Structural study of biotic and abiotic poorly-crystalline manganese oxides using atomic pair distribution function analysis
Introduction
Manganese (Mn) oxides are versatile materials, in terms of their use in industrial applications, and environmentally important role in biogeochemical elemental cycling. They can be used as cathodic materials in lithium batteries, molecular sieves, and catalysts, etc. (Feng et al., 1995, Post, 1999). Mn oxide minerals affect carbon cycling (Shindo and Huang, 1982, Stone and Morgan, 1984), nutrient speciation (Shindo and Huang, 1984, Luther, 2002), and mobility, bioavailability and toxicity of metal(loid)s, e.g., arsenic (Zhu et al., 2009, Lafferty et al., 2010), via sorption and redox reactions in the environment. However, they are generally less abundant than other minerals, such as iron and aluminum (oxyhydr)oxides. Both layered and tunnel structures of Mn oxides have been identified in the environment with the former type more abundant, such as birnessite and birnessite-like minerals (Post, 1999). It is critical to know the detailed crystal structures of Mn oxides in order to better understand their behavior in geochemical processes. However, a major portion of naturally-occurring Mn oxide minerals, such as vernadite, are fine-grained and poorly-crystalline. This makes accurate crystal structure determination almost impossible using X-ray diffraction (XRD) Rietveld refinement since it is only suitable for well-crystallized materials (Post, 1999). Nevertheless, poorly crystalline Mn oxides are more actively involved in many important geochemical reactions than their well-crystallized counterparts due to their large surface area and enrichment of structural cation defects. In short, it is their poorly-crystalline nature that makes them both highly reactive and difficult to study. Therefore, it is important that methods be developed to unravel the complexities of these and similar materials.
Vernadite is the dominant Mn oxide phase precipitated in water bodies (e.g., sediments) and as coatings on other mineral surfaces (Chukhrov et al., 1979, Hochella et al., 2005, Bargar et al., 2009). Vernadite samples typically contain intimately mixed iron (oxyhydr)oxides as minor components, which further complicates their structural determination. Synthetic analogs, such as acid birnessite (AcidBir), δ-MnO2 and polymeric MnO2 (PolyMnO2), have been used in investigating vernadite geochemical behavior in laboratory studies (McKenzie, 1980, Luther, 2002, Feng et al., 2007). These materials are poorly-crystalline and nanostructured with broad diffraction peaks at d-spacings of around 7, 3.5, 2.4 and 1.4 Å (Villalobos et al., 2003). They are generally thought to be layered materials consisting of turbostratically stacked MnO6 octahedral layers (OL). MnO6 octahedra in layers are connected by sharing edges only. Mn atoms in the layers are mainly Mn4+ but Mn3+ can be a minor component. The layers also contain varying amounts of cation defects (i.e., vacant sites). The presence of Mn3+ and vacant sites results in negative charge that is compensated by sorption of cations, such as H+ and metal cations, including Mn2+ and Mn3+, in the interlayer (IL) regions.
Vernadite formation is largely catalyzed by Mn-oxidizing microorganisms in oceanic and terrestrial environments (so called biogenic Mn oxides (BioMnOx)) (Wehrli et al., 1995). BioMnOx, produced by model marine and terrestrial bacteria, i.e., Bacillus sp. SG-1, Pseudomonas putida MnB1/GB-1 and Leptothrix discophora SP-6, have been characterized as layered Mn oxides by XRD and extended X-ray absorption fine structure (EXAFS) spectroscopy (Villalobos et al., 2003, Webb et al., 2005a, Saratovsky et al., 2009, Zhu et al., 2010a). These studies show that bacterial BioMnOx samples have hexagonal layer symmetry, unless formed in the presence of cations, such as calcium, under relatively alkaline pH conditions (Webb et al., 2005b, Zhu et al., 2010a). However, the structure for BioMnOx produced by fungi could be a tunnel type. A recent study, based on EXAFS analysis, proposed that Acremonium sp. KR21-2 produces a todorokite-like BioMnOx (Saratovsky et al., 2009) in solid agar media. However, a layered structure of BioMnOx, produced by the same type of bacteria, but in liquid media, was also reported (Grangeon et al., 2010). Todorokite and other tunnel Mn oxides are believed to be formed from layered Mn oxide precursors in the presence of template cations (Feng et al., 2010). These results suggest a direct formation pathway for todorokite from Mn2+ oxidation without layered Mn oxides as precursors.
EXAFS spectroscopy detects local atomic structure (within 6 Å) and cannot describe long-range order (Webb et al., 2005a). However, XRD Rietveld refinement reveals the detailed long-range ordered structure present in well-crystallized materials. These techniques, alone, are not capable of unambiguously characterizing the crystal structure of poorly-crystalline and nanoparticulate materials (Billinge and Levin, 2007). Consequently, XRD and EXAFS data of such complex materials can be variously interpreted or misinterpreted. Vernadite and δ-MnO2 were proposed to be tunnel structures, i.e., todorokite and cryptomelane-like minerals, decades ago (Giovanoli, 1980, Manceau et al., 1992a, Manceau et al., 1992b). Even recently, Kim et al. (2003) proposed that BioMnOx produced by L. discophora SP-6 is a todorokite-like structure based on EXAFS and Raman spectroscopy. However, this viewpoint has been disputed by more recent studies (Villalobos et al., 2003, Webb et al., 2005a, Saratovsky et al., 2006). Assuming layered structures, trial-and-error simulations of the poorly-defined XRD patterns of δ-MnO2, acid birnessite and BioMnOx provided geometrical dimensions using low angle data. Additionally, unit-cell parameters, atomic coordinates and occupancies of layer atoms and interlayer cations and H2O molecules were determined using high angle data (Villalobos et al., 2006, Drits et al., 2007, Grangeon et al., 2010). These methods were also used for Ni-sorbed δ-MnO2 samples to determine Ni positions (Grangeon et al., 2009). The results indicate that all of the above Mn oxides have hexagonal layer symmetry although discrepancies in positions of interlayer species remain. However, the crystallographic space group cannot be determined using the above methods.
Atomic pair distribution function (PDF) analysis is a total scattering technique, considering both Bragg and diffuse scattering. It probes not only the local structure but also intermediate and long-range order (Egami and Billinge, 2003). This makes PDF analysis an appropriate tool for investigating the structures of poorly-crystalline and nanoparticulate materials that yield diffraction patterns with large amounts of diffuse scattering between poorly-defined Bragg diffraction peaks (Egami and Billinge, 2003). A few structural determinations have been performed on poorly-crystalline Mn oxides using PDF analysis. Gateshki et al. (2004) reported that Mn oxides, synthesized by reduction of KMnO4 with LiI in aqueous and anhydrous solutions, are layered structures belonging to the hexagonal RM space group. A study by Petkov et al. (2009) suggested that BioMnOx produced by L. discophora SP-6 bacteria and the fungus Acremonium sp. KR21-2 have birnessite (triclinic space group, P1) and todorokite-like (monoclinic space group, P12/M1) structures, respectively. This study supports the todorokite-like structure proposed by Saratovsky et al. (2009) using XAFS and XRD analyses. However, the triclinic space group of SP-6 BioMnOx (deviating significantly from hexagonal symmetry) is contradictory to the hexagonal layer symmetry reported for BioMnOx, that was produced under similar solution conditions (Villalobos et al., 2003, Zhu et al., 2010a).
In the present study, we applied X-ray PDF analysis to investigate the structures of the vernadite analogs, including BioMnOx produced by P. putida GB-1, δ-MnO2, PolyMnO2 and AcidBir. The BioMnOx used in this study was produced by a different strain of Mn-oxidizing bacteria than was reported in the previous PDF study (Petkov et al., 2009) to determine whether the structure of BioMnOx depends on bacterial producers. The above materials are widely used model minerals in environmental science and biogeochemistry research. However, structural information is ambiguous, and in the case of PolyMnO2, unknown. The PDF analysis in this study provides more rigid constraints on the Mn oxide crystal structures than previously available.
Section snippets
Preparation of Mn oxides
Pyrolusite (1 × 1 Mn4+ tunnel structure) was purchased from Sigma–Aldrich. Todorokite was provided by the Smithsonian National Museum of Natural History. The sample was originally collected from Monte Negro Mine, Oriente Province, Cuba. Cryptomelane (2 × 2 Mn4+ tunnel structure), manganite (γ-MnOOH, 1 × 1 Mn3+ tunnel structure), triclinic birnessite (TrBir, Mn3/4+, layered structure), hexagonal birnessite (HexBir, Mn2/3/4+, layered structure), Acid birnessite (AcidBir), BioMnOx, δ-MnO2 and PolyMnO2
Chemical composition
The chemical composition of the Mn oxide samples and standards are given in Table 1. The chemical composition information was used in the PDF data reductions and also as a reference for composition of structural models in the PDF simulations. The dominant cations associated with Mn oxides are alkaline and alkaline earth metals (Na, K, Mg and Ca) which are introduced from the chemicals used in their syntheses (Table 1). Since the todorokite sample is a naturally-occurring material, it may
Discussion
The PDF analyses provide direct structural information for the poorly-crystalline Mn oxides studied here. PDFs list atomic pair correlations up to several nanometers which are much more extensive, and subsequently more convincing, than radial structure functions derived from EXAFS spectroscopy, which are valid only up to 0.6 nm. Comparisons of the PDFs of these samples with Mn oxide standards unambiguously indicate that these materials are layered structures. This conclusion is further
Conclusions
More constrained crystal structures of four poorly-crystalline Mn oxides, including a biogenic Mn oxide sample produced by the bacterium P. putida GB-1, were determined using atomic pair distribution function (PDF) analysis. The PDF analyses indicate that these Mn oxides are layered materials with the C12/m1 crystallographic space group, verifying the prevailing viewpoint that these oxides have a layered structure. The C12/m1 space group provides more structural variations, required for
Acknowledgements
The authors would like to thank beamline scientists, Dr. Karena W. Chapman, Dr. Peter J. Chupas and Kevin A. Beyer at beamline 11-ID-B at the Advanced Photon Source (APS) and Dr. Jonathan Hanson, Dr. Laura Barrio-Pliego and Dr. Gong Zhang at beamline X7B at the National Synchrotron Light Source (NSLS) for their technical assistance with data collection and analyses. Use of the NSLS was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract
References (71)
- et al.
Structural characterization of terrestrial microbial Mn oxides from Pinal Creek, AZ
Geochim. Cosmochim. Acta
(2009) - et al.
Adsorption and redox reactions of heavy metals on synthesized Mn oxide minerals
Environ. Pollut.
(2007) - et al.
Formation of nano-crystalline todorokite from biogenic Mn oxides
Geochim. Cosmochim. Acta
(2010) - et al.
Structure of the synthetic K-rich phyllomanganate birnessite obtained by high-temperature decomposition of KMnO4 substructures of K-rich birnessite from 1000 °C experiment
Microporous Mesoporous Mater.
(2007) - et al.
Direct observation of heavy metal-mineral association from the Clark Fork River Superfund Complex: implications for metal transport and bioavailability
Geochim. Cosmoichim. Acta
(2005) The interaction of metal ions at the manganese dioxide-solution interface
Geochim. Cosmochim. Acta
(1975)- et al.
The Structure of manganese oxide formed by the fungus Acremonium sp. strain KR21–2
Geochim. Cosmochim. Acta
(2009) - et al.
Characterization of the manganese oxide produced by Pseudomonas putida strain MnB1
Geochim. Cosmochim. Acta
(2003) - et al.
Data Reduction and Error Analysis for the Physical Sciences
(1992) - et al.
The problem with determining atomic structure at the nanoscale
Science
(2007)