T H E I N F L U E N C E OF T H E C L A Y F R A C T I O N ON THE ENGINEERING PROPERTIES OF SOIL, W I T H S O M E S U G G E S T I O N S FOR F U T U R E RESEARCH By E. K. CLARE. The clay fraction affects the engineering properties of the soil in various ways: Mechanical Strength.---This is normally measured in terms of resistance to shear, which is a function of the internal friction and cohesion of the soil. The internal friction is primarily a characteristic of the coarser soil fractions and results from the interlocking of the particles; its value is a function of the normal stress on the soil. The cohesion of a soil is produced by forces binding the particles together, and is independent of the normal stress. These binding forces are accounted for partly by the, surface tension at the air-water interface of the water films surrounding the particles a n d partly by the intermolecular forces acting between films of water adsorbed on the surfaces of neighbouring particles. The surface tension forces are greatest between clay particles due to the smallness of the voids between them, while the interaction of adsorbed films of physically or chemically bound water is almost entirely a function of the clay fraction. Clay thus enables a soil to develop shear strength by virtue of the cohesive forces between the particles. However these forces also resist the mechanical stress required to displace the particles and pack them closely together to form a dense mass, and more energy has therefore to be supplied to a cohesive than to a non-cohesive soil to obtain adequate compaction. Soil moisture content.--When a soil has completed its free drainage under the action of gravity, it retains a certain amount of water at a reduced pressure in the pores and on the surfaces of the particles by virtue of the surface tension and adhesive forces. Principally as the result of the large surface area of the clay fraction compared with the coarser fractions, the water retained by a soil increases with increasing clay content. Moisture movement.--In the case of saturated flow in the liquid phase, the permeability or rate of moisture movement in a soil decreases with increasing clay content, due to the increasing proportion of small capillary channels. In the case of unsaturated flow in the liquid phase, moisture movements arise from differences in negative pressure due usually to local differences in moisture content. This effect can be explained by the tendency of adsorbed water films to grow at the expense of moisture held at lower tension n adjacent 3O

INFLUENCE OF CLAY FRACTIONON ENGINEERINGPROPERTIES

31

pores. At any given moisture content, the negative pressure is greater the larger the clay fraction. Moisture movements in the vapour phase occur when a difference o f vapour pressure exists in two parts of a soil. Such differences are mainly associated with temperature gradient in the soil. The vapour pressure of a soil at any temperature has been shown experimentally to be a function of both the clay and moisture contents (Purl, Crowther and Keen, 1925). The susceptibility of a soil to disruption by frost action depends on its permeability and hence indirectly on the magnitude of its clay fraction. The consolidation of a soil, or its compressibility under load also depends on its permeability, and hence its clay fraction, since the rate of compression depends on the rate at which moisture is expelled from the soil. The shrinkage and swelling ofsubgrade soils are known to be responsible for seasonal movements of road surfaces. Both these phenomena are due to variations in the thickness of the adsorbed water films causing variations in the bulk volume Qccupied by unit weight o f ' t h e soil particles. Since only adsorbed water films are concerned, these phenomena are mainly characteristic of the clay fraction.

Electrical drainage of soil.--This process has been confined in practice to silty soils, but the suggested mechanism indicates that it is a colloidal phenomenon and should therefore be particularly operative in the clay fraction. The fact that in practice clay soils cannot be easily drained by the process may be due to the low permeability of such soils to moisture movement. Physico-ehemical characteristics of clay partieles.--These are likely to be influenced by size distribution, particle shape, variability in structural dimensions (as e.g., for montmorillonite), and type of adsorbed cation (which will influence the thickness of the adsorbed water film), as well as the chemical and mineralogical characteristics of the clay particles. As regards the bond holding adsorbed cations and water molecules, some authors differentiate between adsorption due to broken chemical bonds on exposed edges, said to be very strong, and planar adsorption on the surfaces of the crystal lattice sheets. Whatever the detailed mechanism, the molecules nearest the particles will be held by "activated" or chemisorption, and the nature of the particle surface will be of importance. In an endeavour to determine the effect of the clay fraction on the flow properties of soil, an analysis has been made of the results of the Atterberg plasticity tests which are carried out as routine classification tests. These comprise the Liquid Limit and the Plastic Limit tests (American Society for Testing Materials, 1944), both of which are standardised in the U.S.A. and are at the moment under consideration by the British Standards Institution. In the Liquid Limit test a groove of fixed dimensions is cut in a pat of damp soil contained in a standard brass cup. By repeated experiment the moisture

32

E.K. CLARE

content is determined at which the groove can be dosed over a fixed part of its length by allowing the cup to drop 25 times through a height of 1 centimetre. In the Plastic Limit test damp soil is rolled into a small thread between the hand and a fiat surface so that the water is evaporated from the soil by the heat of the hand, until the soil just ceases to be plastic and crumbles. The moisture content at which this occurs is the Plastic Limit. A comparison of the results of plasticity tests and mechanical analyses for a wide variety of soils shows that both the Liquid and Plastic Limits are dependent on the Characteristics of the adsorbed water films and hence on the amount and type of the clay fraction in a soil, but the difference between them, the Plasticity Number ( o r "Plasticity Index" as it is termed in Civil Engineering practice) is a function only of the amount of fine material; therefore if the Liquid Limits of a number of soils of different types are plotted against the corresponding Plasticity Indices, the existence of more than one relationship will. be due to differences between the types of clay mineral in the various soils. This forms the basis for some of the soil classification systems in use in Civil Engineering soil work, the most widely known being that due to Casagrande. The results of plasticity tests made on a large number of British soils have been plotted in this way. Although the corresponding equations are very similar, there may be two distinct relationships corresponding to P N ~ 0 . 7 5 (LL--15) and P N = 0 . 8 (LL--15) approximately. It is difficult to say whether these differences are significant, having regard to possible experimental errors and the scatter of the points. Data obtained from loss on ignition determinations have also been examined in an attempt to classify clay soils (Road Research Laboratory, 1945). Samples of soil dried at 105 ~ were heated to temperatures between 700~176 in an electric furnace and the loss in weight determined, and calculated as a percentage of the fraction with equiv, particle diameters < 2 tL. The results suggest that again there may be two main groups, one giving a loss of about 14 per cent. and the other about 20 per cent. on ignition, comparable to the two groups of LL/PN lines, but the divisions do not agree with each other if one takes geological group as a criterion of soil type. These discrepancies may be due to incorrect geological classification, which was not made with the use of the detailed drift map, but merely refers to the underlying strata. Size distribution in the clay fraction has been studied by means of a development of the technique of Puri and Puri (1941) and E. W. Russell (1943) using a hypodermic syringe for sampling. The frame carrying the syringe is fitted with a vernier screw device enabling the point of the nozzle to be inserted to very shallow depths with an accuracy of 0.01 cm. The technique was compared in two cases with the normal method employing a 10 ml. pipette for particles of 0.006 mm. and 0.002 mm. equiv, diameter, and the two methods were found to give results that differ by not more than I per cent. of material, even

INFLUENCE ON CLAY FRACTION ON ENGINEERING PROPERTIES

33

for sampling depths of 2 mm. In the future it is intended to investigate the possibility of using the syringe technique in conjunction with centrifugal separation. A limited study has been made of the size and shape of particles of clay by means of electron microscopephotographs, and for this purpose the Metallurgical Division of the National Physical Laboratory took 48 photographs of clays from eight British soils supplied by the Road Research Laboratory. Lath-shaped particles (halloysite?) have been observed in some of the clays. Another line of research has shown that certain surface active agents such as cetyl trimethyl ammonium bromide and cctyl pyridinium bromide can reduce the water adsorption and swelling of cohesive soils. The heat of wetting of a soil is also materially affected; a sample of sandy loam from Harmondsworth gave 3.4 cals./grn, untreated, and 2.4 after treatment. Since the clay fraction is known to be responsible for most of the heat given out when a soil is wetted, adsorption probably occurs principally in that fraction. The results suggest that the organic cations are adsorbed in layers on to the adsorbed water films surrounding the soil particles when these are fairly thin, that is when the soil has a low moisture content. -The tendency for the chains of polar water molecules to grow will be interrupted, since the relatively large organic ions will be between the adsorbed and the pore water, reducing the orienting electric fields and breakirlg the chain sequence. If the mechanism outlined above is correct, the effect might be developed into a method of controlling the adsorption and swelling characteristics of soils that would be very useful to engineers in tropical and subtropical areas.

Suggestions .for future research.--(1) Investigation of methods of thermal analysis for elucidating structure of, and classifying, c l a y minerals. (2) Classification of soils with predominantly similar clay fractions to help the engineer in predicting mechanical properties. (3) Correlation of the mechanical properties of soils with types of clay minerals and mixtures of these, including effect of exchangeable bases and anions. (4) Continuation of work already done on size and shape of clay particles, including wider use of electron microscope with improvement of dispersion technique and specimen-forming methods. Correlation of sedimentation analysis with particle size observed in electron microscope. (5) Investigation of structure and behaviour of water and other films adsorbed on to surfaces of clay particles with particular reference to mineralogical constitution of the clay. References.

American Society for Testing Materials, 1944. (a) Standard method of test for liquid limit of soils. A.S.T.M. designation D.423-39. (b) Standard method for plastic limit and plasticity index of soils, A.S.T.M. designation D.424-39 (A.S.T.M. Standards, Pt. lI. Non-metallic materials--Constructional).

34

E. FORSIAND

Casagrande--See Engineering Manual, W.D. Office of the Chief of Engineers. Chap. XX---Part ]I, Exhibit 1 . Purl, A. N., E. M. Crowther and B. A. Keen, 1925. J. Agric. Sci. 15 (I) 68-88. Purl, A. N. and B. R. Purl, 1941. J. Agric. Sci. 31, 171-7. Road Research Laboratory, 1915. A rapid method of determining the grain-size analysis of soils for particles less than 0.002 ram. Note No. RN/657/JFR. Russell, E. W., 1943. J. Agric. Sci. 33 (3), 147-54. Road Research Laboratory, Harmondsworth, Middlesex.

NOTE

ON WORK I N PROGRESS A T SWEDISH CEMENT .AND CONCRETE RESEARCH INSTITUTE

THE

BY E. FORSLIND.

(Received but not read at the meeting). One of the most interesting and practically most important questions regarding the behaviour of clays, and particularly montmorillonites, is concerned with the water binding mechanism in relation to the mechanical properties, elasticity and plasticity. A reconsideration of certain known data for clays, in combination with some recent developments in the study of the structure of liquids and of adjacent fields of research, suggest a working hypothesis that may deserve a closer study. If the water films between the single crystal sheets, composing the montmorillonite particles, are considered to be principally responsible for the mechanical deformation properties of clays, the different :factors influencing the degree of order and the stability of the water structure are of obvious importance. Starting from the suggestions of Grim (1942) that the montmorillonite lattice structure proposed by Edelman and Favejee (1939-40) is correct and that the structure of the water layers according to the ideas of Hendricks and Jefferson is crystalline, it seems natural to suppose that the structure of the water layers should be that of ordinary ice. The geometrical relationship between the silicate framework and the ice structure and the negligible deformations that ~/re necessary tbr the formation of hydrogen bonds between the two structures, together with X-ray data, seem to support the idea that the montmorillonite crystal lattice stabilizes the natural water configuration.