Semiconducting hematite facilitates microbial and abiotic reduction of chromium

Semi-conducting Fe oxide minerals, such as hematite, are well known to influence the fate of contaminants and nutrients in many environmental settings through sorption and release of Fe(II) resulting from microbial or abiotic reduction. Studies of Fe oxide reduction by adsorbed Fe(II) have demonstrated that reduction of Fe(III) at one mineral surface can result in the release of Fe(II) on a different one. This process is termed “Fe(II) catalyzed recrystallization” and is believed to be the result of electron transfer through semi-conducting Fe (hydr)oxides. While it is well understood that Fe(II) plays a central role in redox cycling of elements, the environmental implications of Fe(II) catalyzed recrystallization require further exploration. Here, we demonstrate that hematite links physically separated redox reactions by conducting the electrons involved in those reactions. This is shown using an electrochemical setup where Cr reduction is coupled with a potentiostat or Shewanella putrefaciens, a metal reducing microbe, where electrons donated to hematite produce Fe(II) that ultimately reduces Cr. This work demonstrates that mineral semi-conductivity may provide an additional avenue for redox chemistry to occur in natural soils and sediments, because these minerals can link redox active reactants that could not otherwise react due to physical separation.

in samples were measured by ICP-MS using a Sc or Ge internal standard, and samples were diluted by 10 times with 2% HNO3 to reduce wear on the detector. The choice of internal standard depended on which was most stable over a given run. Triplicate samples were run on the ICP-MS, and sample concentrations did not vary more than 10%. A calibration curve was prepared by dilutions of a multi-element standard (Perkin-Elmer), and ranged from 0.02 µM to 2.0 µM Cr. Fe(II) and total Fe were measured by applying the ferrozine method to the diluted samples with concentration ranges from 18 µM to 180 µM Fe. 1,2 The total solution in experimental chambers was tracked by measurements in changes of total chamber mass, which was then used to calculate amounts of Fe and Cr in the experimental chambers based on concentration and chamber volume.
Additional results and discussion

Measurements of Hematite Resistivity
Hematite resistivity was measured by coating two sides of a piece of specular hematite with silver epoxy and gold wire, and then measuring the resistance of the piece using a multimeter.
Hematite samples were of a similar size and shape as the hematite pieces used as electrodes, but were not the samples used to fabricate the electrodes. Further measurements of hematite resistivity as a function of applied voltage were also taken. This method was applied because hematite is a well-established n-type semi-conductor, which will have shifts in resistivity depending on the applied voltage. Increasing applied potentials should therefore result in band bending that enhances conduction of electrons. 3,4 The results of these measurements are given in figure S1 for three hematite samples that were of similar size and shape to those used as electrodes. The sample resistance was calculated by first calculating the circuit current and hematite potential drop by the voltage drop over a measurement resistor and thereby the hematite sample resistance. This, with the sample dimensions were then used to estimate the hematite resistivity. The dimensions used to calculate the resistivity assume an idealized rectangular prism geometry, which was reasonably accurate for the samples used. Sample areas did not vary more than 10%, though the sample thickness did vary up to 50% in one sample. Resistivity of the hematite varied both by the voltage applied to the measurement circuit, as well as according to the sample measured. In one sample, the resistivity changed by nearly an order of magnitude in response to an increasing applied voltage, while for another, little change in resistivity was observed. Hematite orientation in the circuit (both the nearest terminal of the voltage source, and the sample orientation in the circuit) were changed and while some small differences were observed, did not appear to impact the resistivity measured. The measured resistivity by a 2-point resistance measurement across the samples using a multimeter was also a similar value to those produced by resistivity measurements with around 0.5V of applied voltage. The measurements of resistivity highlight the large amount of heterogeneity intrinsic to the specular hematite, which is linked to observed variations in the amount of electrons transferred in each experiment. The measured resistivity of specular hematite samples ranged from 500Ωm to nearly 11000Ωm, which falls within the large range of resistivities reported for natural hematite, which cover from 1 × 10 2 Ωm to 1 × 10 7 Ωm for bulk hematite. 5 Another electrochemical study that used natural hematite as an electrode, however, reported resistivities of only 0.5Ωm to 10Ωm. 4,6 Surface analysis also helps explain these differences. The electron backscatter images and EDS spectra (   Figure S1 shows that depending on the hematite sample measured, that difference could also result in an increase of hematite resistance by a factor 2, which would then further decrease current that bacteria could generate. This resistivity increase matches with what would be expected for a semiconducting material, as smaller applied negative potentials result in less charge accumulation in the conduction layer of the semiconductor that in turns allows for less current. 3,4 Cr control experiments Experiments were performed to analyze potential sources of Cr removal from solution when no current was being provided to the hematite electrode. Two processes were considered: sorption of Cr to chamber surfaces in the absence of an electrical connection between a hematite and carbon electrode, and loss of Cr resulting from any background current that might develop.
Sorption losses were tested in an experimental chamber set up for a biotic experiment, but no electrical connection between the carbon and hematite electrodes was made. Thus, any changes in Cr concentration had to result from sorption to experimental surfaces. To test for the presence of background current, a biotic experiment was performed with an electrical connection between the carbon and hematite electrodes, but the spike of metal reducing bacteria was omitted. This experiment was performed using a chamber that had been used for a sorption experiment previously to reduce any possible sorption that might affect the results. The results of all 4 experiments are plotted in figure S3.
Three experiments to test sorption were performed: The first was performed at a significantly higher Cr concentration (~45 umol total Cr at the start, labeled Sorption A in figure   S3) and 125 hours. This The results of the abiotic control experiment showed up to 5 µmol decrease in Cr from the start to the finish of the experiment. Because this was performed in a chamber that had previously been used in a sorption experiment (specifically, experiment Sorption A), sorption likely does not play a role in this observed decrease. Similarly, because no current was observed, it is not likely that Cr reduction is occurring as a result of electrons delivered from the carbon electrode. The total loss (5 µmol) is within 10% of the total, which is within the uncertainty of the ICP-MS measurement. Another possibility is that exposure to light is leading to Cr reduction, via a photoreduction mechanism, as hematite has some photoreduction capacity. 10,11 In those cases, an external electron source was needed to drive the first step of reduction (i.e. reduction of Cr(VI) to Cr(V)), and the remaining reduction to Cr(III) proceeds with the assistance of light. Since there was no current source recorded in this work, this mechanism is likely not driving Cr reduction.
Similarly, photoreduction is unlikely due to the relatively low purity and thickness of the hematite electrodes used. Since this was not controlled for explicitly, however, the possibility remains.
While it is not clear the exact source of the decrease, the amount is small enough (10% of initial) that even if this separate reduction mechanism was operating in this experimental system, the indirect reduction by microbes or the potentiostat would still occur.

Current measurements
Plots of the measured current are given in figure S4 for      Table S1: Elemental abundances on the unreacted thin electrode determined by XPS.
Abundances are relative and determined by fitting peaks in observed spectra. Blank values mean that no peak was fitted, and the element was not detected.