Efficiency of soil washing in hydrocarbon removal: the role of temperature, surfactant concentration and soil mineralogy

The object of the current PhD research is to evaluate the effect of variations in the type and concentrations of surfactants, washing temperature and most importantly, mineral composition of the soil on washing efficiencies. In order for surfactant-enhanced soil washing to be a viable and efficient remediation strategy a series of requirements must be met. First, the soils being treated cannot have clay contents higher than 30% and organic materials exceeding 2 wt%, as it would result in excessive loss of surfactant to adsorption and subsequent difficulty in the removal of the washing solution. In addition, the soils should also have relatively low cation exchange capacity (CEC). The study was articulated into three distinct phases. In the first phase fourteen samples representative of the various soils present in the Marche Region were collected and their grain size distributions, organic contents, cation exchange capacity, mineral composition and Atterberg limits were determined. Based on their grain size distributions, the soils are characterized by sand fractions between 7% and 49%, silt fractions between 2% and 49% and clay fractions between 2% and 52%. They have organic contents of 0.21 to 6.37 wt%, CEC’s of 7.2 to 35 cmol+/kg and plasticity indexes that range from non-determinable (for non-plastic soils) to 25. The studied soils are rich in calcite, quartz and mica, and have minor amounts of clay minerals, feldspars and plagioclases. It was concluded that, in the event of contamination, most of the sampled soils would be unsuitable for surfactant-enhanced soil washing due to their high clay and organic contents. Of the remaining soils, two were chosen, MF-05 and MF-10, for the experimental phase of the project, based on their different mineralogy and locations in close proximity to population centers. Sample MF-05 has equal proportions of quartz, calcite and mica family minerals, while MF-10 has no calcite and lower CEC, organic contents and plasticity index. The second phase of the project was focused on the development and fine-tuning of the laboratory washing procedure using the procedure described in Urum et al. (2014) as a setting off point. The procedure had to be modified to take into consideration a different contaminant and analytical method. Each step of the washing and extraction of residual hydrocarbon procedure was assessed. The final procedure involved the artificial contamination of the soil samples with 15 wt% diesel followed by washing in a temperature-controlled water bath for 30 minutes with surfactant- based washing solutions in a ratio of 4 to 1. The samples were then rinsed with deionized water, dried in an oven set to 60°C. The diesel remaining in the sample was then extracted using n-Hexane and analyzed using GC-MS (Gas Chromatography – Mass Spectrometry). Two surfactants were used in the experiments: SDS (anionic) and TWEEN80 (non-ionic). The experiments were carried out at two washing temperatures (30°C and 40°C) and three surfactant concentrations. One below the critical micelle concentration (CMC) and two above. Experiments using deionized water as the washing liquid were used as a baseline of comparison. It was found that for sample MF-05, washing with SDS-based washing solutions had little effect at concentrations below the CMC, improving drastically once the CMC was exceeded; in sample MF-10, on the other hand, a drastic improvement in washing efficiency was already evident at concentrations below the CMC and continued to increase, at a lower rate, with increasing SDS concentrations. The initial delay in washing efficiency improvement in sample MF-05 is likely due to the adsorption of the negatively charged surfactant molecules to the positively charged calcite surfaces, absent in sample MF-10. For both samples, washing with TWEEN80-based solutions resulted in a significant improvement in efficiency at concentrations below the CMC, followed by less significant improvement at intermediate concentrations and a decrease in efficiency at the highest concentrations of TWEEN80. The similar performance of the TWEEN80 washing solutions in samples MF-05 and MF-10 is consistent with its neutral charge. Temperature did not produce significant differences in surfactant performance. When compared to each other, as expected, SDS slightly outperformed TWEEN80 and both are more effective at extracting n-Alkanes from sample MF-10 than MF-05. The decrease in TWEEN80 efficiency at higher concentrations may be due to a change in shape or loss of stability of micelles.