— .,.,.....— ~ ‘ !. PUBO1452.Pdf )“ : . ,.. .,,<... .. . ... ,*-, .. .: ,. .“- 1 I .. — The shrinkage by H. E. Bofinger, H. O. Hassan and R. 1.T. Williams THE SHRIN~= OF FINE-~I~D SOIL-~~NT by H E Bofinger, H O Hassan and R I T Williams The incidence of cracks not associated with traffic loads is a feature of roads with soil-c=ent bases. The consequent problem of such cracks reflecting through a bitminous road surfacing has caused concern for many years. The cracks result from restrained thermal and shrinkage movements but, when the stabilized soil is fine-grained, it is believed that this cracking is mainly attributable to shrinkage. An examination of the shrinkage behaviour of soil-caent mixed from Littlehmpton brickearth was carried out to provide a better understanding of one of the major c~uses ‘ of cracking in soil-cement road bases. The effects of the following factors on the shrinkage of this mixture were investigated in this study: i) the method of compaction used to mould the specimens ii) the cement content of the mixture iii) the moisture content of the mixture when the specimen is moulded iv) the average density of the specimens v) the moisture content of the soil immediately before it is mixed with cement and the moulding water vi) the effect of drying after a period of autogenous curing. Three methods of compaction were used to mould specimens, nmely static compaction, kneading compaction and dynamic compaction using the British Standard laborato~ compaction rammer. The shrinkage of specimens moulded by static compaction was measured parallel to the direction of the compactive force and perpendicular to it. When the other methods of compaction were used, the shrinkage was only measured parallel to the direction of the compactive force. The specimens were sealed to prevent loss of moisture thereby simulating ideal conditions of curing. The main conclusions drawn from this investigation are listed below: 1. The shrinkage of laborato~ specimens is anisotropic and is also markedly influenced by the method used to compact them. Hence the prediction of the shrinkage behaviour of full-scale pavements on the basis of tests on laboratory specimens must take accout of the methods used to prepare these specimens, and the direction in which the shrinkage is measured. 2. Autogenous shrinkage in the direction perpendicular to the compacting force reduces when higher proportions of cement are added to clays. It is this type of laboratory test that most closely simulates the structure of a soil-cement pavment layer. 3. Volme changes in clay-cement are thought to be caused by the interaction of soilmoisture suction, ~e re-orientation of water adsorbed on the clay particl-es, the expansion of cement gel as it hydrates, the self desiccation caused by the hydration of the cement and the increase in the strength of the cemented soil skeleton. 4. Shrinkage of clay-cement specimens is profoundly affected by the initj.al condition of the soil prior to moulding, by the moisture content at which they are compacted and by their final density. \ I“’!The shrinkage will be minimised if the soil is processed from a dry state and all the water required for compaction and hydration is added during mixing. In addition, to minimise shrinkage, the mixture should be compacted as quickly as possible to the minimm density that will ensure.that the material will attain adequate strength. 5. If clay-cement specimens are exposed to a drying atmosphere, their shrinkage potential is increased but the longer the period of time they are kept sealed before drying commences, the smaller is the total shrinkage. The work described in this Digest forms part of the programme carried out by the Overseas Unit of T~ for the Minist~ of Overseas Development but any views expressed are not necessarily those of the Ministry. If this information is insufficient for yow needs a copy of the full report, SR398, may be obtained on vritten request to the Technical Information d Libr~ Sewices, Transport and Road Reseoch Laboratory, Old Wokinghm Road, Crmthorne, Berkshire. Crown Copyright. Extracts from the text may be reproduced, except for commercial purposes, provided the source is acknowledged. 6m 4.78 T.A. Ltd. 1093 Printed in Englend. TRANSPORT and ROAD RESEARCH LABORATORY Department of the Environment Department of Transport SUPPLE~NTARY REPORT 398 THE SHR~~m OF FINE-~INED SOIL-CE~NT by H E Bofinger, H O Hassan and R I T Williams The work described in this Report forms part of the progr-e carried out for the Ministry of Overseas Development, but any views expressed are not necessarily those of the Ministry. Overseas Unit Transport and Road Research Laboratory Crowthorne, Berkshire 1978 ISSN 0305-1315 CO~E~S Mstract 1. Introduction 2. Previous research 3. Determination of shrinkage 4. Materials and laboratory procedures 4.1 Materials 4.2 Preparation of the soil 4.3 Preparation of specimens 5. Results 5.1 The effect of the 5.2 The effect of the 5.3 The effect of the 5.4 The effect of the method cement of compaction content moulding moisture content level of compaction 5.5 Influence of the pre-treatment moisture content of the soil 5.6 The influence of drying 6. Supplementary tests 7. Discussion of results 7.1 Pore pressure and suction in unstabilized clay 7.2 Changes in the structure of the water in the clay 7.3 The effect of additions of cement on the behaviour of untreated clay 7.4 The shape of the shrinkage-time curve 7.5 Influence of the method of compaction on the magnitude of shrinkage 7.6 The effect of the cement content 7.7 The effect of the moisture content at which the soil-cement is moulded 7.8 The influence of the density 7.9 The effect of the pre-treatment moisture content 7.10 The effect of drying 8. Implications for road pavement layers 8.1 Cement content 8.2 Moisture content 8.3 Compaction 8.4 Curing Page 1 1 3 4 5 5 7 7 8 8 10 10 11 12 13 13 14 14 15 16 17 18 19 20 20 20 21 22 22 23 23 23 9. Conclusions 10. Acknowledgements 11. References 12. Appendix (C) CRO~ COP~IGHT 1978 Page 24 25 25 28 Extracts from the text may be reproduced, except for commercial purposes, provided the source is acknowledged. THE SHRI~~ OF FINE-~INED SOIL-CE~NT ~STRACT The incidence of cracks not associated with externally applied loads is a feature of soil-cement roadbases, and the consequent problem of reflection crackin~ through the superimposed surfacing has caused concern over the years. The cracks result from restrained thermal and shrinkage movements but, when the sttiilized soil is fine-grained, it is believed that the cracking is mainly attributable to shrinkage. The paper describes a study of the autogenous shrinkage of such a material aimed at improving the understanding of its susceptibility to cracking. In the investigation, specimens were cured under ideal conditions, but, because conditions in the field are often far from ideal, the influence of drying on shrinkage was also studied. The principal finding of the investigation is that the measured value of shrinkage is greatly affected by the test conditions imposed. In particular, major differences were observed in specimens compacted by different methods and, furthermore, the mode of compaction had a striking influence on the way in which the cement content of mixes influenced shrinkage. A tentative explanation for the results obtained is present(>d in terms of the pore pressure in the specimens, the hydration of the cement and the particle orientation. It is clear, however, that further work is necessary before the general validity of the results can be determined. 1. INTRODUCTION In a ntier of areas of the world traditional road building materials of adequate quality are not readily available within a reasonable haulage distance of road construction projects. In such situations the stabilization of local soils and gravels can produce an economical material for building roadbases. Cement has been widely used as a stabilizing agent for locallyoccurring materials but cement-stabilized road-bases have a reputation for cracking. Cracks formed in the base often reflect through the bitm.inous surfacing unless a considertile thickness of bituminous material is provided. The extent and effects of these cracks concern engineers because they may seriously affect the performance of roads if they permit the ingress of water to the sub-base and sub-grade. Many factors influence the degree of cracking of soil-cement in terms of the spacing, direction and width of cracks. Among the more important are: 1 (i) the susceptibility of the material to shrinkage as a result of moisture loss or the internal redistribution of moisture as the cement hydrates, (ii) the contraction of the material as the ambient temperature falls, (iii) the restraint imposed on the roadbase by the sub-base or (iv subgrade, the tensile properties of the hardening soil-cement in terms of its strength, its strain capacity, modulus of elasticity and the ability to The extent to which the surfacing depends on relieve critical conditions through creep. cracks in the soil-cement ultimately reflect through a ntier of factors but principally on the type and thickness of the bituminous surfacing, the effect of climate and age on the viscosity of the binder, and the volume and magnitude of the axle loads of commercial vehicles using the road. (However, not all cracks in bituminous surfacings are attributable to cracks in the base) . This paper considers one factor only that contributes to cracking, namely the shrinkage characteristics of soil-cement, which are of particular 1,2,3 importance in the case of fine-grained soil-cement . The ideal curing condition which allows soil-cement to achieve the maximum increase in strength and to minimise cracking due to differential drying shrinkage is to enclose totally the soil-cement with an impermeable material so that no moisture can escape. Under these ideal conditions the shrinkage that occurs has been called ‘autogenous’ shrifiage. At the other extreme, when moisture is allowed to dry freely from soil-cement, ‘drying’ shrinkage occurs and the potential strength of the material is then substantially reduced and the likelihood of cracking is increased. Desirably, the curing conditions for soil-cement road bases should approach as closely as possible the conditions of autogenous shrinkage. In practice, however, this is difficult to achieve, and frequently inadequate attention is paid to retaining moisture in the. material. 2 To provide a better understanding of a major cause of cracking in soilcement roadbases, an investigation has been undertaken of the shrinkage properties of clayey soil-cement. The following factors (which affect the shrinkage of soil-cement investigated. in this study: were (i) the (ii) the (iii) the (iv) the (v) the method of compaction used in moulding the specimens, cement content of the mixes, moisture content of the mixture when the specimen is moulded, average density of the specimens, moisture content of the soil immediately before it is mixed with cement and the moulding water, (vi) the effect of drying after a period of 2. P~IOUS ~S~CH autogenous curing. A number of laboratory studies of the shrinkage of soil-cement under various conditions have been conducted, some of the more important being those undertaken by Nakayama and Handyl, George2, Wang*, Pretorius and Monismiths and Dunlop6. Shrinkage of soil-cement is generally held to be caused by loss of water due to self desiccation and evaporation. Evaporation from unprotected surfaces causes the largest changes in volume but good curing practices can minimise this 10ss. Self desiccation causes volume changes of a smaller magnitude Thermal shrinkage is believed to be insignificant in soil-cement, George2. 7 Plummer and Dore have suggested that the drying shrinkage in bodies which undergo volume changes smaller than the corresponding loss in the volume of moisture is caused by a capillary phenomenon and, therefore, that the shrinkage obeys capillary tension theory. Means and Parcher8 and Czern’ing have associated the drying shrinkage of cement theog. When cement paste is cured under sealed conditions paste with the same 10 it first expands but after some time, recrystallization occurs which results in a loss of intracrystalline or adsorbed water, thus reducing the volume of the material 11 Bernal . 3 George2 has proposed three hypotheses to explain drying shrinkage in soil-cement. At high humidities he suggests that the capillary tension effect predominates whereas at intermediate humidities, the decrease of the adsorbed water film causes contraction and, when the humidity is reduced further, shrinkage is due to the loss of water from the crystal lattice of the soil. 12 In.,alater study, Wang and Kremmydos proposed that the lattice shrinkage in clay would occur as soon as evaporation begins, contrary to George’s third hypothesis. 13 Expansion has been measured in sealed specimens of sand-cement . These mixtures are composed of volumetrically stable sand particles surrounded by a cement gel which of the specimen. follows that the expansion of the expands when it hydrates, thus causing the overall expansion If the gel in clay-cement mixtures expands similarly, it shrinkage within the clay fraction must be greater than the cement gel for autogenous shrinkage to occur. 3. DETEWINATION OF S~INKAGE In this Paper the term ‘shrinkage’ is used to describe autogenous shrinkage or shrinkage under sealed conditions. When the specimens are permitted to dry the term ‘drying shrinkage’ is used. One of the important aspects of this study was the examination of the anisotropic nature of the shrinkage of soil-cement. Most researchers who have studied the shrinkage characteristics of soil-cement have measured the axial deformation of cylindrical specimens compacted in moulds developed for producing compression strength test specimens. Specimens prepared and monitored in this way do not simulate the critical shrinkage conditions experienced by soil-cement in roadbases since the shrinkage measured will reflect the vertical shrinkage of roadbases. %rinkage in this direction, which causes a change in the thickness of the layerl is probably not very relevant to the formation of the cracks that significantly affect the performance of soil-cement bases. In this study cylindrical shrinkage specimens were prepared using static compaction, kneading compaction and impact compaction, and the shrinkage was measured in the direction in which the compacting force was applied. In addition, prismatic specimens were compacted statically and the shrinkage was measured at right angles to the direction of the compacting force. 4 Most previous researchers in this field have concentrated on the measurement of the drying shrinkage of specimens initially cured in a moist condition for some period of time. This approach does not enable measurements to be made of the volume changes that occur immediately after compaction is completed or during the early stages of the curing period. The behaviour of soil-cement during this early stage of curing may have an important effect on the subsequent performance of tie material. In this investigation, shrinkage measurements were therefore commenced immediately after the compacted specimen had been sealed 4. MATERIALS 4.1 Materials in wax. ~D LABORATORY PRWEDURES The majority of the tests in this investigation were conducted on specimens prepared from Littlehampton brickearth. In addition two other types of soilwere used to check certain of the results obtained with the brickearth. These additional soils were Bagshot sandy clay and Wainscott brown clay. The properties of all three soils are shown in Figure 1 and Table 1. T&le 1 also includes test results of the ordinary Portland cement used in all the tests. 5 Soil property 1. Particle sizes Sand % Silt % Clay % 2. Specific gravity 3. Atterberg limits Liquid limit Plastic limit Plasticity index Compaction: (i) B.S. light(14) Max. dry density lb/ft3 Mg/m3 Optimum moisture content 14 (ii) B.S. heavy Max. dry density lb/ft3 Mg/m3 Optimum moisture content California bearing ratio (c.B.R.) % Linear shrinkage % TABLE la Properties of soils Littlehampton brickearth (L.B.E.) 23 51 26 2.73 39 19 20 111 1.78 17% 123 1.97 13% 13 12.4 Bagshot sandy clay (B.S.C.) 45 18 37 2.72 34 18 16 112 1.79 15% 123 1.97 12% Wainscott brom clay (W.B.C.) 12 39 49 2.73 75 31 44 97 1.55 21% 16 Properties of the cement (BS12)15 Setting times Water 26.3% Initial 120 min Final 165 min Fineness Specific surface 348 sq m/kg Expansion Le Chatelier 2.5mm Compressive strenath . Vibrated mortar cubes 3 days 31 ~/m2 7 days 45 ~/m2 4.2 Preparation of the soil The clays were air-dried and then (2.36 mm) sieve. Water was mixed with moisture content to a level 2 per cent pulverised until they passed the No 7 the pulverised soil to raise its below the optimum for compaction. The moist soil was then sealed in containers and stored for a minimum period of one month to allow the moisture to distribute uniformly throughout the soil. 4.3 Preparation of specimens Pre-wetted soil and cement were mixed, the additional moulding water added, the whole thoroughly mixed and then moulded into specimens as quickly as possible. Three shapes of specimens were used: (i) 4,, x 211 @ cylinders (101.6 X 50.8 mm) (ii) 4.6” x 4“ @ cylinders (116.6 X 101.6 m) (iii) 1+” x 1+” x 6“ bars (38.1 X 38.1 X 152.4 mm) The 4“ x 2“ cylinders were moulded by vertical kneading compaction using a plunger similar to static compaction or by the type developed for the Havard miniature compaction apparatus. The 4.6” x 4“ ~diame.tercylinders were compacted in accordance with the BS 2.5 kg rammer method (BS 1377-1975)14- The 1+” x 1+” x 6“ bars were moulded horizontally by static compaction in three equal iayers, each layer being compacted separately under a pressure of 227 psi (1560 kN/m2). This pressure was chosen because it produced approxi- (14) mately the same dry density as BS compaction with the 2.5 kg rammer . Immediately after they had been compacted, the specimens were mounted on carriers, s’ealed in wax, set up on stands, and the shrinkage measured by dial gauges reading to 0.001 mm as shown in Plate 1. Shrinkage measurements were made at fixed time intervals over a period of 28 days, during which time the temperature of the specimens was maintained at 2o”f20C. At the end of the 28 day testing period, the weight of each specimen was that the wax-seal was intact. men drying shrinkage measurements were required, shrinkage was first measured during the initial period wax was then removed and the drying shrinkage measured temperature. 5. ~SULTS remeasured to check the autogenous of moist curing, the at the same ambient The results of the shrinkage tests are shown in Figures 2 to 5 and 7 to 15, in which shrinkage strain is plotted against time. Each curve is plotted through the average points obtained from five separate tests, and examples of the variability of the complete results of typical groups of tests are given in tie Appendix. 5.1 The effect of the method of compaction Figures 2 to 5 show the results of shrinkage tests carried of Littlehampton brickearth stabilized with various percentages out on samples of cement. 8 Specimens were moulded in four different ways: by vertical static compaction in 4“ x 2“ cylinders, by horizontal static compaction in 1+” x 1+” x 6“ bars, by kneading compaction in 4“ x 2“ cylinders and by impact compaction in a 14 BS standard compaction mould . For each oE.-.thesetests the total. shrinkage strains after twenty-eight days are listed in Table 2 overall pattern of behaviour. The development of shrinkage with time follows a to summarise the similar pattern for each type of specimen and for most values of cement content, except for specimens with cement contents of 10 and 15 per cent moulded horizontally by static compaction in which some expansion occurred. Generally, the rapid shrinkage during the first few hours was followed by a null zone which contin– ued for up to 1 day before further shrinkage occurred. It is clear from Table 2 that the method of compaction influence onthe magnitude of shrinkage, the greatest amount has a major of shrinkage being measured on specimens moulded by impact compaction and the smallest amount by kneading compaction. Specimens moulded by vertical static compaction and horizontal compaction have a similar magnitude of shrinkage. TABLE 2 Total shrinkage strains x 10-6 after 28 days of curing for specimens moulded at O.M.C. to B.S.light density by different methods of compaction I I Cement content % Soil Method of compaction o 4 6 8 10 I 15 Vertical static compaction (cylinders) 1850 1300 300 400 600 - Little- ~eading compaction 960 50 500 - - hampton brickearth Dynamic compaction - 4500 3900 4090 - - Horizontal static compaction (bars) 2100 1230 640 580 440 330 Vertical static compaction Bagshot Sandy (cylinders) 1250 850 950 - - Clay Horizontal static compaction — .- (bars) 1450 1350 950 - - Vertical static compaction Wainscott Brown (cylinders) 2900 1900 3250 - - Clay Horizontal static compaction (bars) - 4150 3850 3500 - - 9 5.2 The effect of the cement content The effect of the cement content on the shrinkage behaviour was influenced by the method of compaction. When the specimens were moulded in vertical cylinders, either by static compaction or by kneading compaction, there was a critical cement content at which the shrinkage was a minimum. In contrast, specimens that were moulded horizontally by static compaction showed a progressive reduction in shrinkage when the cement content was increased. At high values of cement content there was an initial expansion rather than shrinkage. Finally, when specimens were compacted by the standard impact hammer, the results (Figure 5) are less well defined but they suggest that the shrinkage is relatively insensitive to the cement content. It is evident from the shrinkage of sealed specimens which contained cement that the addition of cement to clayey soils reduced the autogenous shrinkage (Figure 6) . This is in contrast with the widely held view that shrinkage will be reduced by reducing the cement content. The results obtained from specimens moulded horizontally by static compaction (which most closely model conditions in roadbases) thus differ from vertically compacted specimens, and , thereby, from previously no published results which have been obtained using cylindrical ‘compression’ specimens. To determine whether this effect was peculiar to specimens moulded from Littlehampton brickearth, comparative tests were made with Bagshot sandy clay and Wainscott brown clay. The results of these tests are shown in Figur~7 to 10. It is clear from these figures that the reduction in shrinkage with increasing cement content for specimens moulded horizontally by static compaction is not peculiar to soil-cement made from Littlehampton brickearth. 5.3 The effect of the moulding moisture content The influence of the moisture content of the moulded mixture on the shrinkage of soil-cement was studied on specimens which were moulded vertically by static compaction at 2 per cent above and 2 per cent below 14 the optimum moisture content for BS compaction (2.5 kg rammermethod ) . The target and groups tests were density was the maximum density for BS compaction (2.5 kg rammer) of specimens were moulded at various cement contents. Additional conducted on specimens containing 8 per cent cement, moulded 10 / horizontally by static compaction at the same moisture contents and density. The values of total shrinkage measured after 28 days of curing are summarised in Table 3. TABLE 3 -6 Total 28-day shrinkage strain x 10 at various moisture contents Type of compaction Vertical Static Compaction Horizontal Static Compaction Moisture contient Cement content 2% Ii 2% % below OMC above OMC OMC 4 I 740 I 1300 I 1590 6 I 280 I 300 I 320 8 420 410 450 10 560 600 650 8 460 580 700 I S.4 The effect of the level of compaction A series of tests was conducted to assess the effect that the magnitude of the compacted density has on the shrinkage characteristics of soil-cement. In the initial tests, specimens containing various percentages of cement were moulded vertically by static compaction at the optimum moisture content and to the maximum dry density for the BS 4.5 kg rammer method for compaction. TWO additional sets of specimens containing 8 per cent of cement were moulded horizontally at the same optimm moisture content, to the same density and to 95 per cent of that density. The results are shown diagrammatically in Figuresll and 12 and a summary of the total shrinkages after 28 days is included in Table 4, together with values obtained from specimens moulded at the optimum moisture content and to the maximum dry density for the BS 2.5 kg rammer method. il TABM 4 Total 28-day shrinkage strain x 10 -6 at different densities and optimum moisture contents Moulding conditions Cement Type of BS light BS hea~ 95% of content density density BS heavy compaction % and OMC and OMC and 13% (17%) (13%) moisture content 4 1300 1700 Vertical Static 6 300 1850 Compaction 8 410 1950 10 600 2150 -’ Horizontal Static Compaction 8 580 2100 1200 There was a marked increase in the shrinkage at the high density, even though the heavier compactive effort is associated with a lower optimum moisture content. It should be noted, however, that when the density was varied without changing the moisture content, shrinkage was greater at the higher density. The shrinkage behaviour of the more dense specimens differed in other respects too. In specimens moulded vertically by static compaction, the shrinkage increased with increasing cement content and it continued to increase progressively with time without going through a ‘null’ period. Specimens moulded horizontally by static compaction also exhibited this latter type of behaviour. 5.5 Influence of the pre-treatment moisture content of the soil It was expected that the moisture content of the soil immediately before it was mixed with cement and moulding water would affect the shrinkage of soil-cement. To investigate this, a small study was conducted on specimens containing 6 per cent of cement which were compacted vertically by static compaction. The results of these tests are shown in Figure 13 and they indicate that the pre-treatment moisture condition has a marked effect on the shrinkage 12 / of soil-cement. men oven-dry soil was used, the specimens developed a small shrinkage strain during the first few hours followed by expansion for up to 7 days before they started to contract again. Specimens prepared from soil which was pre-wetted to the optimw moisture content showed progressive shrinkage with time. men the soil was used at the standard pre-treatment moisture condition of 2 per cent below the optimum moisture content for compaction, the ‘typical’ null period was observed for curing periods of 2 hours to 1 day. 5.6 The influence of drying The major part of this investigation was directed towards the autogenous shrinkage of soil-cement. However, reported studies on specimens which time and then were exposed to allow treated similarly for comparison. because a number were moist cured drying to occur, of researchers have for a short period of some specimens were Specimens containing 8 per cent of cement were moulded vertically by static compaction and then immediately sealed in wax for periods of 1 hour, 1 day and 7 days. The wax seal was then stripped, the drying shrinkage observed and the results compared with the shrinkage of specimens which were sealed in wax for the full period of the test. The results of these tests are shown in Figure 14. It can be seen that, regardless of the initial period of moist cure, almost all of the drying shrinkage occurs within 7 days and its magnitude is reduced as the period In all cases the drying shrinkage is much greater shrinkage. 6. SUPPLE~~MY TESTS of moist cure is increased. than the autogenous The results obtained from the foregoing studies could not be entirely explained by the hypotheses proposed by previous research workers. Accordingly further tests were undertaken in an attempt to provide an explanation for the unexpected shrinkage behaviour that was observed. men the compactive force is removed from an unsaturated clay specimen, 16,17 suction or negative pore pressure is induced (Lambe 18 , Aitchison ). Furthermore Lambe has suggested that surface chemical phenomena give clay the capacity to imbibe water and if the water content of a clay mass is less than this capacity, a water deficiency will exist, increasing the magnitude 13 of the suction. The intergranular stress will be increased by this suction, causing consolidation in the material. The pattern of shrinkage behaviour that was measured on specimens moulded by each of the methods of compaction suggest that the pore pressure condition contributes to the rapid initial shrinkage. Tests were therefore carried out to assess the development and dissipation of pore pressure during and after the compaction of tiree specimens, Qne containing no cement and the other two containing 8 per cent of cement. The specimens were moulded at the optimum moisture content for BS compaction with the 2.5 kg rammer and also at two per cent above this value for specimens containing cement. A small piezometer was placed in the middle of the specimens which were compacted vertically by static compaction. The pore pressure of the specimens was measured for 24 hours and the results are shown in Figure 15. The limited number of tests made do not permit quantitative conclusions to be drawn but it is clear tiat during the compaction phase a positive pore pressure developed in all three specimens. This dropped to zero within 9 minutes of the specimens being extruded. Thereafter, a pore suction was measured and this remained until the end of the period of observation of 24 hours, although, in each case, the maximum pore suction was measured after 5 to 6 hours. 7. DISCUSSION OF ~SULTS 7.1 Pore pressure and suction in unstabilized clay An unexpected result from the investigation was the magnitude of the shrinkage of sealed specimens of untreated clay and the fact that this shrinkage was still increasing at the end of the observation period of 28 days. The pore suction that develops in the specimens shortly after they are extruded from the mould would be expected to contribute to their shrinkage only during the initial life for the following reasons:- 1. men the soil skeleton is stijected to constant additional stress, the maximum drainage path in specimens of this size is 1 inch (25.4 ~) . Based on standard ltioratory consolidation tests, one would expect 14 \ , / that the primary consolidation of these specimens should effectively be completed within 24 to 48 hours. 2. At any age of 5 to 6 hours the pore suction in the specimens reaches a maximum after which it tends to reduce towards zero. It follows that intergranular stress due to pore suction will fall after about 6 hours and the specimens could possibly rebound if the suction reduces sufficiently, resulting in an increase rather than a decrease in their volume. It is suggested that the development and dissipation of pore suction in untreated clay will cause shrinkage ,of the form shown diagrammatically in Figure 16 and hence that another mechanism must contribute to the shrinkage observed, particularly over periods of time greater than 24 hours after the material has been compacted. 7.2 Changes in the structure of the water in the clay The water in clays is not always in a liquid state which is character19 20 ised by a complete lack of orientation or structure. Rosenquist , Low and others have suggested that the water immediately surrounding the clay (21) particles has an orientated structure but researchers cannot agree whether the density of this non-liquid water is greater or less than 1.0. Grim21 hypothesised that when clays are in an ‘undisturbed’ state, the non-liquid water contributes to their shear remolding he postulated that the structure disrupted, thus reducing the shear strength phenomenon of ‘sensitivity’. strength. During the process of in the orientated water is of the clay and causing the The change from non-liquid to liquid water during the remolding process is not irreversible. Disturbed soils gradually regain strength and eventually return to the ‘undisturbed’ state. This change in state suggests a mechanism for the longer term shrinkage of unstabilized clay specimens. men water is added to adjust the moisture content of a sample of soil and it is mixed and compacted, tie soil is in a highly disturbed condition. In a recently moulded specimen, tie water immediately surrounding the clay particles gradually becomes re-orientated, a condition in which it probably 15 21 has a higher density and, therefore, a lower volume . The gradual reduction in the volume of water in the clay causes it to shrink slowly and this process continues for some time as the clay tends to resume the undisturbed state. A possible timing for the shrinkage associated with this process is illustrated in Figure 17, but obviously the hypothesis implies that the density of nonliquid water is greater than 1.0. 7.3 The effect of additions of cement on the behaviour of untreated clay men cement is added to the soil, the material is modified in several ways. The first effect is an increase in the strength and volume stability of the soil skeleton. In the first few hours this effect is produced by the cement flocculating the structure of the clay fraction but the more permanent and more significant effects are attributable to the hydration of the cement which occurs over a much longer period of time. After the initial set has occurred in the cement paste, a more rigid soil structure develops and continues to increase in strength and rigidity as the hydration progresses. The pore suction which develops in a specimen immediately after it is extruded would cause less ‘consolidation’ shrinkage in the stronger, stabilized specimen than occurs in the untreated material. Likewise, less shrinkage can be expected to be caused by re-orientation of the water immediately surrounding the clay particles in the stabilized material. 22 According to Taylor , ordinary Portland cement contains approximately 45 per cent of tricalcium silicate (C3S) and approximately 27 per cent of dicalcium silicate (C2S). The C3S hydrates relatively quickly while the C2S takes a considerable time, suggested values being some 10 hours and 1000 hours respectively. In sealed specimens the water for hydration is drawn from the moist clay and as hydration proceeds, the amount of adsorbed water in the clay will steadily decrease, tius increasing the suction in the pores and consequently, the intergranular stress. Sherwood23 has shown that when Harmondsworth brickearth is stabilized with 10 per cent of cement, approximately 13 per cent of water (based on the dry weight of cement in the mixture) will be used in the hydration reaction within the first 7 days. This increases the pore suction by 1.09 and there is a similar increase in the intergranular stress. men the cement content is lower there would be a smaller but still significant increase in the intergranular stress. 16 Littlehampton brickearth is generally similar and soil-cement specimens moulded from it would be similar increase in the intergranular stress while thereby causing consolidation or shrinkage. to Harmondsworth brickearth expected to exhibit a hydration is proceeding, Another important effect is the expansion in the cement gel while it hydrates. Tests on cement pastes 10,11 have shown,that expansion ceases after a certain time and the hydrated gel starts to shrink as recrystallisation occurs. A similar type of behaviour was observed when sealed specimens of cement paste were moulded at 17 per cent moisture content to the same density as the soil-cement specimens. The maximum value of expansion occurred after 1 day of curing and this was followed bya reduction in the volume of the 24 paste . 13 Bofinger and Duffell showed that sand-cement mixtures, which have volumetrically stable soil particles, expand when they are sealed and do not subsequently shrink. This is probably because the strength of the cemented sand matrix is sufficiently large to resist the shrinkage stresses in the paste after one day. 7.4 The shape of the shrinkage-time curve The results of the experiments clearly indicated that the four methods of compaction used in the study produced specimens that exhibited different magnitudes of shrinkage but which had similar time-dependent patterns of shrinkage behaviour. During the first phase, which lasts for up to one day, the rate of shrinkage is rapid and is thought to be dominated by the pore suction that develops when specimens are extruded. As the pore suction dissipates, the resultant shrinkage will also diminish and within 24 hours this process will be complete. TWO other factors, namely the pore suction caused by the hydration of the cement, and the expansion of the cement paste, will contribute to the overall dimensional changes during this first phase. However, their influence is thought to be less significant than the pore suction developed after extrusion, although further work is required in order to confirm this. 17 The initial rapid shrinkage is followed by a null zone’which lasts for up to 7 days during which it seems likely that the increasing pore suction caused by the hydration of the cement tends to counterbalance the dissipation of the pore suction developed during the extrusion process. After approximately 10 hours the consumption of water in the hydration reaction is well advanced but the water will be taken first from the larger pores, thus causing only a relatively small increase in the pore suction which is insufficient to overcome the strength that the soil structure has attained at that time. As hydration proceeds and water is removed progressively from finer and finer pores, the suction will build up sufficiently to cause the third stage of shrinkage which commences after approximately 7 days of moist curing. 7.5 Influence of the method of compaction on the magnitude of shrinkage The method of compaction was found to have a major influence on the magnitude of shrinkage, the largest amount of shrinkage being measured on specimens moulded by impact compaction. It is likely that the mode of compaction tifects both the structure in the material and the Pore suction when the specimen is extruded. It is expected that the pore pressure developed during in consequence, the pore suction developed after the sample will be greatest when a standard Proctor hammer is used and samples are moulded by static compaction. During the first compaction and, is extruded, least when the 24 hours, therefore, specimens moulded by Proctor hammer will probably shrink more than specimens moulded by kneading or static compaction. 25 El Rawi et al found that the method of compaction influences the magnitude of cohesion in soil-cement specimens as well as its rate of increase with time. Specimens ~at were moulded by kneading compaction had higher cohesion values than those moulded dynamically. It can be surmised that the greater the cohesion, the greater will be the stresses that develop in the soil-cement. The difference in the total shrinkage of resistance to shrinkage specimens moulded statically in the vertical and horizontal directions supports the view that there is some degree of anisotropy in the specimens, particularly in those moulded from Wainscottbrown clay. Of more practical significance is the relationship 18 between the cement content and shrinkage of specimens moulded in each direction. If randomly orientated plate-shaped particles are compact tend to align themselves at right angles to the direction in which th compacting force is applied and tend to adopt a structure similar to ordered pile ofcards (see Figure 18) . men a suction is developed in in such a structure, movement of the particles can be achieved more e in the horizontal direction than in the vertical direction. One woul therefore, that the shrinkage measured at right angles to the compact force would be greater than in the direction of compaction. In a roadbase, the compacting force is at right angles to the di in which the shrinkage is critical, ie, the behaviour in the field is closely modelled by the horizontally compacted specimens. The result tests on horizontally-moulded specimens should therefore be considere carefully when the behaviour of a soil-cement roadbase is being predi from laboratory tests. The method of compaction influences the pore suction, the cohesi the orientation of particles, and thus contributes to the shrinkage c teristics of &oil-cement specimens. It is suggested that they provid insight into the large differences in the magnitude of the shrinkage ( cement specimens compacted by the four methods studied in this invest 7.6 The effect of the cement content Variations in the cement content affected the shrinkage characte: of the different types of specimens in different ways. For vertically compacted specimens moulded s~tatically or by kneading, there was a cr cement content at which the autogenous shrinkage was a minimum. Georl noticed a similar trend in the drying shrinkage of soil-cement. No completely satisfactory explanation can be offered for this behaviour especially since it is completely different from the behaviour of spe~ moulded horizontally by static compaction which displayed lower shrinl with progressive increase in cement content. Furthermore, the shrinki specimens compacted dynamically was found to be insensitive to cement although the compacted in content. 19 conditions leading to the relatively large shrinkage this way may havem’asked the effect of differences in of : cem{ An increase in the cement content will increase the volume of expanding gel and provide a stronger and more stable soil structure to resist the higher pore suction caused by self-desiccation. When the cement content is increased sufficiently one would expect the rigidity of the soil aggregates to approach that of the sand in sand-cement, leading to continued expansion instead of autogenous shrinkage. There is evidence that this is starting to occur when the cement content of the horizontally-moulded specimens is 10 and 15 per cent (Figure 3). 7.7 The The moisture effect of the moisture content at which the soil-cement is moulded shrinkage pattern was not altered by compacting the specimens at contents 2 per cent greater or less than the optimum but the initial shrinkage was higher in the wettest specimens. Figure 15 shows that moulding a specimen at 2 per cent above the optimum moisture content will increase the pore suction developed after extrusion to nearly double the values obtained from a specimen moulded at the optimum moisture content. Consequently the higher initial shrinkage in the wetter specimens is probably due to this increase in the pore suction. 7.8 The influence of the density The specimens compacted to a higher density, where the particles are closer together and the strength is higher, exhibited greater shrinkage than the specimens compacted to the lower density. This was a most unexpected resuit, and the probable explanation is that the water required for the hydration of the cement in a more dense specimen will be taken from finer capillaries and hence a higher pore suction will be induced. One must also consider the increased strength of the matrix due to the closer contact within the cemented skeleton but the results suggest that this higher strength is not sufficient to counteract the shrinkage forces. This is clearly a finding that merits further investigation. It should be remembered,that strength and durability, which are also affected by the density of the material, can be even more important in practice than shrinkage. 7.9 The effect of the pre-treatment” moisture content An explanation for the results shown in Figure 13 of the competition between the soil and the cement for can the be offered in terms moulding water. 20 In specimens prepared from dry soil, the water coating the soil aggregates is simultaneously attracted both by the soil particles and the cement. The cement will start to hydrate and expand, the clay fraction in the soil will swell but, in the initial stage, the tendency for the constituents to expand will not completely counteract the reduction in volume caused by the removal of water from the film surrounding the soil aggregates. As the hydration proceeds and as more water is adsorbed by the clay, expansion of the cement gel and swelling of the clay will predominate and there will be a,noverall expansion in the material. Eventually water for cement hydration will be extracted from the clay, self-desiccation will exceed the expansion in the cement gel, and the material will shrink. tien specimens are prepared from soil pre-treated at the optimum moisture content, the water for hydration will be taken entirely from within the soil causing shrinkage in the clay fraction which is unlikely to be offset by the expansion in the gel. If the soil is pretreated at 2 per cent less than content, the behaviour is intermediate between the two above. 7.10 The effect of drying the optimum moisture extremes described There was an initial sharp rise in the drying shrinkage when the wax was stripped from the specimens, but in every case it was completed wit:hin 7 days of the beginning of the drying cycle. The magnitude of the drying shrinkage decreased when the period of moist curing was increased, supporting the findings of George2 but contrary to the results of Nakayama and Handyl. If soil-cement specimens are allowed to dry, the self-desiccation is augmented by the loss of moisture to the atmosphere, emptying smaller capillaries and increasing the pore suction and shrinkage forces. Ultimately, most of the capillary water will be lost, the shrinkage forces will reach a maximum and no further shrinkage strain will occur. The loss of water due to evaporation is greater than the mount of water required by cement for hydration and, therefore, evaporation drying will contribute more to shrinkage than will self-desiccation. 21 When specimens are moist cured for a longer period of time before being allowed to dry out, more cementation and bonding occursr thereby decreasing the shrinkage potential in the clay fraction. 8. I~LICATIONS FOR RO~PAVE~NT LA=RS This study has produced data which have important practical implications for minimizing the shrinkage in pavement layers constructed from clayey soilcement. The spacing, and more importantly, the width of cracks in a soil-cement pavement layer are important factors influencing overall performance. Under ideal curing conditions, the spacing of cracks is primarily governed by the strength of the material and the subgrade restraint, while the width of the cracks during the early life of the soil-cement layer can mainly be attributed to the shrinkage characteristics of the material. Subsequent changes in the width of the cracks will be influenced by changes in temperature and in the moisture within the layer. In the following discussion the effects of the cement content, moisture content, compaction and curing on the shrinkage of soil-cement are considered, but the way in which the strength and stiffness of soil–cement affects the spacing and width of cracks is not discussed here. 8.1 Cement content It is often suggested that the problem of shrinkage cracking in soilcement can be reduced by lowering the cement content provided that an adequate minimum strength is maintained. Horizontally moulded specimens are appropriate for estimating the shrinkage behaviour of a soil-cement layer because the orientation of the soil particles is similar to that that, contrary to the traditional moulded specimens is minimised by thinking has usually assumed that in a pavement layer. The study showed ideas, autogenous shrinkage of horizontally increasing the cement content [traditional shrinkage is caused by drying] . However the full implications of increasing cement content and the consequent increase in strength on the spacing between the cracks should be considered together with the crack width. Further work is needed to study drying shrinkage after a number of months of moist curing in an attempt to simulate the field. conditions to which a well-cured base may be subjected. 22 8.2 Moisture content The magnitude of the shrinkage can be significantly reduced if the soil is kept in the driest practicable state before it is mixed. The most effective way to achieve this is to use a single-pass mixer and compact the stabilized material as soon as possible after mixing, before any significant swelling occurs within the soil aggregates. An added advantage of using this method is that the moisture will be sucked into the soil aggregates, carrying hydrated cementitious material with it, thus helping to stabilize the interior of the soil lumps. The target moisture content should be the optimum for compaction but it is preferable to err on the low side rather than the high if the criterion to be satisfied is minimizing the air difficult to meet. 8.3 Compaction that of minimizing shrinkage. Other criteria, such as voids in the compacted material, would then be more ‘ Ideally, compaction plant should induce a low pore pressure in the upper part of the layer on which it is working, when compacting a clayey soilcement material. Probably the best plant for this purpose is a pneumatictyred roller, and the least suit~le is likely to be a vibrating roller because the shock loading it would impose would cause a large immediate increase in pore pressure. It would appear from the results obtained that compacting clay-cement mixtures to a high density will increase the shrinkage as well as the strength of the material, resulting in a large spacing between cracks. As a consequence of both strength and shrinkage considerations, the crack formed will therefore be very wide. This is a controversial finding that requires further investigation. 8.4. Curing It is difficult to prevent the gradual loss of moisture from a cement pavement layer into the underlying material, but evaporation surface of a layer can be prevented by sealing it immediately after been compacted. Any period of curing is beneficial in reducing the soilfrom the it has shrinkage and the longer the layer can be maintained in this condition, the smaller will be the final total shrinkage. 23 9. CONCLUSIONS The results obtained in this investigation enable the following main conclusions to be drawn: 1. The method of compaction and the anisotropy induced in specimens has a major influence on the magnitude of shrinkage and hence the prediction of the shrinkage behaviour of full-scale pavements on the basis of laboratory tests must be made with caution. Shrinkage was a maximum 14 in specimens moulded by dynamic compaction (BS co~action ) and a minimum when static compaction was used. 2. The higher the proportion of cement that is added to clays, the smaller will be the autogenous shrinkage of horizontally moulded specimens in which the structure of a soil-cement pavement layer is most closely simulated. Vertical compaction methods which do not simulate field conditions allow other conclusions to be drawn but these should be treated with reservation. When the clay is unstabilized, the larger values of shrinkage that occur are probably due to the combined effects of the soil moisture suction which develops when the specimen is extruded from the mould and the gradual re-orientation of the water adsorbed on the clay particles. The addition of cement introduces other factors, the most important being the expansion of the cement gel as it hydrates, the self-desiccation caused by the hydration of the cement and the increase in the strength of the cemented soil skeleton. 3. Shrinkage of clay-cement specimens is profoundly affected by the initial moisture condition of the soil prior to moulding, by the moisture content at which they are compacted, and by their final density. Specimens moulded to a high density shrink more than those with lower densities and the shrinkage increases when the moulding moisture content is increased. The shrinkage of clay cement will be minimised if the soil is processed in a dry state and the water required for compaction and hydration is added during mixing. Additionally, to minimise shrinkage the mixture should be compacted as quickly as possible to will attain the minimum density “that will ensure that the material adequate strength. 24 4. If clay-cement specimens are exposed to a drying atmosphere at any time after moulding the shrinkage will be greater than that of sealed specimens, but the longer the period of time they are kept Sealed before drying commences, the smaller is the total shrinkage. 10 ACKNOWLEDGEMENTS The work described in this Report was carried out the Trans~rt and Road Research Laboratory. Much by H O Hassan, a voluntary worker in the Overseas in the Overseas Unit of of the work was undertaken Unit, as part of his work for a Ph.D thesis for the University of Surrey. The supervisors of this Ph. D programme were R I T Williams of the University of Surrey and H E Bofinger of the Overseas Unit. 1. 2. 3. 4. 5. 6. 11. ~FERENCES NAKAYAMA, H and R L HANDY, 1965. Factors soil-cement. Highway Research Record 86; Research Board) pp15-27 influencing shrinkage of Washington DC (Highway GEORGE, K P, 1968. Shrinkage characteristics of soil-cement. Research Record 255. Washington DC (Highway Research Board) Highway pp42-57 WEBB, T L, T P CILLIERS and N STUTTERHEIM, 1950. The properties of compacted soils and soil-cement mixtures for use in buildings. South African Council for Sci and Indust Research, Natal Building Research Institute. WANG, J W H, lg73. Use of additives and expansive cement for shrinkage crack control in soil-cement. Highway Research Record 442. Washington DC (Highway Research Board) ppll-20. PRETORIUS, P C and C L MONISMITH, 1971. Prediction of shrinkage stresses in pavement containing soil-cement bases. Highway Research Record 362. Washington DC (Highway Research Board) pp63-86. DUNLOP,R J, 1973. Shrinkage and creep characteristics of soil-cement. Unptilished Ph.D Thesis, Univ of Canterbury, Christchurch, New Zealand. 25 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. PLUmR, F L and S M DORE, 1940. Soil mechanics and foundation. Pitman, New York, pp60-61. MEANS, R G and J V PARCHER, 1963. Physical properties of soils. Merrill Books Inc., ColumbUS. CZERNIN, W, 1962. Cement chemistry and physics for civil engineers. London. Crosby Lockwood. HANSEN, T C 1958. Creep of concrete. Research Institute Bulletin No. 33. BERNAL, J D, lg52. The structure of cement hydration compounds 3rd Intern. Symp. on the Chemistry of Cement. London 1952. pp216-236. WANG, J W H and A H ~MMYDOS, 1970. Use of sodium reducing shrinkage in montmorillonitic soil-cement. Record 315. Washington DC (Highway Research Board) BOFINGER, H E and C G DUFFELL, 1973. The effect of chloride in Highway Research pp81-90. filler on the characteristics of sand-cement mixtures. Department of the Environment, T~L Report LR 527. Crowthorne, 1973 (Road Research L*oratory). BRITISH STANDARDS INSTITUTION. engineering purposes. London British Standard No. 1377. BRITISH STANDARDS INSTITUTION. Methods of testing soils for civil 975 (British Standard Institution). Part 2: Specification for Portland Cement (Ordinary and rapid hardening). London 1971 (British Standards Institution) . British Standard No. 12. LAMBE, T W, 1961. Residual pore-water pressure in compacted clays. Proc. 5th Intr. Conf. on Soil Mech. paris pp207-212. -E, T W, 1958. The structure of Mech and Found. Div. Proc. of ASCE, and Found. Engng. Vol 1. compacted clays. Jour. of Soil May 1958. 26 18. 19. 20. 21. 22. 23. 24. 25. AITCHISON, G D, 1961. Relationship of moisture stress and effective stress functioxls in unsaturated soils. Pore-pressure and suction in soils, Soil Mech. and Found. Engng. Butterworth. ROSENQUIST, I T, 1959. Physico chemical properties of soils: Soil water system. Proc. ASCE 85 SM2. LOW, P F, lg59. Viscosity of water in clay Conference Clays and Clay GRIM, R E, 1962. Applied TAYLOR, H F W, 1964. The London, New York. SHERWOOD, P T, 1968. The Minerals, London. clay mineralogy. chemistry of cement, properties of cement systems: Proc. 8th Nat. McGraw-Hill Book Coy. VO1 1. Academic Press, stabilized materials. Ministry of Transport, RRL Report LR 205. Crowthorne, 1968 (Road Research Laboratory). HASSAN, H O, 1975. Shrinkage and creep in soil cement. Unpublished PhD Thesis. University of Surrey. EL RAWI, N M, T A HALIBURTON and R L JANES, 1968. Effect of compaction method on strength parameters of soil-cement mixtures. Highway Research Record 255, Washington DC (Highway Research Board) pp72-80. 27 12. APPE~IX TABLE 1 Typical shrinkage strains vs time of L.B.E. sealed specimens moulded to B.S. light density and O.M.C. by vertical static compaction Cement content Spec No Time 15 reins 30 reins 1 hr 2 hrs 4 hrs 1 day 2 days 3 day,s 7 days 14 days 28 days 72 days 8% 1 0 92 96 96 126 480 512 512 512 520 542 856 Shrinkage strain x 10-b 2 0 46 46 46 58 232 276 278 278 280 296 536 3 0 36 38 40 48 386 406 410 410 416 424 686 4 0 32 36 38 44 336 358 358 360 364 386 746 5 0 46 46 48 60 272 312 312 318 322 356 716 Average o 50 52 54 67 341 373 374 376 380 401 708 28 TABLE 2 Typical shrinkage strains vs time of L.B.E. sealed specimens moulded to B.S. light density and O.M.C. by horizontal static compaction Cement content Spec No Time 15 reins 30 reins 1 hr 2 hrs 4 hrs 1 day 2 days 3 days 7 days 14 days 28 days 56 days 8% 1 17 53 95 123 124 147 143 127 251 378 547 867 Shrinkage strain x 10-6 2 67 81 95 108 108 167 160 143 231 283 475 760 3 25 59 84 95 95 181 173 173 271 438 614 814 4 31 58 87 106 108 162 162 162 260 374 557 848 5 42 70 96 114 118 176 170 150 258 364 533 827 Average 36 64 91 109 111 167 162 151 254 367 545 823 29 TABLE 3 Typical shrinkage strains vs time of L.B.E. specimens moulded to B.S. light density and by kneading compaction sealed O.M.C. Cement content Spec No Time 15 reins 30 reins 1 hr 2 hrs v 4 hrs 1 day 2 days 3 days 7 days 14 days 28 days 8% 1 275 400 440 443 443 380 380 380 380 380 450 2 Shrinkage strain x 10 -b . 316 439 468 468 468 385 381 381 381 387 473 3 298 431 463 465 460 379 373 373 373 382 488 4 257 389 431 437 433 403 398 398 398 426 560 5 363 553 601 605 600 508 427 427 427 460 543 Average 302 442 481 484 481 411 392 392 392 407 503 30 TABLE 4 Typical shrinkage strains vs time of L.B.E. sealed specimens moulded to B.S. light density and O.M.C. by dynamic compaction Cement content Spec No Time 15 reins 30 reins 1 hr 2 hrs 4 hrs 1 day 2 days 3 days 7 days 14 days 28 days 8% 1 126 260 500 678 786 960 1018 1150 2060 2420 3980 1 2 Shrinkage strain x 10-6 142 304 446 740 900 1070 1124 1260 2920 3230 4340 3 119 263 468 683 800 1013 1080 1186 2217 2650 4015 4 136 293 470 706 831 1043 1113 1209 2500 2936 4209 5 130 279 493 718 853 1056 1128 1237 2660 3086 4256 Average 131 280 475 705 834 1028 1093 1208 2471 2864 4160 31 o 0 [ 1 Ogl -- Loo— 100.0 = 1000.0 — 0 0 E .- -- - -- - -- 10 a 0 0 0 0 0 0 0 0 m b m m u m m. 0 — — Ixllllllllllll - -- -- -- -—- ---- ------ --- ----- --—- t 001 > a d — — — : 0, \ “\ \ \ — o 0 0 0 0 0 0 0 0 0 0 0 8 0 a w w N — 0 9.01 x U!QJIS~6eau!J~Sa6eJaAv / 1000 800 600 400 200 1 2 4 (h) 1 23 7 14 28 (days) I I I I I I I I I f (per cent) 4 / o 10 Fig. 4 RELATIONSHIP 102 ,03 Curing time (rein) BETWEEN SHRINKAGE AND LOG. TIME OF L.B. E. SOIL-CEMENT. 104 KNEAD!NG COMPACTION 105 1 2 4 (h) 1 23 7 14 28 (days) 4200 3600 a 3000 0 x c.-mL 2400 z a. m s c.- : 1800 a ~ i 1200 600 0 I I I I 1 I I [ /’ (per cent) 10 30 60 102 ,03 ,04 2 34 ,.5 Curing time (rein) Fig. 5 RELATIONSHIP BETWEEN SHRINKAGE AND LOG. TIME OF L.B.E. SOIL~EMENT. DYNAMIC COMPACTION 4000 3500 3000 2500 2000 1500 1000 500 0 Dyna mic Static (horizontal) o 2 4 6 8 10 Cement content (per cent) Fig 6 INFLUENCE OF CEMENT CONTENT ON SHRINKAGE OF SPECIMENS MOULDED FROM L.B.E. BY VARIOUS METHODS OF COMPACTION 1400 1200 1000 800 600 400 200 0 — Cement (per cent) 4 —8 6 10 Fig 7 ,.3 104 ,05 Curing time (rein) RELATIONSHIP BE~EEN SHRINKAGE AND LOG. TIME OF BAGSHOT SANDY CLAY SOILCEMENT. VERTICAL STATIC COMPACTION .- \ \ \ \ o 0 0 0 0 0 ~ o 0 0 0 0 w m o 0 0 w w e N 9.01x u!eJls a6eyu!~qs afieJaAV 4800 4000 ‘ 3200 ‘ 2400 1600 800 0 10 ,Oz 103 Curing time (rein) 104 Fig. 9 RELATIONSHIP BETWEEN SHRINKAGE AND LOG. TIME OF WAINSCOTT BROWN CLAY SOIL-CEMENT. VERTICAL STATIC COMPACTION Fig. 10 4800 4000 3200 2400 ‘ 1600 (per cent) 4 —6 8 10 102 ,03 Curing time (rein) RELATIONSHIP BETWEEN SHRINKAGE AND LOG. TIME OF WAINSCOTT BROWN CLAY COMPACTION 104 SOILXEMENT. HORIZONTAL STATIC \ 0 9.01 x u!eJls a6eXu!J~s 2400 2000 1600 1200 800 Fig. 12 10 102 103 ,04 105 Time (rein) RELATIONSHIP BE~EEN SHRINKAGE AND LOG. TIME OF L.B.E. SOIL4EMENT CONTAINING 8 PER CENT CEMENT MOULDED AT VARIOUS DENSITIES. HORIZONTAL STATIC COMPACTION 500 400 300 200 100 0 100 200 10 10Z 103 ,04 105 Curing time (rein) Fig 13 INFLUENCE OF PRE-TREATMENT MOISTURE CONTENT ON SHRINKAGE OF SPECIMENS CONTAINING 6 PER CENT CEMENT COMPACTED AT OPTIMUM MOISTURE CONTENT TO THE (2:5kg RAMMER TEST) MAXIMUM DRY DENSITY — o00 0 o . . ~ No cement at O.M.C. — — 8% cement at O.M.C. ~ 8% cement at 2% above O.M.C. 1 2 4 (rein) 1 2 45 (h) 24 400 I o I E i 300 200 100 I o I I I I I I I I 1’ I 1;.- 1: /8 10 r — 1 II I I I I s I I I I , I / 0 I I I i I I I / I I I I a I 8 6 - 4 0 -2 II A -4 I I II -6 - I II I ) -8 ,02 103 ,.4 ,05 10 Fig 15 Curing time(s) CHANGES IN PORE PRESSURE AND AXIAL STRAIN WITH TIME Fig. 16 I 1 1 56 24 Time (h) POSSIBLE SHRINKAGE OF CLAY WITH TIME DUE TO DEVELOPMENT AND DISSIPATION OF PORE SUCTION 1 2 3 Time (months) Fig. 17 POSSIBLE SHRINKAGE OF CLAY WITH TIME DUE TO RE-ORIENTATION OF ADSORBED WATER Direction of compacting force 1 Fig. 18 CLAY STRUCTURE AFTER COMPACTION 4. -------- 1 — (726) Dd0536316 lM 5/78 HPLtd So’tonG191S PRINTED IN ENGLAND ABSTRACT THE SHRIN~GE OF FINE-G~INED SOIL-CENENT:H E Bofinger, H O Hassan and R I T Williams: Department of the Environment Department of Transprt TRRL Supplementary Report 398: Crowthorne, 1978 (Transport and Road Research Laboratory) . The incidence of cracks not associated with externally applied loads is a feature of soil-cement roadbases” and the consequent problem of reflection cracking through the superimposed surfacing has caused concern over the years. The cracks result from restrained thermal and shrinkage movements but, when the stabilized soil is fine-grained, it is believed that the cracking is mainly attributable to shrinkage. The paper describes a study of the autogenous shrinkage of such a material aimed at improving the understanding of its susceptibility to cracking. In the investigation, specimens were cured under ideal conditions, but, because conditions in the fie{d are often far from ideal,the influenceof dryingon shrinkage was also studied. The principal finding of the investigation is that the measured value of shrinkage is greatly affected by the test conditions imposed. In particular, major differences were observed in specimens compacted by different methods and, furthermore, the mode of compaction had a striking influence on the way in which the cement content of mixes influenced shrinkage. A tentative explanation for the results obtained is presented in terms of the pore pressure in the specimens, the hydration of the cement and the particle orientation. It is clear, however, that further work is necessary before the general validity of the results can be determined. ISSN 0305-1315 ABSTRACT THE SHRIN~GE OF FINE-GRAINED SOIL-CENENT: H E Bofinger, H O Hassan and R I T Williams: Department of the Environment Department of Transport TRRL Supplementary Report 398: Crowthorne, 1978 (Transport and Road Research Laboratory) . The incidence of cracks not associated with externally applied loads is a feature of soil-cement roadbases, and the consequent problem of reflection cracking through the superimposed surfacing has caused concern over the years. The cracks result from restrained thermal and shrinkage movements but, when the stabilized soil is fine-grained, it is believed that the cracking is mainly attributable to shrinkage. The paper describes a study of the autogenous shrinkage of such a material aimed at improving the understanding of its susceptibility to cracking. In the investigation, specimens were cured under ideal conditions, but, because conditions in the field are often far from ideal, the influence of drying on shrinkage was also studied. The principal finding of the investigation is that the measured value of shrinkage is greatly affected by the test conditions imposed. In particular, major differences were observed in specimens compacted by different methods and, furthermore, the mode of compaction had a striking influence on the way in which the cement content of mixes influenced shrinkage. A tentative explanation for the results obtained is presented in terms of the pore pressure in the specimens, the hydration of the cement and the particle orientation. It is clear, however, that further work is necessary before the general validity of the results can be determined. ISSN 0305-1315