American Mineralogist, Volume 96, pages 1809–1820, 2011 Cesium and strontium incorporation into zeolite-type phases during homogeneous nucleation from caustic solutions NELSON RIVERA,1,* SUNKYUNG CHOI,2 CALEB STREPKA,3 KARL MUELLER,3 NICOLAS PERDRIAL,4 JON CHOROVER,4 AND PEGGY A. O’DAY1 15200 North Lake Road, School of Natural Sciences, University of California, Merced, California 95343, U.S.A. 2U.S. Environmental Protection Agency, Natational Risk Management Research Lab, Ada, Oklahoma 74820, U.S.A. 3Department of Chemistry, Pennsylvania State University, University Park, Pennsylvania 16802, U.S.A. 4Department of Soil, Water and Environmental Science, University of Arizona, Tucson, Arizona 85721, U.S.A. ABSTRACT Formation of faujasite- and sodalite/cancrinite-type phases associated with caustic waste reactions in the environment may structurally incorporate contaminant species such as radioactive Sr2+ and Cs+, and thus provide a mechanism of attenuation. To investigate mineral evolution and structural incorporation of cations in simplified experiments, aluminosilicate solids were precipitated homo- geneously at room temperature from batch solutions containing a 1:1 molal ratio of Si to Al and 10-3 molal Sr and/or Cs, and aged for 30 or 548 days. Syntheses were done with solutions in equilibrium with atmospheric CO2 and with gas-purged solutions. Experimental products were characterized by bulk chemical analyses, chemical extractions, XRD, SEM/TEM, TGA, solid-state 27Al NMR, and Sr EXAFS. Chemical analysis showed that solids had a 1:1 Al:Si molar ratio, and that Sr was sequestered at higher amounts than Cs. After 30 days of aging in purged solutions, XRD showed that zeolite X (faujasite-type) was the only crystalline product. After aging 30 and 548 days in solutions equilibrated with atmospheric CO2, a mixture of sodalite, cancrinite, and minor zeolite X were produced. Surface areas of solids at 30 days were much lower than published values for zeolite phases synthesized at high temperature, although particle aging produced more crystalline and less aggregated phases with higher bulk surface areas. Characterization of products by 27Al NMR indicated only tetrahedrally coordinated Al. Measured isotropic shifts of primary resonances did not change substantially with precipitate aging although the primary mineral phase changed from zeolite X to sodalite/cancrinite, indicating local ordering of Al-Si tetrahedra. Analysis of reaction products by Sr EXAFS suggested Sr bonding in hexagonal prisms and six-membered rings of the supercages of zeolite X that may be more site specific than those of monovalent cations. For samples aged for 548 days, interatomic dis- tances from Sr-EXAFS are consistent with partial Sr dehydration and bonding to framework oxygen atoms in sodalite cages or in large channels in cancrinite. Incorporation of Sr into both faujasite and sodalite/cancrinite phases is favored over Cs during room-temperature synthesis, possibly because of increased cation site competition between Cs+ and Na+. Results of this study help to constrain cation incorporation into sodalite/cancrinite mineral assemblages that form at caustic waste-impacted field sites and may aid in the predictive modeling of contaminant release. Keywords: Zeolite X, sodalite, cancrinite, strontium, cesium INTRODUCTION high surface area minerals such as clays or zeolites, or by miner- Zeolite and feldspathoid phases have been proposed to form alogical transformations associated with dissolution-precipitation in the environment at locations where caustic (pH > 13), high processes, have been identified as critical mechanisms for the ionic strength, Al-rich waste solutions containing various metal understanding and prediction of long-term contaminant behavior and radionuclide contaminants have reacted with subsurface (Zachara et al. 2007). sediments (Bickmore et al. 2001; Mashal et al. 2005a; Qafoku Prior laboratory studies examining reaction of simulated et al. 2003a; Zachara et al. 2002). Such solutions have been tank wastes with model systems (quartz sand, biotite, reference unintentionally released to the environment at locations such as clays) (Bickmore et al. 2001; Choi et al. 2005a, 2005b, 2006; the DOE Hanford site (Washington, U.S.A.), a former nuclear Chorover et al. 2003; Crosson et al. 2006; Um et al. 2005) and weapons processing facility (Gephart 2003; Gephart and Lund- natural sediments (Ainsworth et al. 2005; Chorover et al. 2008; gren 1998). Of particular interest are 90Sr and 137Cs, which account Mashal et al. 2005a, 2005b; Perdrial et al. in review; Qafoku et for the majority of radioactivity released at the Hanford site from al. 2003a, 2003b, 2004; Wan et al. 2004a, 2004b) have shown that leaking tanks (Behrens et al. 1998; Serne et al. 2002). Attenuation various aluminosilicate reaction products may form and poten- of these contaminants in the subsurface, either by adsorption by tially incorporate radioactive contaminants. Differences among product phases and cation uptake that have emerged from prior * E-mail: [email protected] experimental investigations depend on (1) whether dissolved Si 0003-004X/11/1112–1809$05.00/DOI: 10.2138/am.2011.3789 1809 1810 RIVERA ET AL.: CESIUM AND STRONTIUM INCORPORATION INTO ZEOLITE-TYPE PHASES in the system is limited by release from weathering of primary prepared using ultra-pure (MilliQ) water and reagent grade NaNO3, NaOH, CsCl, Si-bearing phases (quartz, feldspar, or phyllosilicate), or is added SrCl2·6H2O (J.T. Baker) and NaAlO2·xH2O powder (EM Science) as obtained from the manufacturer. Synthesis solutions similar to the synthetic tank waste leachate to the experimental system at high concentration (similar to the (STWL) described previously (Choi et al. 2006; Chorover et al. 2008; Crosson et concentration of dissolved Al); and (2) whether dissolved CO2 al. 2006; Thompson et al. 2010) were prepared in batch on a mass basis (mol/kg is present at ambient levels or at higher concentrations from the solvent) with a total solvent weight (water) of 25 g. Solutions with atmospheric CO2 dissolution of carbonate phases, or excluded from the aqueous (referred to as +CO2) were not degassed at any step of the synthesis and consisted of 3+ + - - -3 + and gas phases of the experimental system. These two factors, 0.05 m Al , 2.0 m Na , 1.0 m NO3, and 1.0 m OH (pH ~13.7), and either 10 m Cs or 10-3 m Sr2+, both Cs+ and Sr2+ (10-3 m each), or neither (Na only). For both sets of together with reaction time, appear to influence the kinetic path- synthesis, chloride salts were used for Sr and Cs. This produced small differences ways of product formation in addition to expected differences in chloride concentration (0 mm for Na only to 3 mm for Sr+Cs). Each solution was from thermodynamic considerations based on solution compo- then spiked with 0.05 m colloidal SiO2 (from 40% w/w Ludox solution W.R. Grace sition. The studies cited above have identified the feldspathoid and Co.), which resulted in immediate formation of a visible precipitate. Suspensions were shaken initially to ensure homogenization, but no other agitation of the solution phases cancrinite and sodalite, several zeolite phases, and other was done during aging at ambient room temperature. Solutions and their precipitates unidentified aluminosilicate phases as products of these base- were aged in individual batch experiments for 4 h, 3 days, 30 days, and 548 days. neutralization reactions. Typical solid yield was 0.2–0.3 g (8–12 g/kg STWL), which limited the extent of Feldspathoids and zeolites are large, diverse classes of miner- characterizations that could be performed. als characterized by a crystalline framework of tetrahedral Al and Precipitates from batch, low CO2 solutions (referred to as –CO2) consisted of purging the initial water (400 g) for 24 h with N2 gas to remove carbon dioxide. The Si with a three-dimensional pore system, multiple cation sites solution was further purged for 10 min after the addition of each reagent salt (same as within cages and pores, and the ability to host different cations above) to the water. Solutions were prepared with either 10-3 m Cs+ or 10-3 m Sr2+, both with various ion sizes and charges (Barrer 1984; Bonaccorsi Cs+ and Sr2+ at 10-3 m each (duplicate samples denoted A and B), or neither (Na only). and Merlino 2005; Depmeier 2005; Frising and Leflaive 2008). After the addition of the colloidal silica (stock solution not purged), the container was filled with a N gas in the head space and aged for 30 days. No other measures were Because of the importance of zeolite-type phases as industrial 2 taken to exclude CO2 from the system and dissolved CO2 was not measured. Typical catalysts and molecular sieves, their synthesis and stability under solid yield was 2.5–3.0 g (6.25–7.5 g/kg STWL). For all samples after the specified controlled laboratory conditions and often at hydrothermal tem- aging time, suspensions were centrifuged and supernatant solutions decanted. Solid peratures (60–1000 °C) have been widely studied (Barrer 1984; precipitates were washed three times with 95% ethanol and then air dried. Davis and Lobo 1992; Navrotsky et al. 1995). Most investiga- Digestions and extractions tions under controlled conditions, however, have been designed to optimize zeolite performance for industrial applications, and Complete dissolution of all product solids was done by concentrated hydrofluoric acid (HF) digestion at room temperature. Triplicate digestions were performed on therefore do not specifically address