MINERALOGICAL AND GEOCHEMICAL ANALYSES OF SYNTHETIC AND NATURAL BIRNESSITES

The Pennsylvania State University The Graduate School Department of Geosciences

MINERALOGICAL AND GEOCHEMICAL ANALYSES OF SYNTHETIC AND NATURAL BIRNESSITES

A Dissertation in Geosciences by Florence T. Ling

 2016 Florence T. Ling

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

December 2016

ii

The dissertation of Florence T. Ling was reviewed and approved* by the following:

Peter J. Heaney Professor of Mineral Sciences Disseration Advisor Chair of Committee

James D. Kubicki Professor of Geosciences

Christopher H. House Professor of Geosciences

William D. Burgos Professor of Environmental Engineering

Demian Saffer Professor of Geosciences Interim Associate Head of Graduate Programs and Research

*Signatures are on file in the Graduate School

iii

ABSTRACT Manganese (Mn) oxides have long been challenging to study using conventional materials characterization techniques due to their small particle size, poor crystallinity and variable structure at the nanoscale. The phyllomanganates within the birnessite family, for example, exhibit a diversity of Mn oxidation states, interlayer cations, water contents, and octahedral vacancy concentrations. These differences lead to subtle structural modifications that control the redox and cation exchange capacities of birnessite phases in soils, and they determine the usefulness of birnessites for environmental applications, such as the remediation of contaminant metals. Although powder X-ray diffraction (XRD) can offer deep insights into birnessite crystallography, spectroscopic techniques provide complementary information when structural disorder is high. This study introduces Fourier-transform infrared spectroscopy (FTIR) as a method for analyzing birnessite varieties, coupling our analyses with synchrotron X-ray diffraction and absorption spectroscopy for comparison. We found that FTIR can readily differentiate synthetic triclinic Na-birnessite and hexagonal H-birnessite. The region from 400 to 750 cm-1, which is most sensitive to Mn-O bond vibrations, displays the most prominent distinctions between the two varieties, with peaks at ~418, ~478, and ~511 cm-1 in synthetic triclinic Na-birnessite, and peaks at ~440 and ~494 cm-1 in hexagonal H-birnessite. Spectral unmixing of known mixtures of triclinic and hexagonal birnessite yielded errors comparable to results from linear combination fitting (LCF) of extended X-ray absorption fine-structure (EXAFS) data. Spectral unmixing of synthetic cation-exchanged birnessites, including K-birnessite, Ba-birnessite, Ca-birnessite, and two hexagonal birnessite samples prepared in pH 3 and HEPES-buffered pH 7 solutions, also yielded fractions of triclinic and hexagonal birnessite comparable to those calculated by LCF of EXAFS data. Our EXAFS, FTIR, and Rietveld refinements of XRD data all confirmed that as

iv

Mn3+ content increases, the triclinic character of the birnessite also increases, as manifested by an increase in the β angle and in the length of the a-axis of the unit cell. With these synthetic samples providing a baseline for comparison, natural birnessite samples next were interrogated using the same suite of techniques. Contrary to the common assumption that most natural birnessites are hexagonal, our FTIR and EXAFS investigations revealed natural birnessite varieties as either triclinic or hexagonal or, most commonly, a mixture of the two. Further, this work demonstrates that the use of biological buffers such as HEPES and MES can promote the transformation of synthetic triclinic Na-birnessite into hexagonal Hbirnessite, perhaps accounting for the presumption of hexagonal symmetry in natural biogenic samples. When our EXAFS and FTIR analyses were combined with X-ray photoelectron spectroscopy (XPS) data in an allied study (Ilton et al. 2016), a relation between Mn3+ content in birnessite and deviations from hexagonal symmetry appeared, suggesting that when Mn3+ exceeds ~25 mol% in birnessite, Jahn-Teller distortions couple, and the resultant strains generate structures with triclinic symmetry. In order to connect birnessite crystal chemistry with reactivity, we examined dissolved Pb uptake into synthetic triclinic Na- and hexagonal H-birnessite at pH 3 and pH 5 using ~6 hr timeresolved XRD and 112-day batch experiments. We found that Pb sorbs onto the surface of birnessite and into the interlayer, with uptake occurring in the interlayer even after 14 days. In hexagonal H-birnessite the ratio of the (1 0 0) to (1 0 -1) peaks increased as Pb exchanged into the interlayer, as expected from our diffraction simulations. Triclinic Na-birnessite transformed into a turbostratically disordered hexagonal H-birnessite during uptake at pH 3 and pH 5 due to the relative instability of triclinic birnessite at pH < 8.2. As a consequence of this transformation, both triclinic Na-birnessite and hexagonal H-birnessite sequestered similar amounts of Pb after 112 days.

v

TABLE OF CONTENTS List of Figures .......................................................................................................................... viii List of Tables ........................................................................................................................... xiv Acknowledgements .................................................................................................................. xvi Chapter 1 Introduction ............................................................................................................. 1 References Cited .............................................................................................................. 7 Chapter 2 Transformations from triclinic to hexagonal birnessite at circumneutral pH induced through pH control by common biological buffers ............................................ 10 Abstract ............................................................................................................................ 10 Introduction ...................................................................................................................... 11 Methods............................................................................................................................ 14 Materials synthesis ................................................................................................... 14 Batch experiments involving triclinic Na-birnessite and buffers ............................. 15 pH-monitored batch experiments with birnessites and unbuffered solutions........... 16 Synchrotron X-ray diffraction .................................................................................. 17 Structure refinement ................................................................................................. 17 Energy-dispersive x-ray spectroscopy (EDS) .......................................................... 19 Results .............................................................................................................................. 19 Reactions of triclinic birnessite in solutions buffered to circumneutral pH ............. 19 Reaction of birnessites in unbuffered solutions ....................................................... 21 Discussion ........................................................................................................................ 22 Cause of buffer-induced transformations in birnessite............................................. 22 Cause of birnessite transformation in unbuffered solutions ..................................... 25 Implications of reversibility of pH change ............................................................... 26 Hydrolysis vs. water oxidation ................................................................................. 28 Conclusions ...................................................................................................................... 29 Acknowledgements .......................................................................................................... 29 References ........................................................................................................................ 30 Figures.............................................................................................................................. 35 Tables ............................................................................................................................... 42 Chapter 3 Sorption of contaminant Pb by triclinic and hexagonal birnessite .......................... 45 Abstract ............................................................................................................................ 45 Introduction ...................................................................................................................... 46 Experimental Section ....................................................................................................... 49 Materials synthesis ................................................................................................... 49 Time-resolved x-ray diffraction experiments ........................................................... 50 Batch experiments .................................................................................................... 50 Structure Refinements .............................................................................................. 52

vi

X-ray diffraction simulations ................................................................................... 53 Results and Discussion..................................................................................................... 53 Intercalation of Pb in triclinic birnessite .................................................................. 53 Intercalation of Pb in hexagonal birnessite .............................................................. 55 Mechnisms of Pb Sorption for Triclinic and Hexagonal Birnessites ....................... 57 Acknowledgements .......................................................................................................... 59 References ........................................................................................................................ 60 Figures.............................................................................................................................. 63 Tables ............................................................................................................................... 77 Chapter 4 Fourier-transform infrared spectroscopy (FTIR) analysis of triclinic and hexagonal birnessites ....................................................................................................... 81 Abstract ............................................................................................................................ 81 Introduction ...................................................................................................................... 81 Methods............................................................................................................................ 84 Materials synthesis ................................................................................................... 84 X-ray diffraction (XRD) and structure refinements ................................................. 85 X-ray absorption spectroscopy/extended X-ray absorption fine structure (XAS/EXAFS).................................................................................................. 87 Fourier-transform infrared spectroscopy (FTIR) ..................................................... 88 Results & Discussion ....................................................................................................... 91 Distinguishing triclinic and hexagonal birnessite with XRD and FTIR ................... 91 Quantitative analysis of known triclinic-hexagonal birnessite mixtures: FTIR, EXAFS, and XRD ............................................................................................ 92 Comparison of observed FTIR spectra to DFT calculations .................................... 93 Conclusions ...................................................................................................................... 97 Acknowledgements .......................................................................................................... 98 References ........................................................................................................................ 99 Figures.............................................................................................................................. 104 Tables ............................................................................................................................... 115 Supplementary Information A.......................................................................................... 120 Chapter 5 The relationship between Mn oxidation state and structure in triclinic and hexagonal birnessites ....................................................................................................... 121 Abstract ............................................................................................................................ 121 Introduction ...................................................................................................................... 121 Methods............................................................................................................................ 124 Materials synthesis ................................................................................................... 124 X-ray diffraction (XRD) and structure refinements ................................................. 126 X-ray absorption spectroscopy/extended X-ray absorption fine structure (XAS/EXAFS).................................................................................................. 127 Fourier transform-infrared spectroscopy (FTIR) ..................................................... 128 X-ray photoelectron spectroscopy (XPS) ................................................................. 128 Results & Discussion ....................................................................................................... 129 Qualitative comparison of phases using XRD, FTIR, and EXAFS ......................... 129 Effects of Mn3+ on birnessite structure as revealed by Rietveld analysis................. 132 Trends from quantitative EXAFS and IR analysis with Mn3+.................................. 134

vii

Distorted triclinic birnessite structures vs. intergrown triclinic-hexagonal birnessite layers ................................................................................................ 135 Conclusions ...................................................................................................................... 136 Acknowledgements .......................................................................................................... 137 References ........................................................................................................................ 138 Figures.............................................................................................................................. 143 Tables ............................................................................................................................... 155 Supplementary Information B .......................................................................................... 162 Chapter 6 A characterization of natural terrestrial birnessites ................................................. 163 Abstract ............................................................................................................................ 163 Introduction ...................................................................................................................... 164 Methods............................................................................................................................ 167 Samples .................................................................................................................... 167 X-ray diffraction (XRD)........................................................................................... 169 Scanning electron microscopy/energy dispersive spectroscopy (SEM/EDS) .......... 169 Electron probe microanalysis (EPMA) .................................................................... 170 Fourier transform infrared spectroscopy (FTIR) ...................................................... 171 X-ray absorption spectroscopy/extended X-ray absorption fine structure (XAS/EXAFS).................................................................................................. 171 X-ray photoelectron spectroscopy (XPS) ................................................................. 172 Results and Discussion..................................................................................................... 173 X-ray diffraction of natural birnessites .................................................................... 173 EXAFS provides additional phase information for samples unidentifiable by XRD ................................................................................................................. 175 Implications of the mismatch between XRD analyses and LCF analyses of EXAFS spectra ................................................................................................. 177 Fourier-transform infrared spectroscopy (FTIR) of natural 3-line birnessites ......... 179 X-ray photoelectron spectroscopy (XPS) of synthetic birnessites and natural birnessites ......................................................................................................... 182 Structural differences in birnessite samples with similar formation environments .................................................................................................... 183 Conclusions ...................................................................................................................... 184 Acknowledgments ............................................................................................................ 185 References ........................................................................................................................ 186 Figures.............................................................................................................................. 192 Tables ............................................................................................................................... 210 Supplementary Information C .......................................................................................... 215 Appendix XPS Determination of Mn Oxidation States in Mn (Hydr)oxides ......................... 219

viii

LIST OF FIGURES Figure 2-1. Schematic diagram of (a) triclinic Na-birnessite and (b) hexagonal Hbirnessite (after Lanson et al. 2000). ................................................................................ 35 Figure 2-2. Rietveld refinement fits and difference curves for (a) initial triclinic Nabirnessite, (b) hexagonal birnessite formed from triclinic Na-birnessite in a pH 7 HEPES-buffered solution after 2 weeks, and (c) mixed triclinic and hexagonal birnessite formed from triclinic Na-birnessite in a pH 7 MES-buffered solution after 2 weeks. Crosses represent observed intensities; the solid green lines are the calculated patterns. The solid red lines are the calculated backgrounds. The solid aqua lines represent the difference between the observed and calculated patterns. Only data between the dashed vertical lines were used for refinements. ......................... 36 Figure. 2-3. X-ray diffraction patterns of (a) the initial triclinic Na-birnessite powder, (b) mixed hexagonal and triclinic birnessite after reacting in a pH 7 solution for 14 days, (c) the mixed hexagonal and triclinic birnessite after reacting in a pH 7 MES solution for 14 days, (d) hexagonal birnessite after reacting in a pH 7 HEPES solution for 14 days, and (e) hexagonal H-birnessite synthesized by placing triclinic Na-birnessite in pH 3 solution for 24 hrs. ........................................................................ 37 Figure 2-4. EDS spectra of (a) initial triclinic Na-birnessite, (b) birnessite in the pH 7 unbuffered solution, (c) birnessite in the pH 7 MES-buffered solution, and (d) birnessite in the pH 7 HEPES-buffered solution.............................................................. 38 Figure 2-5. X-ray diffraction patterns of (a) the initial triclinic Na-birnessite powder, (b) mixed triclinic and hexagonal birnessite after reacting in a pH 7 MES solution for 12 hrs, (c) mixed hexagonal and triclinic birnessite after reacting in a pH 7 MES solution for 3 days, (d) mixed triclinic and hexagonal birnessite after reacting in a pH 7 MES solution for 14 days, (e) hexagonal birnessite after reacting in a pH 6.30 MES solution for 16 hrs, and (f) hexagonal H-birnessite. ............................................... 39 Figure 2-6. The change in pH over time for batch experiments with triclinic Na-birnessite and pH 7 solution, and another with hexagonal H-birnessite in pH 11 solution. ............. 40 Figure 2-7. X-ray diffraction patterns of (a) the initial triclinic Na-birnessite powder, (b) mixed hexagonal and triclinic birnessite after reacting in a pH 7 solution for 22 days, (c) the initial hexagonal H-birnessite powder, and (d) the hexagonal H-birnessite after reacting in a pH 11 solution for 22 days. ................................................................. 41 Figure 3-1. Schematic diagrams of synthetic (a) triclinic Na-birnessite (after Post et al. 2002) and hexagonal H-birnessite (after Lanson et al. 2000). ......................................... 63 Figure 3-2. Experimental set-up for time-resolved X-ray diffraction flow-through experiments. ..................................................................................................................... 64 Figure 3-3. Rietveld refinement fits from the batch experiments of (a) the initial hexagonal H-birnessite, (b) the hexagonal H-birnessite after 14 days of reacting with

ix

pH 3 0.1 M Pb(NO3)2, and (c) the reacted hexagonal H-birnessite after 114 days of reacting with pH 3 0.1 M Pb(NO3)2. ................................................................................ 65 Figure 3-4. Triclinic Na-birnessite was reacted with a pH 3 HCl solution, and TRXRD patterns were collected every 30 s for 3 hrs. It transformed into hexagonal birnessite within ~30 min. ................................................................................................................ 66 Figure 3-5. Triclinic Na-birnessite was reacted with a 0.1 M Pb(NO3)2 solution at pH 3, and TRXRD patterns were collected every 2 min for ~6.5 hrs. Within 1.25 hrs, it transformed into a turbostratically disordered hexagonal-like phase............................... 67 Figure 3-6. Overlay of XRD patterns for the 7, 14, and 112-day batch experiments with triclinic Na-birnessite in pH 3 0.1 M Pb(NO3)2................................................................ 68 Figure 3-7. XRD patterns of (a) the initial triclinic Na-birnessite, (b) triclinic Nabirnessite (NaB) after reacting in a batch experiment in pH 5 HNO3, Pb-free solution for 5.5 hrs (c) triclinic Na-birnessite after reacting in a batch experiment with 0.1 M Pb(NO3)2 at pH 5 for 112 days, and (d) triclinic Na-birnessite reacting with 0.1 M Pb(NO3)2 at pH 3 in a 6-hr flow-through experiment for comparison. ............................ 69 Figure 3-8. Concentrations for Pb, Mn, and Na were measured over time for the 112-day batch experiments with triclinic Na-birnessite and 0.1 M Pb(NO3)2 at pH 3 and 5. (a) The moles of Pb/g birnessite uptaken was calculated by subtracting the measured Pb concentration from the initial Pb concentration/g birnessite. (b) The mmol of Mn/g birnessite and (c) the mmol of Na/g birnessite released into solution were directly measured. Note that the data point for Na-birnessite reacting with a pH 3 0.1 M Pb(NO3)2 solution at 56 days was deleted in all Pb, Mn, and Na plots due to their persistence as outliers. ............................................................................................. 70 Figure 3-9. TR-XRD results for reacting hexagonal H-birnessite with pH 3 0.1 M Pb(NO3)2. Patterns were collected every 30 s for ~2 hrs. Little change occurred in the hexagonal birnessite. .................................................................................................. 71 Figure 3-10. Overlay of XRD patterns for the 7, 14, and 112-day batch experiments with hexagonal H-birnessite in pH 3 0.1 M Pb(NO3)2. ............................................................ 72 Figure 3-11. XRD results from batch reactions of hexagonal H-birnessite (HB) at pH 5, including the (a) initial hexagonal H-birnessite, (b) hexagonal H-birnessite reacting in a Pb-free pH 5 HNO3 solution for 5.5 hrs, and (c) hexagonal H-birnessite reacting in a 112-day batch reaction with pH 5 0.1 M Pb(NO3)2. .................................................. 73 Figure 3-12. Simulations of XRD patterns on CrystalDiffract of two hexagonal Hbirnessites with different fractional occupancies of 0.40 and 0.18 for interlayer Mn. The XRD simulation with a high Mn(int) occupancy of 0.40 has a higher (1 0 0) to (1 0 -1) peak intensity ratio compared to the simulation with a low Mn(int) occupancy of 0.18. ........................................................................................................... 74 Figure 3-13. Rietveld refinement results from the initial, 7, 14, and 112-day batch experiments using Mn as the interlayer cation, and then Pb as the interlayer cation for the (a) fractional occupancy for interlayer Mn or Pb position, (b) fractional

x

occupancy for interlayer O, representing interlayer H2O, (c) lattice parameter a, (d) lattice parameter c, and (e) volume. ................................................................................. 75 Figure 3-14. Concentrations for Pb, Mn, and Na were measured over time for the 112-day batch experiments with hexagonal H-birnessite and 0.1 M Pb(NO3)2 at pH 3 and 5. (a) The moles of Pb/g birnessite uptaken was calculated by subtracting the measured Pb concentration from the initial Pb concentration/g birnessite. (b) The mmol of Mn/g birnessite and (c) the mmol of Na/g birnessite released into solution were directly measured. ............................................................................................................ 76 Figure 4-1. Schematic diagrams of (a) synthetic triclinic Na-birnessite and (b) synthetic hexagonal H-birnessite after Lanson et al. (2000). .......................................................... 104 Figure 4-2. Geometry optimized models of nanoclusters representing (a) triclinic birnessite Mn7O24H22, ( b) hexagonal birnessite without a vacancy, Mn7O24H23, and (c) hexagonal birnessite with a vacancy, Mn7O26H26. Figures were drawn in Materials Studio (Accelerys Inc., San Diego CA). .......................................................... 105 Figure 4-3. Rietveld refinement fits for (a) synthetic triclinic Na-birnessite and (b) synthetic hexagonal H-birnessite. .................................................................................... 106 Figure 4-4. (a) Full IR spectra of synthetic triclinic Na-birnessite and synthetic hexagonal H-birnessite, (b) enlarged portion of the IR spectra over the OH-vacancy vibration range, and (c) enlarged portion of the IR spectra for the Mn-O bond vibration range. ... 107 Figure 4-5. FTIR spectra of pure triclinic and hexagonal birnessite and the first set of mechanical mixtures with weighted ratios of 25:75, 50:50, and 75:25 triclinichexagonal birnessite. ........................................................................................................ 108 Figure 4-6. Rietveld refinement fits for (a) 25:75, (b) 50:50, and (c) 75:25 triclinic-tohexagonal birnessite mixtures. ......................................................................................... 109 Figure 4-7. Comparison for volume fractions of mechanical mixtures of triclinic and hexagonal birnessite from (a) FTIR compared with XRD volume fractions, (b) FTIR compared with EXAFS volume fractions, and (c) EXAFS compared with XRD volume fractions. The dashed lines represent the 1:1 correlation line. ........................... 110 Figure 4-8. (a) Full calculated IR spectra for nanoclusters representing triclinic birnessite, hexagonal birnessite with no vacancy site, and hexagonal birnessite with a vacancy site, along with collected FTIR spectra for endmember triclinic Na-birnessite and hexagonal H-birnessite with intensities multiplied by a scale factor of 3000 for comparison. (b) The data in the range from 1300 to 1800 cm-1 are enlarged to highlight differences resulting from varying vacancy concentration in the birnessite structure. (c) The data in the range of 400 to 1000 are enlarged for easier comparison of peaks related to Mn-O bonding in the DFT calculations and observed FTIR spectra..................................................................................................................... 111 Figure 4-9. Exemplary vibrations in a birnessite nanocluster for (a) inner Mn-O lattice vibrations leading to vibrations in the 400 to 700 cm-1 range, (b) peripheral Mn-O vibrations leading to vibrations in the 700 to 1400 cm-1 range, (c) H2O scissoring

xi

vibrations leading to peaks in the 1400 to 1800 cm-1 range, and (d) OH stretches leading to peaks in the 1800 to 4000 cm-1 range. The larger, thick arrows represent the main vibrations, while smaller arrows represent resulting motions due to the close vicinity or connectivity in the nanocluster to the main vibrations occurring simultaneously. ................................................................................................................ 112 Figure 4-10. All vibrations for hexagonal birnessite nanocluster with a vacancy site in the 1300 to 1800 cm-1 range, including the OH vibrations at (a) 1653, (b) 1599, (c) 1569, and (d) 1498 cm-1. .................................................................................................. 113 Figure 4-11. Gaussian-Lorentzian peak shapes for peaks in the OH vacancy-related range from 1300 to 1800 cm-1 for (a) hexagonal H-birnessite and (b) triclinic Nabirnessite. ......................................................................................................................... 114 SI-A Figure 1. Exemplary movements for vibrational modes a.) Mn-Obr, b.) Mn3-Obr, c.) Mn-Obr asymm, d.) Mn-Obr symm, and e.) defect site. .................................................... 120 Figure 5-1. Schematic diagram of (a) synthetic triclinic Na-birnessite and (b) synthetic hexagonal H-birnessite after Lanson et al. (2000). .......................................................... 143 Figure 5-2. XRD patterns of synthetic birnessites. .................................................................. 144 Figure 5-3. FTIR spectra of (a) hexagonal birnessites and (b) triclinic birnessites in the Mn-O lattice range from 400 to 750 cm-1. ....................................................................... 145 Figure 5-4. EXAFS k2 χ(k) data for (a) standard triclinic Na-birnessite, (b) standard hexagonal pH 2 birnessite, (c) hexagonal HEPES birnessite, and (d) hexagonal pH 3 birnessite. ......................................................................................................................... 146 Figure 5-5. Radial distribution functions for (a) standard triclinic Na-birnessite, (b) standard hexagonal pH 2 birnessite, (c) hexagonal HEPES birnessite, and (d) hexagonal pH 3 birnessite. ............................................................................................... 147 Figure 5-6. EXAFS k2 χ(k) data for (a) standard triclinic Na-birnessite, (b) standard hexagonal pH 2 birnessite, (c) K-birnessite, (d) Ca-birnessite, and (e) Ba-birnessite. .... 148 Figure 5-7. Radial distribution functions for (a) standard triclinic Na-birnessite, (b) standard hexagonal pH 2 birnessite, (c) K-birnessite, (d) Ca-birnessite, and (e) Babirnessite. ......................................................................................................................... 149 Figure 5-8. XPS results showing fractions of Mn4+, Mn3+, and Mn2+ for each synthetic sample from fitting the Mn3p peak, along with an average Mn average oxidation state. ................................................................................................................................. 150 Figure 5-9. Representative Rietveld refinements and difference curves for (a) Cabirnessite and (b) pH 3 hexagonal H-birnessite. .............................................................. 151 Figure 5-10. Correlations between Mn3+ and the refined parameters: (a) β angle, (b) the a parameter, (c) volume, and (d) the a:c ratio..................................................................... 152

xii

Figure 5-11. Comparison of the fitted fraction of triclinic Na-birnessite for EXAFS data vs. FTIR data using endmember triclinic Na-birnessite and end-member pH 2 hexagonal H-birnessite. .................................................................................................... 153 Figure 5-12. Correlation between the fraction of triclinic Na-birnessite determined by EXAFS and FTIR to the fraction of Mn3+ determined from XPS in a sample................. 154 Figure 6-1. Schematic diagram of (a) triclinic and (b) hexagonal birnessite after Lanson et al. (2000). ......................................................................................................................... 192 Figure 6-2. XRD pattern of the same synthetic triclinic Na-birnessite using a (a) Cu tube and (b) Mo tube during data collection. ........................................................................... 193 Figure 6-3. A metal removal unit at the passive coal mine drainage treatment site near Glasgow, PA. ................................................................................................................... 194 Figure 6-4. XRD patterns for (a) synthetic triclinic Na-birnessite, (b) synthetic hexagonal H-birnessite, (c) France 128319, (d) rancieite Spain, and (3) Paxton Cave. These natural samples from France, Spain, and Paxton Cave were identified as hexagonal birnessite using XRD. ...................................................................................................... 195 Figure 6-5. Select SEM images of XRD-identified hexagonal birnessite samples for (a) rancieite Spain, (b) France 128319, and (c) Paxton Cave. ............................................... 196 Figure 6-6. Moles of cations measured from EPMA in natural samples, excluding Mn. All values were calculated assuming that only Mn4+ existed in the sample, unless XPS measurements were performed on the sample to determine Mn oxidation state ratios. ................................................................................................................................ 197 Figure 6-7. XRD patterns for (a) synthetic triclinic Na-birnessite, (b) synthetic hexagonal H-birnessite, the fungally precipitated 3-line birnessite (c) fungal Stag50Ca1.6Mn, and natural 3-line birnessite samples (d) Glasgow, (e) Vermilion, (f) DS1-M3f, (g) DS2-M3f, (h) PBS-M2f-1, (i) PBS-M2f-2, and (j) Spring Branch. The dotted lines represent the identifying d-spacings for 3-line birnessite. ............................................... 198 Figure 6-8. Select SEM images of 3-line birnessite samples (a) DS1-M3f, (b) Glasgow, (c) DS2-M3f, (d) Spring Branch, (e) PBS-M2f-1, and (f) Vermilion. ............................. 199 Figure 6-9. χ(k) plots of 3-line birnessite samples with overlays of synthetic triclinic Nabirnessite and synthetic hexagonal H-birnessite χ(k) plots. ............................................. 200 Figure 6-10. Radial distribution functions (RDFs) of 3-line birnessite samples with overlays of synthetic triclinic Na-birnessite and synthetic hexagonal H-birnessite RDFs. ............................................................................................................................... 201 Figure 6-11. χ(k) plots of XRD-identified hexagonal birnessite samples with overlays of synthetic triclinic Na-birnessite and synthetic hexagonal H-birnessite χ(k) plots. .......... 202

xiii

Figure 6-12. Radial distribution functions (RDFs) of XRD-identified hexagonal birnessite with overlays of synthetic triclinic Na-birnessite and synthetic hexagonal Hbirnessite RDFs. ............................................................................................................... 203 Figure 6-13. Mn oxidation state ratios determined from XPS for select samples Paxton Cave, Glasgow, and Spring Branch compared to synthetic triclinic and synthetic hexagonal birnessites. ...................................................................................................... 204 Figure 6-14. FTIR spectra of (a) synthetic triclinic Na-birnessite, and XRD-identified hexagonal (b) Paxton Cave, (c) rancieite Spain, (d) France 128319 samples, and (d) synthetic hexagonal H-birnessite. .................................................................................... 205 Figure 6-15. FTIR spectra of (a) synthetic triclinic Na-birnessite, the 3-line birnessites (b) Spring Branch, (c) PBS-M2f-1, (d) PBS-M2f-2, (e) DS2-M3f, (f) Vermilion, (g) DS1-M3f, (h) Glasgow, (i) fungal Stag50Ca1.5Mn, and (j) synthetic hexagonal Hbirnessite. ......................................................................................................................... 206 Figure 6-16. FTIR spectra of original buserites and their spectra after drying at 110°C to become birnessite for (a-b) Spring Branch, (c-d) PBS-M2f-1, (e-f) PBS-M2f-2, (g-h) DS2-M3f, (i-j) DS1-M3f. ................................................................................................. 207 Figure 6-17. (a) Trends in Peak 1, Peak 2, and Peak 3 observed from the FTIR for 3-line birnessites as hexagonality increases according to LCF of EXAFS data. (b) The peak shifts are depicted in the FTIR spectra in the range from 400 to 650 cm-1. ..................... 208 Figure 6-18. Trends in Peak 1, Peak 2, and Peak 3 observed from FTIR according to their Mn3+ concentrations for the select samples on which XPS was conducted. Samples include synthetic hexagonal birnessite (HB), synthetic triclinic Na-birnessite (NaB), Paxton Cave (PC), Glasgow (GL), and Spring Branch (SB). .......................................... 209 SI-C Figure 1. XRD of original buserites and dehydrated buserites........................................ 215 SI-C Figure 2. Additional SEM images of natural Mn oxides. ................................................ 216 SI-C Figure 3. Full FTIR spectra of well-crystalline samples with non-birnessite phases labeled. ............................................................................................................................. 217 SI-C Figure 4. Full FTIR of poorly-crystalline samples, with non-birnessite phases labeled. ............................................................................................................................. 218

xiv

LIST OF TABLES Table 2-1. Rietveld refinement results for the original triclinic Na-birnessite. ....................... 42 Table 2-2. Rietveld refinement results for the final hexagonal birnessite created from adding triclinic Na-birnessite to a 20 mM HEPES solution at pH 7. ............................... 43 Table 2-3. Rietveld refinement results for the final hexagonal birnessite created from adding hexagonal H-birnessite to a pH 11 solution. ........................................................ 44 Table 3-1. Rietveld refinements for the unit cell and atom positions for hexagonal Hbirnessite. ......................................................................................................................... 77 Table 3-2. Rietveld refinements for the unit cell and atom positions for hexagonal birnessite with interlayer Mn after a 112-day batch reaction with Pb(NO3)2................... 78 Table 3-3. Rietveld refinements for the unit cell and atom positions for hexagonal birnessite with interlayer Pb after a 112-day batch reaction with Pb(NO3)2. ................... 79 Table 3-4. Bond lengths from final refinements of hexagonal H-birnessite reacting in a 112-day batch reaction with pH 3 0.1 M Pb(NO3)2 using either Mn(int) or Pb(int) as the interlayer cations. ....................................................................................................... 80 Table 4-1. Rietveld refinement results for unit cell parameter for end-member synthetic triclinic Na-birnessite and hexagonal H-birnessite. ......................................................... 115 Table 4-2. Atom positions for end-member synthetic triclinic Na-birnessite and hexagonal H-birnessite from Rietveld analysis................................................................ 116 Table 4-3. Peak locations in the Mn-O range (400 to 750 cm-1) for synthetic birnessites, and their respective peak labels to compare with Fig. 8. ................................................. 117 Table 4-4. Volume fractions of triclinic Na- and hexagonal H-birnessite in mixtures determined with FTIR, EXAFS, and XRD. ..................................................................... 118 Table 4-5. Peak positions from calculated FTIR spectra of birnessite nanoclusters................ 119 Table 6-1. Peak locations in the Mn-O range (400 to 750 cm-1) for synthetic birnessites. ...... 155 Table 6-2. Rietveld refinement results of unit cell parameters for triclinic cationexchanged birnessites using a triclinic unit cell. .............................................................. 156 Table 6-3. Atom positions for triclinic cation-exchanged birnessites from Rietveld refinements. ...................................................................................................................... 157 Table 5-4. Rietveld refinement results of unit cell parameters for hexagonal birnessites using a hexagonal unit cell (P -3) and a triclinic unit cell (C -1). .................................... 158 Table 5-5. Atom positions for hexagonal birnessites from Rietveld refinements using both hexagonal and triclinic unit cells...................................................................................... 159

xv

Table 5-6. EXAFS linear combination fitting results for fractions of end-member triclinic Na-birnessite and end-member pH 2 hexagonal H-birnessite. ......................................... 160 Table 5-7. FTIR spectral unmixing results for fractions of end-member triclinic Nabirnessite and end-member pH 2 hexagonal H-birnessite. ............................................... 161 Table 6-1. Analytical details for elements measured using the electron probe microanalyzer. .................................................................................................................. 210 Table 6-2. List of birnessite samples, categorized as either well-crystalline or poorlycrystalline. ........................................................................................................................ 211 Table 6-3. EXAFS linear combination fitting results. ............................................................. 213 Table 6-4. FTIR peaks for birnessites. ..................................................................................... 214

xvi

ACKNOWLEDGEMENTS This research was made possible through funding from the National Science Foundation (NSF Grant EAR-1147728 and EAR-1552211), the Committee on Institutional Cooperation (CIC) and Smithsonian Institution Fellowship, the Penn State Institutes of Energy and the Environment (PSIEE) Seed Grant Program, the Hiroshi and Koya Ohmoto Graduate Fellowship, and Krynine fund grants. A portion of this work was carried out at beamline GeoSoilEnviroCARS (Sector 13) and 12-BM at the Advanced Photon Source (APS), Argonne National Laboratory. GeoSoilEnviroCARS is supported by the National Science Foundation Earth Sciences (EAR-1128799) and Department of Energy - GeoSciences (DE-FG0294ER14466). This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357. The staff members at APS were also critical to my success, including Joanne Stubbs, Peter Eng, Nancy Lazarz, Sungsik Kim, and Benjamin Reinhart. Their support and around-the-clock assistance during my numerous trips to the beamline has made my work so much easier. Parts of this research were conducted with Advanced CyberInfrastructure computational resources provided by The Institute for CyberScience at The Pennsylvania State University (http://ics.psu.edu). The XRD and ICP-MS at the Materials Characterization Laboratory at the Pennsylvania State University were also utilized in this work, with the assistance of Nichole Wonderling and Melanie Saffer. The research that was completed at the Smithsonian Institution included use of the XRD, the SEM, the EPMA, and the FTIR. The FTIR laboratory at the Smithsonian Institution was established with generous support from Stephen Turner. Individuals at the Smithsonian who helped tremendously in my work there include Timothy Gooding, Timothy Rose, Alexandre

xvii

Fowler, Keal Byrne, and Margaret Anne Hinkle. I would especially like to thank Jeffrey Post, my supervisor at the Smithsonian, whose ideas and guidance made this work possible, and to Cara Santelli, for her mentorship and advice. I would like to acknowledge the co-authors for various chapters of this dissertation for their contribution to this work. The co-authors for Chapter 2 include Peter J. Heaney, Jeffrey E. Post, and Xiang Gao. The co-authors for Chapter 3 are Peter J. Heaney and Jeffrey E. Post. The co-authors for Chapter 4 are Jeffrey E. Post, Peter J. Heaney, James D. Kubicki, and Cara M. Santelli. The co-authors of Chapter 5 include Jeffrey E. Post, Peter J. Heaney, Eugene S. Ilton, and Cara M. Santelli. The co-authors of Chapter 6 include Jeffrey E. Post, Peter J. Heaney, Cara M. Santelli, Eugene S. Ilton, William D. Burgos, and Arthur W. Rose. Past and present members of our research group have also acted as great resources, teachers, labmates, and friends. These include Claire Fleeger, Kristina Peterson, Xiang Gao, Sunny Lin, Matt Oxman, and Phil Kong. I would like to thank Claire, in particular, for answering my emails on birnessite even several years after she had left. Lastly, my committee has provided invaluable advice and input to this work: James Kubicki, William Burgos, and Chris House. In regards to my advisor, Peter Heaney, I could not have asked for a better mentor. I would like to thank him for allowing me to pursue opportunities to help reach my career aspirations, and for providing the positive feedback I needed to feel confident about my work. He has always been supportive and understanding of his students’ career and personal choices, and I truly appreciate that.

xviii

DEDICATION I dedicate this work to my family and friends. To my parents, especially, who raised me to value my education and to reach for my dreams, I can only imagine the sacrifices you have made and the obstacles you endured when you hopped on that plane or boat in search for a better future. The best I can do is listen closely and remember the bedtime stories of hardships long past, and remind myself of how incredibly privileged I am to have parents as supportive and loving as the both of you, parents who have given me every possible opportunity in their power for me to succeed in life. To my brother Norman who taught me the value of hard work and self-discipline, I know I will probably never meet someone as motivated as you. You taught me the persistence I needed to get through graduate school. To my sister Felice, who inspired me to always climb higher and follow my passion, who reminded me in tough times that her little sister does not fail, and that she will most certainly pass her candidacy exam. To Matt Mizuhara, whose companionship has made graduate school a million times better than I thought it would be. You are the reason the past few years have gone so quickly.

1

Chapter 1 Introduction Manganese (Mn) oxides are found in a wide range of geological settings as more than 30 mineral varieties, from minerals in fine-grained aggregates in sediments and soils, to desert varnishes, and in lake and ocean nodules (Post 1999). Mn itself occurs naturally as Mn2+, Mn3+, and Mn4+, resulting in the large variety of Mn minerals that form with distinct oxidation states and structures. Mn oxides exist naturally as both tunnel and layer structures, with MnO6 octahedra as their building blocks. Tunnel structures vary in the size of their tunnels, ranging from 1x1 tunnels to 4x3 tunnels, and in the cations occupying their tunnels. Layer structures can also vary by the types of cations occupying the interlayer. In addition, layered Mn oxides vary by stacking Mn octahedral sheets with other metal hydroxide layers, by having different degrees of disorder in their stacking arrangements, and by holding variable amounts of water in the interlayer. In this work, we focus on the layered Mn oxide birnessite, studied for its common occurrence and high reactivity (Post 1999; Weaver and Hochella 2003). Birnessite was first discovered near Birness, Scotland by Jones and Milne (1956). Birnessite can occur with triclinic and hexagonal structures. Post and Veblen (1990) and Post et al. (2002) describe synthetic triclinic Na-birnessite (Na0.58(Mn4+1.42Mn3+0.58)O4 • 1.5H2O) as a layered Mn oxide with Mn octahedral sheets containing ~29% Mn3+ and ~71% Mn4+ cations. The interlayer holds hydrated Na+ cations that partially occupy the interlayer sites. In contrast, synthetic hexagonal H-birnessite (H0.33Mn3+0.111Mn2+0.055(Mn4+0.722Mn3+0.111☐0.167)O2), according to Silvester et al. (1997), consists

2

of Mn octahedral sheets with ~72% Mn4+ cations, ~11% Mn3+ cations, and ~17% vacancies, while Mn2+, Mn3+, and H+ cations occupy locations over the vacancy sites. Studies of Mn oxides such as birnessite are challenged by their small particle size and poor crystallinity, and multiple techniques such as X-ray diffraction (XRD), transmission electron microscopy, microprobe analysis, and infrared spectroscopy are often necessary for a full characterization (Post 1999). More recently, X-ray absorption spectroscopy (XAS) has become the standard technique for identifying poorly crystalline Mn oxides. Bargar et al. (2009), for example, characterized Mn oxides from Pinal Creek, AZ with XRD, providing evidence that it resembled a buserite structure that collapsed to a birnessite structure upon heating at 110°C. However, peaks in the XRD pattern were too broad to reveal whether the material was triclinic or hexagonal. By applying XAS, however, Bargar et al. (2009) claim to have identified the birnessite as hexagonal. The relationship between triclinic and hexagonal birnessite is an area of interest as a consequence of their chemical and structural differences. Triclinic birnessite transforms to hexagonal birnessite in a low pH solution (Giovanoli et al. 1970). According to Silvester et al. (1997) and Drits et al. (1997), the change in symmetry occurs alongside the disproportionation of 2Mn3+  Mn2+ + Mn4+ as triclinic birnessite transforms to hexagonal birnessite. In triclinic birnessite, the Mn3+ cation creates Jahn-Teller distortions in the Mn octahedral sheet, such that the Mn octahedra have longer axial bonds, and shorter equatorial bonds. However, upon transitioning to hexagonal birnessite with the loss of Mn3+, the Mn octahedra become less distorted and more symmetrical, creating the higher-symmetry structure of hexagonal birnessite. Because Mn2+ is too large to fit into the octahedral sheet, it moves into the interlayer, creating a vacancy, and forcing other interlayer cations in the original triclinic birnessite such as Na+ to exchange out. H+ is thought to exist as an interlayer cation for charge balance, with the amount of structural H+ dependent on the pH.

3

The relation between crystal structure and cation exchange kinetics remains an area of active exploration. Lopano et al. (2007, 2009, 2011), for instance, showed that K+, Cs+, and Ba2+ readily exchange into the interlayer of synthetic Na-birnessite to replace Na+. Fleeger et al. (2013) showed that hexagonal birnessite also exchanges Cs+ into the structure with higher interlayer Cs taken up at pH’s above 6.5. In addition to cation-exchange reactions, birnessites sorb metals including Pb, Zn, and Ni. Extended X-ray absorption fine structure (EXAFS) studies show that Pb, Zn, and Ni all sorb to the vacancy sites in birnessite (Kwon et al. 2010; Toner et al. 2006; Peacock 2009). The ability of Mn oxides to sorb and cation-exchange metals makes Mn oxides not only important as metal cyclers in the environment, but also useful in the remediation of contaminant metals. Several studies have described the capacity of Mn oxides to oxidize such metals as Co, As, and Cr (Weaver and Hochella 2003; Lafferty et al. 2011; Simanova and Peña 2015). Fischer (2011) observed that the oxidation of Cr3+ to Cr6+ is correlated with a phase transformation from triclinic to hexagonal birnessite. Interestingly, Mn3+ appears to be a key player in the mechanisms for several redox reactions. For both Cr3+ and Co2+, for example, the rate of oxidation by Mn oxides appears dependent on the amount of Mn3+ present (Manceau et al. 1992; Weaver and Hochella 2003; Simanova and Peña 2015). Similarly, Mn3+ plays a crucial role in the oxidation of water. Water oxidation is a ratelimiting step in plant photosynthesis, which operates via a CaMn4O5 cluster in Photosystem II (Ferreira et al. 2004; McEvoy and Brudvig 2006; Umena et al. 2011). In order to develop a synthetic analogue for the natural CaMn4O5 cluster to incorporate into solar cell technology, scientists have turned to Mn oxides such as Ca-birnessite, and have found that the Mn3+ content plays a critical role in the ability of Mn oxide to oxidize water (Takashima et al. 2012a,b; McKendry et al. 2015).

4

Most studies have focused on synthetic birnessites because of their greater crystallinity, but some studies have shown that natural birnessites also exist as both triclinic and hexagonal varieties. Tan et al. (2010), for example, identified a mix of triclinic and hexagonal birnessite in an acid mine drainage remediation site in Pennsylvania. Similarly, Shiraishi et al. (2016) identified co-existing triclinic and hexagonal birnessite in an ocean nodule from the Pacific floor. These studies took advantage of synchrotron X-ray absorption spectroscopy as a means of interrogating poorly crystalline samples. In addition to phase identification, several studies have focused on the formation of Mn oxides in natural environments. Although abiotic precipitation may sometimes be responsible for Mn oxide formation, the kinetics of biotic precipitation are 4 times faster, suggesting that bioprecipitation plays a major role in the crystallization of natural Mn oxides (Hastings and Emerson, 1986). Most studies have found that the initial phase bioprecipitated by bacteria is a hexagonal birnessite, according to EXAFS analysis (Villalobos et al. 2003, 2006; Bargar et al. 2005; Saratovsky et al. 2006). Fungi also have been shown to bioprecipitate hexagonal birnessite and even todorokite (Saratovsky et al. 2009; Santelli et al. 2011). Bacterially produced hexagonal birnessite appears to age over time, transforming to “pseudo-orthogonal” birnesite (Webb et al. 2005), and through a process of refluxing, birnessites can become the tunnel-structured todorokite (Feng et al., 2010). This dissertation consists of five separate studies, presented in Chapters 2 through 6. A paper published with Dr. Eugene Ilton (Pacific Northwest National Laboratory) is included as an appendix. In Chapter 2, we explore the effects of pH on the formation of hexagonal and triclinic birnessite. Buffers such as HEPES and MES typically are used to maintain circumneutral pH in bioprecipitation experiments, optimizing growth conditions for bacteria and fungi. It is unclear whether the buffers themselves influence birnessite precipitation and over which pH ranges

5

hexagonal and triclinic birnessite are favored. My batch experiments demonstrated that triclinic birnessite is stabilized relative to the hexagonal variety above pH 8.2, but that the kinetics of transformation from triclinic to hexagonal birnessite are exceedingly slow at circumneutral pH (with only partial transformation after ?? months at pH 7). However, when synthetic triclinic Nabirnessite was reacted with a pH 7 20 mM HEPES or MES buffered solution, the powders transformed to well-ordered hexagonal birnessite within 24 hr. Consequently, the buffers in bioprecipitation experiments likely are forcing the formation of hexagonal birnessite. This chapter was published in Chemical Geology. Chapter 3 examines the sorption of Pb into synthetic triclinic and hexagonal birnessite, testing whether differences in birnessite phase influence the effectiveness of birnessite as a remediator for contaminant Pb. Time-resolved X-ray diffraction (TR-XRD) experiments were conducted to examine the changes in crystal structure as both triclinic and hexagonal birnessite reacted with Pb(NO3)2 solutions at pH 3. Triclinic Na-birnessite transformed into a turbostratically disordered hexagonal-like phase upon Pb sorption within 6 hrs, while hexagonal H-birnessite showed little change in that time. However, extended batch reactions at pH 3 and pH 5 that ran for 7, 14, and 112-days indicated that Pb sorption into hexagonal H-birnessite manifests itself through an increase in the peak ratios of the (1 0 0) and (1 0 -1) peaks. Triclinic Na-birnessite displayed similar turbostratically disordered phases at pH 3 and pH 5 after 112-days in the batch reactions. This chapter will be submitted to American Mineralogist. Chapter 4 establishes Fourier-transform infrared spectroscopy (FTIR) as an alternative to synchrotron X-ray absorption and diffraction techniques for studying and quantifying phase fractions of triclinic and hexagonal birnessite. For the first time, we identify differences in the FTIR spectra of synthetic triclinic Na-birnessite and synthetic hexagonal H-birnessite. We also perform spectral unmixing to determine the phase fractions of triclinic and hexagonal birnessite in mixtures of known, pre-weighted ratios of the two phases. This method gave errors within 13

6

vol% of the known values, and were comparable to results obtained from linear combination fitting (LCF) of extended X-ray absorption fine structure (EXAFS) data, a common synchrotron method of determining volume fractions of different Mn oxide phases. Density functional theory (DFT) calculations of birnessite nanoparticles also provided insight into the vibrational modes that are responsible for the differences in the FTIR spectra of synthetic triclinic and hexagonal birnessite. Chapter 4 is in review with Chemical Geology. In Chapter 5, we analyze a series of synthetic birnessites, including triclinic Na-birnessite and hexagonal H-birnessite, along with cation-exchanged K-, Ca-, and Ba-birnessites to further test the abilities of FTIR to characterize birnessites in combination with X-ray diffraction (XRD), EXAFS, and X-ray photoelectron spectroscopy (XPS). Using FTIR and EXAFS, we fit the cation-exchanged birnessites with end-member triclinic Na-birnessite and end-member hexagonal H-birnessite in linear combination procedures. We then raise the question of whether a range of birnessite structures exist between an end-member triclinic and an end-member hexagonal structure, or whether most mixed birnessites are actually interlayered mixtures of endmember triclinic and hexagonal birnessite. To further explore this question, Rietveld refinements of XRD patterns of all birnessites were performed, showing systematic trends in the unit-cell parameters with the degree of triclinicity. When combined with XPS Mn oxidation state ratios from Ilton et al. (2016), these analyses revealed that the β angle increases with increasing Mn3+ content. This observation suggests that Mn3+ distorts the Mn octahedra as Mn3+ concentration increases, as hypothesized by Silvester et al. (1997) and Drits et al. (1997). We also found similar correlations between Mn3+ content and the fraction of triclinic birnessite as determined by FTIR and EXAFS analyses. This chapter will be submitted to American Mineralogist. The last chapter turns its focus towards natural birnessites. By applying the same techniques used in the previous chapter to natural birnessites, we successfully explored the crystal structures of birnessites that are not tractable using XRD alone. Although several samples

7

presented as hexagonal ranciéite-like phases using XRD, most of the material exhibited what we call “3-line” patterns with broad, asymmetric peaks. Linear combination fitting of EXAFS spectra yielded fractions of triclinic and hexagonal birnessite for a given sample, but we were skeptical of the results. We also applied FTIR to the natural “3-line” birnessites and obtained spectra that seemed more directly interpretable than the EXAFS spectra. FTIR peak positions in the 450 to 750 cm-1 range suggested that an entire range or mix of triclinic and hexagonal birnessites exist in natural environments. When combined with XPS, the EXAFS and FTIR showed that Mn3+ content correlates with increasing triclinic character in natural birnessites, in agreement with our observations of synthetic birnessites. Chapter 6 will be submitted to American Mineralogist.

References Cited Bargar, J. R., Fuller, C. C., Marcus, M. A., Brearley, A. J., Perez De la Rosa, M., Webb, S. M., & Caldwell, W. A. (2009). Structural characterization of terrestrial microbial Mn oxides from Pinal Creek, AZ. Geochimica et Cosmochimica Acta, 73, 889–910. http://doi.org/10.1016/j.gca.2008.10.036 Bargar, J. R., Tebo, B. M., Bergmann, U., Webb, S. M., Glatzel, P., Chiu, V. Q., & Villalobos, M. (2005). Biotic and abiotic products of Mn(II) oxidation by spores of the marine Bacillus sp. strain SG-1. American Mineralogist, 90(Ii), 143–154. http://doi.org/10.2138/am.2005.1557 Drits, V. A., Silvester, E., Gorshkov, A. I., & Manceau, A. (1997). Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite: II. Results from chemical studies and EXAFS spectroscopy. American Mineralogist, 82(9-10), 962–978. Feng, X. H., Zhu, M., Ginder-Vogel, M., Ni, C., Parikh, S. J., & Sparks, D. L. (2010). Formation of nano-crystalline todorokite from biogenic Mn oxides. Geochimica et Cosmochimica Acta, 74, 3232–3245. http://doi.org/10.1016/j.gca.2010.03.005 Ferreira, K. N., Iverson, T. M., Maghlaoui, K., Barber, J., & Iwata, S. (2004). Architecture of the Photosynthetic Oxygen-Evolving Center. Science, 303, 1831–1838. Fischer, T. B. (2011). Structural transformations of birnessite (δ-MnO2) during biological and abiological reduction. Fleeger, C. R., Heaney, P. J., & Post, J. E. (2013). A time-resolved X-ray diffraction study of Cs exchange into hexagonal H-birnessite. American Mineralogist, 98(2000), 671–679. http://doi.org/10.2138/am.2013.4287 Giovanoli, R., Stähli, E., & Feitknecht, W. (1970). Über Oxidhydroxide des vierwertigen Mangans mit Schichtengitter. 1. Mitteilung: Natrium-mangan(II,III)manganat(IV).

8

Helvetica Chimica Acta, 53, 454 – 464. Hastings, D., & Emerson, S. (1986). Oxidation of manganese by spores of a marine bacillus: Kinetic and thermodynamic considerations. Geochimica et Cosmochimica Acta, 50, 1819– 1824. http://doi.org/10.1016/0016-7037(86)90141-9 Jones, L. H. P., & Milne, A. A. (1956). Birnessite, a new manganese oxide mineral from Aberdeenshire, Scotland. The Mineralogical Magazine, XXXI(235), 283 – 288. Kwon, K. D., Refson, K., & Sposito, G. (2010). Surface complexation of Pb(II) by hexagonal birnessite nanoparticles. Geochimica et Cosmochimica Acta, 74(23), 6731–6740. http://doi.org/10.1016/j.gca.2010.09.002 Lafferty, B. J., Ginder-Vogel, M., & Sparks, D. L. (2011). Arsenite Oxidation by a PoorlyCrystalline Manganese Oxide. 3. Arsenic and Manganese Desorption. Environmental Science & Technology, 45, 9218–9223. http://doi.org/10.1021/es201281u Lopano, C. L., Heaney, P. J., Bandstra, J. Z., Post, J. E., & Brantley, S. L. (2011). Kinetic analysis of cation exchange in birnessite using time-resolved synchrotron X-ray diffraction. Geochimica et Cosmochimica Acta, 75(14), 3973–3981. http://doi.org/10.1016/j.gca.2011.04.021 Lopano, C. L., Heaney, P. J., & Post, J. E. (2009). Cs-exchange in birnessite: Reaction mechanisms inferred from time-resolved X-ray diffraction and transmission electron microscopy. American Mineralogist, 94, 816–826. http://doi.org/10.2138/am.2009.3068 Lopano, C. L., Heaney, P. J., Post, J. E., Hanson, J., & Komarneni, S. (2007). Time-resolved structural analysis of K- and Ba-exchange reactions with synthetic Na-birnessite using synchrotron X-ray diffraction. American Mineralogist, 92, 380–387. http://doi.org/10.2138/am.2007.2242 Manceau, A., Gorshkov, A. I., & Drits, V. A. (1992). Structural chemistry of Mn, Fe, Co, and Ni in manganese hydrous oxides: Part I. Information from XANES spectroscopy. American Mineralogist, 77, 1133–1143. McEvoy, J. P., & Brudvig, G. W. (2006). Water-Splitting Chemistry of Photosystem II. Chemical Reviews, 106(11), 4455–4483. http://doi.org/10.1021/cr0204294 McKendry, I. G., Kondaveeti, S. K., Shumlas, S. L., Strongin, D. R., & Zdilla, M. J. (2015). Decoration of the layered manganese oxide birnessite with Mn(II/III) gives a new water oxidation catalyst with fifty-fold turnover number enhancement. Dalton Transactions (Cambridge, England : 2003). http://doi.org/10.1039/c5dt01436k Peacock, C. L. (2009). Physiochemical controls on the crystal-chemistry of Ni in birnessite: Genetic implications for ferromanganese precipitates. Geochimica et Cosmochimica Acta, 73(12), 3568–3578. http://doi.org/10.1016/j.gca.2009.03.020 Post, J. E. (1999). Manganese oxide minerals: Crystal structures and economic and environmental significance. Proceedings of the National Academy of Sciences, 96, 3447–3454. http://doi.org/10.1073/pnas.96.7.3447 Post, J. E., Heaney, P. J., & Hanson, J. (2002). Rietveld refinement of a triclinic structure for synthetic Na-birnessite using synchrotron powder diffraction. Powder Diffraction, 17(3), 218–221. Post, J. E., & Veblen, D. R. (1990). Crystal structure determinations of synthetic sodium, magnesium, and postassium birnessite using TEM and the Rietveld Method. American Mineralogist, 75, 477–489. Santelli, C. M., Webb, S. M., Dohnalkova, A. C., & Hansel, C. M. (2011). Diversity of Mn oxides produced by Mn(II)-oxidizing fungi. Geochimica et Cosmochimica Acta, 75, 2762– 2776. http://doi.org/10.1016/j.gca.2011.02.022 Saratovsky, I., Gurr, S. J., & Hayward, M. A. (2009). The structure of manganese oxide formed by the fungus Acremonium sp. strain KR21-2. Geochimica et Cosmochimica Acta, 73,

9

3291–3300. http://doi.org/10.1016/j.gca.2009.03.005 Saratovsky, I., Wightman, P. G., Pastén, P. A., Gaillard, J.-F., & Poeppelmeier, K. R. (2006). Manganese Oxides: Parallels between Abiotic and Biotic Structures. Journal of the American Chemical Society, 128, 11188–11198. http://doi.org/10.1021/ja062097g Shiraishi, F., Mitsunobu, S., Suzuki, K., Hoshino, T., Morono, Y., & Inagaki, F. (2016). Dense microbial community on a ferromanganese nodule from the ultra-oligotrophic South Pacific Gyre: Implications for biogeochemical cycles. Earth and Planetary Science Letters, 447, 10–20. http://doi.org/10.1016/j.epsl.2016.04.021 Silvester, E., Manceau, A., & Drits, V. A. (1997). Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite: II. Results from chemical studies and EXAFS spectroscopy. American Mineralogist, 82, 962–978. Simanova, A. a., & Peña, J. (2015). Time-resolved investigation of cobalt oxidation by Mn(III)rich δ-MnO2 using quick X-ray absorption spectroscopy. Environmental Science & Technology, 49, 10867–10876. http://doi.org/10.1021/acs.est.5b01088 Takashima, T., Hashimoto, K., & Nakamura, R. (2012a). Inhibition of Charge Disproportionation of MnO2 Electrocatalysts for Efficient Water Oxidation under Neutral Conditions. Journal of the American Chemical Society, 134, 18153–18156. http://doi.org/10.1021/ja306499n Takashima, T., Hashimoto, K., & Nakamura, R. (2012b). Mechanisms of pH-Dependent Activity for Water Oxidation to Molecular Oxygen by MnO2 Electrocatalysts. Journal of the American Chemical Society, 134, 1519–1527. http://doi.org/10.1021/ja206511w Tan, H., Zhang, G., Heaney, P. J., Webb, S. M., & Burgos, W. D. (2010). Characterization of manganese oxide precipitates from Appalachian coal mine drainage treatment systems. Applied Geochemistry, 25, 389–399. http://doi.org/10.1016/j.apgeochem.2009.12.006 Toner, B., Manceau, A., Webb, S. M., & Sposito, G. (2006). Zinc sorption to biogenic hexagonalbirnessite particles within a hydrated bacterial biofilm. Geochimica et Cosmochimica Acta, 70, 27–43. http://doi.org/10.1016/j.gca.2005.08.029 Umena, Y., Kawakami, K., Shen, J.-R., & Kamiya, N. (2011). Crystal structure of oxygenevolving photosystem II at a resolution of 1.9 Å. Nature, 473, 55–60. http://doi.org/10.1038/nature09913 Villalobos, M., Lanson, B., Manceau, A., Toner, B., & Sposito, G. (2006). Structural model for the biogenic Mn oxide produced by Pseudomonas putida. American Mineralogist, 91, 489– 502. http://doi.org/10.2138/am.2006.1925 Villalobos, M., Toner, B., Bargar, J., & Sposito, G. (2003). Characterization of the manganese oxide produced by Pseudomonas putida strain MnB1. Geochimica et Cosmochimica Acta, 67(14), 2649–2662. http://doi.org/10.1016/S0016-7037(03)00217-5 Weaver, R. M., & Hochella, M. F. J. (2003). The reactivity of seven Mn-oxides with Cr3+aq : A comparative analysis of a complex, environmentally important redox reaction. American Mineralogist, 88, 2016–2027. Webb, S. M., Tebo, B. M., & Bargar, J. R. (2005). Structural characterization of biogenic Mn oxides produced in seawater by the marine bacillus sp. strain SG-1. American Mineralogist, 90(8-9), 1342–1357. http://doi.org/10.2138/am.2005.1669

10

Chapter 2 Transformations from triclinic to hexagonal birnessite at circumneutral pH induced through pH control by common biological buffers

Abstract Laboratory experiments that explore the bioprecipitation or redox transformations of layered Mn oxides commonly employ buffers, such as the HEPES and MES buffers, to maintain solution pH to near neutrality. The assumption is that holding solution pH constant does not serve as the primary control for the variety of Mn oxide produced. To test this assumption, synthetic triclinic Na-birnessite was reacted in batch experiments with a pH 7 HEPES buffer, a pH 7 MES buffer, and an unbuffered pH 7 solution for 14 days in total darkness. At the end of the experimental run, the Mn oxide solids were analyzed by conventional and synchrotron X-ray powder diffraction. These assays revealed that in the presence of the HEPES buffer, triclinic Na-birnessite completely transformed into highly crystalline hexagonal H-birnessite. In unbuffered solutions starting at pH 7 and in the presence of MES, which offers a lower buffering capacity than does HEPES, triclinic Na-birnessite partially transformed to poorly crystalline hexagonal H-birnessite. The unbuffered pH 7 solution exhibited an increase in pH to 8.03. We interpret the results to indicate that: 1) buffers can indirectly promote the transformation of triclinic Na-birnessite to hexagonal H-birnessite by serving as a source of H+, even at circumneutral pH; 2) triclinic Na-birnessite alone can stimulate hydrolysis, which in turn induces an exchange of H+ for interlayer Na+; and 3) H-birnessite and Na-birnessite operate as an acid-conjugate base pair.

11

Introduction Manganese (Mn) oxides occur in a wide variety of soil environments, typically as trace minerals that coat grain surfaces and occasionally as fine-grained concretions, veins, and nodules (Post 1999). Despite their relatively low abundance in most soils, Mn oxides can exert strong controls on redox activity, and they also exhibit a high capacity for cation exchange (Paterson et al. 1986; Lopano et al. 2007, 2009, 2011; Fleeger et al. 2013). Therefore, researchers have investigated Mn oxides for their role in the natural cycling of metals between soils and groundwater, as well as for their potential applications in solid-state batteries, microbial fuel cells, and even in solar cells (Zhang et al. 2011; Deng et al. 2014; Frey et al. 2014; Najafpour et al. 2014). In many natural low-temperature environments such as soils and exposed rock surfaces, bacteria and fungi are thought either to facilitate or to induce the precipitation of many varieties of Mn oxide phases (Northup et al. 2003; Webb et al. 2005; Bargar et al. 2009; Spiro et al. 2010; Santelli et al. 2010, 2011; Tang et al. 2013; Yang et al. 2013). These natural minerals typically are poorly crystalline, but among the most commonly observed biogenic Mn oxides are layered structures belonging to the birnessite family (Potter and Rossman 1979; McKeown and Post 2001; Manceau et al. 2002, 2007). X-ray diffraction studies of synthetic triclinic Na-birnessite, Na0.58(Mn4+1.42Mn3+0.58)O4 • 1.5H2O, have revealed that it consists of octahedral sheets containing Mn3+ and Mn4+ cations, with hydrated Na+ cations between the sheets to compensate for charge (Post and Veblen 1990) (Fig. 1a). In low pH (~3) solutions, triclinic birnessite rapidly transforms to a structure with hexagonal symmetry (Drits et al. 1997; Silvester et al. 1997; Lanson et al. 2000). The hexagonal H-birnessite structure reported by Lanson et al. (2000) has a proposed formula Mn2+0.05Mn3+0.12(Mn4+0.74Mn3+0.10☐0.17)O1.7(OH)0.3, so that it consists of octahedral sheets

12

with ~74% Mn4+ cations, 10% Mn3+ cations, and ~17% Mn vacancies, while Mn2+, Mn3+, and H+ cations occupy the interlayer (Fig. 1b). Efforts to generate Mn oxides biotically in laboratory experiments have predominantly reported the production of hexagonal rather than triclinic birnessite (Villalobos et al. 2003, 2006; Bargar et al. 2005; Saratovsky et al. 2006; Learman et al. 2011; Santelli et al. 2011; Hansel et al. 2012). Webb et al. (2005) note that the biogenic production of hexagonal birnessite by bacillus sp. strain SG-1 is followed by its transformation into a “pseudo-orthogonal” birnessite within 12 hrs. Feng et al. (2010) observe that a biogenic birnessite-like phase produced by Pseudomonas putida strain GB-1 can evolve into todorokite, a Mn oxide structure with 3x3 octahedral tunnels, after refluxing for 48 hrs. Nevertheless, even when birnessite is not the final reaction product, hexagonal birnessite is nearly always reported as the initial precipitate when a biological agency is responsible for crystallization. These experiments typically require that the bacteria or fungi are cultured at circumneutral pH, achieved by the utilization of HEPES [C8H16(OH)N2SO2OH] or similar buffers (e.g., Villalobos et al. 2003, 2006; Jürgensen et al. 2004; Bargar et al. 2005; Saratovsky et al. 2006, 2009; Feng et al. 2010; Santelli et al. 2010; Learman et al. 2011a,b; Hansel et al. 2012). Similarly, experiments aimed at understanding the interactions of birnessite-like phases with organic molecules, such as the siderophore desferrioxamine-B and citric acid, are usually buffered with HEPES, MES [C6H12NOSO2OH], or MOPS [C7H14NOSO2OH] to maintain nearneutral pH (e.g., Wang and Stone 2006; Duckworth and Sposito 2007; Duckworth et al. 2008). Moreover, many studies that have probed the reductive transformation of birnessite to other phases and the adsorptive characteristics of birnessite have employed HEPES and other buffers to maintain a constant pH (Elzinga 2011; Pena et al. 2011, 2015). In addition, phosphate and bicarbonate buffers are commonly employed in Mn oxide research involving technological applications of birnessite-like phases. For example, studies that

13

focus on the capacity of Mn oxides to oxidize water, with potential applications for artificial photosynthesis in solar energy storage, sometimes involve bicarbonate buffers (Iyer et al 2012; Robinson et al. 2013). Microbial fuel cell researchers have also utilized phosphate buffers when studying the effectiveness of Mn oxides as a cathode catalyst for the oxygen reduction reaction (Zhang et al. 2009; Zhang et al. 2011). Since a buffer is employed to maintain a constant and specific pH, the expectation is that the use of the buffer will not dictate the mineral phase that forms during bioprecipitation reactions. However, some hints that buffering plays a more deterministic role are evident from published studies. For example, investigations of water oxidation for artificial photosynthesis have shown inconsistent results relating Mn oxide structure and the rate of water oxidation. Iyer et al. (2012) find that amorphous Mn oxides produce more O2 during water oxidation than does a 2x2 tunnel-structured cryptomelane while using a bicarbonate buffer, whereas Meng et al. (2014), who did not use a buffer, report the opposite behavior for a similar 2x2 tunnel-structured hollandite. Both Meng et al. (2014) and Iyer et al. (2012) agree that layered Mn oxides are the weakest catalysts for water oxidation. However, Deibert et al. (2015) find that layered Mn oxides act as efficient water oxidizers when a bicarbonate buffer is used. Robinson et al. (2013), on the other hand, describe no electrocatalytic activity for 2x2 tunnel-structured Mn oxides and layered Mn3+,4+ oxides, but they do observe electrocatalytic activity for phases containing Mn3+ when using a bicarbonate buffer. Although the presence of a buffer may not be the only cause for these discrepancies, Menezes et al. (2014) argue that the electrocatalytic activity for Mn oxide water oxidation in solutions that contain carbonate, acetate, phosphate, and borate buffers will vary depending on the buffer used, and that water oxidation only occurs in buffered solutions. In addition, a recent radiotracer study, in which Mn-54 was used to characterize the extent of Mn atom exchange

14

between aqueous Mn2+ and vernadite (δ-Mn4+O2) at pH 7.5 under anoxic conditions, argues that the HEPES buffer is a reductant of structural Mn4+ in vernadite (Elzinga and Kustka 2015). In order to explore effects that pH buffering exerts on the products of reactions that involve Mn oxides, we have used conventional and synchrotron X-ray diffraction to study the interaction of HEPES and MES buffers with triclinic Na-birnessite in circumneutral aqueous solutions.

Methods

Materials synthesis Triclinic Na-birnessite was synthesized according to the procedure described in Golden et al. (1986). A 200 ml solution of 0.5 M MnCl2 (Mallinckrodt Baker) was mixed with 250 ml of 5.5 M NaOH (J.T. Baker). The mixture was oxygenated through a glass frit for ~5 hrs at a rate of 1.5 L/min. The precipitate was divided evenly and centrifuged in 14 centrifuge tubes. The solution was decanted and replaced with fresh deionized (DI) water to rinse. The rinse cycle was repeated five times. Na-birnessite was stored in ~350 ml DI water until experimental use. For experiments, aliquots of Na-birnessite were filtered with a 0.05 μm Nuclepore Track-Etched polycarbonate membrane filter (Whatman), rinsed three times with 100 mL DI water, and left to air-dry. The Na-birnessite was then ground in an agate mortar under acetone to disaggregate clumps. The synthesis of triclinic Na-birnessite with pH, as in the case for HEPES (where pKa 7.55 > pH 7), then [HA] > [A-]. On the other hand, for MES, a pKa of 6.15 is less than pH 7, indicating that [HA] < [A-]. Thus, when triclinic Na-binessite is placed in a pH 7 solution with HEPES, more undissociated acid is initially available relative to the base than is the case with MES at pH 7. HEPES consequently can supply more H+ than MES as the reaction progresses, inducing a complete transformation to hexagonal H-birnessite. Whereas the HEPES buffer supplied a sufficient concentration of H+ to fully convert triclinic Na-birnessite to hexagonal H-birnessite with no change in pH, the incompleteness of the transformation in the presence of MES allows the quantification of the amount of H+ taken up by the birnessite in that reaction. We recall that β = 0.0050 for the MES-buffered experiment, so from Eqn. 1 the total number of moles of acid removed per liter of solution to cause the observed increase in pH is represented by dn = 7.0 x 10-4 M H+. Additionally, by measuring the pH increase from 7.00 to 7.14 in 100 ml of solution, we can calculate that a loss of 2.76 x 10-9 mol H+ contributed to the pH change. Thus, the remaining protons were consumed by the powder, yielding an uptake of 0.00123 mol H+/g birnessite. Based on the chemical formula for triclinic Na-birnessite [Na0.58(Mn4+1.42Mn3+0.58)O4 • 1.5H2O with a molecular weight of 214.20 g/mol], the calculation reveals that 0.26 mol H+/mol birnessite were consumed. Interestingly, this value is approximately half the 0.58 mol Na+/mol of the starting triclinic birnessite, supporting our observation that ~53% of the triclinic birnessite transformed to hexagonal birnessite and providing evidence for a one-Na+ to one-H+ exchange.

25

When triclinic Na-birnessite was placed in a pH 6.30 MES-buffered solution, it transformed completely to hexagonal birnessite within 16 hrs (Fig. 5e). The pH of 6.30 is close to the pKa of MES, 6.15, yielding a calculated buffering capacity β = 0.011, greater even than that of the HEPES buffer at pH 7.00. This high buffering capacity for MES at pH 6.30 thus explains the efficiency of the transformation from triclinic to hexagonal birnessite in this experiment. The lower pH of the starting solution also may have contributed to the faster reaction kinetics relative the MES-buffered solution at pH 7.00. Nevertheless, the observation of a complete phase change close to the pKa of MES once again provides evidence that the buffering capacity controls the amount of hexagonal birnessite produced.

Cause of birnessite transformation in unbuffered solutions Even in the absence of buffers, at pH 7 triclinic Na-birnessite slowly transformed to hexagonal birnessite. We suggest that this occurred through hydrolysis followed by exchange of H+: H2O  H+ + OHH+(aq) + Na-(triclinic)birnessite(s)  Na+ (aq) + H-(hexagonal)birnessite(s)

(3) (4)

As suggested by Fig. 6, the uptake of H+ by triclinic Na-birnessite occurred in two stages. The reaction of triclinic Na-birnessite in an unbuffered pH 7 solution started with a sudden sharp increase in pH to ~9.3 followed by a more gradual decrease to ~8.2. Murray (1974) similarly observed that for hydrous Mn oxides, the addition of an aliquot of base would produce a rapid pH increase, followed by a slow decrease. These results were also similar to observations for ferric oxides in experiments conducted by Onoda and De Bruyn (1966). Both attributed this behavior to proton adsorption at the solid-solution interface. Thus, the rapid increase in pH when triclinic

26

Na-birnessite was first added to water probably resulted from the initial hydration of the Mn oxide surface, accompanied by rapid protonation of the surface. The subsequent gradual decrease to a steady-state pH of ~8.2 possibly resulted from a slow release of H+ or sorption of OH- on the hydration layer, with some potential contributions from acidification by atmospheric CO2 (Fig. 6) (Onoda and DeBruyn 1966; Murray 1974). Over the duration of our experiment, the uptake of H+ produced through hydrolysis initiated a partial transformation from triclinic to hexagonal birnessite.

Implications of reversibility of pH change When the addition of triclinic Na-birnessite to an unbuffered solution at pH 7 raised the final pH to ~8.2, the solid transformed to a mixture of 57 ± 1 wt% triclinic and 43 ± 2 wt% hexagonal birnessite after 14 days (Fig. 3b). That behavior suggests that hexagonal H-birnessite is favored below pH 8.2, triclinic Na-birnessite above pH 8.2, and that they co-exist at pH 8.2. We tested this hypothesis by exploring the reversibility of the reaction. We added hexagonal Hbirnessite to a solution at pH 11 – well above the co-existence pH of 8.2 – and monitored the change in pH. Over 22 days, the solution pH gradually decreased from 11 to a steady-state value of 8.19 (Fig. 6). Although some of the decrease may have resulted from acidification of the solution through atmospheric CO2, the magnitude of the pH decrease suggests that the release of H+ by hexagonal H-birnessite into solution was the dominant control. These results collectively indicate that hexagonal H-birnessite and triclinic Na-birnessite operate as an acid-conjugate base pair. After hexagonal H-birnessite was placed in a pH 11 solution for 22 days and the solution pH had stabilized at 8.19, X-ray diffraction of the reaction products revealed peaks corresponding only to hexagonal birnessite (Fig. 7d). In this respect, the lack of a co-existence of triclinic Na-

27

birnessite with hexagonal H-birnessite did not parallel the reverse reaction when triclinic Nabirnessite increased the pH of the host solution from 7.00 to ~8.20. However, some structural modification of hexagonal H-birnessite must have occurred as the pH lowered from 11.00 to 8.19, because the ratio of the (1 0 0) and (1 0 -1) peak intensities decreased to a value similar to that exhibited by triclinic Na-birnessite. Rietveld analysis indicated that this change in X-ray diffraction patterns resulted from a decrease in the occupancy of interlayer Mn to 0.088(2) (Table 3), in comparison with a refined occupancy of 0.127(2) for interlayer Mn in the HEPES hexagonal birnessite (Table 2). Aqueous Na+ was present in solution from the NaOH used to adjust pH, and it has a lower electron density than does Mn2+. Thus, it seems likely that the site occupancy decreased as Na+ replaced Mn2+ in the interlayer. This hypothesis is supported by observations by Fleeger (2012), who found that when hexagonal birnessite was added to a 0.1 M NaOH solution at pH 13, triclinic Na-birnessite crystallized within 20 min. Thus, the results reported here raise a question: Why does storing birnessite in DI water solutions that are initially at pH 7 not transform the synthetic triclinic birnessite to a mixture of triclinic and hexagonal varieties, with a final solution pH of ~8.2? We tested the pH of the storage solutions that had achieved a steady-state equilibrium with triclinic Na-birnessite for ~4 years, and our analyses revealed the pH of the supernatant to be 12.03, a value at which triclinic Na-birnessite indeed is expected to be favored relative to hexagonal H-birnessite. The storage solutions contained a birnessite-to-water ratio of ~31.03 gbirn/LH2O, much greater than the birnessite-to-water ratio of ~0.50 gbirn/LH2O that characterized our batch experiments. This observation suggests that birnessite deprotonation is characterized not by one but by two steadystate equilibria conditions. Thus, birnessite must host distinct hydroxyl sites with different acidities, and we speculate that the dissimilar behaviors can be attributed to disparities in the surface and interlayer hydroxyl groups. When the birnessite-to-fluid ratio is high, the total surface area is large and the surface hydroxyl sites maintain pH near 12. On the other hand, when

28

the birnessite-to-fluid ratio is low, the surface sites are saturated with OH and the interlayer hydroxyl groups control the acidity to near pH 8.2.

Hydrolysis vs. water oxidation The net consumption of H+ by triclinic Na-birnessite in a neutral solution suggests that the triclinic birnessite splits water into H+ and OH-; after the surface sites are saturated with respect to H+, dissolved H+ then exchanges for interlayer Na+, and the residual OH- induces an increase in solution pH according to the reactions described in equations (3) and (4). In a similar manner, HEPES and MES buffers promote the transformation from triclinic to hexagonal birnessite by providing a reservoir of H+ at circumneutral pH. Silvester et al. (1997) have argued on the basis of EXAFS that, in acidic conditions, Mn3+ disproportionates to Mn2+ and Mn4+ (2Mn3+  Mn2+ + Mn4+) during this phase transformation. Since we observed the same phase transformation at circumneutral pH in the presence of HEPES and MES, it seems likely that Mn3+ in triclinic Na-birnessite also undergoes disproportionation when placed in circumneutral solutions with buffers. Indeed, our observation that triclinic Na-birnessite in an unbuffered solution at pH 7 partially transforms to hexagonal H-birnessite indicates that disproportionation of Mn3+ occurs even in the absence of buffers. The hydrolytic behavior described above is distinct from the oxidative reactions catalyzed by birnessite-like phases in other chemical systems. Many Mn oxides are known to promote the splitting of water into molecular H2 and O2 according to the reaction: 2H2O  2H2 + O2

(5)

and this Mn-mediated electrocatalysis has been intensively studied as the basis for natural and artificial photosynthesis (Takashima et al. 2012a,b; Iyer et al. 2012; Robinson et al. 2013). Interestingly, Takashima et al. (2012a,b) have argued that Mn3+ is integral to the electrocatalytic

29

activity of Mn oxides, and that the disproportionation of Mn3+ to Mn2+ and Mn4+ inhibits the efficiency of birnessite for water oxidation. Based on our observation that triclinic birnessite transforms to hexagonal birnessite in both buffered and unbuffered solutions at pH 7, we suggest that even at circumneutral pH, an H+-mediated transition from triclinic to hexagonal birnessite will lead to disproportionation of Mn3+, impeding water oxidation. Consequently, research studying Mn oxide water oxidation should take into account the effect of buffers on the disproportionation of Mn3+ and on reaction rates.

Conclusions Since triclinic Na-birnessite can react with H+ and transform to hexagonal birnessite, even in neutral solutions, we must interpret Mn oxide studies that utilize buffers with caution. The use of a buffer in experiments that focus on bacterial or fungal biomineralization of birnessite-like phases may force the formation of hexagonal birnessite rather than allowing triclinic birnessite to crystallize. Indeed, natural buffers may function as a control on the specific Mn oxide phase that precipitates in soil and groundwater environments. Moreover, experiments that measure the efficiency of water oxidation in the presence of buffers may be influenced by the loss of Mn3+ as it disproportionates to Mn2+ and Mn4+ at circumneutral and lower pH values.

Acknowledgements Funding was provided by NSF grant EAR-1147728. Portions of this work were performed at GeoSoilEnviroCARS (Sector 13), Advanced Photon Source (APS), Argonne National Laboratory. GeoSoilEnviroCARS is supported by the National Science Foundation Earth Sciences (EAR-1128799) and Department of Energy - GeoSciences (DE-FG02-

30

94ER14466). This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357.

References Bargar, J.R., Tebo, B.M., Bergmann, U., Webb, S.M., Glatzel, P., Chiu, V.Q., Villalobos, M. (2005) Biotic and abiotic products of Mn(II) oxidation by spores of the marine Bacillus sp. Strain SG-1. American Mineralogist. 90. 143 – 154. Bargar, J.R., Fuller, C.C., Marcus, M.A., Brearley, A.J., Perez De la Rosa, M., Webb, S.M., Caldwell, W.A. (2009) Structural characterization of terrestrial microbial Mn oxides from Pinal Creek, AZ. Geochimica et Cosmochimica Acta. 73. 889 – 910. Boppana, V.B.R., Jiao, F. (2011) Nanostructured MnO2: an efficient and robust water oxidation catalyst. Chemical Communications. 47. 8973 – 8975. Butler, J.N. (1964) Solubility and pH Calculations. 62 – 69. Deibert, B.J., Zhang, J., Smith, P.F., Chapman, K.W., Rangan, S., Banerjee, D., Tan, K., Wang, H., Paquale, N., Chen, F., Lee, K., Dismukes, G.C., Chabal, Y.J., Li, J. (2015) Surface and Structural Investigaion of a MnOx Birnessite-Type Water Oxidation Catalyst Formed under Photocatalytic Conditions. Chemistry, A European Journal. 21. 1 – 12. Deng, Y., Wan, L., Xie, Y., Qin, X., Chen, G. (2014) Recent advances in Mn-based oxides as anode materials for lithium ion batteries. RSC Advances. 4. 23914 – 23935. Drits, V.A., Silvester, E., Gorkshkov, A.I., Manceau, A. (1997) Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite: I. Results from X-ray diffraction and selected-area electron diffraction. American Mineralogist. 82. 946 – 961. Duckworth, O.W., Sposito, G. (2007) Siderophore-promoted dissolution of synthetic and biogenic layer-type Mn oxides. Chemical Geology. 242. 497 – 508. Duckworth, O.W., Bargar, J.R., Sposito, G. (2008) Sorption of ferric iron from ferrioxamine B to synthetic and biogenic layer type manganese oxides. Geochimica et Cosmochimca Acta. 72. 3371 – 3380. Elzinga, E.J. (2011) Reductive Transformation of Birnessite by Aqueous Mn(II). Environmental Science & Technology. 45. 6366 – 6372. Elzinga, E.J., Kustka, A.B. (2015) A Mn-54 radiotracer study of Mn isotope solid-liquid exchange during reductive transformation of vernadite (δ-MnO2) by aqueous Mn (II). Environmental science & technology. 49. 4310–4316. Feng, X.H., Zhu, M., Ginder-Vogel, M., Ni, C., Parikh, S.J., Sparks, D. L. (2010) Formation of nano-crystalline todorokite from biogenic Mn oxides. Geochimica et Cosmochimica Acta. 74. 3232 – 3245. Fischer, T.B. (2011) Structural Transformations of Birnessite (δ-MnO2) During Biological and Abiological Reduction. Doctoral dissertation. The Pennsylvania State University. Finger, L.W., Cox, D.E., Jephcoat, A.P. (1994) A Correction for Powder Diffraction Asymmetry Due to Axial Divergence. Journal of Applied Crystallography. 27. 892 – 900. Fleeger, C. R. (2012). Contaminant sequestration and phase transformation properties of birnessite-like phases (-MnO2) Doctoral dissertation. The Pennsylvania State University.

31

Fleeger, C.R., Heaney, P.J., Post, J.E. (2013) A time-resolved X-ray diffraction study of Cs exchange into hexagonal H-birnessite. American Mineralogist. 98. 671 – 679. Frey, C.E., Wiechen, M., Kurz, P. (2014) Water-oxidation catalysis by synthetic manganese oxides – systematic variations of the calcium birnessite theme. Dalton Transactions. 43. 4370 – 4379. Giovanoli, R., Stähli, E., and Feitknecht, W. (1970) Über Oxidhydroxide des vierwertigen Mangans mit Schichtengitter. 1. Mitteilung: Natriummangan(II,III) manganat(IV). Helvetica Chimica Acta. 53. 209–220. Glover, E.D. (1977) Characterization of a marine birnessite. American Mineralogist. 62. 278– 285. Golden, D.C., Chen, C.C., Dixon, J.B. (1986) Synthesis of Todorokite. Science. 231. 4739. 717 – 719. Good, N.E., Winget, G.D., Winter, W., Connolly, T.N., Izawa, S., Singh, R.M.M. (1966) Hydrogen Ion Buffers for Biological Research. Biochemistry. 5. 2. 467 – 477. Hansel, C.M., Zeiner, C.A., Santelli, C.M., Webb, S.M. (2012) Mn(II) oxidation by an ascomycete fungus is linked to superoxide production during asexual reproduction. Proceedings of the National Academy of Sciences of the United States of America. 1 – 5. Iyer, A., Del-Pilar, J., King’ondu, C.K., Kissel, E., Garces, H.F., Huang, H., El-Sawy, A.M., Dutta, P.K., Suib, S.L. (2012) Water Oxidation Catalysis using Amorphouse Manganese Oxides, Octahedral Molecular Sieves (OMS-2), and Octahedral Layered (OL-1) Manganese Oxide Structures. The Journal of Physical Chemistry. 116. 6474 – 6483. Jürgensen, A., Widmeyer, J.R., Gordon, R.A., Bendell-Young, L.I., Moore, M.M., Crozier, E.D. (2004) The structure of the manganese oxide on the sheath of the bacterium Leptothrix discophora: An XAFS study. American Mineralogist. 89. 1110 – 1118. Lanson, B., Drits, V., Silvester, E., Manceau, A. (2000) Structure of H-exchanged hexagonal birnessite and its mechanism of formation from Na-rich monoclinic buserite at low pH. American Mineralogist. 85. 826-838. Learman, D.R., Voelker, B.M., Vazquez-Rodriguez, A.I., Hansel, C.M. (2011) Formation of manganese oxides by bacterially generated superoxide. Nature Geoscience. 4. 95 – 98. Learman, D.R., Wankel, S.D., Webb, S.M., Martinez, N., Madden, A.S., Hansel, C.M. (2011) Coupled biotic-abiotic Mn(II) oxidation pathway mediates the formation and structural evolution of biogenic Mn oxides. Geochimica et Cosmochimica Acta. 75. 6048 – 6063. Lopano, C.L., Heaney, P.J., Post, J.E., Hanson, J, Komarneni, S.K. (2007) Time-resolved structural analysis of K- and Ba-exchange reactions with synthetic Na-birnessite using synchrotron X-ray diffraction. American Mineralogist. 92. 380 – 387. Lopano, C.L., Heaney, P.J., Post, J.E. (2009) Cs-exchange in birnessite: Reaction mechanisms inferred from time-resolved X-ray diffraction and transmission electron microscopy. American Mineralogist. 94. 816 – 826. Lopano, C.L., Heaney, P.J., Bandstra, J.Z., Post, J.E., Brantley, S.L. (2011) Kinetic analysis of cation exchange in birnessite using time-resolved synchrotron X-ray diffraction. Geochimica et Cosmochimica Acta. 75. 3973 – 3981. Manceau, A., Marcus, M.A., Tamura, N. (2002) Quantitative speciation of heavy metals in soils and sediments by synchrotron X-ray techniques. Applications of Synchrotron Radiation in Low-Temperature Geochemistry and Environmental Sciences. 49. 341 – 428. Manceau, A., Lanson, B., Geoffrey, N. (2007) Natural speciation of Ni, Zn, Ba, and As in ferromanganese coatings on quartz using X-ray fluorescence, absorption, and diffraction. Geochimica et Cosmochimica Acta. 71. 95-128.

32

McKeown, D.A., Post, J.E. (2001) Characterization of manganese oxide mineralogy in rock varnish and dendrites using X-ray absorption spectroscopy. American Mineralogist. 86. 701 – 713. Menezes, P.W., Indra, A., Littlewood, P., Schwarze, M., Göbel, C., Schomäcker R., Driess, M. (2014) Nanostructured Manganese Oxides as Highly Active Water Oxidation Catalysts: A Boost from Manganese Precursor Chemistry. ChemSusChem. 7. 2202 – 2211. Meng, Y., Song, W., Huang, H., Ren, Z., Chen, S., Suib, S.L. (2014) Structure-Property Relationship of Bifunctional MnO2 Nanostructures: Highly Efficient, Ultra-Stable Electrochemical Water Oxidation and Oxygen Reduction Reaction Catalysts Identified in Alkaline Media. Journal of the American Chemical Society. 136. 114352 – 11464. Murray, J. (1974) The Surface Chemistry of Hydrous Manganese Dioxide. Journal of Colloid and Interface Science. 46. 3. Najafpour, M.M., Holyńska, M., Salimi, S. (2014) Applications of the “nano to bulk” Mn oxides: Mn oxide as a Swiss army knife. Coordination Chemistry Reviews. . Northup, D.E., Barns, S.M., Yu, L.E., Spilde, M.N., Schelble, R.T., Dano, K.E., Crossey, L.J., Connolly, C.A., Boston, P.J., Natvig, D.O., Dahm, C.N. (2003) Diverse microbial communities inhabiting ferromanganese deposits in Lechuguilla and Spider Caves. Environmental Microbiology. 5. 11. 1071 – 1086. Onoda, G.Y., De Bruyn, P.L. (1966) Proton Adsorption At The Ferric Oxide/Aqueous Solution Interface, I. A Kinetic Study of Adsorption. Surface Science. 4. 48 – 63. Paterson, E., Clark, D.R., Russell, J.D., Swaffield, R. (1986) Cation-exchange in synthetic manganates. 2. The structure of an alkylammonium-saturated phyllomanganate. Clay Minerals. 21. 5. 957 – 964. Pena, J., Bargar, J.R., Sposito, G. (2011) Role of Bacterial Biomass in the Sorption of Ni by Biomass-Birnessite Assemblages. Environmental Science & Technology. 45. 7338 – 7344. Pena, J., Bargar, J.R., Sposito, G. (2015) Copper sorption by the edge surfaces of synthetic birnessite nanoparticles. Chemical Geology. 396. 196 – 207. Post, J.E., Veblen, D.R. (1990) Crystal structure determinations of synthetic sodium, magnesium, and potassium birnessite using TEM and the Rietveld method. American Mineralogist. 75. 477 – 489. Post, J.E. (1999) Manganese oxide minerals: Crystal structures and economic and environmental significance. Proceedings of the National Academy of Sciences of the United States of America. 96. 3447 – 3454. Post, J.E., Heaney, P.J., Hanson, J. (2002) Rietveld refinement of a triclinic structure for synthetic Na-birnessite using synchrotron powder diffraction data. Powder Diffraction. 17. 3. 218 – 221. Potter, R.M., Rossman, G.R. (1979) The tetravalent manganese oxides: identification, hydration, and structural relationships by infrared spectroscopy. American Mineralogist. 64. 1199 – 1218. Prescher, C. and Prakapenka, V.B. (2015) DIOPTAS: a program for reduction of two-dimensional X-ray diffraction data and data exploration. High Pressure Research. 1 – 8. Robinson, D.M., Go, Y.B., Mui, M., Gardner, G., Zhang, Z., Mastrogiovannia, D., Garfunkel, E., Li, J., Greenblatt, M., Dismukes, G.C. (2013) Photochemical Water Oxidation by Crystalline Polymorphs of Manganese Oxides: Structural Requirements for Catalysis. Journal of the American Chemical Society. 135. 3494 – 3501. Santelli, C.M., Pfister, D.H., Lazarus, D., Sun, L., Burgos, W.D., Hansel, C.M. (2010) Promotion of Mn(II) Oxidation and Remediation of Coal Mine Drainage in Passive Treatmet

33

Systems by Diverse Fungal and Bacterial Communities. Applied and Environmental Microbiology. 76. 14. 4871 – 4875. Santelli, C.M., Webb, S.M., Dohnalkova, A.C., Hansel, C.M. (2011) Diversity of Mn oxides produced by Mn(II)-oxidizing fungi. Geochimica et Cosmochimica Acta. 75. 2762 – 2766. Saratovsky, I., Wightman, P.G., Pastén, P.A., Gaillard, J., Poeppelmeier, K.R. (2006) Manganese Oxides: Parallels between Abiotic and Biotic Structures. Journal of the American Chemical Society. 128. 11188 – 11198. Saratovsky, I., Gurr, S.J., Hayward, M.A. (2009) The Structure of manganese oxide formed by the fungus Acremonium sp. Strain KR21-2. Geochimica et Cosmochimica Acta. 73. 3291 – 3300. Silvester, E., Manceau, A., Drits, V.A. (1997) Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite: II. Results from chemical studies and EXAFS spectroscopy. American Mineralogist. 82. 962 – 978. Spiro, T.G., Bargar, J.R., Sposito, G., Tebo, B.M. (2010) Bacteriogenic Manganese Oxides. Accounts of Chemical Research. 43. 1. 2 – 9. Stephens, P.W. (1999) Phenomenological model of anisotropic peak broadening in powder diffraction. Journal of Applied Crystallography. 32. 281-289. Takashima, T., Hashimoto, K., & Nakamura, R. (2012a). Inhibition of charge disproportionation of MnO2 electrocatalysts for efficient water oxidation under neutral conditions. Journal of the American Chemical Society. 134. 18153-18156. Takashima, T., Hashimoto, K., Nakamura, R. (2012b) Mechanisms of pH-Dependent Activity for Water Oxidation to Molecular Oxygen by MnO2 Electrocatalysts. Journal of the American Chemical Society. 134. 1419 – 1527. Tang, Y., Zeiner, C.A., Santelli, C.M., Hansel, C.M. (2013) Fungal oxidative dissolution of the Mn(II)-bearing mineral rhodochrosite and the role of metabolites in manganese oxide formation. Environmental Microbiology. 15. 4. 1063 – 1077. Theppakorn, T., Kanasawud, P., Halling, P.J. (2003) Effect of solid-state buffers on the catalytic activity of papain in low water media. Enzyme and Microbial Technology. 32. 828 – 836. Thompson, P., Cox, D.E., Hastings, J.B. (1987) Rietveld Refinement of Debye-Scherrer Synchrotron X-Ray Data from Al2O3. Journal of Applied Crystallography. 34. 210 – 213. Toby, B.H., Von Dreele, R.B. (2013) GSAS-II: the genesis of a modern open-source all purpose crystallography software package. Journal of Applied Crystallography. 46. 544 – 549. Villalobos, M., Toner, B., Bargar, J., Sposito, G. (2003) Characterization of the manganese oxide produced by Pseudomonas putida strain MnB1. Geochomica et Cosmochimica Acta. 67. 14. 2649 – 2662. Villalobos, M., Lanson, B., Manceau, A., Toner, B., Sposito, G. (2006) Structural model for the biogenic Mn oxide produced by Pseudomonas putida. American Mineralogist. 91. 489 – 502. Wang, Y., Stone, A.T. (2006) The citric acid-MnIII,IVO2(birnessite) reaction. Electron transfer, complex formation, and autocatalytic feedback. Geochimica et Cosmochimica Acta. 70. 4463 – 4476. Webb, S.M., Tebo, B.M., Bargar, J.R. (2005) Structural characterization of biogenic Mn oxides produced in seawater by the marine bacillus sp. Strain SG-1. American Mineralogist. 90. 1342 – 1357. Yang, W., Zhang, Z., Zhang, Z., Chen, H., Liu, J., Ali, M., Liu, F., Li, L. (2013) Population Structure of Manganese-Oxidizing Bacteria in Stratified Soils and Properties of

34

Manganese Oxide Aggregates under Manganese-Complex Medium Enrichment. PLoS ONE. 8. 9. 1 – 13. Zhang, L., Liu, C., Zhuang, L., Li, W., Zhou, S., Zhang, J. (2009) Manganese dioxide as an alternative cathodic catalyst to platinum in microbial fuel cells. Biosensors and Bioelectronics. 24. 2825 – 2829. Zhang, Y., Hu, Y., Li, S., Sun, J., Hou, B. (2011) Manganese dioxide-coated carbon nanotubes as an improved cathodic catalyst for oxygen reduction in a microbial fuel cell. Journal of Power Sources. 196. 9284 – 9289.

35

Figures

Figure 2-1. Schematic diagram of (a) triclinic Na-birnessite and (b) hexagonal H-birnessite (after Lanson et al. 2000).

36

Figure 2-2. Rietveld refinement fits and difference curves for (a) initial triclinic Na-birnessite, (b) hexagonal birnessite formed from triclinic Na-birnessite in a pH 7 HEPES-buffered solution after 2 weeks, and (c) mixed triclinic and hexagonal birnessite formed from triclinic Na-birnessite in a pH 7 MES-buffered solution after 2 weeks. Crosses represent observed intensities; the solid green lines are the calculated patterns. The solid red lines are the calculated backgrounds. The solid aqua lines represent the difference between the observed and calculated patterns. Only data between the dashed vertical lines were used for refinements.

37

Figure. 2-3. X-ray diffraction patterns of (a) the initial triclinic Na-birnessite powder, (b) mixed hexagonal and triclinic birnessite after reacting in a pH 7 solution for 14 days, (c) the mixed hexagonal and triclinic birnessite after reacting in a pH 7 MES solution for 14 days, (d) hexagonal birnessite after reacting in a pH 7 HEPES solution for 14 days, and (e) hexagonal Hbirnessite synthesized by placing triclinic Na-birnessite in pH 3 solution for 24 hrs.

38

Figure 2-4. EDS spectra of (a) initial triclinic Na-birnessite, (b) birnessite in the pH 7 unbuffered solution, (c) birnessite in the pH 7 MES-buffered solution, and (d) birnessite in the pH 7 HEPESbuffered solution.

39

Figure 2-5. X-ray diffraction patterns of (a) the initial triclinic Na-birnessite powder, (b) mixed triclinic and hexagonal birnessite after reacting in a pH 7 MES solution for 12 hrs, (c) mixed hexagonal and triclinic birnessite after reacting in a pH 7 MES solution for 3 days, (d) mixed triclinic and hexagonal birnessite after reacting in a pH 7 MES solution for 14 days, (e) hexagonal birnessite after reacting in a pH 6.30 MES solution for 16 hrs, and (f) hexagonal H-birnessite.

40

Figure 2-6. The change in pH over time for batch experiments with triclinic Na-birnessite and pH 7 solution, and another with hexagonal H-birnessite in pH 11 solution.

41

Figure 2-7. X-ray diffraction patterns of (a) the initial triclinic Na-birnessite powder, (b) mixed hexagonal and triclinic birnessite after reacting in a pH 7 solution for 22 days, (c) the initial hexagonal H-birnessite powder, and (d) the hexagonal H-birnessite after reacting in a pH 11 solution for 22 days.

42

Tables Table 2-1. Rietveld refinement results for the original triclinic Na-birnessite. Space group: C -1 a = 5.183(4) Å b = 2.8561(4) Å α = 89.8(1)° β = 103.21(4)° Atom positions x Mn(oct) 0 O(oct) 0.3762(9) Na or H2O(int) 0.520(7) Refinement No. of observations No. of reflections Diffraction range (°2θ), λ = 0.8308 Å No. of variables R(F2) R wp χ2

c = 7.344(5) Å γ = 89.78(2)° y z 0 0 0.01(1) 0.1418(8) 0.178(3) 0.493(8)

Occupancy 1.000 1.000 0.93(3)

967 44 10.000 – 35.700 32 0.0382 0.0208 4.15

Uiso 0.005(2) 0.007(2) 0.19(1)

43

Table 2-2. Rietveld refinement results for the final hexagonal birnessite created from adding triclinic Na-birnessite to a 20 mM HEPES solution at pH 7. Space group: P -3 a = b = 2.8527(8) Å α = β = 90° Atom positions Mn(oct) O(oct) O(int) Mn(int) Refinement No. of observations No. of reflections

c = 7.217(2) Å γ = 120° x y 0 0 0.3333 0.6667 0.6667 0.3333 0 0

Diffraction range (°2θ), λ = 0.8308 Å No. of variables R(F2) R wp χ2

z 0 0.135(1) 0.467(3) 0.692(2)

Occupancy 0.95(1) 1.000 0.552(5) 0.127(2)

1174 17 4.496 – 35.700 24 0.0089 0.0189 1.82

Uiso 0.008(1) 0.004(2) 0.065(8) 0.014(6)

44

Table 2-3. Rietveld refinement results for the final hexagonal birnessite created from adding hexagonal H-birnessite to a pH 11 solution. Space group: P -3 a = b = 2.8452(7) Å α = β = 90° Atom positions Mn(oct) O(oct) O(int) Mn(int) Refinement No. of observations No. of reflections

c = 7.248(2) Å γ = 120° x y 0 0 0.3333 0.6667 0.6667 0.3333 0 0

Diffraction range (°2θ), λ = 0.8308 Å No. of variables R(F2) R wp χ2

z 0 0.115(1) 0.591(2) 0.654(2)

Occupancy 0.80(1) 1.000 0.537(7) 0.088(2)

2002 42 3.620 – 43.660 21 0.0590 0.0188 2.59

Uiso 0.006 0.002 0.093 0.015

45

Chapter 3 Sorption of contaminant Pb by triclinic and hexagonal birnessite

Abstract The layered manganese oxide, birnessite, can potentially be used to remediate contaminant lead. Past studies suggest that as the average oxidation state of Mn within the octahedral sheet increases, with a concomitant increase in vacancy concentration, Pb sorption also increases. We have explored this relationship by comparing the uptake of Pb by hexagonal Hbirnessite and triclinic Na-birnessite using time-resolved X-ray diffraction (TR-XRD). In addition, we have analyzed the products of 7-, 14-, and 112-day batch experiments in which Naand H-birnessite powders were mixed with 0.1 M Pb(NO3)2 solutions at pH 3 and 5. In our TR-XRD experiments involving pH 3, Pb-free solutions, triclinic Na-birnessite fully transformed to well-crystalline hexagonal H-birnessite within 30 min, and in pH 5 Pb-free solutions, it partially transformed to hexagonal H-birnessite within 5.5 hrs. In the presence of Pb, however, triclinic Na-birnessite transformed into a turbostratically disordered hexagonal-like birnessite phase within 1 hr at pH 3, and only partially transformed at pH 5. Although hexagonal H-birnessite exhibited no structural change when exposed to Pb solutions after 2 hr, it did display an increase in the ratio of the (1 0 0) to (1 0 -1) peak after 14 days at both pH 3 and 5, signifying an increase in electron density in the interlayer. Rietveld refinements and chemical analysis confirm that Pb continued to sorb to both hexagonal and triclinic birnessite over the course of several months, and that Pb entered the birnessite interlayer, replacing Mn2+. We attribute the slow kinetics of Pb uptake relative to larger alkali metals (e.g., Cs) to repulsions involving the lone pair of electrons on Pb. After 112 days at both pH 3 and pH 5, the transformation of triclinic Na-birnessite into hexagonal-like birnessite ultimately allowed for equal amounts of Pb uptake.

46

Introduction In humans, high Pb levels can cause long-term neurological disorders, such as mental retardation, learning disorders, and attention-deficit/hyperactivity disorder (Nigg et al. 2008). Fifteen to twenty percent of children in cities across the U.S. have lead (Pb) blood levels above 10 μg/dL, the initial screening level set by the U.S. Centers for Disease Control and Prevention. For urban children, Pb ingestion is largely attributed to chipped paint in poorly maintained homes and to the past use of leaded gasoline. Lead derived from these and other sources has accumulated in soils (Filipelli and Laidlaw 2010), and when those soils are infiltrated by stormwater runoff, Pb and other heavy metals can mobilize and contaminate groundwater (Weiss et al. 2008). Moreover, dangerous Pb levels have been found in groundwaters around the world due to increased agriculturalization and industrialization (Singh et al. 2014; Wongsasuluk et al. 2014), as well as mining operations (Ruby et al. 1999). Researchers have attempted to cover Pb-contaminated soils with clean soil, but they have found that treated areas become re-contaminated within several months due to re-suspension of Pb from neighboring soils (Mielke et al. 2006). Phytoremediation techniques have also been tested, for instance, by using plants combined with chelates in the soil to uptake Pb. Chelates such as ethylenediaminetetraacetic acid increase the solubility of Pb, allowing accumulation into the shoots of plants (Lee 2013). Fe and Mn (hydr)oxides are recognized as potential Pb remediators for their ability to uptake lead and other deleterious heavy metals (Gasparatos 2013). For example, Beak et al. (2008) have argued that Pb sorbed to birnessite in contaminated media will have a near-zero bioaccessibility, minimizing any risk associated with incidental ingestion of soil. The high cation exchange capacities of the layered Mn oxides within the birnessite family make them effective environmental sinks for many heavy metals (Post et al. 2002), and prior

47

studies have demonstrated that birnessite shows a strong affinity for Pb (McKenzie 1980; Lanson et al. 2002). Birnessite commonly occurs in soils as fine-grained aggregates and coatings, and in Fe-Mn nodules and concretions (Gasparatos 2013). Because it is naturally widespread, the ability of birnessite to mitigate the effects of Pb contamination is being investigated, and its use for remediation has shown promise. For example, 10-day toxicity tests on the organism Hyalella azteca in an aquatic system containing both dissolved Pb and birnessite resulted in a 95% increase in survival rate relative to a birnessite-free, Pb-contaminated environment (Lee et al. 2010). Furthermore, Lee et al. (2013) showed that the sorption capacity for Pb by birnessite is improved by the intercalation of tetramethylammonium. The Pb “removal efficiency” of this pillared birnessite was 78.9% at pH 5, exceeding the removal efficiency of 57.6% for untreated Na-birnessite. Birnessite exists as both hexagonal and triclinic varieties. Triclinic Na-birnessite is composed of octahedral sheets containing Mn3+ and Mn4+ cations, with hydrated Na+ cations within the interlayer to compensate for charge (Fig. 1a) (Post et al. 2002). Hexagonal Hbirnessite is composed of octahedral sheets containing Mn4+ cations and Mn vacancies, with Mn2+, Mn3+, and H+ cations in the interlayer (Fig. 1b) (Drits et al. 1997; Silvester et al. 1997; Lanson et al. 2000). Most published studies of the interactions between dissolved Pb2+ and birnessite do not account for possible distinctions between triclinic and hexagonal birnessite in their sorption of heavy metal cations. However, in light of the heightened reactivity associated with octahedral vacancies in birnessite (Kwon et al. 2010), crystallography may be crucial in understanding the mechanism and rates of Pb sorption to birnessite. Specifically, X-ray diffraction can reveal whether Pb is adsorbed only to surface sites or intercalated within the bulk structure for both birnessite phases. Researchers have attempted to clarify the role that octahedral vacancies play with respect to the uptake of Pb by birnessite. Zhao et al. (2009, 2011) showed that the sorption of Pb2+

48

increased with increasing average oxidation state of the Mn oxide layers. These authors interpreted the average oxidation state (AOS) as an indicator for the relative amounts of Mn2+, Mn3+, and Mn4+ within the octahedral sheets of birnessite. A high average oxidation state was construed to suggest high concentrations of Mn4+ within the sheets, and therefore more octahedral vacancies to balance the layer charge. Wang et al. (2012) also studied the sorption of Pb2+, Cu2+, Zn2+, and Cd2+ to hexagonal birnessite with varying Mn average oxidation states. They found that for all metals studied, sorption increased with increasing AOS. Like Della Puppa et al. (2013), they discovered that birnessite sorbed significantly higher amounts of Pb2+ compared to Cu2+, Zn2+, and Cd2+, which exhibited rates of sorption in the order Pb2+ >> Cu2+ > Zn2+ > Cd2+. They attributed the faster sorption of Pb2+ to birnessite to the bonding of Pb onto both edge and interlayer sites, whereas the other metal ions sorbed mostly to edge sites. Further supporting the hypothesis that vacancies control Pb sorption, structural studies have shown the association of Pb to vacancy sites. For example, Lanson et al. (2002) simulated X-ray diffraction patterns of Pb-sorbed Na-buserite, the hydrated relative of Na-birnessite, and found Pb located either above or below vacancies. X-ray absorption fine structure (EXAFS) analysis and density functional theory calculations also showed that Pb forms triple cornersharing complexes above and below vacancies located in hexagonal birnessite (Beak et al. 2008; Kwon et al. 2010; Zhao et al. 2011). More recently, van Genuchten and Peña (2016) calculated differential pair distribution functions (d-PDFs) using high-energy X-ray scattering data of Pbreacted δ-MnO2 and found that models of both triple-corner sharing complexes with a single Pb over a vacancy site and triple-corner sharing complexes with two Pb atoms above and below a vacancy were necessary in reproducing the experimental data, leading to the conclusion that both external and internal sorption sites contained Pb.

49

The goal of the present study is to compare the sorption of Pb to both triclinic and hexagonal birnessite using time-resolved X-ray diffraction to clarify the relationships among birnessite structure, vacancy concentration, and Pb sorption. In this study, sorption will refer to the uptake of Pb on both surface sites and through exchange in the interlayer of birnessite. We hoped that the high electron density associated with Pb would allow us to monitor its proposed entry into interlayer sites. An improved understanding of the mechanism by which Pb is sequestered by birnessite may suggest treatments that will enhance the capacity of Pb uptake in birnessite for remediation practices. In addition, these insights may constrain the role that Mn oxides play in metal cycling in natural environments (Grigoriev et al. 2013).

Experimental Section

Materials synthesis Triclinic Na-birnessite was synthesized according to the procedure described in Golden et al. (1986). A mixture of chilled 250 mL 5.5 M NaOH (J.T. Baker) and 200 mL 0.5 M MnCl2 (Mallinckrodt) was oxygenated for 5 hrs at a rate of 1.5 L/min. The precipitate was divided evenly and centrifuged in 14 centrifuge tubes. The solution was decanted and replaced with fresh DI water to rinse. The rinse cycle was repeated five times. Birnessite was stored in fresh DI water until experimental use. For the experiments, aliquots of stored triclinic Na-birnessite were filtered through 47 mm, 0.05 µm polycarbonate nuclepore track-etched membranes (Whatman) and allowed to air dry. The Na-birnessite was then ground in an agate mortar under acetone to disaggregate clumps. To make hexagonal H-birnessite, dried triclinic Na-birnessite was reacted with 250 mL of 0.001 M HCl (OmniTrace) in 250 ml glass beakers for 6 hours. The H-birnessite was then

50

filtered with a 0.05 μm filter and rinsed with 300 ml DI water. The material was allowed to air dry. Both triclinic Na-birnessite and hexagonal H-birnessite were characterized by a Rigaku II D/MAX-RAPID microdiffractometer (Materials Characterization Laboratory, Pennsylvania State University) with a Mo tube source (λ = 0.7093 Å).

Time-resolved x-ray diffraction experiments Time-resolved X-ray diffraction experiments (TR-XRD) were conducted at the Advanced Photon Source (APS) at Argonne National Laboratory at beamline 13-BM-C. Approximately 2 – 4 mg of birnessite were packed in 0.7 mm quartz capillaries (Charles Supper) between two cotton plugs and attached to a flow-through apparatus designed by Wall et al. (2011) (Fig. 2). Solutions of 0.1 M Pb(NO3)2 (Alfa-Aesar) adjusted to pH 3 with 0.1 M HNO3 (Fisher Scientific) and 0.1 M NaOH (Fisher Scientific) flowed through the quartz capillaries at a rate of ~1 drop/min. X-ray diffraction patterns were collected every 30 – 60 s for periods of 2 – 6 hrs. The wavelength was 0.8315 Å. At Brookhaven National Laboratory beamline X7B, similar flow-through experiments were conducted with pH 3 HCl (Fisher Scientific) solution flowing through 0.7 mm quartz capillaries of ~1 mg triclinic Na-birnessite. X-ray diffraction patterns were collected every 30 s for periods of 2 – 4 hrs. The wavelength was 0.9200 Å.

Batch experiments Based on our results from the TR-XRD experiments, batch experiments were conducted imitating the previous flow-through experiments at pH 3. In one set of experiments, ~80 mg of triclinic Na-birnessite powder were placed in ~100 to 150 ml of solution at pH 3 and 0.1 M

51

Pb(NO3)2 (ACS reagent), and allowed to run for 7 days, 14 days, and 112 days. In another set of batch experiments, ~80 mg of hexagonal H-birnessite powder were placed in ~100 to 150 ml of solution at pH 3 and 0.1 M Pb(NO3)2 for 7 days, 14 days, and 112 days. In addition to reactions at pH 3, we also ran batch reactions at pH 5, placing ~80 mg each of triclinic and hexagonal birnessite powders separately in solutions of pH 5 0.1 M Pb(NO3)2 for 112 days. Because a control experiment of hexagonal birnessite in a pH 5 solution was not conducted with TR-XRD flow-through experiments at the synchrotron, a batch experiment with ~80 mg hexagonal birnessite in a pH 5 HNO3 Pb-free solution was also conducted for 5.5 hrs. Solutions were initially adjusted to their appropriate pH with 0.05 M HNO3 (ACS reagent) and 0.1 M NaOH (J.T. Baker). The solutions for each 112-day batch experiment were sampled over time with 0.45 μm cellulose acetate syringe filters (VWR), weighed, and acidified with 68.0-70.0% HNO3. At the end of all batch experiments, the reacted samples were also filtered with 0.05 μm polypropylene membrane filters (Whatman), and the solutions were collected for further analysis. All solutions were diluted with 2% HNO3 (ACS reagent) and analyzed for Pb, Mn, and Na using an X-Series II Thermo Scientific Quadrupole inductively coupled-mass spectrometer (ICP-MS) with Thermo Scientific PlasmaLab software (Materials Characterization Laboratory, Pennsylvania State University). The reacted birnessites were rinsed with 300 ml DI water and left to air dry. Birnessite samples were then characterized with X-ray diffraction at Argonne National Laboratory beamline 13-BM-C for comparison with the results obtained from the flow-through experiments. The

wavelength for the 7-day and 14-day batch experiments was 0.8315 Å. The wavelength for the 112-day batch experiments was 0.8308 Å. The control batch experiment with hexagonal birnessite in a pH 5 Pb-free solution was characterized with a Rigaku D/MAX-RAPID

52

microdiffractometer (Materials Characterization Laboratory, Pennsylvania State University) with a Mo tube source (λ = 0.71 Å). The 2θ values were then converted to the corresponding 2θ values for a 0.8308 Å wavelength for ease of comparison.

Structure Refinements The X-ray diffraction data obtained from the time-resolved experiments and from the 7-, 14-, and 112-day batch experiments were analyzed by Rietveld methods (Rietveld 1969) using the General Structure Analysis System-II (GSAS-II) software (Toby and Von Dreele 2013). The starting structure and atom positions for the Na-birnessite refinements were drawn from Post et al. (2002). The starting structure and atom positions for H-birnessite refinements were drawn from Heaney et al. (2003). Peak profile parameters that captured instrumental broadening were determined by Rietveld analysis of a diffraction pattern generated by a LaB6 standard (NIST SRM 660a) as collected using the same sample holder as was used for birnessite. The background intensities were fitted with a Chebyshev function using 5 to 17 terms. One to two background peaks were added to represent amorphous material contributing to background intensity if necessary. Peak profiles were fitted with a pseudo-Voigt function as parameterized by Thompson et al. (1987), with asymmetry correction by Finger et al. (1994), and microstrain anisotropic broadening terms by Stephens (1999). The background parameters, scale factor, unitcell parameters, and peak profile coefficients were allowed to vary. Final refinements allowed the atomic positions and occupancies to vary. Temperature factors were also varied in the refinements of pure triclinic Na-birnessite and pure hexagonal birnessite, but kept constant for all other refinements.

53

X-ray diffraction simulations CrystalDiffract 1.4.0 (CrystalMaker Software Ltd.) was used to simulate XRD patterns from known crystal structures. The refined structure for hexagonal H-birnessite was built in CrystalMaker 2.3.2 (CrystalMaker Software Ltd.) and imported into CrystalDiffract 1.4.0. The fractional occupancy of the interlayer Mn atom for the structure was adjusted to see the resulting changes in the XRD pattern. The wavelength was set at 0.83154 Å for comparison with other XRD patterns collected at the synchrotron.

Results and Discussion

Intercalation of Pb in triclinic birnessite When triclinic Na-birnessite was added to an aqueous solution at pH 3 with no dissolved Pb, within 30 min it transformed to hexagonal H-birnessite (Fig. 4). As proposed by Silvester et al. (1997), this reaction likely results from the disproportionation of 2Mn3+ to Mn4+ and Mn2+ at low pH. Since the ionic radius of Mn2+ is too large to be accommodated by the octahedral sheet, Mn2+ cations migrated out of the octahedral sites, creating vacancies. The loss of the Jahn-Teller distortions associated with Mn3+ increased the symmetry from triclinic to hexagonal. Indeed, our Rietveld analyses of the hexagonal H-birnessite produced in this fashion yielded an occupancy for Mn of 0.83(2), or a vacancy concentration of ~17% within the octahedral sites (Table 1), with average Mn-Ooct bond lengths of 1.93(6), as is consistent with Mn4+-O bonds (Table 4). In contrast, when triclinic Na-birnessite reacted with a pH 3 solution that contained 0.1 M Pb(NO3)2, the (0 0 l) diffraction intensities decreased to three-fourths of their original intensities within 5 minutes, and all diffraction peaks became strongly asymmetrical within 1.25 hrs (Fig. 5). The peaks shifted to positions characteristic of hexagonal birnessite. Thus, we interpret the

54

combined effects of low pH and 0.1 M dissolved Pb2+ as having transformed well-crystalline triclinic birnessite into a turbostratically disordered hexagonal phase. Longer durations for the reactions did not alter this result. When triclinic Na-birnessite was exposed to 0.1 M Pb(NO3)2 solutions for 7, 14, and 112 days, the final solids were characterized by weaker intensities and asymmetric peaks (Fig. 6). Because of the high asymmetry of the peaks, we were unable to complete a high-quality Rietveld analysis of this phase. Experiments in which triclinic Na-birnessite was placed within pH 5 solutions with and without Pb showed similarities to the experiments conducted at pH 3. However, reactions occurred at slower rates, and full transformations to hexagonal-like phases did not reach completion. For instance, when triclinic Na-birnessite was added to a pH 5 solution with no dissolved Pb2+ in a batch reaction, the triclinic Na-birnessite only partially transformed to hexagonal birnessite after 5.5 hrs (Fig. 7b). Similarly, when triclinic Na-birnessite was added to a pH 5 solution with 0.1 M Pb(NO3)2, the kinetics of transformation to a hexagonal phase were so slow that even after 112 days, the products represented a mix of triclinic and hexagonal-like birnessite (Fig. 7c). ICP-MS analysis of the 112-day batch experiments showed that triclinic Na-birnessite sorbed 0.05 ± 0.02 mol Pb/g birnessite at pH 3 and 0.06 ± 0.02 mol Pb/g birnessite at pH 5 over 112 days, and the rate of sorption decreased significantly after ~14 days (Fig. 8a). The triclinic Na-birnessites initially released Na into solution (Fig. 8b). Surprisingly, triclinic birnessite appeared to begin uptaking Na by day 2 at pH 3 and 5. It is possible that Na is re-entering the interlayer or adsorbing to the surfaces. However, the analyses of Na must be treated with caution, because our calculations could not accurately account for the background Na concentration from the initial pH adjustment of solutions. In contrast to the behaviors of Na and Pb, the Mn concentration in solution increased over time for all birnessites (Fig. 8c). Mn release could be a

55

result of either Mn dissolution, the substitution of Mn with Pb in the interlayer of the structure, or both.

Intercalation of Pb in hexagonal birnessite Surprisingly, for the experiments in which H-birnessite was immersed in solutions with 0.1 M Pb(NO3)2 at pH 3, no obvious changes were noticeable from the TR-XRD patterns after a reaction time of 2 hrs (Fig. 9). However, the XRD patterns from the 7, 14, and 112-day batch experiments at pH 3 indicated a subtle, but definite variation between the initial and final birnessite patterns (Fig. 10). The hexagonal symmetry was maintained, but the intensity of the (1 0 0) peak increased, while the intensity of the (1 0 -1) peak decreased relative to (1 0 0). The same result occurred in the 112-day pH 5 0.1 M Pb(NO3)2 batch experiment with hexagonal Hbirnessite (Fig. 11c). For these experiments in which H-birnessite was exposed to 0.1 M Pb(NO3)2 solutions at pH 3, Rietveld analyses shed light on the cation exchange processes occurring in the interlayer (Tables 1 - 3). Examples of Rietveld refinements for initial, 14-day, and 112-day batch reactions when hexagonal H-birnessite reacted in solutions at pH 3 with 0.1 M Pb(NO3)2 are shown in Fig. 3. Refinement results and atom positions for selected experiments are presented in Tables 1 – 3. Bond distances for hexagonal H-birnessite reacting in solutions at pH 3 with 0.1 M Pb(NO3)2 for 112-days are presented in Table 4. Initially, H-birnessite was refined with Mn in the interlayer along with interlayer O to represent H2O molecules, according to the structure provided by Heaney et al. (2003). Most notably, the fractional occupancy for interlayer Mn at position (0 0 0.68) – above the octahedral vacancy – increased over 112 days from 0.114(3) to 0.190(9), and that for H2O at position (2/3 1/3 0.56) increased from 0.48(1) to 0.83(4) (Fig. 13a, 13b). When interlayer Mn was replaced with Pb in the refinement, the fractional occupancy for interlayer Pb

56

increased from 0.037(1) to 0.060(0) (Fig. 13a). The fractional occupancy for interlayer H2O also increased to 0.63(4) (Fig. 13b). These results provide some support for the entry of Pb into the interlayer, although the position of the Pb may be more disordered than prior models have proposed. The increase in electron density over the vacancy site falls outside the calculated error, but it also seems significant that the occupancy of the water site also increases. The increase in the fractional occupancy for interlayer H2O may indicate that either more Pb or more Mn atoms are occupying the site typically occupied by the H2O molecules, which are located slightly above or below the halfway point on the c-axis. Our Rietveld refinements revealed some unexpected changes in the lattice parameters with the introduction of Pb into the interlayer. Despite the larger ionic radius of Pb2+ (0.98 Å) relative to Mn2+ (0.66 Å) (Shannon 1976), the a axis decreased from 2.865(2) to 2.805(5) Å after hexagonal H-birnessite reacted in a solution at pH 3 with 0.1 M Pb(NO3)2 for 112-days (Fig. 13c), while the c-axis notably decreased from 7.316(2) to 7.119(8) Å (Fig. 13d). These changes coincide with a fairly linear decrease in unit-cell volume with time, consistent with Vegard’s Law if increasing concentrations of Pb2+ substituted into the interlayer as a linear function of time (Fig. 13e). When hexagonal birnessite was exposed to Pb-free solutions at pH 5, the structure showed little transformation, except for a slight decrease in the ratio of the (1 0 0) to (1 0 -1) peak intensities (Fig. 11b), indicating perhaps a decrease in Mn atoms occupying interlayer sites due to dissolution. Simulations on CrystalDiffract support this interpretation, in which the diffraction pattern of hexagonal birnessite with a low Mn(int) occupancy of 0.18 had a decreased ratio of the (1 0 0) to (1 0 -1) relative to the high Mn(Int) occupancy of 0.40 (Fig. 12). Because triclinic Nabirnessite transforms to hexagonal H-birnessite when the pH is less than ~8.2 (Ling et al. 2016), the relative stability of hexagonal birnessite in a pH 5 solution was not surprising.

57

ICP-MS analysis of the 112-day batch experiments with hexagonal H-birnessite indicated that, like triclinic birnessite, hexagonal H-birnessite sorbed Pb over time at both pH 3 and pH 5 with a significant decrease in the rate of sorption after ~14 days (Fig. 14a). At pH 3, hexagonal H-birnessite sorbed 0.05 ± 0.02 mol Pb/g birnessite , and at pH 5, it sorbed 0.06 ± 0.02 mol Pb/g birnessite after 112 days. Like triclinic Na-birnessite, hexagonal H-birnessites also seemed to start uptaking Na by day 2 at both pH 3 and 5, but because the background concentration of Na again was uncertain because NaOH was used for pH adjustment, we once again hesitate to conclude that Na re-sorbs to the structure. The Mn concentration in solution increased over time at both pH 3 and 5 for hexagonal H-birnessite, perhaps due to Mn dissolution or substitution of Mn by Pb into the interlayer (Fig. 14c).

Mechnisms of Pb Sorption for Triclinic and Hexagonal Birnessites After 112 days, triclinic Na-birnessite and hexagonal H-birnessite sorbed essentially equal amounts of Pb at both pH 3 and 5. However, the mechanisms of sorption to the two birnessites appeared to differ, yielding structurally distinct products. Triclinic Na-birnessite transformed to a hexagonal-like phase with extreme stacking disorder. We interpret this result to indicate that triclinic birnessite sorbed Pb through a process of sequential delamination, Pb cation insertion into the interlayer, and relamination as part of the phase transformation to a hexagonal structure. Based on the release of Mn2+ to the solution during this reaction, as revealed by ICPMS (Fig. 8c), we suggest that Pb2+ cations replaced Mn2+ cations that migrated from the octahedral sheet during the disproportionation of Mn3+ to Mn2+ and Mn4+ during the triclinic-tohexagonal phase transformation. Electrostatic repulsions between the lone pair electrons of the sorbed interlayer Pb2+ cations may have been responsible for the turbostratic disorder.

58

Hexagonal H-birnessite, on the other hand, did not experience a symmetry-breaking phase change when it was exposed to dissolved Pb2+ in low pH solutions, but an increase in the ratio of the (0 0 1) to (1 0 -1) peaks argues for an exchange of interlayer Mn2+,3+ for Pb2+. The lack of stacking disorder in these Pb-exchange experiments with hexagonal birnessite strikes us as support that delamination-relamination did not occur here. In the absence of a symmetrychanging cation exchange reaction, the layers did not disorder and reorder as they would in a typical cation-exchange transformation in birnessite. In contrast, hexagonal H-birnessite, which does not undergo phase changes at pH 3 and 5, could be uptaking its interlayer Pb through a diffusive process in which the layers of birnessite did not peel apart for cation exchange, unlike the delamination-relamination model for the cation-exchange of Cs+ into birnessite (Lopano et al. 2009).. Both triclinic and hexagonal birnessite released less Mn than expected based on their chemical formulas, if all interlayer Mn2+,3+ was substituted by Pb2+, but triclinic Na-birnessites released more Mn into solution than did hexagonal H-birnessites. The chemical formula of Nabirnessite, Na0.29(Mn4+0.71Mn3+0.29) O2 •0.75H2O (Post et al. 2002), suggests that 0.747 mmol Mn/g Na-birnessite can be released during transformation to H-birnessite into solution. However, in Na-birnessite batch experiments after 112 days, only 48.91 mol% to 71.17 mol% of interlayer Mn was released for the experiments at pH 5 and pH 3 respectively. These calculations assume that all interlayer Mn2+ is released by cation exchange, when in fact we know that during the transformation from triclinic to hexagonal birnessite at pH 3 and 5, ~1 to 6 ppm of Mn2+ can be released through dissolution, which for ~100 ml of solution would be < 0.011 mmol of Mn2+ (Fleeger 2012). The chemical formula of H-birnessite, Mn0.12(Mn4+0.80□0.20)O2 •0.75H2O (Fischer 2011), suggests that 1.25 mmol Mn/g H-birn can be released from the interlayer if all of the interlayer Mn is exchanged. By the final day of the 112-day batch experiments, however, the

59

hexagonal H-birnessites released only 19.03 mol% to 26.17 mol% of interlayer Mn, again not accounting for release by dissolution. The relatively low release of Mn into solution by both birnessites can be explained by limited Pb sorption into the interlayer, with most of the sorption occurring on surface and edge sites. Comparisons to other Pb sorption studies indicate that hexagonal and/or triclinic birnessites sorb ~0.90 to ~2.13 mmol Pb/g birnessite in pH ranges from 3 to 5 within 12 to 24 hrs (Matocha et al. 2001; Della Puppa et al. 2013; Lee et al. 2013). After 2 days, our results indicate that ~2 mmol Pb/g birnessite is taken up at pH 3 and ~8 mmol Pb/g birnessite is taken up at pH 5 for both triclinic and hexagonal birnessites. Matocha et al. (2001) also found that Pb uptake continued to occur at a much slower rate after 3 hrs of reaction time with ~0.08 mM Mn release/g birnessite into solution after 24 hrs, only slightly lower than our 0.14 mM Mn release/g birnessite after 18 hrs for hexagonal birnessite. Our experiments ran 112 days and showed that Pb uptake into the interlayer continued even after 24 hrs. Limited Pb uptake into the interlayer agrees with our low Pb fractional occupancies from the Rietveld refinements. The Pb cations may sterically hinder other Pb atoms from entering into the interlayer, consequently limiting the number of occupied interlayer sites and the number of cations released from the interlayer. This inhibition of interlayer cation exchange also explains the slow rate of the reaction, a process that was measurable in our experiments even after 14 days, much slower than rates for other metals, such as K, Ba, and Cs, which fully exchange within hours (Lopano et al. 2011; Fleeger et al. 2013).

Acknowledgements This project was funded by NSF Grant 1147728 and NSF Grant EAR-1552211. Experiments were carried out at GeoSoilEnviroCARS (Sector 13), Advanced Photon Source

60

(APS), Argonne National Laboratory, and at the National Synchrotron Light Source, Brookhaven National Laboratory. GeoSoilEnviroCARS is supported by the National Science Foundation Earth Sciences (EAR-1128799) and Department of Energy - GeoSciences (DE-FG0294ER14466). This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357.Use of the National Synchrotron Light Source, Brookhaven National Laboratory, was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-98CH10886. Special thanks to Joanne Stubbs and Peter Eng for their assistance at the beamline and improvements in the design of the flow-through setup.

References Beak, D. G., Basta, N. T., Scheckel, K. G., & Traina, S. J. (2007). Linking solid phase speciation of Pb sequestered to birnessite to oral Pb bioaccessibility: Implications for soil remediation. Environmental science & technology, 42, 779-785. Biscoe, J., Warren, B.E. (1942) An X-Ray Study of Carbon Black. Journal of Applied Physics. 13. 364 – 371. Della Puppa, L., Komárek, M., Bordas, F., Bollinger, J. C., & Joussein, E. (2013). Adsorption of copper, cadmium, lead and zinc onto a synthetic manganese oxide. Journal of colloid and interface science, 399, 99-106. Drits, V.A., Silvester, E., Gorkshkov, A.I., Manceau, A. (1997) Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite: I. Results from X-ray diffraction and selected-area electron diffraction. American Mineralogist. 82. 946 – 961. Filipelli, G.M., Laidlaw, M.A.S. (2010) The Elephant in the Playground: Confronting LeadContaminated Soils as an Important Source of Lead Burdens to Urban Populations. Perspectives in Biology and Medicine. 53. 41-45. Finger, L.W., Cox, D.E., Jephcoat, A.P. (1994) A Correction for Powder Diffraction Asymmetry Due to Axial Divergence. Journal of Applied Crystallography. 27. 892 – 900. Fisher, T.B. (2011) Structural Transformations of Birnessite (δ-MnO2) During Biological and Abiological Reduction. The Pennsylvania State University. 111. Fleeger, C.R., Heaney, P.J., Post, J.E. (2013) A time-resolved X-ray diffraction study of Cs exchange into hexagonal H-birnessite. American Mineralogist. 98. 671 – 679. Fleeger, C.R. (2012) Contaminant Sequestration and Phase Transformation Properties of Birnessite-like Phases (δ-MnO2). The Pennsylvania State University. 134.

61

Gasparatos, D. (2013) Sequestration of heavy metals from soil with Fe-Mn concretions and nodules. Environmental Chemistry Letters. 11. 1 – 9. Golden, D.C., Chen, C.C., Dixon, J.B. (1986) Synthesis of Torokite. Science. 231. 4739. 717 – 719. Grigoriev, A. G., Zhamoida, V. A., Gruzdov, K. A., & Krymsky, R. S. (2013). Age and growth rates of ferromanganese concretions from the gulf of Finland derived from 210Pb measurements. Oceanology, 53(3), 345-351. Heaney, P.J., Post, J.E., Lopano, C.L., and Hanson, J.C. (2003) Hydrogen exchange in Nabirnessite: A time-resolved synchrotron X-ray diffraction analysis. Geological Society of America Abstracts with Programs. 35. 620. Kwon, K., Refson, K., Sposito, G. (2010) Surface complexation of Pb(II) by hexagonal birnessite nanoparticles. Geochimica et Cosmochimica Acta. 74. 6731 – 6740. Laidlaw, M.A.S., Filipelli, G.M. (2008) Resuspension of urban soils as a source of lead poisoning in children: A review and new directions. Applied Geochemistry. 23. 2021 – 2039. Lanson, B., Drits, V.A., Silvester, E., Manceau, A. (2000) Structure of H-exchanged hexagonal birnessite and its mechanisms of formation from Na-rich monoclinic buserite at low pH. American Mineralogist. 85. 826 – 838. Lanson, B., Drits, V.A., Gaillot, A., Silvester, E., Plancon, A., Manceau, A. (2002) Structure of heavy metal sorbed birnessite. American Mineralogist. 1631 – 1645. Lee, S., An, J., Kim, Y., Nam, K. (2010) Binding strength-associated toxicity reduction by birnessite and hydroxyapatite in Pb and Cd contaminated sediments. Journal of Hazardous Materials. 186. 2117 – 2122. Lee, J.H. (2013) An Overview of Phytoremediation as a Potentially Promising Technology for Environmental Pollution Control. Biotechnology and Bioprocess Engineering. 18. 431 – 439. Lee, C., Kim, T., Komarneni, S., Han, S., Cho, Y. (2013) Sorption characteristics of lead cations on microporous organo-birnessite. Applied Clay Science. 83 – 84. 263 – 269. Lopano, C.L., Heaney, P.J., Post, J.E. (2009) Cs-exchange in birnessite: Reaction mechanisms inferred from time-resolved X-ray diffraction and transmission electron microscopy. American Mineralogist. 94. 816 – 826. Lopano, C.L., Heaney, P.J., Bandstra, J.Z., Post, J.E., Brantley, S.L. (2011) Kinetic analysis of cation exchange in birnessite using time-resolved synchrotron X-ray diffraction. Geochimica et Cosmochimica Acta. 75. 3973 – 3981. Matcoha, C.J., Elzinga, E.J., Sparks, D.L. (2001) Reactivity of Pb(II) at the Mn(III,IV) (Oxyhydr)Oxide-Water Interface. Environmental Science & Technology. 35. 2967 – 2972. McKenzie, R.M. (1980) The adsorption of lead and other heavy metals on oxides of manganese and iron. Journal of Soil Research. 18. 61 – 73. Mielke, H.W., Powell, E.T., Gonzales, C.R., Mielke Jr., P.W., Otteson, R.T., Langedal, M. (2006) New Orleans Soil Lead (Pb) Cleanup Using Mississippi River Alluvium: Need, Feasibility, and Cost. Environmental Science and Technology. 40. 8. 2784 – 2789. Nigg, J.T., Knottnerus, G.M., Martel, M.M., Nikolas, M., Cavanagh, K., Karmaus, W., Rappley, M.D. (2008) Low blood lead levels associated with clinically diagnosed attentiondeficit/hyperactivity disorder and mediated by weak cognitive control. Biological Psychiatry. 63. 3. 325 – 331. Post, J.E., Heaney, P.J., Hanson, J. (2002) Rietveld refinement of a triclinic structure for synthetic Na-birnessite using synchrotron powder diffraction data. Powder Diffraction. 17. 3. 218 – 221.

62

Rietveld, H.M. (1969) A Profile Refinement Method for Nuclear and Magnetic Structures. Journal of Applied Crystallography. 2. 65 – 71. Ruby, M.V., Schoof, R., Brattin, W., Goldade, M., Post, G., Harnois, M., Mosby, D.E., Casteel, S.W., Berti, W., Carpenter, M., Edwards, D., Cragin, D., Chappell, W. (1999) Advances in Evaluating the Oral Bioavailability of Inorganics in Soil for Use in Human Health Risk Assessment. Environmental Science and Technology. 33. 21. 3697 – 3705. Scheckel, K.G., Diamond, G.L., Burgess, M.F., Klotzbach, J.M., Maddaloni, M., Miller, B.W., Partridge, C.R., Serda, S.M. (2013) Amending Soils With Phosphate As Means To Mitigate Soil Lead Hazard: A Critical Review Of The State Of The Science. Journal of Toxicology and Environmental Helath, Part B: Critical Reviews. 16. 337 – 380. Shannon, R.D. (1976) Revised Effective Ionic Radii and Systematic Studies of Interatomic Distances in Halides and Chalcogenides. Acta Crysta. A32. 751 – 767. Silvester, E., Manceau, A., Drits, V.A. (1997) Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite: II. Results from chemical studies and EXAFS spectroscopy. American Mineralogist. 82. 962 – 978. Singh, C.K., Rina, K., Singh, R. P., Mukherjee, S. (2014) Geochemical characterization and heavy metal contamination of groundwater in Satluj River Basin. Environmental Earth Science. 71. 201 – 216. Stephens, P.W. (1999) Phenomenological model of anisotropic peak broadening in powder diffraction. Journal of Applied Crystallography. 32. 281-289. Thompson, P., Cox, D.E., Hastings, J.B. (1987) Rietveld Refinement of Debye-Scherrer Synchrotron X-Ray Data from Al2O3. Journal of Applied Crystallography. 34. 210 – 213. Toby, B.H., Von Dreele, R.B. (2013) GSAS-II: the genesis of a modern open-source all purpose crystallography software package. Journal of Applied Crystallography. 46. 544 – 549. van Genuchten, C.M., Peña, J. (2016) Sorption Selectivity of Birnessite Particle Edges: a d-PDF Analysis of Cd(II) and Pb(II) Sorption by δ-MnO2 and Ferrihydrite. Environmental Science: Processes & Impacts. 1 – 37. Wall, A.J., Heaney, P.J., Mathur, R., Post, J.E., Hanson, J.C., Eng, P.J. (2011) A flow-through reaction cell that couples time-resolved X-ray diffraction with stable isotope analysis. Journal of Applied Crystallography. 44. 429 – 432. Wang, Y., Feng., X., Villalobos, M., Tan, W., Liu, F. (2012) Sorption behavior of heavy metals on birnessite: Relationship with its Mn average oxidation state and implications for types of sorption sites. Chemical Geology. 292 – 293. 25 – 34. Weiss, P.T., LeFevre, G., Gulliver, J.S. (2008) Contamination of Soil and Groundwater Due to Stormwater Infiltration Practices, A Literature Review. St. Anthony Falls Laboratory. Minnesota Pollution Control Agency. Wongsasuluk, P., Chotpantarat, S., Siriwong, W., Robson, M. (2014) Heavy metal contamination and human health risk assessment in drinking water from shallow groundwater wells in an agricultural area in Ubon Ratchathani province, Thailand. Environmental Geochemistry and Health. 36. 169 – 182. Zhao, W., Tan, W., Feng, X., Liu, Fan., Xie, Y., Xie, Z. (2011) XAFS studies on surface coordination of Pb2+ on birnessites with different average oxidation states. Colloids and Surfaces A: Physicochemical and Engineering Aspects. 86 – 92. Zhao, W., Cui, H., Liu, F., Tan, W., Feng, X. (2009) Relationship between Pb2+ adsorption and average Mn oxidation state in synthetic birnessites. Clays and Clay Minerals. 57. 5. 513 – 520.

63

Figures

Figure 3-1. Schematic diagrams of synthetic (a) triclinic Na-birnessite (after Post et al. 2002) and hexagonal H-birnessite (after Lanson et al. 2000).

64

Figure 3-2. Experimental set-up for time-resolved X-ray diffraction flow-through experiments.

65

Figure 3-3. Rietveld refinement fits from the batch experiments of (a) the initial hexagonal Hbirnessite, (b) the hexagonal H-birnessite after 14 days of reacting with pH 3 0.1 M Pb(NO3)2, and (c) the reacted hexagonal H-birnessite after 114 days of reacting with pH 3 0.1 M Pb(NO3)2.

66

Figure 3-4. Triclinic Na-birnessite was reacted with a pH 3 HCl solution, and TRXRD patterns were collected every 30 s for 3 hrs. It transformed into hexagonal birnessite within ~30 min.

67

Figure 3-5. Triclinic Na-birnessite was reacted with a 0.1 M Pb(NO3)2 solution at pH 3, and TRXRD patterns were collected every 2 min for ~6.5 hrs. Within 1.25 hrs, it transformed into a turbostratically disordered hexagonal-like phase.

68

Figure 3-6. Overlay of XRD patterns for the 7, 14, and 112-day batch experiments with triclinic Na-birnessite in pH 3 0.1 M Pb(NO3)2.

69

Figure 3-7. XRD patterns of (a) the initial triclinic Na-birnessite, (b) triclinic Na-birnessite (NaB) after reacting in a batch experiment in pH 5 HNO3, Pb-free solution for 5.5 hrs (c) triclinic Nabirnessite after reacting in a batch experiment with 0.1 M Pb(NO3)2 at pH 5 for 112 days, and (d) triclinic Na-birnessite reacting with 0.1 M Pb(NO3)2 at pH 3 in a 6-hr flow-through experiment for comparison.

70

(a)

NaB ++pH M Pb(NO3)2 NaB pH3 0.1 3 Pb(NO3)2

NaB ++pH M Pb(NO3)2 NaB pH5 50.1Pb(NO3)2

mol Pb/g birnessite sorbed

0.09

0.07

0.05

0.03

0.01

-0.01

0

28

56

84

112

Days

(b)

mmol Na in solution/g birnessite

-0.03 9 8 7 6 5 4 3 2 1 0 0

28

56

84

112

(c)

mmol Mn in solution/g birnessite

Days

1 0.9 0.8

Calculated interlayer Mn

0.7 0.6 0.5

0.4 0.3 0.2 0.1 0 0

28

56

84

112

Days

Figure 3-8. Concentrations for Pb, Mn, and Na were measured over time for the 112-day batch experiments with triclinic Na-birnessite and 0.1 M Pb(NO3)2 at pH 3 and 5. (a) The moles of Pb/g birnessite uptaken was calculated by subtracting the measured Pb concentration from the initial Pb concentration/g birnessite. (b) The mmol of Mn/g birnessite and (c) the mmol of Na/g birnessite released into solution were directly measured. Note that the data point for Nabirnessite reacting with a pH 3 0.1 M Pb(NO3)2 solution at 56 days was deleted in all Pb, Mn, and Na plots due to their persistence as outliers.

71

Figure 3-9. TR-XRD results for reacting hexagonal H-birnessite with pH 3 0.1 M Pb(NO3)2. Patterns were collected every 30 s for ~2 hrs. Little change occurred in the hexagonal birnessite.

72

Figure 3-10. Overlay of XRD patterns for the 7, 14, and 112-day batch experiments with hexagonal H-birnessite in pH 3 0.1 M Pb(NO3)2.

73

Figure 3-11. XRD results from batch reactions of hexagonal H-birnessite (HB) at pH 5, including the (a) initial hexagonal H-birnessite, (b) hexagonal H-birnessite reacting in a Pb-free pH 5 HNO3 solution for 5.5 hrs, and (c) hexagonal H-birnessite reacting in a 112-day batch reaction with pH 5 0.1 M Pb(NO3)2.

74

Mn(int) Occ 0.40

Mn(int) Occ 0.18

6.00E-01

5.00E-01

Intensity

4.00E-01

3.00E-01

2.00E-01

_ (1 0 1)

(1 0 0)

1.00E-01

0.00E+00 5

10

15

20

25

30

35

40

45

Two-Theta (λ = 0.83154 Å)

Figure 3-12. Simulations of XRD patterns on CrystalDiffract of two hexagonal H-birnessites with different fractional occupancies of 0.40 and 0.18 for interlayer Mn. The XRD simulation with a high Mn(int) occupancy of 0.40 has a higher (1 0 0) to (1 0 -1) peak intensity ratio compared to the simulation with a low Mn(int) occupancy of 0.18.

75

(c)

2.88

a

2.86

0.2 Pb(int) or Mn(int) fractional occupancy

(a)

2.84 2.82

0.16 2.8

0.12

0

0.08

(d)

0.04

50 Time (days)

100

50 Time (days)

100

7.35 7.3

0 50 Time (days)

100 7.25 c

0

(b)

7.2

O(int) fractional occupancy

0.9

7.15

0.8 0.7

7.1

0.6

0

0.5

0.4

53

(e)

0.3

52

0.2 50 Time (days)

100 51

vol

0

50 49 48 0

50

100

Time (days)

Figure 3-13. Rietveld refinement results from the initial, 7, 14, and 112-day batch experiments using Mn as the interlayer cation, and then Pb as the interlayer cation for the (a) fractional occupancy for interlayer Mn or Pb position, (b) fractional occupancy for interlayer O, representing interlayer H2O, (c) lattice parameter a, (d) lattice parameter c, and (e) volume.

76

(a)

HB +3pH 3 0.1 M Pb(NO pH Pb(NO3)2 - HB 3)2

HB 5 + pH 5 0.1 M Pb(NO pH Pb(NO3)2 - HB 3)2

mol Pb/g birnessite sorbed

0.09

0.07

0.05

0.03

0.01

-0.01

0

28

56

84

112

84

112

Days

(b)

mmol Na in solution/g birnessite

-0.03 9 8

7 6 5 4 3 2 1 0 0

28

56

(c)

mmol Mn in solution/g birnessite

Days 1 0.9

0.8

Calculated interlayer Mn

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

28

56

84

112

Days

Figure 3-14. Concentrations for Pb, Mn, and Na were measured over time for the 112-day batch experiments with hexagonal H-birnessite and 0.1 M Pb(NO3)2 at pH 3 and 5. (a) The moles of Pb/g birnessite uptaken was calculated by subtracting the measured Pb concentration from the initial Pb concentration/g birnessite. (b) The mmol of Mn/g birnessite and (c) the mmol of Na/g birnessite released into solution were directly measured.

77

Tables Table 3-1. Rietveld refinements for the unit cell and atom positions for hexagonal H-birnessite. Hexagonal H-Birnessite Space group: P -3 a = b = 2.865(2) Å Atom positions Mn(oct) O(oct) O(int)/H2O Mn(int) Refinement No. of observations No. of reflections

c = 7.316(2) Å x y 0 0 0.3333 0.6667 0.6667 0.3333 0 0

Diffraction range (°2θ), λ = 0.8308 Å No. of variables R(F2) R wp

z 0 0.136(2) 0.562(6) 0.676(5)

Occupancy 0.83(2) 1.0000 0.48(1) 0.114(3)

1557 17 4.575 – 35.700 38 0.01392 0.02489

Uiso 0.0028 0.0094 0.1362 0.0379

78

Table 3-2. Rietveld refinements for the unit cell and atom positions for hexagonal birnessite with interlayer Mn after a 112-day batch reaction with Pb(NO3)2. Hexagonal Mn-Birnessite Space group: P -3 a = b = 2.831(3) Å c = 7.115(5) Å Atom positions x y Mn(oct) 0 0 O(oct) 0.3333 0.6667 O(int)/H2O 0.6667 0.3333 Mn(int) 0 0 Refinement No. of diffraction points No. of reflections Diffraction range (°2θ), λ = 0.8308 Å No. of variables R(F2) R wp

z 0 0.116(5) 0.580(6) 0.658(7)

Occupancy 0.69(4) 1.0000 0.83(4) 0.190(9)

1567 17 4.575 – 35.900 34 0.01380 0.01332

Uiso 0.0028 0.0094 0.1362 0.0379

79

Table 3-3. Rietveld refinements for the unit cell and atom positions for hexagonal birnessite with interlayer Pb after a 112-day batch reaction with Pb(NO3)2. Hexagonal Pb-Birnessite Space group: P -3 a = b = 2.805(5) Å Atom positions Mn(oct) O(oct) 0.3333 O(int)/H2O 0.6667 Pb(int) Refinement No. of diffraction points No. of reflections

c = 7.119(8) Å x y 0 0 0.6667 0.3333 0 0

Diffraction range (°2θ), λ = 0.8308 Å No. of variables R(F2) R wp

z 0 0.130(3) 0.587(6) 0.645(4)

Occupancy 0.51(3) 1.0000 0.63(4) 0.060(2)

1557 17 4.575 – 35.700 29 0.0221 0.01788

Uiso 0.0028 0.0094 0.1362 0.0379

80

Table 3-4. Bond lengths from final refinements of hexagonal H-birnessite reacting in a 112-day batch reaction with pH 3 0.1 M Pb(NO3)2 using either Mn(int) or Pb(int) as the interlayer cations. Initial H-Birnessite (HB) HB, Mn(int) Refinement, HB, Pb(int) Refinement, Refinement Batch Exp 112 days Batch Exp 112 days Mn(oct) - O(oct) 1.93(6) Mn(oct) - O(oct) 1.83(2) Mn(oct) - O(oct) 1.87(1) - O(oct) 1.86(1) O(oct) - O(oct) - Mn(int) - H2O

2.58(2) O(oct) - O(oct) 2.15(2) - Mn(int) 2.2(4) - H2O

2.3(5) O(oct) - O(oct) 2.28(3) - Pb(int) 2.2(7) - H2O(int)

2.5(3) 2.28(2) 2.0(5)

H2O - H2O - Mn(int)

1.9(5) 1.9(3) 2.4(3)

2.0(5) H2O - H2O 1.72(3) - Pb(int) 2.36(2)

2.0(5) 1.67(2) 2.32(2)

Mn(int) - O(oct) - Mn(oct) - Mn(int)

2.15(2) Mn(int) - O(oct) 2.4(4) - Mn(oct) 2.6(7) - Mn(int)

2.29(3) Pb(int) - O(oct) 2.4(5) - Mn(oct) 2.3(1) - Pb(int) - Pb(int)

2.28(2) 2.53(3) 2.1(5) 2.806 3.5(3)

H2O - H2O - Mn(int)

Note: H2O molecules were refined as O(int).

81

Chapter 4 Fourier-transform infrared spectroscopy (FTIR) analysis of triclinic and hexagonal birnessites

Abstract The characterization of birnessite structures is particularly challenging for poorly crystalline materials of biogenic origin, and a determination of the relative concentrations of triclinic and hexagonal birnessite in a mixed assemblage has typically required synchrotron-based spectroscopy and diffraction approaches. In this study, Fourier- transform infrared spectroscopy (FTIR) is demonstrated to be capable of differentiating synthetic triclinic Na-birnessite and synthetic hexagonal H-birnessite. Furthermore, IR spectral deconvolution of peaks resulting from Mn-O lattice vibrations between 400 and 750 cm-1 yield results comparable to those obtained by linear combination fitting of synchrotron X-ray absorption fine structure (EXAFS) data when applied to known mixtures of triclinic and hexagonal birnessites. Density functional theory (DFT) calculations suggest that an infrared absorbance peak at ~1628 cm-1 may be related to OH vibrations near vacancy sites. The integrated intensity of this peak may show sensitivity to vacancy concentrations in the Mn octahedral sheet for different birnessites.

Introduction Manganese (Mn) oxides form in a wide variety of natural environments, sometimes as nodules in fresh-water lakes and oceans and also as coatings on soil and sediment particles (Post 1999). Within the Critical Zone, Mn oxides are involved in the cycling of heavy metals. Pb, Ni, and Zn, for example, commonly sorb to surfaces or intercalate within the interlayers and tunnels of naturally occurring manganates (Chao and Theobald 1976; Usui and Mita 1994). Mn oxides

82

also promote redox reactions, and have been implicated in the oxidation of Cr3+ to Cr6+, As3+ to As5+, and Tl1+ to Tl3+, among other redox active metals in natural environments (Bartlett and James 1979; Murray and Dillard 1979; Oscarson et al. 1981; Scott and Morgan 1996; Kay et al. 2001; Fandeur et al. 2009; Lafferty et al. 2011; Peacock and Moon 2012; Kazakis et al. 2015). Although a wide variety of Mn oxide phases have been described in natural environments, the layered Mn oxides in the birnessite family have been the target of numerous studies for several reasons. Birnessite-like phases are among the most commonly occurring and chemically active natural Mn oxides (Post 1999; Weaver and Hochella 2003) and exists in two phases, triclinic and hexagonal birnessite, both of which are layer structures. Post and Veblen (1990) and Post et al. (2002) describe a synthetic triclinic Na-birnessite, with formula Na0.58(Mn4+1.42Mn3+0.58)O4 • 1.5H2O, whose octahedral sheets are completely filled with Mn3+and Mn4+ cations. Their Rietveld refinements suggest that Mn3+ occupies ~29% of the octahedral sites and Mn4+ the remaining ~71%. Hydrated Na+ cations partially occupy sites in the interlayer (Fig. 1a). In contrast, Silvester et al. (1997) characterized synthetic hexagonal H-birnessite, with proposed formula H0.33Mn3+0.111Mn2+0.055(Mn4+0.722Mn3+0.111☐0.167)O2, and they argue that the octahedral sheets contain ~72% Mn4+ cations, ~11% Mn3+ cations, and ~17% vacancies. Mn2+, Mn3+, and H+ cations are located in the interlayer over vacancy sites (Fig. 1b). For the synthetic hexagonal H-birnessite used in this study, Ilton et al. (2016) determined the sample to consist of 68% Mn4+, 22% Mn3+, and 10% Mn2+ using X-ray photoelectron spectroscopy by fitting the Mn3p peaks. These differences in Mn valence states and vacancy concentrations are integral in establishing their stability with respect to pH and in predicting their chemical reactivity in natural environments (Ling et al. 2015). The characterization of birnessite-like phases, however, is complicated by poor crystallinity and/or small particle size. The most common method of material identification, Xray diffraction (XRD), is especially challenged by the absence of long-range order in many Mn

83

oxide samples, requiring complementary techniques such as transmission electron microscopy, electron microprobe analysis, and IR spectroscopy for proper identification (Post 1999). X-ray absorption spectroscopy (XAS) has instead become a standard method for differentiating triclinic and hexagonal birnessite. Analysis of the EXAFS region has been used extensively for phase identification along with the determination of short-range structure (Jürgensen et al. 2004; Webb et al. 2005; Villalobos et al. 2006; Bargar et al. 2009; Saratovsky et al. 2009; Learman et al. 2011; Santelli et al. 2011). EXAFS also has been applied to quantify mixtures of triclinic and hexagonal birnessite through linear combination fitting (LCF) of the EXAFS region using endmember standards of triclinic and hexagonal birnessite (Learman et al. 2011, 2013; Santelli et al. 2011; Zhao et al. 2016). Because this approach requires a synchrotron source, data collection can be logistically difficult. Also, the significance of the quantization is strongly dependent upon the choice and quality of the characterization of the end-member phases. Clearly, it is desirable to utilize relatively simple, laboratory based available techniques that can be applied in parallel with synchrotron-based spectroscopy and diffraction in the structural characterization of Mn oxide minerals. Fourier-transform infrared spectroscopy (FTIR) has long been used to identify Mn oxides (Potter and Rossman 1979; Krishnamurti and Huang 1988; Parikh and Chorover 2005; White et al. 2009; Palchik et al. 2014). A few scientists have attempted to quantify chemical properties of birnessite using FTIR, making efforts to attribute specific IR bands to certain vibrations, such as OH, H2O, and Mn-O vibrations (Tsyganenko 1975; Potter and Rossman 1979; Yang and Wang 2001; Julien et al. 2004). Other FTIR investigations have examined differences in IR spectra before and after Pb sorption reactions with birnessite (Zhao et al. 2012), acid treatments (Kang et al. 2007), crystallization reactions (Luo et al. 1998), cation-exchange (Golden et al. 1987), and birnessite transformations to the tunnelstructured todorokite (Atkins et al. 2014; Zhao et al. 2015). To the best of our knowledge, no

84

studies have focused on understanding and applying differences in IR spectra for triclinic and hexagonal birnessites. In this study, we employed FTIR to differentiate between analogues for natural birnessites, examining a triclinic and a hexagonal birnessite synthesized and characterized by Rietveld analysis of X-ray diffraction patterns. The data obtained from these samples were used to determine the relative concentrations of triclinic and hexagonal birnessites in known mixtures. The results of FTIR analyses were compared with those from EXAFS and XRD analyses to explore the consistency of these techniques for quantification of the birnessite mixtures. By examining synthetic birnessite-like phases, we hope to lay the groundwork for identifying natural triclinic and hexagonal birnessites with FTIR, which we plan to explore in a future study. Finally, we explored the usefulness of density functional theory (DFT) to calculate, and help interpret, infrared absorbance behavior for the model structures, and for hexagonal birnessite, with and without the vacancies that populate the octahedral sheets.

Methods

Materials synthesis Synthesis of standard triclinic birnessite. Triclinic Na-birnessite was synthesized according to the procedure described in Golden et al. (1986). A 200 ml solution of 0.5 M MnCl2 (Mallinckrodt Baker) was mixed with 250 ml of 5.5 M NaOH (J.T. Baker). The mixture was oxygenated through a glass frit for ~5 hrs at a rate of 1.5 L/min. The precipitate was divided evenly and centrifuged in 14 centrifuge tubes. The solution was decanted and replaced with pH 6.49 deionized (DI) water to rinse. The rinse cycle was repeated five times. Na-birnessite was stored in ~350 ml DI water until experimental use. For experiments, aliquots of Na-birnessite

85

were filtered with a 0.05 μm Nuclepore Track-Etched polycarbonate membrane filter (Whatman), rinsed three times with 100 ml DI water, and left to air-dry at room temperature. Synthesis of standard hexagonal birnessite. Our hexagonal birnessite standard was synthesized by reacting ~100 mg of dried triclinic Na-birnessite in 100 ml of 0.01 M HCl for 24 hrs as described in Giovanoli et al. (1970). The H-birnessite then was filtered with a 0.05 μm Nuclepore Track-Etched polycarbonate membrane filter (Whatman), rinsed three times with 100 ml DI water, and left to air-dry. X-ray diffraction with a Rigaku D/MAX-RAPID

microdiffractometer (Smithsonian Institute, Mineral Sciences Department) using a Mo tube source (λα1α2 = 0.71069 Å) confirmed the synthesis of both triclinic and hexagonal birnessite. Known mixtures of standard triclinic and hexagonal birnessite. To test the effectiveness of linear spectral unmixing for quantitative analysis of birnessite mixtures, we stirred the triclinic Na-birnessite and the pH 2 hexagonal birnessite standards to prepare samples with weighted ratios of 25:75, 50:50, and 75:25. All mixtures were then lightly ground under acetone in a mortar and pestle to disaggregate and blend the samples. Two sets of mixtures were created. The first suite of mixtures was analyzed by FTIR, XRD, and EXAFS. The second suite was analyzed by FTIR only.

X-ray diffraction (XRD) and structure refinements The structures of the triclinic and hexagonal birnessites were refined using powder X-ray diffraction (XRD) and the Rietveld method. All samples were ground in an agate mortar under acetone to disaggregate clumps. For XRD analysis, ~2 mg of sample were mounted on glass fibers using superglue, which in turn were loaded into a Rigaku D/MAX-RAPID

microdiffractometer with an imaging plate detector and a Mo tube. Samples were rotated 360° around the phi axis at 1° s-1 during data collection with a 10 min exposure time.

86

Rietveld refinements (Rietveld, 1969) were completed for all birnessite samples using the General Structure Analysis System-II (GSAS-II) software (Toby and Von Dreele 2013). The starting atom positions for the triclinic Na-birnessite refinements were drawn from Post et al. (2002), and those for the H-birnessite refinements were taken from Fischer (2011). Peak profile parameters that described instrumental broadening were determined by Rietveld analysis of a diffraction pattern generated by a LaB6 standard (NIST SRM 660a) as collected using the same sample holder as was used for birnessite. The background parameters, scale factor, unit-cell parameters, and peak profile coefficients were allowed to vary. The background intensities were fitted with a Chebyshev function using 6 terms. Diffraction peak profiles were fitted with a pseudo-Voigt function as parameterized by Thompson et al. (1987), with asymmetry correction by Finger et al. (1994), and microstrain anisotropic broadening terms by Stephens (1999). Final refinements for the pure phases only allowed the atomic positions, occupancies, and temperature factors to vary. For all refinements, the (0 0 1) peak was omitted due to its high relative intensity and problems with integrating the diffraction images near the direct beam. Because of the similarity of the structures of triclinic and hexagonal birnessite, and consequent strong correlation among certain structural parameters, it can be challenging for Rietveld analysis to extract accurate phase concentrations from X-ray diffraction patterns of mixtures of triclinic and hexagonal birnessite. Therefore, we developed a structure refinement strategy that yielded good results for the 50:50 mixture that was then applied to the 75:25 and 25:75 triclinic and hexagonal birnessite mixtures to test its robustness for other compositions. The approach was similar to that used for single-phase birnessite with the following exceptions. Phase fractions rather than single scale factors were refined. Atomic positions, occupancies, and temperature factors were held constant. Instead of the general anisotropic microstrain broadening model, an isotropic microstrain term initially was varied, and then a uniaxial model was introduced. The equatorial and axial microstrain parameters were then refined separately for each

87

phase and held constant. The Lorentzian Y peak profile parameter was varied after the microstrain terms had been refined, due to initial instability.

X-ray absorption spectroscopy/extended X-ray absorption fine structure (XAS/EXAFS) For X-ray absorption spectroscopy, samples were ground under acetone in a mortar and pestle, and sieved with a 425 or 500 mesh sieve into a thin layer onto kapton tape. The kapton tape was then folded over to seal the sample. Manganese K-edge XAS spectra were collected using a synchrotron source at Beamline 12-BM of the Advanced Photon Source (APS), Argonne National Laboratory, using a Si(111) double-crystal, fixed exit monochromator and a double mirror system (flat plus torroidal) with an energy cutoff of 23 keV. The pre-edge peak of a Mn foil was used for energy calibration (6539 eV). Fluorescence data were collected with a 13element Ge detector and Cr(III) foil in front of the Ge detector to eliminate scatter. Three to six scans were collected per sample at room temperature from -200 eV to about +800 eV around the Mn K-edge (6539 keV). During data collection, the peak positions and line forms in the near edge region (XANES) were examined to check for photochemical reduction with successive scans, and no changes were observed. Care was also taken to avoid self-absorption, and samples were re-prepared if dampening of EXAFS oscillations was observed. Analyses of spectra were conducted using the ATHENA software (Ravel & Newville 2005). XAS spectra were calibrated using a Mn foil, averaged, background-subtracted, normalized, and deglitched if outlier points were present. Analysis of the Mn K-edge EXAFS region was used for the structural identification and quantification of phase fractions for samples (Webb et al. 2005; Villalobos et al. 2006; Saratovsky et al. 2009; Feng et al. 2010). The χ(k) spectra were converted to k (Å-1) (Sayers & Bunkers, 1988). The resulting χ(k) data were k2weighted and analyzed using the k-range from 2.3 to 11.3 Å-1. Principal component analysis

88

(PCA) was used to determine the number of components representing the entire data set (Manceau et al. 2002). The linear combination fitting (LCF) of χ(k) spectra collected from synthetic birnessites yielded phase fractions of triclinic and hexagonal birnessite (Webb et al. 2005; Villalobos et al. 2006; Saratovsky et al. 2009; Feng et al. 2010). All weights were set between 0 and 1, and all combinations were fit with at most 2 standards. Weights of standards were not forced to sum to 1. The synthetic triclinic Na-birnessite and synthetic pH 2 hexagonal birnessite were used as standards to fit the EXAFS data.

Fourier-transform infrared spectroscopy (FTIR) FTIR spectra data collection and processing. Samples were ground under acetone in a mortar and pestle and sieved through a 325 mesh sieve, so particle sizes were less than 44 µm, but individual crystal diameters were ~300 nm based on SEM observations. Then, 0.5 to 1 mg of Mn oxide sample was milled with ~250 mg KBr using a SPECAC ball mixing mill for 1-2 min, and pressed into a pellet. Up to six pellets of each sample were prepared to test reproducibility of the analyses. Vibrational spectra were collected on a Nicolet 6700 Analytical FTIR Spectrometer from 400 to 4000 cm-1. The resolution was set at 3.86 cm-1 and 64 scans were coadded for each spectrum. The Omnic 8 software (Nicolet) was used to view data during data collection. In order to ensure similar data collection conditions, FTIR spectra of pure triclinic Na-birnessite and hexagonal H-birnessite standards were collected within ~1 hr of data collection for the other samples. Background spectra for samples showed no significant shift within that time. In order to quantify the amount of hexagonal and triclinic birnessite in a sample, FTIR spectra were normalized by setting the intensities of the highest peaks in the 400 to 750 cm-1 region to 1, and then linear backgrounds were subtracted from the spectra using MATLAB R2015a (Mathworks). For the pure triclinic and hexagonal birnessite samples, the spectra were

89

re-scaled after background subtraction with the maximum intensities reset to 1 to adjust for any decreases in the maximum intensity that may have occurred during background subtraction. The 50:50 mixture of triclinic and hexagonal birnessite was also used to develop FTIR reference intensity ratios (RIRs) for triclinic and hexagonal birnessite, since the intensity of the strongest peak in the hexagonal birnessite spectrum was 90% of that for the triclinic birnessite spectrum. Because of the small relative difference, however, application of this correction factor did not result in significant changes to the fits. Linear spectral unmixing of normalized, background-subtracted FTIR spectra was performed using the Least Squares Fitting algorithm with SIXpack (Webb 2005). We used the observed spectra in the range of 400 to 750 cm-1 for spectral decomposition of birnessite mixtures, fitting the spectra with components obtained from our pure triclinic birnessite and hexagonal birnessite standards (Thomson and Salisbury 1993; Ramsey and Christensen 1998). Components were not summed to 1 to allow slight adjustments to scaling from not re-normalizing sample spectra after background subtraction, which was performed only for the pure standards. Peak deconvolution was also performed for peaks in the 1300 to 1800 cm-1 range using OPUS 7.0 spectroscopy software (Bruker). The peaks were deconvoluted with the Levenberg-Marquardt algorithm using a combination of Gaussian and Lorentzian peak shapes. Results of phase fractions from spectral unmixing are reported as volume fractions. The known weight fractions of the mixtures measured during sample preparation were also converted to volume fractions for comparison by using the densities determined for triclinic and hexagonal birnessite from Rietveld analyses. DFT Computations of IR spectra. Using Materials Studio (Accelrys Inc., San Diego, CA), three models (Fig. 2) of birnessite nanoparticles were created as clusters to represent a triclinic birnessite nanoparticle, a hexagonal birnessite nanoparticle free of defects, and a hexagonal birnessite with a vacancy site. The triclinic birnessite nanocluster was built with 7 Mn

90

atoms bonded together with 12 O atoms and terminated with 12 H2O groups. The nanocluster represented 5 Mn4+ and 2 Mn3+ atoms to yield a mean formal charge of +3.71 for Mn, similar to that of triclinic Na-birnessite (Post and Veblen 1990) and producing a final formula of Mn7O24H22 (Fig. 2a). The defect-free hexagonal birnessite nanocluster consisted of 7 Mn atoms, including 4 Mn4+ and 3 Mn3+, bonded together with 12 O atoms and terminated with 11 H2O and 1 OH group, giving a final formula of Mn7O24H23 (Fig. 2b). This cluster produced a mean formal charge of +3.57 to approximate the total average valence (+3.66) in hexagonal H-birnessite according to Silvester et al. (1997). The hexagonal birnessite nanocluster with a vacancy had 7 Mn atoms and 11 O atoms with a mean formal charge of +3.71 to represent 6 Mn4+ atoms and 1 Mn2+ adsorbed near a vacancy site (Fig. 2c). Its chemical formula upon addition of terminating H2O and OH groups was Mn7O26H26. Energy minimizations for the birnessite nanoclusters were performed with Gaussian 09 (Frisch et al. 2010). Mn atoms were initially frozen and then permitted to relax after minimization of the O and H atomic positions. The B3LYP exchange-correlation functional (Lee et al. 1988; Becke 1993) was used along with the 6-31G(d,p) basis set (Rassolov et al. 2001). After energy minimizations were completed, frequency calculations were performed. Model harmonic frequencies were scaled by 0.96 for comparison with experimental data to account for basis set effects, anharmonicity, and approximations in electron correlation based on the approximate scale factor of 0.9614 from Scott and Radom (1996) for B3LYP/6-31G(d). Normal modes were visualized using the program Molden (Schaftenaar and Noordik, 2000).

91

Results & Discussion

Distinguishing triclinic and hexagonal birnessite with XRD and FTIR We present here the detailed structure refinement results for the triclinic and hexagonal birnessites used in this study to fully characterize their structures. Rietveld analyses of triclinic Na- and hexagonal H-birnessite confirmed the differences in crystal structure previously established for these phases (Silvester et al., 1997; Post et al., 2002). The refined fit parameters for these analyses are included in Tables 1 and 2, and the excellent fits between the observed and calculated XRD patterns are illustrated in Figure 3. Birnessites prepared using different methods or having different compositions may not have identical structures, making comparisons among studies sometimes confusing, or impossible. We also present the distinctions between triclinic and hexagonal birnessite apparent from FTIR in Figure 4. Most strikingly, triclinic Na-birnessite displays three prominent absorbances at 418, 478, and 511 cm-1, and a low-intensity peak at 639 cm-1, whereas hexagonal H-birnessite shows two broad, high-intensity peaks at 440 and 494 cm-1, and a low-intensity peak at 659 cm-1 (Fig. 4c, Table 3). FTIR absorbances in the 400 to 650 cm-1 region represent contributions from Mn-O bond vibrations (Julien et al. 2004). More subtle distinctions are also visible for peaks in the regions of ~2950 to ~3670 cm-1 and ~1330 to ~1725 cm-1 (Fig. 4) that represent bending and stretching of H2O and OH bonds, although it is important to note that absorbed water also contributes to these regions (Potter & Rossman 1979). Interestingly, in the region from ~2950 to ~3670 cm-1, the FTIR triclinic Na-birnessite shows more distinct peaks than hexagonal Hbirnessite, indicating more ordering of water and possibly multiple water sites.

92

Quantitative analysis of known triclinic-hexagonal birnessite mixtures: FTIR, EXAFS, and XRD Quantitative analyses for known mechanical mixtures of end-member birnessite standards were performed to compare the relative abilities of FTIR, XRD, and EXAFS to reproduce the known phase fractions. FTIR spectral analyses yielded generally good results, with a maximum average error factor of 0.0021 (Table 4). The calculated ratios of triclinic-tohexagonal birnessite were within 13 vol% of the known value for the first 25:75 triclinic-tohexagonal birnessite mixture (Mix 1 22:78 by vol%), and within 3 vol% for all other mixtures. The quality of the quantitative analysis obtained by FTIR was slightly worse than the results obtained by Rietveld analysis of XRD data for birnessite mixtures (Table 4). Our XRD protocol for quantitative analysis yielded phase fractions within 3 wt% of the known values (or 4 vol% for better comparison to FTIR and EXAFS results). Refinements results for the three triclinic-to-hexagonal birnessite mixtures are shown in Fig. 6; a strong 1:1 correlation exists between the FTIR and XRD results (Fig. 7a). Because many studies employ EXAFS to quantify the relative concentrations of hexagonal and triclinic (or “pseudo-orthogonal”) birnessite in their samples (Learman et al. 2011, 2013; Santelli et al. 2011; Zhao et al. 2016), we applied linear combination fitting (LCF) of the EXAFS region using end-member standards to our mechanical mixtures of birnessite (Table 4). Results from LCF gave errors within 7 vol% for the known physical mixtures (Table 4). When comparing the results from EXAFS with the FTIR and XRD results, we find that strong 1:1 correlations exist among the various methods (Fig. 7b, 7c). In general, Rietveld refinements gave the most accurate results, although spectral fitting of FTIR and LCF of EXAFS still yielded reasonable values.

93

Comparison of observed FTIR spectra to DFT calculations DFT calculations of birnessite-like clusters were performed to gain insights into the disparities between the observed FTIR patterns of triclinic and hexagonal birnessite. An unavoidable discrepancy between our modeled IR spectra and the observed patterns involved a shift in frequencies (Cramer 2004). The harmonic oscillator approximation used in our calculation of vibrational frequencies has an infinite number of vibrational levels that are all equally spaced. In contrast, the variable energy changes during transitions to other vibrational levels in real bonds are finite. Consequently, we considered both absorption frequencies and relative IR intensities when correlating calculated and observed peaks. Mn-O vibrations. Over the spectral range representing Mn-O bond vibrations, relative peak positions for the calculated FTIR patterns were shifted to higher values by ~135 cm-1 for the vacancy-free nanoclusters relative to the corresponding peaks observed for triclinic birnessite (Fig. 8). Correlative peaks between the spectrum for the nanocluster with a vacancy site and that displayed by hexagonal birnessite were shifted upward by ~160 cm-1. Specifically, the three calculated high-intensity peaks at 559, 611, and 642 cm-1 (Table 5) of the triclinic birnessite-like nanocluster (Mn7O24H22), labeled T1, T2, and T3, correlated to observed peaks at 418, 478, and 511 cm-1 in actual triclinic Na-birnessite (Table 3, Fig. 8c). Similar to the FTIR spectrum for actual hexagonal birnessite, the hexagonal birnessite nanoclusters exhibited two rather than three high-intensity peaks over the range of Mn-O vibrations labeled H2 and H3. The calculated pattern revealed peaks at 578 and 636 cm-1 for the hexagonal birnessite-like cluster with no vacancy (Mn7O24H23), and at 612 and 650 cm-1 for the hexagonal birnessite-like cluster with a vacancy (Mn7O26H26) (Fig. 8c). Correlative peaks for the observed hexagonal birnessite FTIR spectra occurred at 440 and 494 cm-1 (Fig. 4c). A calculated peak at 463 cm-1 in both hexagonal birnessite nanoclusters (labeled H1 in Fig. 8) was not seen in the observed data, most likely due

94

to the energy cutoff of the FTIR spectrometer at 400 cm-1 (when the ~160 cm-1 peak shift is factored into the expected peak position). For the H2 and H3 peaks, the observed hexagonal birnessite had peak ratios similar to those of the calculated hexagonal birnessite with a vacancy site. The calculated Mn-O vibrations associated with bonds not immediately attached to the central Mn atom in the birnessite nanoclusters fell within the range of 761 to 908 cm-1 (Table 5b). These peaks were the result of Mn-O vibrations moving in conjunction with OH vibrations on the edges of the nanoclusters (Fig. 9b), in contrast to the vibrations associated with the interior Mn-O bonds, which were calculated to lie within the 384 to 758 cm-1 range (Fig. 9a). The number of peaks calculated over the higher range exceeded the number of peaks observed in the range of ~550 to ~900 cm-1 for actual triclinic and hexagonal birnessite. We attribute this discrepancy to the fact that the nanocluster models included disproportionately more water molecules and OH bonds relative to Mn-O bonds than the synthetic birnessites analyzed (Fig. 2). Our calculations suggest two possibilities for why triclinic birnessite exhibits three prominent peaks over the Mn-O vibrational range whereas hexagonal birnessite displays two. One possibility attributes the differences to Jahn-Teller distortions in the octahedral sheets of triclinic Na-birnessite, in which the symmetry-breaking associated with the distortions may account for the additional peak in triclinic birnessite (Drits et al. 1997). Although calculated MnO bond lengths in all three nanoclusters range from 1.88 to 2.23 Å (Fig. 2), the presence of JahnTeller distortions is not immediately obvious based on the relative locations of short and long bond lengths. Bond lengths in the triclinic birnessite nanocluster point to the possibility of distortions, with the 2.11 and 2.16 Å bond lengths located on opposite sides of the central Mn atoms and the short bond lengths doing the same. However, the hexagonal birnessite without a vacancy and the hexagonal birnessite with a vacancy site exhibit equally wide ranges of Mn-O bond lengths from 1.88 to 2.23 Å surrounding their central Mn atom. The uncertainty in whether

95

Jahn-Teller distortions exist in the DFT models does not rule out whether distortions contribute to the differences between the observed FTIR spectra in the Mn-O range. A second explanation for the difference in the FTIR spectra lies in the possibility that structural disorder in birnessite could lead to band broadening and peak overlap. The DFT peaks in the synthetic spectra for the triclinic birnessite nanocluster Mn7O24H22 are combinations of 5 modes: 492-507 Mn-Obr+ Mn3-Obr, 531-537 Mn-Obr symm, 559-571 Mn-Obr asymm, 588-630 Mn-Obr+ Mn3-Obr, and 642-646 Mn-Obr+ Mn3-Obr. For Mn7O24H23, the hexagonal birnessite nanocluster without a vacancy, the modes consist of combinations of 530-544 Mn-Obr symm, 557-566 Mn-Obr+ Mn3-Obr, 577-622 Mn-Obr+ Mn3-Obr, 630-637 Mn-Obr+ Mn3-Obr, and 649-664 Mn-Obr+ Mn3-Obr. In the hexagonal birnessite nanocluster with a vacancy, Mn7O26H26, the modes include combinations of 551-560 Mn-Obr+ Mn3-Obr, 580-597 Mn-Obr+ Mn3-Obr, 612-641 MnObr+ Mn3-Obr, 663-680 Mn-Obr+ Mn3-Obr, and 763 on the defect site. Exemplary depictions of each vibrational mode can be found in Supplementary Information A. Each individual mode listed often contained multiple vibration types, such as in the case of Mn-Obr + Mn3-Obr, complicating the distinctions between modes. Furthermore, most modes are separated by a range of 8 to 22 cm-1, suggesting that significant peak overlap and peak broadening could prevent proper identification of peaks in the observed spectra. H2O and OH vibrations. The region from 1534 to 3732 cm-1 in the calculated spectra reflects vibrations related to H2O and OH movement (Table 5c, 5d), and the high ratio of OH bonds to Mn-O bonds in the clusters also accounted for additional peaks in the calculated relative to the observed spectra. Specifically, the hexagonal birnessite cluster with a vacancy site exhibited additional peaks at ~2152 cm-1 and at ~2720 cm-1, and these peaks have no correlations in the observed spectra. These vibrations were all related to OH vibrations surrounding the vacancy site, but were likely not present in high enough concentrations throughout the actual birnessite sample to be observed in the FTIR spectra.

96

In the observed spectra, the region from ~1330 to ~1725 cm-1 represents vibrations from H2O bending (Potter and Rossman 1979). In the corresponding region in the simulated IR spectra, 4 distinguishable peaks were calculated over the ~1498 to ~1661 cm-1 range when a vacancy site was present as opposed to only 2 peaks when all octahedral sites were occupied (Fig. 8b). The 2 peaks calculated for these vacancy-free nanoclusters were associated with H2O bending motions on the edge of the nanocluster similar to that depicted in Fig. 9c. In the hexagonal birnessite with a vacancy, the motions related to each of the 4 peaks in the calculated spectra are depicted in Fig. 10. As seen in Figure 10a, the peak at 1653 cm-1 was related to H2O bending around the vacancy site. The observed peaks in the FTIR spectra in the region from ~1300 to ~1800 cm-1 may correspond to the calculated peaks from H2O bending, although the broadness of the peaks prevents clear identification of the quantity and positions of peaks (Fig. 8b). More important, the peak splitting seen in the hexagonal birnessite nanocluster with a vacancy site may provide information for future studies concerning vacancy concentrations in samples. Because the absorbance, or more accurately, the integrated measurement of the peak area, is linearly related with concentration and sample thickness according to the Beer-Lambert law (Libowitzky and Beran 2004), the integrated area of the peak at 1653 cm-1 in the DFT calculations could be directly proportional to the number of OH vibrations surrounding a vacancy site in a specific birnessite structure. If the peak(s) in the observed spectra at ~1628 cm-1 is correlated with the 1653 DFT peak, then perhaps the intensity of that peak may be related to the vacancy-related OH vibrations, and we may be able to use the peak area in the observed birnessite samples to compare relative concentrations of this vacancy-related OH vibration. Although this study did not have a sufficiently large sample set of birnessites with varying vacancy concentrations to test this relationship, it is noteworthy that the peak area at ~1628 cm-1 for triclinic Na-birnessite was 1.36 absorbance units • cm-1, whereas the

97

corresponding peak area for hexagonal H-birnessite was 5.41 absorbance units • cm-1 as determined from peak-fitting (Fig. 11). The larger peak area observed experimentally for hexagonal H-birnessite than triclinic Na-birnessite (Lanson et al. 2000) suggested that the ~1628 cm-1 peak(s) may be sensitive to vacancy concentration, especially because Rietveld analyses indicated that the Mn octahedral sheets in hexagonal H-birnessite have an occupancy of only ~0.80, whereas the octahedral sites in triclinic Na-birnessite are fully occupied. At the same time, the broadness of the observed peak could potentially be fit by several smaller peaks, and no clear conclusions concerning the correspondence of the 1628 cm-1 peak(s) can be made with the calculated 1653 cm-1 peak in the hexagonal birnessite nanocluster with a vacancy site until additional samples are analyzed. Peaks in the ~1300 to ~1800 cm-1 region for both triclinic Na-birnessite and hexagonal Hbirnessite had similar peak positions and shapes (Fig. 11), although at lower intensities for triclinic Na-birnessite. This observation implies the presence of a small number of vacancyrelated vibrations in both materials, as seen in the peaks calculated for the birnessite model with a vacancy site. Rietveld analysis of XRD data, however, revealed no vacancies for triclinic Nabirnessite, suggesting that perhaps FTIR is a more sensitive method for determining the concentration of vacancies in a sample.

Conclusions FTIR serves as a relatively simple, in-house technique for differentiating between synthetic triclinic and hexagonal birnessite samples. Linear spectral unmixing can be used to calculate fractions of two birnessite end-members with a degree of accuracy that is comparable to EXAFS. Rietveld analyses of X-ray diffraction data remains the most accurate means of quantifying mixtures of triclinic and hexagonal birnessite, but this technique is only applicable to

98

reasonably well crystalline samples. Because the two phases differ in their Mn oxidation state ratios and vacancy concentrations, determining their relative proportions in a sample can help predict sorption and redox reactions. In the future, FTIR could potentially characterize very poorly crystalline, natural birnessite samples whose triclinicity or hexagonality cannot be easily quantified using XRD, and perhaps provide insight into the concentration and distribution of vacancies in the structure. Discrepancies that exist between natural birnessites found in different formation environments could also be studied without a trip to the synchrotron. In general, FTIR acts as a convenient method of comparing synthetic and natural birnessites, and for analyzing the evolution of birnessite-like phases.

Acknowledgements Funding for this work was provided by NSF Grant EAR-1147728 and EAR-1552211 and the Committee on Institutional Cooperation (CIC) and Smithsonian Institution Fellowship. This research also utilized samples from the Smithsonian Mineral Research Collection at the Museum of Natural History. The FTIR laboratory at the Smithsonian Institution was established with generous support from Stephen Turner. This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC0206CH11357. Portions of this research were conducted with Advanced CyberInfrastructure computational resources provided by The Institute for CyberScience at The Pennsylvania State University (http://ics.psu.edu). Special thanks to Alexandre Fowler for her contribution to data collection.

99

References Atkins, A., Shaw, S., & Peacock, C. (2014). Nucleation and growth of todorokite from birnessite: Implications for trace-metal cycling in marine sediments. Geochimica et Cosmochimica Acta, 144, 109–125. http://doi.org/10.1017/CBO9781107415324.004 Bargar, J. R., Fuller, C. C., Marcus, M. A., Brearley, A. J., Perez De la Rosa, M., Webb, S. M., & Caldwell, W. A. (2009). Structural characterization of terrestrial microbial Mn oxides from Pinal Creek, AZ. Geochimica et Cosmochimica Acta, 73, 889–910. http://doi.org/10.1016/j.gca.2008.10.036 Bartlett, R. J., & James, B. (1979). Behavior of Chromium in Soils: III. Oxidation. Journal of Environment Quality, 8(1), 31–35. http://doi.org/10.2134/jeq1976.00472425000500040010x Becke, A. D. (1993). Density-functional thermochemistry.III. The role of exact exchange. The Journal of Chemical Physics, 98, 5648–5652. http://doi.org/10.1063/1.464913 Chao, T. T., & Theobald, P. K. (1976). The significance of secondary iron and manganese oxides in geochemical exploration. Economic Geology, 71, 1560–1569. http://doi.org/10.2113/gsecongeo.71.8.1560 Cramer, C. J. (2004). Essentials of Computational Chemistry Theories and Models. Essentials of Computational Chemistry. http://doi.org/10.1021/ci010445m Drits, V. A., Silvester, E., Gorshkov, A. I., & Manceau, A. (1997). Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite: II. Results from chemical studies and EXAFS spectroscopy. American Mineralogist, 82(9-10), 962–978. Fandeur, D., Juillot, F., Morin, G., Olivi, L., Cognigni, A., Webb, S. M., … Brown, G. E. (2009). XANES evidence for oxidation of Cr(III) to Cr(VI) by Mn-oxides in a lateritic regolith developed on serpentinized ultramafic rocks of New Caledonia. Environmental Science & Technology, 43, 7384–7390. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19848150 Feng, X. H., Zhu, M., Ginder-Vogel, M., Ni, C., Parikh, S. J., & Sparks, D. L. (2010). Formation of nano-crystalline todorokite from biogenic Mn oxides. Geochimica et Cosmochimica Acta, 74, 3232–3245. http://doi.org/10.1016/j.gca.2010.03.005 Finger, L. W., Cox, D. E., & Jephcoat, A. P. (1994). A Correction for Powder Diffraction Peak Asymmetry Due to Axial Divergence. Journal of Applied Crystallography, 27, 892–900. http://doi.org/10.1107/S0021889894004218 Fischer, T. B. (2011). Structural transformations of birnessite (δ-MnO2) during biological and abiological reduction. Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. R., … Pople, J. A. (2004). Gaussian 03. Giovanoli, R., Stähli, E., & Feitknecht, W. (1970). Über Oxidhydroxide des vierwertigen Mangans mit Schichtengitter. 1. Mitteilung: Natrium-mangan(II,III)manganat(IV). Helvetica Chimica Acta, 53, 454 – 464. Golden, D. C., Chen, C. C., & Dixon, J. B. (1986). Synthesis of todorokite. Science (New York, N.Y.), 231(4739), 717–719. http://doi.org/10.1126/science.231.4739.717 Golden, D. C., Chen, C. C., & Dixon, J. B. (1987). Transformation of birnessite to buserite, todorokite, and manganite under mild hydrothermal treatment. Clays and Clay Minerals, 35(4), 271–280. Julien, C. M., Massot, M., & Poinsignon, C. (2004). Lattice vibrations of manganese oxides: Part I. Periodic structures. Spectrochimica Acta - Part A: Molecular and Biomolecular Spectroscopy, 60, 689–700. http://doi.org/10.1016/S1386-1425(03)00279-8

100

Jürgensen, A., Widmeyer, J. R., Gordon, R. A., Bendell-Young, L. I., Moore, M. M., & Crozier, E. D. (2004). The structure of manganese oxide on the sehath of the bacterium Leptothrix discophora: An XAFS study. American Mineralogist, 89, 1110–1118. Kang, L., Zhang, M., Liu, Z. H., & Ooi, K. (2007). IR spectra of manganese oxides with either layered or tunnel structures. Spectrochimica Acta - Part A: Molecular and Biomolecular Spectroscopy, 67, 864–869. http://doi.org/10.1016/j.saa.2006.09.001 Kay, J. T., Conklin, M. H., Fuller, C. C., & O’Day, P. A. (2001). Processes of nickel and cobalt uptake by a manganese oxide forming sediment in Pinal Creek, Globe mining district, Arizona. Environmental Science & Technology, 35(24), 4719–4725. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11775144 Kazakis, N., Kantiranis, N., Voudouris, K. S., Mitrakas, M., Kaprara, E., & Pavlou, a. (2015). Geogenic Cr oxidation on the surface of mafic minerals and the hydrogeological conditions influencing hexavalent chromium concentrations in groundwater. Science of The Total Environment, 514, 224–238. http://doi.org/10.1016/j.scitotenv.2015.01.080 Krishnamurti, G. S. ., & Huang, P. M. (1988). Influence of manganese oxide minerals on the formation of iron oxides. Clays and Clay Minerals, 36(5), 467–475. Lafferty, B. J., Ginder-Vogel, M., & Sparks, D. L. (2011). Arsenite Oxidation by a PoorlyCrystalline Manganese Oxide. 3. Arsenic and Manganese Desorption. Environmental Science & Technology, 45, 9218–9223. http://doi.org/10.1021/es201281u Lanson, B., Drits, V. A., Silvester, E., & Manceau, A. (2000). Structure of H-exchanged hexagonal birnessite and its mechanism of formation from Na-rich monoclinic buserite at low pH. American Mineralogist, 85, 826–838. Learman, D. R., Voelker, B. M., Madden, A. S., & Hansel, C. M. (2013). Constraints on superoxide mediated formation of manganese oxides. Frontiers in Microbiology, 4, 1–11. http://doi.org/10.3389/fmicb.2013.00262 Learman, D. R., Wankel, S. D., Webb, S. M., Martinez, N., Madden, A. S., & Hansel, C. M. (2011). Coupled biotic–abiotic Mn(II) oxidation pathway mediates the formation and structural evolution of biogenic Mn oxides. Geochimica et Cosmochimica Acta, 75, 6048– 6063. http://doi.org/10.1016/j.gca.2011.07.026 Lee, C., Yang, W., & Parr, R. G. (1988). Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Physical Review B, 37(2), 785–789. Libowitzky, E., & Beran, A. (2004). IR spectroscopic characterization of hydrous species in minerals. In EMU Notes in Mineralogy: Spectroscopic Methods in Mineralogy (pp. 227 – 279). Ling, F. T., Heaney, P. J., Post, J. E., & Gao, X. (2015). Transformations from triclinic to hexagonal birnessite at circumneutral pH induced through pH control by common biological buffers. Chemical Geology, 416, 1–10. http://doi.org/10.1016/j.chemgeo.2015.10.007 Luo, J., Huang, A., Park, S. H., Suib, S. L., & Young, C. O. (1998). Crystallization of Sodium Birnessite and Accompanied Phase Transformation. Chemistry of Materials, 10, 1561–1568. Manceau, A., Marcus, M. A., & Tamura, N. (2002). Quantitative Speciation of Heavy Metals in Soils and Sediments by Synchrotron X-ray Techniques. Applications of Synchrotron Radiation in Low-Temperature Geochemistry and Environmental Science, 49, 341–428. Murray, J. W., & Dillard, J. G. (1979). The oxidation of cobalt(II) adsorbed on manganese dioxide. Geochimica et Cosmochimica Acta, 43, 781–787. Oscarson, D. W., Huang, P. M., Defosse, C., & Herbillon, A. (1981). Oxidative power of Mn(IV) and Fe(III) oxides with respect to As(III) in terrestrial and aquatic environments. Nature, 291, 50–51. Palchik, N. A., Moroz, T. N., Grigorieva, T. N., & Miroshnichenko, L. V. (2014). Manganese Minerals from the Miassovo Freshwater Lake: Composition and Structure. Russian Journal

101

of Inorganic Chemistry, 59(5), 511–518. http://doi.org/10.1134/S0036023614050131 Parikh, S. J., & Chorover, J. (2005). FTIR spectroscopic Study of Biogenic Mn-Oxide Formation by Pseudomonas putida GB-1. Geomicrobiology Journal, 22, 207–218. http://doi.org/10.1080/01490450590947724 Peacock, C. L., & Moon, E. M. (2012). Oxidative scavenging of thallium by birnessite: Explanation for thallium enrichment and stable isotope fractionation in marine ferromanganese precipitates. Geochimica et Cosmochimica Acta, 84, 297–313. http://doi.org/10.1016/j.gca.2012.01.036 Post, J. E. (1999). Manganese oxide minerals: Crystal structures and economic and environmental significance. Proceedings of the National Academy of Sciences, 96, 3447–3454. http://doi.org/10.1073/pnas.96.7.3447 Post, J. E., Heaney, P. J., & Hanson, J. (2002). Rietveld refinement of a triclinic structure for synthetic Na-birnessite using synchrotron powder diffraction. Powder Diffraction, 17(3), 218–221. Post, J. E., & Veblen, D. R. (1990). Crystal structure determinations of synthetic sodium, magnesium, and postassium birnessite using TEM and the Rietveld Method. American Mineralogist, 75, 477–489. Potter, R. M., & Rossman, G. R. (1979). The tetravalent manganese oxides : identification , hydration , and structural relationships by infrared spectroscopy. American Mineralogist, 64, 1199–1218. Retrieved from http://www.minsocam.org/ammin/AM64/AM64_1199.pdf?ref=PandoraISP Ramsey, M. S., & Christensen, P. R. (1998). Mineral abundance determination: Quantitative deconvolution of thermal emission spectra. Journal of Geophysical Research, 103(B1), 577–596. http://doi.org/10.1029/97JB02784 Rassolov, V. A., Ratner, M. A., Pople, J. A., Redfern, P. C., & Curtiss, L. A. (2001). 6-31G* Basis Set for Third-Row Atoms. Journal of Computational Chemistry, 22(9), 976–984. http://doi.org/10.1002/jcc.1058 Ravel, B., & Newville, M. (2005). ATHENA , ARTEMIS , HEPHAESTUS : data analysis for Xray absorption spectroscopy using IFEFFIT. Journal of Synchrotron Radiation, 12, 537– 541. http://doi.org/10.1107/S0909049505012719 Rietveld, H. M. (1969). A profile Refinement Method for Nuclear and Magnetic Structures. Journal of Applied Crystallography, 2, 65–71. http://doi.org/10.1107/S0021889869006558 Santelli, C. M., Webb, S. M., Dohnalkova, A. C., & Hansel, C. M. (2011). Diversity of Mn oxides produced by Mn(II)-oxidizing fungi. Geochimica et Cosmochimica Acta, 75, 2762– 2776. http://doi.org/10.1016/j.gca.2011.02.022 Saratovsky, I., Gurr, S. J., & Hayward, M. A. (2009). The Structure of manganese oxide formed by the fungus Acremonium sp. strain KR21-2. Geochimica et Cosmochimica Acta, 73, 3291–3300. http://doi.org/10.1016/j.gca.2009.03.005 Sayers, D. E., & Bunkers, B. A. (1988). X-ray Absorption Principles, Applications, Techniques of EXAFS, SEXAFS, and XANES. New York: Wiley. Schaftenaar, G., & Noordik, J. H. (2000). Molden: a pre-and post-processing program for molecular and electronic structures. Journal of Computer-Aided Molecular Design, 14, 123– 134. http://doi.org/10.1023/A:1008193805436 Scott, A. P., & Radom, L. (1996). Harmonic Vibrational Frequencies: An Evaluation of Hartree Fock, Møller - Plesset, Quadratic Configuration Interaction, Density Functional Theory, and Semiempirical Scale Factors. Journal of Physical Chemistry, 100, 16502–16513. Scott, J. M., & Morgan, J. J. (1996). Reactions at oxide surfaces. 2. Oxidation of Se(IV) by synthetic Birnessite. Environmental Science & Technology, 30, 1990–1996. http://doi.org/10.1021/es950741d

102

Silvester, E., Manceau, A., & Drits, V. A. (1997). Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite: II. Results from chemical studies and EXAFS spectroscopy. American Mineralogist, 82, 962–978. Stephens, P. W. (1999). Phenomenological model of anisotropic peak broadening in powder diffraction. Journal of Applied Crystallography, 32, 281–289. http://doi.org/10.1107/S0021889898006001 Thompson, P., Cox, D. E., & Hastings, J. B. (1987). Rietveld Refinement of Debye-Scherrer Synchrotron X-ray Data from A1203. Journal of Applied Crystallography, 20(2), 79–83. http://doi.org/10.1107/S0021889887087090 Thomson, J. L., & Salisbury, J. W. (1993). The Mid-Infrared Reflectance of Mineral Mixtures (714 µm). Remote Sensing of Environment, 45, 1–13. http://doi.org/10.1016/00344257(93)90077-B Toby, B. H., & Von Dreele, R. B. (2013). GSAS-II: the genesis of a modern open-source all purpose crystallography software package. Journal of Applied Crystallography, 46(2), 544– 549. http://doi.org/10.1107/S0021889813003531 Tsyganenko, A. A. (1975). IR spectra and structure of the hydroxyl covering of oxides comparison with the spectra of hydroxides and silicates. Journal of Structural Chemistry, 16(4), 536–540. Usui, A., & Mita, N. (1994). IX. Mineralogy, geochemistry and internal growth structure of manganese nodules in the western part of the Penrhyn Basin, South Pacific (GH83-3 area). Geological Survey of Japan Cruise Report, (23), 165–185. Villalobos, M., Lanson, B., Manceau, A., Toner, B., & Sposito, G. (2006). Structural model for the biogenic Mn oxide produced by Pseudomonas putida. American Mineralogist, 91, 489– 502. http://doi.org/10.2138/am.2006.1925 Weaver, R. M., & Hochella, M. F. J. (2003). The reactivity of seven Mn-oxides with Cr 3 + aq : A comparative analysis of a complex , environmentally important redox reaction. American Mineralogist, 88, 2016–2027. Webb, S. M. (2005). SIXPack : a Graphical User Interface for XAS Analysis Using IFEFFIT. Physica Scripta, T115, 1011–1014. Webb, S. M., Tebo, B. M., & Bargar, J. R. (2005). Structural characterization of biogenic Mn oxides produced in seawater by the marine bacillus sp. strain SG-1. American Mineralogist, 90, 1342–1357. http://doi.org/10.2138/am.2005.1669 Webb, S. M., Tebo, B. M., & Bargar, J. R. (2005). Structural characterization of biogenic Mn oxides produced in seawater by the marine bacillus sp. strain SG-1. American Mineralogist, 90, 1342–1357. http://doi.org/10.2138/am.2005.1669 White, W. B., Vito, C., & Scheetz, B. E. (2009). The mineralogy and trace element chemistry of black manganese oxide deposits from caves. Journal of Cave and Karst Studies, 71(2), 136– 143. Retrieved from ://WOS:000270607100005 Yang, D. S., & Wang, M. K. (2001). Syntheses and Characterization of Well-Crystallized Birnessite. Chemistry of Materials, 13, 2589–2594. Zhao, H., Liang, X., Yin, H., Liu, F., Tan, W., Qiu, G., & Feng, X. (2015). Formation of todorokite from “c-disordered” H+-birnessites: the roles of average manganese oxidation state and interlayer cations. Geochemical Transactions, 16, 1–11. http://doi.org/10.1186/s12932-015-0023-3 Zhao, H., Zhu, M., Li, W., Elzinga, E. J., Villalobos, M., Liu, F., … Sparks, D. L. (2016). Redox Reactions between Mn(II) and Hexagonal Birnessite Change its Layer Symmetry. Environmental Science & Technology, 1–33. http://doi.org/10.1021/acs.est.5b04436 Zhao, W., Liu, F., Feng, X., Tan, W., Qiu, G., & Chen, X. (2012). Fourier transform infrared spectroscopy study of acid birnessites before and after Pb2+ adsorption. Clay Minerals, 47,

103

191–204. http://doi.org/10.1180/claymin.2012.047.2.04

104

Figures

Figure 4-1. Schematic diagrams of (a) synthetic triclinic Na-birnessite and (b) synthetic hexagonal H-birnessite after Lanson et al. (2000).

105

H

H

O

O

Mn

Mn

a.) Mn7O24H22, triclinic birnessite

b.) Mn7O24H23, hexagonal birnessite without a vacancy H

O

Mn

vacancy

c.) Mn7O26H26, hexagonal birnessite with a vacancy

Figure 4-2. Geometry optimized models of nanoclusters representing (a) triclinic birnessite Mn7O24H22, ( b) hexagonal birnessite without a vacancy, Mn7O24H23, and (c) hexagonal birnessite with a vacancy, Mn7O26H26. Figures were drawn in Materials Studio (Accelerys Inc., San Diego CA).

106

Figure 4-3. Rietveld refinement fits for (a) synthetic triclinic Na-birnessite and (b) synthetic hexagonal H-birnessite.

107

Figure 4-4. (a) Full IR spectra of synthetic triclinic Na-birnessite and synthetic hexagonal Hbirnessite, (b) enlarged portion of the IR spectra over the OH-vacancy vibration range, and (c) enlarged portion of the IR spectra for the Mn-O bond vibration range.

108

Figure 4-5. FTIR spectra of pure triclinic and hexagonal birnessite and the first set of mechanical mixtures with weighted ratios of 25:75, 50:50, and 75:25 triclinic-hexagonal birnessite.

109

Figure 4-6. Rietveld refinement fits for (a) 25:75, (b) 50:50, and (c) 75:25 triclinic-to-hexagonal birnessite mixtures.

110

a.)

0.8

NaHB 75:25

0.6 NaHB 50:50

0.4

c.) 1.0

NaHB 25:75

0.2

0.0 0.0

0.2

0.4

0.6

0.8

1.0

IR Fraction Na-Birnessite

b.) XAFS Fraction Na-Birnessite

1.0

0.8 NaHB 75:25

0.6

0.8 NaHB 75:25

0.6

0.4

NaHB 50:50

NaHB 25:75

0.2

0.0 0.0

NaHB 50:50

0.4

XAFS Fraction Na-Birnessite

XRD Fraction Na-Birnessite

1.0

0.2

0.4

0.6

0.8

1.0

XRD Fraction Na-Birnessite

NaHB 25:75

0.2

0.0 0.0

0.2

0.4

0.6

0.8

1.0

IR Fraction Na-Birnessite

Figure 4-7. Comparison for volume fractions of mechanical mixtures of triclinic and hexagonal birnessite from (a) FTIR compared with XRD volume fractions, (b) FTIR compared with EXAFS volume fractions, and (c) EXAFS compared with XRD volume fractions. The dashed lines represent the 1:1 correlation line.

111

Figure 4-8. (a) Full calculated IR spectra for nanoclusters representing triclinic birnessite, hexagonal birnessite with no vacancy site, and hexagonal birnessite with a vacancy site, along with collected FTIR spectra for endmember triclinic Na-birnessite and hexagonal H-birnessite with intensities multiplied by a scale factor of 3000 for comparison. (b) The data in the range from 1300 to 1800 cm-1 are enlarged to highlight differences resulting from varying vacancy concentration in the birnessite structure. (c) The data in the range of 400 to 1000 are enlarged for easier comparison of peaks related to Mn-O bonding in the DFT calculations and observed FTIR spectra.

112

a.) 618 cm-1

c.) 1550 cm-1

H

vacancy

Mn

Mn

Mn

Mn

O

O H

b.) 995 cm-1

d.) 3570 cm-1

vacancy

Mn

Mn

H

Mn

O H

Mn

Mn O O

H

Figure 4-9. Exemplary vibrations in a birnessite nanocluster for (a) inner Mn-O lattice vibrations leading to vibrations in the 400 to 700 cm-1 range, (b) peripheral Mn-O vibrations leading to vibrations in the 700 to 1400 cm-1 range, (c) H2O scissoring vibrations leading to peaks in the 1400 to 1800 cm-1 range, and (d) OH stretches leading to peaks in the 1800 to 4000 cm-1 range. The larger, thick arrows represent the main vibrations, while smaller arrows represent resulting motions due to the close vicinity or connectivity in the nanocluster to the main vibrations occurring simultaneously.

113

a.) 1653 cm-1

c.) 1569 cm-1 O H vacancy vacancy

Mn Mn H

Mn O

H b.) 1599 cm-1

d.) 1498 cm-1 H O Mn vacancy

vacancy

Mn

Mn

Mn

H

O H

Figure 4-10. All vibrations for hexagonal birnessite nanocluster with a vacancy site in the 1300 to 1800 cm-1 range, including the OH vibrations at (a) 1653, (b) 1599, (c) 1569, and (d) 1498 cm-1.

114

Figure 4-11. Gaussian-Lorentzian peak shapes for peaks in the OH vacancy-related range from 1300 to 1800 cm-1 for (a) hexagonal H-birnessite and (b) triclinic Na-birnessite.

115

Tables Table 4-1. Rietveld refinement results for unit cell parameter for end-member synthetic triclinic Na-birnessite and hexagonal H-birnessite.

Space group Unit cell a (Å) b (Å) c (Å) α (°) β (°) γ (°) vol (Å) Density (g/cm3) Refinement No. of diffraction points No. of reflections Diffraction range (2θ), λ = 0.71069 Å No. of variables R(F2) Rwp

Triclinic Na-birnessite C -1

Hexagonal H-birnessite P -3

5.167(6) 2.849(1) 7.347(9) 90.1(1) 103.14(6) 90.1(1) 105.3(1) 3.573

2.873(3) 7.432(4) 90 90 120 53.12(7) 2.917

1820 125

1820 43

8.00 – 44.41 24 0.0488 0.0232

8.00 - 44.41 22 0.0804 0.0151

116

Table 4-2. Atom positions for end-member synthetic triclinic Na-birnessite and hexagonal Hbirnessite from Rietveld analysis.

Triclinic Nabirnessite

Hexagonal Hbirnessite

Atom Mn(oct) O(oct) Na or H2O(int) Mn(oct) O(oct) H2O(int) Mn(int)

Site occupancy x y z factor 0 0 0 1 0.379(1) 0.01(1) 0.143(1) 1 0.579(2) 0.200(3) 0.489(2) 0.83(2) 0 0 0 0.727 0.3333 0.6667 0.099(2) 1 0.6667 0.3333 0.643(3) 0.444 0 0 0.670(3) 0.065

Uiso 0.00286 0.01871 0.05580 0.02447 0.01731 0.14121 0.02840

117

Table 4-3. Peak locations in the Mn-O range (400 to 750 cm-1) for synthetic birnessites, and their respective peak labels to compare with Fig. 8. Sample Na-birnessite Hexagonal Birnessite

Peak 1 418 (T1)

Peak 2 478 (T2) 440 (H2)

Peak 3 511 (T3) 494 (H3)

Peak 4 639 (T4) 659 (H4)

118

Table 4-4. Volume fractions of triclinic Na- and hexagonal H-birnessite in mixtures determined with FTIR, EXAFS, and XRD. Wt. % ratio of triclinicto-hexagonal birnessite

Vol. % ratio of triclinicto-hexagonal birnessite

Mix 1 25:75 Mix 1 50:50 Mix 1 75:25 Mix 2 25:75 Mix 2 50:50 Mix 2 75:25

Mix 1 22:78 Mix 1 45:55 Mix 1 71:29* Mix 2 22:78* Mix 2 45:55* Mix 2 70:30

Wt. % ratio of triclinicto-hexagonal birnessite

Vol. % ratio of triclinicto-hexagonal birnessite

Mix 1 25:75 Mix 1 50:50 Mix 1 75:25

Mix 1 22:78 Mix 1 45:55 Mix 1 71:29

Wt. % ratio of triclinicto-hexagonal birnessite

Vol. % ratio of triclinicto-hexagonal birnessite

IR Spectral Unmixing Vol Fraction Vol Fraction Na-birnessite H-birnessite R-factor 0.347 ± 0.002 0.653 ± 0.002 0.0009 0.482 ± 0.002 0.518 ± 0.002 0.0015 0.700 ± 0.002 0.300 ± 0.002 0.0014 0.232 ± 0.003 0.768 ± 0.003 0.0021 0.468 ± 0.004 0.552 ± 0.004 0.0021 0.719 ± 0.002 0.281 ± 0.002 0.0010 XAFS Linear Combination Fitting Vol Fraction Vol Fraction Na-birnessite H-birnessite R-factor 0.180 ± 0.021 0.820 ± 0.023 0.0156 0.436 ± 0.022 0.564 ± 0.025 0.0184 0.639 ± 0.022 0.361 ± 0.024 0.0144 XRD Structure Refinements

Vol Fraction Na-birnessite Mix 1 25:75 Mix 1 22:78 0.217 ± 0.006 Mix 1 50:50 Mix 1 45:55 0.490 ± 0.004 Mix 1 75:25 Mix 1 71:29 0.702 ± 0.004 *results averaged from spectral unmixing of 2 – 4 KBr pellets

Vol Fraction H-birnessite 0.783 ± 0.010 0.510 ± 0.005 0.298 ± 0.005

Rwp 0.0407 0.0204 0.0294

119

Table 4-5. Peak positions from calculated FTIR spectra of birnessite nanoclusters. (a.) Model Mn7O24H22 Mn7O24H23 Mn7O26H26

Inner Mn-O lattice vibrations Peak 1 Peak 2 Peak 3 559(T1) 611(T2) 642(T3) 463(H1) 578(H2) 636(H3) 463(H1) 612(H2) 650(H3)

(b.) Model Mn7O24H22 Mn7O24H23 Mn7O26H26

Peak 4 761 756 763

Peak 5 848 803

(c.) Model Mn7O24H22 Mn7O24H23 Mn7O26H26

Peak 13 1536(T13) 1544(H13) 1536(H13)

Peak 6 885 866

H2O vibrations Peak 14 Peak 15 1572(T14) 1569(H14) 1569(H14) 1599(H15)

(d.) Model Peak 17 Mn7O24H22 Mn7O24H23 Mn7O26H26 Model Mn7O24H22 Mn7O24H23 Mn7O26H26

Peripheral Mn-O vibrations Peak 7 Peak 8 Peak 9 1045 957 907 995 1051

Peak 18

2152

2187

Peak 23 3425 3448

Peak 24 3521 3533 3528

Peak 10

Peak 11

Peak 12

1122

1167

1218

Peak 16

1653(H16)

OH vibrations Peak 19 Peak 20 2881 3083 2685 2720 OH vibrations (cont.) Peak 25 Peak 26 3571 3542 3575 3589

Peak 21

Peak 22

3124

3237

Peak 27 3690 3698

Peak 28 3731 3732 3733

120

Supplementary Information A

c.) Mn-Obr asymm (in Mn7O24H22 at 582 cm-1)

O

O Mn

a.) Mn-Obr (in Mn7O24H22 at 511 cm-1)

H

O d.) Mn-Obr symm (in Mn7O24H22 at 559 cm-1)

Mn

O H H O Mn b.) Mn3-Obr (in Mn7O24H22 at 528 cm-1)

O

H e.) Defect site (in Mn7O26H26 at 795 cm-1) Mn

O

O vacancy

Mn O O

H

SI-A Figure 1. Exemplary movements for vibrational modes a.) Mn-Obr, b.) Mn3-Obr, c.) Mn-Obr asymm, d.) Mn-Obr symm, and e.) defect site.

121

Chapter 5 The relationship between Mn oxidation state and structure in triclinic and hexagonal birnessites

Abstract Through a combination of Fourier transform infrared spectroscopy (FTIR), extended Xray absorption fine structure (EXAFS), X-ray photoelectron spectroscopy (XPS) and X-ray diffraction (XRD), we have examined the relationship between structural symmetry and the Mn oxidation state ratios in triclinic birnessites with Na, K, Ba, and Ca in the interlayer, and hexagonal birnessites synthesized at different pH conditions. Our results suggest that as the concentration of Mn3+ as measured by XPS increases, the departure from hexagonal symmetry, as revealed by EXAFS and FTIR spectroscopy, also increases. Likewise, the Jahn-Teller distortions associated with Mn3+ induce systematic variations in unit-cell parameters refined from X-ray diffraction data, particularly an increase in the a-axis and the β angle of the unit cell as determined from Rietveld refinements of XRD data.

Introduction Manganese (Mn) oxides exist in a wide range of natural environments such as ocean nodules and soil coatings (Post, 1999). They are heavily involved in the cycling of metals and in redox reactions, including the sorption of Pb, Mo, and Cu (Chao and Theobald 1976; Usui and Mita 1994) and the oxidation of Cr3+, Co2+ and Se4+ (Bartlett and James 1979; Murray and Dillard 1979; Scott and Morgan 1996). In addition to their significance in nature, Mn oxides offer technological applications in environmental remediation and energy storage and extraction. For instance, McCann et al. (2015) successfully tested the use of a natural Mn oxide as an effective

122

stabilizing agent for Pb in soil. Mn oxides also have been explored for the sequestration of Cs, Cu, and Cd (Cho et al. 2011; Fleeger et al. 2013; Peña et al. 2015). Moreover, Mn oxides may be critical to clean-energy technologies, particularly with respect to the development of efficient solar cells. The oxidation of water in natural photosynthesis occurs via an oxygen-evolving complex with a CaMn4O5 cluster with a structural affinity to ranciéite (Ferreira et al. 2004; McEvoy and Brudvig 2006; Umena et al. 2011). Researchers are working to develop a synthetic Mn oxide to serve as an analogously efficient water oxidation catalyst in solar cells (Iyer et al. 2012; Robinson et al. 2013; Meng et al. 2014; Najafpour and Isaloo 2015). Although various Mn oxide phases have been discovered in the natural environment, the layered Mn oxides in the birnessite family are some of the most widely studied due to their common occurrence and chemical reactivity (Post 1999; Weaver and Hochella 2003). Birnessites exist with either triclinic or hexagonal structural symmetry (Fig. 1). Synthetic triclinic Nabirnessite, or Na0.58(Mn4+1.42Mn3+0.58)O4 • 1.5H2O, is described as a layered Mn oxide with Mn octahedral sheets consisting of ~29% Mn3+ and ~71% Mn4+ cations with hydrated Na+ cations partially occupying interlayer sites (Post and Veblen 1990; Post et al. 2002). Synthetic hexagonal H-birnessite, on the other hand, has the proposed formula H0.33Mn3+0.111Mn2+0.055(Mn4+0.722Mn3+0.111☐0.167)O2, with Mn octahedral sheets containing ~72% Mn4+ cations, ~11% Mn3+ cations, and ~17% vacancies. Mn2+, Mn3+, and H+ cations occupy locations over the vacancy sites in the interlayer (Silvester et al. 1997). Variations in Mn oxidation state influence the reactivity of birnessites. Octahedral Mn3+, for example, is thought to control the oxidation of Cr3+ and Co2+ (Manceau et al. 1992; Weaver and Hochella 2003; Simanova and Peña 2015). Mn3+ also appears to perform an integral function in the oxidation of water by the oxygen-evolving complex during photosynthesis (Takashima et al. 2012a,b; McKendry et al. 2015). However, the existence of three valence states for Mn in natural Mn oxides (Mn2+, Mn3+, and Mn4+) complicates our ability to tease out the role that Mn

123

oxidation state plays in mediating chemical reactions (Post 1999; Manceau et al. 2012). Improving our ability to relate the valence state of Mn in birnessite to its crystal structure will allow us to better understand the role of birnessite in the natural environment and to develop applications for environmental remediation and energy technologies. Many studies of birnessite report only the Mn average oxidation state (AOS) as an indicator of birnessite reactivity. In the present study, we explore the assumption that Mn AOS is sufficient to predict the properties of a given birnessite phase. Wet-chemical titration is the most common method for quantifying the Mn valence state in solids, but most studies have employed this technique to measure only the AOS (Lingane and Karplus 1987; Freeman and Chapman 1971; Murray et al. 1984; Kijima et al. 2001). It is often assumed that little to no Mn2+ exists in the birnessite structure, yielding an AOS that is based primarily on the ratio of Mn3+ to Mn4+ (Wang et al. 2012; Zhao et al. 2011). Moreover, scientists typically have focused their attention on hexagonal rather than triclinic birnessite based on a presumption that hexagonal birnessite is more prevalent in the natural environment. Laboratory experiments have consistently suggested that both bacteria and fungi initially produce hexagonal H-birnessite at circum-neutral pH (Villalobos et al. 2003, 2006; Bargar et al. 2005; Saratovsky et al. 2006; Learman et al. 2011; Santelli et al. 2011; Hansel et al. 2012). Ling et al. (2015), however, have argued that common biological buffers can force the precipitation of hexagonal rather than triclinic birnessite. Natural birnessites typically contain Ca, K, and other interlayer cations (Jones and Milne 1956; McKeown and Post 2001; Tan et al. 2010), and the interlayer cations also are likely to influence the symmetry of the birnessite structure. For example, Ca-rich birnessite has been reported in sediments from passive treatment systems designed to remediate excessive Mn oxide in acid-mine wastes, and both hexagonal and triclinic constituents are identified in these birnessite samples (Tan et al. 2010). Thus, it is

124

important to understand how different interlayer cations affect birnessite structure and Mn oxidation state. In order to gain insights into these questions, we applied Fourier transform infrared spectroscopy (FTIR), X-ray diffraction (XRD), and extended X-ray absorption fine structure (EXAFS) to analyze the structure of triclinic Na-birnessite samples that were cation exchanged with K, Ca , and Ba, and we compared those results with investigations of hexagonal birnessite samples that were synthesized by a variety of preparations. The structural data so obtained were correlated with Mn oxidation states, as determined from X-ray photoelectron spectroscopy (XPS), to explore the effects of Mn3+ on short- and long-range crystal structure.

Methods

Materials synthesis Synthesis of standard triclinic birnessite. A 7 Å, triclinic Na-birnessite was synthesized according to the procedure described in Golden et al. (1986). A 200 ml solution of 0.5 M MnCl2 (Mallinckrodt Baker) was mixed with 250 ml of 5.5 M NaOH (J.T. Baker). The mixture was oxygenated through a glass frit for ~5 hrs at a rate of 1.5 L/min. The precipitate was divided evenly and centrifuged in 14 centrifuge tubes. The solution was decanted and replaced with fresh deionized (DI) water to rinse. The rinse cycle was repeated five times. Na-birnessite was stored in ~350 ml DI water until experimental use. For experiments, aliquots of Nabirnessite were filtered with a 0.05 μm Nuclepore Track-Etched polycarbonate membrane filter (Whatman), rinsed three times with 100 mL DI water, and left to air-dry. Synthesis of standard hexagonal birnessite. Our hexagonal birnessite standard was synthesized by reacting ~100 mg of dried triclinic Na-birnessite in 100 ml of 0.01 M HCl for 24

125

hrs. The H-birnessite then was filtered with a 0.05 μm Nuclepore Track-Etched polycarbonate membrane filter (Whatman), rinsed three times with 100 mL DI water, and left to air-dry. Synthesis of cation-exchanged birnessites. K-, Ca-, and Ba-birnessites were created by exchanging interlayer Na in the triclinic Na-birnessite standard with K, Ca, and Ba. A 0.1 M KCl solution was prepared with 1.4951 g KCl (Fluka) diluted to 200 ml. A 0.01 M CaCl2 solution was made from 0.3932 g CaCl2 • 2H2O (Amresco) diluted to 250 ml with DI water. A 0.1 M BaCl2 solution was mixed using 4.1718 g BaCl2 (J.T. Baker) in 200 ml DI water. All solutions were adjusted to pH 7 using 0.1 M HCl (OmniTrace) and 0.1 M NaOH (J.T. Baker). For each, ~70 to 100 mg of triclinic Na-birnessite was placed in a 250 ml glass beaker with 100 ml of 0.1 M KCl, CaCl2, or BaCl2 solution for ~24 hrs. Each sample was investigated with XRD (described below) and analyzed using an FEI Nova NanoSEM 600 (15 keV, 1 – 2 nA) equipped with a ThermoFisher energy dispersive spectroscopy X-ray detector (EDS) (Department of Mineral Sciences, Smithsonian Institution) to confirm full cation exchange. Synthesis of non-standard hexagonal birnessites. In addition, we synthesized birnessite powders under conditions that we thought might promote different ratios of Mn4+:Mn3+ within the octahedral sites relative to the standard hexagonal birnessite. In one preparation, H-birnessite was synthesized at pH 3 rather than pH 2, as was used for our hexagonal birnessite standard, by reacting ~100 mg of dried triclinic Na-birnessite in 100 ml of 0.001 M HCl for 24 hrs. Since Ling et al. (2015) demonstrated that biological buffers at circumneutral pH can promote the formation of hexagonal birnessite, we also synthesized H-birnessite using a HEPES buffer at pH 7. A 20 mM pH 7 HEPES-buffered solution was prepared by dissolving ~2.38 g HEPES (SigmaAldrich) in ~300 ml DI water. The pH was then adjusted with 0.1 M NaOH (J.T. Baker) until pH 7.00 was achieved, and the volume was brought up to 500 mL. About 100 mg of dried triclinic Na-birnessite were then submersed and stirred in 100 ml of a 20 mM pH 7 HEPES-buffered

126

solution for 14 days. All samples were filtered with a 0.05 μm polycarbonate membrane filter, rinsed three times with 100 mL DI water, and left to air-dry.

X-ray diffraction (XRD) and structure refinements For XRD investigations, all samples were ground in an agate mortar under acetone to disaggregate clumps. For analysis, ~2 mg of sample were mounted on glass fibers into a Rigaku D/MAX-RAPID microdiffractometer with an imaging plate detector (Smithsonian Institution, Dept. of Mineral Sciences) and a Mo tube source (λ = 0.7093 Å). Samples were rotated 360° around the phi axis at 1° s-1 during data collection with a 10 min exposure time. Rietveld refinements (Rietveld 1969) were completed for all birnessite specimens using the General Structure Analysis System-II (GSAS-II) software (Toby & Von Dreele 2013). The starting atom positions for the Na-birnessite refinements were drawn from Post et al. (2002). The starting atom positions for the H-birnessite refinements were drawn from Fischer (2011). Peak profile parameters that captured instrumental broadening were determined by Rietveld analysis of a diffraction pattern generated by a LaB6 standard (NIST SRM 660a) as collected using the same sample holder as was used for birnessite. The background parameters, scale factor, unit-cell parameters, and peak profile coefficients were allowed to vary. The background intensities were fitted with a Chebyshev function using 6 terms. Diffraction peak profiles were fitted with a pseudo-Voigt function as parameterized by Thompson et al. (1987), with asymmetry correction by Finger et al. (1994), and microstrain anisotropic broadening terms by Stephens (1999). Final refinements allowed the atomic positions and occupancies to vary. Temperature factors were also varied in the refinements of pure triclinic Na-birnessite and pure hexagonal birnessite, but kept constant for all other refinements. For all refinements, the (0 0 1) peak was omitted due to its high relative intensity and problems with integrating the diffraction images near the direct beam.

127

X-ray absorption spectroscopy/extended X-ray absorption fine structure (XAS/EXAFS) For X-ray absorption spectroscopy (XAS), samples were ground under acetone in a mortar and pestle, and sieved into a thin layer with a 425 or 500 mesh sieve onto kapton tape. The kapton tape was then folded over to seal in the sample. Manganese K-edge XAS spectra were collected using a synchrotron source at Beamline 12-BM of the Advanced Photon Source (APS), Argonne National Laboratory using a Si(111) double-crystal, fixed exit monochromator and a double mirror system (flat plus torroidal) with an energy cutoff of 23 keV. The pre-edge peak of a Mn foil was used for energy calibration (6539 eV). Fluorescence data were collected with a 13-element Ge detector and Cr(III) foil in front of the Ge detector to cut out scatter in all samples. Three to six scans were collected per sample at room temperature from -200 to about +800 eV around the Mn K-edge (6539 keV). Data analysis of spectra was conducted using the ATHENA software program (Ravel & Newville, 2005). XAS spectra were calibrated using Mn foil, averaged, background-subtracted, normalized, and deglitched if obvious outliers existed in the data. Analysis of the Mn K-edge EXAFS region was used for linear combination fitting (LCF) of samples (Webb et al. 2005a,b; Villalobos et al. 2006; Saratovsky et al. 2009; Feng et al. 2010). The χ(k) spectra were converted to k (Å-1) (Sayers & Bunkers, 1988). The resulting χ(k) data were k2-weighted and analyzed using the k-range from 2.3 to 11.3 Å-1. The k2χ(k) data were then Fourier transformed over the 2.5 < k < ~11.5 Å-1 range using a 2 Å-1 Kaiser-Bessel window for the resulting partial radial distribution function (RDF). Linear combination fitting (LCF) of the k2χ(k) data was used to determine the phase fractions of triclinic and hexagonal birnessite for the synthetic birnessites with mixed symmetry. The synthetic triclinic Na-birnessite and synthetic hexagonal pH 2 birnessite standards were used to fit the EXAFS data. All weights were set between 0 and 1.

128

Fourier transform-infrared spectroscopy (FTIR) Samples were ground under acetone in a mortar and pestle and sieved through a 325 mesh sieve. Then, 2 to 4 mg of Mn oxide sample were milled with ~250 mg KBr for 1-2 min, and pressed into a pellet. Vibrational spectra were collected on a Nicolet 6700 Analytical FTIR Spectrometer from 400 to 4000 cm-1. The resolution was set at 3.86 cm-1 with a scan number of 64. Omnic 8 (Nicolet) was used to view data during data collection. A linear baseline was used for background subtraction, and all data were normalized with the highest peak in the 400 to 750 cm-1 region set to 1 in MATLAB R2015a (Mathworks). If peak deconvolution was performed, the peaks were deconvoluted in the OPUS 7.0 spectroscopy software (Bruker) with the Levenberg-Marquardt algorithm using a combination of Gaussian and Lorentzian peak shapes. To quantify the relative concentrations of hexagonal and triclinic birnessite in a sample, linear spectral unmixing of normalized, background-subtracted FTIR spectra was also performed using the Least Squares Fitting algorithm with SIXpack (Webb 2005). We used the observed spectra in the range of 400 to 750 cm-1 as obtained from our triclinic birnessite and hexagonal birnessite standards for spectral decomposition of those synthetic phases with mixed symmetry character (Thomson and Salisbury 1993; Ramsey and Christensen 1998).

X-ray photoelectron spectroscopy (XPS) For XPS analysis, powder samples were covered with a strip of conductive copper tape and pressed with clean borosilicate glass blocks onto copper stubs. Measurements were conducted with a Kratos Axis Ultra DLD spectrometer with an Al Kα X-ray source (1486.7 eV) operating at 10 mA and 15 kV. Magnetic immersion lenses were used to improve collection efficiency. The instrument work function was calibrated to give a binding energy (BE) of 83.96

129

eV ± 0.05 eV for the 4f7/2 line of metallic gold. The spectrometer dispersion was adjusted to yield a BE of 932.62 eV for the Cu2p3/2 line of metallic copper. Measurements of the Mn2p, Mn3s, Mn3p, O1s, C1s, and various alkali and alkaline Earth lines were conducted with a step size of 0.1 eV, an analysis area of 300 x 700 microns, and pass energies (PE) of 20 or 40 eV. The resultant full-width-at-half-maximums (FWHM) for the Ag3d5/2 line were 0.54 and 0.77 eV, respectively. The low sensitivity of the Mn3s line resulted in measurements only with PE = 40 eV. Survey scans were conducted at PE = 160 eV and step size = 0.5 eV. XPS spectra were fit by non-linear least squares after Shirley background subtractions with the CasaXPS curve resolution software package. Gaussian/Lorentzian contributions to line shapes were numerically convoluted with a Voigt function.

Results & Discussion

Qualitative comparison of phases using XRD, FTIR, and EXAFS Hexagonal birnessites. When examined by XRD, FTIR, and EXAFS, the two synthetic hexagonal birnessite powders synthesized at pH 3 and also in a pH 7 HEPES solution closely resembled the standard hexagonal birnessite synthesized at pH 2. For example, all three preparations generated birnessite samples that yielded powder X-ray diffraction patterns that were virtually indistinguishable (Fig. 2). The similarity of the XRD patterns suggests that the structures of these samples were nearly identical over the ~10-Å correlation length associated with diffraction analysis (Buerger 1960). Moreover, the FTIR spectra for these hexagonal birnessite samples were also very similar to that of standard hexagonal birnessite (Ling et al. in prep.). Absorption peaks over the energy range that is most sensitive to Mn-O vibrations (400 to 750 cm-1) were close to those of the standard hexagonal birnessite (Fig. 3a, Table 1).

130

Nevertheless, we note that the peak positions for the hexagonal birnessite powders prepared by non-standard protocols were shifted slightly relative to the hexagonal birnessite standard. Specifically, H-birnessite synthesized at pH 3 exhibited absorption peaks at 440 and 494 cm-1, whereas the H-birnessite synthesized in the presence of the HEPES buffer exhibited peaks at 447 and 491 cm-1 (Table 1). Since FTIR is more sensitive to local structure than is X-ray diffraction, the FTIR analysis may be detecting distortions that fell outside the resolution of XRD. The full FTIR spectra of hexagonal birnessites can be found in Supplementary Information B. Qualitative analysis of the EXAFS spectra for these samples offered further support for hexagonality of the pH 3 and pH 7 HEPES birnessites. The k2χ(k) plots of all three show the distinct, single antinode at ~8.0 Å-1 (Fig. 4b-d), indicative of hexagonal birnessite (Webb et al. 2005a). The peak at ~8.0 Å-1 splits into a double antinode for triclinic birnessite, due to the ordering of structural Mn3+ (Webb et al. 2005a). Similarly, the RDFs of the birnessite synthesized at pH 3 and in a HEPES-buffered solution at pH 7 matched well with that of the standard pH 2 hexagonal birnessite (Fig. 5). The HEPES-buffered birnessite sample did, however, exhibit an extra peak at ~3.8 Å-1 relative to the hexagonal birnessite synthesized at pH 2 and 3. Triclinic birnessites. The FTIR spectra for K-, Ca-, and Ba-birnessite synthesized at pH 7 revealed absorption peaks that are characteristic of triclinic birnessite, which shows strong absorbances at 418, 478, and 511 cm-1 (Ling et al., in prep., Fig. 3b, Table 1). However, these spectra offer hints that the triclinic symmetry in the K- and Ba-exchanged birnessites is trending towards hexagonality. Specifically, the tendency towards a hexagonal character in the K- and Baexchanged birnessites was suggested by shifts of an absorption peak that appears at 478 cm-1 in triclinic Na-birnessite to 479 cm-1 in K-birnessite and 483 cm-1 in Ba-birnessite, thus approaching the position of the 496 cm-1 peak in the hexagonal birnessite standard (Table 1). The FTIR spectrum for Ca-birnessite showed the most significant shift of this Mn-O vibration peak, to 487

131

cm-1, appearing in close alignment with the 496 cm-1 peak in the hexagonal birnessite standard (Fig. 3b). Moreover, the Ca-birnessite absorbance at 419 cm-1 was asymmetric, and decomposition of this peak into two components yielded a peak at 438 cm-1, which approximates the position of a 453 cm-1 absorbance in the hexagonal birnessite standard. Full FTIR spectra of cation-exchanged triclinic birnessites can be found in the Supplementary Information B. Analysis of the EXAFS data showed similar behaviors. In the k2χ(k) spectrum of triclinic Na-birnessite, two prominent antinodes occur at ~8.0 Å-1. These antinodes were a little less prominent in the EXAFS spectrum of Ba-birnessite (Fig. 6). K-birnessite exhibited a broad peak in this range, without a clear distinction between the two antinodes. In Ca-birnessite, the antinode approaches the appearance of a single antinode, as is seen in the EXAFS spectrum of the hexagonal H-birnessite standard, although the intensity and sharpness of the antinode for Cabirnessite was lower that observed in hexagonal birnessite. The radial distribution functions (RDFs) extracted from the EXAFS spectra further supported this pattern. The peak positions of the RDF for Ba-birnessite closely correlated with those of triclinic Na-birnessite (Fig. 7). On the other hand, K- and Ca-birnessite exhibited features over the range of 2.9 to 4.0 Å that more closely resemble the RDF of hexagonal Hbirnessite than triclinic Na-birnessite. For example, the ratios of the peak intensity at ~5.3 Å to that at ~5.7 Å for K- and Ca-birnessite are 1.05 and 1.06, respectively, close to the ratio of 1.07 exhibited by hexagonal H-birnessite and far from the ratio of 0.94 for the same peaks in triclinic Na-birnessite. The peak ratio for Ba-birnessite was 0.39, even lower than that of triclinic Nabirnessite. Structural modeling of EXAFS data by Webb et al. (2005a) suggests that the effective scattering distance over the range of ~5.6 – 5.8 Å is highly sensitive Mn distortions in the octahedral sheet, perhaps leading to the differences we see between cation-exchanged birnessites. Our results for Na- and Ca-birnessite differ from those reported in Webb et al. (2005b), whose EXAFS analyses revealed the structure of a synthetic, biogenically precipitated Na-

132

birnessite to be hexagonal and that of Ca-birnessite to be triclinic through EXAFS analyses. The birnessites in Webb et al. (2005b) were produced from marine Bacillus sp., strain SG-1, incubated in 10 uM Mn(II) in 5 mM HEPES buffer around pH 7.7 to 7.8 with either 50 mM CaCl2 or NaCl. This discrepancy suggests that the environment and mechanisms of formation can easily influence the final triclinicity or hexagonality of birnessite.

Effects of Mn3+ on birnessite structure as revealed by Rietveld analysis Silvester et al. (1997) and Drits et al. (1997) propose that as Mn3+ content within the octahedral sheet increases, Jahn-Teller distortions associated with Mn3+ deform the octahedra from ideality, with a concomitant loss of hexagonal symmetry. This interpretation implies that the extent of the distortions that lead to triclinicity correlate with the content of structural Mn3+. In Ilton et al. (2015), we determined the ratios of Mn2+:Mn3+:Mn4+ for the samples used in this present study (Fig. 8). The mole fraction of Mn3+ in our triclinic Na-birnessite standard (0.38 mol%) was 1.8 times higher than the amount in a hexagonal birnessite prepared at pH 3 (0.21 mol%). Consistent with the model of Silvester et al. (1997), the cation-exchanged triclinic birnessites exhibited high Mn3+ fractions, ranging from a low of 0.29 (Ca-birnessite) up to 0.40 ± 0.02 mol% (Ba-birnessite), whereas the Mn3+ fractions of the hexagonal birnessites were the same within error (0.21 to 0.22 ± 0.02 mol%). Importantly, we note that birnessite samples with different chemistries and symmetries exhibited the same average oxidation state for Mn. The Mn AOS for triclinic Na-birnessite, Cabirnessite, and pH 2 hexagonal birnessite fell within error of each other (3.58 ±0.02 mol%), as did the AOS for Ba-birnessite, pH 3 hexagonal birnessite, and HEPES pH 7 birnessite (3.50 ±0.02 mol%). Therefore, it is clear that Mn AOS alone does not adequately register Mn2+:Mn3+:Mn4+ stoichiometries, structural symmetry, or interlayer composition.

133

Our Rietveld analyses of the synchrotron X-ray diffraction patterns for these phases allowed us to look for correlations between Mn3+ content and unit-cell parameters. We refined the triclinic Na-birnessite and cation-exchanged K-, Ba-, and Ca-birnessites using S.G. C -1 and obtained reasonable fits to the data (Fig. 9a, Tables 2 – 3). All three hexagonal H-birnessites were refined using both a P -3 unit cell and a C -1 unit cell for direct comparison with refined parameters from a triclinic unit cell (Fig. 9b, Tables 4 – 5). This analysis revealed a strong correlation (R2 ~0.90) between Mn3+ content and the a-axis and the  angle (in the C -1 setting), with both parameters increasing with increasing Mn3+ concentration (Fig. 10). It is challenging to infer a relationship between the orientation of the Jahn-Teller distortions associated with Mn3+ and the increase in a and  because Lopano at el. (2007) have shown that the size of the interlayer cation can influence unit-cell dimensions, particularly with respect to the length of the c-axis and the unit-cell volume. Drits et al. (1997) explored ratios between Mn3+ content and lattice parameters, and they argued for a direct relationship between the a:b ratio and Mn3+ concentration, with the a parameter increasing significantly relative to the b parameter as Jahn-Teller distortions increased. Our analysis of these relationships, however, yielded only weak correlations between Mn3+ content and ratios of unit cell parameters. For example, the correlation coefficient (R2) between Mn3+ content and a:c was 0.5805 (Fig. 10c). Correlations with unit-cell volume (Fig. 10d) were similarly weak (R2 = 0.5763). All other unit cell ratios and angle parameters examined, including the a:b ratio (R2 = 0.4908), b:c ratio (R2 = 0.0125), α (R2 = 0.4051), and γ (R2 = 0.0859), correlated with Mn3+ even more weakly. Nevertheless, if we take at face value the correlations observed in this study between Mn3+ content and the a-axis and the  angle, we may infer that the Jahn-Teller distortions are oriented such that the octahedral elongation lies primarily parallel to a. When Mn3+ concentrations are sufficiently low, the distorted octahedra are isolated from one another and do not generate a bulk strain that violates the hexagonal symmetry. When Mn3+ concentrations

134

exceed ~25 mol%, then the strains couple and reinforce over longer scales such that the symmetry transitions from hexagonal to triclinic. As Mn3+ concentration increases further, the value of β moves further away from 90°. It is also interesting to note that the hexagonal birnessites, when refined with a triclinic unit cell, yielded γ angles ranging from 89.792° to 90.499°, not quite the value of 90° expected for a hexagonal unit cell. This result indicates that perhaps hexagonal birnessites may not be truly hexagonal, but that the broadness of the peaks allows for successful refinement within a hexagonal space group. Most importantly, these correlations suggest that a range of birnessite structures exist between end-member triclinic and hexagonal birnessite. As discussed below, however, the structural averaging associated with XRD allows the possibility that layers of endmember triclinic and hexagonal birnessite are inter-stratified beyond the resolution of XRD, yielding a refined triclinic cell that is intermediate between the endmembers in its dimensions.

Trends from quantitative EXAFS and IR analysis with Mn3+ Instead of modeling the local coordination environment around Mn using our EXAFS data, we applied linear combination fitting (LCF). LCF is frequently applied to natural or biologically cultured birnessite samples that tend to be poorly crystalline to determine the fraction of Mn oxide phases in mixtures of hexagonal birnessite, triclinic birnessite, and todorokite (Learman et al. 2011, 2013; Santelli et al. 2011; Zhao et al. 2016). Sspectral unmixing of FTIR spectra in the 400 to 750 cm-1 region is similarly sensitive to local structure and can yield comparable results (Ling et al. 2016, in prep). These techniques are necessary when poor crystallinity inhibits XRD differentiation of triclinic and hexagonal birnessite. As it is unclear whether our synthetic birnessites are intergrown mixtures of triclinic and hexagonal birnessite, or distinct structures of their own, LCF and spectral unmixing can offer a

135

sense of the triclinic and hexagonal character of a sample at the local scale. Therefore, we applied LCF analysis to our EXAFS data and spectral unmixing to our FTIR data in order to search for relationships between the ratio of triclinic to hexagonal birnessite in a mixture as a function of mole fraction of Mn3+, as our XRD suggested should exist. This approach yielded similar results for both the EXAFS and the FTIR datasets (Fig. 11, Table 6 – 7). Both analyses revealed that the fraction of Na-birnessite in a mixture correlated positively with the concentration of Mn3+, with R2 values of 0.860 and 0.943 for the EXAFS and IR datasets, respectively (Fig. 12). When the datasets are combined, the relationship between end-member triclinic Na-birnessite with Mn3+ is as follows: (1) where

is the wt fraction of triclinic Na-birnessite and

is the fraction of Mn3+. Based

on our limited data set, the error on the fraction of Mn3+ is within ±0.04 for FTIR, and within ±0.05 for EXAFS.

Distorted triclinic birnessite structures vs. intergrown triclinic-hexagonal birnessite layers The methods used to analyze birnessites in this study raise an interesting question: Are the birnessites studied here intergrown mixtures of triclinic and hexagonal birnessite endmembers, or does each sample possess a homogeneous structure? Among phyllosilicates, for instance, illite-smectite is commonly modeled as an intergrowth of two end-members according to the MacEwan crystallite model (Moore and Reynolds 1997). The averaging of multiple unit cells by XRD may make it impossible to distinguish between the two phases, instead yielding an averaged, distorted triclinic cell for cation-exchanged birnessites. Similar to XRD, EXAFS and FTIR provide averages of structures, but at smaller length scales over local atomic coordination environments. Consequently, as described by Webb et al.

136

(2005a), EXAFS structure models of the evolution from hexagonal to what they term “pseudoorthogonal” birnessite over time can be interpreted either as changes in proportions of triclinic and hexagonal birnessite or as distortions of the Mn-O bonds. Yet, no matter the method of interpretation, our LCF of EXAFS and spectral unmixing of FTIR data for our synthetic birnessites rendered reasonable fits and strong correlations with Mn3+ contents, as did the unit cell parameters obtained from Rietveld analyses. Consequently, either of two mechanisms may be operative for cation-exchange in synthetic birnessites. The first scenario involves delamination of sequential layers followed by cation exchange and then relamination, as described in Lopano et al. (2009) for Cs+ exchange into Na-birnessite. If this model applies to the present study, then we infer that only some Nacontaining interlayers exchanged with the aqueous K+, Ba2+, or Ca2+ introduced in our batch reactions while the remaining interlayers exchanged with H+. This hypothesis is plausible, since Ling et al. (2015) showed that triclinic Na-birnessite is unstable below pH ~8.2 in solutions in contact with an open atmosphere, implying that H+ competes with other cations for interlayer sites in solution at pH 7. Alternatively, if partial cation exchange occurred within all layers such that K+, Ba2+, or Ca2+ cations were uniformly distributed with H+ throughout the interlayer regions, it may be more appropriate to consider the fraction of triclinic birnessite, as described in Tables 6 and 7, as the degree of triclinic character in a strained structure. This degree of triclinicity therefore provides insight into the extent of the Jahn-Teller distortions of the Mn-O octahedra with increasing Mn3+ content.

Conclusions In this study, we show using a host of complementary diffraction and spectroscopy techniques that strong relationships among structure, symmetry, and Mn3+ content in synthetic

137

birnessite phases. Our X-ray diffraction analyses indicated that as Mn3+ concentration increases, the birnessite structure deforms, ultimately breaking the hexagonal symmetry when Mn3+ concentrations exceed ~25 mol%, to adopt a triclinic symmetry. EXAFS and FTIR spectroscopy re-inforced this connection between triclinicity and Mn3+ content, to a degree that we could posit a linear equation between the Mn3+ concentration and the mole fraction of triclinic Na-birnessite in a triclinic/hexagonal mixture. These observations raise the question of whether birnessite phases exist as triclinic and hexagonal end-members intergrown at the nanoscale, or as distinct structures with unit-cell dimensions that are intermediate between those of hexagonal or triclinic birnessite. Despite our inability to resolve this issue here, the relationships that we have developed for understanding Mn3+ in the context of crystal structure should prove useful for studies of reactions that are sensitive to Mn3+ content, such as transition metal redox reactions and photochemical oxidation of water (Takashima et al. 2012a,b; Simanova & Peña 2015).

Acknowledgements Funding for this work was provided by NSF Grant EAR-1147728, NSF Grant EAR1552211, and the Committee on Institutional Cooperation (CIC) and Smithsonian Institution Fellowship. The FTIR laboratory at the Smithsonian Institution was established with generous support from Stephen Turner. This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357.

138

References Bargar, J. R., Tebo, B. M., Bergman, U., Webb, S. M., Glatzel, P., Chiu, V. Q., & Villalobos, M. (2005). Biotic and abiotic products of Mn(II) oxidation by spores of the marine Bacillus sp. strain SG-1. American Mineralogist, 90, 143–154. http://doi.org/10.2138/am.2005.1557 Bartlett, R. J., & James, B. (1979). Behavior of Chromium in Soils: III. Oxidation. Journal of Environment Quality, 8(1), 31–35. http://doi.org/10.2134/jeq1976.00472425000500040010x Buerger, M. (1960). Crystal-structure Analysis. New York: Wiley & Sons, Inc. Chao, T. T., & Theobald, P. K. (1976). The significance of secondary iron and manganese oxides in geochemical exploration. Economic Geology, 71, 1560–1569. http://doi.org/10.2113/gsecongeo.71.8.1560 Cho, Y., Jang, S., Kim, Y., Komarneni, S., & Kim, S. (2011). Uptake of cadmium, copper, and lead by microporous synthetic Na-birnessite. Journal of Porous Materials, 18, 125–131. http://doi.org/10.1007/s10934-010-9430-0 Drits, V. A., Silvester, E., Gorshkov, A. I., & Manceau, A. (1997). Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite: II. Results from chemical studies and EXAFS spectroscopy. American Mineralogist, 82(9-10), 962–978. Feng, X. H., Zhu, M., Ginder-Vogel, M., Ni, C., Parikh, S. J., & Sparks, D. L. (2010). Formation of nano-crystalline todorokite from biogenic Mn oxides. Geochimica et Cosmochimica Acta, 74, 3232–3245. http://doi.org/10.1016/j.gca.2010.03.005 Ferreira, K. N., Iverson, T. M., Maghlaoui, K., Barber, J., & Iwata, S. (2004). Architecture of the Photosynthetic Oxygen-Evolving Center. Science, 303, 1831–1838. Finger, L. W., Cox, D. E., & Jephcoat, A. P. (1994). A Correction for Powder Diffraction Peak Asymmetry Due to Axial Divergence. Journal of Applied Crystallography, 27, 892–900. http://doi.org/10.1107/S0021889894004218 Fischer, T. B. (2011). Structural transformations of birnessite (δ-MnO2) during biological and abiological reduction. Fleeger, C. R., Heaney, P. J., & Post, J. E. (2013). A time-resolved X-ray diffraction study of Cs exchange into hexagonal H-birnessite. American Mineralogist, 98, 671–679. Freeman, D. S., & Chapman, W. G. (1971). An Improved Oxalate Method for the Determination of Active Oxygen in Manganese Dioxide. Analyst, 96, 865–869. Golden, D. C., Chen, C. C., & Dixon, J. B. (1986). Synthesis of todorokite. Science (New York, N.Y.), 231(4739), 717–719. http://doi.org/10.1126/science.231.4739.717 Hansel, C. M., Zeiner, C. A., Santelli, C. M., & Webb, S. M. (2012). Mn(II) oxidation by an ascomycete fungus is linked to superoxide production during asexual reproduction. Proceedings of the National Academy of Sciences, 109(31), 12621–12625. http://doi.org/10.1073/pnas.1203885109 Iyer, A., Del-Pilar, J., King’Ondu, C. K., Kissel, E., Garces, H. F., Huang, H., … Suib, S. L. (2012). Water Oxidation Catalysis using Amorphous Manganese Oxides, Octahedral Molecular Sieves (OMS-2), and Octahedral Layered (OL-1) Manganese Oxide Structures. Journal of Physical Chemistry, 116, 6474–6483. http://doi.org/10.1021/jp2120737 Jones, L. H. P., & Milne, A. A. (1956). Birnessite, a new manganese oxide mineral from Aberdeenshire, Scotland. The Mineralogical Magazine, XXXI(235), 283 – 288. Kijima, N., Yasuda, H., Sato, T., & Yoshimura, Y. (2001). Preparation and Characterization of Open Tunnel Oxide α-MnO2 Precipitated by Ozone Oxidation. Journal of Solid State Chemistry, 159, 94–102. http://doi.org/10.1006/jssc.2001.9136 Learman, D. R., Voelker, B. M., Madden, A. S., & Hansel, C. M. (2013). Constraints on

139

superoxide mediated formation of manganese oxides. Frontiers in Microbiology, 4, 1–11. http://doi.org/10.3389/fmicb.2013.00262 Learman, D. R., Voelker, B. M., Vazquez-Rodriguez, A. I., & Hansel, C. M. (2011). Formation of manganese oxides by bacterially generated superoxide. Nature Geoscience, 4, 95–98. http://doi.org/10.1038/ngeo1055 Learman, D. R., Wankel, S. D., Webb, S. M., Martinez, N., Madden, A. S., & Hansel, C. M. (2011). Coupled biotic–abiotic Mn(II) oxidation pathway mediates the formation and structural evolution of biogenic Mn oxides. Geochimica et Cosmochimica Acta, 75, 6048– 6063. http://doi.org/10.1016/j.gca.2011.07.026 Ling, F. T., Heaney, P. J., Post, J. E., & Gao, X. (2015). Transformations from triclinic to hexagonal birnessite at circumneutral pH induced through pH control by common biological buffers. Chemical Geology, 416, 1–10. http://doi.org/10.1016/j.chemgeo.2015.10.007 Lingane, J. J., & Karplus, R. (1987). New Method for Determination of Manganese. Industrial and Engineering Chemistry, 18(3), 191–194. Lopano, C. L., Heaney, P. J., & Post, J. E. (2009). Cs-exchange in birnessite: Reaction mechanisms inferred from time-resolved X-ray diffraction and transmission electron microscopy. American Mineralogist, 94, 816–826. http://doi.org/10.2138/am.2009.3068 Manceau, A., Gorshkov, A. I., & Drits, V. A. (1992). Structural chemistry of Mn, Fe, Co, and Ni in manganese hydrous oxides: Part I. Information from XANES spectroscopy. American Mineralogist, 77, 1133–1143. Manceau, A., Marcus, M. A., & Grangeon, S. (2012). Determination of Mn valence states in mixed-valent manganates by XANES spectroscopy. American Mineralogist, 97, 816–827. http://doi.org/10.2138/am.2012.3903 McEvoy, J. P., & Brudvig, G. W. (2006). Water-Splitting Chemistry of Photosystem II. Chemical Reviews, 106(11), 4455–4483. http://doi.org/10.1021/cr0204294 McKendry, I. G., Kondaveeti, S. K., Shumlas, S. L., Strongin, D. R., & Zdilla, M. J. (2015). Decoration of the layered manganese oxide birnessite with Mn(ii/iii) gives a new water oxidation catalyst with fifty-fold turnover number enhancement. Dalton Transactions (Cambridge, England : 2003). http://doi.org/10.1039/c5dt01436k McKeown, D. A., & Post, J. E. (2001). Characterization of manganese oxide mineralogy in rock varnish and dendrites using X-ray absorption spectroscopy. American Mineralogist, 86, 701–713. Retrieved from \\Scancyp\Literatur\Lit6-pdf\AmerMineral 86 McKeown_701713_01.pdf Meng, Y., Song, W., Huang, H., Ren, Z., Chen, S., & Suib, S. L. (2014). Structure − Property Relationship of Bifunctional MnO2 Nanostructures : Highly Efficient , Ultra-Stable Electrochemical Water Oxidation and Oxygen Reduction Reaction Catalysts Identified in Alkaline Media. Journal of the American Chemical Society, 136, 11452–11464. http://doi.org/10.1021/ja505186m Moore, D.M., Reynolds, R. C. J. (1997). Chapter 5: Individual Clay Minerals. In X-Ray Diffraction and the Identification and Analysis of Clay Minerals. 2nd Edition. (pp. 138– 203). Murray, J. W., Balistrieri, L. S., & Paul, B. (1984). The oxidation-state of manganese in marine sediments and ferromanganese nodules. Geochimica et Cosmochimica Acta, 48, 1237–1247. Retrieved from dx.doi.org/10.1016/0016-7037(84)90058-9 Murray, J. W., & Dillard, J. G. (1979). The oxidation of cobalt(II) adsorbed on manganese dioxide. Geochimica et Cosmochimica Acta, 43, 781–787. Najafpour, M. M., & Isaloo, M. A. (2015). The mechanism of water oxidation catalyzed by nanolayered manganese oxides: New insights. Journal of Photochemistry and Photobiology. B-Biology, 152, 133–138. http://doi.org/10.1016/j.jphotobiol.2015.01.009

140

Peña, J., Bargar, J. R., & Sposito, G. (2015). Copper sorption by the edge surfaces of synthetic birnessite nanoparticles. Chemical Geology, 396, 196–207. http://doi.org/10.1016/j.chemgeo.2014.12.021 Post, J. E. (1999). Manganese oxide minerals: Crystal structures and economic and environmental significance. Proceedings of the National Academy of Sciences, 96, 3447–3454. http://doi.org/10.1073/pnas.96.7.3447 Post, J. E., Heaney, P. J., & Hanson, J. (2002). Rietveld refinement of a triclinic structure for synthetic Na-birnessite using synchrotron powder diffraction. Powder Diffraction, 17(3), 218–221. Post, J. E., & Veblen, D. R. (1990). Crystal structure determinations of synthetic sodium, magnesium, and postassium birnessite using TEM and the Rietveld Method. American Mineralogist, 75, 477–489. Ramsey, M. S., & Christensen, P. R. (1998). Mineral abundance determination: Quantitative deconvolution of thermal emission spectra. Journal of Geophysical Research, 103(B1), 577–596. http://doi.org/10.1029/97JB02784 Ravel, B., & Newville, M. (2005). ATHENA , ARTEMIS , HEPHAESTUS : data analysis for Xray absorption spectroscopy using IFEFFIT. Journal of Synchrotron Radiation, 12, 537– 541. http://doi.org/10.1107/S0909049505012719 Rietveld, H. M. (1969). A profile Refinement Method for Nuclear and Magnetic Structures. Journal of Applied Crystallography, 2, 65–71. http://doi.org/10.1107/S0021889869006558 Robinson, D. M., Go, Y. B., Mui, M., Gardner, G., Zhang, Z., Mastrogiovanni, D., … Dismukes, G. C. (2013). Photochemical Water Oxidation by Crystalline Polymorphs of Manganese Oxides: Structural Requirements for Catalysis. Journal of the American Chemical Society, 135, 3494–3501. http://doi.org/10.1021/ja310286h Santelli, C. M., Webb, S. M., Dohnalkova, A. C., & Hansel, C. M. (2011). Diversity of Mn oxides produced by Mn(II)-oxidizing fungi. Geochimica et Cosmochimica Acta, 75, 2762– 2776. http://doi.org/10.1016/j.gca.2011.02.022 Saratovsky, I., Gurr, S. J., & Hayward, M. A. (2009). The Structure of manganese oxide formed by the fungus Acremonium sp. strain KR21-2. Geochimica et Cosmochimica Acta, 73, 3291–3300. http://doi.org/10.1016/j.gca.2009.03.005 Saratovsky, I., Wightman, P. G., Pastén, P. A., Gaillard, J.-F., & Poeppelmeier, K. R. (2006). Manganese Oxides: Parallels between Abiotic and Biotic Structures. Journal of the American Chemical Society, 128, 11188–11198. http://doi.org/10.1021/ja062097g Sayers, D. E., & Bunkers, B. A. (1988). X-ray Absorption Principles, Applications, Techniques of EXAFS, SEXAFS, and XANES. New York: Wiley. Scott, J. M., & Morgan, J. J. (1996). Reactions at oxide surfaces. 2. Oxidation of Se(IV) by synthetic Birnessite. Environmental Science & Technology, 30, 1990–1996. http://doi.org/10.1021/es950741d Silvester, E., Manceau, A., & Drits, V. A. (1997). Structure of synthetic monoclinic Na-rich birnessite and hexagonal birnessite: II. Results from chemical studies and EXAFS spectroscopy. American Mineralogist, 82, 962–978. Simanova, A. a., & Peña, J. (2015). Time-resolved investigation of cobalt oxidation by Mn(III)rich δ-MnO2 using quick X-ray absorption spectroscopy. Environmental Science & Technology, 49, 10867–10876. http://doi.org/10.1021/acs.est.5b01088 Stephens, P. W. (1999). Phenomenological model of anisotropic peak broadening in powder diffraction. Journal of Applied Crystallography, 32, 281–289. http://doi.org/10.1107/S0021889898006001 Takashima, T., Hashimoto, K., & Nakamura, R. (2012a). Inhibition of Charge Disproportionation of MnO2 Electrocatalysts for Efficient Water Oxidation under Neutral Conditions. Journal

141

of the American Chemical Society, 134, 18153–18156. http://doi.org/10.1021/ja306499n Takashima, T., Hashimoto, K., & Nakamura, R. (2012b). Mechanisms of pH-Dependent Activity for Water Oxidation to Molecular Oxygen by MnO2 Electrocatalysts. Journal of the American Chemical Society, 134, 1519–1527. http://doi.org/10.1021/ja206511w Tan, H., Zhang, G., Heaney, P. J., Webb, S. M., & Burgos, W. D. (2010). Characterization of manganese oxide precipitates from Appalachian coal mine drainage treatment systems. Applied Geochemistry, 25, 389–399. http://doi.org/10.1016/j.apgeochem.2009.12.006 Thompson, P., Cox, D. E., & Hastings, J. B. (1987). Rietveld Refinement of Debye-Scherrer Synchrotron X-ray Data from A1203. Journal of Applied Crystallography, 20(2), 79–83. http://doi.org/10.1107/S0021889887087090 Thomson, J. L., & Salisbury, J. W. (1993). The Mid-Infrared Reflectance of Mineral Mixtures (714 µm). Remote Sensing of Environment, 45, 1–13. http://doi.org/10.1016/00344257(93)90077-B Toby, B. H., & Von Dreele, R. B. (2013). GSAS-II: the genesis of a modern open-source all purpose crystallography software package. Journal of Applied Crystallography, 46(2), 544– 549. http://doi.org/10.1107/S0021889813003531 Umena, Y., Kawakami, K., Shen, J.-R., & Kamiya, N. (2011). Crystal structure of oxygenevolving photosystem II at a resolution of 1.9 Å. Nature, 473, 55–60. http://doi.org/10.1038/nature09913 Usui, A., & Mita, N. (1994). IX. Mineralogy, geochemistry and internal growth structure of manganese nodules in the western part of the Penrhyn Basin, South Pacific (GH83-3 area). Geological Survey of Japan Cruise Report, (23), 165–185. Villalobos, M., Lanson, B., Manceau, A., Toner, B., & Sposito, G. (2006). Structural model for the biogenic Mn oxide produced by Pseudomonas putida. American Mineralogist, 91, 489– 502. http://doi.org/10.2138/am.2006.1925 Villalobos, M., Toner, B., Bargar, J., & Sposito, G. (2003). Characterization of the manganese oxide produced by Pseudomonas putida strain MnB1. Geochimica et Cosmochimica Acta, 67(14), 2649–2662. http://doi.org/10.1016/S0016-7037(03)00217-5 Wang, Y., Feng, X., Villalobos, M., Tan, W., & Liu, F. (2012). Sorption behavior of heavy metals on birnessite: Relationship with its Mn average oxidation state and implications for types of sorption sites. Chemical Geology, 292-293, 25–34. http://doi.org/10.1016/j.chemgeo.2011.11.001 Weaver, R. M., & Hochella, M. F. J. (2003). The reactivity of seven Mn-oxides with Cr 3 + aq : A comparative analysis of a complex , environmentally important redox reaction. American Mineralogist, 88, 2016–2027. Webb, S. M. (2005). SIXPack : a Graphical User Interface for XAS Analysis Using IFEFFIT. Physica Scripta, T115, 1011–1014. Webb, S. M., Tebo, B. M., & Bargar, J. R. (2005). Structural characterization of biogenic Mn oxides produced in seawater by the marine bacillus sp. strain SG-1. American Mineralogist, 90(8-9), 1342–1357. http://doi.org/10.2138/am.2005.1669 Webb, S. M., Tebo, B. M., & Barger, J. R. (2005). Structural Influences of Sodium and Calcium Ions on the Biogenic Manganese Oxides Produced by the Marine Bacillus Sp., Strain SG-1. Geomicrobiology Journal, 22, 181–193. http://doi.org/10.1080/01490450590946013 Zhao, H., Zhu, M., Li, W., Elzinga, E. J., Villalobos, M., Liu, F., … Sparks, D. L. (2016). Redox Reactions between Mn(II) and Hexagonal Birnessite Change its Layer Symmetry. Environmental Science & Technology, 1–33. http://doi.org/10.1021/acs.est.5b04436 Zhao, W., Tan, W., Feng, X., Liu, F., Xie, Y., & Xie, Z. (2011). XAFS studies on surface coordination of Pb2+ on birnessites with different average oxidation states. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 379, 86–92.

142

http://doi.org/10.1016/j.colsurfa.2010.11.040

143

Figures

Figure 5-1. Schematic diagram of (a) synthetic triclinic Na-birnessite and (b) synthetic hexagonal H-birnessite after Lanson et al. (2000).

144

Figure 5-2. XRD patterns of synthetic birnessites.

145

* hexagonal-like peaks in Ca-birnessite Figure 5-3. FTIR spectra of (a) hexagonal birnessites and (b) triclinic birnessites in the Mn-O lattice range from 400 to 750 cm-1.

146

Figure 5-4. EXAFS k2 χ(k) data for (a) standard triclinic Na-birnessite, (b) standard hexagonal pH 2 birnessite, (c) hexagonal HEPES birnessite, and (d) hexagonal pH 3 birnessite.

147

Figure 5-5. Radial distribution functions for (a) standard triclinic Na-birnessite, (b) standard hexagonal pH 2 birnessite, (c) hexagonal HEPES birnessite, and (d) hexagonal pH 3 birnessite.

148

Figure 5-6. EXAFS k2 χ(k) data for (a) standard triclinic Na-birnessite, (b) standard hexagonal pH 2 birnessite, (c) K-birnessite, (d) Ca-birnessite, and (e) Ba-birnessite.

149

Figure 5-7. Radial distribution functions for (a) standard triclinic Na-birnessite, (b) standard hexagonal pH 2 birnessite, (c) K-birnessite, (d) Ca-birnessite, and (e) Ba-birnessite.

150

Mn4+

Mn3+

HB pH 2

0.68

HB pH 3

0.646

HEPES HB

0.63

CaB

0.64

Mn AOS

Mn2+ 0.1

3.58

0.21

0.15

3.50

0.22

0.15

3.48

0.22

0.29

0.07

3.57

KB

0.58

0.38

0.01 3.55

BaB

0.56

0.40

0.01 3.52

NaB

0.60 0

0.2

0.38 0.4

0.6

0.8

0.02 3.58 1

Figure 5-8. XPS results showing fractions of Mn4+, Mn3+, and Mn2+ for each synthetic sample from fitting the Mn3p peak, along with an average Mn average oxidation state.

151

Figure 5-9. Representative Rietveld refinements and difference curves for (a) Ca-birnessite and (b) pH 3 hexagonal H-birnessite.

152

104

(a)

y = 51.88x + 80.86 R² = 0.90

102 100

0.72

(c)

NaB BaB

0.71

y = 0.11x + 0.66 R² = 0.58

0.7

CaB

KB

β 96

0.69

94

HB2 HB3

92 0

0.1

0.3

HB3

0.68

HEPESB

0.2

HB2 0.4

0.67

0.5

0

0.1

Mn3+ (d)

0.3

0.4

0.5

110

NaB

y = 1.11x + 4.69 R² = 0.89

y = 20.00x + 96.96 R² = 0.58

108

BaB

5.1

KB

106

Vol

KB

a

0.2

Mn3+

5.2 5.15

KB

HEPESB

CaB

90

(b)

NaB

a:c

98

BaB

5.05

NaB 104

5 102

HB3

4.95

HEPESB HB2

4.9 0

0.1

0.2

0.3

Mn3+

BaB

HB3

CaB

HB2

CaB

HEPESB 100 0.4

0.5

0

0.1

0.2

0.3

0.4

0.5

Mn3+

Figure 5-10. Correlations between Mn3+ and the refined parameters: (a) β angle, (b) the a parameter, (c) volume, and (d) the a:c ratio.

153

BaB

1.00

FTIR Fraction Na-Birnessite

NaB KB

0.80

0.60 CaB 0.40

0.20 HB pH 3

HEPES HB

HB pH 2 0.00 0.00 0.20

0.40

0.60

0.80

1.00

EXAFS Fraction Na-birnessite

Figure 5-11. Comparison of the fitted fraction of triclinic Na-birnessite for EXAFS data vs. FTIR data using endmember triclinic Na-birnessite and end-member pH 2 hexagonal H-birnessite.

154

Fraction Na-birnessite

1.0 0.8 0.6 0.4 0.2 0.0 0.20

0.30 XPS Fraction Mn3+ EXAFS

y = 4.30x - 0.83 R² = 0.8359

0.40

FTIR y = 5.13x - 1.04 R² = 0.9779

Figure 5-12. Correlation between the fraction of triclinic Na-birnessite determined by EXAFS and FTIR to the fraction of Mn3+ determined from XPS in a sample.

155

Tables Table 6-1. Peak locations in the Mn-O range (400 to 750 cm-1) for synthetic birnessites. Hexagonal

Peak 1

Peak 2

Peak 3

Hexagonal Birnessite pH 2 Hexagonal Birnessite pH 3 HEPES Birnessite pH 7 Triclinic Na-birnessite K-birnessite Ba-birnessite Ca-birnessite

453 440 447 Peak 2 478 479 483 487

496 494 491 Peak 3 511 513 512 515

659 659 661 Peak 4 639 635 640 641

Peak 1 418 415 418 419

156

Table 6-2. Rietveld refinement results of unit cell parameters for triclinic cation-exchanged birnessites using a triclinic unit cell.

Space group Unit cell a (Å) b (Å) c (Å) α (°) β (°) γ (°) vol (Å) Refinement No. of diffraction points No. of reflections Diffraction range (dspacing, Å) No. of variables R(F2) Rwp

Na-birnessite K-birnessite Ba-birnessite Ca-birnessite C -1 C -1 C -1 C -1 5.166(5) 2.8479(9) 7.343(7) 89.61(9) 103.15(5) 90.07(9) 105.2(1)

5.076(7) 2.915(2) 7.36(2) 89.8(1) 98.6(1) 88.80(9) 107.7(3)

5.13(1) 2.838(2) 7.24(2) 90.2(1) 101.86(9) 89.92(9) 103.2(2)

4.9799 2.914(5) 7.07(1) 90.8(2) 94.5(1) 92.25(9) 102.2(2)

2046 175

1783 116

1783 115

1821 124

0.86 - 5.09 25 0.04762 0.02438

0.95 - 5.09 26 0.0244 0.03116

0.95 - 5.09 28 0.0531 0.01601

0.94 - 5.09 27 0.07535 0.01105

157

Table 6-3. Atom positions for triclinic cation-exchanged birnessites from Rietveld refinements.

Atom Mn(oct) O(oct) Na or H2O(int)

x 0 0.3766(9) 0.567(2)

y 0 0.018(7) 0.196(2)

z 0 0.1366(5) 0.502(2)

Site occupancy factor 1 1 0.6378

K-birnessite

Mn(oct) O(oct) K or H2O(int)

0 0.350(2) 0.699(8)

0 0.041(6) 0.201(7)

0 0.122(2) 0.512(4)

0.92(2) 1 0.44(3)

0.00583 0.02585 0.03153

Ba-birnessite

Mn(oct) O(oct) Ba or H2O(int)

0 0.362(2) 0.677(3)

0 0.00(1) 0.406(9)

0 0.119(1) 0.497(1)

0.84(1) 1 0.82(2)

0.00583 0.02585 0.03153

Ca-birnessite

Mn(oct) O(oct) Ca or H2O(int)

0 0.341(2) 0.664(5)

0 0.023(4) 0.30(1)

0 0.104(2) 0.407(6)

0.79(2) 1 0.38(2)

0.00583 0.02585 0.03153

Na-birnessite

Uiso 0.00583 0.02585 0.03153

158

Table 5-4. Rietveld refinement results of unit cell parameters for hexagonal birnessites using a hexagonal unit cell (P -3) and a triclinic unit cell (C -1). Hexagonal unit cell

Triclinic unit cell

pH 2 Hbirnessite P -3

pH 3 Hbirnessite P -3

HEPES pH 7 birnessite P -2

pH 2 Hbirnessite C -1

pH 3 Hbirnessite C -1

HEPES pH 7 birnessite C -1

2.863(3) 7.389(5)

2.890(2) 7.352(4)

52.45(7)

2.860(2) 7.323(4) 90 90 120 51.87(5)

53.19(6)

4.9205(6) 2.845(1) 7.29(1) 88.7(1) 92.5(1) 90.2(1) 101.9(2)

4.952(3) 2.837(3) 7.265(7) 88.72(9) 92.3(1) 89.8(1) 102.0(2)

4.92344 2.856(1) 7.170(9) 89.1(1) 92.0(1) 90.5(1) 100.7(2)

Refinement No. of diffraction points No. of reflections

1820 43

1821 43

1820 45

1820 115

1821 116

1820 111

Diffraction range (d-spacing, Å) No. of variables R(F2) Rwp

0.94 - 5.09 22 0.08414 0.01391

0.94 - 5.09 23 0.0894 0.01363

0.94 - 5.09 23 0.02987 0.02333

0.94 - 5.09 26 0.15801 0.01829

0.94 - 5.09 30 0.17679 0.0182

0.94 - 5.09 27 0.12156 0.0332

Space group Unit cell a (Å) b (Å) c (Å) α (°) β (°) γ (°) vol (Å)

159

Table 5-5. Atom positions for hexagonal birnessites from Rietveld refinements using both hexagonal and triclinic unit cells.

Atom

x y Hexagonal unit cell 0 0 0.3333 0.6667 0.6667 0.3333 0 0

z

Site occupancy factor

Uiso

0 0.114(2) 0.55(2) 0.695(6)

0.780(2) 1 0.30(1) 0.068(3)

0.00637 0.00082 0.10105 0.0088

0 0.116(2) 0.51(2) 0.696(4)

0.79(1) 1 0.34(3) 0.074(6)

0.00637 0.00082 0.10105 0.0088

pH 2 H-birnessite

Mn(oct) O(oct) H2O(int) Mn(int)

pH 3 H-birnessite

Mn(oct) O(oct) H2O(int) Mn(int)

HEPES pH 7 birnessite

Mn(oct) O(oct) H2O(int) Mn(int)

pH 2 H-birnessite

Mn(oct) O(oct) Mn or H2O(int)

pH 3 H-birnessite

Mn(oct) O(oct) Mn or H2O(int)

0 0.322(3) 0.769(4)

0 0.055(4) 0.355(8)

0 0.091(2) 0.391(4)

0.57(3) 1 0.47(3)

0.00583 0.02585 0.03153

HEPES pH 7 birnessite

Mn(oct) O(oct) Mn or H2O(int)

0 0.303(3) 0.772(3)

0 0.074(3) 0.332(6)

0 0.118(2) 0.408(3)

0.57 1 0.56(2)

0.00583 0.02585 0.03153

0 0.3333 0.6667 0

0 0.6667 0.3333 0

0 0 0 0.82(2) 0.00637 0.3333 0.6667 0.127(2) 1 0.00082 0.6667 0.3333 0.56(1) 0.32(2) 0.10105 0 0 0.679(4) 0.089(2) 0.0088 Triclinic unit cell 0 0 0 0.57(3) 0.00583 0.327(2) 0.058(3) 0.084(2) 1 0.02585 0.797(5) 0.324(7) 0.380(3) 0.40(2) 0.03153

160

Table 5-6. EXAFS linear combination fitting results for fractions of end-member triclinic Nabirnessite and end-member pH 2 hexagonal H-birnessite. XAFS Linear Combination Fitting Sample Triclinic Na-birnessite (endmember) K-birnessite Ba-binessite Ca-birnessite Hexagonal pH 2 H-birnessite (endmember) HEPES pH 7 birnessite pH 3 H-birnessite

Fraction Nabirnessite

Fraction pH 2 H-birnessite

R-factor

1.000 0.60 ± 0.02 0.81 ± 0.02 0.66 ± 0.02

0.000 0.40 ± 0.02 0.19 ± 0.03 0.33 ± 0.03

N/A 0.0144 0.0178 0.0171

0.000 0.14 ± 0.02 0.00 ± 0.00

1.000 0.86 ± 0.03 1.00 ± 0.00

N/A 0.0221 0.0246

161

Table 5-7. FTIR spectral unmixing results for fractions of end-member triclinic Na-birnessite and end-member pH 2 hexagonal H-birnessite. IR Spectral Unmixing Sample

Fraction Na- Fraction pH 2 Hbirnessite birnessite

R-factor

Triclinic Na-birnessite (endmember) K-birnessite Ba-binessite Ca-birnessite

1.000 0.891 ± 0.005 1.000 ± 0.008 0.406 ± 0.005

0.000 0.109 ± 0.004 0.000 ± 0.007 0.594 ± 0.005

N/A 0.0058 0.0219 0.0058

Hexagonal pH 2 H-birnessite (endmember) pH 3 H-birnessite HEPES pH 7 birnessite

0.000 0.101 ± 0.003 0.118 ± 0.004

1.000 0.899 ± 0.003 0.882 ± 0.003

N/A 0.0024 0.0038

162

Supplementary Information B

163

Chapter 6 A characterization of natural terrestrial birnessites

Abstract With a focus on a large set of natural birnessites collected from terrestrial, freshwater systems, we applied and compared the capabilities of X-ray diffraction (XRD), extended X-ray absorption fine structure (EXAFS), Fourier-transform infrared spectroscopy (FTIR), and X-ray photoelectron spectroscopy (XPS) to characterize crystal structure and chemistry. Using XRD, we successfully identified 3 of the 11 natural birnessite samples as hexagonal ranciéite-like phases, but the remaining samples yielded less interpretable “3-line” diffraction patterns with broad, asymmetrical peaks at d-spacings of ~7.2 Å, ~2.4 Å, and ~1.4 Å. EXAFS analysis suggested that many of these samples had characteristics of both triclinic and hexagonal birnessite. However, application of EXAFS to the ranciéite-like phases yielded unreasonably high concentrations of triclinic birnessite as an intergrowth, calling into question the use of synthetic hexagonal H-birnessite as an appropriate standard in the linear combination fitting of EXAFS data for natural birnessites. FTIR spectroscopy of the “3-line” birnessite samples successfully distinguished triclinic and hexagonal constituents, and analyses of peak positions suggested that natural birnessites occur as a full spectrum of triclinic and hexagonal intergrowths. XPS analysis of these samples revealed that higher Mn3+ concentrations relative to Mn2+ and Mn4+ are correlated to increased proportions of triclinic birnessite.

164

Introduction Manganese oxides exist naturally as fine-grained sediments in soils, as nodules in lakes and oceans, and as coatings on rock surfaces (J.E. Post, 1999). Their fine grain size and high surface areas result in their high reactivity, and Mn oxides consequently play a vital role in heavy metal cycling and redox reactions occurring in the environment (Paterson et al. 1986; Lopano et al. 2007, 2009, 2011; Fleeger et al. 2013). For instance, Mn oxides can sorb Zn, Sn, and Ni (Chao and Theobald 1976; Usui and Mita 1994), and they can oxidize Se4+ to Se6+ and Co2+ to Co3+, just a few of the redox reactions in which Mn oxides are involved (Bartlett and James 1979; Murray and Dillard 1979; Oscarson et al. 1981; Scott and Morgan 1996; Kay et al. 2001; Fandeur et al. 2009; Lafferty et al. 2011; Peacock and Moon 2012; Kazakis et al. 2015). The layered Mn oxide minerals in the birnessite family are widely studied for their natural occurrence in a range of terrestrial settings and for their chemical reactivity (Potter and Rossman 1979; Post 1999; McKeown and Post 2001; Manceau et al. 2002, 2007; Weaver and Hochella 2003). Birnessites are thought to exist as two possible symmetries, triclinic and hexagonal (Fig. 1). According to Post and Veblen (1990) and Post et al. (2002), synthetic triclinic Na-birnessite (Na0.58(Mn4+1.42Mn3+0.58)O4 • 1.5H2O) consists of layered Mn octahedral sheets with ~29% Mn3+ and ~71% Mn4+, along with hydrated Na+ cations in the interlayer sites. In contrast, Silvester et al. (1997) describe synthetic hexagonal H-birnessite (H0.33Mn3+0.111Mn2+0.055(Mn4+0.722Mn3+0.111☐0.167)O2) as a Mn oxide with Mn octahedral sheets consisting of ~72% Mn4+ cations, ~11% Mn3+ cations, and ~17% vacancies, along with Mn2+, Mn3+, and H+ cations that occupy the interlayer near the octahedral vacancy sites. Since the Mn oxidation states and the vacancy concentrations are different for triclinic and hexagonal birnessite, they can behave differently in similar geochemical environments. The vacancies in hexagonal birnessite, for example, act as sites for the sorption of Pb and Zn (Kwon

165

et al. 2010; Toner et al. 2006). Similarly, Mn3+ in triclinic birnessite plays a role in the oxidation of Cr3+ to Cr6+ and Co2+ to Co3+ (Manceau et al. 1992; Weaver and Hochella 2003; Simanova and Peña 2015). Identifying the structural characteristics of a natural birnessite sample can help predict which chemical processes will predominate. The characterization of natural birnessites has been challenged by their small particle size and poor crystallinity. X-ray diffraction (XRD), a standard method for phase identification, often falls short of properly identifying natural Mn oxides due to the absence of long-range order (Post 1999). Birnessite, specifically, is distinguished by its ~7 Å d-spacing. In many natural samples, the birnessite may initially be identified as buserite, the hydrated relative of birnessite that has a distinct ~10 Å spacing. Upon heating buserite to 110°C, the 10 Å spacing collapses as interlayer water is lost, yielding the ~7 Å spacing of birnessite. This collapse distinguishes birnessite from todorokite, which has a ~10 Å peak that does not collapse upon heating at 110°C (Usui & Mita, 1994). This method of identifying natural birnessites is widely utilized (Post 1999; Bilinski et al. 2002; Zhao et al. 2012; Frierdich et al. 2011), but when diagnostic X-ray diffraction peaks are not evident, XRD may not be able to reveal whether a birnessite is triclinic or hexagonal. The situation is complicated further by the tendency of researchers to employ Cu rather than Mo Xray radiation, since the absorption edge for Cu is so close to that of Mn that absorption and fluorescence greatly degrade the signal-to-noise ratio (Fig. 2). Instead, synchrotron X-ray absorption spectroscopy (XAS) has been employed extensively for the structural characterization of Mn oxides by focusing on the bonding environment surrounding Mn atoms. XAS has been used to identify the phase distributions in poorly crystalline Mn oxides precipitated by bacteria and fungi in laboratory cultures. Unlike Xray diffraction analysis of “3-line” birnessite, XAS can offer insights into the relative concentrations of triclinic and hexagonal birnessite through linear combination fitting (LCF) and/or modeling of the extended absorption fine structure (EXAFS) or X-ray absorption near

166

edge structure (XANES) regions (Jürgensen et al. 2004; Webb et al. 2005a, 2005b; Villalobos et al. 2006; Bargar et al. 2009; Saratovsky et al. 2009; Learman et al. 2011; Santelli et al. 2011). Most studies of natural birnessites in low-temperature environments presume that the phases are directly or indirectly biogenic, and that biogenic birnessite precipitates at least initially as hexagonal birnessite. This belief originates from laboratory experiments in which hexagonal birnessite has consistently been observed as the first phase produced (Villalobos et al. 2003, 2006; Bargar et al. 2005; Saratovsky et al. 2006; Learman et al. 2011; Santelli et al. 2011; Hansel et al. 2012), although “pseudo-orthogonal” birnessite and the tunnel-structured Mn oxide todorokite have been reported as secondary crystallization phases (Webb et al. 2005b; Feng et al. 2010). The validity of these presumptions is challenged by observations that both triclinic and hexagonal birnessite can co-exist in natural environments (Tan et al. 2010). Moreover, Ling et al. (2015) showed that biological buffers can force the formation of hexagonal birnessite, potentially skewing results of bioprecipitation experiments with bacteria and fungi. By characterizing eleven birnessite samples from a variety of natural environments using a suite of analytical approaches, here we explore the assumption that most natural birnessites are hexagonal. We interrogated the samples with scanning electron microscopy coupled with electron dispersive spectroscopy (SEM/EDS), electron microprobe analysis (EPMA), X-ray diffraction, X-ray absorption spectroscopy (EXAFS and XANS), and X-ray photoelectron spectroscopy (XPS) to survey the chemical compositions and Mn oxidation state ratios of naturally occurring birnessites. Following the demonstration by Ling et al. (2016) that Fouriertransform infrared spectroscopy (FTIR) can differentiate between synthetic triclinic and hexagonal birnessites, we also applied FTIR to natural birnessites, and we evaluated the relative accuracy and ease of these techniques to gauge their suitability for birnessite characterization.

167

Methods

Samples Synthesis of standard triclinic birnessite. Triclinic Na-birnessite was synthesized according to the procedure described in Golden et al. (1986). A 200 ml solution of 0.5 M MnCl2 (Mallinckrodt Baker) was mixed with 250 ml of 5.5 M NaOH (J.T. Baker). The mixture was oxygenated through a glass frit for ~5 hrs at a rate of 1.5 L/min. The precipitate was divided evenly and centrifuged in 14 centrifuge tubes. The solution was decanted and replaced with pH 6.49 deionized (DI) water to rinse. The rinse cycle was repeated five times. Na-birnessite was stored in ~350 ml DI water until experimental use. For experiments, aliquots of Na-birnessite were filtered with a 0.05 μm Nuclepore Track-Etched polycarbonate membrane filter (Whatman), rinsed three times with 100 ml DI water, and left to air-dry at room temperature. Synthesis of standard hexagonal birnessite. Our hexagonal birnessite standard was synthesized by reacting ~100 mg of dried triclinic Na-birnessite in 100 ml of 0.01 M HCl for 24 hrs. The H-birnessite then was filtered with a 0.05 μm Nuclepore Track-Etched polycarbonate membrane filter (Whatman), rinsed three times with 100 ml DI water, and left to air-dry. X-ray diffraction with a Rigaku D/MAX-RAPID microdiffractometer using a Mo tube source (λ = 0.71 Å) confirmed the synthesis of both triclinic and hexagonal birnessite. Synthetic fungal sample. The fungal species Stagnospora sp. SRC11sM3a was obtained from a passive acid mine drainage remediation system in Central Pennsylvania that effectively removed high concentrations of Mn through precipitation of Mn oxides as described in Santelli et al. (2010). The culture was inoculated in a 20 mM pH 7 HEPES-buffered AY medium that consisted of 0.25 g L-1 sodium acetate, 0.15 g L-1 yeast extract, 1 mL L-1 trace elements stock (10 mg L-1 CuSO4•5H2O, 44 mg L-1 ZnSO4•7H2O, 20 mg L-1 CoCl2•6H2O, 13 mg L-1

168

Na2MoO4•2H2O), and supplemented with 1.5 mM MnCl2 and 50 mM CaCl2. Mn was held constant during growth for ~20 days, but not supplemented after. The selected conditions were chosen to simulate the average geochemical conditions measured in the treatment systems during sampling. The culture was grown in 50 mL of liquid AY medium without agitation to resemble a submerged environment, and incubated in the dark at room temperature for ~20 – 24 days and stored in dark conditions to prevent photoreduction of Mn oxides. Natural samples. We collected four birnessite samples from passive coal mine drainage treatment sites near the central Pennsylvania towns of De Sale (near Eau Claire, Butler Co.) (labeled DS1-M3f and DS2-M3f) and Central City in Somerset Co. (labeled PBS-M3f-1 and PBS-M3f-2 ) (Table 2). The treatment systems consist of rectangular beds filled with crushed limestone to ~1 m in depth. The coal mine drainage horizontally flowed through the limestone beds on which Mn oxide sediments precipitated. Further descriptions of sampling methods and localities for the De Sale and the Central City specimens are described in Tan et al. (2010) and Luan et al. (2012). Likewise, we collected another birnessite sample (labeled Glasgow) from a similar passive coal mine drainage treatment site in Glasgow near Tyrone, Blair Co., PA. The powder was recovered from a Metal Removal Unit (MRU) built by EcoIslands LLC. The MRUs consisted of ~6 ft x 4 ft plywood boxes, with each box divided into 3 sections containing coconut coir (Fig. 3). The MRUs were partially drained to allow settling of sediment particles prior to sample collection. Mn oxide coatings were scraped off the coconut coir for analysis. Another birnessite sample (labeled Spring Branch) originated as a Mn oxide coating on limestone cobbles from a stream in Spring Branch, TN, and this site was deemed a natural environmental analog to the Pennsylvania acid mine remediation sites. In contrast, a sample from Vermilion, MN (kindly provided by Michael G. Sommers) derived from Lake Vermilion, MN. The sampling method and site description are detailed in Sommers et al. (2002), with this

169

particular sample described as a “reef-type stromatolite” based on its reef-like shape and resemblance to biotically precipitated specimens. The remaining samples were provided by the Smithsonian’s Mineral Reference Collection in the Department of Mineral Sciences, National Museum of Natural History. Mn oxides from Paxton Cave, VA (Smithsonian Institution NMNH #160078) were identified as ranciéite in previous studies (Richmond et al. 1969). Ranciéite from Spain originated in an abandoned open pit Fe/Mn mine near Trevelez, Las Alpujaras, Granada, Spain. Finally, the one birnessite sample that might have formed at slightly elevated temperatures (Smithsonian Institution NMNH #128319) derived from a hydrothermal vein in the French Pyrenees.

X-ray diffraction (XRD) All samples were ground in an agate mortar under acetone to disaggregate clumps. For XRD analysis, ~2 mg of sample were mounted on glass fibers into a Rigaku D/MAX-RAPID

microdiffractometer with an imaging plate detector (Smithsonian Institute, Mineral Sciences Department) and a Mo tube source (λ = 0.71 Å). Samples were rotated 360° around the phi axis at 1° s-1 during data collection with a 10 min exposure time.

Scanning electron microscopy/energy dispersive spectroscopy (SEM/EDS) A FEI Nova NanoSEM 600 (Department of Mineral Sciences, Smithsonian Institution) operating at an accelerating voltage of 15 keV and a beam current of 1 to 2 nA, and equipped with a ThermoFisher energy dispersive X-ray detector (EDS), was used for elemental analyses of the birnessite samples. The data were processed using the Noran System Six 3 (NSS 3) software. Samples were initially prepared by putting carbon tape on top of an aluminum stub, and placing

170

the sample onto the tape for chemical analysis and imaging. The samples were also prepared by creating polished sections using cold-set epoxy to avoid heating. Samples were not C coated, and the EDS data were collected using the low-vacuum mode for chemical analysis.

Electron probe microanalysis (EPMA) The polished sections of samples used for SEM/EDS were carbon coated for analysis on an electron probe microanalyzer. A JEOL 8900 electron probe microanalyzer (Department of Mineral Sciences, Smithsonian Institution) was used to determine chemical compositions. The instrument was equipped with five wavelength-dispersive spectrometers (WDS), and was operated at 15 kV accelerating voltage and 20 nA beam current. Electron probe point analyses were collected for 15 elements using a 10 x 12 µm beam, with analytical details described in Table 1. Chemical formulas for all birnessite samples are reported in Table 2. Note that for all birnessites, EPMA measurements were complicated by the lack of smooth, flat surfaces due to small particle sizes and the flaky nature of the materials, making surface polishing difficult during sample preparation. Consequently, weight percents often totaled to ~60%, although the low totals in part resulted from the initial assumption that Mn was divalent as MnO during analyses, the presence of water in the sample, and the inability to measure H+. In general, the calculations of chemical formulas (Table 2) assumed that all Mn had a valence of 4+ , except for those samples that we also analyzed by X-ray photoelectron spectroscopy, which provided additional information regarding Mn oxidation state ratios.

171

Fourier transform infrared spectroscopy (FTIR) Samples were ground under acetone in a mortar and pestle and sieved through a 325 mesh sieve. Then, 0.5 to 1 mg of Mn oxide sample was milled with ~250 mg KBr using a SPECAC ball mixing mill for 1-2 min, and pressed into a pellet. Vibrational spectra were collected on a Nicolet 6700 Analytical FTIR Spectrometer from 400 to 4000 cm-1. The resolution was set at 3.86 cm-1 and 64 scans were co-added for each spectrum. The Omnic 8 software (Nicolet) was used to view data during data collection.

X-ray absorption spectroscopy/extended X-ray absorption fine structure (XAS/EXAFS) For X-ray absorption spectroscopy, samples were ground under acetone in a mortar and pestle, and sieved into a thin layer with a 425 or 500 mesh sieve onto kapton tape. The kapton tape was then folded over to seal in the sample. Manganese K-edge XAS spectra were collected using a synchrotron source at Beamline 12-BM of the Advanced Photon Source (APS), Argonne National Laboratory using a Si(111) double-crystal, fixed exit monochromator and a double mirror system (flat plus torroidal) with an energy cutoff of 23 keV. The pre-edge peak of a Mn foil was used for energy calibration (6539 eV). Fluorescence data were collected with a 13element Ge detector and Cr(III) foil in front of the Ge detector to eliminate scatter in all samples. Three to six scans were collected per sample at room temperature from -200 eV to about +800 eV around the Mn K-edge (6539 keV). During data collection, the peak positions and line forms in the near edge region (XANES) were examined to check for photochemical reduction with successive scans, and no changes were observed. Data analyses of spectra were conducted using the ATHENA software (Ravel & Newville 2005). XAS spectra were calibrated using a Mn foil, averaged, background-subtracted,

172

normalized, and deglitched if outlier points were present. Analysis of the Mn K-edge EXAFS region was used for identification and phase fraction quantification of samples (Webb et al. 2005a, 2005b; Villalobos et al. 2006; Saratovsky et al. 2009; Feng et al. 2010). The χ(k) spectra were converted to k (Å-1) (Sayers & Bunkers, 1988). The resulting χ(k) data were k2-weighted and analyzed using the k-range from 2.3 to 11.3 Å-1. Principal component analysis (PCA) was used to determine the number of components representing the entire data set (Manceau et al. 2002). Linear combination fitting (LCF) of the χ(k) spectra collected for natural birnessites yielded phase fractions of Mn oxides. The following standards were used to fit the EXAFS data: todorokite from South Africa (Smithsonian Insitution #NMNH R15434), romanechite from Van Horne, TX (Smithsonian Insitution #NMNH 97618), lithiophorite from South Africa (Smithsonian Insitution #NMNH 162391), cryptomelane from India (Smithsonian Insitution #NMNH 89104), chalcophanite from Sterling, NJ (Smithsonian Insitution #NMNH C1814), manganosite from Franklin, New Jersey (Smithsonian Insitution #NMNH #C6088), manganite from Germany (Smithsonian Insitution #NMNH 157872), manganese carbonate (Sigma-Aldrich), pyrolusite from Rossbach, Germany (Smithsonian Insitution #NMNH B6724), synthetic triclinic Na-birnessite, and synthetic pH 2 hexagonal birnessite. All weights were set between 0 and 1, and all combinations were fit with at most 2 standards. Weights of standards were not forced to sum to 1.

X-ray photoelectron spectroscopy (XPS) For XPS analysis, powder samples were covered with a strip of conductive copper tape and pressed with clean borosilicate glass blocks onto copper stubs. Measurements were conducted with a Kratos Axis Ultra DLD spectrometer with an Al Kα X-ray source (1486.7 eV) operating at 10 mA and 15 kV. Magnetic immersion lenses were used to improve collection

173

efficiency. The instrument work function was calibrated to give a binding energy (BE) of 83.96 eV ± 0.05 eV for the 4f7/2 line of metallic gold. The spectrometer dispersion was adjusted to yield a BE of 932.62 eV for the Cu2p3/2 line of metallic copper. Measurements of the Mn2p, Mn3s, Mn3p, O1s, C1s, and various alkali and alkaline Earth lines were conducted with a step size of 0.1 eV, an analysis area of 300 x 700 microns, and pass energies (PE) of 20 or 40 eV. The resultant full-width-at-half-maximums (FWHM) for the Ag3d5/2 line were 0.54 and 0.77 eV, respectively. The low sensitivity of the Mn3s line resulted in measurements only with PE = 40 eV. Survey scans were conducted at PE = 160 eV and step size = 0.5 eV. XPS spectra were fit by non-linear least squares after Shirley background subtractions with the CasaXPS curve resolution software package. Gaussian/Lorentzian contributions to line shapes were numerically convoluted with a Voigt function.

Results and Discussion

X-ray diffraction of natural birnessites Our analyses of the XRD patterns classified the natural birnessites as either hexagonal ranciéite-like materials or as less interpretable “3-line” birnessite varieties. The natural hexagonal samples included specimens from Paxton Cave, VA, Granada, Spain, and the French Pyrenees, all of which yielded patterns whose peaks closely matched those of a synthetic hexagonal H-birnessite (Fig. 4). The SEM images of the ranciéite samples from France and Spain showed layered flakes that extended ~100 µm across but were only tens of nm in thickness (Fig. 5). The Paxton Cave sample, on the other hand, appeared as fluffy orbicules of birnessite mixed with diatoms (Fig. 5c). EPMA analyses revealed that Ca was the dominant interlayer cation in all three of these birnessite samples, although a host of other metals, including Mg, Fe,

174

Zn, Al, Ba, K, Na and possibly Si, also were observed above trace concentrations (Fig. 6). Since the natural birnessite samples often were intergrown with kaolinite, Al and Si may have been contributed from aluminosilicate clays that fell below the EPMA imaging resolution. In contrast, we grouped “3-line” birnessite varieties on the basis of 3 broad, relatively asymmetric peaks in their XRD patterns with d-spacings of ~7.2 Å, ~2.4 Å, and ~1.4 Å (Fig. 7). Several of these samples presented initially as buserite with a 10 Å d-spacing – namely, the sample from Spring Branch, TN and three of the central Pennsylvania birnessites (DS1-M3f, DS2-M3f, and PBS-M2f-2) – and these were dried at 110°C to collapse the layer spacing to 7 Å to allow a more direct comparison with other natural 7 Å birnessites. The original XRD patterns prior to drying can be found in SI Fig. 1. Interestingly, one of the central Pennsylvanian samples (PBS-M2f-1) also exhibited a 10 Å d-spacing, but it failed to collapse completely upon heating. However, linear combination fitting of EXAFS did not identify any todorokite in this sample that could produce the 10 Å spacing observed by XRD. Ca2+, Mg2+, Ni2+, and Co2+ cations have been observed to stabilize the buserite structure, and all of these (with the exception of Ni 2+) were detected by EPMA in these samples (Fig. 6). It still is unclear why some buserites collapse upon heat treatment at 110 oC, while others do not. When the “3-line” birnessite samples were observed with the SEM, they consisted mostly of nanometer-thick flakes that measured