Environ. Sci. Technol. 2007, 41, 8447–8452
Results of Field Tests on Radio-Wave Heating for Soil Remediation U. ROLAND,* F. HOLZER, D. BUCHENHORST, AND F.-D. KOPINKE UFZ - Helmholtz Centre for Environmental Research, Department of Environmental Technology, Permoserstrasse 15, D-04318 Leipzig, Germany
Received June 8, 2007. Revised manuscript received September 14, 2007. Accepted October 17, 2007.
After developing the radio-wave technique for various conditions in laboratory-scale and technical plant-scale experiments, field tests in combination with biodegradation and soil vapor extraction were carried out at three sites: (i) a bioremediation facility for ex situ cleaning of soil, (ii) in situ remediation of contamination at a former storage facility for organic solvents, and (iii) a polluted soil under a former petrol station. Various electrode arrangements such as parallel plates, rod arrays, and coaxial antenna were applied in order to meet the site-specific requirements optimally. Soil temperatures between 35 and 100 °C were established. The successful tests gave much insight into the engineering, physical, biological, and chemical aspects of radio-wave application. General conclusions on the appropriateness and competitiveness of the radio-wave method as well as on preferred application fields are drawn.
Introduction The application of electromagnetic waves in the radio frequency range to heat soil formations was described more than 20 years ago (1). Starting from the utilization of dielectric heating to better exploit a reservoir of oil shale, radio-wave (RW) heating was also used to thermally support soil remediation, namely by thermally enhanced soil vapor extraction (TESVE) (2–6). In principle, two different types of electrodes were used to introduce RW energy into the soil and to realize heating: horizontally or vertically positioned rod-shaped electrodes or RW antennae (3–7). To treat larger volumes, various arrangements have been developed. In the case of rod electrodes, three parallel sets of so-called “cold” and “hot” electrodes are a preferred option. In principle, this arrangement (mostly with the “hot” electrodes in the middle row, cf (4, 5).) can be compared with a plate capacitor. However, field strength and consequently temperature distribution are far from being homogeneous in the soil between the electrodes, because the electric field is concentrated around the rod electrodes. The same problem is inherent to coaxial antennae as RW emitters (7). Nevertheless, they allow the RW energy to be transferred into the desired soil volume in a relatively simple way and the heat transport in the soil can be utilized to obtain more homogeneous temperature * Corresponding author phone: ++49 341 235 2581; fax: ++49 341 235 2492; e-mail:
[email protected]. 10.1021/es071369s CCC: $37.00
Published on Web 11/15/2007
2007 American Chemical Society
profiles. Such a locally positioned energy input can be realized by rod electrodes only after modification with an air gap (see ref (8)). Over the years, a number of applications of RW heating have been described in the literature (1–9). Most of these represent rather pragmatic and empirical studies. Detailed analyses concerning temperature profiles, distribution of pollutants before and after treatment, and variation of the electrode geometry are often missing. Therefore, investigations giving more insight into the engineering and physicochemical aspects are still necessary. Additionally, studies related to the combination of RW heating with biological processes are still lacking at a larger scale. Laboratory tests have shown that the application of RW does not inhibit the microbial activity (10), thus allowing a utilization of RW heating to support biodegradation. In the present paper, the suitability of RW heating for optimizing soil remediation is demonstrated for various situations at the field scale. The studies include a detailed analysis of temperature profiles and heating rates, conversion and removal of organic pollutants and activity of microorganisms (MO) in the case of biodegradation. Fiber-optical temperature sensors (11) allowed monitoring of the temperature at various positions in the soil continuously, i.e., during RW application. The desired temperatures were stabilized by applying RW pulses when the temperature was below a defined value. The power was switched off or reduced to a very low value when the desired temperature level was reached (details in ref (8)). This was essential for the biodegradation experiments. The details of the experimental setup were different for the distinct field sites and are, therefore, described in the corresponding sections. In the following, the respective conditions at every field site including the properties of the contaminated soil, the arrangement for thermally enhanced remediation, the process course, and the results are presented for the three reference projects. The discussion will clarify general aspects, engineering options, and conclusions for RW-supported remediation tasks. Due to the validation and demonstration character, the tests cannot be understood as fully commercial reference remediation projects.
Field Site 1: Thermally Enhanced Bioremediation Description of the Site and the RW Arrangement. The longterm tests for thermally enhanced microbial degradation were carried out in a full-scale plant for soil bioremediation at Hirschfeld (Saxony, Germany; facility of Bauer & Mourik Umwelttechnik, Schrobenhausen). Two soil reactors with a maximum volume of approximately 20 m3 (3 × 3 m2 base, about 2.5 m height) were constructed, one of them having the RW heating option (RW) and the other being a reference (REF). The complete apparatus is schematically presented in Figure S1, Supporting Information. The apparatus consisted of the RW power source (RW generator IS 15 with a maximum power output of 15 kW working at a fixed frequency of 13.56 MHz), the matching network (matchbox PFM 30000A; both from Hüttinger Elektronik, Freiburg/Brsg., Germany), two parallel plate electrodes in the heatable soil reactor, the electromagnetic shielding (copper gauze), fiber-optical temperature sensors, measuring instruments for electrical field strength, off-gas cleaning by charcoal as an option for the thermodesorption mode, devices for the analysis of the gaseous effluent (with respect to hydrocarbons, CO2, O2), and the data transfer and controlling system. VOL. 41, NO. 24, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8447
A perforated tube made of PTFE was placed in the soil and connected to a pump realizing a flow between 60 L/h and 5 m3/h taking gas from the soil volume thus ensuring the aeration of the soil. Concentrations of hydrocarbons, CO2, and O2 were measured by bypassing a small gas stream from the gas effluent of the soil through gas monitoring devices. Characterization of the Contaminated Soil before Treatment. The treated gravely sandy soil originated from a heavily contaminated industrial site (fuel oil leakage, ground mixed with filling material). Prior to transport and remediation, lime was added to the soil by the site owner leading to an extremely high pH value of 11.6. The total carbon (TC) content (C-MAT 5500 analyzer from Ströhlein, Karst, Germany) was about 19 wt.% including a significant organic contamination which will be characterized later. Largely varying TC values revealed a significant heterogeneity of the soil (8). Elementary analysis showed nitrogen contents between 0.2 and 0.3 wt.% which led to the necessity of fertilization (Basfoliar 12, BASF; 12% N, 4% P2O5, 6% K2O) during the long-term experiment (details in ref (8)). The extent of hydrocarbon contamination was about 3 wt.% in total. The spectrum of pollutants ranged from mineral oil constituents (approximately 2.2 wt.%, mainly n-alkanes from C13 to C30) over BTEX aromatics ( 0.2 wt.%) and phenols (0.3 wt.%) to PAHs ( 250 ppm, consisting mainly of naphthalene, acenaphthylene, and phenanthrene). The concentration of alkenes was unexpectedly high and represented 30% of that of the corresponding n-alkanes. Initially, the microbiological activity and the degradation rate were negligibly low. The long-term experiment began in winter with ambient temperatures between 0 and 10 °C also being characteristic for the treatment hall. Temperature Regime during the Long-Term Test. During the long-term experiment with the real contaminated soil, several options of RW-supported remediation were to be tested. Therefore, the temperature in the RW reactor was first established at a mean value between 35 and 37 °C to initiate microbial degradation. After about 55 days, microbiological self-heating led to even slightly higher temperatures in the RW reactor which made external heating unnecessary. After about 230 days, the temperature was enhanced to about 60 °C to evaluate the option of utilizing themophilic microorganisms for biodegradation. After this phase, the temperature was decreased again to about 35 °C. Subsequently, heating was intensified to change the method of remediation from biodegradation to thermodesorption. This flexibility of RW heating can allow the accepted residual concentrations to be obtained under various conditions using only one technical arrangement. Depending on the humidity of the soil, final temperatures above 100 °C were achieved. The mean temperature in the RW reactor in comparison to the temperatures in the reference reactor (REF) and the ambient temperature in the hall are represented in Figure S2, Supporting Information. Oscillations in the ambient temperature were due to the daily course. It was obvious from the temperature rises that during the summer months the biological processes were accelerated in the fertilized soils (RW, REF), whereas a deposited soil without aeration and fertilization (CONV) was not spontaneously warmed up above the ambient temperature. Influence of RW Heating on Soil Respiration. Soil respiration was quantified by both CO2 formation and O2 consumption (Figure 1). The extension of the gas analysis to oxygen was essential in the present case because of the initially high pH value (CO2 absorption). The influence of RW heating on soil respiration is very clear: the oxygen consumption was raised, indicating stimulation of microbiological activity by heating. The production of CO2 and its binding in water led to a decrease of the pH 8448
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 24, 2007
FIGURE 1. Carbon dioxide emission and oxygen consumption for RW-heated and reference reactor during the long-term experiment: (a) initial period with moderate heating to about 35 °C and (b) further course during the phase of maintained temperature.
FIGURE 2. Development of the pH value in the soil during the long-term experiment for the two reactors. value (Figure 2) which results in the release of CO2 in the course of time (consider the different scales in Figure 1 a and b). As expected for biological oxidation processes, soil respiration is influenced by fertilization overcoming nutrient limitations. The observable respiratory quotient (ratio of CO2 formation and O2 consumption in mol/mol) increased during the remediation process, finally reaching a value of about 0.8 which is in the characteristic range for complete mineralization of hydrocarbons (e.g., 0.67 for long-chain alkanes). Although measurable respiration could also be observed in the REF reactor starting in the summer months (i.e., about two months later), the initiating effect of RW heating on respiration could be clearly demonstrated. Of course, the advantages of external heating are most pronounced in those cases where the ambient temperature and
FIGURE 3. Emission of CO as ppmv in the purge gas flow during the long-term experiment and correlation with the temperature course in the RW-heated reactor.
FIGURE 4. Biodegradation of the mineral oil contaminants as represented by the decay of the concentration of the n-alkanes in different vertical parts of the RW-heated and the reference reactors. the available carbon sources are not sufficient for a spontaneous microbiological self-heating. Interestingly, a detectable amount of CO was also formed in the RW reactor (Figure 3) indicating an at least local deficiency of oxygen. This enhanced CO concentration was only found after a temperature leap initiated by RW heating. The start of CO2 emission (indicating a sufficiently low pH) correlates with the end of CO formation. Influence of RW Heating on Biodegradation of Pollutants. To quantify the microbial degradation of the various pollutants, a multitude of soil samples were taken during the long-term study, extracted using a mixture of dichloromethane, n-hexane, and phosphoric acid, and analyzed by GC-MS (Shimadzu, QP 5000) after precleaning of the extracts by means of a silica gel column. Details of the corresponding state-of-the-art procedures have been described elsewhere (8). Notwithstanding considerable inhomogeneity of the contaminant distribution within the soil, representative analytical results could be obtained. The influence of the enhanced temperature realized by RW heating can be seen in Figure 4 for the main contaminants. Mineral oil hydrocarbons are represented by the pattern of n-alkanes. Their degradation proceeds from the very beginning of the thermally supported treatment. From the practical point of view, the fact that initial RW heating is sufficient to establish the microbial activity (also represented by soil respiration as discussed above) leading to strong self-heating is essential. The technology thus can act as a “catalyst” for the inherent remediation process based on autochthonous microorganisms. This dramatically decreases the energy costs for a long-term treatment.
FIGURE 5. Preferred elimination of nonbranched aliphatic hydrocarbons during thermal treatment exemplarily shown by the pristane and n-heptadecane contents of the soil in the RW reactor. To ascertain a microbiological origin of the decrease of the alkane concentrations unambiguously and to exclude a dominant influence of thermodesorption, the relative reduction of isoprenoides (highly branched iso-alkanes) compared to n-alkanes can be used. It is well-known that branched hydrocarbons are harder to degrade than the corresponding nonbranched compounds (12). As an example from the present study, the concentrations of n-heptadecane (bp 302 °C) and pristane (2,6,10,14-tetramethylpentadecane, bp 296 °C) were compared as a function of treatment time. As shown in Figure 5, the relative content of the isoprenoide pristane increased. This result is as expected for biodegradation. The same result was obtained for the comparison of phytane (i-C20H42) with n-octadecane. In contrast, due to similar boiling points, a similar removal rate would be expected in the case of thermodesorption. The positive effect of external heating was also observed for the degradation of phenolic compounds. However, the concentration of these pollutants was below the detection limit (
