Environ. Sci. Technol. 2010, 44, 129–135
Microfungal Alkylation and Volatilization of Selenium Adsorbed by Goethite MIRKO PEITZSCH,† DANIEL KREMER,‡ AND MICHAEL KERSTEN* Geosciences Institute, Johannes Gutenberg-University, Mainz 55099, Germany
Received March 3, 2009. Revised manuscript received September 21, 2009. Accepted September 30, 2009.
Selenium adsorbed in the oxyanionic form by Fe-oxides like goethite is considered of benefit for long-term stabilization of 79 Se under near field conditions of radionuclide waste disposal sites. However, microbe-mediated volatilization of the uranium fission product 79Se has not yet been considered for risk assessment based on the use of the water-solid distribution coefficient KD. We have performed incubation experiments in a ternary system selenium-microbe-goethite and show that mycobiota including the common black microfungi genera Alternaria alternata are capable of volatilizing the Se even if immobilized by goethite. The microfungi were incubated in a standardized nutrient broth suspension with 10 g L-1 of the oxide target under defined conditions. Volatile organic selenium (VOSe) species formed in the head space of the culture flasks were sampled and measured directly by a cryotrapping cryofocusing gas chromatographic system coupled with ICPMS detection (CT-CF-GC-ICP-MS). Alkylated VOSe species were found at the tens to hundreds ng m-3 levels dominated by dimethyl selenide (DMSe) and dimethyl diselenide (DMDSe). The total amount of DMSe released into the 80-mL headspace volume within the 21 days of incubation was up to 1.12 ( 0.17 nmol and 0.48 ( 0.12 nmol for systems without and with goethite amendment, respectively. Alkylation rates of up to 0.1 µmol Se per day and g biomass cannot be neglected as a potential fission product mobilization pathway, unless the inherent radioactivity is proven to prevent any such microbial activity on the long-term. Otherwise it may lead to an onsite accumulation of 79Se through evapoconcentration in the enclosed underground caverns.
Introduction Radioactive waste repository performance assessments demonstrated that 79Se (half-life about 3.77 × 105 years) as a uranium fission product might contribute significantly to the total dose during the next million years (1). Selenium in nature exists in four different oxidation states: Se(-II), Se(-I), elemental selenium Se(0), Se(IV) and Se(VI), but also in a variety of organoselenium species. Under the pH and redox conditions in the natural aqueous environment, Se exists as the oxyanions selenate (SeO42-), selenite (SeO32-), or bise* Corresponding author e-mail:
[email protected]. † Present address: Division of Medical Microbiology, Dept. of Laboratory Medicine, Lund University, Lund 22363, Sweden. ‡ Present address: LGA Immissions- and Arbeitsschutz GmbH, Nu ¨ rnberg 90431, Germany. 10.1021/es9006492
2010 American Chemical Society
Published on Web 10/09/2009
lenite (HSeO3-), but the latter two species dominate under moderately oxidizing conditions such as in surface soils and groundwater aquifers (2). Se oxyanions are particle reactive and adsorb to surfaces of a variety of soil minerals, including Fe and Al oxyhydroxides, and clay minerals. Adsorption and desorption reactions on hydrous oxide surfaces are particularly important mobility-controlling reactions because they are widespread in the hydrogeologic environment as coatings on other solids. In anoxic environments, selenium may become reduced abiotically to oxidation states 0, -I, and -II by Fe(II)-containing solids (3). Reduced selenium chemistry parallels sulfur chemistry, and may form highly insoluble metal selenide precipitates such as FeSe, Fe7Se8, and FeSe2. The local structures of these solid redox reaction products suggest formation of nanoscale clusters, which may be prone to colloid-facilitated transport (3). Modeling radionuclide mobility for long-term risk assessment is usually based on the use of the distribution coefficient KD. This coefficient is a “macroscopic” parameter, used as a gross estimate of the distribution of an element between the solid and aqueous phases, irrespectively whether due to adsorption or other processes involved. A KD value is not species-specific and depends on many environmental factors or experimental conditions (pH, equilibration time, soil-solution composition, etc.). Surface complexation models which may account for variation of species and physicochemical conditions are not yet commonly applied in longterm repository performance assessment. What matters more, however, is that microbiological activity does not yet play any significant role in these considerations. The problem is that any biochemical transformations involved require most sophisticated analytical identification and quantification of selenium species possible. This is because in the living organisms, Se is not coordinated but rather forms covalent C-Se bonds which make speciation a challenging task (4). Formation of these C-Se bonds may also challenge the current KD concept if to consider 79Se fate and mobilization pathways. Extensive literature on selenite interactions with microorganisms in soils and waste deposits exists (5-9). It reveals that selenite is always an intermediary product in the microbially mediated selenate reduction, because it can undergo a lot of transformations into more reduced selenium forms. Fe´vrier et al. (10) have shown that soil microbial activity induces an increase of the KD value of selenium from 16 L kg-1 in sterile conditions to 130 L kg-1 when the soil was amended with glucose and nitrate during incubation. They have not studied the biochemical transformation mechanisms in detail but argued that the loss of aqueous selenite in the batch supernatants could be explained by the formation of (i) organoselenium species assimilated into the microorganisms, or (ii) particulate Se(0) particles due to biochemical reduction of the selenite. The Se(0) immobilization pathway has been evidenced simply by the typical intensive red coloration induced by precipitation of monoclinic Se(0) in samples with low opacity (11), or by X-ray absorption spectroscopy in case of occurrence of the gray trigonal Se(0) polytype (12). Further reduction to selenide may occur in anaerobic incubations with excess organic inocula amendments (12-16). Both microbial metabolic processes (e.g., dissimilatory reduction) and microbially mediated physicochemical mechanisms (insoluble metal selenide precipitation) may therefore contribute to the increase of KD beyond that involving adsorption processes only, and hence more effective immobilization of selenium. However, all these VOL. 44, NO. 1, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
129
studies fall short because they have considered dissolved and solid Se species only. In fact, microbial processes are not that simple. Se(0) and selenide precipitation is not the only mechanism responsible for the microbially mediated Se oxyanion removal. Microbial reduction of Se is known also to induce alkylation which results in volatile organic selenium (VOSe) species such as dimethyl selenide (DMSe) and dimethyl diselenide (DMDSe) (9, 17, 18). VOSe species are by 3 orders of magnitude less toxic than the Se oxyanions (19). Such a biovolatilization process serves primarily the detoxification of the microenvironment of the microorganisms and should therefore be only of local importance. At best it may be exploited as a useful strategy for bioremediation and natural attenuation of Se-contaminated soils (8, 18, 20). Acquiring KD values from dissolved Se without any VOSe speciation analysis is therefore irrelevant as long as ecotoxicity is of concern. This biovolatilization process may, however, lead to onsite accumulation of VO79Se and build-up of a radiotoxicity threat if to occur via evapoconcentration in a spent nuclear fuel underground repository cavern. Our study aimed at evaluating VOSe formation in microfungal inoculates amended with typical selenite sorbents such as goethite. There are as yet no data in the scientific literature that demonstrate fungi to be natural inhabitants of nuclear waste repositories. However, there is no conceptual argument against their dwelling underground. Fungi are capable of surviving anaerobic conditions, which makes them suitable for life in the narrow aquifers of hard rock. Recent ¨ spo¨ HRL (Sweden) investigations of groundwater from the A strongly suggest that fungi are a natural part of the subterranean biosphere in igneous rock aquifers in contact with such repositories (21). Classical laboratory batch experiments were performed to derive the adsorption as a function of pH of selenite between the aqueous phase and the solids. Then the impact of the microbial activity on the immobilization of stable Se isotopes by goethite was studied in incubation experiments. For analysis of VOSe species in the head space, a hyphenated technique based on online coupling of cryotrapping and cryofocusing for accumulation, chromatographic species separation, and ICP-MS detection was developed. We will show a dramatic impact by introduction of common microfungal biofilm-forming cultures on the selenium distribution and transformations between the aqueous and solid phases, and also the Se alkylation and hence volatilization rates in equilibrium with the solids.
Materials and Methods Microbial Incubation Experiments. A pure microfungal culture of Alternaria alternata was purchased from the “Deutsche Sammlung von Mikroorganism und Zellkulturen GmbH, Braunschweig” (DSMZ). The DSMZ catalogue no. 1102 is an Alternaria alternata strain of the typical soilinhabiting nonpathogenic fungi of the melanin containing Dematiaceae genera, which do appear as common filamentous colonizers of biofilms. The strain was cultivated in modified malt extract medium in butyl rubber-stoppered glass bottles with shaking (120 rpm) in the dark at 30 °C and a pH of 6.5 as recommended by Thompson-Eagle et al. (22). This medium was composed of 1.0 g L-1 malt extract (Merck), 1.0 g L-1 glucose, and 0.1 g L-1 peptone from soy beans (Merck). All subsequent culture experiments were performed in the dark using aseptic techniques of a BioSafety level 1 laboratory. Batch experiments were conducted in sterile 100mL glass incubation flasks containing 20 mL of nutrient broth and 200 mg of a commercial goethite product (Fluka). First the goethite was weighted in the flasks and sterilized by autoclaving at 120 °C for 15 min. Then the sterile nutrient broth spiked with the target compound Na2SeO3 (Fluka) to a final Se(IV) concentration of 100 µg L-1 and an initial pH 130
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 1, 2010
) 6.0 was added. The flasks were capped with sterile rubber stoppers and equilibrated on a rotary shaker at 30 ( 1 °C for 48 h to adsorb the dissolved Se(IV) target compound onto the mineral. After the equilibration a set of flasks was used for estimation of Se adsorption as a function of pH by titration with HNO3 or NaOH, respectively. Another subset of the flasks was ultimately spiked with 1% of a 72-h incubated isolate culture of A. alternata, and capped again. Control experiments were performed with the same concentrations of total Se(IV) target spike and inoculates, but without the sorbent mineral. All incubations were performed on a rotary shaker (120 rpm) at 30 ( 1 °C. In total 24 flasks (+24 controls) were prepared from each incubation experiment to allow for triplicate sampling at 8 time intervals. The VOSe production in the headspace was followed over an incubation time of 3 weeks (21 days). After incubating all flasks for 2 days, 3 culture flasks of each incubation series were sampled at times for selenium species in the headspace, together with subsamples of the nutrient broth for analysis of the amount of biomass, pH, Eh, the glucose content, and total dissolved selenium. The residual glucose content in the nutrient broth throughout all incubations was photometrically analyzed with the hexokinase-glucose-6-phosphate dehydrogenase assay according the manufacturers description (Sigma-Aldrich). The amount of biomass in terms of mg formed in the 20-mL samples without the sorbent mineral was determined gravimetrically after filtration and drying at 105 °C. For the mineral-containing samples, differential thermo-analysis (DTA, simultaneous thermal analyzer STA 429, Netzsch) was used. The dried samples were heated in a temperature range from 40 to 1200 °C. Quantification (mg per 20 mL) was performed on the basis of the additional exothermic peak compared to goethite samples without biomass. Characterization of Sorbent. A subsample of the commercial goethite (R-FeOOH) was used for BET surface area measurement, which gave a relatively low value of 10 m2 g-1 but after extended sterilization at 120 °C. Surface characterization in terms of acidity and charging behavior was determined by potentiometric titration as described in detail elsewhere (23). A pHPZC ) 8.8 was determined in presence of ambient atmosphere (i.e., of CO2). This value is therefore somewhat lower than theoretically predicted for pure goethite under inert gas probably due to carbonate adsorption, but still in the range of experimental values reported in the literature (24). Selenite Adsorption Experiments. For characterization of the sorption behavior of Se(IV) on goethite, subsamples were prepared as described above but without the addition of microbe inoculate over a pH range from 3 to 8. After equilibration for 48 h, the final pH was measured in the suspensions using a combination electrode (SenTix81, WTW) calibrated with three commercial buffer solutions (CertiPur, Merck). The subsamples were then centrifuged at 4000 rpm for 20 min, and the supernatant was filtered through a 0.2µm membrane filter and acidified by HCl (Merck, p.a.). Dissolved Se(IV) concentrations were analyzed using a hydride generation technique (HG-AAS Varian SpectrAAS/ VGA 76) with a determination limit of 0.5 µg L-1. A solution of 0.6 g of NaBH4 (Riedel-de-Hae¨n, p.a.) and 0.5 g of NaOH (Merck, p.a.) dissolved in 100 mL of distilled water (ultrapure grade, conductivity 0.055 µS cm-1) and 10 M HCl (Merck, p.a.) were mixed with the sample to form Se hydrides. The amount of adsorbed selenite (% adsorbed) was calculated as the difference between Se added and final dissolved (i.e.,
