Solubility and dissimilatory reduction kinetics of ... - Semantic Scholar

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 73 (2009) 5273–5282 www.elsevier.com/locate/gca

Solubility and dissimilatory reduction kinetics of iron(III) oxyhydroxides: A linear free energy relationship Steeve Bonneville *, Thilo Behrends, Philippe Van Cappellen 1 Department of Earth Sciences – Geochemistry, Utrecht University, P.O. Box 80.021, TA 3508 Utrecht, The Netherlands Received 24 February 2009; accepted in revised form 4 June 2009; available online 16 June 2009

Abstract Rates of reduction of Fe(III) oxyhydroxides by the bacterium Shewanella putrefaciens were measured as a function of the bacterial density and the Fe(III) substrate concentration. The results show that an earlier reported positive correlation between the solubility products (*Kso) and the maximum cell-specific reduction rates (vmax) of predominantly poorly crystalline Fe(III) oxyhydroxides also applies to insoluble and crystalline Fe(III) oxyhydroxides. The mineral solubilities were measured by a dialysis bag technique under acidic conditions (pH 1 up to 2.5) at 25 °C. Initial iron reduction rates by S. putrefaciens were determined in the presence of excess lactate as electron donor. In all cases, the microbial reduction rate exhibited saturation behavior with respect to the Fe(III) oxyhydroxide concentration. On a double logarithmic scale, the maximum rates vmax and the solubility products defined a single linear free energy relationship (LFER) for all the Fe(III) oxyhydroxides considered. The solubility provided a better predictor of vmax than the specific surface area of the mineral phase. A rate limitation by the electron transfer between an iron reductase and a Fe(III) center, or by the subsequent desorption of Fe2+ from the iron oxide mineral surface, are both consistent with the observed LFER. Ó 2009 Elsevier Ltd. All rights reserved.

1. INTRODUCTION Iron(III) oxyhydroxides (collectively referring here to Fe(III) oxide phases) are major terminal electron acceptors in aquifers (Lyngkilde and Christensen, 1992; Jakobsen and Postma, 1999), soils (Van Breemen, 1988), marine sediments (Canfield et al., 1993) and estuarine sediments (Hyacinthe et al., 2006; Lin et al., 2007). Evidence gathered over the last decades shows that microbial (enzymatic) catalysis dominates Fe(III) oxyhydroxides reduction in non-sulfidogenic sediments (Lovley et al., 1991). Iron(III) oxyhydroxides occur in soils and sediments in a variety of forms, ranging from amorphous phases, such as ferrihydrite, to well-crystallized minerals, such as hematite or goethite. * Corresponding author. Present address: School of Earth and Environment, University of Leeds, LS2 9JT Leeds, United Kingdom. E-mail address: [email protected] (S. Bonneville). 1 Present address: School of Earth and Atmospheric Sciences, Georgia Institute of Technology, Atlanta, GA 30332-0340, United States.

0016-7037/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.gca.2009.06.006

Even a single Fe(III) oxyhydroxide mineral may exhibit a range of crystallinity, particle size, specific surface area, and solubility. Such variations in mineral properties result in a continuum of reactivity. Larsen and Postma (2001), for instance, showed that the abiotic reduction kinetics by ascorbate decrease systematically in the order ferrihydrite > lepidocrocite > goethite > hematite. The exact dependence of enzymatic Fe(III) reduction kinetics on mineral properties is still a matter of debate. Munch and Ottow (1983) showed that amorphous ferric compounds are reduced preferentially over crystalline Fe(III) oxyhydroxides. These authors further observed that, among the crystalline ferric phases, the reactivity towards microbial iron reduction decreases according to the sequence lepidocrocite > hematite > goethite. In a later study, Roden and Zachara (1996) proposed that the specific surface area of Fe(III) oxyhydroxides exerts a major control on the rate and extent of microbial Fe(III) oxyhydroxide reduction. More recently, we demonstrated that microbial iron reduction exhibits a Michaelis–Menten kinetic dependence with respect to the concentration of the

5274

S. Bonneville et al. / Geochimica et Cosmochimica Acta 73 (2009) 5273–5282

Fe(III) substrate (Bonneville et al., 2004). Moreover, for Fe(III) oxyhydroxides the maximum reduction rate per cell, vmax, was found to correlate positively with the solubility product of the mineral. Solubility products of Fe(III) oxyhydroxides reported in the literature cover more than three orders of magnitude. Even for a single mineral, it is common to find a range of 1–2 orders of magnitude (Schwertmann and Cornell, 1991). This variability arises from differences in mineral properties, for example differences in crystallinity, grain size and impurity content, but also from difficulties associated with the experimental determination of solubility products. In our previous study, we performed, a series of pe–pH titrations in the pH range 4–7 in presence of Fe2þ ðaqÞ , and derive the solubility products of several Fe(III) oxyhydroxides including lepidocrocite (Bayferrox), nanohematite, 6-line ferrihydrite and hydrous ferric oxide (HFO) (Bonneville et al., 2004). The high specific surface area coupled with the relatively high solubility of the latter Fe(III) oxyhydroxides permit fast reversible electron transfer with the platinum electrode allowing for equilibrium redox potential specific of each Fe(III) oxyhydroxide. However, when tested with crystalline and insoluble Fe(III) oxyhydroxides exhibiting low surface (e.g., hematite or goethite), the redox potential measured between pH 4 and 7 could not stabilize even after several days of equilibration. As the pe–pH failed for crystalline Fe(III) oxyhydroxides, the positive correlation between vmax and solubility products was mainly restricted to nanoparticulate Fe(III) phases (i.e. HFO, nanohematite and six lines ferrihydrite) and a lepidocrocite (Bonneville et al., 2004). However, other more crystalline Fe(III) oxyhydroxides, such hematite and goethite, are ubiquitous in subsurface and often serve as electron acceptor for dissimilatory iron reduction. The goal of present paper is, therefore, to assess to what extent the dependencies of the microbial Fe(III) reduction kinetics on the amount and solubility of the Fe(III) substrate observed for the soluble and poorly crystalline Fe(III) oxyhydroxides also apply to the more insoluble and crystalline Fe(III) phases. In order to do so, the dialysis method proposed by Kuma and coworkers (1992, 1993) was used to measure precisely the solubility product, * K Kso, of a number of Fe(III) crystalline oxyhydroxides, including hematite and goethite. In addition, even though a lepidocrocite was already used previously, we included in the present study two other lepidocrocites to complement the range of Fe(III) oxyhydroxides considered. The same Fe(III) solids were incubated anaerobically with the bacterium Shewanella putrefaciens in the presence of excess lactate as electron donor. The kinetic results were then compared to those obtained previously with less crystalline, more soluble Fe(III) oxyhydroxides. 2. MATERIALS AND METHODS 2.1. Fe(III) oxyhydroxides Three crystalline Fe(III) oxyhydroxides (two hematites and one goethite) and two lepidocrocites were used in the experiments. Low surface area (LSA) hematite (Bayferrox

105M) and goethite (Bayferrox 910) were purchased from Scholz & Co. GmbH. Hematite (Merck) and lepidocrocite (Alfa Aeser) were purchased from Merck and Alfa Aeser, respectively. The solids were used as received. Lepidocrocite 6.8 was prepared according to the method described in Schwertmann and Cornell (1991): 11.93 g of unoxidized crystals of FeCl24H2O was dissolved in 300 mL of demineralized water. The pH was adjusted to 6.8 using a 1 M solution of NaOH and exposed to air at a flow rate of 100 mL min1. Additional NaOH was added to neutralize proton production during oxidation and maintain the pH at 6.8. Hydrous ferric oxide (HFO) was used as a representative poorly crystalline nanoparticulate phase. It was synthesized by neutralizing a 0.4 M solution of FeCl36H2O with 1 M NaOH and subsequently dialyzing against demineralized water to remove Na+ and Cl ions. The specific surface areas of the solids, measured by the N2 BET method, are given in Table 1. 2.2. Solubility measurements Solubilities of Fe(III) oxyhydroxides were derived by determining the activities of aqueous ferric iron, Fe3þ ðaqÞ , in equilibrium with the solids over a range of pH (1–2.5) at 25 °C), using a method inspired from Kuma et al. (1992). Cellulosic dialysis bags were rinsed for a few hours with running demineralized water and further washed with a 10 mM NaNO3 solution adjusted to the pH of the dissolution experiment. The dialysis membrane bags were filled with 20 mL of concentrated Fe(III) oxyhydroxide suspension, and then placed in 100 mL glass bottles. Total iron concentrations were around 100 mM. The glass bottles contained 10 mM sterilized NaNO3 solution and the pH was adjusted to 1, 1.5, 2 and 2.5 with 1 M HNO3. The dialysis bags were completely submerged and kept in the dark under aerobic conditions in a thermo-stated water bath at 25 °C. We chose cellulosic dialysis bags because they exhibit minimal adsorption affinity for di- and tri-valent metallic cations at the pH range considered in our experiments (Farrah and Pickering, 1978; Weltje et al., 2003). Aliquots from the solution surrounding the dialysis membrane bags were collected regularly to monitor the build up of total dissolved Fe and dissolved Fe2þ ðaqÞ . The concentrations of Fe2þ ðaqÞ in the samples were measured using the ferrozine assay method (Viollier et al., 2000). Total dissolved Fe concentrations were determined by mixing 25– 300 lL of sample with a hydroxylamine hydrochloride solution (1.4 M in 2 M HCl). The hydroxylamine reduced Fe(III) to Fe(II) and, after 30 min, the total dissolved Fe(II) was measured by the ferrozine assay. To check whether particulate Fe(III) may have breached the dialysis membrane, 200 lL of sample was mixed with 100 lL of 2 M HCl and placed in an oven overnight to dissolve any Fe(III) particles. The acidified solution was then mixed with hydroxylamine solution, and the total dissolved Fe concentration was measured by ferrozine assay. The solubility experiments were terminated when the total dissolved Fe concentration no longer increased. The final pH was recorded using a Ross glass electrode calibrated with pH 1, 2 and 3 solutions freshly prepared

Solubility and dissimilatory reduction kinetics of Fe(III) oxyhydroxides

5275

Table 1 Fe(III) oxyhydroxides used in this study(1) and in Bonneville et al. (2004)(2). Errors on cell densities measurements and on the reduction rate calculations were accounted into the error intervals reported for the vmax values. Iron oxyhydroxide

Surface area (m2 g1)

LSA hematite Hematite (Merck) Goethite (Bayferrox) Lepidocrocite (Alfa Aeser) Lepidocrocite 6.8 HFO (dialysis)

12 3 15 83 75 600*

Lepidocrocite (Bayferrox) Nanohematite 6 line Ferrihydrite HFO (pe–pH)

15 125 175 600*

*

Ratio (n)

vmax (1017 mol h1 cell1)

1.22 ± 0.38 1.05 ± 0.20 0.66 ± 0.17 0.68 ± 0.50 1.11 ± 0.40 1.78 ± 0.04

2.60 2.75 2.84 2.61 2.83 2.74

0.28 ± 0.04 0.99 ± 0.13 1.28 ± 0.03 3.39 ± 0.41 5.48 ± 0.67 6.50 ± 0.30

(1) (1) (1) (1) (1) (1)

0.46 ± 0.10 0.52 ± 0.15 1.62 ± 0.27 1.90 ± 0.13

2.65 2.85 2.75 2.71

2.00 ± 0.10 2.40 ± 0.15 8.10 ± 0.30 6.50 ± 0.30

(2) (2) (2) (2)

log *Kso

Specific surface area reported in Roden and Zachara. (1996).

by dilution of a 1 M HCl solution (Titrisol). The final pH values were used to calculate the solubility products (see Section 4.1). The total Fe3þ ðaqÞ concentration was obtained by subtracting the measured Fe2þ ðaqÞ concentration from the measured total dissolved Fe concentration. The formation of Fe(III) solution complexes, predominantly FeðOHÞ2þ ðaqÞ in the pH range considered, were accounted for assuming equilibrium speciation. For the latter calculation, we used the pH measured in the dissolution experiments and a log K of 2.187 as a value for the formation constant of 3þ FeðOHÞ2þ ðaqÞ . The free FeðaqÞ ion concentration was therefore estimated as follows: diss 2þ 2þ ½Fe3þ ðaqÞ ¼ ½Fetot ½FeðaqÞ ½FeðOHÞðaqÞ

ð1Þ

½Fediss tot

is the total dissolved Fe concentration, ½Fe2þ where aq the dissolved ferrous iron concentration, and ½FeðOHÞ2þ ðaqÞ the concentration of the aqueous Fe3+ hydroxyl complex. The activity of Fe3þ ðaqÞ was then calculated using the extended Debye–Hu¨ckel equation, with a, the size of Fe3+ ion, equal ˚ . The variation in ionic strength due to the pH range to 9 A in the dissolution experiments (varying from 0.013 to 0.11 M for pH 1 to pH 2.5, respectively) were accounted for the calculation of Fe3þ ðaqÞ activity. 2.3. Microbial rate measurements Cultures of S. putrefaciens 200R were provided by Dr. T. DiChristina from the Georgia Institute of Technology (DiChristina et al., 1988, 2002). Cells were kept aerobically on Luria Bertani medium (LB) and routinely cultivated in liquid LB medium on a rotary shaker at room temperature. The bacteria were washed, concentrated in 50 mL of 0.1 M NaCl, and the cell densities were measured by direct countings after acridine staining using epifluorescence microscopy (Hobbie et al., 1977) following the procedure described in Bonneville et al. (2006). For each Fe(III) oxyhydroxide, 24 mineral-bacteria incubation experiments were performed, using three cell densities (from 8.5 107 to 109 cells mL1) at eight different Fe(III) substrate concentrations (0–60 mM Fe(III)). The incubations were performed in air-tight 100 mL glass bottles under anaerobic conditions. The composition of the medium was: 10 mM lactate, 10 mM

Hepes buffer, 28 mM NH4Cl and 1 mM CaCl22H2O. After suspending the Fe(III) oxyhydroxide in the medium, the pH was adjusted to 7 and the final ionic strength was calculated to be around 0.05 M. The mineral suspensions were degassed with N2 for 2 h prior to being inoculated with bacteria. The bottles were kept on a rotary table in a temperature-controlled room (22–23 °C) and aliquots were collected to monitor the production of Fe2+ for 24 h after inoculation to insure that measurements are representative of initial rates of microbial iron reduction. In doing so, we also avoid the precipitation of secondary Fe(II) minerals, the potential transformation of Fe(III) oxyhydroxides and also limit the loss of viability of the bacteria suspension. Aliquots were extracted for 1 h with 0.5 M HCl, this procedure was showed to re-solubilize efficiently the adsorbed Fe(II) from the surface of Fe(III) oxyhydroxides (Hyacinthe et al., 2008). After the extraction step, the dissolved Fe2+ concentrations in the aliquots were measured with the ferrozine method (Viollier et al., 2000). In Bonneville et al. (2004), control experiments without bacteria with the same medium as described above were performed with a range of Fe(III) oxyhydroxides. Over 24 h incubations, the dissolved iron concentrations was below the detection limit of the ferrozine method.

3. RESULTS 3.1. Fe(III) oxyhydroxide solubilities The total Fe concentrations measured on acid-digested samples from the solutions surrounding the dialysis bags were indistinguishable from the dissolved Fe concentrations (results not shown), implying the effective retention of the Fe(III) oxyhydroxide particles by the dialysis membrane. For all Fe(III) oxyhydroxides, except HFO, the measured total dissolved Fe concentrations reached a plateau at the four pH values, as illustrated for one of the lepidocrocites in Fig. 1a. The noticeable exception was HFO, for which total dissolved Fe concentrations reached a maximum after about 400 h, and then slowly decreased with time (Fig. 1a and b). The time required to achieve dissolution equilib-

5276


and 2. In contrast, higher contributions of Fe2þ ðaqÞ were observed at pH 2.5, in particular for the experiments with hematite (Merck), LSA hematite and goethite, where up to 50%–70% of the total dissolved Fe occurred in the reduced state. The free Fe3þ ðaqÞ ion concentrations, calculated with Eq. (1), followed the same trends as the total dissolved Fe concentrations (Fig. 1a). They systematically decreased with increasing pH (Fig. 1b). On a log-pH plot, the computed Fe3þ ðaqÞ activities defined linear relationships with slopes close

rium varied from around 500 h (20 days) for HFO to about 4000 h (166 days) for LSA hematite. The fraction of Fe(III) oxyhydroxide dissolved to reach equilibrium ranged from 0.01% (LSA hematite, pH 2.5) to about 50% (HFO, pH 1). In all the solubility experiments, dissolved Fe2þ ðaqÞ was detected in the solution surrounding the dialysis bags. The highest Fe2þ ðaqÞ concentration was observed for HFO at pH 1 (82 lM). Ferrous iron represented only a small fraction (610%) of total dissolved Fe in experiments at pH 1, 1.5

(a)

(b) 105

103

102

101

Fe

3+ (aq)

concentration in µmol L

-1

104

10

pH 1 pH 1.5 pH 2

0

pH 2.5

10-1 0

500

1000

2000

2500

Time (h)

Fig. 1. (a) Total dissolved Fe (open symbols) and dissolved Fe3þ ðaqÞ (closed symbols) versus time in triplicate dialysis bag dissolution experiments with HFO and lepidocrocite (Alfa Aeser), at pH 2. The Fe3þ ðaqÞ concentrations plotted correspond to the hexaaquo ferric iron ion 2þ tot 2þ 3þ concentrations (½Fe3þ ðaqÞ ¼ ½FeðaqÞ ½FeðaqÞ ½FeðOHÞðaqÞ ; see text). (b) Build-up of FeðaqÞ during HFO dissolution at pH 1, 1.5, 2 and 2.5. The

dotted lines represent the maximum Fe3þ ðaqÞ concentrations used to calculate the solubility product of HFO (see text for discussion). Note the logarithmic concentration scale.


5277

mineral phase defined a single trend when normalizing the rates to the cell density (Fig. 2). Hence, the iron reduction rates could be fitted to the Michaelis–Menten rate expression: R ¼ vmax B

Fig. 2. Equilibrium activities of Fe3þ ðaqÞ in the dialysis bag experiments as a function of pH for the various Fe(III) oxyhydroxides, at 25 °C. The dotted lines are linear fits of log aFe3þ versus pH, for the individual minerals. The corresponding slopes of the linear fits are given in Table 1.

to 3 (Fig. 2). The significance of the slopes is discussed in Section 4.1. 3.2. Microbial Fe(III) reduction kinetics The build up of total Fe(II) was quasi-linear (i.e. r2 > 0.9) during the 24 h following inoculation of the mineral suspensions with the iron-reducing bacteria (data not shown here – examples of Fe2+ production are presented in Bonneville et al. (2004, 2006) and Hyacinthe et al. (2008)). Because of the relatively low initial concentrations of Fe(III) oxyhdyroxides, cell densities and the short period of incubation, no precipitation of secondary Fe(II) or mixed Fe(II)/Fe(III) phases was observed in our experiments. Typically, precipitates are forming in systems with high initial loads of Fe(III), electrons donors and cells, for instance, magnetite and green rusts were observed in experiments with 300 mM of lepidocrocite with 75 mM of lactate incubated with 1010 cells mL1 in presence of electron shuttle and after 6 days of incubations (Ona-Nguema et al., 2002). Pedersen et al. (2005) reported transformations of two line ferrihydrite into goethite within two days of exposure with high dissolved Fe(II) concentrations relative to Fe(III) oxyhydroxides (Fe(II)/Fe(III) ratio of 0.5–2). Of all our 24 experiments with HFO, two incubations reached a Fe(II)/Fe(III) oxyhydroxide ratio of 0.5, however only at the end of the 24 h incubation period. Therefore, even though it is difficult to rule out any transformation of HFO, the relatively low Fe(II) production over our shortterm incubations are likely to have prevented a significant transformation of HFO. Initial iron reduction rates were calculated by linear least square regression of the total Fe(II) concentrations versus time. The reduction rates exhibited saturation with respect to the initial Fe(III) substrate concentration (Fig. 2). Each

½FeIII tot K m þ ½FeIII tot

ð2Þ

where R is the iron reduction rate per unit volume suspension, vmax is the maximum cell-specific Fe(III) reduction rate, B is the cell density of the suspension, Km is the affinity constant for the substrate (i.e. the Fe(III) concentration at half vmax), and ½FeIII tot is the (initial) Fe(III) concentration per unit volume of suspension. The optimized values of the maximum cell-specific iron reduction rate for the minerals considered here are listed in Table 1, together with those obtained previously (Bonneville et al., 2004). (Note: the interpretation of the affinity constant, Km, is discussed in detail in Bonneville et al. (2006); it falls outside the scope of the present study.) 4. DISCUSSION 4.1. Fe(III) oxyhydroxide solubilities The presence of Fe2þ ðaqÞ in the dialysis bag experiments must be due to reduction of ferric iron in contact with the cellulosic membrane. Several studies have demonstrated that a variety of organic compounds are able to reduce Fe3þ ðaqÞ in the dark at low pH (Fukushima and Tatsumi, 1999; van Schaik et al., 2008; Weber et al., 2006). In order to obtain meaningful solubility products, it is thus important to account for the redox speciation of dissolved Fe in the equilibration experiments. Under the acidic conditions of the experiments (pH 62.5), the mineral suspensions all reach steady state total dissolved Fe and corresponding dissolved Fe3þ ðaqÞ concentrations, with the exception of HFO (Fig. 1a). We can thus reasonably assume that the steady state concentrations correspond to solubility equilibrium values. In the case of HFO, the drop in dissolved Fe concentration beyond 500 h (Fig. 1a) most likely reflect the aging of the mineral phase, for example, due to Ostwald ripening or recrystallization (Steefel and Van Cappellen, 1990). In what follows, the maximum Fe3þ ðaqÞ concentrations are used to compute the solubility product of HFO. Note that the decrease in solubility of HFO occurs on time scales that largely exceed the 24-h period over which the microbial reduction rates are measured. Using hematite as an example, the solubility equilibrium reaction of a pure, stoichiometric Fe(III) oxyhydroxide mineral can be written as 1 3 3þ Fe2 O3ðsÞ þ 3Hþ ðaqÞ () FeðaqÞ þ H2 O 2 2

ð3Þ

and the solubility product is expressed as

K so ¼ aFe3þ a3 Hþ

ð4Þ

which can be rearranged to give log K so ¼ log aFe3þ þ 3 pH

ð5Þ

5278


Thus, only the activity of the hexaquo ferric iron complex * (i.e. Fe3þ ðaqÞ ) and the pH are required to derive Kso. Eqs. (4) and (5) predict that log aFe3þ values plotted versus pH should define a straight line with a slope of 3. In practice, deviations from the theoretical value of 3 occur because of non-stoichiometric compositions (Bonneville et al., 2004). The slopes in Fig. 2 are indeed close, but not equal to 3. They vary from 2.60 (LSA hematite) to 2.84 (goethite). Previous studies have shown that stoichiometric O(H)/Fe ratios of Fe(III) oxyhydroxides mostly fall between 2 and 2.7 (Biedermann and Chow, 1966; Spiro et al., 1966; Murphy et al., 1976; Dousma and De Bruyn, 1978; Fox, 1988), due to the incorporation of counter ions, such as NO3, OH, SO42 or Cl, in the mineral structure. To account for non-stoichiometric compositions, Fox (1988) proposes to use the alternative solubility product:

K so ¼ aFe3þ an Hþ

ð6Þ

where n is the observed slope between log aFe3þ and pH. The experimentally derived solubility products and their standard deviations, as well as the values of n can be found in Table 1, together with the values obtained previously for soluble and poorly crystalline Fe(III) oxyhydroxides. The solubility of the same HFO as used here was previously determined by measuring the redox potentials of a Fe2þ ðaqÞ -amended mineral suspension over a pH range of 4– 7 (Bonneville et al., 2004). The values of log *Kso (1.90 ± 0.13) and n (2.71) obtained using this pe–pH titration method agree well with those derived from the dialysis bag equilibration experiments (log *Kso = 1.78 ± 0.04; n = 2.74), despite the very different experimental approaches and pH conditions. The range of stoichiometric ratios n are also similar for both methods (Table 1), and comparable to n values reported in the literature for a variety of Fe(III) oxyhydroxides (Biedermann and Chow, 1966; Spiro et al., 1966; Murphy et al., 1976; Dousma and De Bruyn, 1978; Fox, 1988; Byrne and Luo, 2000; Byrne et al., 2000). In our earlier study, we estimated the solubility of LSA hematite from redox potential measurements of a Fe2þ ðaqÞ amended mineral suspension at pH 2 after an equilibration period of 4 days (Bonneville et al., 2004). Because the measurements were restricted to a single pH, no value for the stoichiometric ratio, n, was obtained. Using a range of n values (2.65–3), we derived log *Kso values for LSA hematite from 0.82 to 0.02. These values deviate substantially from the solubility product derived from the dialysis equilibration experiments (1.22, Table 1). Most likely, the discrepancy reflects the inability of measuring true equilibrium redox potentials in suspensions of crystalline and insoluble iron(III) oxyhydroxides. The solubility products reported in Table 1 range over more than three orders of magnitude. Significant variability is also seen for the same mineral phase: nearly two orders of magnitude separate the solubility product of LSA hematite from that of nanohematite. The three different lepidocrocites exhibit smaller, but significant, differences in solubility. Taken together, the solubilitiy of the Fe(III) oxyhydroxides in Table 1 increase in the order: (crystalline) hema-

tites < goethite < nanohematite 6 lepidocrocites < 6 ferrihydrite < HFO.

line

4.2. Microbial reduction rates In the absence of limitation by the electron donor, the initial reduction kinetics of all the Fe(III) oxyhydroxides considered can be described by Eq. (2). We developed a kinetic model explaining this behavior for HFO, 6 line ferrihydrite, lepidocrocite (Bayferrox), nanohematite and LSA hematite (Bonneville et al., 2006). The model assumes that the reduction rate per cell is proportional to the amount of cell surface in contact with Fe(III) particles. The maximum cell-specific reduction rate, vmax, then corresponds to the rate when the entire cell surface is in contact with Fe(III) particles. Under these conditions, Fe(III) reductases, presumably located on the outer membrane (Myers and Myers, 1997; Beliaev and Saffarini, 1998; Blakeney et al., 2000), are transferring electrons to Fe(III) centers at the mineral surface at the maximum possible rate. Increasing the concentration of Fe(III) substrate beyond this point no longer causes a further increase in the reduction rate. In the Michaelis–Menten model for enzymatic reactions, the rate of reaction is proportional to the concentration of the enzyme–substrate complex, [ES], according to: v ¼ k ½ES

ð7Þ

where k is the first-order rate constant describing the dissociation of the complex. In the kinetic model for cell-surface catalyzed reduction of Fe(III) particles, the enzyme E corresponds to the Fe(III) reductases and the substrate S to the Fe(III) centers at the cell-mineral interface. This is consistent with the much lower average density of c-type cytochromes at the surface of S. putrefaciens (0.05 nm2) compared to that of Fe(III) centers at the surface of Fe(III) oxyhydroxides (2 nm2, see Bonneville et al., 2006, for details). It further implies that the maximum value of [ES] (i.e. when the cell surfaces are saturated with Fe(III) colloids) is an intrinsic property of the iron reducing cells, while the properties of the mineral therefore enter the reduction kinetics via the rate constant, k. The kinetic results presented here (Fig. 3) and in our earlier study (Bonneville et al., 2004) show that the Michaelis– Menten rate dependence on the availability of the Fe(III) substrate is a robust feature of iron reduction by S. putrefaciens. In particular, it applies to a large spectrum of Fe(III) oxyhydroxides which, together, exhibits a huge range of mineral, morphology and surface properties. 4.3. Linear free energy relationship On a log–log plot, the maximum cell-normalized rates of iron reduction by S. putrefaciens and the solubility products of the corresponding Fe(III) oxyhydroxides exhibit a linear relationship (Fig. 4): log vmax ¼ 0:40 log K so 16:79

ðr2 ¼ 0:90Þ

ð8Þ

where vmax is expressed in mol iron reduced per hour and per cell. The relationship in Fig. 4 is qualitatively similar

Hematite Merck 12x10-12 10x10-12 8x10-12 6x10-12 4x10-12

2.6x108 cells mL-1 3.7x108 cells mL-1

-12

7x108 cells mL-1

2x10

Michaelis-Menten kinetic model

0 0

10

20

30

40

50

Reduction rate in micromol h-1 cell-1



Lepidocrocite Alfa Aeser 30x10-12 25x10-12 20x10-12 15x10-12 10x10-12

8.5x107 cells mL-1 2.5x108 cells mL-1

5x10

5.7x108 cells mL-1

-12


0

0

10

Fe(III) initial in mM

12x10-12 10x10-12 8x10-12 6x10-12 1.5x108 cells mL-1 4.8x108 cells mL-1 -1

8.2x10 cells mL

2x10-12


0 0

10

20

30

40

50

60




Goethite Bayferrox

8

20

30

40

50


14x10-12

4x10-12

5279

Lepidocrocite 6.8 80x10-12

60x10-12

40x10-12 1.8x108 cells mL-1

20x10-12

5x108 cells mL-1 109 cells mL-1 Michaelis-Menten kinetic model

0 0

10

20

30

40

50

60


Fig. 3. Initial, cell-normalized rates of Fe(III) reduction by S. putrefaciens, as a function of the initial concentrations of Fe(III) substrate. Errors bars for the rates correspond to the standard deviations on the slopes of the linear regressions of total Fe2+ concentration versus time, for the 24 h duration of the incubations. The solid lines are the best fits of Eq. (2) to the data.

to the one presented in Bonneville et al. (2004). The latter, however, only included one insoluble and crystalline endmember mineral phase, LSA hematite, whose previously estimated solubility product was likely overestimated (see above), thereby yielding a significantly different slope and intercept of the relationship. The log *Kso values of the Fe(III) oxyhydroxide are related to the standard Gibbs energies of the corresponding dissolution reactions and also reflect the thermodynamic stability associated to the lattice structure of each Fe(III) oxyhydroxides. Therefore, the log *Kso values of the Fe(III) oxyhydroxides do not depend on the chemical composition or pH of the medium and can be used to describe the thermodynamic stability of the Fe(III) phases within our incubations experiments. In this view, the Eq. (8) thus defines a linear free energy relationship (LFER) for the initial reduction of Fe(III) oxyhydroxides by S. putrefaciens. A large number of LFERs have been reported in the literature, especially for organic reaction series (Leffler and Grunwald, 1989), but also for (abiotic) mineral dissolution processes (Wollast, 1974; Wieland et al., 1988; Duckworth et al., 2004). A LFER is usually assumed to reflect the sensitivity of a series of reactions proceeding via the same

mechanism to the composition of the reaction medium or to structural changes in the reactants or products. Roden and Zachara (1996) and Roden (2003) proposed that the specific surface area of Fe(III) oxyhydroxides is the dominant factor controlling microbial iron reduction rates, not the mineral solubility. As pointed out by these authors, solubility and specific surface area of Fe(III) oxyhydroxides are usually correlated to one another. For instance, with the data in Table 1 we obtain a linear trend between log *Kso and the logarithm of the specific surface area, with a correlation coefficient r2 = 0.8. Nonetheless, for the data set presented here, the correlation of the maximum, cell-specific iron reducing activity with the mineral solubility is much stronger (r2 = 0.90 on a log–log scale) than with the specific surface area (r2 = 0.54 on a log–log scale). The solubility is thus a better predictor of the potential iron reduction rate than the specific surface area. Reported slopes of LFERs are almost always comprised between zero and one (Leffler and Grunwald, 1989). For simple reaction mechanisms, the slope can be predicted on theoretical grounds. For instance, a series of elementary reactions where the transition state is the same for all the reactions results in a slope of one. This is the model proposed by Rimstidt and Barnes (1980) to

5280


Fig. 4. Linear free energy relationship for Fe(III) oxyhydroxide reduction by S. putrefaciens. The X-axis corresponds to the logarithm of the solubility product defined by Eq. (6) in the text, the Y-axis to the logarithm of the maximum initial reduction rate per cell, vmax. The slope of the linear free relationship is 0.40 ± 0.05 with an intercept value of 16.79 ± 0.05. The vmax and log *Kso values for 6 line ferrihydrite, nanohematite, lepidocrocite and HFO were obtained from Bonneville et al. (2004).

relate the dissolution rate constants and solubility products of silica polymorphs in water. Another example is the Marcus theory for outer sphere electron transfer, which yields a slope between the logarithms of rate constants and equilibrium constants of around 0.5 (see Moore and Pearson (1981), p. 359 for details). The slope of the LFER obtained here (0.40 ± 0.05) falls within the expected range. The value would suggest that the transition state of the rate-limiting step resembles the reactant (i.e. the Fe(III) mineral) more than the product (i.e. dissolved Fe2+). Given the complex nature of the microbial reduction of iron(III) oxyhydroxides, the LFER in Fig. 4 should be primarily considered as an empirical relationship, however. For the experimental conditions under which we carried out the bacteria-mineral incubations, the saturation behavior of the initial microbial reduction rate with respect to the concentration of the Fe(III) substrate (Fig. 3) can be modeled by assuming that the rate is proportional to the amount of cell–mineral contact area (Bonneville et al., 2006). If we further assume that the reaction mechanism involves the formation of a complex between a membranebound Fe(III) reductase and a Fe(III) center at the iron oxide surface, then the reaction steps should include the electron transfer (ET) from the reductase to the Fe(III) center, followed by the dissociation of the enzyme–Fe(II) complex and the release of the Fe2+ ion from the mineral surface, where the last two steps may or may not be taking place simultaneously.

The ET kinetics between outer-membrane cytochromes and hematite have been the subject of a number of recent studies (Neal et al., 2003; Xiong et al., 2006; Kerisit et al., 2007). Kerisit et al. (2007) suggest that the ET may be the rate-limiting step in the microbial reduction of hematite. In this case, a LFER linking the iron reduction rate and the solubility of the Fe(III) oxyhydroxide minerals is not entirely unexpected. The situation is in fact similar to that encountered in many organic reaction series, where an LFER describes the response of the reaction rate to the Gibbs energy of variously substituted reactant molecules. For the microbial iron reduction reaction series, the substituent would be the mineral surface lattice in which the Fe(III) center is embedded. The LFER for the microbial reduction of Fe(III) oxyhydroxides is consistent with similar trends observed for the abiotic dissolution of Fe(III) oxyhydroxides by ascorbate (Postma, 1993; Larsen and Postma, 2001), cysteine (Amirbahman et al., 1997) and sulfide (Dos Santos Afonso and Stumm, 1992; Keresztes et al., 2001; Poulton et al., 2004). In this view, one may hypothesize that the reductive dissolution of Fe(III) phases by siderophore, electron shuttle or exudate potentially produced by dissimilatory iron reducing bacteria may also be influenced by the thermodynamic stability of Fe(III) oxyhydroxides. Indeed, the detachment of the Fe2+ ion following ET has been suggested as the potential rate-limiting step for the abiotic reductive dissolution of Fe(III) oxyhydroxides (Pyzik and Sommer, 1981; Zinder et al., 1986; Hering and Stumm,


1990; Stumm and Wieland, 1990; Suter et al., 1991; Dos Santos Afonso and Stumm, 1992; Roden, 2003; Poulton et al., 2004). Using Fe istopes, Pedersen et al. (2005) further showed that the Fe2+ detachment kinetics from Fe(III) oxyhydroxide surfaces decreases in the order ferrihydrite > lepidocrocite > goethite > hematite, implying that the desorption kinetics are a function of the properties of the mineral substrate. In particular, the observed order corresponds to that of decreasing solubility of the minerals. Thus, whether the ET or the detachment of the Fe2+ ions (in combination with the dissociation of the enzyme– Fe(II) complex, or not) is controlling the initial reduction rate of Fe(III) by S. putrefaciens, both reaction steps are likely to be influenced by the thermodynamic stability of the Fe(III) solid phase. That is, both rate-limiting steps would be expected to yield the observed LFER. 5. CONCLUSIONS The initial reduction rates of a range of Fe(III) oxyhydroxides by the iron reducing bacterium S. putrefaciens are related to the solubilities of the Fe(III) mineral phases via a linear free energy relationship (LFER). The LFER includes all commonly occurring Fe(III) oxyhydroxides, ranging from fairly soluble, poorly crystalline phases, such as ferrihydrite and hydrous ferric oxide, to insoluble, crystalline minerals, such as goethite and hematite. A rate limitation of the microbial reduction process both by the electron transfer to Fe(III) centers at the mineral surface and by the detachment of reduced ferrous iron ions from the mineral surface can explain the observed LFER. The LFER emphasizes the key role of Fe(III) mineral properties in controlling the potential iron reducing activity of microbial communities in sediments and soils. ACKNOWLEDGMENTS This study was part of TRIAS project 835.80.004, co-funded by the Centre for Soil Quality Management and Knowledge Transfer (SKB), Delft Cluster (DC) and the Council for the Earth and Life Sciences (ALW) of the Netherlands Organisation for Scientific Research (NWO). S.B. also beneficiated from fundings from the UK NERC ‘‘Weathering Science Consortium” (NE/C004566/1). The authors thank C. Daughney and two anonymous reviewers for their contructive reviews.

REFERENCES Amirbahman A., Sigg L. and von Gunten U. (1997) Reductive dissolution of Fe(III) (hydr)oxides by cysteine: kinetics and mechanism. J. Colloid Interface Sci. 194, 194–206. Beliaev A. S. and Saffarini D. A. (1998) Shewanella putrefaciens mtrB encodes an outer membrane protein required for Fe(III) and Mn(IV) reduction. J. Bacteriol. 180, 6292–6297. Biedermann G. and Chow J. T. (1966) Hydrolysis of metal ions. LVII. Hydrolysis of the Fe(III) ion and the solubility product of the Fe(OH)2.70Cl0.30 in 0.5M (Na+)Cl medium. Acta Chem. Scand. 20, 1376–1388. Blakeney M. D., Moulaei T. and DiChristina T. J. (2000) Fe(III) reduction activity and cytochrome content of Shewanella

5281

putrefaciens grown on ten compounds as sole terminal electron acceptor. Microbiol. Res. 155, 87–94. Bonneville S., Behrends T., Van Cappellen P., Hyacinthe C. and Roling W. F. M. (2006) Reduction of Fe(III) colloids by Shewanella putrefaciens: a kinetic model. Geochim. Cosmochim. Acta 70, 5842–5854. Bonneville S., Van Cappellen P. and Behrends T. (2004) Microbial reduction of iron(III) oxyhydroxides: effects of mineral solubility and availability. Chem. Geol. 212, 255–268. Byrne R. H. and Luo Y.-R. (2000) Direct observations of nonintegral hydrous ferric oxide solubility products: K so = [Fe3+][H+]2.86. Geochim. Cosmochim. Acta 64, 1873–1877. Byrne R. H., Luo Y.-R. and Young R. W. (2000) Iron hydrolysis and solubility revisited: observations and comments on iron hydrolysis characterizations. Mar. Chem. 70, 23–35. Canfield D. E., Thamdrup B. and Hansen J. W. (1993) The anaerobic degradation of organic matter in Danish coastal sediments: iron reduction, manganese reduction, and sulfate reduction. Geochim. Cosmochim. Acta 57, 3867–3883. DiChristina T. J., Arnold R. G., Lidstrom M. E. and Hoffmann M. R. (1988) Dissimilative iron reduction by the marine eubacterium Alteromonas putrefaciens strain 200. Water Sci. Technol. 20, 69–79. DiChristina T. J., Moore C. M. and Haller C. A. (2002) Dissimilatory Fe(III) and Mn(IV) reduction by Shewanella putrefaciens requires ferE, a homolog of the pulE (gspE) type II protein secretion gene. J. Bacteriol. 184, 142–151. Dos Santos Afonso M. and Stumm W. (1992) Reductive dissolution of Iron(III) (Hydr)oxides by hydrogen sulfide. Langmuir 8, 1671–1675. Dousma J. and De Bruyn P. L. (1978) Hydrolysis–precipitation studies of iron solutions. II. Aging studies and the model for precipitation from iron(III) nitrate solutions. J. Colloid Interface Sci. 64, 154–170. Duckworth O. W., Cygan R. T. and Martin S. T. (2004) Linear free energy relationships between dissolution rates and molecular modeling energies of rhombohedral carbonates. Langmuir 20, 2938–2946. Farrah H. and Pickering W. F. (1978) The effect of pH and ligands on the sorption of heavy metal ions by cellulose. Aust. J. Chem. 31, 1501–1509. Fox L. E. (1988) The solubility of colloidal ferric hydroxide and its relevance to iron concentrations in river water. Geochim. Cosmochim. Acta 52, 771–777. Fukushima M. and Tatsumi K. (1999) Light acceleration of iron(III) reduction by humic acid in the aqueous solution. Colloids and Surf. A. 155, 249–258. Hering J. G. and Stumm W. (1990) Oxidative and reductive dissolution of minerals. Rev. Mineral. 23, 427–465. Hobbie J. E., Daley R. J. and Jasper S. (1977) Use of nuclepore filters for counting bacteria by fluorescence microscopy. Appl. Environ. Microbiol. 33, 1225–1228. Hyacinthe C., Bonneville S. and Van Cappellen P. (2006) Reactive iron(III) in sediments: chemical versus microbial extractions. Geochim. Cosmochim. Acta 70, 4166–4180. Hyacinthe C., Bonneville S. and Van Cappellen P. (2008) Effect of sorbed Fe(II) on initial reduction kinetics of 6-line ferrihydrite and amorphous ferric phosphates by Shewanella putrefaciens. Geomicrobiol. J. 25, 181–192. Jakobsen R. and Postma D. (1999) Redox zoning, rates of sulfate reduction and interactions with Fe-reduction and methanogenesis in a shallow sandy aquifer, Romo, Denmark. Geochim. Cosmochim. Acta 63, 137–151. Keresztes Z., Felhosi I. and Kalman E. (2001) Role of redox properties of biofilms in corrosion processes. Electrochim. Acta 46, 3841–3849.

5282


Kerisit S., Rosso K. M., Dupuis M. and Valiev M. (2007) Molecular computational investigation of electron-transfer kinetics across cytochrome-iron oxide interfaces. J. Phys. Chem. C 111, 11363–11375. Kuma K., Nakabayashi S., Suzuki Y. and Matsunaga K. (1992) Dissolution rate and solubility of colloidal hydrous ferric oxide in seawater. Mar. Chem. 38, 133–143. Kuma K., Suzuki Y. and Matsunaga K. (1993) Solubility and dissolution rate of colloidal c-iron oxyhydroxide in seawater. Water Res. 27, 651–657. Larsen O. and Postma D. (2001) Kinetics of reductive bulk dissolution of lepidocrocite, ferrihydrite and goethite. Geochim. Cosmochim. Acta 65, 1367–1379. Leffler J. E. and Grunwald E. (1989) Rates and Equilibria of Organic Reactions. Courier Dover Publications. Lin B., Hyacinthe C., Bonneville S., Braster M., Van Cappellen P. and Roling W. F. M. (2007) Phylogenetic and physiological diversity of dissimilatory ferric iron reducers in sediments of the polluted Scheldt estuary, Northwest Europe. Environ. Microbiol. 9, 1956–1968. Lovley D. R., Phillips E. J. P. and Lonergan D. J. (1991) Enzymic versus nonenzymic mechanisms for iron(III) reduction in aquatic sediments. Environ. Sci. Technol. 25, 1062–1067. Lyngkilde J. and Christensen T. H. (1992) Redox zones of a landfill leachate pollution plume (Vejen, Denmark). J. Contam. Hydrol. 10, 273–289. Moore J. W. and Pearson R. G. (1981) Kinetics and Mechanism. John Wiley and Sons. Munch J. C. and Ottow J. C. G. (1983) Reductive transformation mechanism of ferric oxides in hydromorphic soils. Environ. Biogeochem. Ecol. Bull. 35, 383–394. Murphy P. J., Posner A. M. and Quirk J. P. (1976) Characterization of hydrolyzed ferric ion solutions. A comparison of the effects of various anions on the solutions. J. Colloid Interface Sci. 56, 312–319. Myers C. R. and Myers J. M. (1997) Outer membrane cytochromes of Shewanella putrefaciens MR-1: spectral analysis, and purification of the 83-kDa c-type cytochrome. Biochim. Biophys. Acta 1326, 307–318. Neal A. L., Rosso K. M., Geesey G. G., Gorby Y. A. and Little B. J. (2003) Surface structure effects on direct reduction of iron oxides by Shewanella oneidensis. Geochim. Cosmochim. Acta 67, 4489–4503. Ona-Nguema G., Abdelmoula M., Jorand F., Benali O., Gehin A., Block J.-C. and Genin J.-M. R. (2002) Iron(II, III) hydroxycarbonate green rust formation and stabilization from lepidocrocite bioreduction. Environ. Sci. Technol. 36, 16–20. Pedersen H. D., Postma D., Jakobsen R. and Larsen O. (2005) Fast transformation of iron oxyhydroxides by the catalytic action of aqueous Fe(II). Geochim. Cosmochim. Acta 69, 3967–3977. Postma D. (1993) The reactivity of iron oxides in sediments: a kinetic approach. Geochim. Cosmochim. Acta 57, 5027–5034. Poulton S. W., Krom M. D. and Raiswell R. (2004) A revised scheme for the reactivity of iron (oxyhydr)oxide minerals towards dissolved sulfide. Geochim. Cosmochim. Acta 68, 3703– 3715. Pyzik A. J. and Sommer S. E. (1981) Sedimentary iron monosulfides: kinetics and mechanism of formation. Geochim. Cosmochim. Acta 45, 687–698.

Rimstidt J. D. and Barnes H. L. (1980) The kinetics of silica–water reactions. Geochim. Cosmochim. Acta 44, 1683–1699. Roden E. E. (2003) Fe(III) oxide reactivity toward biological versus chemical reduction. Environ. Sci. Technol. 37, 1319– 1324. Roden E. E. and Zachara J. M. (1996) Microbial reduction of crystalline iron(III) oxides: influence of oxide surface area and potential for cell growth. Environ. Sci. Technol. 30, 1618–1628. Schwertmann U. and Cornell R. M. (1991) Iron Oxides in the Laboratory: Preparation and Characterization. Weinheim. Spiro T. G., Allerton S. E., Renner J., Terzis A., Bils R. and Saltman P. (1966) The hydrolytic polymerization of iron(III). J. Am. Chem. Soc. 88, 2721–2726. Steefel C. I. and Van Cappellen P. (1990) A new kinetic approach to modeling water–rock interaction: the role of nucleation, precursors, and Ostwald ripening. Geochim. Cosmochim. Acta 54, 2657–2677. Stumm W. and Wieland E. (1990) Dissolution of oxide and silicate minerals: rates depend on surface speciation. In Aquatic Chemical Kinetics: Reaction Rates of Processes in Natural Waters (ed. W. Stumm). John Wiley and Sons, Inc., New York, NY, pp. 367–400. Suter D., Banwart S. and Stumm W. (1991) Dissolution of hydrous iron(III) oxides by reductive mechanisms. Langmuir 7, 809–813. Van Breemen N. (1988) Effects of seasonal redox processes involving iron on the chemistry of periodically reduced soils. NATO ASI Ser. Ser. C 217, 797–809. van Schaik J. W. J., Persson I., Berggren Kleja D. and Gustafsson J. P. (2008) EXAFS study on the reactions between iron and fulvic acid in acid aqueous solutions. Environ. Sci. Technol. 42, 2367–2373. Viollier E., Inglett P. W., Hunter K., Roychoudhury A. N. and Van Cappellen P. (2000) The ferrozine method revisited: Fe(II)/ Fe(III) determination in natural waters. Appl. Geochem. 15, 785–790. Weber T., Allard T., Tipping E. and Benedetti M. F. (2006) Modeling iron binding to organic matter. Environ. Sci. Technol. 40, 7488–7493. Weltje L., den Hollander W. and Wolterbeek H. T. (2003) Adsorption of metals to membrane filters in view of their speciation in nutrient solution. Environ. Toxicol. Chem. 22, 265–271. Wieland E., Wehrli B. and Stumm W. (1988) The coordination chemistry of weathering: III. A generalization on the dissolution rates of minerals. Geochim. Cosmochim. Acta 52, 1969– 1981. Wollast R. (1974) Silica problem. In The Sea (eds. A. E. Maxwell, E. D. Goldberg and M. N. Hill). Wiley, pp. 5359–5392. Xiong Y., Shi L., Chen B., Mayer M. U., Lower B. H., Londer Y., Bose S., Hochella M. F., Fredrickson J. K. and Squier T. C. (2006) High-affinity binding and direct electron transfer to solid metals by the Shewanella oneidensis MR-1 outer membrane ctype cytochrome OmcA. J. Am. Chem. Soc. 128, 13978–13979. Zinder B., Furrer G. and Stumm W. (1986) The coordination chemistry of weathering: II. Dissolution of iron(III) oxides. Geochim. Cosmochim. Acta 50, 1861–1869. Associate editor: Christopher John Daughney