Vapor Condensation Technique for Measuring Stress-Strain Relation of Unsaturated Soil

All tests for stress-strain laws should be conducted under uniform stress conditions within the entire testing specimen. This is rarely the case in most conventional tests for saturated soil. Nevertheless, such nonuniformity has been minimized in all standardized tests so that the stress-strain data are meaningful for the development of constitutive laws. However, in the common unsaturated soil tests such as the wetting in odometer and triaxial setups, the uniformity condition has been seriously violated. Upon the abrupt introduction of liquid water or decrease in suction to the specimen, a sharp wetting front in the specimen is often unavoidable, resulting in nonuniform suction stress changes. Often under some preloaded stresses, localized failure along the wetting front could occur and propagate, resulting in nonuniform strain within the specimen. Unfortunately, such behavior has been used to invalidate the use of effective stress (e.g., Jennings and Burland 1962), and to establish some popular concepts for the soil’s constitutive laws under unsaturated conditions.

To ensure a uniform suction stress or water content change within a specimen, a vapor condensation (VC) technique is explored to imbibe a cake-shaped specimen. The theory and the working principle of the VC are similar to the drying cake (DC) technique (Lu and Kaya 2013) except here the wetting is regulated by injecting saturated vapor into a chamber enclosing the specimen. The VC technique minimizes nonuniformity of suction stress or water content change in the specimen. Equilibrium for an increment of water content is ensured before the corresponding displacement field is measured by the particle image velocimetry. Because the test is under zero total stress, the radial displacement field shown in Fig. 1 is solely due to the changes in suction stress and is used to compute the drying and wetting suction stress characteristic curves (SSCCs).

Striking results, countering to many previous interpretations and conceptions, are observed in different soils and are illustrated in Fig. 1 for a specimen of compacted Georgia kaolin. Drying from full saturation, the volume continuously reduces (cake a). At and below a water content approximately $0.29$, the volume starts to increase (cakes b and c). This switch in displacement direction indicates that effective stress first decreases and then increases during the drying process (Fig. 1). A reverse trend is observed during subsequent imbibition, where the volume first decreases (cakes d and e), and then increases after the water content increases to above 0.29 (cake f). These reversals in displacement direction can be accurately captured by the SSCC-based effective stress theory (Lu and Kaya 2013) and soil water retention transition between hydration and capillary mechanisms (Lu 2016), indicating the validity of effective stress representation with the SSCC. The occurrence of a minimum suction stress ${\sigma }_{\min }^s$ of $-216\mathrm{kPa}$ for the drying in this soil has been previously validated by several independent methods (Lu and Kaya 2013).