Suction and strength in heavily compressed carbonate sand

Abstract

When oven-dried powdered clays are heavily compressed they acquire tensile strength, which is variously attributed to bonding or to suctions in water not evaporated during oven drying. When coarse-grained soils are heavily compressed the grains fracture (Coop, 1990; McDowell & Bolton, 1998). The tests described here show that, after compression under stresses of the order of 100 MPa, oven-dried carbonate sand develops tensile and compressive strengths of the order of 1 MPa due to relatively large suctions in small quantities of water released from within the pores of the carbonate material. Tests were carried out on Dogs Bay sand. This soil has been described in detail by Coop (1990). It is a biogenic carbonate sand consisting largely of foraminifera and mollusc shells; the carbonate content was reported by Houlsby et al. (1988) as around 88–94%. In its natural state it is poorly graded, with sizes in the range 0.05–1 mm. Samples were oven dried and compressed one-dimensionally in a thick-wall steel cylinder, nominally 38 mm diameter, to vertical stresses up to about 170 MPa. (This corresponds to depths of burial of the order of 20 km.) The steel cylinder was lined with a thin plastic tube to reduce side friction and to aid extraction after compression. Thin samples were made in one layer and, after one-dimensional compression, they were tested by loading across a diameter (the Brazil test). Other samples, 76 mm long, were made in several layers and, after one-dimensional compression, they were loaded in triaxial compression tests with a confining pressure of 100 kPa and with atmospheric pore air pressure. Although they had been initially oven dried, after heavy one-dimensional compression the samples had acquired small but measurable water contents, generally in the region of 0.05–0.2%. The most probable explanation is that there were small quantities of water trapped within self-contained pores in the carbonate material, which was released as particles fractured. Also, the samples may have gained small quantities of water from humidity in the laboratory atmosphere. All samples developed considerable strength after heavy compression, but on immersion in water in a beaker they immediately disintegrated and formed a cone of sand in the bottom of the beaker. It is unlikely that any significant bonding would be developed during the relatively short period (about 5 min) for which the samples were under heavy loading, and strong bonding would not be expected to disappear immediately on immersion in water. Hence the measured strengths must have been largely due to suctions. Figure 1 shows grading curves determined by wet sieving of samples before and after one-dimensional compression to 70 MPa. These show that the effect of compression was to reduce particle sizes by fracturing, except that the maximum size was unchanged. This is characteristic of the change of grading of coarse-grained soils (McDowell & Bolton, 1998). From the results of the Brazil tests the tensile strength was calculated from