American Mineralogist, Volume 94, pages 816–826, 2009 0003-004X/09/0506–816$05.00/DOI: 10.2138/am.2009.3068 816 Cs-exchange in birnessite: Reaction mechanisms inferred from time-resolved X-ray diffraction and transmission electron microscopy Christina L. Lopano, 1, * peter J. heaney, 1 and Jeffrey e. post 2 1 Department of Geosciences, Pennsylvania State University, University Park, Pennsylvania 16802, U.S.A. 2 Department of Mineral Sciences, Smithsonian Institution, Washington, D.C. 20013-7012, U.S.A. abstraCt We have explored the exchange of Cs for interlayer Na in birnessite using several techniques, including transmission electron microscopy (TEM) and time-resolved synchrotron X-ray diffraction (XRD). Our goal was to test which of two possible exchange mechanisms is operative during the reaction: (1) diffusion of cations in and out of the interlayer or (2) dissolution of Na-birnessite and reprecipitation of Cs-birnessite. The appearance of distinct XRD peaks for Na- and Cs-rich phases in partially exchanged samples offered support for a simple diffusion model, but it was inconsistent with the compositional and crystallographic homogeneity of (Na,Cs)-birnessite platelets from core to rim as ascertained by TEM. Time-resolved XRD revealed systematic changes in the structure of the emergent Cs-rich birnessite phase during exchange, in conflict with a dissolution and reprecipitation model. Instead, we propose that exchange occurred by sequential delamination of Mn oxide octahe- dral sheets. Exfoliation of a given interlayer region allowed for wholesale replacement of Na by Cs and was rapidly followed by reassembly. This model accounts for the rapidity of metal exchange in birnessite, the co-existence of distinct Na- and Cs-birnessite phases during the process of exchange, and the uniformly mixed Na- and Cs-compositions ascertained from point analyses by selected area electron diffraction and energy dispersive spectroscopy of partially exchanged grains. Keywords: Cation exchange, cesium, birnessite, synchrotron, X-ray diffraction, transmission electron microscopy introduCtion Groundwater contamination by radionuclides (including U, Pu, Np, and Cs) is a serious problem at several national labora- tories that were involved in the production of components for nuclear weapons in the United States, such as the Hanford Site in Washington state (McKinley et al. 2001). Leaks from high- level waste (HLW) storage tanks in the 200 Area of the Hanford Site have released appreciable quantities of 137 Cs into the vadose zone, and migration of this contaminant has extended to depths that are significantly greater than expected (Serne et al. 2001a, 2001b). Radioactive 137 Cs is a fission product of irradiated U and Pu with a half-life of 35.7 years (Zachara et al. 2002). Because it is highly soluble, 137 Cs can be extremely mobile in soil envi- ronments (Bostick et al. 2002). However, its rate of migration through soils is difficult to model because the transport of 137 Cs depends on numerous factors, particularly fluid composition and soil type (Almgren and Isaksson 2006). For example, several researchers have demonstrated that 137 Cs will readily sorb to various aluminosilicate clay minerals (Comans et al. 1991; Sut- ton and Sposito 2001; Bostick et al. 2002; Zachara et al. 2002), dramatically inhibiting transport through the subsurface. In the present study, we explored the interaction of dissolved Cs ions with layered Mn oxides, which are ubiquitous in a wide range of soils, from arid desert varnishes to temperate soil pre- cipitates (Waychunas 1991; Post 1992, 1999; Yang et al. 2003), including the Ringold Formation, which underlies the Hanford formation and comprises a mixture of poorly consolidated clays, silts, sands, and gravels (Fredrickson et al. 2004). Even when Mn oxides occur at the 1 wt% level or lower in soils, these phases can act as the controlling players in contaminant migration. The high reactivity of these minerals can be attributed to several fac- tors. Many Mn oxide phases occur as particles that are only a micrometer in diameter or smaller, particularly when they grow authigenically within soils. Consequently, the ratio of reactive surface area to volume is extremely high (Murray 1974, 1975). In addition, Mn oxides can occur in various structural topologies (Fritsch et al. 1997), and many of the phases that are commonly found in soils have structural architectures (e.g., layer-type or tunnel-type) that are especially amenable to solid-state diffusion (Balachandran et al. 2002). Birnessite-like phases are the most common natural phyllo- manganates. The birnessite structure (Fig. 1) consists of sheets of edge-sharing Mn 4+ O 6 octahedra where Mn 3+ or vacancies substi- tute for Mn 4+ in the octahedral layers, resulting in a net negative layer charge, which is balanced by various hydrated cations in the interlayer region (commonly Na and Ca). Many studies have quantitatively demonstrated that Mn oxides (particularly birnes- site) are sinks for a host of transition metals (Loganathan and Burau 1973; Singh and Subramanian 1984; Burns et al. 1985; Nicholson and Eley 1997), and even transuranic radionuclides have been shown to exhibit a strong affinity for Mn oxides (Triay et al. 1991; Duff et al. 2001, 2002; Powell et al. 2006). * Present address: RJ Lee Group, Inc., 350 Hochberg Road, Monroeville, Pennsylvania 15146, U.S.A. E-mail: clopano@ rjlg.com