Electronic Structures and Bonding in Fe-Mn Oxide and Oxide Hydroxide MineralsXAS and RIXS Spectroscopy using Station 8.01 at ALS Berkeley.
David M. Sherman, C. L. Peacock* and E.G. Todd*
Department of Earth Sciences, University of Bristol, Bristol BS8 IRJ
* NERC Ph.D Students
Introduction. Naturally occurring iron(ll,lll) and manganese(lll,IV) oxide minerals (Table1) are among the most important phases in environmental chemistry since they sorb a variety of heavy metals and pollutants. Fe-Mn oxides also act as electron sinks so that oxidation-reduction reactions that are kinetically unfavored in aqueous solutions are facilitated at the oxide-water interface. However, the mechanisms by which iron and manganese oxide mineral surfaces promote electron transfer processes are not understood
Table 1: Some of the Fe-Mn oxide minerals important in aqueous environmental chemistry
Iron and manganese oxides are also involved in various photochemical processes occurring in sunlight. Such reactions are important for nutrient recycling in the photic zone of the oceans. A complex aspect of these minerals is that they are often mixed-valence phases (e.g., Mn3+ replacing Mn4+) this allows electron hopping via intervalence charge transfer (small polarons) and must play a significant role in the electrochemistry of these oxides. The structures of Fe-Mn oxides also allow incorporation of other cations in the tunnels and sheets. Isomorphous replacement of Mn4+ and Fe3+ by other cations such as Mn3+, Co3+, and V is the primary reason for the geochemical and economic importance of Fe-Mn oxides.
If we knew the electronic structures of iron and manganese oxide minerals, we could begin to understand how their surfaces mediate electron-transfer reactions. We could also begin to understand the nature of the photochemical reactions involving FeOOH and MnO2. A detailed structure of the d-band orbitals, in particular, would shed great insight on the stability of heavy metals such as Cu, Ni, Co and Pb that are sorbed to Fe-Mn oxides.
Iron(III) and manganese(II,III,IV) oxide minerals have partially occupied Fe(3d) and Mn(3d) orbitals which yield a large exchange splitting of the valence band (mostly O(2p)) and conduction band (Fe or Mn(3d)). The relative energies of metal d-bands will determine if electron transfer between adsorbed cations and the bulk oxide is possible. The absolute energies of the partially occupied- orbitals can be related to the electrochemical scale since we know the energy of the reaction
H+ (aq) + e- = H2(g) to be -4.5 eV.
At present, we do not have a satisfactory understanding of the actual band gaps in these semiconducting oxide minerals. Optical spectra are plagued by a complex variety of exchange- enhanced transitions that interfere with the ligand-to metal charge transfer bands (Sherman, 1984; Sherman and Waite, 1985). Metal-metal charge transfer processes in Mn(III,IV) oxides make optical band gaps impossible to measure. The electronic structures of Fe-Mn oxides has been calculated from finite clusters (Sherman, 1984; 1985) using density functional theory. One of the most fundamental questions is the physical nature of calculated orbital eigenvalues determined using density functional theory. Band gaps calculated from ground state orbital energies are usually different from those derived-from optical spectra
Station8.0.1at the Advanced Light Source, Lawrence Berkeley Laboratory, Berkeley CA (USA):
Beamline 8.0.1 provides photons over an energy range 64-1400 eV with a flux of 1011- 1015 photons/s. The high source brightness at 8.0.1 allows measurements of Soft X-ray emission spectra; for this, a Soft X-ray Fluorescence spectrometer (SXF) endstation has been developed to measure fluorescence over the range 4—1000 eV. at ALS. Emission spectra allow us to determine the nature of occupied electronic states in minerals and is a necessary complement to the probe of unoccupied states provided by the Total Electron Yield (surface sensitive) and Fluorescence Yield (bulk sensitive) measurements. Moreover, by selectively tuning the excitation wavelength, we can do Resonant Inelastic Scattering which can provide more detailed electronic structure information from the X-ray spectra.
O K-edge XAS. Our first (3 day) visit to ALS station 8.0.1 allowed us to obtain some preliminary O K-edge X-ray absorption and emission spectra of MnO2 and FeOOH phases (Sherman et al., 2002). The results were very encouraging. We find that the ground state electronic structure of MnO2, calculated using a localised basis set and the GGA exchange- correlation functional, is in remarkable agreement with the measured XAS and XES spectrum (Figure 1). This is consistent with Slater's original idea of DFT eigenvalues corresponding to "orbital electronegativities". The O(2p)-Mn(3d) band gap and splitting of the Mn(3d) orbitals are in good agreement with calculated results. However the data obtained does not tell us the position of the occupied Mn(3d) t2g orbital.
Figure1: Electronic structure of MnO2:(a) Calculated O(2p) projected density of states using GGA approximation for exchange-correlation, (b) O K-edge absorption spectrum to unoccupied states of Mn(3d) character, (c) Emission spectrum shows complex structure due to O(2p) band; structure in the emission spectra between 527-537 eV we assign to O(2p)-Mn(3d) bonding orbitals and the filled Mn(3d)-O(2p) antibonding orbitals (majority-spin t2g Mn(3d) orbitals).
Resonant Inelastic X-ray spectroscopy
(RIXS). The electronic structures of transition metals such as Co and V in Mn/Fe oxide phase can be investigated from the RIXS spectra associated with metal L-edge X-ray emission. A completely elastic emission occurs when the core-hole pair recombine with the same energy as the excitation. An inelastic peak, however, occurs when the excited electron loses energy by exciting to a higher energy state before recombining. This allows us to probe the d-d transitions of transition metal oxides and resolve the d-d excitation from the broad ligand-to-metal charge transfer (LMCT) transitions. Our second visit to ALS (14 days) was to measure the RIXS spectra of metals (Co, V and Mn) incorporated into FeOOH and layer phyllomanganite minerals.
Figure 2: X-ray absorption and RIXS spectra of CoOOH. The cobalt is low spin and has a nominal orbital electronic configuration of t2g6eg0; this agrees with the lack of crystal field splitting in the L-edge absorption spectrum.
The RIXS of Co3+ in CoOOH (Fig. 2) show that the loss features are not of the Co3+ (d-d) transition type but more resemble LMCT transitions. This may reflect the very covalent nature of the Co3+-O bonding.
RIXS of Co2+ in goethite (FeOOH), bimessite and buserite (Fig. 3) are very similar to those of CoO (Butorin, 2000) except for the absence of the LMCT peak in the emission spectra of goethite and the manganese oxides, ll has been shown that the spectrum of CoO can be well modelled in terms of the atomic multiplets of Co2+ This data also shows that Co2+ can be incorporated into goethite, bimessite and buserite without oxidation to Co3+ This is at odds with previous work (Manceau et al., 1997) that appeared to show oxidation of Co2+ in bimessite and buserite to Co3+ It has been argued that the oxidation of Co on manganese oxides controls the marine chemistry of Co and its uptake into ferromanganse nodules. The crystal chemistry of Co in manganese oxides clearly needs further work. In particular, we need to supplement our RIXS measurements with EXAFS spectroscopy at Daresbury.
We were initially setting out to determine the 3d energies of Co, V and Mn substituted impurities in FeOOH in order to understand the stability of different oxidation states of metals in FeOOH. However, this does not yet appear possible with the current instrumental configuration. Our initial study of the RIXS data, however, does appear to show that the O->Mn3+ ligand to metal charge transfer energy for Mn3+ in FeOOH is different from that in MnOOH and Mn2O3.
Figure3: X-ray absorption and RIXS spectra of Co2+ exchanged into the buserite interlayer. The cobalt is still Co2+ as evidenced by the crystal field splitting of the Co L-edge peak in the absorption spectrum. The emission spectra are very similar to those of CoO (Butorin, 2000)
References:
Butorin, S.M. (2000) Resonant inelastic X-ray scattering as a probe of optical scale excitations in strongly electron-correlated systems: quasi-localised view. J. Electron. Spectros. Rel. Phenom., 110-111, 213-233.
Manceau A., Drits V.A., Silvester E., Bartoli C., and Lanson B. (1997) Structural mechanism of Co2+ oxidation by the phyllomanganate buserite. American Mineralogist 82,1150-1175.
Sherman, D.M. (1984) Electronic structures of manganese oxide minerals. American Mineralogist 69, 788-799.
Sherman, D.M. (1985) Electronic structures of Fe3+ coordination sites in iron oxides; applications to spectra, bonding and magnetism. Physics and Chemistry of Minerals 12,161-175.
Sherman, D.M., and T.D. Waite (1985) Electronic spectra of Fe3+ oxides and oxide hydroxides in the near-IR to near- UV. American Mineralogist 70, 1262-12696.
Sherman, D.M. (1990) Molecular orbital (SCF-Xa-SW) theory of Fe2+-Mn3+, Mn2+-Fe3+ and Fe3+-Mn3+ charge transfer and magnetic exchange interactions in oxides and silicates. American Mineralogist 75, 256-261.
Sherman, D.M., Todd E.G., and Purton J.A. (2002) Electronic structures of Fe-Mn oxide minerals: results from X-ray spectroscopy and density functional theory. Abstracts of Papers submitted to the American Chemical Society. Orlando FL April 2002.
Todd, E.G., Sherman D.M., and Purton J.P. (2002) Surface oxidation of pyrite under ambient atmospheric and aqueous (pH = 2-12) conditions: electronic structure and mineralogy from X-ray absorption spectroscopy. Geochimica et Cosmochimica Acta (in press).