Article pubs.acs.org/est
In Situ X-ray Diffraction Study of Na+ Saturated Montmorillonite Exposed to Variably Wet Super Critical CO2 Eugene S. Ilton,* H. Todd Schaef, Odeta Qafoku, Kevin M. Rosso, and Andrew R. Felmy Pacific Northwest National Laboratory, P.O. Box 999, MS K8-96, Richland, Washington 99352, United States ABSTRACT: Reactions involving variably hydrated super critical CO2 (scCO2) and a Na saturated dioctahedral smectite (Na-STX-1) were examined by in situ high-pressure X-ray diffraction at 50 °C and 90 bar, conditions that are relevant to long-term geologic storage of CO2. Both hydration and dehydration reactions were rapid with appreciable reaction occurring in minutes and near steady state occurring within an hour. Hydration occurred stepwise as a function of increasing H2O in the system; 1W, 2W-3W, and >3W clay hydration states were stable from ∼2−30%, ∼31−55 < 64%, and ≥ ∼71% H2O saturation in scCO2, respectively. Exposure of sub 1W clay to anhydrous scCO2 caused interlayer expansion, not contraction as expected for dehydration, suggesting that CO2 intercalated the interlayer region of the sub 1W clay, which might provide a secondary trapping mechanism for CO2. In contrast, control experiments using pressurized N2 and similar initial conditions as in the scCO2 study, showed little to no change in the d001 spacing, or hydration states, of the clay. A salient implication for cap rock integrity is that clays can dehydrate when exposed to wet scCO2. For example, a clay in the ∼3W hydration state could collapse by ∼3 Å in the c* direction, or ∼15%, if exposed to scCO2 at less than or equal to about 64% H2O saturation.
■
INTRODUCTION Ameliorating the impact of anthropogenic carbon dioxide (CO2) on climate has focused the attention of Earth scientists on the possibility of large scale geologic sequestration of CO2 by injecting it as a super critical fluid into depleted oil and gas, or more generally saline brine, formations.1 To a first approximation, storage strategies rely on physical retention of the buoyant super critical CO2 (scCO2) by impermeable caprock layers, followed by solubility trapping through contact with formation waters and, ideally over time, permanent fixation via precipitation of carbonate minerals.2 The latter mechanism should be viable in basalt terrains3−5 which contain highly reactive alkaline Earth-bearing silicates. In particular, laboratory based studies on the interaction of forsterite (Mg2SiO4) with CO2 saturated brine has shown that Mg-carbonates readily form at conditions relevant to geologic storage of CO2.6−9 Further, recent experimental results have shown that reaction of variably wet scCO2 (i.e., the single phase fluid) with forsterite also yields magnesium carbonates.10−12 However, the majority of potential repositories presently being considered do not contain sufficient reactive alkaline Earth containing minerals for metal carbonization to be a viable natural trapping mechanism.1,13,14 Here, both long-term physical and solubility trapping are required. Regardless, under most injection scenarios, it is imperative that the repository have an impermeable caprock.15,16 In this regard, modeling and experimental studies have focused on the effects of mineral reactivity and over pressurization on the integrity of overlying © 2012 American Chemical Society
caprocks when exposed to one fluid phase (aqueous brines containing dissolved CO2) or two phase fluids (aqueous brine + scCO2).17−21 The same concerns apply to the integrity of seals around monitoring and abandoned wells that penetrate the caprock.22−24 The low permeability of most caprocks and engineered barriers is due to the presence of phyllosilicates including the micas and smectites. Recent experimental work on micas, high layer charge nonexpandable 2:1 phyllosilicates, has shown that the basal plane is susceptible to acid attack, with near-field precipitation of nanosize reaction products through contact with H2O saturated scCO2 or CO2 saturated brines at potential repository conditions.21,25 Also interesting is the apparent illitization of chlorites in CO2−H2O experiments with a natural shale material.26 Such mineral transformations, also discussed by Credoz et al.,27 modify mineral volumes and ultimately the mechanical response of a caprock.26 Whether the net result is one of porosity increase or decrease is the subject of ongoing research. Another group of 2:1 phyllosilicates falls under the generic term smectites which have a lower layer charge than the micas. This allows the interlayer region between the T-O-T structural units to expand or contract, the degree to which depends on layer charge, the interlayer cation, and the activity Received: Revised: Accepted: Published: 4241
January 19, 2012 March 8, 2012 March 11, 2012 March 11, 2012 dx.doi.org/10.1021/es300234v | Environ. Sci. Technol. 2012, 46, 4241−4248
Environmental Science & Technology
Article
fraction suspensions were flocculated by addition of NaCl (5M). To further ensure saturation of clay sites with Na, the concentrated clay was maintained in the NaCl solution for 7 days. Finally, the excess salt was removed by dialyzing the suspension in DI, followed by freeze-drying. After saturation with ethylene glycol, the d001 basal reflection expanded from 12.23 to 17.52 Å, which is typical of this type of Na-saturated swelling clay.41 Experimental Setup. An experimental high pressure X-ray diffraction reactor, previously described in Schaef et al.,42 was used to examine the clay at pressure and temperature. The reactor consists of a stainless steel base, small beryllium inserts or posts, and a beryllium cap that is pressure rated to 200 bar. Small aliquots of dilute clay suspensions (1.9 mg) were deposited on the Be post and allowed to air-dry. The post was attached to the base with a small flat head screw and placed on a custom built XYZ environmental stage attached to a BrukerAXS D8 Discover XRD unit equipped with a rotating Cu anode (1.54 Å), göbel mirror, 0.5 mm collimator, and 0.5 mm pinhole (Madison, WI). Initially, the clay films were aligned using a video laser alignment system. Several patterns were collected while adjusting the XYZ sample stage to optimize the intensity of the d001 reflection before securing the Be cap. Patterns were collected with a GADDS area detector positioned at 21.5° 2θ with a measured distance from the sample of 15 cm. Collection of individual XRD tracings required 200 s with power settings of 45 kV and 200 mA. Images were processed with Bruker-AXS GADDS software before importing into JADE XRD software to obtain peak positions, intensities, and full width at halfmaximum (fwhm) values, reported in °2θ. Pressurization with CO2 or N2 was accomplished using a high pressure syringe pump (ISCO) rated to 482 bar. Reagent grade CO2 and N2 were purchased from Oxarc (Pasco WA). Characterization. Hydration properties of Na-STx-1 were established through characteristic features of the d001 series, where the d001 positions correlate to the intercalation of 0, 1, 2, or 3 planes of H2O molecules,43 which are defined as dehydrated (0W), monohydrated (1W), bihydrated (2W), and trihydrated (3W), respectively. At ambient conditions (∼27 °C and 1.01 bar), Na-saturated montmorillonite produces a range of interplanar spacings depending on the hydration state: 0W (9.6−10.1 Å), 1W (12.3−12.7 Å), 2W (15.0−15.8 Å), and 3W (18.5−19.0 Å).28,29 Thermogravimetric analysis coupled to mass spectrometry (TGA-MS) was performed on a TG-209F1 (Netzsch Instruments) by heating an unreacted sample (equilibrated at ambient conditions; 1W hydration state with d001 = 12.45 Å) to 900 °C with a heating ramp of 2 °C/min under N2 flow (20 mL/min). Mass spectrometry was done simultaneously on an Aeolos QMS-403C. Generally, between 1 and 5 mg of sample was analyzed while monitoring masses corresponding to H2O (18 m/z). Characterization of the unreacted clay indicated three weight loss steps at ∼60 °C, ∼110 °C, and ∼650 °C representing a loss of adsorbed water, chemically bound water associated with the exchangeable Na+ cation, and structural water equal to ∼9.5 wt %. These results are consistent with values reported by Shilling35 for STx-1. Experimental Approach. Two different experimental approaches were followed. In experimental setup 1, variable amounts of water were placed in the reactor but not in direct contact with the clay. The system was sealed and then equilibrated for 1 h at 50 °C prior to pressurization with CO2 or N2. These experiments were designed to correlate the d001
of H2O, as well as pressure and temperature. It follows that caprock and well-seal integrity should also be a function of the swelling characteristics of expandable clays, in particular for locations where there is a high degree of dry scCO 2 advection.16 There is extensive literature on the relationship between clay composition, relative humidity, and interlayer (or interlamellar) expansion/contraction at ambient conditions, e.g., refs 28 and 29. In contrast, there are very few in situ experimental studies on the swelling characteristics of clays exposed to the low-water environment of anhydrous to wet scCO2 under potential geologic repository conditions. Recently, Schaef et al.30 reported an in situ XRD study on the interaction of scCO2 with a Ca saturated montmorillonite (CaSTx-1) from 90 to 180 bar and 50° to 100 °C. It was demonstrated that the 1W hydration state was stable when exposed to anhydrous scCO2 under all conditions, regardless of whether the initial hydration state was 1W or 2W. Further, exposure of slightly sub 1W Ca-STx-1 to anhydrous scCO2 yielded a small expansion of the interlayer spacing that was consistent with intercalation of CO2. Giesting et al.31 described in situ transmission mode XRD characterization of the interaction of Na montmorillonite (SWy-1) with gaseous CO2 and dry scCO2 at 45 °C and 50−650 bar. The response of d001 to CO2 was dependent on the initial water content, where nearly dry clay (d001 = 10.1 Å) showed no significant response but partially hydrated clay (d001 = 10.5 Å) expanded to 12.25 Å, with similar results over the entire pressure range. The focus of these previous studies was the interaction of dry scCO2 with variably hydrated clay. In contrast, the objective of this study is to characterize the swelling properties of Na+ saturated montmorillonite (Na-STx-1) exposed to variably wet scCO2 at conditions relevant to carbon sequestration in geologic formations. Throughout the paper we use the term wet-scCO2 to mean the one phase fluid dominated by CO2 but containing variable amounts of dissolved H2O. Basal reflections were monitored by in situ high pressure X-ray diffraction (HXRD) at 50 °C and 90 bar. Peak positions were compared to published values to ascertain the hydration state of the clay. Control experiments using N2 helped assess whether CO2 was sorbed in the interlayer region and more generally whether any observed structural changes were due to CO2 or temperature and pressure effects.
■
MATERIALS AND METHODS Materials. Naturally occurring low iron montmorillonite (STx-1), obtained from the Clay Mineral Repository (Columbia, MO), was used for all of the experiments and has been previously well characterized.32−39 STx-1, as received, has virtually all the layer charge emanating from the octahedral sites a n d th e f o llo w i ng p u b l i s he d s tr u c t u r a l for m u l a (Ca0.27Na0.04K0.01)(Al2.41Fe3+0.09MntrMg0.71Ti0.03)2Si8O20(OH)4. The clay was chosen precisely because of its very low Fe content so that we could eventually compare this in situ XRD study with ongoing parallel in situ NMR experiments using the same material under similar conditions. To remove traces of calcite, the STx-1 was reacted with 1 M sodium acetate buffer at pH 5 for 2 weeks, following the procedure outlined by Kunze.40 The acetate was removed by dialyzing the clay in deionized water (DI) until the conductivity of the washes was 18 MΩ cm. Separation of the
