4.General Points

Expansive soils undergo volume changes upon wetting and drying. For a volume change to occur these soils must be initially unsaturated at some water content wo. When the water content changes to a new value w1, the volume increases if w1 > wo or decreases if w1 < wo, unless wo is the shrinkage limit where wo = ws. These soils occur in an active zone, which starts at the ground surface and goes down to the saturated part of the zone of capillary rise above the ground water table. Expansive soils are mostly found in arid and semiarid areas worldwide and contain large amounts of lightly weathered clay minerals. Low rainfall has hindered the weathering of more active clay minerals such as the smectite family to less active clay types such as illite or kaolinite, and the rainfall has not been enough to leach the clay particles far enough into the strata that the overburden pressure can control the swell. In general, all clayey soils tend to shrink on drying and expand when the degree of saturation S increases. Usually, the lower the shrinkage limit ws and the wider the range of the plasticity index Ip, the more likely is volume change to occur (Table 7-1) and the greater the amount of such change. Volume change is particularly troublesome in large areas of the southwestern United States, India, and Australia, and in parts of Africa and the Middle East that are subject to long dry periods and periodic heavy rains of short duration. The dry periods tend to desiccate the soil; then the sudden rainy season(s) cause large amounts of swelling near the ground surface. There is not enough regular rainfall to leach and weather the troublesome clay minerals to greater depths; thus, they remain unaltered near the ground surface. During the rainy season(s) they are rapidly wetted and quickly swell to form a water barrier to further water entry, thus keeping the problem near the ground surface. Soils in these areas are particularly troublesome to build on as they appear competent during dry periods (with the possible exception of surface tension cracks). They would of course remain competent if their water content is controlled in some manner, a difficult task. What happens is that water vapor migrating from the ground water table, which may be at a depth of many meters, condenses on the bottom sides of the floor slabs and footings. Anyone can readily observe this phenomenon by turning over a flat rock in the field (even after a prolonged dry spell) and noting the dampness on the underside. Since a building is somewhat impermeable, similar to the flat rock in the field, the soil in the interior zone eventually becomes wet to saturated from the condensation of upward-rising water vapor. The soil will then swell unless the building provides sufficient weight to restrain the swelling pressure, and buildings seldom provide the huge restraining pressures required to control swelling. Cold storage buildings with an uninsulated thermal gradient may cause condensation of the water vapor in the soil or create an upward flow of water vapor from the water table. Ice lenses may form. These typically are more serious from either of two reasons: amplifying swell or producing a semifluid zone when they melt if the temperature is sufficiently low. With buildings, in addition to possible evaporation of soil water from the perimeter zone, there is also the problem of soil in this region becoming desiccated from water absorption through the roots of adjacent shrubbery and/or trees used in landscaping. Loss in soil moisture by evaporation from heating the building or from beneath or adjacent to heating units such as boilers can also create shrinkage volume changes. Shrinkage tends to produce perimeter settlements (unless from heating units where the interior may settle), which, in combination with any interior swell, develop larger differential settlements than would be obtained from swelling action alone. By the way, the field rock also has this problem (it is seldom wet from edge to edge) but is of no consequence. In all of the shrinkage cases the amount of volume change is referenced to the initial natural water content wo of the soil and the current natural water content w\. Volume changes produced from shrinkage {w1 < wo) are usually smaller than volume changes from swell (w1 > wo). Table below may be used as a guide in evaluating the potential for volume change of soils based on easily determined index properties. In part, this table is a summary of data from Holtz (1959) on several soils, which are correlated with some 50 soils from other areas, including a large number of Indian black cotton soils by Dakshanamurthy and Raman (1973). In terms of relative values a “low” volume change might be taken as not more than 5 percent whereas “very high” could be interpreted as over 25 percent.

5.Designing Structures on Soils Susceptible to Volume Change

Structures built on expansive soils require special construction techniques for their foundations. When the problem is identified, one may address it in several ways:

1. Alter the soil. For example, the addition of lime, cement, or other admixture will reduce or eliminate the volume change on wetting or drying.

2. Compact the soil well on the wet side of the optimum moisture content (OMC). This process produces a lower than maximum dry density and if the water content does not change until the structure is built there should be little swell. Remember, it is not the water content but the change in water content that produces soil volume change. Soils compacted well on the wet side of the OMC are usually nearly saturated [see Gromko (1974), with large number of references]. Often, however, this compaction state may not have sufficient strength for the design requirements.

3. Control the direction of expansion. By allowing the soil to expand into cavities built in the foundation, the foundation movements may be reduced to tolerable amounts. A common practice is to build “waffle” slabs (see Fig. 7-8) so that the ribs support the structure while the waffle voids allow soil expansion [BRAB (1968), Dawson (1959)]. It may be possible to build foundation walls to some depth into the ground using tiles placed such that the soil can expand laterally into the tile cavities.

4. Control the soil water. The soil may be excavated to a depth such that the excavated overburden mass of soil will control heave, lay a plastic fabric within the excavation, and then backfill. The rising water vapor is trapped by the geotextile, and any subsequent volume change is controlled by the weight of overlying material and construction. The surface moisture will also have to be controlled by paving, grading, etc.

5. Check whether a granular blanket of 0.3 to 1 m or more depth will control capillary water and maintain a nearly constant water content in the clay [Gogoll (1970)].

6. Ignore the heave. By placing the footings at a sufficient depth and leaving an adequate expansion zone between the ground surface and the building, swell can take place without causing detrimental movement. A common procedure is to use belled piers (Fig. 7-9) with the bell at sufficient depth in the ground that the soil swell produces pull-out tension on the shaft or the whole system heaves. Small-diameter pipes with end plates for bearing can also be used to isolate smaller structures from expansive soil. The pier or pile shaft should be as small as possible to minimize the perimeter of the shaft so that soil expansion against the shaft does not produce tension or compression friction/adhesion stresses from vertical movement sufficient to pull the shaft apart or crush

1. Adhesion on the pier shaft can be minimized by using a slightly oversized hole or by surrounding the shaft with straw or other porous material such as sawdust to reduce adhesion. The foundation system movements should be stabilized by the time any organic material used has rotted. Load the soil to sufficient pressure intensity to balance swell pressures. This method is used in many fills where the fill weight balances the swell pressure. This technique can also be used beneath buildings either by using spread footings of high pressure intensity or by excavating several feet of the clay and backfilling with granular material. The backfill in combination with foundation pressures may contain the swell. This method may not be practical for one-story commercial buildings and residences because of the small soil pressures developed from foundation loads. Heave of expansive soils is difficult to predict, since the amount depends on the clay mineralogy, particle orientation, confining overburden pressure, and the difference between the current and reference (initial) in situ water content. Estimates of heave may be obtained from standard 1-D consolidation tests (ASTM D 2435) in which the sample is compressed and then allowed to rebound. The slope of the rebound curve is related to swell. The ASTM D 4546 method using consolidation test equipment also can be used to obtain swell estimates. It should be used instead of D 2435 where possible since the data obtained is more directly related to swell. Basically we can obtain a swell curve by confining the sample in the ring using a very small confining pressure of about 7 kPa (1 psi) and allowing it to absorb water and swell. If we measure the volume change in these conditions, we have a free swell test. If we apply sufficient consolidation pressure to prevent the sample from expanding, we can measure the swell pressure required to maintain the zero volume change. These data can be directly extrapolated to the expected heave or to the footing/overburden pressure required to eliminate, or at least reduce, the swell movements. The resulting estimates improve with sample quality and careful attention to test details. The estimate also improves if the current natural water content wo, and the degree of saturation for the laboratory volume change, are representative of the long-term in situ value. The latter is a very important consideration since the laboratory sample is thin and has access to sufficient free water to obtain S = 100 percent in a short time; this may never occur in the field, at least through the full depth id of the zone (Fig. 7-6) with potential to expand or shrink.