Article pubs.acs.org/est
Colloidal Properties of Nanoparticular Biogenic Selenium Govern Environmental Fate and Bioremediation Effectiveness Benjamin Buchs,† Michael W. H. Evangelou,‡ Lenny H. E. Winkel,§,∥ and Markus Lenz*,†,⊥ †
Institute for Ecopreneurship, University of Applied Sciences and Arts Northwestern Switzerland (FHNW), School of Life Sciences, Gründenstrasse 40, 4132 Muttenz, Switzerland ‡ Institute of Terrestrial Ecosystems, ETH Zurich, CH-8092 Zurich, Switzerland § Institute of Biogeochemistry and Pollutant Dynamics, Department of Environmental Sciences, ETH Zurich, CH-8092 Zurich, Switzerland ∥ Swiss Federal Institute of Aquatic Science and Technology (Eawag), Ü berlandstrasse 133, Postfach 611, 8600 Dübendorf, Switzerland ⊥ Sub-Department of Environmental Technology, Wageningen University, 6700 EV Wageningen, The Netherlands S Supporting Information *
ABSTRACT: Microbial selenium (Se) bioremediation is based on conversion of water soluble, toxic Se oxyanions to water insoluble, elemental Se. Formed biogenic elemental Se is of nanometer size, hampering straightforward separation from the aqueous phase. This study represents the first systematic investigation on colloidal properties of pure biogenic Se suspensions, linking electrophoretic mobility (ζpotential) to column settling behavior. It was demonstrated that circumneutral pH, commonly applied in bioremediation, is not appropriate for gravitational separation due to the negative ζ-potential preventing agglomeration. Mono/di/trivalent counter cations and acidity (protons) were used to screen efficiently the intrinsic negative charge of biogenic Se suspensions at circumneutral pH. Fast settling was induced by La3+ addition in the micromolar range (86.2 ± 3.5% within 0.5 h), whereas considerably higher concentrations were needed when Ca2+ or Na+ was used. Colloidal stability was furthermore studied in different model waters. It was demonstrated that surface waters as such represent a fragile system regarding colloidal stability of biogenic Se suspensions (ζ-potential ∼ −30 mV), whereas dissolved organic matter increases colloidal stability. In marine waters, biogenic Se is colloidally destabilized and is thus expected to settle, representing a potential sink for Se during transport in the aquatic environment. suspension for prolonged times.11,12 This is due to the fact that, regardless of the microbial strain used, biogenic elemental Se is typically of nanoparticulate size (∼50−500 nm; biogenic nanoselenium, BioNSe).13,11,14 Thus a decisive factor for cost effectiveness of microbial based bioremediation is the possibility to reduce effluent concentrations below legislative guidelines by simple technical means,7 such as gravitational settling. It is worth noticing that microbes convert Se oxyanions to high purity elemental Se (e.g., 15), which should be exploited for recovery, both in the view of worldwide low dietary levels16 and increasing scarcity of Se as a critical element for high tech applications.17 This study therefore systematically investigated conditions in which BioNSe suspensions can be destabilized by balancing
1. INTRODUCTION Selenium (Se) is a trace element that is both essential and highly toxic to human and animals. In aquatic environments, Se acts as a powerful reproductive toxicant (being both gonadotoxic and teratogenic1). Selenium contamination has received considerable public attention in the past due to the dramatic effects on trophic end members of aquatic food chains.2 The predominant route of exposure for aquatic organisms is via the food web, after complex processes of biotransformation, bioconcentration, and biomagnification of mainly dissolved Se oxyanions, selenite (HSeO3−) and selenate (SeO42−) (e.g., refs 3 and 4). The ability of microorganisms to convert these dissolved Se species to insoluble elemental Se has been intensively studied in the frame of bioremediation.5−7 Several studies have confirmed that an efficient reduction of Se oxyanions to elemental Se can be achieved in continuously operated bioreactors (e.g., refs 8−10). However, one major drawback of microbially based bioremediation is the fact that biogenic Se remains in aqueous © 2013 American Chemical Society
Received: Revised: Accepted: Published: 2401
December 3, 2012 January 19, 2013 January 29, 2013 January 31, 2013 dx.doi.org/10.1021/es304940s | Environ. Sci. Technol. 2013, 47, 2401−2407
Environmental Science & Technology
Article
Information SI1). For dissolved organic matter (DOM) containing waters, humic acid standards “Elliott soil” (1S102H), “Pahokee Peat” (1S103H), and “Leonardite” (1S104H) were purchased from International Humic Substances Society (IHSS, St. Paul, MN). Humic acids stock solutions were prepared in 10−5 M NaCl under addition of a few drops of concentrated NaOH to yield final concentrations of 1 g/L. From the supernatants, dilutions were made daily to yield a concentration of 1 mg/L in 10−5 M NaCl at 7.0 ± 0.1. For electrophoretic measurements at different pH values, a background electrolyte (10−5 M NaCl) was used and pH was carefully adjusted using minimal amounts of diluted NaOH/ HCl. For the experiments using counterions, stock solutions (0.16 M) of sodium, calcium, and lanthanum counter cations were prepared in MilliQ water by weighing their respective chloride salts (all >99% purity, Sigma Aldrich, Buchs, Switzerland). Sodium, calcium and lanthanum were chosen in this study since these elements are present as truly mono(Na+), di- (Ca2+), and trivalent (La3+) cations in all Eh/pH conditions applied. The pH of the stock solutions was set to 7.0 ± 0.1 using NaOH/HCl before further dilution. Data of the electrophoretic measurements were only considered as to a concentration that could still be confirmed by ICP-MS measurements (i.e., the background equivalent concentrations of 10−9 M La3+, 10−7 M Ca2+, 10−5 M Na+, respectively). The pH of the diluted salt solutions was readjusted (if applicable) to 7.0 ± 0.1 using diluted NaOH/HCl. 2.4. Settling Experiments. Counter cation concentrations and pH values were adjusted to match a ζ-potential of maximal colloidal stability (i.e., maximal modulus ζ-potential observed), of destabilized colloidal stability (ζ-potential = −15 mV), and of minimal colloidal stability (ζ-potential = 0 mV) (Figures 1
intrinsic electrostatic forces preventing agglomeration and settling. Electrophoretic measurements were conducted investigating effect of pH, the nature, and concentration of counterions on the zeta- (ζ-) potential of BioNSe suspensions. Furthermore, to assess BioNSe settling tendency in the environment, the electrophoretic measurements were conducted in different artificial waters (seawater, fresh waters) and in the presence of humic substances. Finally, the colloidal stability fields derived from the electrophoretic measurements were confirmed by direct settling column experiments. By adjusting mono/di/trivalent countercation concentration and pH, the ζ-potential of BioNSe suspensions was set to values that define colloidally stable (−30 mV), destabilized (−15 mV), and minimally stable (0 mV) suspensions.18 BioNSe settling was then monitored using inductively coupled plasma mass spectrometry (ICP-MS) measurements.
2. MATERIALS AND METHODS 2.1. Instrumentation. Electrophoretic measurements (ζpotential) were conducted on a Zetasizer Nano ZS (Malvern Instrument Ltd., Worcestershire, UK) using a laser beam at 633 nm and a scattering angle of 173° at 25 °C. Raw data were processed by the DTS software (Malvern Instrument Ltd.), which converted the measured electrophoretic mobility (i.e., velocity of a charged species per electric field strength) into ζpotential (i.e., the electric potential difference on the hydrodynamic slip plane) by Smoluchowski approximation. A stock solution of BioNSe was prepared in MilliQ (>18.2 mΩ cm) water with a concentration of 1200 μg Se/mL. Selenium concentrations were determined on a 7500cx ICP-MS (Agilent, Basel, Switzerland), operated as previously described,13 using H2 as reaction gas, 78Se for quantification, and 80Se for verification. Immediately before the electrophoretic measurements, 30 μL of this stock BioNSe suspension was added to 2.97 mL of the corresponding solutions of defined composition. Each sample was measured three times. After each measurement the cuvette was flushed with several mL of MilliQ water and 2 mL of ethanol to clean remaining residues. Particle size distribution (PSD) of BioNSe was determined using Scanning Ion Occlusion Sensing (SIOS19) on a qNano machine (Izon, Christchurch, New Zealand). The machine was calibrated using a 220 nm polystyrol standard (Izon), suspended in 10−5 M NaCl (pH 7). For PSD, 4 times approximately 1500 BioNSe particles in 10−5 M NaCl were counted, with the cell rinsed with NaCl solution between the single measurements. Before, during, and after single measurements, the accuracy of the PSD calibration was verified using a 400 nm polystyrol standard, with the deviations of mean size of less than 6% (i.e., 376 nm). The final PSD was determined by cumulating the approximately 6000 individual particle sizes determined. 2.2. Source and Purification of BioNSe Suspensions. Bacillus selenatarsenatis (strain 18680, German Collection of Microorganisms and Cell Cultures, Braunschweig, Germany), a dissimilatory Se reducer, was grown on liquid medium (DSMZ medium 220, without agar) containing 10−2 M selenate. BioNSe nanoparticles were harvested in the late stationary phase. BioNSe was then purified from all residual biomass and growth media components by repeated density based centrifugation in sodium polytungstate solution as previously described.20 2.3. Preparation of Artificial Waters. Three artificial waters (neutral lake water, pH = 8.0; acidic lake water, pH = 4.97; seawater, pH = 7.87) were used in the study (Supporting
Figure 1. ζ-Potential of biogenic selenium suspensions versus pH (background electrolyte 10−5 M NaCl). The potential stability field (i.e., −30 mV < ζ-potential < 30 mV) of the colloidal system is marked by striped line. Error bars represent standard deviation (n = 3).
and 2; Table 1), defining the colloidal stability field. To set the ζ-potential to −15 mV and 0 mV, respectively, pH and countercation concentrations were derived off a best fit straight line through 3 measurement points closest to the desired value (Table 1). For BioNSe suspensions in Na+ solutions, conditions for minimal colloidal stability could not be determined, since the conductivity was too high for the ζ-potential determination. For the settling experiments, suspensions of defined Se (10 mg/L) and pH/countercation concentrations (Table 1) were prepared by pipetting the Se stock suspension to a measuring 2402
dx.doi.org/10.1021/es304940s | Environ. Sci. Technol. 2013, 47, 2401−2407
Environmental Science & Technology
Article
During the addition of the bivalent cation, Ca2+, BioNSe suspensions were only indicated to be colloidally stable at concentrations lower than 10−5 M. At the highest concentration of Ca2+ applied (10−1 M), the isoelectric point was met (−0.1 ± 3.8 mV, Figure 2). La3+ addition had a destabilizing effect on BioNSe suspensions at concentrations higher than 10−6 M, whereas addition of concentrations higher than 4 × 10−4 M (i.e., point of zero charge) could reverse the intrinsic negative charge (Figure 2). At even higher La3+ concentrations, the ζpotential leveled off at around +12 mV, with no significant differences observed above 10−2 M. 3.2. Colloidal Properties of BioNSe Suspensions in Artificial Natural Waters. BioNSe suspensions were between stable and destabilized state in both neutral and acidic synthetic lake waters (ζ = −30.4 ± 1.4 mV and −29.9 ± 0.8 mV, respectively; Figure 3). In seawater, BioNSe suspensions were
Figure 2. ζ-Potential of biogenic nanoselenium suspensions in the presence of monovalent (Na+; circles), divalent (Ca2+, squares) and trivalent (La3+, triangles) counter cations. The potential stability field (i.e., −30 mV < ζ-potential < 30 mV) of the colloidal system is marked by striped line. Error bars represent standard deviation (n = 3).
flask and filling up with the respective salt/pH solution. The resulting suspensions were then homogenized in an ultrasonic bath (Bandelin Sonorex RK31, Berlin, Germany) at 35 kHz and 240 W for 20 min before aliquots (triplicate) were filled in 25mL graduated glass cylinders (height 220 mm, outside diameter 17 mm). The glass cylinders were sealed by a glass joint and shaken overhead for 16 h before start of the experiment. Samples (0.5 mL) were taken 1 cm below liquid surface. BioNSe was then pelleted by centrifugation (21 500g), dissolved in aqua regia (25 μL), and diluted (1% HNO3, semiconductor grade) for ICP-MS measurement.
Figure 3. ζ-Potential of biogenic nanoselenium suspensions in neutral lake water (NLW), acidic lake water (ALW), seawater (SW), and waters containing dissolved organic matter (humic acids “Leonardite”, LEO; “Pahokee peat”, PAH; “Elliot soil”, ELI) at 1 mg/L. Error bars represent standard deviation (n = 3).
3. RESULTS 3.1. Colloidal Stability Fields of BioNSe Suspensions. In experiments using different pH, BioNSe suspensions were colloidally stable above pH = 5.4 due to their intrinsic negative charge of < −30 mV (Figure 1). Above pH 6.4, there was no significant difference in the observed ζ-potentials (ζ = −34.5 ± 1.5 mV at pH 6.4 and ζ = −35.0 ± 2 mV at pH = 10). The isoelectric point was extrapolated to be at pH 3.5 (ζ = 0 mV, Table 1). In solutions lower than pH = 3.5, the intrinsic negative charge of BioNSe particles was reversed with a maximum of ζ = +23.0 ± 0.4 mV (Figure 1). In the presence of the monovalent countercation, Na+, BioNSe suspensions were indicated to be stable at concentrations lower than 2.5 × 10−2 M, whereas even at the highest concentration applied (10−1 M), the isoelectric point was not reached (ζ = −16.6 ± 1.6 mV, Figure 2). Lower Na+ concentrations resulted in ζ-potentials leveling off at approximately −38 mV (minimum of −39.1 ± 1.2 mV, 10−4 M).
not colloidally stable, yet still negatively charged (−11 mV). All DOM rich waters studied favored colloidally stable suspensions (all lower than ζ = −53.6), whereas no considerable differences existed between the various DOM sources (Figure 3). 3.3. Settling Properties of BioNSe Suspensions in Different Colloidal Stability Regimes. SIOS measurements showed a particle size distribution with number average mean diameter of 360 nm and a volume average mean diameter of 405 nm (Figure 4). Using Stoke’s law, the volume average diameter size was used to derive the average terminal settling velocity of 2.93 cm/day, assuming discrete, nonflocculating, spherical particles, laminar flow conditions (Reynolds number pH 3.5, Figure 1) and low ionic strengths (Table 1, Figure 2). At low counterion concentrations, a plateau in ζ-potential = ∼ −38 mV was approached (Figure 2), which corresponds well with the plateau observed at low proton concentrations (i.e., higher pH values, Figure 1). It can be assumed that in these conditions, ionizable groups on the surface of the particles are fully dissociated. Conversely, a plateau at positive ζ-potential was observed at high concentrations of the trivalent countercation and high proton concentrations (Figure 1, 2). The fact that proteins are strongly associated with BioNSe particles20 suggests that these effects are associable to charges conferred by amino groups (plateau at positive values) and carboxyl groups (plateau at negative values) within the proteins. In biogenic elemental sulfur similar effects of countercation addition on ζ-potential were observed.23 For both BioNSe and biogenic sulfur, ζ-potential (or electrophoretic mobility) remained at negative values in high (10 −1 M) Na + concentrations. In both cases, the point of zero charge was approached at the same concentration of Ca2+ and at fairly comparable concentrations of La3+ (∼2.9 × 10−4 M La3+ for sulfur, 4 × 10−4 M La3+ for BioNSe).23 Much like BioNSe, biogenic sulfur spheres are composed of a sulfur core modified by organic moieties (proteins or incorporation of polar groups such as thionates24). From this it may be concluded that for both chalcogens, the particular physicochemical properties (e.g., electrophoretic mobility) are conferred by similar organic moieties. The determined colloidal stability fields (Figures 1 and 2, Table 1) related well to settling experiments. Settling times could be significantly shorted by setting the ζ-potential to values of 0 mV (i.e., minimally stable suspensions). Manipulated in this way, the majority of Se (>80%) settled within 0.5 h (Ca2+, La3+; Figure 6) and 5 h (pH experiment, Figure 5), respectively. Colloidally stable BioNSe suspensions did hardly settle within 12 h of experiment (max. 17.4%, Figures 5 and 6). When suspensions were manipulated to have a residual charge of −15 mV (i.e., destabilized suspensions), again, most Se (>80%) was settled upon termination of the experiment (with exception of La3+), yet with a different kinetic behavior (Figure 5, Figure 6B). The fact that at ζ-potential < −30 mV considerable amounts of Se were found in suspension upon termination of the experiments was not expected when considering the terminal settling velocity (∼1.5 cm/12 h; samples taken at 1 cm below settling column surface) derived from theoretical calculations based on volume average mean particle diameter of 405 nm (Figure 4). Thus, direct determination of settling properties is indispensable when designing a gravitational settler for bioremediation (see Section 4.2). 4.2. Implications for Environmental Fate of Colloidal Selenium. This study demonstrates that BioNSe suspensions will behave considerably dissimilar in different aqueous environments such as lake, marine, DOM, and cation (metal) rich waters. In both neutral and acidic surface waters, BioNSe suspensions were balancing on the boundary of their colloidal stability fields (i.e., at zeta-potentials of ∼ −30 mV, Figure 3). This demonstrates the fragility of these systems, in which a slight change in chemical composition (i.e., pH and/or cation concentration) can induce a significant change in settling and eventually transport behavior. Elemental Se displays a large pE/ pH stability field and is thermodynamically favored in these aquatic environments.45 As a consequence, elemental Se 2405
dx.doi.org/10.1021/es304940s | Environ. Sci. Technol. 2013, 47, 2401−2407
Environmental Science & Technology
Article
awaiting legislation, with a rules proposal expected late 2012 and a final regulation in the spring of 2014.42 In addition to potential discharge fines, a still little considered fact that can influence the choice toward microbial reduction with improved gravitational settling is the increasing resource scarcity of Se. The dramatic effect of (so far temporary) scarcity on the market price has been observed between the years 2002 and 2005, when Se prices skyrocketed by a factor of 12.43 Such increased market value favors systems that not only serve the purpose of reducing total Se effluent concentrations, but allow for recovery of Se. The use in several high tech applications (e.g., thin film photovoltaics44) is anticipated to further increase demand for Se from the engineering sector, which competes with the demand as an essential element for animal and human healtha global problem with a magnitude only coming to awareness recently.16,45 In summary, biological treatment with gravitational settling may well represent the best symbiosis from the perspectives of cost, achievable removal efficiency, and the possibility to recover Se as a resource.7
effective removal by small reactor footprint, gravitational settling). Currently, there are several systems based on biological reduction of Se oxyanions, either still in development or already in use.7,32−34 This study represents a systematic explanation of why these systems are often challenged by insufficient Se removal. Besides a few exceptions (e.g., 35), most bioremediation processes make use of moderately saline and circumneutral conditions. When such conditions are applied, BioNSe particles will have a tendency to remain in suspension (Figures 1 and 2) for extended times. Since BioNSe does not settle from aqueous suspensions it contributes to the legally binding, maximal Se effluent concentrations. Therefore, different costly and/or high reactor footprint post-treatment steps such as ultracentrifugation,36 dissolved air flotation/slow sand filtration,37 or polymer coagulation/flocculation8 have been used. This study shows that simple addition of salts or acid makes these approaches obsolete. The estimation of costs associated with biological Se treatment is still uncertain, relying on few publicly available data. However, one field study of relevance here used a series of anaerobic solid bed reactors with back flushing and slow sand filtration to treat up to 2 mg/L Se from a contaminated artesian flow.34 For this case, a net present value of 0.35 US $/m3 treated water was calculated. Even in the most conservative assumption, i.e. using high purity (99.99%) La3+ as the most expensive countercation tested, the price of 30 US $/kg La38 would add only 0.025 US $/m3 to the overall net present value at destabilized conditions found here (6 × 10−6 M, Table 1). The resulting overall net present value remains a factor of 6 (catalyzed cementation) and factor of 10 (ferrihydrite adsorption) lower in comparison to conventional treatment,34 not taking into account the obsolete back-flushing and slow sand filtration. Certainly, using lower grade La sources or even cheaper trivalent counter cations (in particular iron or aluminum) will result in less than a 0.025 US $/m3 increase in net present value. It should be noted that this study used one cation at a time to induce settling. Therefore some adjustment in absolute cation amounts may be necessary, since other ions (and DOM) may be already present in the wastewaters, influencing colloidal stability. More stringent legislation regarding permissible Se emissions may be the main determinant for cost-effectiveness of Se (bio)remediation in the future. For instance, conventional techniques based on ion exchange can be certainly cheap (0.06 US $/m3), yet suffer from poor Se removal (
