The Khorramabad-Zal highway is a road constructed in Lorestan province (southwest of Iran). With a length of 104 km, it connects Khorramabad and Andimeshk cities. Performing many excavations in sandstone outcrops along this highway has created stable and unstable slopes. The release of Na+ and SO42- ions due to weathering and dissolution of the marl and gypsum rocks has formed Na2SO4 salt in this area. Permeating this salt into the sandstones has led to the salt crystallization process in them. The present study investigates sandstones’ physico-mechanical properties in the stable and unstable slopes before and after salt crystallization test up to 15 cycles. To this end, 14 sandstone samples belonging to stable and unstable slopes were collected. Next, their physico-mechanical properties are characterized by determining their porosity (n), point load index (PLI), Schmidt hammer hardness (SHH), and slake durability index (SDI) before and after salt crystallization test in Na2SO4 solution (i.e., untreated and treated samples, respectively). The results indicate that samples of unstable slopes in untreated conditions generally have a higher n, and lower PLI, SHH, and SDI than those from the stable slopes. With increasing the salt crystallization test cycles, the samples’ n increased, while their PLI, SHH, and SDI decreased with different levels depending on the slope types. Overall, the deterioration effects of salt crystallization were more significant on the samples’ physico-mechanical properties belonging to unstable slopes, especially those subjected to salt crystallization under field conditions. This result was confirmed by the rate of loss (RL) of the samples’ physico-mechanical properties, and microscopic and macroscopic observations after the salt crystallization test. The deterioration in the physico-mechanical properties of sandstone samples caused by salt crystallization in field conditions increased slope instability near marl and gypsum rock outcrops. Finally, regression analyses showed that the SDI has good correlations with n, PLI, and SHH with determination coefficients (R2) of 0.919, 0.882, and 0.837, respectively. These correlations are accurate enough for use in the preliminary assessment of samples SDI from their n, PLI, and SHH. Finally, these correlations obviate the need to perform SDI as a laborious and time-consuming test.
Read full abstract