Environ. Sci. Technol. 2003, 37, 5671-5677
Surface Chemical Heterogeneity of Bacteriogenic Iron Oxides from a Subterranean Environment R A U L E . M A R T I N E Z , † D . S C O T T S M I T H , #,† KARSTEN PEDERSEN,‡ AND F . G R A N T F E R R I S * ,† Microbial Geochemistry Laboratory, Department of Geology, University of Toronto, Toronto, Ontario, M5S 3B1, Canada, and Department of Cell and Molecular Biology, Microbiology Section, University of Go¨tebo¨rg, Go¨tebo¨rg, Box 462, S-405 30, Sweden
This study quantifies the surface chemical heterogeneity of bacteriogenic iron oxides (BIOS) and its end-members (2-line ferrihydrite and intermixed intact and fragmented bacteria). On a dry weight basis, BIOS consisted of 64.5 ( 1.8% ferrihydrite and 34.5 ( 1.8% organic matter. Enrichment of Al, Cu, Cr, Mn, Sr, and Zn was shown in the solid versus the aqueous phase (1.9 < log Kd < 4.2). Within the solid-phase Al (69.5%), Cu (78.7%), and Zn (77.9%) were associated with the bacteria, whereas Cr (59.8%), Mn (99.8%), and Sr (79.4%) preferred ferrihydrite. Acidbase titration data from the BIOS and bacteria were fitted using FOCUS pKa spectroscopy. The bacteria spectrum with pKa’s of 4.18 ( 0.37, 4.80 ( 0.54, 6.98 ( 0.45, and 9.75 ( 0.68 was similar to discrete and continuous spectra for intact and fragmented bacteria. The BIOS spectrum recorded pKa’s of 4.27 ( 0.51, 6.61 ( 0.51, 7.89 ( 1.10, and 9.65 ( 0.66 and was deconvoluted to remove overlapping binding site contributions from the bacteria. The resulting residual iron oxide spectrum coincided with discrete MUSIC spectra for goethite and lepidocrocite with pKa values of 4.10 ( 0.43, 6.53 ( 0.45, 7.81 ( 0.76, and 9.51 ( 0.68. Surface site density analysis showed that acidic sites (pKa < 6) were contributed by the bacteria (37%), whereas neutral sites (6 < pKa < 8) were characteristic of the iron oxide fraction (35%). Basic sites (8 < pKa) were higher in the bacteria (57%), than in the BIOS (44%) or iron oxide fractions (47%). This analysis suggested a high degree of bacterial group masking and a similarity between the BIOS and goethite surface reactivity. An understanding of the BIOS surface chemical heterogeneity and inherent proton and metal binding capacity was obtained through the use of FOCUS apparent pKa spectroscopy.
Introduction Bacteriogenic iron oxides (BIOS), abundant in surface water and groundwater systems, are potent sorbents of dissolved metal cations (1-4). These properties emphasize their * Corresponding author phone: (416)978-0526; fax: (416)978-3938; e-mail:
[email protected]. † University of Toronto. ‡ University of Go ¨ tebo¨rg. # Present address: Department of Chemistry, Wilfred Laurier University, Waterloo, Ontario, Canada, N2L 3C5. 10.1021/es0342603 CCC: $25.00 Published on Web 10/31/2003
2003 American Chemical Society
importance in regulating the dispersion of metals in pristine and contaminated environments. BIOS are complex systems which are composed of poorly ordered iron oxyhydroxide (i.e., 2-line ferrihydrite) intermixed with intact and partly degraded bacterial cells, as shown previously by HR-TEM and SEM analysis (2, 3). They are formed in anoxic/oxic water mixing zones, at low pO2, high Fe2+ concentration, and circumneutral pH by the metabolic activity of Gallionella ferruginea and other iron oxidizing microorganisms (1-3). The BIOS metal binding potential is a function of the type, concentration, and availability of iron oxyhydroxide and bacterial cell surface functional groups and their respective acidity constants (Ka) defined previously (5). Chemical equilibrium models of acid-base titration data have suitably described the surface reactivity of bacterial cells in terms of a sum of monoprotic acids (6). Calculated pKa values for bacterial surfaces have been attributed to carboxyl, phosphoryl, and amine functional groups (7-10). Because the surface of bacterial cells is a three-dimensional cross-linked structure (11, 12), a distribution of pKa values for each type of reactive site may be observed as a result of steric and electronic interactions (8). The complex molecular architecture of bacterial surfaces underscores the utility of applying discrete or continuous pKa spectrum models to interpret surface chemical heterogeneity (7-10). The crystal structure of 2-line ferrihydrite has not been well resolved owing mainly to the absence of well-defined maxima in XRD spectra (13). EXAFS studies indicate that the Fe-(O,OH) bond lengths in ferrihydrite are characteristic of Fe3+ octahedral coordination as in goethite (13). Other works suggest that 25% of the Fe3+ in ferrihydrite is in tetrahedral coordination at the surface, while the bulk iron has an arrangement similar to that of octahedral oxyhydroxides (13). Surface oxygen atom coordinations on iron oxyhydroxides have been assigned IUPAC designations, such as oxo (hydroxo), µ-oxo (hydroxo), and µ3-oxo (hydroxo) which refer to an oxide (hydroxide) ion bound to one FeO(H), two Fe2O(H), and three Fe3O(H), Fe atoms, as well as aquo and µ-aquo denominations for (Fe-OH2) and (Fe2OH2), respectively (1417). The CD-MUSIC and quantum mechanical models propose that each of these configurations has a characteristic acidity constant (Ka), which depends on the mineral surface being considered (14-17). In the BIOS mixture, all or part of the reactive functional groups on the bacteria may be either occupied or sorbed by Fe3+, and/or masked by iron oxyhydroxides (18, 19). The extent to which the reactive sites on each BIOS fraction contribute to the mixture’s surface reactivity could be assessed by determining the percent concentrations of the total normalized group density corresponding to each binding site type, as a function of acidity. Comparison of the BIOS end-member site contributions with those of the mixed sorbent would be indicative of the nature of the BIOS surface reactivity. In this study, the above phenomena were investigated by unraveling the surface chemical heterogeneity of the BIOS and its end-member phases. FOCUS pKa spectroscopy (6) enabled deconvolution of the mixed BIOS spectra to remove overlapping binding site density contributions from the bacteria and recover the reactive surface heterogeneity of the ferrihydrite fraction. This procedure was able to provide an estimate of the degree of specific chemical interactions among the BIOS phases and quantify the mixture’s surface capacity for proton and metal cation sorption. VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5671
Materials and Methods BIOS Sample Collection. BIOS precipitates develop in the A¨ spo¨ HRL at anoxic/oxic mixing zones where groundwater enters the tunnel through hydraulically conductive fractures at circumneutral pH (pH 7.0-8.0) (2). Iron oxide precipitation occurs due to the metabolic oxidation of Fe2+ to Fe3+ by the stalk forming bacteria Gallionella ferruguinea (1-3). BIOS samples were recovered using sterile plastic spatulas and 50 mL sterile polypropylene centrifuge tubes which were filled to approximately 30 mL with BIOS precipitate. Chemical and Mineralogical Analysis of BIOS Samples. The bacterial cell fraction of the BIOS was isolated using hydroxylamine treatment to reductively dissolve the intermixed iron oxyhydroxide phase. This involved initially washing 5 mL of BIOS three times with ultrapure water (UPW) (18 MΩ) by centrifugation at 10 000 × g for 10 min. The washed BIOS pellet was then resuspended in 20 mL of UPW. A 500 µL aliquot of the 20 mL suspension was subsequently reacted for 6 h at 65 °C in a final 8.0 mL volume of 0.04 M NH2OH‚HCl in 25% (v/v) acetic acid and filtered through 0.2 µm filters to recover the insoluble bacterial cell fraction (3). The filtrate was retained to determine the trace metal content of the iron oxyhydroxide phase using an ICP-OES system (Perkin-Elmer Optima 3000). To determine the trace metal content of the isolated bacterial cell fraction, samples were digested in 8 mL of concentrated trace metal grade HNO3 and 30% (v/v) H2O2 before analysis. Dissolved trace metals in the water from which BIOS samples were collected were measured after filtration with 0.2 µm filters and acidification to 2.0% (v/v) with HNO3. The mineralogy of the BIOS was assessed after washing by centrifugation with UPW, drying the pellet to a constant weight at 40 °C, and grinding to a fine powder for X-ray diffraction. Stock Solutions for Potentiometric Analysis. Glassware and plasticware were soaked for 24 h in a 20% (v/v) trace metal grade nitric acid and rinsed with UPW. Stock solutions of 0.1 M NaOH and 0.1 M KNO3 were prepared in an anaerobic chamber under a positive pressure of nitrogen gas using N2purged degassed sterilized UPW. The 0.1 M NaOH solution was made from a 10 N volumetric standard (ACP Scientific, Montreal, Quebec) according to standard analytical methods (5) and was standardized for titration as described previously (8, 10). Salts and solutions were stored under N2 when not in use. Preparation of BIOS Samples for Acid/Base Titration. A 30 mL BIOS sample was transferred to a 50 mL conical bottom Corning sterile centrifuge tube and centrifuged at 20 °C for 8 min at 6200 × g. The pellet was resuspended in 25 mL of sterile degassed UPW. The centrifugation process was repeated four times to wash the sample, three times with UPW and once with a sterile and degassed solution of 0.1 M KNO3. Finally, the remaining pellet was resuspended in 20 mL of 0.1 M KNO3. A 1000 µL aliquot was then extracted from the washed BIOS and resuspended in 10 mL of 0.1 M KNO3 for acid-base titration. The pH of each aliquot was lowered to 3.5 using 200 µL of a 0.2 M sterile HNO3 stock. The process was repeated for BIOS bacterial cell fractions by treating 30 mL whole sample pellets with 0.04 M NH2OH.HCl in 25% (v/v) acetic acid, at 25 °C, to reductively dissolve the iron oxyhydroxide phase, before washing by centrifugation. Titration of BIOS and Automatic Titrator Settings. The sample vessel, containing BIOS or bacterial cell fraction suspensions, was placed inside a Metrohm glass titration vessel (6.1415.310) and covered with a Metrohm lid (6.1414.010) through which a pH electrode and a custom designed N2 gas line interface were fitted. The pH electrode was connected to a Metrohm GP 736 Titrino autotitrator system interfaced with Titrino Workcell 4.3 software to a personal computer. The pH electrode was three point 5672
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003
calibrated with fresh buffer (pH 4, 7, and 10) before each experiment, and the slope in all cases was >95% of the Nernstian value. Before the start of the titration, the system was allowed to reach equilibrium by maintaining a constant pH reading for a period of at least 180 min. The autotitrator system settings were adjusted to maintain a fixed interval size of 0.15 pH units. The drift criteria for the pH electrode was set such that titrant additions occurred only if the change in potential was less than 0.1 mV over 1 min (0.1 mV/min). The titrations were conducted in the pH range of 3.5-11. The mass of the BIOS or bacterial cell material present in solution was determined by vacuum filtration of the suspension through a 0.2 µm filter (Gelman Laboratory, Supor 200 Membrane Filter) dried to a constant weight at 65 °C. An average solid concentration of 0.90 ( 0.01 mg/mL was recovered from mixed BIOS and bacteria fraction titration experiments. Modeling of Acid-Base Titration Data. Acid-base titration data from the BIOS mixture and bacterial cell fractions were analyzed using FOCUS pKa spectroscopy (6, 8). Proton dissociation mechanisms for a single protonated site were modeled using the reaction below Ka
HL 98 H+ + L-
(1)
where HL refers to a protonated binding site, H+ ∼ H3O+ is the hydronium ion species, whose activity, {H+}, was measured with a pH electrode. {H+} was converted to concentration through the relationship {H+} ) γH‚[H+], where γH is the tabulated value for the proton activity coefficient in ref 5 for H+ at the relevant ionic strength and L- describes a deprotonated reactive site with a net negative charge. Finally, Ka is the apparent proton dissociation constant for HL in eq 1. Ka is conditional on ionic strength and implicitly embodies electrostatic parameters. It is described by the relationship
Ka )
[L-]‚[H+] [HL]
(2)
where pKa ) -log Ka is a measure of functional group acid strength as described previously (5). Acid-base titration data must be transformed to remove the effects of the dissociation of water, which dominate the shape of the raw titration curve (6, 8). Once this is done, the mechanisms of deprotonation and protonation reactions on the substrate of interest can be monitored directly (6, 8). The transformation generates the experimental net surface charge excess, which can then be fitted as a function of measured ([H+]) and adjustable (Ka and [LT]) speciation parameters, where [LT] refers to the total binding site density for a particular site (6, 8). Continuous apparent pKa spectra and binding site densities were then calculated for the BIOS and bacterial cell fractions using FOCUS. Details of the fitting method have been described previously (6, 8).
Results and Discussion The mineralogical analysis of the BIOS mixture was done by powder X-ray diffraction, using Cu KR radiation and a step size of 0.02 at a rate of 0.85 s step-1 from a 2θ of 4° to 90° (20). This analysis revealed two broad peaks with maxima at approximate 2θ values of 64.7° and 34.9°. This pattern is consistent with the presence of 2-line ferrihydrite (20, 21), a poorly ordered fine-grained iron oxyhydroxide that precipitates in response to microbial iron oxidation. Inspection of BIOS by light microscopy (data not shown) revealed loosely aggregated masses of helical ferrihydrite-encrusted Gallionella ferruginea stalks, as described previously (2, 23, 24). This chemoautotrophic species of bacteria acquires the
FIGURE 1. Replicate acid-base titration experiments for the whole BIOS sample (subplot a) and the organic residue (subplot b) at 0.1 M KNO3. The charge excess b is plotted against pH. Values of apparent pHpzc of 9.6 ( 0.1 and 4.1 ( 0.1 are observed for subplots (a) and (b), respectively.
TABLE 1. Dissolved and Solid Phase Metal Concentrations in the BIOS, Oxide (Hydroxylamine Soluble), and Bacterial Cell (Hydroxylamine Insoluble) Fractions concentration (ppm)a metal Al Ca Cr Cu K Mg Mn Na Sr Zn
aqueous phaseb
BIOS
oxide (%)c
0.975 6708 2048 (30.5) 869 29170 26795 (92.0) 0.003 42.6 25.4 (59.8) 0.032 67.3 14.3 (21.3) 20.2 4797 1058 (22.1) 113 5573 820 (14.7) 1.42 15745 15722 (99.8) 1600 2796 1087 (38.9) 15 1153 915 (79.4) 0.047 141 31.2 (22.1)
bacterial (%)c
log Kd BIOSd
4659 (69.5) 2374 (8.0) 17.1 (40.2) 53.0 (78.7) 3738 (77.9) 4752 (85.3) 22.5 (0.2) 1708 (61.1) 237 (20.6) 109 (77.9)
3.8 1.5 4.2 3.3 2.4 1.7 4.0 0.2 1.9 3.5
a Relative standard deviations are e10% of reported mean values. Aqueous phase refers to the filtrate after removal of solid-phase BIOS by filtration. c Percentage of corresponding metal concentration in the BIOS composite. d Kd is solid-phase metal concentration in the BIOS divided by the corresponding dissolved concentration in the aqueous phase. b
energy needed for growth from metabolic oxidation of Fe2+ (1), which is then followed by chemical hydrolysis and precipitation of evolved Fe3+ to form ferrihydrite (2, 3). The results from the BIOS trace metal analyses are shown in Table 1. From the hydroxylamine extractions, the BIOS consisted on a dry weight basis of 64.5 ( 1.8% ferrihydrite and 34.5 ( 1.8% bacterial organic matter. Solid-phase partition coefficients (Kd values, presented in Table 1 as log Kd) show up to 3 or 4 orders of magnitude enrichment, notably for Mn, Cr, Cu, Zn, and Al in the BIOS. Of these metals, Cr (60%) and Mn (99%) were associated mainly with the
ferrihydrite oxide fraction, whereas Cu (79%), Zn (78%), and Al (70%) displayed a stronger association for the bacterial fraction. Cu and Zn in particular have been shown to exhibit a high affinity for organic matter in solid particulates (3). These observations are consistent with reported BIOS Kd values, which indicate that the incorporation of bacterial organic matter into inorganic particulate solids contributes to enhance solid-phase metal partitioning (2, 3, 18). Figure 1 shows the FOCUS optimization results for the transformed experimental acid-base titration data. Subplots (a) and (b) correspond to the plots of charge excess as a function of pH for the BIOS mixture and the bacterial cell fraction, respectively. Apparent points of zero charge (pHpzc) were estimated at the pH where the surface charge excess is equivalent to zero. These correspond to values of 9.6 ( 0.1 and 4.1 ( 0.1 for the BIOS mixture and bacterial fraction, respectively. These estimates are not fully equivalent to actual points of zero charge, as these are conventionally determined from acid-base titration experiments at different ionic strengths (6, 8). The apparent pHpzc value of 9.6 ( 0.1 for the BIOS mixture differs from an average pHpzc of ∼8.0 for synthetic, freshly precipitated 2-line ferrihydrite reported previously (25, 26). The higher pHpzc for the BIOS can be explained by the presence of metal cations, such as Ca2+, Cr3+, Mn2+, and Sr2+ partitioned within the BIOS iron oxide fraction, as indicated in Table 1. As described by Stumm and Morgan (27), the binding of specifically adsorbed metal cations to a hydrous oxide fraction may cause the fixed surface charge to increase or become less negative, shifting the pHpzc to higher pH values, as observed in Figure 1a for the BIOS mixture (27). The apparent pHpzc of 9.6 ( 0.1 parallels values of 9.4 and 8.6 to 9.3 reported for goethite and hematite, respectively (15, 17, 28, 29). This pHpzc value and evidence of Fe octahedral VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5673
FIGURE 2. Continuous pKa affinity spectra generated by the FOCUS fit represented as solid and dotted lines. Subplots (a) and (b) show continuous pKa spectra corresponding to the same subplots in Figure 1 as solid lines. In subplot (c) the solid line represents the residual deconvoluted spectrum for the iron oxide phase. The dotted lines in all cases represent the Gaussian distribution fits used to calculate the pKa values recorded in Table 2. The discrete spectra are that of goethite (white bars) and lepidocrocite (black bars) as calculated by the MUSIC model. These are normalized to the highest concentration peak so that the comparison is for spectral shape and not absolute site density values. coordination in 2-line ferrihydrite (13) infer a correlation between the surface charging behavior of the iron oxyhydroxides and that of the BIOS 2-line ferrihydrite (partitioned as shown in Table 1). The pHpzc of 9.6 ( 0.1 further suggests that the BIOS mixture surface charge is dominated by contributions from reactive groups associated with ferrihydrite rather than the bacterial cell phase. This implies that acidic (i.e., low pKa) groups of the bacterial cell fraction interact chemically with ferrihydrite and thus are not subject to titration in mixed BIOS samples (2-4, 18). The low apparent pHpzc value of 4.1 ( 0.1 for the bacterial fraction, arising from contributing exposed acidic groups on the cell surface, is directly comparable to the value of 4.5 ( 0.1 reported by Plette et al. for Gram-positive bacterial cell wall fragments (22). This implies a similar charging behavior between bacteria grown in pure culture and the bacterial cell fraction of the BIOS, emphasizing their biogenic origin. The apparent pHpzc value of 4.1 ( 0.1 is also in close agreement with the intrinsic pHpzc of 4.5 reported for dissolved organic matter by Croue´ et al. (30). Figure 2 shows the pKa spectra obtained from FOCUS optimization. Subplots (a) and (b) correspond to the mixed BIOS and bacterial cell fractions, respectively. The spectrum in subplot (c) represents the residual deconvolution of the BIOS (a) after removal of overlapping peak pKa contributions from the bacteria (b). The deconvoluted spectrum (c) was determined using the weighted mass fraction of the hydroxylamine insoluble phase (0.345) and the FOCUS optimized site density vectors (dotted lines in Figure 2) for the BIOS (LBIOS) and bacteria (Lbacteria) fractions. The deconvoluted residual (iron oxide fraction) site density, Lres, was then 5674
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003
TABLE 2. Summary of pKa Values for Each Sitea sample
site 1
site 2
site 3
site 4
BIOS 4.27 ( 0.51 6.61 ( 0.55 7.89 ( 0.88 9.65 ( 0.66 bacteria 4.18 ( 0.37 4.80 ( 0.54 6.98 ( 0.45 9.75 ( 0.68 residual 4.10 ( 0.43 6.53 ( 0.45 7.81 ( 0.76 9.51 ( 0.68 a For each site the column entries correspond to the pK value plus a or minus the width ((2σ) of the mean pKa value determined from a Gaussian model for the peak shapes.
calculated as Lres ) LBIOS - 0.345xLbacteria. This plot (c) thus portrays the concentration and pKa distribution of proton binding sites present on the 2-line ferrihydrite portion of the BIOS sample (i.e., positive values) as well as sites lost in the composite (i.e., negative values) owing to surface chemical interactions between the bacteria and ferrihydrite. The apparent pKa results for the BIOS are summarized in Table 2. The value of 4.27 ( 0.51 is in good agreement with acidity constants of 3.80 and 4.00, and 4.30 for lepidocrocite and the µ3-hydroxo coordination on the (110) goethite surface, respectively, as calculated by the MUSIC model (15). Values of 7.89 ( 0.88 and 9.65 ( 0.66 are consistent with those reported for the aquo (110) and µ-hydroxo (021) coordination with pKa values of 8.05 and 9.79, respectively. The experimental pKa of 7.89 ( 0.88 is also comparable to other acidity constants calculated for µ3-hydroxo in the (110) goethite surface ranging from 8.61 to 8.84 (14, 15, 31, 32). These observations would be consistent with the proposed octahedral arrangement of the Fe atoms in ferrihydrite, as mentioned in Jambor et al. 1998 (13). The value of 6.61 ( 0.55 in Table 1 for the BIOS may be attributable to the
FIGURE 3. Pie charts (a), (b), and (c) correspond to the percent binding site density contribution from the BIOS, bacteria and ferrihydrite’s acidic (pKa < 6), moderately acidic (6 < pKa < 8), and basic (pKa > 8) functional groups, as per Table 2. µ-hydroxo conformation on the (110) surface of goethite, as suggested by Rustad et al. (14). Figure 2b shows the FOCUS result for the bacterial cell fraction. Acidity constants of 4.18 ( 0.37, 4.80 ( 0.54, 6.98 ( 0.45, and 9.75 ( 0.68 (Table 2) are consistent with ranges of 2-6, 5.6-7.2, and 9-11 for carboxyl, phosphate, and amine functional groups reported for intact and fragmented bacterial cells (7-10, 22). These values differ from those of the mixed BIOS reported in Figure 2a and Table 2. This observation would implicate masking of cell surface reactive groups, which were not detected through acid-base titration methods due to specific interactions of the iron oxide (2-line ferrihydrite) phase with the bacterial cell surface (18, 19). The spectral pattern in Figure 2c was determined from the deconvolution of the BIOS spectrum in Figure 2a as explained earlier. Ideally, if the mixed BIOS could be represented as a sum of the independent site densities from ferrihydrite and bacteria, then the deconvoluted spectrum should be that of the iron oxide phase. This spectrum is in good agreement with the discrete MUSIC pKa spectra for goethite and lepidocrocite (33). The pKa value of 6.53 ( 0.45 (Table 2) for the residual spectrum coincides with those at 6.2 and 6.4 in the lepidocrocite spectrum (Figure 2c). This emphasizes the nonorganic nature of the pKa at 6.61 ( 0.55 in the BIOS spectrum (Figure 2a). pKa values of 7.89 ( 0.88 and 7.81 ( 0.76, for the BIOS and ferrihydrite fractions respectively, resemble values reported for the goethite discrete spectra (Figure 2c) and were not observed for the bacterial fraction. This suggests specific bacterial functional group masking by iron oxyhydroxides. This is further emphasized by the agreement of BIOS apparent pKa’s with intrinsic constant values of 7.4 and 8.6 for HFO as explained in ref 25 and similar values of 7.29 and 8.93 reported by Dzombak and Morel (26). The similarities observed between the pKa values calculated using FOCUS and those from molecular static calculations (14, 15) emphasize the use of continuous pKa spectra
TABLE 3. Summary of Total and Individual Binding Site Density (LT) Valuesa sample BIOS bacteria residual
- µmol/mg BIOS - µmol/mg bacteria - µmol/mg BIOS - µmol/mg HFO - µmol/mg BIOS
L1
L2
L3
L4
total LT
0.95 0.25 0.09 0.84 0.55
0.15 1.10 0.38 0.31 0.20
1.25 0.20 0.07 1.31 0.85
1.85 2.10 0.72 2.15 1.40
4.20 3.65 1.26 4.62 3.00
a Site densities in bold are in units of µmol/mg of BIOS fraction and were calculated using the area under the Gaussian curves. The L1-L4 labels correspond to sites 1-4 in Table 1. The total LT column is the sum of the individual L’s for sites 1-4. Total and individual site densities (italics) were normalized to µmol/mg of BIOS using the mass fractions of 0.345 and 0.655 for bacteria and iron oxide, respectively.
as a means to determine acidity constants of natural iron oxide surfaces. However, because the structure of 2-line ferrihydrite is not well resolved (13), and due to the uncertainty that persists in iron oxide acid dissociation constant calculations (31), only a qualitative comparison of FOCUS pKa values with those of molecular modeling approaches should be feasible. Nonetheless, the apparent FOCUS spectra do give underlying mechanistic information about the surface reactive heterogeneity of BIOS (34). The inclusion of additional adjustable electrostatic parameters in FOCUS would translate into increased flexibility in the data optimization routine. This, in turn, may provide an unrealistic view of the intrinsic BIOS reactivity due to the sorbent’s chemical complexity (34, 35). Table 3 summarizes individual (L1-L4) and total site densities (LT) for the corresponding pKa’s in Table 2. Boldface values represent site concentrations in units of µmol/mg of BIOS, bacteria, and ferrihydrite, respectively. LTs of 4.20, 3.65, and 4.62 are consistent with those reported separately for bacteria and iron oxides (6, 8, 34). Total and individual site densities were normalized to µmol/mg of BIOS using mass fractions of 0.345 and 0.655 for bacteria and ferrihydrite, VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5675
respectively, as shown in italics. For reasons of clarity, the reader is referred to Table 3 for a summary of L and LT values. To probe the reactive surface properties of the BIOS mixture, acidity constants (pKa’s - Table 2) were grouped as acidic (pKa < 6), neutral (6 < pKa < 8), and basic (pKa > 8) following a general pH classification (5). Normalized individual site densities (L1-L4 - Table 3), from each BIOS fraction, were assigned to one the above classifications as per Table 2. Because the BIOS surface is composed of masked bacterial functional groups and exposed iron oxyhydroxide surfaces, which may contain different oxygen coordinations, a comparison of the percent contributions of each site type (acidic, neutral, or basic) in each BIOS phase would provide a simple way of inferring the surface characteristics of the mixed BIOS from acid-base titrations and empirical equilibrium models. Figure 3 shows the percent site density contributions in the form of pie charts. Subplots a, b, and c correspond to the BIOS, bacteria, and ferrihydrite fractions, respectively, while d corresponds to goethite. Subplot a, indicates an acidic, neutral, and basic site contribution of 23, 33, and 44% respectively. Subplot b shows the proportion of the same site types in the bacteria to be 37, 6, and 57%. In the ferrihydrite fraction the acidic site contribution (18%) was slightly lower than that in the mixed BIOS (23%). The neutral and basic sites, in subplot c, contributed 35 and 47% respectively. Basic sites in the bacteria constituted 57% of the total normalized site density, compared to 44 and 47% in the BIOS and ferrihydrite fractions, respectively. Subplot b in Figure 3, for the bacterial BIOS fraction, differs considerably from subplots a, c, and d corresponding to the mixed BIOS, ferrihydrite, and goethite, respectively. This discrepancy and the agreement in percent site density contributions in subplot a, with those of c and d, suggest masking of bacterial surface functional groups by iron oxyhydroxides (18, 19). The results for goethite (subplot d) calculated using MUSIC model values (6, 33) and comparable to those in subplot a emphasize the iron oxyhdroxide nature of the BIOS surface. This was further supported by the apparent BIOS pHpzc (9.6 ( 0.1), which is within the range of previously reported pHpzc values for other iron oxyhydroxide phases (26, 29, 32). The results shown in this study attest to the complicated structural nature of the BIOS mixture. Such an arrangement, composed of intact and fragmented bacteria embedded within a poorly ordered iron oxide phase, is bound to have uncertainties in size, geometry, and penetrability by counterions, as suggested previously for humics (35). These characteristics would limit the ability of electrostatic models (CD-MUSIC and the classic 2-pK methods) to predict intrinsic speciation chemistry (26, 35, 36). An empirical model such as FOCUS, which implicitly embodies electrostatic parameters, would provide a more realistic approximation of the BIOS surface chemistry (35, 36). The analysis of FOCUS apparent pKa spectra and percent binding site contributions for the BIOS and its fractions indicated masking of bacterial surface functional groups by iron oxyhydroxides. This observation and a pHpzc of 9.6 further imply that the reactive surface of the mixed BIOS is comparable to that of pure iron oxyhydroxides. As a result, the mixture’s metal binding capacity should not be regarded as a direct combination of the reactive surface heterogeneity of the organic and mineral fractions. The mixed BIOS, along with humics, organic ligands, and mineral colloids, are considered to play an essential role in the control of metal partitioning in pristine and contaminated aquatic environments. Bacterial surfaces serve as nucleation sites for mineral oxide colloids; however, their ability to influence BIOS metal partitioning is largely attenuated by specific surface chemical interactions with the iron oxyhydroxide phase. 5676
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003
Although the FOCUS model does not explicitly account for electrostatic interactions, the method is useful for the analysis of the surface heterogeneity of a natural sample, such as the BIOS, under largely invariant environmental conditions of solution composition and ionic strength. The FOCUS model allows for mathematical manipulation of pKa spectra (deconvolution) as demonstrated for the surface reactivity analysis of the BIOS inorganic phase. Martinez et al. combined the FOCUS model with a master curve approach to remove electrostatic effects and determine intrinsic parameters for intact bacterial cells (8). In addition, this model determined the number of BIOS binding sites as well as their pKa’s and concentrations without the need of a close initial guess for pKa or binding site concentration values (6, 8, 34).
Acknowledgments This work was founded by the Natural Sciences and Engineering Research Council (NSERC) of Canada and an Ontario Premier’s Research Excellence Award (PREA).
Note Added after ASAP Posting This paper was released ASAP on 10/31/2003 with a minor error in the caption of Figure 2. The correct version was posted on 11/06/2003.
Literature Cited (1) Lovley, D. R. Environmental Microbe-Metal Interactions; ASM Press: Washington, DC, 2000. (2) Ferris, F. G.; Konhauser, K. O.; Lyve´n, B.; Pedersen, K. Geomicrobiol. J. 1999, 16, 181. (3) Ferris, F. G.; Hallberg, R. O.; Lyve´n, B.; Pedersen, K. Appl. Geochem. 2000, 15, 1035. (4) Warren, L. A.; Ferris, F. G. Environ. Sci. Technol. 1998, 32, 2331. (5) Harris, D. C. Quantitative Chemical Analysis; Freeman: New York, 1995. (6) Smith, D. S.; Ferris, F. G. Environ. Sci. Technol. 2001, 35, 4637. (7) Fein, B. J.; Daughney, C. J.; Yee, N.; Davis, T. A. Geochim. Cosmochim. Acta 1997, 61, 3319. (8) Martinez, R. E.; Smith, D. S.; Kulczycki, E.; Ferris, F. G. J. Colloid Interface Sci. 2002, 253, 130. (9) Sokolov, I.; Smith, D. S.; Henderson, G. S.; Gorby, Y. A.; Ferris, F. G. Environ. Sci. Technol. 2001, 35, 341. (10) Cox, J. S.; Smith, D. S.; Warren, L. A.; Ferris, F. G. Environ. Sci. Technol. 1999, 33, 4514. (11) Beveridge, T. J. Int. Rev. Cytol. 1981, 72, 229. (12) Beveridge, T. J. In Antibiotic Inhibition of Bacterial Cell: Surface Assembly and Funcion; Actor, P., Daneo-Moore, L., Higgins, M. L., Salton, M. R. J., Shockman, G. D., Eds.; Soc. Microbiol: Washington, DC, 1988; p 3. (13) Jambor, J. L.; Dutrizac, J. E. Chem. Rev. 1998, 98, 2549. (14) Rustad, J. R.; Felmy, A. R.; Hay, B. P. Geochim. Cosmochim. Acta 1996, 60, 1563. (15) Felmy, A. R.; Rustad, J. R. Geochim. Cosmochim. Acta 1998, 62, 25. (16) Hiemstra, T.; Van Riemsdijk, W. H. J. Colloid Interface Sci. 1996, 179, 488. (17) Hiemstra, T.; Venema, P.; Van Riemsdijk, W. H. J. Colloid Interface Sci. 1996, 184, 680. (18) Small, T. D.; Warren, L. A.; Ferris, F. G. Appl. Geochem. 2001, 16, 939. (19) Vermeer, A. W. P.; McCulloch, J. K.; Van Riemsdijk, W. H.; Koopal, L. K. Environ. Sci. Technol. 1999, 33, 3892. (20) Kennedy, C. B.; Martinez, R. E.; Scott, S. D.; Ferris, F. G. Geobiology 2003, 1, 59. (21) Cornell, R. M.; Schwertmann, U. The Iron Oxides; VCH: New York, 1996. (22) Plette, A. C. C.; Van Riemsdijk, W. H.; Benedetti, M. F.; Van der Wal, A. J. Colloid Interface Sci. 1995, 175, 354. (23) Emerson, D.; Moyer, C. L. Appl. Environ. Microbiol. 2002, 68, 3085. (24) Santelli, C. M.; Welch, S. A.; Westrich, H. R.; Banfield, J. F. Chem. Geol. 2001, 180, 99. (25) Kraepiel, A. M. L.; Keller, K.; Morel, F. M. M. Environ. Sci. Technol. 1998, 32, 2829. (26) Dzombak, D. A.; Morel, F. M. M. Surface Complexation Modeling: Hydrous Ferric Oxide; Wiley: New York, 1990.
(27) Stumm, W.; Morgan, J. J. Aquatic Chemistry; Wiley-Interscience: New York, 1996. (28) Rustad, J. R.; Wasserman, E.; Felmy, A. R. Surf. Sci. 1999, 424, 28. (29) Boily, J.-F.; Lu ¨ tzenkirchen, J.; Balme`s, O.; Beattie, J.; Sjo¨berg, S. Colloid Surf. A 2001, 179, 11. (30) Croue´, J.-P.; Benedetti, M. F.; Violleau, D.; Leenheer, J. A. Environ. Sci. Technol. 2003, 37, 328. (31) Contescu, C.; Jagiello, J.; Schwarz, J. A. Langmuir 1993, 9, 1754. (32) Rustad, J. R.; Dixon, D. A.; Felmy, A. R. Geochim. Cosmochim. Acta 2000, 64, 1675. (33) Venema, P.; Hiemstra, T.; Weidler, P. G.; Van Riemsdijk, W. H. J. Colloid Interface Sci. 1998, 198, 282.
(34) Smith, D. S.; Ferris, F. G. J. Colloid Interface Sci. 2003, in press. (35) Westall, J. C.; Jones, J. D.; Turner, G. D.; Zachara, J. M. Environ. Sci. Technol. 1995, 29, 951. (36) Ganguly, C.; Huang, C.; Rabideau, A. J.; Van Benschoten, J. E. J. Environ. Eng. 1999, 125, 712.
Received for review March 21, 2003. Revised manuscript received August 12, 2003. Accepted September 15, 2003. ES0342603
VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5677
