874 178 163 141 152 1.25 1.15 0.99 1.07 28 187 188 956 178 179 188 184 0.95 0 96 1 00 0 99 29 161 181 998 178 187 181 184 1.10 1.16 1.12 1.14 30 188 156 1,009 178 189 156 173 0.95 1.01 0.83 0.92

Avg. 1.03 1.01 0.96 0.99 Std. Dev. ±0.18 ±0.08 ±0 08 ±0.06 Conf. Umlt

The average value for the ratio M .̂ was determined to be 0.99 with a standard deviation of ± 0.06 and a 95 percent confidence limit of ± 0.04. The fourth method (Mj.g4) of estimating the Mr was determined to be the most reliable.

EVALUATION OF FATIGUE BEHAVIOR Procedure

The apparatus used for the repeated load tests has been described. Loading details for the specimens subjected to repeated loads are shown in Figure 3. The general type of repeated load that was applied was a continuous cycle f luctuatmg between a maximum and a minimum value, with the stress-time pattern forming a sine wave as shown in Figure 4. The actual type of cycle that the material might be subjected to when used as a highway base course material is shown in Figure 5. This figure shows the random frequencies and stresses applied to the material. Effects of rest periods and varying magnitudes of repeated loads are extremely complex and require a basic understanding of the fatigue behavior of the material.

The minimum applied stress was due to two factors, the stress caused by the dead weight of the specimen and a small load that was constantly held on the specimen to prevent the loading apparatus from applying an impact load. The total mimmum stress level applied was held at approximately 0.12 for all tests.

After the specimens had failed by re-peated loading, the portions of the speci-men that remained were subjected to the static tests as previously described m this paper so that the modulus of rupture of the specimen could be estimated.

Test Results Results of the repeated load tests are

given in Table 3. Columns 2 and 3 give the maximum and minimum applied stresses, Oniax *^min' respectively. The estimated modulus of rupture for each specimen is shown in column 4. The ratios of the maximum and minimum applied stresses to the estimated modu-lus of rupture, a xaax/^re ° min/^re. respectively, are shown in columns 5 and 6. Column 7 gives the number of ap-plications (N) required to produce failure.

. ' . i i m r . i . i j

Specimen

Figure 3. toadlng deta i l s for load t e s t s .

repeated

^ Maximum load

Tests 1 through 15 were all completed in less than 32 hr after the loading was initiated, but tests 16, 17, and 18 required 17, 11, and 10 days, respectively, for completion. Inasmuch as a lime-fly ash-aggregate mixture continues to gain strength at a relatively rapid rate for a long period of time, the estimation of the modulus of rupture made at the end of the test would be higher than the modulus of rupture of the specimen during the testing. Specimens made by identical procedures and tested at an age of 40 days, or 12 days after the 28-day strength had been reached, gave a strength

T A B L E 3

FATIGUE T E S T RESULTS

Test "max "mln Mre "max "min N ""max No. (psi) (psi) (psl) Mre Mre

(1) (2) (3) (4) (5) (6) (7) (8)

1 124 23 173 0.718 0.133 1,900 86,800

0.670 2 96 20 168 0.571 0.119

1,900 86,800 0.510

3 83 20 187 0.444 0.107 819,600 0.376 4 124 24 194 0.640 0.124 63,100

454,000 0.585

S 112 23 201 0.557 0.114 63,100

454,000 0.497 6 95 23 172 0.552 0.134 135,100 0.484 7 97 22 170 0.570 0.129 301,600

77,300 0.510

8 119 24 167 0.714 0.144 301,600

77,300 0.661 9 127 23 183 0.694 0.126 10,000

204,300 0.650

10 106 22 190 0.558 0.116 10,000

204,300 0.498 11 109 23 165 0.660 0.140 16,100 0.600 12 115 23 169 0.681 0.136 48,600

542,200 0.624

13 105 22 195 0.538 0.113 48,600

542,200 0.479 14 130 24 206 0.631 0.116 26,000 0.581 15 131 23 192 0.684 0.120 19,800

10,432,900* 0.640

16 64 23 204 0.317 0.113 19,800

10,432,900* 0.228 17 89 21 194 0.459 0.108 6,599,700 0.395 18 88 23 192 0.459 0.120 4,828, 700 0.388 Specimen did not f a l l under repeated loading.

10

.8

max

Mre • •

Did not fail

3 4 5 6 Millions of cycles to fracture

Figure 6. Relationship between maximum stress l eve l and number of cycles to fracture (minimum stress l e v e l approximately 0.12).

equal to 118 percent of the 28-day strength. If a linear rate of gain is assumed, this is equal to a gain of 1% percent of the 28-day strength for each day after 28 days. Because this would make a significant difference for specimens 16, 17, and 18, the modulus of rupture of these specimens was estimated by using the average modulus of rupture (M^ei) and adding the estimated strength gain that would have occurred after half of the test time. For all other specimens, the modulus of rupture was estimated by method 4.

Figure 6 shows a plot of the relationship between the number of cycles required

10

.6

max

re

•

• •

Di d not fail

Number of cycles to fracture

Figure 7. Relationship between maximum stress l eve l and logarithm of number of cycles to fracture (minimum stress l eve l approximately 0.12) .

to produce failure and the ratio of the applied stress to the estimated static strength. The same relationship is shown in Figure 7 except that the number of cycles required to produce failure are plotted on a logarithmic scale.

The method of least squares was used to determine the straight line of best f i t for the data as shown in Figure 7. The equation determined is

' max M, re

= 1.00 - 0.0798 log N.

Because all tests were conducted with a minimum stress of about 12 percent of the modulus of rupture, it was necessary to determine the effect of this stress on the test results. The minimum applied stress has the effect of increasing the

^0

mm

max Mean stress level.

Figure 8. Modified Goodman diagram.

number of cycles required to cause fracture at some given maximum applied stress. A modified Goodman diagram (such as shown in Fig. 8) was used to determine the

maximum stress level corresponding to a minimum stress of zero. For each test, the mean stress level (o joax

10

fa i lu re , that at some number of cycles in the range of 1 to 10 mi l l ion that an applied stress level is reached below which fatigue f racture w i l l not occur.

In attempting to select a design stress level, the interrelationship of several factors must be considered. The tests reported in this paper were performed on 28-day old specimens at rates of loading (450 cycles per min) f a r i n excess of what might be ex-pected f o r service loading. The age at which loading is initiated and the rate of applica-tion of loading would appear to cause considerable effect on the number of cycles required to cause f rac ture . Both laboratory and f i e l d investigations of the effect of time on the strength of l i m e - f l y ash-aggregate mixtures have shown that there is a significant gain in strength f o r a long period of time (6). The rate of strength gain is dependent on the composition of the mixture, the age of material , moisture conditions, and the prevailing temperatures. The interrelationship of this strength gain and the progressive f rac ture by repeated loading is not linown. Though supporting data is not available, i t appears reasonable to assume that the mcrease m strength of a l i m e - f l y ash-aggre-gate mixture that i s developed over a considerable period of t ime would influence the fatigue behavior by greatly extendmg the number of cycles required to cause f rac ture .

CONCLUSIONS

Repeated f lexura l loads applied to the l i m e - f l y ash-aggregate mixture investigated caused the material to f rac ture as a rigid-type structure at stress levels considerably below the static strength of the material .

The relationship between the applied stress level and the number of cycles required to f racture the material can be linearized in range of 1 to 1,000,000 cycles of applied stress by relating the applied stress level to the logarithm of the number of cycles required to f racture the material . A small reduction in the applied stress can greatly increase the number of cycles required to cause f rac ture .

Further investigation should be conducted to determine the effects of varying magnitudes and frequencies of applied stress, rest periods, and the interrelationship of the continued gain in strength with t ime of a l i m e - f l y ash-aggregate mixture and the progressive f rac ture by repeated loading.

The concept of fatigue should be included as one of the c r i t e r i a i n the development of a rational method of pavement design f o r the use of l i m e - f l y ash-aggregate mixtures.

ACKNOWLEDGMENTS

This study was conducted as a part of the Lime-Pozzolan Mixtures Research Pro-gram being conducted by the Engineering Experiment Station of the University of I l l ino i s . Sponsors of the program during the period in which this study was conducted were the National Lime Association and Commonwealth Edison Company. W.W. McVinnie was a participant in the National Science Foundation Undergraduate Research Participation Program durmg the t ime he worked on this study. The investigation was under the general direction of E l l i s Danner, Professor of Highway Engineering. The authors gratefully acknowledge a l l the people who have contributed assistance and advice to this study.

REFERENCES

1. Hollon, G.W. , andAhlberg, H . L . , " F i r s t Progress Report of Lime-Pozzolan-Aggregate Mixtures Highway Base Course Research." Department of Civ i l Engineering, Univ. of n i . , Urbana (1961).

2. Kesler, C . E . , Taylor, C . E , , Corten, H . T . , and Wetenkamp, H . R . , "Mechanical Behavior of Solids." Stipes Publishing Co. , Champaign, m. (1959).

3. Hilsdorf, H . , and Kesler, C . E . , "The Behavior of Concrete in Flexure Under Varying Repeated Loads." Department of Theoretical and Applied Mechanics, Univ. of m . , Urbana (1960).

4. "ASTM Book of Standards." Pt. 4, pp. 1119-1129 (1958). 5. "ASTM Book of Standards." Pt. 4, pp. 1152-1157 (1958). 6. Hollon, G .W. , and Marks, B . A . , "A Correlation of Published Data on Lime-Pozzo-

lan-Aggregate Mixtures f o r Highway Base Course Construction." Department of CivU Engineering, Univ. of n i . , Urbana (1960).

Effects of Lime on Plasticity and Compressive Strength of Representative Iowa Soils PAUL E. PIETSCH, Engineer, Caswell Engineering Company, usseo, Minn . , and DONALD T . DAVIDSON, Professor of Civi l Engineering, Iowa State University

This paper considers the selection of 20 representative Iowa soils and the results of laboratory tests to determine the effects of a dolomitic monohydrate l ime on the plastic l i m i t and unconfined compressive strength of these soils. This is a step toward the ultimate goal of the development of a system of soi l - l ime stabilization in Iowa based on soil series.

The plastic l imi t s of a l l the soils increased with the addition of small amounts of l ime up to the l ime fixation point, after which there was l i t t l e change m the plastic l i m i t s . Although the late Wisconsin age t i l l s showed strength gains with the f i r s t additions of l ime, the older t i l l s and loess C-horizon materials gained strength only after the l ime fixat ion point had been reached. The major i ty of the A-horizon soils exhibited l i t t l e or no strength gain.

• L I M E has a long and varied history as a stabilization agent f o r soil (5, 8, 13, 14). Its use in road building, f o r example, began with the Romans and the Appian Way about 312 B. C. (15) and continues today in the building of the Interstate h i ^ w a y network (12).

The three basic mechanisms of soi l - l ime stabilization have been reported by Davidson and Handy (10). They are aggregation or flocculation of the clay particles, carbonation of the l ime by carbon dioxide f r o m the a i r , and the pozzolanic reactions.

Increasing unconfined compressive strength with the addition of l ime to soils has been reported by many authors. Increases in the plastic l i m i t s of clayey soils wi th the addition of l ime have also been reported. Hi l t (6) related the increases in strength and plastic l i m i t s i n clayey soils and reported on what he termed l ime f ixat ion. Using clay soils with a variety of clay minerals and various percentages of reagent grade calcitic l ime , he reported the increase i n plastic l i m i t wi th the addition of l ime unti l a point was reached at which there was l i t t l e or no fur ther Increase. This is the point at which l ime f ixat ion is complete. He reported that i n the same soils the unconfined compressive strengths remained constant as the plastic l i m i t s increased, after which the strengths increased and the plastic l imi t s remained nearly constant.

For this study, a number of representative Iowa soils were treated with dolomitic l ime with the objective of establishing relationships working toward a system of design-ing soi l - l ime mixes f o r road construction based on soil series. In addition, the fo l low-ing lesser objectives were also in mind:

1. To confi rm the expected relationship between plastic l i m i t and the l ime f ixat ion point.

2. To establish a relationship between the percentage of clay size material present in the soils and their l ime fixat ion points.

3. To conf i rm the expected relationship between l ime content and strength up to the l ime fixat ion point.

4. To establish the relationship between l ime content and strength above the l ime fixation point.

U

12

SOILS

Most of the bedrock of Iowa is mantled by Pleistocene glacial d r i f t deposits f r o m a l l of the major glacial stages, shown on the Iowa Geological Survey map (Fig. 1). These stages are the Nebraskan, Kansan, ni inoian, and Wisconsin, wi th the latter divided into lowan, Tazewell, Gary I , and Gary I I substages. The Gary I and Gary n substages were fo rmer ly Gary and Mankato, respectively. The largest portion of the d r i f t i s t i l l , but deposits of s t rat i f ied d r i f t are associated with i t .

Much of the d r i f t of western, southern, and southeastern Iowa is covered by loess. There are also deposits of loess, peat, volcanic ash, and alluvial materials buried within the d r i f t . At the surface of the d r i f t and loess, there are alluvial deposits associated with the present stream valleys. In the northeastern corner of the State, there i s some residual mantle, which resulted f r o m the weathering of the underlying bedrock.

In the interval since deposition, weathering has taken place on the exposed surfaces of the d r i f t and loess. This weathering has produced the soil p rof i l e . Buried soil prof i les are also present within the d r i f t and loess, indicating times of past exposure to weathering. Five factors i n the formation of soils prof i les are considered by Jenney (9): climate, l iv ing organisms, rel ief (topography), parent material and t ime. The development of Iowa soils in light of these f ive factors is considered by Simonson, et a l . (17). Individual soils prof i les exist f o r each combination of the f ive factors . This concept began wi th the Russian school of soil science and was later broadened and adopted in the United States under the leadership of Marbut (18).

Parent material was used as the basic c r i t e r ia f o r selection of the representative soils used in this study. The distribution of principal soi l parent materials in the State is d r i f t , 39 percent; loess, 42 percent; alluvium, 18 percent; and residual material , 1 percent. In view of the small percentage of residual parent material , only soils of the f i r s t three groups were considered. I t was fur ther decided to use A - h o r i -zon soils i n the study, but only those occurring in f l a t te r ra in , where i t would be more reasonable to use them than to remove them or bury them under better f i l l . In addi-t ion, soi l series, geologic age, areal extent, and vegetation were considered in the selection. The locations of the sample sites are shown on the soU association area map of Iowa (Fig . 2).

LEGEND

I lAUUVUK BSSLACUSTRME DEPOSITS

HMSCONSM Cory I

Coryl

f ^ K A M S A W CZZmicaRASiuN

Figure 1. Prelijiiinary map of the glacial geology of Iowa.

13

The soils selected f o r the study are given in Table 1. The three loess C-horizon and one loess B-horizon samples were obtained f r o m the southwestern portion of the State, where there appears to be a systematic variation in particle-size distribution with distance f r o m the Missouri River (7). They have approximately the following percent si l t -clay distributions: 80-20, 70-30, 60-40, and 50-50.

Because of the l imi ted information about Iowa t i l l s , they were sampled in random fashion, based on geological age. Two t i l l s were not sampled, Nebraskan because of l imi ted exposure in Iowa and niinoian because of l imi ted occurrence. Because the youngest Gary I I d r i f t i s mapped in greater detail, the samples were obtained f r o m an area of ground moraine. Kansan gumbotil f r o m southeastern Iowa was also sampled to obtain a soil wi th high clay content and because of i ts troublesome nature.

Although alluvium accounts f o r about 10 percent of the parent material, i t i s widely scattered. However, the Missouri River f lood plain is the largest single area in the State, and accounts f o r a large portion of the total alluvial material . A sample repre-sentative of the high clay content overbank material was selected f r o m this area.

MIX MATERIALS AND LABORATORY WORK

Commercially available dolomitic monohydrate l ime , sold under the trade name Kemidol, was used throughout the study. I t was manufactured by the U.S. Gypsum Gompany at Genoa, Ohio. Dist i l led water was used in a l l the mixes and testing procedures to eliminate experimental variables.

Sample Preparation Af te r drying the f ie ld ' samples and brealcing the larger soil clods, representative

r / T o z e w e l l ^ / .nGolvo | . '

H a i ^ u r g

Kenyon ebster

Clarion

Dnton^

/ A a n d /

•psburgy Shorpsburg A.B and C

j iEdlna

CKC:Cre8co,Kasson, Clyde CL: Clinton ondLlndley CW: Clarion and Webster F: Faye t t e FDS: Fayette, Dubuque, Stony Land 6H: Grundy and Hoig GPS- Golva.Primghor.Sac

Wafi"''^"'*^ MIHtMonono, I d a , Hamburg

8: So i l* of b o t t e m t o n d s

Konson

Mo- Moody MPS: Morcus, Primghor, Sac MT- Mahaska and Taintor SCW: S to rden ,C la r ion .Webs te r S6H: Shelby,Grundy, Hoig SSE: Shelby, Seymour, Edlno SSW: Slielby, Shorpsburg, Winterset TD: Tamo and Downs , TM: Tama and Muscat ine WL: Weller and LIndley

X: Sample s i t e s

Figure 2 . Map of principal soil associations of Iowa and locations of samples used, from Simonson et a l . (17, p. 36).

14

samples of each of the 20 soils were obtained. The remaining portion of the soil was passed through a No. 4 sieve and used in molding the 2- by 2- in . specimens f o r the moisture-density and strength tests.

Descriptive Tests

The following descriptive tests were performed on each of the 20 soils except where differently indicated:

1. Particle size analysis. Standard mechanical analysis (ASTM Designation; D 422-54T) (1); sodium metaphosphate dispersing agent; Iowa State A i r jet dispersion device (2).

2. Organic matter. A horizon soils only; potassium dichrornate t i t ra t ion method (3).

3. X-ray diffractometer analysis. To determine the predominant clay mineral present in the soils.

TABLE 1 SAMPLE SITES

Parent Sample Depth Plant Tier and Material No. Series Horizon In . Gover County Range Section Kansan t i l l 423-1 Lindley A 0-15 Trees Appanoose T70N,

R16W NW% NW'/i 2

423-5 Lindley C 157-205 Trees Appanoose T70N, R16W

NWVi NWy, 21

528-8 Gumbotil Fossil B 91-107 Grass Keokuk T75N, NWV. NWy« 7 RlOW

lowan t i l l - Kenyon A 2-14 Grass Butier T91N, R16W

SW% NW'/i 14

G 36-60 Grass BuUer T91N, R16W

SWVt NWV, 14

Tazewell - -t i l l C 36-48 Grass O'Brien T94N,

R39W NW% NE'/4 27

Gary I t l U - Glarion C 36-72 Grass Story T83N, R24W

NE% SW'/i 5 Story T83N, R24W

Gary n t i l l - Glarion A 0-12 Grass Calhoun T87N, R32W

SE% SEV, 4

C 72-96 Grass Calhoun T87N, R32W

NE'/« SE%30 T87N, R32W

Webster A 0-15 Grass Calhoun T88N, R44W

SW% SE'/« 28 T88N, R44W

Wisconsin loess 15-2 Hamburg G 120-132 Grass Monona T83N,

R44W NW% NW% 10

Marshall A 2-12 Grass Shelby T79N, R37W

NW'A NW% 13

28-1 MarshaU C 72-84 Grass Shelby T79N R37W

tm% NW% 13

512-1 Sharpsburg A 1-12 Grass Clarke T71N, R27W

NE% N E % 4

512-2,3 Sharpsburg B 12-46 Grass Clarke T71N, NEVi N E % 4 R27W

512-4,5 6 Sharpsburg G 46-94 Grass Clarke T71N,

R27W NEVi NEy4 4

- Edina A 0-15 Grass Wayne T69N NE% N E % 22 Wayne R23W

524-1 Clinton A - 0-6 Trees Mahaska T77N, R17W

swy« SWVt 29 319-1 Galva A 0-8 Grass Plymouth T92N,

R43W N E % NE'/« 7

Missouri River alluvium 627-1 - - 0-48 Trees Harrison T79N,

R45W SE'/< SW'/t 21

15

4. Garbonate content. Material passing No. 40 sieve; leaching and t i t rat ion with versenate and treatment wi th dilute HC 1 (4).

5. Deternunation of pH. Material passing No. 40 sieve; Leeds and Northrup C o m -pany Universal pH meter.

Atterberg L imi t s

Liquid l i m i t s , plastic l im i t s , and plasticity indexes were determined fo r each of the 20 untreated soils. In addition, the plastic l imi t s were determined f o r each soil wi th 1, 2, 3, 4, and 7 percent l ime by dry weight of soil added. The l imi t s were determined according to standard ASTM procedures, except that the soil-water and so i l - l ime-water mixes were cured f o r two days in a moisture room.

Moisture Density Relationships and Strength Tests

Two- in . high by 2- in . diameter specimens were prepared and cured in accordance with procedures described by Hi l t (6). Unconfined compressive strength tests were carr ied out i n the Soiltest, Inc . , stability testing machine.

Nine soils were selected f o r the prel iminary studies of moisture-density and moisture-strength relationships. For two of these soils, nine 2- by 2- in . specimens were molded at varying moisture contents, f o r each of 0,6, and 12 percent l ime by oven dry weight of so i l . The dry density of each group of nine specimens was deter-mined f r o m their height and weight, and the moisture content of the mix. Three of the specimens were tested f o r unconfined compressive strength at the end of curing periods of 7 days, 28 days and 28 days plus 1 day immersed.

For the seven other soils in the f i r s t group, six specimens were molded at varying moisture contents, with l ime contents of 0, 6, and 12 percent. Strength tests were made on three of these specimens at the end of 7- and 28-day curing periods.

Only moisture-density relationships were determined f o r the eleven remaining soils. Three specimens were molded at each different moisture content.

Final unconfined compressive strength tests were made on a l l 20 soils, wi th 0, 1, 2, 4, 8, and 12 percent l ime added. For each l ime content of each soi l , 12 specimens were molded at a chosen optimum moisture content. Three of these specimens were tested after curing periods of 7 days plus one day immersed, 28 days, and 28 days plus 1 day immersed.

RESULTS

Descriptive Tests

The results of the particle-size analyses are given in Table 2. The soils are grouped according to parent material and horizon, and numbered f o r future reference. The groups containmg more than one soil are fur ther arranged according to the percent of 5-fi clay present i n the whole sample.

The results of the analyses f o r carbonate content are given in Table 3. The results of the versenate test are reported as a percent of the oven dry weight of soil passing the No. 40 sieve. The results of the test wi th dilute H C l are expressed as calcareous or noncalcareous, wi th the major i ty of soils noncalcareous.

F r o m the X-ray analyses, i t was determined that montmorillomte was the pre-dominant clay mineral present i n each of the 20 soils. In addition, each X-ray trace was checked f o r the presence of carbonates in the f o r m of calcium or magnesium car-bonate peaks. The presence of one or both of these peaks in a noticeable intensity corresponded in every case to the soils having a carbonate content c£ more than 9 per-cent as determined by the versenate test.

The results of the tests f o r organic matter and pH are also given in Table 3. The organic matter content i s expressed as a percent of the oven dry weight of soil passing the No. 4 sieve, wi th variations in organic matter present in each of the A-horizon groups. Also, the values of pH l ie in the 5 to 9 range, wi th the majori ty of the values in the 6 to 8 range.

16

TABLE 2 PARTICLE-SIZE mSTRIBUTION OF SOIL SAMPLES

Sou Gravel

Whole Sample (%) 5-11

Sand SUt Clay 2-11

Clay Percent Passing

Group Name Number Gravel

Whole Sample (%) 5-11

Sand SUt Clay 2-11

Clay No. 4 No. 10 No. 40 I AUuvial 627-1 0 0 2.4 28.8 68.8 57.4 100 0 100.0 99 4

n Sbarpsburg A 512-1 0.0 1 7 56.5 41.8 33.4 100.0 100.0 99.6 Galva A 319-1 0.0 1.8 61.4 36.8 28.8 100.0 100.0 99.8 MarsbaU A - 0.0 0.7 68.3 31.0 24 8 100.0 100.0 99.9 CUnton A 524-1 0.0 1 2 69 8 29.0 24.0 100 0 100 0 99.8 Edina A - 0 0 2 3 69.3 28.4 19.0 100.0 100.0 99.0

m Sbarpsburg B 512-2,3 0.0 0.7 52.9 46.4 38.0 100.0 100.0 99 7 IV Sbarpsburg C 512-4,5,6 0.0 0.6 57.3 42.1 33.8 100.0 100.0 99.9

Marshall C 28-1 0.0 0.3 70 5 29 2 23.0 100 0 100 0 100.0 Hamburg C 15-1 0.0 0.0 81 0 19.0 15.0 100.0 100 0 100.0

V Webster A - 0.0 11.7 39.7 48.6 38.4 100.0 100.0 98 2 Clarion A - 0.2 18.1 42.5 39.2 28.0 100.0 99.8 95.9 Kenyon A - 0.5 37.5 38 6 23.4 15 8 100.0 99.5 90.6 Undley A 423-1 0 9 38.8 47.7 12 6 8.4 99.8 99 1 93.1

VI Gumbotll 528-8 1.0 21.2 15.2 62.6 58 8 99.8 99.0 92.7 v n Kansan ' 423-5 1.8 29.1 33.5 35.6 28.0 99.6 98.2 90.8

low ail - 3.4 33.2 30.2 33.2 28.8 99.0 96.6 91.4 Tazewell - 8.0 26 5 33.1 32.4 24.6 97.7 92.0 83.5 Cary n - 9.1 29.9 31 0 30.0 22 2 97 1 90 9 81.7 Cary I - 10.0 41.2 30 2 18.6 12.8 95 3 90 0 78.4

TABLE 3 DESCRIPTIVE TEST RESULTS AND U M E FIXATION POINTS

Sou Organic Carbonates bonates Group Name Number Matter LFPa

(i) RA^ PH (Ji) I A l luv ia l 627-1 1.33 3.1 NC 8.05 3

n Sharpsburg A 512-1 2.41 1.3 NC 6.72 2 Galva A 319-1 4.61 2.1 NC 7.31 2 Marshall A - 0.55 2.6 NC 6.92 2 Clinton A 524-1 2.04 1.2 NC 6.45 3 Edina A - 3.50 1.0 NC 5.19 1

m Sharpsburg B 512-2,3 2.2 NC 6.28 3 IV Sharpsburg C 512-4,5,6 2.3 NC 6.88 4

Marshall C 28-1 1.4 NC 6.98 3 Hamburg C 15-2 10.8 C 8.40 2

V Webster A - 3.76 9.0 c 8.04 3 Clarion A - 4.77 1.9 NC 6.17 4 Kenyon A - 3.97 1.2 NC 6.58 1 Llndley A 423-1 1.62 0.9 NC 6.58 0

V I Gumbotll 528-8 1.9 NC 7.03 4 vn Kansan 423-5 9.6 C 8.24 2

lowan - 1.6 NC 6.83 2 Tazewell - 26.2 C 8.49 3 Cary n - 15.6 c 8.50 3 Cary I - 16.2 c 8.27 2

l̂ime fixation point, percent liae based on oven dry weight of soi l . ^Calculated carbonate content from amount of calcium determined from versenate test, percent carbonate based on oven drjr weight of soil .

*Tlelative amount of caAonate present by dilute HCl test, reported as calcareous (C) or noncalcareous (NC).

17

Glassification of the 20 soils according to the Highway Research Board System and the Unified System is given in Table 4.

Atterberg L i m i t s The l iquid and plastic l i m i t s and the plasticity indexes of the 20 soils used are given in

Table 4. When the Atterberg l imi t s are compared w i f h the soils arranged in decreasing

amount of 5-^ clay present i n the whole sample, a general relationship between the two is apparent. Plots of these l i m i t values vs the amount of 2-fi clay present in the portion of the sample passing the No. 10 sieve are shown in Figure 3. In these graphs, the single soils i n groups m and V I have been combined wi th their respective G-horizon groups, I V and Vn.

Gonsidering groups I I I and IV , a straight line relationship exists f o r the l i m i t s and the plasticity indexes of the three group IV soils. Both the plastic and l iquid l i m i t s of the group m B-horizon soil l ie above the lines connecting these same values f o r the

TABLE 4

ENGINEERING SOIL GLASSIFIGATIONS AND ATTERBERG LIMITS

Soil Glassification Liquid Plastic Plasticity Group Soil HRB Unified L i m i t L i m i t Index

I A l luv ia l A-7-6 (20) GH 72.0 26.0 46.0 n Sharpsburg

20.0 A A-7-6 (14) OL 47.5 27.5 20.0 Galva A A-7-5 (14) OL or OH 50.0 31.0 19.0 Marshall

A A-7-6 (11) GL 40.5 23.5 17.0 Glinton A A-6 (9) OL 37.0 24.0 13.0 Edina A A-7-6 (9) OL 40.5 28.5 12.0

m Sharps-32.0 burg B A-7-6 (19) GH 56.0 24.0 32.0

rv Sharps-28.0 burg G A-7-6 (17) GL 48.0 20.0 28.0

Marshall G A-6 (10) CL 37.5 23.0 14.5

Hamburg G A-4 (8) M L 31.5 23.5 8.0

V Webster A A-7-5 (20) GH 60.0 30.5 29.5

Glarion A A-7-5 (15) OH 54.0 33.5 20.5

Kenyon 22.0 A A-7-6 (11) GL 47.5 25.5 22.0

Lindley 3.5 A A-4 (5) M L 21.0 17.5 3.5

V I Gumbotil A-7-6 (20) GH 76.0 22.5 53.5 vn Kansan A-6 (9) GL 34.0 17.0 17.0

lowan A-6 (10) GL 39.0 18.0 21.0 Tazewell A-6 (8) GL 34.5 18.0 16.5 Gary I I A-6 (8) GL 37.0 19.0 18.0 Gary I A-4 (3) SG 24.0 14.5 9.5 •>

18

G-horizon soils. However, the plasticity index does lie on the line connecting the plasticity indexes for the C-horizon soils. Inasmuch as only one B-horizon soil was used, no direct conclusions can be drawn. These straight line relationships exist between the soils of group IV in spite of the nonsystematic variation of carbonate content, indicating little effect of this variable on the Atterberg limits of the untreated soils. The equations for the lines connecting the various points of the group IV soils are

Liquid limits (LL) = 0.87 x G + 18.0 Plastic limits (PL) = -0.188 x G + 26,9 Plasticity indexes (PI) = 1.08 x G -8.8

in which G = percent of 2-n clay of the whole sample passing the No. 10 sieve. In the curves for the soils of group H, it would seem that the same general relation-

ship exists between the Atterberg limits and the clay content. However, when plotted, the points are much more scattered. In this group other variables are introduced, especially that of organic matter. Because the Sharpsburg and Glinton soils have approximately the same amount of organic matter, lines were drawn connecting their limits and plasticity indexes. These same points were connected for the Galva and Edina soils, which also have about the same organic matter content. The equations for the Sharpsburg-Glinton and Galva-Edina lines, respectively, are

Liquid limits: 1. LL = 1.10 x G + 10.8 2. LL = 0.95 X G + 22.5

Plastic limits: 1. PL = 0.37 x C + 15.3 2. PL= 0.25xG + 23.8

Plasticity indexes: 1. PI = 0.73 X G - 4.6 2. PI = 0.71 X G - 1.6

in which G = percent of 2-fi clay of the whole sample passing the No. 10 sieve. In group V, the general relationship again appears, although there are great varia-

tions on plottiJig. Points for similar organic matter content Webster and Kenyon soils were again connected, and the following equations were obtained:

Liquid limits (LL) = 0.53 x G + 39.5 PlasUc limits (PL) = 0.27 x G + 20.5 PlasUcity indexes (PI) = 0.27 x G + 18.8

in which G = percent of 2-ji clay of the whole sample passing the No. 10 sieve. Four of the group VH soils are closely related in clay content, liquid and plastic

limits, and plasticity indexes, though varying greatly in geological age and carbonate content. Althoî h the Gary I till provides a point of lower clay content, no soil of group vn has a high clay content. In view of the relations found in groups HI and IV, it would seem highly questionable to use the group VI fossil B-horizon gumbotil for a point of high clay content. The equations for the lines drawn from the Gary I points through the four bunched points are

Liquid limits (LL) = 0.94 x G + 10.7 PlasUc limits (PL) = 0.35 x C = 8.3 PlasUcity indexes (PI) = 0.72 x G = 1.1

in which G = percent of 2-fi clay of the whole sample passing the No. 10 sieve. Although these lines extended pass near the gumbotil points, their relationship is suggestive of that found m the soils of groups m and IV.

The loess G-horizon soils of group IV exhibit the best straight line relationship for all the soils of a group. The till soils of group VII also seem to show this single straight line relationship, but this is somewhat uncertain because of the bunching of four of the points and the lack of a point of high clay content in the group. Both of these groups have these relationships in spite of unsystematic variation in carbonate

19

content, leading to the belief that i n parent material this variable is unimportant in the Atterberg l i m i t s of untreated soils.

The A-horizon soils, on the other hand, do not exhibit the smgle straight line r e -lationship f o r the soils of given parent material . In the soils of group H, two groups of two soils, each two having close to the same amount of or game matter, yielded two lines f o r each combination of l iquid l i m i t , plastic l i m i t , and plasticity index points, both of the lines having approximately the same slope. This leads to the theory of a fami ly of lines f o r both of the l i m i t s and the plasticity index, f o r each type of parent material . The line famil ies would then f a l l into some l imi t ing ranges, defined by the soils found in Iowa. Certainly, many more points would be needed to prove this theory. In addition, work would be needed on the B-horizon soils to determine i f they would follow the single line approach or if they would be dominated by the variables which lead to the f ami ly of lines in the A-horizon soils.

Moisture-Density Relationships and Strength Tests

Optimum Moisture Content Determinations.—From the results of the prel iminary moisture-density and moisture-strength studies on nine soils, curves were plotted showing dry density and unconfined compressive strengths vs moisture content f o r 0, 6, and 12 percent l ime added. Unconfined compressive strengths after curing periods of 7 days, 28 days, and 28 days plus 1 day immersed were obtained f o r the Hamburg C-horizon and Sharpsburg B-horizon soils. Curves f o r the f i r s t two curing periods only were obtained f o r the seven remaining soils. In addition, a compromise moisture content (CMC) curve was plotted f o r each of the nine soils. This curve was determined according to procedures given by Katt i et a l . (11). In determining the CMC, the strength vs moisture content curves f o r 0 percent l ime were used only f o r the Hamburg C-horizon and Sharpsburg B-horizon soils. A representative graph, i l lustrat ing the dry density, strength and CMC curves is shown in Figure 4.

Table 5 gives the optimum moisture content fo r maximum dry density (OMC) and the CMC f o r each of the nine soils. In addition. Table 5 gives the correction factor and the f i na l compromise moisture content (FCMC). The correction factor was de-termined in i t ia l ly f o r the nine soils as the difference between the OMC and the CMC, corrected to the nearest 0. 5 percent. Other slight adjustments were made in some of the correction factors to better f i t a given group or to eliminate excessive adjustment of the OMC in the rounding-off processes.

Good curves were obtained f o r the alluvial soi l , shown i n Figure 4. The value of the CMC was determined and a correction factor of 2.0 was selected.

Both loess A-horizon soils of group H exhibited less than ideal curves, the maxima on the moisture-strength curves occurring at appreciably lower moisture contents than the OMC. The differences between the OMC and CMC were 5.2 percent f o r the higher clay content Sharpsburg A-horizon soil and 7.3 percent f o r the lower clay content Clmton A-horizon so i l . The correction factors were placed at 5.0 and 7.0 percent respectively.

Standard moisture-density and moisture-strength curves were obtained for the two soils of groups IH and I V . For both soils, the minimum of the CMC curve occurred within 0.5 percent moisture content of the OMC. Therefore, the correction factor in both cases was selected as zero.

The curves f o r the soils of group V were somewhat errat ic , wi th l i t t l e difference between OMC and CMC f o r the Webster soil and 5. 7 percent difference fo r the Lindley soi l . The correction factors selected were 0 and 5.5 respectively.

Group V I and vn soils yielded generally good curves. With a difference of 1.4 percent between the OMC and CMC for the gumbotil and 1.7 percent f o r the Cary I I , 1.5 percent was selected f o r the correction factor.

Following the extensive prel iminary tests to obtain the correction factor by relating the OMC of the untreated soil to the CMC of the same soil , the OMC's were determined fo r the remaining eleven untreated soils. Correction factors were also selected f o r these eleven soils. In group n, the correction factors were selected on the basis of clay content of the soils considermg two already determined, except the factor f o r the

20

LIHEnXATION POINT, % U m t '

MOISTURE

CONTENT

WOgbl et Sdl

UME FIXATION POINT,%Limi *

MOISTURE

CONTENT,

% Oven Dry 40 WelgbtetSell

ptaltic Unit - * plot!ki t , indta - ^

L o e t i A Horlion

Til l A Nerl ien

Leete B onil c Herliee

Sombelll end Tin C Herb Her l iee

2 MICRON CLAY. « ef Perllen Peteing NalO Sieve

Figure 3. Atterberg limits and lime fixation points at varying clsy contents.

TABLE 5

MOISTURE CONTENTS

Sou OMG CMC Correction FCMC Group Name Number (55) (5«) Factor {%) m

I A l luvia l 627-1 25.4 27.7 2.0 27.5 n Sharpsburg A 512-1 20.5 15.3 -5.0 15.5

Galva A 319-1 24.1 - -5.5 18.5 Marshall A - 19.0 - -6.5 12.5 Clinton A 524-1 18.4 11.1 -7.0 11.5 Edina A - 24.6 - -6.5 18.0

m Sharpsburg B 512-2,3 21.8 21.3 0 22.0 IV Sharpsburg C 512-4,5,6 19.5 - 0 19.5

Marshall C 28-1 19.0 - 0 19.0 Hamburg G 15-2 18.0 18.3 0 18.0

21,

TABLE 5 (Continued)

Soil OMC CMC Correction FCMC Group Name Number m Factor 0>) (^)

V Webster A 25.5 25.3 0 25.5 Clarion A - 20.3 - 0 20.5 Kenyon A - 20.4 - 0 20.5 Lindley A 423-1 13.0 7.3 -5.5 7.5

V I Gumbotil 528-8 23.2 24.6 1.5 24.5 vn Kansan 423-5 14.4 - 1.5 16.0

lowan - 12.4 - 1.5 14.0 Tazewell - 15.2 - 1.5 16.5 Cary n - 16.1 17.8 1.5 17.5 Cary I - 11.9 - 1.5 13.5

200i

SUMMATION OF DEVIATIONS,

% lOOh

0

4001 |_ 0 % l i m e - o 6 % l i m e - x 12% l i m e - A

UNCONFiNED

COMPRESSIVE

STRENGTH.

320

2 4 0

160

DRY DENSITY,

pcf

0

9 8

9 4

9 0

8 6

T 1 r Alluvial Soil

-CMC

L 1-28 day moist x cured

7 d o y moist cured / -28 day moist cured

OMC

MOISTURE CONTENT. % Dry Weight of Mis

Figure h. Unconflned compressive strengths, dry densities, and summations of de-viations for alluvial soil mixed with varying amounts of lime at varying moisture con-

tents.

22

Edina soil was placed below 7.0 percent because i ts OMG more closely resembled the soils in this range. The Marshall and Sharpsburg C-horizon soils of group IV were given correctioiT factors of zero. The Clarion and Kenyon soils of group V were given correction factors of zero, as they appeared to more closely resemble the Webster soi l . The remaining t i l l soils of group VII were given correction factors of 1.5 per-cent.

The FCMC was then determined f o r each soil by applying the correction factor to the OMG of the untreated soi l . The FCMC was used as the molding moisture content f o r the f ina l strength tests.

Strength Tests. - T h e plotted curves of strength at FCMC vs l ime content f o r the various curing periods are shown f o r each of the 20 soils in Figures 5 through 8. Figure 9 shows strength at different l ime contents plotted against 2-/ i clay content of the portion passing the No. 10 sieve. Groups HI and IV and VI and vn are shown to -gether.

The alluvial soil of group I , shown in Figure 5, gains a maximum 28-day dry un-confined compressive strength of 320 psi with 12 percent l ime added, with a closeness of the dry and immersed strengths, part icularly at the higher l ime percentages. This soil was one of the f ive tested that showed immersed strengths with no l ime added.

UNCONFINEO

COMPRESSIVE

STRENOTH,

pel

ealeo A Ed no A

Cllntos A

Merehel l A

7 do, a o i t t cartd - o do, neUt cored - x

tooted dry -24 beoro Iramorood - - —

S 10 12 0 2 4

LIME CONTENT, % Oeen Dry Weight at Sel

F i g u r e 5. U n c o n f i n e d compress ive s t r e n g t h s o f a l l u v i a l s o i l and f i v e l o e s s A - h o r i z o n s o i l s a t v a r y i n g l i n e c o n t e n t s .

23

The soils of group n, also shown in Figure 5, gain l i t t l e strength with the addition of l ime, the Marshall showing the highest 28-day dry unconfined compressive strength of 275 psi . A l l of the group H soils have their highest strength at 12 percent l ime, with the exception of Clinton, which has a maximum strength at 8 percent. There is l i t t l e relation between clay content and strength, as shown in Figure 9. The Marshall has the highest carbonate content and lowest organic matter content of the group, apparently accounting f o r i ts much higher strength.

Soils of groups m and IV appear to behave most systematically. In the curves of strength vs clay content (Fig. 9), there is the least variation in strength with the lower l ime contents. However, as the l ime content increases, the inverse relationship of strength to clay content becomes more apparent. There is also a systematic change in the shape of the strength vs l ime content curves (Fig. 6) as the clay content of the soil changes.

The soils of group V, shown in Figure 7, exhibit generally poor strengths, wi th the Webster soil having the high 28-day dry unconfmed compressive strength of 280 psi at a l ime content of 12 percent. There is a general decrease in strength with de-creasing clay content f o r the three soils that appear to be most s imilar , with the Lindley exhibiting an mcrease in strength, as shown in Figure 10. Also there is a much higher carbonate content f o r the Webster soi l . Further, though the Lindley and Kenyon soils contain approximately the same amounts of sand and gravel, the Lindley has less clay and more s i l t size material .

5oor

2S0h

UNCONFINED Q

COMPRESSIVE

STRENGTH.soO,

p t i I Morsholl C

250

Shorpsbura B

rdoymoltlcured - o ZBdoymoisI cured- x tested dry 24 hours Immersed- -

Sharpsburg C

Homburg C

LIME CONTENT, % Oven Ory Weight of Sell

Figure 6. Unconfined compressive strengths of the loess B-horizon soil and three loess G-horizon soils at varying lime contents.

24

As mentioned earl ier , the soils of group V H are alike i n many respects and d i f fe r greatiy in others, wi th some overlapping between groups depending on the point of division. For example, four of the soils have very high carbonate contents, whereas another grouping of four show much s imi la r i ty in clay content and Atterberg l i m i t s . If the gumbotil of group V I is also considered, three of the soils are of relatively recent age (16,000 years or younger) whereas three are of lowan age or older.

I t is part icularly in the consideration of strength that a division on the basis of geological age shows up to the best advantage. The younger t i l l s (Tazewell, Gary I , and Gary n) gain strength immediately with the addition of small amounts of l ime, r ise to a peak strength, and then decline in strength as more l ime is added. The strength of the Kansan t i l l r ises rather abruptly wi th additions of l ime , after remain-ing constant f o r a short period, and the gumbotil and lowan t i l l strengUis r ise steadily as l ime is added. Although the younger t i l l s have a 28-day dry unconfined compressive strength of no less than 520 psi with 4 percent l ime added, the older t i l l s and the gum-botil exhibit a maximum strength of 285 ps i with the addition of 12 percent l ime, as shown in Figures 8 and 9.

In general, the shapes of the curves f o r 28-day unconfined compressive strength vs l ime content f i t into one of the following four cases:

1. Gase A . —Strength gains begin immediately with the addition of small amounts of l ime , and the curve rises abruptiy to a peak strength wi th the possibility of a slight decrease in slope before the peak is reached. Strength decreases after the peak with fur ther additions of l ime . The Tazewell, Gary I , and Gary H t i l l s follow this pattern.

lOOr

Clarion A

UNCONFINED

COMPRESSIVE

STRENGTH.2oo|

Webster A

10 12

Lindley A

7 aof moot curad - o 28 tfojr mobi cured - x la«tad dry 24 hour* Immeraad—

10 12

LIME CONTENT, % Oven Dry Weiobt of Sell

Figure 7. Unconfined compressive strengths of four t i l l A-horizon soils at varying lime contents.

25

2. Case B. —Strength increases slightly, remains the same, or decreases slightly as the f i r s t amounts of l ime are added. Af te r a point, an abrupt strength increase takes place with the addition of more l ime, unt i l another break point is reached, after which strength remains constant or increases or decreases slightly with fur ther addi-tions of l ime . This case includes a l l soils of groups I , m, and IV and the Kansan t i l l soil of group Vn.

3. Case C. —Strength tends to increase slightly, remain the same, or decrease slightly with the addition of small amounts of l ime. Thereafter, strength tends to i n -crease continuously with the fur ther addition of l ime. Soils in this case include lowan t i l l , gumbotil, Webster A-horizon and Marshall A-horizon.

4. Case D. —Strength shows very l i t t l e increase regardless of the amount of l ime

4O0t-UNCONRNED COHPRCSSIve

STRCNOTH,

p*< 2ao|

C a n I till

Cory I till

2 4 e B to 12

Xoneoii t i l l

Tdoyaol t te i i fed -2edo,molatcorod -tooted dr, 24 hearo lotHiorood-

UME CONTENT, % Oeea Dry WelgM ef Sail

Figure 8. Unconfined compressive strengths of gumbotil and five t i l l C-horizon soils at varying lime contents.

26

added. This case includes all the loess and t i l l A-horizon soils with the exception of the Webster and Marshall soils.

It is particularly evident cases B, C, and D are gradational. The peaks on the curves of case B smooth out to approach more closely a straight line (as in Case C), and the straight line gradually decreases in slope as the total strength gain becomes less until there is little total strength gain (as in Case D).

Lime Fixation Point

Plastic limits at various lime contents were plotted against lime content to ascertain the lime fixation point. These curves are shown in Figure 11. The lime fixation point was selected from the curves as the point at which plastic limits no longer mcreased with the addition of more lime. As a check, plots were also made of equivalent 28-day dry unconf ined compressive strength vs plastic limit, with selected curves shown in Figure 10.

Because lime does not react with material of greater than silt size (16), there would be a greater concentration of lime to reaction size material in the strength specimens than m the material for the plastic linut tests if the same lime content was used in both cases. Therefore, the equivalent lime content was f i r s t determined according to

in which

L e = L X P«_ P4

L e = equivalent lime content; L = original lime content;

le D A Y

C O M P R E S S I V E

S T R E N G T H .

LOESS B gad C H o r l l M S o i l i

T I L L * H o r i i o a S o l K eUMBOTIL end T I L L C H o r l l M S o i l s

^ ' | * ^ l f m t conttnt

-is is I memN C L A T , » O< tatlM P a t H o g N & I O S l a n

Figure 9. Unconfined compresBive strengths at varying lime contents after 28 days.

27

P40 = Percentage of soil passing No. 40 sieve; and

percentage of soil passing No. 4 sieve.

P4

The equivalent unconfined compressive strength was then determined using the equivalent lime content, which was lower than the lime content for the corresponding plastic l imit test. Corrections were made only for groups V, VI, and vn, because in the other five groups at least 99 percent of the sample passed the No. 40 sieve.

To eliminate the problem of bias in the selection of the LFP from the plastic l imit vs lime content curves, the LFP was chosen as the point at which there was an abrupt change to a slope of opposite sign or the point after which the slope of the curve was one or less. The values determined for the LFP are shown in Table 3.

The values obtained from the plastic l imit vs lime content curves were then com-pared to those obtained f rom the strength vs plastic l imit curves. Most of the values compared rather well. In the Sharpsburg and Edina A-horizon soils, there was a lag between the percent lime at which the plastic l imit stopped increasing and the percent lime at which strength started increasing. Also, the Tazewell, Gary I and Gary n soils plots of strength vs plastic limit were irregular, in that the strength began in -creasing immediately on the addition of lime, rather than after an amount of lime sufficient for lime fixation had been added. There were also some discrepancies in

1 1 1 1 1 1 Lindlcy A

1

1 ^ l i R t f l content

4% 14% >-3 % %

1%

1 1 Shorptbttrg A

—1 1 r 1 ^12% i s % 1^2%

tJr* * fr^ —t tn t« cont«nt 0% 1% ^

EQUIVALENT UNCONFINED

COMPRESSIVE STRENOTH,

Cory B

e % f 1

J2% i lima coot tnt^ . _, 1

1 \ \

z%y V*

0%5-

' 1 1 1 1 1

Alluvial 7 tfor moUl ciirttf - • 28 4oy noUt eurtd - •

lime contttat

PLASTIC LIMIT, % 0»an Dry Weight el M i l

Figure 10. Comparison of equivalent unconfined compressive strengths and pl a s t i c l i m i t s for four s o i l s .

28

the Kenyon and.Lindley A-horizon soils, but the smaller, more practical values of LFP from the f i r s t curves were used because of the very small total increase in the plastic limits of both soils.

The lime fixation points were also plotted vs clay content for each of the soils, except the alluvial soil, as shown in Figure 3. Again, the group IV loess C-horizon soils yield the best straight line relationship. The range of LFP values for the soils of groups n and I I I is 1 to 3 percent; for group IV soils, 2 to 4 percent.

The group V t i l l A-horizon soils had lime fixation points m the 0 to 4 percent range, with the group VII soils in the 2 to 3 percent range. The LFP of the alluvial soil was 3 percent and that of the gumbotil 4 percent.

P L A S T I C LIMIT,

% Dry W«i«M

o « M i . , 0

, - '>Galv

29

CONCLUSIONS 1. A straight line relationship exists in the loess C-horizon soils of Iowa between

the Atterberg limits and the 2-iiclay content. The general trend of this relationship continues into other Iowa soil groups studied, but no definite conclusions can be drawn about these groups.

2. Additions of lime increase the plastic limits of Iowa soils up to the lime fixa-tion pomt, even though the total increase in plastic limit may be small or the leveling off of plastic limit values after the lime fixation point is reached may not be as apparent in some soils as m others.

3. Lime fixation occurs in the loess C-ho'rizon soils of Iowa in the 2 to 4 percent lime range, the amount required being proportional to the amount of clay size ma-terial in the soil and independent of carbonate content of the soil. The range of lime fixation for loess A- and B-horizons is 1 to 3 percent, with no definite relation to clay content.

4. Lime fixation occurs in t i l l C-horizon soils of Iowa in the 2 to 3 percent lime range, and appears to be interrelated to particle size and geological age. The range of lime fixation in t i l l A-horizons is 0 to 4 percent.

5. Iowa loess B- and C-horizon soils exhibited marked strength gains with the addition of lime in amounts above the lime fixation point. The strength gain was in-versely proportional to the clay content. Loess A-horizon soils had small strength gains, not directly related to clay content or other single variables.

6. The gumbotil and t i l l C-horizon Iowa soils treated with lime can be placed in two general strength categories on the basis of geological age. Relatively younger tills had far better maximum strengths than the lowan and older ti l ls and gumbotil. T i l l A-horizon soils gave generally low strengths.

7. It would appear that loess C-horizon soils of Iowa would better f i t a soil-lime design system for road construction based on particle-size distribution than one based on soil series. T i l l C-horizon soils of Iowa would seem to best f i t into a design system based on geological age. However, i t would seem that modification to f i t into a system based on soil series would be possible for both groups with further study.

8. Much further work would be needed to f i t the loess A- and B-horizon soils, t i l l A- and B-horizon soils and alluvial soils of Iowa into a soil-lime stabilization design system for road construction purposes.

ACKNOWLEDGMENTS The subject matter of this investigation was obtained as part of the research being

done under Project 449-S at the Iowa Engineering Experiment Station of the Iowa State Umversity of Science and Technology. Project 449-S is under contract with the Iowa Highway Research Board of the Iowa State Highway Commission.

The lime used in this work was generously furnished by the U. S. Gypsum Company, Chicago, 111.

The authors express their appreciation to members of the staff of the Engineering Experiment Station for their advice and assistance m this investigation and in the preparation of this paper.

REFERENCES 1. "Procedures for Testing Soils." American Society for Testing Materials, Philadel-

phia, (1955). 2. Chu, T. Y. , and E>avidson, D. T. , "A Simplified Air-Jet Dispersion Apparatus for

Mechanical Analysis of Soils." HRB Proc. 32:541-547 (1953). 3. Clare, K . E . , "Some Chemical Tests for Soil Engineering Purposes." Roads and

Road Construction, 27:43-49, 83-86 (1949). 4. Diehl, K . E . , and Smith G. E. , "Quantitative Analysis." WUey (1955). 5. Dockery, W.D. , and Manigault, D.E.H. , "Lime Stabilization and Low Cost Road

Construction." Roads and Streets, 90: No. 8, 91-95 (Aug. 1947). 6. Hilt, G. H. and Davidson, D.T. , "Lime Fixation in Clayey Soils." HRB Bull.

262, 20-32 (1960).

30

7. Button, C.E., "Studies of Loess Derived Soils in Southwestern Iowa." Proc., Soil Science Soc. of America, 12:424-431 (1947).

8. "Screenings from the Soil Research Lab. 1, No. 5." Iowa State University of Science and Technology. Engineering Experiment Station (Sept. -Oct. 1957).

9. Jenny, H . , "Factors of SoU Formation." McGraw-ffill (1941). 10. Johnson, A.W. , Herrin, M . , Davidson, D.T . , and Handy, R . L . , "Soil Stabiliza-

t ion." In Woods, K . B . , Berry, D.S., and Goetz, W.H. , (Eds.) "Highway Engineering Handbook." 1st Ed., pp. 21-1 through 21-133, McGraw-Hill (1960).

11. KatU, R.K, , Davidson, D.T . , and Sheeler, J .B. , "Water in Cutback Asphalt StabilizaUon of Soil." HRB BuU. 241, 14-48 (1960).

12. "Lime Subgrade Stabilization on Texas Interstate Projects." Roads and Streets, 10: No. 7, 75-76, 83-84, 87, 99 (July 1957).

13. McDowell, C., "Hydrated Lime for Stabilizing Roadway Materials." Roads and Streets, 92: No. 2, 81, 82, 84 (Feb. 1949).

14. National Lime Association, "Lime Stabilization of Roads." National Lime Assoc. BuU. 323 (1954).

15. Rose, A. C., "Public Roads of the Past." American Association of State Highway Officials, Washington, D. C.

16. SeU, O.E., "Lime-Fly Ash Ratio and Admixture Content Versus Strength of StabUized Sandy and Silty SoUs." Unpublished MS thesis Iowa State Univ. of Science and Technology (1957).

17. Simonson, R.W., Riecken, F .F . , and Smith, G.D., "Understanding Iowa SoUs." William C. Brown Co. (1952).

18. U. S. Department of Agriculture, Bureau of Plant Industry, Soils and Agricultural Engineering, Soil Survey Staff. "Soil Survey Manual." U. S. Department of Agriculture Handbook 18 (Aug. 1951).

Formation of New Minerals with Lime Stabilization as Proven by Field Experiments in Virginia JAMES L . EADES, F. P. NICHOLS, JR., and RALPH E. GRIM, respectively, Research Assistant, University of Illinois, Urbana; Highway Research Engmeer, Virginia Council of Highway Investigation and Research, Charlottesville; and Research Professor, University of Illinois, Urbana

In the fa l l of 1956 and the spring of 1957 subgrade soils on three projects in Virginia were stabilized with hydrated lime (Ca(OH)2). The projects, located approximately 150 miles apart, were constructed on three different soil types. The clay fraction of each of the three soils was composed of different clay minerals; therefore, there was a consider-able difference in the physical properties of the soils.

The three projects were sampled during 1960 for the purpose of studying the effects of addition of hydrated lime. X-ray diffraction data of the treated soils revealed that new minerals—calcium silicate-hydrates and calcium carbonate—had been formed. Petrographic analysis of thm sections prepared f rom the treated soils showed that the source of additional strength was a cementing ma-terial, hydrated calcium silicates, which now interlaces the soil grains.

• SINCE World War n much attention has been directed to the use of stabilizing agents to improve the engineering properties of cohesive soils. Lime (Ca(OH)3), one of the many chemicals tried, has proven to be an effective and economical additive for im-proving many soils used in road construction. Although lime is generally used to stabilize clayey type soils, such soils as micaceous silts and loess have been stabilized by the addition of lime.

The beneficial effects of lime are generally attributed to the interaction of lime and the clay minerals ui the soil. The reactions mentioned most often in the literature (1, 2) are (a) aggradation caused by flocculation of the clays, (b) cation exchange (that is, replacement of sodium, hydrogen or potassium ions by calcium), (c) cementing or bondmg action of indefimte character, (d) reaction of the lime and CO2 f rom the atmosphere to form calcium carbonate and thereby cement the soil particles together. It is believed ttiat any or all of these phenomena may occur under the right conditions and with the right soil.

Although Eades and Grim (3) in 1960 reported the formation of calcium silicate-hydrates from pure clay minerals in the laboratory, many engineers have questioned whether or not this same reaction occurred in the field. Thus the question of per-manency of lime stabilization stil l remamed. It is for this reason, as well as others, that data are needed concerning the true nature of the reactions of lime and soil under field stabilization conditions.

The Virginia Council of Highway Investigation and Research, the research division of the Virgima Highway Department, in its soils laboratory added lime to its soil stabilization program in 1954. In the early fa l l of 1956 and the spring of 1957, field experiments involving use of lime, along with other additives, for treatment of sub-

31

32

grade materials were constructed. The objective of the subgrade stabilization phase of the experiment was to determine the feasibility of improving the bearmg value of even the poorest subgrade soils to the extent that a minimum base thickness would prevent subgrade failure. The sections have been kept under surveillance since they were completed and periodic tests were made throughout the f i r s t year to determine strength characteristics and moisture-density changes.

During June 1960 the lime-treated sections of the two experimental projects, as weU as a lime-stabilized project built in 1957 under regular construction, were sampled so that mineralogical studies could be made. The aim of the study was to note any mineralogical differences between the control and the lime-treated sections. The three projects studied were built in three different soil areas. The clay minerals were radically different, so the engmeering characteristics were different.

DESCRIPTION OF FIELD EXPERIMENTS The two experimental projects are located in the Piedmont Province of Virginia.

One of the projects is located at the foot of the Blue Ridge Mountains m Patrick County on Route 58. The experimental sections, constructed with 3 percent and 5 percent lime, are 500 f t long. One group of sections was built in September 1956, a second group adjoining the f i rs t was completed in May 1957. The area is underlain by a deeply weathered micaceous schist, which seldom offers a clay horizon thicker than 24 in. Therefore, a good clay borrow material with which to cover the resilient micaceous material is scarce.

The second experimental project, also located in the Piedmont about 160 nu east of the experiments in Patrick County, is about 20 mi west of Petersburg on Route 460 in Dinwiddle County. The area is underlain by a deeply weathered granite, which has resulted in a nucaceous silty clay. On this short project the soil varied from A-2-4 containing an appreciable percentage of coarse sand to the highly plastic A-7-5 (20) soil according to the AASHO classification (4).

A third project, built by contract in a privately owned subdivision called Blrchwood Gardens, was located north of Route 58 in Princess Anne County, in the Coastal Plains Province of Virginia. The area is underlain by a heavy clay which classified as A-7-6 (12). A 6-in. 5 percent lime-stabilized subgrade with a 6-in. base was used in place of 12 in. of select borrow with a 6-in. base.

LABORATORY AND FIELD TEST RESULTS The California Bearing Ratio test, made on the subgrade soil, is used to aid m

determining the design of Virgmia pavements. Therefore, early evaluation of the suc-cess of lime for subgrade stabilization was obtained from CBR tests of the soil-lime mixtures, both in the laboratory and in the field.

The laboratory samples were compacted into standard CBR molds which had tight fitting lids with rubber gaskets to prevent moisture loss. This permitted the specimens to be given an accelerated oven cure at 140 F, usually for 7 days, after which the lids were removed and the specimens were given the usual 4-day soaking period prior to testing. Later some of the samples were tested after only a 3-day oven cure, inas-much as this seemed to produce better correlation between laboratory CBR and field CBR after 1 year. Table 1 presents data on laboratory CBR values for one typical sample f rom each of the three field projects.

During and for some time after construction of the lime-treated subgrades, in -place CBR tests were made in the field to record the progressive effects of the lime.

The field CBR test is known to have distinct limitations. Attempts to correlate i t with the standard laboratory CBR test have generally proven unsuccessful. Field test values may vary unpredictably with variations in moisture content and percentage d coarse aggregate in the soils. However, this much can be said of CBR values of un-treated subgrade soils: experience gained from testing a number of projects in Virgima has indicated generally that the field CBR value tends to go down as the moisture con-tent of the soil reaches equilibrium. For a good many soils, this in-place value a year or two after construction has been found to be lower than the laboratory soaked value for the same soil.

33

TABLE 1 RESULTS OF LABORATORY CBR TESTS

CBR ii) Soil Raw Sou + Soil + SoU +

Source Class. Sou 3i Lime^ 5̂ Lime* 55SLime*

Dinwiddle Co. A-5-(2) 2.5 - 99.0 -Patrick Co. A-4-1 8.8 39.0 110.0 66.0 Princess

58.0 Anne Co. A-7-6 (12) 6.0 42.0 87.0 58.0

1 Cured 7 d ^ s at lliO F. ^Cured 3 days at lUO F.

For those field CBR tests made on the lime treated subgrades, however, the m-dications are that the treated soUs tend, rather, to gain strength with age. Figures 1 and 2 show the strength gained over periods up to 1 year on the experimental projects on Routes 58 and 460.

Crude attempts were made to secure a later series of field CBR data as late as 1960, but the results were of questionable value because only 6-m. auger holes were available through which the tests might be made, and the CBR piston could not be assured a positive and uniform contact with the subgrade. In spite of this, the average field CBR values remained in approximately the same neighborhood as those determined earlier from test pits and shown in Figures 1 and 2.

X-RAY DIFFRACTION ANALYSIS Samples were taken from each of the auger holes made in 1960 for mineral analysis.

As the samples were taken from the subgrade they were sealed in plastic bags to pro-tect any unreacted lime from carbonatmg until they could be transported to the labora-tory.

Portions of the samples were ground to pass a 200-mesh screen and were analyzed by X-ray diffraction. When the diffraction tracings of the treated soils were compared with those from the control sections i t was evident that a reaction had taken place form-ing new compounds. Closer analysis showed that calcium carbonate was present and also some additional lines. The next step was to identify the compounds causing these Imes, and to determine the percentage of carbonate present.

The soils f rom the projects prior to stabUization did not contain calcium carbonate; therefore, it was a simple matter to determine how much of the lime had carbonated. Chemical analysis showed that the samples from the 5 percent treated sections now contained about 2.5 percent calcium carbonate.

To better determine the reaction products of the soU and Ume, samples of the soU from the control sections were treated with 20 percent lime and the mixtures were cured at 140 F for 60 days m sealed containers.

When these laboratory-treated soils were analyzed it was learned that calcium silicate-hydrates had formed. By comparing the X-ray tr