Date of Award


Document Type

Open Access Dissertation


Civil and Environmental Engineering

First Advisor

Sarah Gassman


Soils of Pleistocene age in the South Carolina Coastal Plain have experienced liquefaction due to historic and pre-historic earthquakes. Numerous field and laboratory studies have shown that aged soil deposits maintain a greater resistance to liquefaction than younger soil deposits. The currently available methods for assessing liquefaction potential are based on cases in which soils are of Holocene age or younger (< 10,000 yrs). The Pleistocene age soils that were tested and characterized varied in age from about 200,000 years old to 1,400,000 years old. Several sites were investigated using field methods that included the seismic cone penetration test, cone penetration test, standard penetration test, and flat plate dilatometer. Piezometers were installed at the sites. Undisturbed soil samples were retrieved from the subsurface and frozen ex situ to minimize sample disturbance during transportation and laboratory handling. The undisturbed samples were used for cyclic triaxial testing in the laboratory and were tested for shear wave velocity and compression wave velocity using in-cell transducers. Laboratory tests were performed to determine the specific gravity, grain size distribution, moisture content, unit weight, Atterberg limits, Unified Soil Classification, and visual-manual description. Optical petrography and scanning electron microscopy were used to determine the mineral content of the soils, to view grain characteristics, and to view microscopic features that were part of the soil aging process.

Laboratory index tests showed that Pleistocene soils consisted predominately of poorly-graded fine sands, silty sands, and clayey sands. Shear wave velocities from the cyclic triaxial test specimens were comparable to the in situ shear wave velocities measured using the seismic cone penetration test. Compression wave velocities from the cyclic triaxial specimens were indicative of a saturated state in the soil prior to cyclic triaxial testing. The optical petrography showed that the dominant mineral in the sands consisted of quartz, which was accompanied by minor amounts of mica, feldspar, and opaque minerals. Scanning electron microscopy indicated the presence of kaolin, showed alteration features on quartz sand surfaces, and also showed the presence of soil fabric in the form of preferred grain orientation. Field testing using the standard penetration test and the cone penetration test indicated that the Pleistocene soils maintain a higher cyclic resistance ratio than the Holocene soils used in the current methods of analysis, however, the soils remain susceptible to liquefaction given expected peak ground accelerations where the cyclic stress ratio exceeds the cyclic resistance ratio.

Based on the known ages of the soils, the two methods of analysis using the cone penetration test (Idriss and Boulanger, 2008 and Youd et al., 2001) showed out-of-sequence age versus cyclic resistance ratio for the Idriss and Boulanger method and a properly sequenced age versus cyclic resistance ratio for the Youd et al. method. The standard penetration test showed out-of-sequence age versus cyclic resistance ratio for all methods and the difference between the Holocene liquefaction curve and the Pleistocene liquefaction curve was less than the difference for the cone penetration tests. Field cyclic resistance ratios derived from the laboratory cyclic triaxial tests, which were adjusted for bi-directional motion and in situ stress, resided at or below the cyclic resistance ratios determined for the Pleistocene soils from the field tests and in some cases below the Holocene liquefaction curve.