Department of Civil Engineering, King Fahd University of Petroleum and Minerals, KFUPM, P.O. Box 368, Dhahran 31261, Saudi Arabia

Abstract

The behavior of unsaturated clayey soil is highly influenced by the coupled interaction between water and clay content. Various aspects of the behavior of artificial clay–sand mixtures with variable water content were experimentally studied. Laboratory tests were utilized for the determination of consistency limits, the stress–strain relationship, strength parameters, hydraulic conductivity, and volume change characteristics for various combinations of water and clay content in soil mixtures.

Results presented for various clay–sand mixtures include: new normalized consistency limits; the combined effect of clay content and water content on the stress–strain relationship and on the strength parameters (c and φ); and the effect of clay content on hydraulic conductivity and swelling potential. The cohesion of clayey sand is found to increase with increasing water content to a certain limit, above which it decreases. The angle of internal friction for clayey sand is found generally to decrease with increasing water content. The degree of saturation is found to be better than the water content in explaining the strength behavior. The hydraulic conductivity sharply decreases with increasing clay content up to 40% beyond which the reduction becomes less significant. Simple empirical equations are proposed for predicting the swelling potential of clayey soils as a function of either the clay content or plasticity index.

Author Keywords: Soil behavior; Unsaturated soils; Clayey sand; Swelling potential; Environmental factors

φ

peak angle of internal friction

ε

axial strain

ΔH

change in height

ρ_w

water density

A

activity

A′

corrected area

A_o

original area

A_s

specific area

C

clay content

c

cohesion

d

thickness of adsorbed water layer

e

void ratio

G_s

specific gravity

H_o

original height

k

hydraulic conductivity

LL

liquid limit

PI

plasticity index

PL

plastic limit

R²

coefficient of determination

S

degree of saturation

SP

swelling potential (swell %)

w

water content (%) and

w_c

water content absorbed by the clay (%).

Article Outline

Nomenclature

1. Introduction

2. Material properties

3. Experimental investigation

4. Results and discussion

4.1. Geotechnical properties

4.2. Stress–strain relationships

4.3. Strength parameters

4.4. Hydraulic conductivity

4.5. Swelling potential

5. Conclusions

Acknowledgements

References

1. Introduction

The behavior of a soil depends on the composite effects of several interacting factors, namely compositional and environmental factors. Compositional factors include: the amount and type of soil minerals, the shape and size distribution of soil particles, adsorbed cations, and pore water composition. Environmental factors include: water content, density, confining pressure, fabric, and temperature. Additionally, compositional factors determine the potential range of values for any soil properties, whereas environmental factors dictate the actual value (Mitchell, 1993). Most of the engineering properties of soils are not intrinsic; they depend on different environmental conditions. Considering both kinds of factors gives a more meaningful assessment of soil behavior and a full spectrum of the variation of various soil properties, especially when dealing with unsaturated soils in arid regions, or dealing with some geoenvironmental problems.

Clay minerals produce very important soil types (a compositional factor), which are known to have high plasticity, cohesion, and swelling potential, but low hydraulic conductivity and friction angle. They are also known for their dominating influence on the behavior of the entire soil mass even if they exist only as a small fraction of its constituents. These clay properties are only mobilized by the presence of water (an environmental factor). The physicochemical interactions between clay minerals and water resulting from surface hydration and the diffused double layer, affect many of the properties of clayey soils. The amount of clay fraction in a soil is crucial in determining its properties required for all geotechnical and geoenvironmental applications, especially hydraulic conductivity, volume change, and strength. The hydraulic conductivity of clayey soils is very important in the design of more economical and stronger clay liners and cut-off walls for waste containment. Shelly and Daniel (1993) found that a gravel content of up to 60% in a clayey gravel liner did not significantly increase the hydraulic conductivity.

The shear strength parameters of a soil (cohesion, c, and angle of internal friction, φ) are usually reported either at optimum moisture content, or at saturated (soaked) condition. Such values may lead to catastrophic failures, especially in arid regions where soils are unsaturated most of the time; e.g. Hight et al. (1998). The shear strength of unsaturated soils can be related to the degree of saturation ( Fredlund; Fredlund, 1985 and Toll).

This paper is an attempt to quantify the combined effect of water content and clay content on the behavior of synthetic mixtures of selected clay and sand. An experimental approach was employed, in which an extensive laboratory investigation program was conducted using various synthetic clay–sand mixtures prepared especially for this purpose. To minimize the number of variables, only one type of clay and one type of sand were used in these mixtures. Various aspects of soil behavior were studied, including: consistency limits, the stress–strain relationship, strength parameters (c and φ), hydraulic conductivity, and volume change characteristics. A triaxial test was used to determine stress–strain behavior and strength parameters. Hydraulic conductivity was measured using the constant head method, whereas swelling potential was determined using the conventional one-dimensional oedometer. Comprehensive results were obtained for all the studied aspects of soil behavior. The obtained results are important for remolded and compacted soils; such as embankments and subbases for roads and foundations.

2. Material properties

Synthetic soil mixtures made of various proportions of clay and sand were used in this investigation. Both soils were collected from Eastern Province of Saudi Arabia. The geology of this area consists of Quaternary and recent deposits resting on Tertiary strata. These geological formations reflect the successive alternation between marine and non-marine conditions that the Arabian shelf has experienced since Tertiary times. This resulted in carbonate deposition with some marl, mudstone, and evaporate interbeds (Powers et al., 1963). Clayey strata are presented in the Rus and Dammam Formations of the Eocene age, and in Hadrukh, Dam and Hofuf Formations of the Miocene age ( Al-Sayari and Zötl, 1978). The Quaternary deposits are generally represented in sabkha areas and beach and dune sands, which are occasionally cemented.

The investigated clay material is a greenish brown clay representing an extensive local strata. The sampling site is located near the city of Al-Qatif (latitude between 26°35′, and 26°40′N, and longitude between 49°34′ and 49°37′E), which is located 5 km from the shore of the Arabian Gulf, in the Eastern Province of Saudi Arabia. Clays at this site are weathered from marl and limestone rocks of the Tertiary and Quaternary ages (Abduljauwad, 1993). The clay has been diagnosed as extremely problematic due to its expansive and swelling characteristics, mainly because of its high smectite content ( Abduljauwad and Al-Sulaimani, 1993). The mineralogical identification of clay material was obtained using X-ray diffraction (XRD) technique. Semi-quantitative analysis of XRD results revealed that the main constituents of this clay are: smectite (52%), illite (23%), palygorskite (5%), dolomite (9%), gypsum (6%), and quartz (5%). Such high affinity of this clay to water led to its selection.

The sand material was yellowish-brown in color, obtained from the beach of the Arabian Gulf, and used after modifying its grain size distribution. This sand was washed with distilled water, to get rid of all salts precipitated on the grains and hence to eliminate the possibility of any chemical interaction with the clay–water-electrolyte system. Then, the sand was sieved to separate sizes as retained in the various sieves. A grain size distribution was pre-selected, thereby producing a ‘well-graded’ fine sand with the narrowest void ratio range. The engineered grain size distribution was achieved by mixing equal amounts of sand passing ASTM sieve #40, and retained in sieves #45, 50, 60, 70, 80, 100, 120, and 140. Material retained in #40 sieve was rejected, so as to abide by the standard test procedure for the determination of consistency limits (ASTM, 1993). Sieve analysis was performed according to ASTM D422. The resulting sand is classified as SP according to the Unified Soil Classification System and as A-3 according to AASHTO Soil Classification System. The XRD results for sand indicated that the used sand consists mainly of quartz, with traces of calcite, dolomite, and ferroan.

Synthetic clay–sand mixtures were prepared by mixing different proportions of clay and sand together. These mixtures were proportioned according to the dry weights of both constituents. Different amounts of water were added to the various mixtures, to produce mixtures with variable water contents. It was noticed that the clay in its ‘natural dry’ state has a moisture content of around 11.5%.

3. Experimental investigation

An extensive laboratory investigation program was conducted on the various clay–sand mixtures. This program consisted of the determination of consistency limits, strength parameters, hydraulic conductivity, and swelling potential. The clay–sand mixtures were prepared by adding particular amounts of clay and sand together, which were blended to form a homogeneous mix. Consistency limits were determined according to ASTM D4318, for various clay–sand mixtures with a clay content ranging from 10 to 95% as well as for the pure clay. The specific gravity of the sand and clay materials, and the maximum and minimum density of the sand were also determined.

The stress–strain relationship and shear strength parameters (i.e. cohesion and angle of internal friction) were determined using the consolidated drained (CD) triaxial compression test for samples with a clay content ranging from 10 to 40% at 10% increments, and with an ‘added’ water content ranging from about 5–20% at 5% increments. The actual water content of various mixtures was experimentally determined, which is found to be more than the added water content as shown in Table 1. The clay–sand samples were thoroughly mixed with the required amount of water, then kept in plastic bags for about one day before they were tested, in order to achieve a uniform distribution of water. Samples were prepared according to Proctor compaction test. Samples, at the same clay and moisture contents, were compacted to the same density. For different sets of samples, at different clay and water content, samples were compacted with the same effort rather than to the same density, because it was not possible to obtain a sound specimen at a higher clay content when density was kept constant. Weight and volume of samples were recorded before testing. Triaxial tests were conducted using 5.1×10.2 cm cylindrical specimens at three different confining pressures of 70, 140, and 210 kPa (10, 20, and 30 psi). One set of triaxial tests were made for each condition. Data was obtained by a portable data logger connected to a computer.

The hydraulic conductivity of the pure sand, and the sand–clay mixtures with 10, 20, 30 and 40% clay content were determined using the constant head method according to ASTM D2434. Dry samples were compacted to the same respective densities as those of the triaxial test. Measurements were recorded until the readings stabilized. The hydraulic conductivity of the pure clay was determined via a consolidation test conducted on an undisturbed clay specimen sampled from a natural strata.

The swelling potential was determined for the various sand–clay mixtures by using a fixed ring oedometer according to ASTM D4546. Two samples of sand–clay mixture at each clay content were tested. Samples were prepared by filling the oedometer ring (7.0 cm diameter and 1.9 cm height) with the required amount of soil and applying a static force of 1400 kPa for half an hour. Then, the compaction force was removed and the soil sample was subjected to a seating pressure of 7 kPa (1 psi) for 24 h. Thereafter, the soil sample was inundated with distilled water and allowed to swell vertically under the seating pressure. The swelling potential of the pure clay was determined using an undisturbed sample (70 mm diameter and 19 mm height). Readings were collected by a portable data logger till complete swelling was achieved. The water content prior to the test (initial) was also determined.

4. Results and discussion

4.1. Geotechnical properties

The results of the consistency limits are plotted in Fig. 1, which shows the values of the liquid limit (LL), the plastic limit (PL), and the plasticity index (PI) of the various clay–sand mixtures. The LL and PL for clay contents lower than 10% could not be determined in the laboratory. The pure clay exhibits high Atterberg limits due to the high smectite content which increases the intake of water molecules by the clay, facilitated by the negatively charged clay surfaces and the large specific surface area of the clay mineral. The addition of sand to clay reduces these limits, because the sand particles act as an inert filler and do not interact electrochemically with water.

Fig. 1 shows that as the amount of clay is increased in the mixture, PL, LL, and PI increase almost linearly with respect to the clay content, except at a very low and very high clay content. This linear variation of consistency limits with clay content is expected, because the sand fraction behaves as an inert filler that does not have any physicochemical interaction with the clay fraction that can affect its plasticity. These results are similar to those obtained by Skempton (1953) for four clayey soils, by Seed et al. (1964a) for kaolinite–bentonite mixtures, and by Nagaraj and Han, 1998 for bentonite–sand mixtures.

The values of PL, LL, and PI for the sand–clay mixtures were normalized to the corresponding values of pure clay, and the normalized values plotted versus clay content are shown in Fig. 2. A theoretical straight line is drawn which represents the expected variation of the normalized consistency limits for clay–sand mixtures, if each mineral contributed in proportion to the amount present. This line passes through the origin, assuming that sand has a PL, LL, and PI all equal to zero. Based on that, all the three limits collapse to one line. A similar normalization was made for results of Nagaraj and Han, 1998, and the normalized values were included in Fig. 3. The results for data obtained from Nagaraj et al. (1991) lie exactly on the theoretical line. But those from Han (1998) lie away from the theoretical line, which can be attributed to the fact that the sand he used was not pure quartz, but contained feldspar and mica among the main minerals. This observation is similar to that found by Seed et al. (1964a) for illite–bentonite mixtures, which is attributed to the physicochemical interaction between clay and other active soil minerals.

The results are also plotted on a modified plasticity chart, Fig. 3, and all of them lie between the A-line and the U-line. Notice that a sand–clay mixture with a clay content of 10% falls in the CL–ML region. The plasticity chart was modified by adding vertical and horizontal axes representing the clay content.

The specific gravity for pure sand was found to be 2.69, and that for pure clay was 2.80. The specific gravity of the clay–sand mixtures was taken as the weighted average value of both ingredients. The maximum and minimum values of the dry density of sand were found to be 1.845 and 1.553 Mg/m³, respectively, which correspond to minimum and maximum values of the void ratio equal to 0.458 and 0.732, respectively.

4.2. Stress–strain relationships

Fig. 4 shows the stress–strain relationship for a confining pressure (σ₃) of 140 kPa (20 psi) at various water contents and constant clay contents. The stress plotted is the deviator stress, and the strain is the axial strain. The plots show distinct peaks at low strain for samples with low water contents for all the clay contents, whereas at high water contents, the curves are flat and do not show sharp peaks. The exception to this is the curve of the 10% clay content with 11.29% water content, where the peak is not sharp. The highest strength for the clay content of 10 and 20% was achieved at water content of 6.29 and 7.31%, respectively; and that for the clay content of 30 and 40% was at water content of 13.41 and 14.6%, respectively. Each set of curves shown in Fig. 4 depicts the effect of wetting and drying on the stress–strain behavior of soil having a certain clay content. Although these triaxial tests were drained, no water was observed to drain from any of the tested specimens, indicating that complete saturation was not achieved. All the triaxial specimens had defined inclined failure planes; no specimen had bulged.

Results showing the dry density versus water content and clay content for various clay–sand specimens are presented in Fig. 5. Dry density is obtained for the compacted specimen from the weight, volume, and water content in a manner similar to Proctor test. Fig. 5a represents the compaction curves for various clay–sand mixtures. Fig. 5b gives a 3-D variation of dry density as a function of both the clay and water contents. The plot shows that dry density increases with increasing water and clay content, up to a certain limit, beyond which the dry density decreases. The increase in dry density at low clay content is attributed to the fact that clay fills the voids between sand particles. At a high clay content, sand particles will be replaced by clay material which has a low dry density of 1.17 Mg/m³, while the water content has a role similar to that in the typical compaction curve. Using basic phase relations, the degree of saturation (S) and void ratio (e) were found for the different clay–sand mixtures. Fig. 6 shows the variation in the degree of saturation as a function of water content for the various clay–sand mixtures. Fig. 7 presents the variation in the void ratio for the various clay–sand mixtures at different water contents. The void ratio decreases with increasing clay and water contents, upto a certain limit, beyond which it starts to increase. This can be explained in a similar way to the variation in dry density.

Fig. 8 gives the stress–strain relationship for ‘dry’ clay–sand mixtures in addition to pure clay and pure sand samples under a confining pressure of 140 kPa (20 psi). Notice that the ‘natural dry’ clay contains some moisture adsorbed from the atmosphere as shown in Table 1. The dry pure clay is stronger than the dry pure sand, because dry clay is found to have moisture content of 11.5%. However, this strength can only be achieved at a strain level approximately four times higher than that of the sand. It is evident from these curves that the strain required to reach the peak strength increases as the clay content in the mixture is increased. Fig. 9 presents the peak deviator stress and the corresponding axial strain for various clay contents. Fig. 9a gives the peak deviator stress using both the original cross-sectional area of the specimen (A_o), and the adjusted area (A′=A_o/1−ε), where ε is the axial strain. The density of the dry clay–sand specimens is plotted as a function of clay content in Fig. 10a. Using basic phase relations, the void ratio for dry clay–sand specimens was calculated and plotted as a function of clay content in Fig. 10b. The variations of both density and void ratio with clay content are not linear, owing to the fact that at a low clay content most of the clay fills the voids between the sand grains and less coats the sand grains. This increases the density and decreases the void ratio. At high clay content, the portion of the clay coating the sand grains is more than that needed to fill the voids between the sand grains. This causes a reduction in dry density, and an increase in void ratio.

4.3. Strength parameters

The cohesion (c) was determined from the data obtained from different sets of triaxial tests, using the Mohr–Coulomb shear strength criterion. Fig. 11a illustrates the variation in cohesion with differing water content for various clay contents. Cohesion increases to a maximum and then decreases for all clay contents, similar to the shape of the compaction curve. Increasing clay content results in higher values for the cohesion and in sharper curves, which require a higher water content to reach maximum cohesion. Cohesion can be attributed to a combination of true and apparent types, including: cementation and adhesion due to compaction, electrostatic and electromagnetic attractions, and capillary suction. All of these cohesion sources increase with increasing clay content. They also increase with increasing water content, but only to certain limits, above which they start to decrease. As the water content increases, the separation distance between the clay particles increases and the electrostatic and electromagnetic attractions (van der Walls forces) decrease. Mitchell (1993) reported that these forces become significant for separation distances <2.5 nm. Also, the cementation and adhesion forces increase with increasing water content, but only up to a certain value, above which these forces decrease because of excessive water content. Furthermore, capillary suction will be completely lost by increasing the water content to a saturation condition. Notice that the water content at the peak cohesion (Fig. 11a) is much less than the water content required for full saturation, as can be seen from Fig. 6. The cohesion of 100% clay specimens with the optimum water content of 43% was found by Abduljauwad and Rehman (1996) to be 225 kPa.

Fig. 11b shows a three-dimensional variation of the cohesion as a function of both clay-content and water-content, simultaneously. This plot depicts the combined effect of clay-content and water-content on the cohesion parameter. For naturally dry pure clay, the cohesion is found to be only 5 kPa. For dry clay–sand mixtures, there is no significant variation of cohesion with clay content. For all the other water contents, cohesion increases with increasing clay content.

Water content include free water and absorbed water. The absorbed water is strongly attracted to the surface of clay particles up to a water thickness (d) of about 1.0 nm, i.e. three molecular layers of water, (Sposito, 1984). The water content (w_c) corresponding to that thickness of the absorbed water can be found from basic phase relationships to be:

Notice that this water content (w_c) is for the clay fraction, not for the entire soil. For values of A_s equal to 100, 200, 400, and 800 m²/gm, w_c can be found from Eq. (1) to be 10, 20, 40 and 80%, respectively. The moisture content adsorbed by this smectite clay is about 11.5%, which indicates that this clay is still not satisfying its affinity for water. It is important to notice that the water content of the clay fraction (w_c) is different than the water content of the entire mix (w). Because water intake of clay fraction is more, the water content of the entire soil is expected to be lower than w_c, but not in proportion to the weight of different constituents.

The peak angle of internal friction (φ) was also determined from the data obtained from different sets of triaxial tests, using the Mohr–Coulomb shear strength criterion. Fig. 12a illustrates the variation in φ with water content for various clay contents. The angle of internal friction generally decreases with increasing water content for all clay contents, except for the 10% clay content. This reduction is due to the increased lubrication of the clay paste following water addition causing sand grains to slip and slide, resulting in a reduced friction angle. For high clay content, the clay fraction dominates the behavior of the soil mixture, and the water acts as a lubricant, which decreases the friction angle as the water content increases. The lubrication effect of water on clay minerals has been observed by Horn and Deere (1962). This lubricating occurs when the surface of the clay particles is wetted, causing the mobility of the absorbed film to increase due to increased thickness and greater surface ion hydration and dissociation ( Mitchell, 1993). For a low clay content (10%), the angle of internal friction decreases with increasing water content to a certain value, beyond which it start to increase. This increase can be attributed to the fact that excess water results in cleaning some sand particles from the adhering clay, and the lubrication effect reduces, which in turn, increases the angle of internal friction. For this low clay content (10%), the sand fraction dominates the behavior of the soil mixture. The antilubrication effect of water on quartz minerals has been reported by Tschebotarioff and Welch, 1948; Horn; Bromwell, 1966; Dickey, 1966 and Procter. This antilubricating effect of water apparently results from the disruptive effect of water on the adsorbed clay film coating the sand particles that may have served as a lubricant at lower water content.

Fig. 12b shows a three-dimensional variation of the angle of internal friction as a function of both clay and water content, simultaneously. This plot depicts the combined effect of clay and water content on the friction parameter. For naturally dry pure clay, the angle of internal friction is found to be 46.7°. For dry clay–sand mixtures, there is no significant variation of φ with clay content. For low water contents, the value of φ increase to a maximum and then decrease. For intermediate water contents, the curve decreases to a stable minimum, which is similar to the curve obtained by Lupini et al. (1981) for sand–bentonite mixtures, in which they identified three zones termed rolling shear (at a low clay content), transitional shear, and sliding shear (at a high clay content). Similar curves were also reported by Kenney (1967) for the residual friction angles of different clay–quartz mixtures. For high water content, the curve increases and is expected to reach a stable maximum at a higher clay content. The angle of internal friction for 100% clay specimens, compacted at the optimum water content of 43% to a maximum dry density of 1.17 Mg/m³, was found by Abduljauwad and Rehman (1996) to be 24.5°.

Using basic phase relationships, the degree of saturation (S) was found for various water contents. Fig. 11 and Fig. 12 are replotted using degree of saturation in Fig. 13 and Fig. 14, respectively. These figures show that degree of saturation S is a better strength predictor than the moisture content (w).

4.4. Hydraulic conductivity

The coefficient of permeability (k) was determined from permeability tests for various clay–sand mixtures at different times. Fig. 15a gives the variation of k with elapsed time, which decreases initially with time, then reaches a steady state. The reduction in k with time is because, as the clay minerals absorb water and swell, they reduce the pore space and block the water paths. The absorbed water is strongly attracted to the clay particles, and therefore immobilized. Further, the reduction in k is more pronounced at low compared to high clay contents, owing to the fact that at a low clay content, pore spaces are larger, which allow more freedom for clay particles to freely expand and block relatively larger water paths. Fig. 15b gives the variation in the steady-state value of k with clay content, which indicates a sharp decrease in k with clay content up to a value of about 10%, after which the curve flattens. The value of k for pure sand was obtained as 0.928 cm/s, and that for an undisturbed sample of pure clay was found from a consolidation test to be about 2×10⁻⁹ cm/s. The reduction in k with clay content was observed by Nagaraj and Han, 1998 for bentonite–sand mixtures. The value of k for a 10% clay content is five to six orders of magnitude less than that of a pure sand. The value of k for a 40% clay content is only two orders of magnitude more than that for an undisturbed sample of pure clay. The high value of hydraulic conductivity of samples, with 40% clay content, relative to undisturbed clay can be attributed to the fact that the samples were prepared in a dry state. Also, the value of k for pure clay was obtained using undisturbed sample, and by a different method, i.e. from consolidation test. Lambe and Mitchell showed that the value of k for silty clay compacted at a molding water content dry of optimum is two to three orders of magnitude higher than that compacted at a molding water content wet of optimum. The fabric and structure of the soil are responsible for this variation. Accordingly, it is expected that hydraulic conductivity will only be insignificantly reduced by increasing the clay content above 40%, because non-clay materials become a filler suspended in the clay matrix.

4.5. Swelling potential

Fig. 16 presents the results of swelling tests in the form of percent swell (ΔH/H_o) as a function of elapsed time. The curves for soil with a clay content of 20% or more can be distinctly divided into three stages, representing the initial, primary, and secondary swelling. A high rate of increase in percent swell is observed in the primary stage followed by a low rate in the secondary one. This can be attributed to high water adsorptive forces during the primary stage (Abduljauwad, 1993). Further, a low swelling rate is observed in the initial stage due to the low permeability of the specimens, which reduces the rate of flow of water.

Fig. 17a plots percent swell versus clay content. At low clay content (10%), the primary swelling is not pronounced, due to the fact that the swollen clay particles just occupy the voids between the sand grains causing a relatively non-significant amount of swell. However, for clay content of 20% or more, the percent swell increases at a much higher rate, caused by clay particles dislocating the sand grains and causing the entire soil to heave. It is expected that this high rate of swelling (not the amount of swelling) with respect to clay content will reduce slightly at very high clay content, due to the interaction between diffuse-double-layers. For an undisturbed sample of pure clay (100%) with an initial water content of 38.58%, the percent swell was found to be 40.1%. The percent swell with respect to clay content (Fig. 17a) can be sub-divided into two distinct straight line segments: a flat part corresponding to a clay content <20% and a steep part for a clay content >20%. For a clay content (C%) <20%, the percent swell or swelling potential (SP%) can be approximated by:

For a clay content >20% but not exceeding say 60%:

The swelling potential correlates well with a compositional factor that reflects both the amount and type of clay, i.e. the plasticity index (PI). Fig. 17b shows the results of percent swell versus PI which are fitted by a third-degree polynomial, a second degree polynomial, and an exponential function, with the following equations respectively:

(4)

(5)

Results given in Fig. 17b are similar to those obtained by Seed et al. (1962) for artificial sand–clay mixtures, compacted at an optimum water content using standard AASHTO compactive effort and allowed to swell under a surcharge of 7 kPa (1 psi). The variation of swelling potential as a function of clay content is approximately similar to its variation as a function of PI, Fig. 18.

The effect of clay content on the PI of the mixtures shown in Fig. 1 suggests a straight line variation of PI with clay content (C%), with a slope of about 1.02 representing the activity (A) of the clay. The straight line has the following equation:

This line intercepts the horizontal axis at a clay content of 3.05, which is used to correct the activity for soil mixtures with a clay content of 40% and below, as suggested by Seed et al. (1964b). The activity of soil mixtures as a function of clay content is plotted in Fig. 19, which is superimposed on typical swelling potential curves given by Seed et al. (1962). This figure shows the increase in the swelling potential of soil mixtures as the percentage of expansive clay increases. Moreover, a clay content of 20% or less falls in the range of low swelling potential (≤1.5%) suggesting that when the clay fraction expands it will mainly fill the voids between the sand grains.

5. Conclusions

The properties of artificial clay–sand mixtures are highly influenced by the clay content and the moisture content. The consistency limits (LL, PL, and PI) are found to vary in proportional with the clay content. This variation is found to be almost linear, owing to the fact that the cleaned sand does not actively interact with the clay minerals. An introduction of the new concept of normalized consistency limits makes it easy to collapse three curves (for the variation of LL, PL, and PI with clay content) into one line, and to help detect any possible physicochemical interaction between the different soil constituents.

The variations in the stress–strain relationship and strength parameters (c and φ) as a function of moisture content of a clayey sand were presented. For dry mixtures, the cohesion and angle of internal friction are not significantly affected by the amount of clay. For other values of water content, cohesion increases with increasing clay content, and it also increases with increasing water content to a certain limit, above which it decreases. Generally, the angle of internal friction decreases with increasing clay content or water content. The degree of saturation was found to give better interpretation of the variation of the strength parameters than the water content. The combined effect of clay content and moisture content or degree of saturation on the cohesion and friction angle, presented as 3-D plots, highlights the importance of the coupling effect of compositional and environmental factors even in simple geotechnical and foundation engineering problems. These illustrate that neither c nor φ are inherent properties of the soil material but, on the contrary, they are dependent on the conditions operative in the laboratory or on the environmental conditions in the field.

Hydraulic conductivity is found to decrease by five to six orders of magnitude with the addition of 10% clay to the fine sand, but it is expected to have an insignificant reduction above a clay content of 40%.

The variation of the swelling potential of a clay–sand mixture is bi-linear with respect to clay content, as shown in (2) and (3). These equations do not include a parameter for the clay type (e.g. activity) because the activity of this clay is found to be almost one. Better relations in terms of PI (which represents both the amount and type of clay) are also given in (4), (5) and (6).

The results obtained in this investigation show the dominating influence of a clay fraction on the behavior of soil even if it is as low as 10% of the entire soil. They also show the effect of water content on mobilizing the clayey behavior of the soil. Although these results were obtained for a specific clay mineral and particular sand, and therefore may not be generalized, the basic concept of the coupling effect of clay content and moisture content (or degree of saturation) on various soil properties can still be utilized to interpret the behavior of unsaturated clayey soils. Practical applications of this study include a wide spectrum of problems in geotechnical, foundation, and geoenvinronmental engineering which involve either the effect of wetting and drying on the behavior of natural clayey sands and sandy clays, or the effect of the clay content on the behavior of unsaturated soils.

Acknowledgements

The author would like to acknowledge the support of King Fahd University of Petroleum and Minerals in providing computing and laboratory facilities. The assistance of Mr Hasan Zakaria and Mr Shahid Azam is also appreciated.

Fax: +966-3-860-2879; email: nshayea@kfupm.edu.sa