Ambivalent Role of Water in Thermodesorption of Hydrocarbons from

Dec 6, 2010 - The ambivalent role of soil moisture in thermodesorption of ... alkanes) have revealed an ambivalent influence of water on desorption ra...
1 downloads 0 Views 774KB Size
Environ. Sci. Technol. 2011, 45, 732–737

Ambivalent Role of Water in Thermodesorption of Hydrocarbons from Contaminated Soil ULF ROLAND,* FRANK HOLZER, AND FRANK-DIETER KOPINKE Helmholtz Centre for Environmental Research - UFZ, Department of Environmental Engineering, Permoserstrasse 15, 04318 Leipzig, Germany

Received September 3, 2010. Revised manuscript received November 18, 2010. Accepted November 18, 2010.

Thermodesorption studies with soil samples from a former filling station for light crude oil contaminated with mineral oil hydrocarbons (mainly benzene, toluene, ethylbenzene, xylenes, naphthalene, alkylnaphthalenes, and C10 to C14 alkanes) have revealed an ambivalent influence of water on desorption rates. Particularly, the influences of soil moisture content, humidity of the purge gas, temperature, and content of soil organic matter (SOM) were studied. At low temperature, purge gas humidity strongly affected the mobility of hydrocarbons in the soil organic matter (SOM) leading to an enhanced release of contaminants at higher moisture contents. Heating resulted in a decrease of thermodesorption when connected with desiccation of soil, in spite of the strong temperature impact on the vapor pressure of contaminants. At high water content of the SOM, the transfer of the pollutant molecules into the gas phase was found to be markedly hindered by the formation of water films or pore-filling by bulk water, both acting as diffusion barriers.

Introduction Sorption of anthropogenic hydrophobic organic compounds (HOC), often representing hazardous soil contaminants, is mainly related to soil organic matter (SOM). The complex nature of SOM with respect to its chemical and physical structure (aromatic and aliphatic moieties, functional groups, hydrogen bonds, porosity, size, and metal content) results in a wide spectrum of factors influencing the interactions between HOC and SOM. Hence, different models are used to describe the observed phenomena (partitioning, adsorption, absorption, dual phase model of SOM, and phase transitions 1-4). Besides their sorption state, the fate of organic pollutants in soil during thermal treatment is determined by their mobility and their ability to be transferred into the water or gas phase, which are related to the factors mentioned before (5, 6). Attempts to clean the soil by soil vapor extraction (SVE) directly depend on these conditions, which are therefore also in the focus of remediation research. In the case of thermally enhanced SVE (e.g., using thermal wells, steam injection, low- or radio frequency heating 7-10), both temperature and soil moisture content can be controlled to a certain extent in order to optimize the cleanup process. However, for this purpose the combined influence of these parameters on the behavior of HOC in soil has to be clarified. * Corresponding author phone: ++49 (341) 235 1762; fax: + +49 (341) 235 45 1762; e-mail: [email protected]. 732

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 2, 2011

For hydrocarbon release, at least two potentially ratedetermining steps have to be considered: diffusion within the SOM and transfer into the gas phase. According to a model proposed by Xing and Pignatello (1) the SOM polymer may consist of two structural states: a glassy one and a more elastic, rubbery one. In the glassy state, almost-isolated holes with a diameter in the range of a few nanometers occur, strongly restricting the mobility of HOC. In contrast, the structure in the rubbery state is more flexible and the interaction of HOC with the SOM can be adequately described by the partitioning model. As a result, HOC bound to the rubbery fraction of the SOM can be more easily desorbed whereas organic molecules in the glassy fraction can hardly leave the holes via diffusion through the glassy matrix. The transition from the glassy to the rubbery state (sometimes described as a second-order phase transition) can be initiated either by temperature increase or by addition of water thus acting as a plasticizer. The first mechanism could be shown by differential scanning calorimetry (DSC) measurements for some natural polymers (3, 4). Unfortunately for studying SOM this method is constrained by the natural moisture content of SOM. Alternatively, the structure-modifying effect of water was indirectly proven by an increase of mobility during desorption measurements, e.g., for toluene on a humic substance provided by Roth (11). Pulsed field gradient nuclear magnetic resonance (PFG NMR) measurements also showed a discontinuous temperature effect on the HOC mobility in a humic substance at a microscopic scale (12). However, the results of kinetic desorption measurements for characterization of phase transitions in SOM are not completely consistent, thus inspiring an ongoing discussion in the literature. For many materials, DSC measurements gave no indication for a second-order phase transition. Additionally, it was shown by NMR spectroscopy that water molecules can also bridge molecular segments of SOM via intramolecular and intermolecular cross-links, thus leading to higher matrix rigidity, hindering diffusion (13). The various observations presented in the literature reflect the wide variety of SOM structures and properties present in nature (4, 5, 14, 15). Many studies on aging and sorption hysteresis have shown that the SOM should be considered more as having a quasi-stationary state rather than a fixed structure (5, 16, 17). The objective of the present work was to evaluate the interplay between soil moisture and temperature for the thermodesorption of organic contaminants, using an original polluted soil originating from a former industrial site.

Materials and Methods Origin of the Samples. The contaminated sandy soil samples originated from a filling facility for light crude oil of a former lignite pyrolysis factory in Espenhain south of Leipzig, Germany. The soil was heavily contaminated with a wide spectrum of BTEX aromatics (benzene, toluene, ethylbenzene, xylenes), naphthalene, alkylnaphthalenes, and alkanes (mainly C10 to C14). A number of samples were taken from a depth of about 1 m below ground. They were stored in a sealed container at about 0 °C to minimize loss of volatile organic compounds (VOCs). In the following, all contents of water and hydrocarbons are related to the mass of the dry samples. Thermal Treatment and Desorption Experiments. In this study, two types of investigations were carried out. The first series of experiments was conducted using a shallow bed of soil to directly monitor the desorption process with limited macroscopic migration of the hydrocarbons through the soil 10.1021/es102778h

 2011 American Chemical Society

Published on Web 12/06/2010

bed. For different sample amounts, either a carbon analyzer or a heatable sample holder connected to a mass spectrometer were employed. In a second series, a soil column was used to study desorption of contaminants which is more comparable with thermally enhanced soil remediation in practice. Experiments with Carbon Analyzer and Mass Spectrometer. Thermal treatment of the samples and desorption studies were carried out in a commercial carbon analyzer C-MAT 5500 (Stro¨hlein, Karst) equipped with a CuO furnace operated at 750 °C for total oxidation of desorbed compounds. The resulting CO2 was measured by a nondispersive infrared detector. A shallow bed of the soil sample (depth about 3 mm) was purged by air, oxygen, nitrogen, or helium as carrier gases. Although the analyzer allowed heating the samples up to 1200 °C with various heating rates, heating in the temperature programmed desorption (TPD) experiments was usually limited to 500 °C in this study. The sample mass was in the range of some grams, thus enabling representative characterization of soil samples with a certain inhomogeneity. Additionally, a tube furnace coupled with a quadrupole mass spectrometer (Shimadzu QP 1100) was used for smaller sample amounts (thermodesorption-mass spectrometry, TD-MS). The sample holder consisted of a thin quartz tube (200 mm × 2 mm) placed inside the heating zone. The sample (10-30 mg, filling height