Structure-Dependent Interactions between Alkali Feldspars and

Sep 14, 2012 - Department of Energy, Environmental and Chemical Engineering, Washington University in St. Louis, St. Louis, Missouri 63130, United Sta...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/est

Structure-Dependent Interactions between Alkali Feldspars and Organic Compounds: Implications for Reactions in Geologic Carbon Sequestration Yi Yang,† Yujia Min,† and Young-Shin Jun*,† †

Department of Energy, Environmental and Chemical Engineering, Washington University in St. Louis, St. Louis, Missouri 63130, United States S Supporting Information *

ABSTRACT: Organic compounds in deep saline aquifers may change supercritical CO2 (scCO2)-induced geochemical processes by attacking specific components in a mineral’s crystal structure. Here we investigate effects of acetate and oxalate on alkali feldspar−brine interactions in a simulated geologic carbon sequestration (GCS) environment at 100 atm of CO2 and 90 °C. We show that both organics enhance the net extent of feldspar’s dissolution, with oxalate showing a more prominent effect than acetate. Further, we demonstrate that the increased reactivity of Al−O−Si linkages due to the presence of oxalate results in the promotion of both Al and Si release from feldspars. As a consequence, the degree of Al−Si order may affect the effect of oxalate on feldspar dissolution: a promotion of ∼500% in terms of cumulative Si concentration was observed after 75 h of dissolution for sanidine (a highly disordered feldspar) owing to oxalate, while the corresponding increase for albite (a highly ordered feldspar) was ∼90%. These results provide new insights into the dependence of feldspar dissolution kinetics on the crystallographic properties of the mineral under GCS conditions.



INTRODUCTION The geologic sequestration of anthropogenic carbon dioxide (CO2) in deep saline aquifers is a promising method for storing CO2 emissions from fossil fuel combustion and therefore for remediating global warming due to elevated concentrations of greenhouse gases.1 Compared to typical atmospheric conditions for water−rock interactions, geochemical processes in deep saline aquifers are characterized by elevated temperature, pressure, and high salinity of formation waters. If the temperature and pressure of an aquifer exceed the critical point of CO2 (73.8 atm and 31.1 °C), injected CO2 will be initially stored as a supercritical fluid. Supercritical CO2 (scCO2) is a good solvent for organic matter and may cause redistribution of organic constituents between the formation waters and the porous media in deep saline aquifers and thus affect subsequent water−rock interactions.2 A report from the Frio formation (Texas, USA) showed that 20 days after the termination of scCO2 injection, the concentration of dissolved organic carbon had increased unexpectedly by a factor of 100, corresponding to enriched concentrations of organic constituents such as formate, acetate, and toluene.3 A question remains to be answered is whether organic compounds, either enriched by CO2 extraction or pre-existing in the formation waters, affect water−rock interactions under geologic carbon sequestration (GCS) conditions in a way similar to what occurs under atmospheric conditions. A comprehensive investigation of water−mineral interactions under GCS conditions is required © 2012 American Chemical Society

as a basis for comparison with studies conducted under nonGCS conditions. In this work, we conducted a systematic investigation of brine−feldspar interactions in the presence of acetate and oxalate under GCS conditions. Feldspar is the most abundant mineral group constituting 60% of the earth’s crust and is present in both caprock and sandstone of typical reservoirs.4 Acetate is the most abundant organic component in subsurface waters, with reported concentrations ranging from 5 × 10−5 M to 0.25 M.5 Oxalate has long been known to facilitate aluminosilicate dissolution under aqueous and low temperature conditions6−17 and has a reported concentration of up to 5 × 10−3 M in formation waters.18 The dependence of feldspar dissolution kinetics on several parameters, such as pH,10,12,13,19−25 chemical affinity,26−30 organic and inorganic ligands, 5,7,11,31−34 and temperature,22,33,35−38 has been intensively studied. Depending on the element of concern (and thus the bond between the atom and the crystal’s framework), two general mechanisms are involved:13,22,39,40 the release of interstitial cations through ion Special Issue: Carbon Sequestration Received: Revised: Accepted: Published: 150

June 10, 2012 September 11, 2012 September 14, 2012 September 14, 2012 dx.doi.org/10.1021/es302324m | Environ. Sci. Technol. 2013, 47, 150−158

Environmental Science & Technology

Article

feldspars (plagioclases)43 and thus are focused here. Two Krich feldspars with different parageneses and appearances (sanidine and microcline) and one Na-rich feldspar (albite) are chosen as model samples. Sanidine and microcline are similar in their chemical compositions but differ in the degree of Al/Si order in their constituent tetrahedra. Microcline and albite have a highly ordered Al/Si distribution in their frameworks, whereas for sanidine the Al/Si distribution is substantially more random.

exchange, and the release of framework elements (Al or Si) through the hydrolysis of intertetrahedra linkages (T−O−T). The rate of ion-exchange can be affected by the aqueous concentration of the cations, whereas the rate of hydrolysis can be changed by the presence of protons, hydroxyl groups, and organic ligands. Organic ligands can promote feldspar dissolution by the formation of aqueous complexes with dissolution products, increasing the apparent solubility of the mineral, and hence the chemical affinity for the feldspardissolution reaction.27 Organic matter may also promote or inhibit dissolution by binding to reactive sites on the surface.41 Under GCS conditions, however, experimental determinations of the effects of organic constituents on feldspar−brine interactions are scarce.42 The overall objective of this work was to study the effects of representative organics on mineral−brine interactions under GCS conditions and to explore potential dependence of such effects on the mineral structure. Specifically, we studied the effects of acetate and oxalate on alkali feldspar−brine interactions to reveal how the aqueous composition in a deep saline aquifer might respond to scCO2 injection in the presence of organic matter. We further interpreted the mineral-specific effects of organic constituents in terms of the reactivity difference between linkages in the feldspar framework.



MATERIALS AND METHODS Sodium acetate trihydrate (NaOAc·3H2O, pKA = 4.9 at 90 °C),44 sodium oxalate (Na2Ox, pKA1 = 1.5, pKA2 = 6.2 at 90 °C),44 sodium chloride (NaCl), and deionized water (18.2 MΩ resistivity, Barnstead Ultrapure water systems) were used to create solutions with the compositions listed in Table S1. The system pH was adjusted using trace metal hydrochloric acid (HCl, 34%−37%, BDH). All chemicals used in this study were at least ACS grade. The sanidine used in this study was obtained from Daichi, Wakayama prefecture, Japan, and consisted of mostly homogeneous crystal fragments. The microcline was from 80 km southeast of Virginia City, Madison County, Montana, and consisted of pure pale blue-green masses. The albite was from Minas Gerais, Brazil, and the samples were semitransparent large chunks with lamellae and apparent cleavages. The crystal phase, chemical composition, and the degree of Al−Si ordering of the natural samples were characterized using X-ray powder diffraction (XRD, Rigaku Geigerflex D-MAS/S, Japan; Figure S1 in the Supporting Information (SI)), X-ray fluorescence (XRF) analysis (Table S2), and Fourier transform infrared spectroscope (FTIR, Thermo Scientific, Nicolet Nexus 470; Figure S2), respectively. The XRF analysis suggested that the chemical formulas for the microcline, sanidine, and albite were Na 0 . 3 K 0 . 7 AlSi 3 O 8 , Na 0 . 2 3 K 0 . 6 6 Ca 0 . 0 1 Al 1 . 0 3 Si 2 . 9 7 O 8 , and Na0.98K0.02AlSi3O8, respectively. The Na contents in K-feldspars indicated that both samples contained ∼30% exsolved albite. We use the terms sanidine and microcline below to describe the two K-rich alkali feldspar samples because their different dissolution behaviors stemmed from their K-feldspar portions. FTIR analysis suggested that the albite and microcline samples were substantially more ordered than sanidine, manifested by increased intensity, number, and sharpness of peaks, and shifts in band frequencies in the spectra (Figure S2).45−47 Quartz (purum p.a., acid purified) of 40−150 mesh was purchased from Fluka Analytical. The feldspar samples were first ultrasonicated in deionized water for 15 min. Purer crystal pieces were picked out and manually ground into a fine powder. Sieves were used to bracket particle sizes (250−425 μm for module I experiments, and 53−106 μm for module II experiments; see Figure S3 and Section S1 in the SI for detailed module descriptions). The specific surface areas of the powders used in module II were measured using BET (AX1C-MP-LP, Quantachrome Instrument) and were 0.2089 m2/g for albite, 0.4478 m2/g for microcline, and 0.2058 m2/g for sanidine. The higher value for microcline may be attributable to the exsolution of albite because the strained-interfaces between exsolution-induced domains is sensitive to mechanical grinding and can thus contribute to the measured surface area. The experimental setup has been described in several previous publications.2,42,48,49 Two reactor modules have been used to mimic different mixing states in a GCS system. A



BACKGROUND INFORMATION Feldspars belong to the tectosilicate group and, as such, it shares structural similarity with quartz (the most abundant constituent of sandstone), and it has both Al−O−Si linkages and Si−O−Si linkages in its framework (Figure 1). Whereas

Figure 1. Crystal structure of alkali feldspar viewed from the b direction. C1̅ nomenclature is used, and a unit cell is boxed. Green balls mark locations for interstitial cations (Na for albite, K for sanidine and microcline). Red balls show the locations of tetrahedral sites. Blue balls are bridging oxygens. Ball sizes do not reflect effective atomic radius. Tetrahedral sites are occupied by either Al or Si. Within one four-membered tetrahedral ring there are four types of tetrahedral sites (T1o, T1m, T2o, and T2m), which differ in the probability of their being occupied by Al. For a perfectly ordered feldspar (e.g., albite), Al can only be found in T1o.

Al−O−Si is important in aluminosilicates, Si−O−Si exists in all tectosilicates. Using feldspar as the model mineral allows us to compare the reactivities of both structures. There are three naturally occurring feldspar end-members, which differ by their interstitial cations. In many reservoirs, sodic and potassic feldspars (alkali feldspars) are more abundant than calcic 151

dx.doi.org/10.1021/es302324m | Environ. Sci. Technol. 2013, 47, 150−158

Environmental Science & Technology

Article

Figure 2. Evolution of cumulative aqueous concentrations in alkali feldspar−brine interaction experiments. The feldspars (250−400 μm) were reacted with brines under 100 atm of scCO2 at 90 °C in module I. The pH values were adjusted for all solutions so that they had the same initial pH (pH = 3.1) after pressurization. The concentration profiles show the net result of primary feldspar dissolution and subsequent secondary mineral formation.



RESULTS The Effects of Organic Compounds on Feldspar− Brine Interactions under GCS Conditions Showed Mineral Dependency. Figure 2 shows the evolution of the cumulative concentrations of Al and Si during the reactions of microcline, sanidine, and albite with brines at 100 atm of CO2 and 90 °C in reactor module I (stagnant system. Results of control experiments with 100 atm of N2 are given in Figure S4). The K concentrations were also measured for the potassic feldspars (microcline and sanidine). The evolution of Na concentration was not monitored due to its high background level. Cumulative aqueous concentrations reflect the overall effects of the organic ligands on feldspar−brine interactions, including feldspar dissolution and the potential subsequent formation of secondary mineral phases. The effects of organic compounds are both mineral- and element-specific. For potassic feldspars, the cumulative concentration of K, the major interstitial cation, slightly increased in the presence of organics. For Al and Si, the framework elements of feldspars, both acetate and oxalate resulted in higher cumulative concentrations, and the effect of oxalate was more prominent than that of acetate. Also, the qualitative patterns of concentration evolution (the shape of the lines) and of netdissolution enhancement (the distances between data lines: the effectiveness of organic constituents on net dissolution) are similar for Al and Si, although the actual concentrations differ.

detailed description with a schematic diagram can be found in the SI. All solution samples were analyzed with an inductively coupled plasma-mass spectrometer (ICP-MS) (7500ce, Agilent Technologies, CA). The temperature (90 °C), pressure (100 atm), and NaCl concentration (1.0 M) used in this study fall into the range of conditions of deep saline aquifers suitable for a GCS implementation.48,50 Ten millimolar of organic compounds were used. The concentrations of oxalate and acetate may differ in the field, but they were kept identical in this study for easy comparison of their effectiveness on feldspar dissolution. Thermal decomposition of acetate and oxalate was assumed negligible under the experimental conditions.2 The pH of each solution was adjusted so that it had the same initial pH after pressurization. We used an initial pH of 3.1. Without CO2, the pH of typical formation waters in equilibrium with silica and aluminosilicate assemblage would be circumneutral. However, after CO2 injection, sharp decreases in the pH have been reported.3,51,52 On the basis of our previous in situ pH measurement42 and a thermodynamics calculation using Geochemist’s Workbench (GWB, Release 8.0, RockWare, Inc.), 3.1 was the equilibrium pH of the simulated brine under the chosen conditions without organic compounds. More discussion regarding pH drifting during the experiments can be found in Section S1. 152

dx.doi.org/10.1021/es302324m | Environ. Sci. Technol. 2013, 47, 150−158

Environmental Science & Technology

Article

interaction with brine under GCS conditions has been most significantly affected by the presence of oxalate, whereas albite is the least affected. Second, in terms of mineralogy, sanidine and albite represent potassium-rich and sodium-rich alkali feldspars, respectively, and have different characteristics of Al− Si ordering. Increases in dissolution rates in the presence of oxalate were observed for both feldspars, although to different extents. To test the effect of oxalate on the breakdown of Si−O−Si linkages, the evolutions of the cumulative concentrations of Si in quartz dissolution experiments were monitored (Figure 4).

The effectiveness of oxalate in increasing the cumulative concentrations of feldspar framework elements is similar for albite and microcline (both Al/Si-ordered feldspars, with different interstitial cations), whereas a much stronger promotion in terms of the extent of net framework dissolution was observed for sanidine (disordered potassic feldspar). The increased cumulative aqueous concentrations in the presence of organic constituents, as reflected in Figure 2, may stem from an increase in feldspar’s dissolution rate or an inhibition of secondary mineral formation. To delineate between these two possible causes, we used reactor module II (equipped with a mechanical stirring mechanism). Here we focused on oxalate, because it showed more significant influence on the feldspar−brine interactions. Similar to mineral weathering under ambient conditions,53 the difference between the effectivenesses of oxalate and acetate can be attributed to their interactions with Al/Si centers on the mineral surface. The surface Al/Si sites act as Lewis acid centers whose covalent binding to the mineral bulk can be polarized by the approach of organic ligands to the coordination sphere. Oxalate, which can form mononuclear bidenatation with the acidic centers, may thus bring more negative charge to surface sites and result in greater extents of polarization. Acetate, on the other hand, showed less significant influence owing to a weaker interaction with surface sites. Figure 3 shows the evolution of cumulative Al and Si concentrations in dissolution experiments of sanidine and albite, conducted in reactor module II. We focused on comparing the results for sanidine and albite for two reasons: First, based in Figure 2, sanidine is the feldspar whose

Figure 4. Evolution of Si concentration during quartz dissolution experiments under 100 atm of scCO2 at 90 °C in reactor module II. The pH value was adjusted for the solution with oxalate so that both experiments had the same initial pH (pH = 3.1) after pressurization. Open symbols, reaction without oxalate; closed symbols, reaction with 0.01 M oxalate. 1.0 M NaCl was used in both experiments to simulate the salinity of formation waters.

The experiments were conducted at 100 atm of CO2 and 90 °C (pH = 3.1, with 1.0 M NaCl) using reactor module II. No appreciable dissolution-promoting effect of oxalate was observed up to 80 h. Between 20 to 40 h, the experiment without oxalate showed even higher Si concentrations than the experiment with oxalate, although the concentrations in both experiments approached a similar level after 70 h. Oxalate May Affect Secondary Mineral Formation Indirectly. The formation of secondary mineral(s) during feldspar−brine interactions was suggested by a comparison between the stoichiometries of dissolution products in module I and in module II. Figure 5a shows the cumulative concentrations of Si plotted against those of Al for sanidine, which showed the greatest extent of dissolution among all three feldspars. The results for albite are also included (Figure 5b) for comparison. For experiments conducted in module II, a preferential release of Al was observed for the first 40 h for sanidine, and released Al and Si were in a ratio of approximately 1:2.5 (solid green line). The ratio gradually increased to a value of approximately 3, indicating an apparent congruent dissolution after 60 h of reaction. In module I, a significantly lowered Si to Al ratio was observed, suggesting an enrichment of Si in the solid phase due to the formation of Si-containing precipitates. In contrast, for albite (Figure 5b), whose net extent of dissolution was lower, the stoichiometric differences between results from the different modules were not as significant, suggesting a relatively smaller amount of precipitation. The introduction of oxalate, however, did not appreciably change the reaction stoichiometry in either module. This invariability suggested that oxalate itself either did not alter the phase of the secondary mineral directly or resulted in a phase alteration which followed approximately the same Al/Si

Figure 3. Evolution of framework element (Al and Si) concentrations during alkali feldspar dissolution experiments. Both feldspars (53−106 μm) were reacted under 100 atm of scCO2 at 90 °C in reactor module II. The pH values were adjusted for all solutions so that they had the same initial pH (pH = 3.1) after pressurization. The data w/o oxalate were obtained in experiments without any organic. Oxalate’s promotion of feldspar dissolution can be observed for both minerals. However, the effect was more prominent for sanidine than for albite. Triangles, Al concentrations; circles, Si concentrations; open symbols, reaction without oxalate; closed symbols, reaction with 0.01 M oxalate. 1.0 M NaCl was used in all experiments to simulate the salinity of formation waters. 153

dx.doi.org/10.1021/es302324m | Environ. Sci. Technol. 2013, 47, 150−158

Environmental Science & Technology

Article

(Figure 3). Hence, the effect of the instability on dissolution, if any, was independent of the effect of oxalate. Also, in the presence of oxalate, microcline dissolved slower than sanidine. The instability thus indicates a thermodynamically favorable transformation from a more reactive phase to a less reactive one. Moreover, the phase transformation is slow. For example, Goldsmith and Laves54 reported that for the transformation between microcline and sanidine, no conversion was detectable after 504 h of annealing at 500 °C. Our experimental temperature was much lower (90 °C), thus we did not expect the occurrence of significant phase transition within the time frame of our dissolution experiments (160 h). Hence, the organic-specific effect is less likely attributable to the instability of sanidine under the temperature and the pressure of the reaction environement. The change in the chemical affinity of the feldspar dissolution reaction due to the presence of aqueous complexation agents is an important mechanism for dissolution enhancement. This mechanism, however, is less likely to be responsible for the mineral-dependent effect of oxalate observed in this study for two reasons. First, the oxalate concentrations used in dissolution experiments with different minerals were identical, yet the oxalate’s effects on mineral dissolution (i.e., the net extent of dissolution) are significantly different. If the increase in the net extent of dissolution is due to the aqueous complexation of oxalate with dissolution products and is independent of the bulk structures of the primary minerals, we would expect similar extents of dissolution enhancement for all three feldspars. However, that is not true, as shown in both Figures 2 and 3. If the varied effects of oxalate on mineral dissolution are to be attributed to the differences in the dissolution rates as functions of chemical affinities, then it essentially means that these effects stem from the structuredependent nature of these functional relationships. The second reason is that although the exact relationship between dissolution rate and chemical affinity is unknown (without assuming the applicability of transition-state-theory rate law to the overall reaction), our systems were all far from equilibrium with respect to the primary mineral when the effect of oxalate became evident. Moreover, the mineral with the greater extent of dissolution (e.g., sanidine) did not show a slowed-down dissolution in the later phase of the dissolution experiment (40−80 h, Figures 2 and 3), where the aqueous concentrations of dissolution products were high. According to previous experimental studies for albite dissolution, there is a “dissolution-plateau” where the system is far from equilibrium,55 during which the rate is not very sensitive to chemical affinity. For K-feldspar, it has been reported that no dissolution plateau exists, and so chemical affinity may always play a role.55 However, the two K-feldspars we used show quite different behaviors in this study, which means that chemical affinity was not the predominant factor in our experimental systems. In summary, interstitial cation content, surface area, therymodynamic stability of the starting material, and aqueous complexation cannot fully explain the observed dissolution behaviors of feldspars in our experiments. The observed diverse effects of oxalate can be interpreted by the structural differences in the feldspars’ frameworks. As discussed in the introduction, two mechanisms are involved in feldspar dissolution. Interstitial cations (Na, K) are released through ion exchange between feldspars and the solution, whereas the decomposition of feldspar framework, usually measured by the release rate of its tetrahedral site occupants (Al

Figure 5. Stoichiometries of sanidine (a) and albite (b) dissolution products. Open symbols, data from reactor module I (dissolution and precipitation); closed symbols, data from reactor module II (dissolution only); circles, 1 M NaCl; squares, 1 M NaCl + 0.01 M NaOAc; triangles, 1 M NaCl + 0.01 M Na2Ox. The solid green line has a slope of 2.5. The dotted blue line has a slope of 1.5 and is drawn to guide the reader’s eyes.

ratio as those minerals that formed in a system without organic constituents. In either case, oxalate changed the saturation state of potential secondary mineral(s) indirectly by facilitating feldspar dissolution. More discussion regarding the identification of secondary mineral phases can be found in Section S3.



DISCUSSION Several factors, such as the interstitial cation content, surface area, phase stability of feldspars, and aqueous complexation, affect the observed dissolution behaviors of feldspars. In this study, the Na content of the structure did not show a significant influence on the effect of organic constituent: The effects of oxalate are similar for albite and microcline, which had a difference of approximately 70% in terms of Na content. Microcline and sanidine, which had a difference of less than 7% in the Na content, showed significantly different dissolution behaviors in the presence of oxalate. Surface area is another key parameter determining the apparent dissolution rates of minerals. On the basis of the BET measurements, we see that the surface areas are similar for albite and sanidine (0.2089 m2/ g vs 0.2058 m2/g), whereas the mineral with the highest surface area (microcline, 0.4478 m2/g) did not show more significant oxalate-related changes in the experiments. Thus the differences in surface areas cannot explain the varying effectivness of oxalate on the dissolution of alkali feldspar. Similarly, concerns may arise regarding the stability of sanidine under the hydrothermal reaction conditions. Sanidine underwent a slow transformation to a more ordered phase (C2/M ⇔ C1̅ transition) which might have had an effect on its dissolutoin. However, a more rapid dissolution of sanidine was observed only when oxalate was present. In experiments under the same temperature and pressure but without oxalate, sanidine did not show a significantly faster dissolution rate 154

dx.doi.org/10.1021/es302324m | Environ. Sci. Technol. 2013, 47, 150−158

Environmental Science & Technology

Article

disscussion to the whole tectosilicate mineral group and consider that the Si−O−Si linkages of minerals in the group share a similar breakdown mechanism. Quartz thus serves as a good model tectosilicate, with neither interstitial cation nor Al− O−Si linkage, and its dissolution depends solely on the reactivity of Si−O−Si linkages.65,66 Figure 4 shows the evolution of Si concentration in quartz dissolution experiments under the same GCS conditions. As described above, no appreciable dissolution-promoting effect of oxalate was observed. The result indicates that should there be any interaction between oxalate and Si−O−Si linkages under GCS conditions, it would not be identical or proportional to oxalate’s promoting effect on the breakdown of Al−O−Si linkages. Therefore, it is more plausible to consider the increased reactivity of Al−O−Si linkages as the origin for the promotions of both Al and Si release rates from feldspars in the presence of oxalate. The reason for the temporal difference in Si concentrations in Figure 4 requires further investigation, which is beyond the scope of our current manuscript. The differences in the minerals’ crystallographic properties can explain the varied effectiveness of oxalate on sanidine and albite dissolution. It is noted that the sanidine sample used in this study contained ∼30% of Na-feldspar, presumably due to the albite/sanidine exsolution. The observed differences between the sanidine sample and the albite sample can thus be attributed predominatly to the K-feldspar portion of the natural sample. Albite features a more ordered Al and Si distribution than sanidine.59 Using the C1̅ nomenclature for illustration (Figure 1), Si occupants of the T1m sites in albite are almost always bound to the bulk structure by Si−O−Si linkages because of the very low probability of finding Al in T2o and T2m sites. If an albite is “perfectly ordered”, aluminum atoms appear only in T1o sites, one-third of the silicon atoms are completely isolated from aluminum, and any change in the reactivity of Al−O−Si linkages can partially affect the release of only two-thirds of the Si atoms. Sanidine, on the other hand, is substantially more disordered in its Al−Si distribution. A direct consequence of the randomness in the Al−Si substitution is that every Si in the sanidine framework, regardless of the type of the tetrahedral site it is occupying, has a chance of being connected to an Al occupant (the exact probability is computable given site-specific Al occupancy).59,67 Hence, a promotion in the breakdown of Al−O−Si linkages can potentially exert an influence on the release of all Si atoms in a sanidine’s framework. The dissolution stoichiometry analysis in Figure 5 is consistent with the speculation that the promoted Si release is associated mechanistically with oxalate’s effect on Al−O−Si linkages. Without organic reactants, the predominant mechanism for Si release from feldspar in the acidic regime is the proton-mediated breakdown of Al−O−Si linkages.5,40,58 Figure 5 shows that the introduction of oxalate did not change the Al/ Si ratio in the feldspar dissolution experiments, which may suggest that the effect of oxalate on Si release is similar to that of protons, in that they both polarize the covalent bond between Al and O. Although there are certain differences: oxalate acts as a electron pair donor which interacts with surface Al as Lewis acid centers, while protons serve as an electron pair acceptor and interact with bridging oxygens directly. Al−O−Si linkage is thus a key reactive component in feldspar’s framework, and its interaction with both protons and oxalate determines the kinetics of feldspar dissolution under GCS conditions.

or Si), involves the breakdown of Al−O−Si or Si−O−Si linkages (T−O−T linkages) by hydrolysis. The crystal structure of feldspar dictates that each intact tetrahedral site is connected to the bulk strucutre by four such T−O−T linkages (Figure 1), and the removal of one T-atom (Al or Si) is essentially a process in which the connectedness (Q, the number of intact T− O−T linkages)40 of a tetrahedral site is reduced from four to zero. Where GCS is concerned, the two linkage types (Al−O− Si and Si−O−Si) may show very different reactivities: the hydrolysis of Al−O−Si can be greatly facilitated by protonation,5,56 whereas the rate of Si−O−Si linkage breakdown may not be very sensitive to pH in the acidic region (seen from the dependence of quartz dissolution rate on pH in the acidic region).57 Interstitial cations are released faster than framework elements. In Figure 2, the changes in the K concentration features a steep initial increase between 0 and 5 h, regardless of the mineral type, suggesting the formation of an interfacial layer depleted of interstitial cations. It has been proposed that the ion exchange which leads to interstitial cation release is a prerequisite step for the protonation of Al sites.58 However, because of the distinct difference in rates, the release of interstitial cation and of framework elements appeared to be independent.39 After the rapid initial stage the K concentration, nevertheless, reflected the renewal of the interfacial layer and hence the decomposition of feldspar framework.42 The K concentrations were only slightly affected by the presence of organic ligands. Organic ligands may change the release rate of tetrahedral site occupants by modifying the reactivities of T−O−T linkages. Aluminum is connected to the bulk structure of feldspar by four Al−O−Si linkages (Al−O−Al linkage is considered energetically unfavorable),59 whereas Si can be connected to the bulk by different combinations of Al−O−Si and Si−O−Si linkages, depending on the mineral’s aluminum content and the state of Al−Si order.59 Thus, any modification in the reactivity of Al−O−Si linkages will be reflected in the cumulative concentrations of both Al and Si, whereas changes in the reactivity of Si−O−Si can be observed only in Si concentration. Figure 2 shows that oxalate promotes Si release from feldspars in a very similar pattern to that for the promotion of Al release, especially for sanidine. Two potential explanations follow. Oxalate may facilitate only the breakdown of Al−O−Si linkages, promoting the Al release rate and subsequently the release rate of those Si atoms that are connected to the crystal bulk by Al−O−Si linkages.12 Some Si atoms are connected to the crystal bulk by four Al−O−Si linkages; thus an increase in the reactivity of Al−O−Si linkages will directly facilitate the release of these Si atoms. Many Si atoms are linked to the crystal bulk by n Al−O−Si linkages and 4-n Si−O−Si linkages (n = 0, 1, 2, or 3). For these Si, the breakdown of Al−O−Si linkages results in partial removal of the tetrahedral site and consequently a reconfiguration of the solid structure, as suggested by observations of Si-enriched layers at the solid−solution interface.60−62 The Si−O−Si linkages in the altered structure can differ in reactivity from the intact Si−O−Si linkages in unreacted feldspar and thus are indirectly affected by the modified reactivity of Al−O−Si.63 Alternatively, oxalate may interact directly with both Al−O−Si and Si−O−Si linkages and facilitate their breakdown with an exact proportionality. The essential difference between the two explanations is whether the inner sphere bidentate complex64 can form between oxalate and surface silicon sites under GCS conditions. A differentiation is possible if we extend the 155

dx.doi.org/10.1021/es302324m | Environ. Sci. Technol. 2013, 47, 150−158

Environmental Science & Technology



Article

Frio-I Brine Pilot test, Texas, USA. Appl. Geochem. 2009, 24 (6), 1106−1112. (4) Gaus, I. Role and impact of CO2-rock interactions during CO2 storage in sedimentary rocks. Int. J. Greenhouse Gas Control 2010, 4 (1), 73−89. (5) Ullman, W. J.; Welch, S. A., Organic ligands and feldspar dissolution. In Water-Rock Interactions, Ore Deposits, and Environmental Geochemistry: A Tribute to David A. Crerar; Hellmann, R., Wood, S. A., Eds.; Geochemical Society: St. Louis, 2002; pp 3−36. (6) Pokrovsky, O. S.; Shirokova, L. S.; Bénézeth, P.; Schott, J.; Golubev, S. V. Effect of organic ligands and heterotrophic bacteria on wollastonite dissolution kinetics. Am. J. Sci. 2009, 309 (8), 731−772. (7) Pokrovsky, O. S.; Golubev, S. V.; Jordan, G. Effect of organic and inorganic ligands on calcite and magnesite dissolution rates at 60 °C and 30 atm pCO2. Chem. Geol. 2009, 265 (1−2), 33−43. (8) Krevor, S. C.; Lackner, K. S. Enhancing process kinetics for mineral carbon sequestration. Energy Procedia 2009, 1 (1), 4867− 4871. (9) Olsen, A. A.; Rimstidt, J. D. Oxalate-promoted forsterite dissolution at low pH. Geochim. Cosmochim. Acta 2008, 72 (7), 1758−1766. (10) Berg, A.; Banwart, S. A. Carbon dioxide mediated dissolution of Ca-feldspar: implications for silicate weathering. Chem. Geol. 2000, 163 (1−4), 25−42. (11) Prapaipong, P.; Shock, E. L.; Koretsky, C. M. Metal-organic complexes in geochemical processes: Temperature dependence of the standard thermodynamic properties of aqueous complexes between metal cations and dicarboxylate ligands. Geochim. Cosmochim. Acta 1999, 63 (17), 2547−2577. (12) Blake, R. E.; Walter, L. M. Kinetics of feldspar and quartz dissolution at 70−80 degrees C and near-neutral pH: Effects of organic acids and NaCl. Geochim. Cosmochim. Acta 1999, 63 (13−14), 2043− 2059. (13) Stillings, L. L.; Drever, J. I.; Brantley, S. L.; Sun, Y. T.; Oxburgh, R. Rates of feldspar dissolution at pH 3−7 with 0−8 mM oxalic acid. Chem. Geol. 1996, 132 (1−4), 79−89. (14) Blake, R. E.; Walter, L. M. Effects of organic acids on the dissolution of orthoclase at 80 degrees C and pH 6. Chem. Geol. 1996, 132 (1−4), 91−102. (15) Fein, J. B.; Hestrin, J. E. Experimental studies of oxalate complexation at 80-degrees-c - gibbsite, amorphous silica, and quartz solubilities in oxalate-bearing fluids. Geochim. Cosmochim. Acta 1994, 58 (22), 4817−4829. (16) Fein, J. B. Porosity enhancement during clastic diagenesis as a result of aqueous metal-carboxylate complexation - experimental studies. Chem. Geol. 1994, 115 (3−4), 263−279. (17) Harrison, W. J.; Thyne, G. D. Predictions of diagenetic reactions in the presence of organic-acids. Geochim. Cosmochim. Acta 1992, 56 (2), 565−586. (18) Kharaka, Y. K.; Hanor, J. S.; Heinrich, D. H.; Karl, K. T. Deep fluids in the continents: I. Sedimentary basins. In Treatise on Geochemistry; Pergamon: Oxford, 2007; pp 1−48. (19) Amrhein, C.; Suarez, D. L. Some factors affecting the dissolution kinetics of anorthite at 25°C. Geochim. Cosmochim. Acta 1992, 56 (5), 1815−1826. (20) Berner, R. A.; Holdren, G. R. Mechanism of feldspar weathering: Some observational evidence. Geology 1977, 5 (6), 369−372. (21) Casey, W. H.; Westrich, H. R.; Arnold, G. W. Surface chemistry of labradorite feldspar reacted with aqueous solutions at pH = 2, 3, and 12. Geochim. Cosmochim. Acta 1988, 52 (12), 2795−2807. (22) Chen, Y.; Brantley, S. L. Temperature- and pH-dependence of albite dissolution rate at acid pH. Chem. Geol. 1997, 135 (3−4), 275− 290. (23) Holdren, G. R.; Speyer, P. M. pH dependent changes in the rates and stoichiometry of dissolution of an alkali feldspar at room temperature. Am. J. Sci. 1985, 285 (10), 994−1026. (24) Oelkers, E. H.; Schott, J.; Devidal, J.-L. The effect of aluminum, pH, and chemical affinity on the rates of aluminosilicate dissolution reactions. Geochim. Cosmochim. Acta 1994, 58 (9), 2011−2024.

ENVIRONMENTAL IMPLICATIONS A better understanding of the mechanisms of water−rock interactions under pertinent conditions is necessary to develop safer and more sustainable GCS operations. Here we demonstrated that organic reactants that may be present in deep saline aquifers caused feldspar−brine interactions and that the organic-sensitive component in feldspar’s framework under GCS conditions is the Al−O−Si linkage. Several implications follow: First, the effects of many environmental variables on mineral dissolution can ultimately be linked to a structurespecific mechanism. For example, the effects of pH (in acidic regimes) and of oxalate on feldspar dissolution are both associated with the reactivity of Al−O−Si linkages, although these two environmental variables differ in their effectiveness. Hence, structure-based source terms for typical sandstones and caprocks can be used in reactive-transport modeling as a supplement to current mineral-based methodologies. Second, deep saline aquifers with higher Al contents are more likely to be affected by the presence of organic constituents. An analogy can be found in feldspar mineralogy, where the effects of organic phases on calcic feldspars are more prominent than on alkali feldspars5,58 because feldspars with higher anorthite contents have more Al−O−Si linkages in their structure. Characterizing a potential sequestration site by its elemental composition should provide an estimation of the sensitivity of its scCO2-induced geochemical alterations to its formation water composition.



ASSOCIATED CONTENT

S Supporting Information *

All experimental conditions, detailed experimental setup descriptions, XRD, XRF, and FTIR analyses of feldspar specimens used in this study, scCO2 vs N2 control experiments and pertinent discussions, and HR-TEM images with electron diffraction patterns and matching results for potential secondary minerals. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Phone: (314) 935-4539. Fax: (314) 935-7211. E-mail: ysjun@ seas.wustl.edu, http://encl.engineering.wustl.edu/. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by the National Science Foundation’s Career Award (EAR-1057117) and the Consortium for Clean Coal Utilization. We thank Ms. Jessica Ray for HR-TEM analysis, Dr. Glenn Waychunas for albite specimens, and Mr. Yunke Liu for experimental assistances.



REFERENCES

(1) Matter, J. M.; Kelemen, P. B. Permanent storage of carbon dioxide in geological reservoirs by mineral carbonation. Nat. Geosci. 2009, 2 (12), 837−841. (2) Shao, H.; Ray, J. R.; Jun, Y.-S. Effects of organic ligands on supercritical CO2-induced phlogopite dissolution and secondary mineral formation. Chem. Geol. 2011, 290 (3−4), 121−132. (3) Kharaka, Y. K.; Thordsen, J. J.; Hovorka, S. D.; Seay Nance, H.; Cole, D. R.; Phelps, T. J.; Knauss, K. G. Potential environmental issues of CO2 storage in deep saline aquifers: Geochemical results from the 156

dx.doi.org/10.1021/es302324m | Environ. Sci. Technol. 2013, 47, 150−158

Environmental Science & Technology

Article

(25) Welch, S. A.; Ullman, W. J. Feldspar dissolution in acidic and organic solutions: Compositional and pH dependence of dissolution rate. Geochim. Cosmochim. Acta 1996, 60 (16), 2939−2948. (26) Beig, M. S.; Luttge, A. Albite dissolution kinetics as a function of distance from equilibrium: Implications for natural feldspar weathering. Geochim. Cosmochim. Acta 2006, 70 (6), 1402−1420. (27) Burch, T. E.; Nagy, K. L.; Lasaga, A. C. Free energy dependence of albite dissolution kinetics at 80°C and pH 8.8. Chem. Geol. 1993, 105 (1−3), 137−162. (28) Hellmann, R.; Tisserand, D. Dissolution kinetics as a function of the Gibbs free energy of reaction: An experimental study based on albite feldspar. Geochim. Cosmochim. Acta 2006, 70 (2), 364−383. (29) Luttge, A. Crystal dissolution kinetics and Gibbs free energy. J. Electron Spectrosc. Relat. Phenom. 2006, 150 (2−3), 248−259. (30) Sorai, M.; Sasaki, M. Dissolution kinetics of anorthite in a supercritical CO2-water system. Am. Mineral. 2010, 95 (5−6), 853− 862. (31) Knauss, K. G.; Copenhaver, S. A. The effect of malonate on the dissolution kinetics of albite, quartz, and microcline as a function of pH at 70-degree-C. Appl. Geochem. 1995, 10 (1), 17−33. (32) Kubicki, J. D.; Schroeter, L. M.; Itoh, M. J.; Nguyen, B. N.; Apitz, S. E. Attenuated total reflectance Fourier-transform infrared spectroscopy of carboxylic acids adsorbed onto mineral surfaces. Geochim. Cosmochim. Acta 1999, 63 (18), 2709−2725. (33) Welch, S. A.; Ullman, W. J. The temperature dependence of bytownite feldspar dissolution in neutral aqueous solutions of inorganic and organic ligands at low temperature (5−35°C). Chem. Geol. 2000, 167 (3−4), 337−354. (34) Alisa Mast, M.; Drever, J. I. The effect of oxalate on the dissolution rates of oligoclase and tremolite. Geochim. Cosmochim. Acta 1987, 51 (9), 2559−2568. (35) Carroll, S. A.; Knauss, K. G. Dependence of labradorite dissolution kinetics on CO2(aq), Al(aq), and temperature. Chem. Geol. 2005, 217 (3−4), 213−225. (36) Casey, W. H.; Sposito, G. On the temperature dependence of mineral dissolution rates. Geochim. Cosmochim. Acta 1992, 56 (10), 3825−3830. (37) Chou, L.; Wollast, R. Study of the weathering of albite at room temperature and pressure with a fluidized bed reactor. Geochim. Cosmochim. Acta 1984, 48 (11), 2205−2217. (38) Hellmann, R. The albite-water system 0.1. The kinetics of dissolution as a function of pH at 100-degrees-C, 200-degrees-C, and 300-degrees-C. Geochim. Cosmochim. Acta 1994, 58 (2), 595−611. (39) Schott, J.; Pokrovsky, O. S.; Oelkers, E. H. The link between mineral dissolution/precipitation kinetics and solution chemistry. Rev. Mineral. Geochem. 2009, 70 (1), 207−258. (40) Brantley, S. L.; Kubicki, J. D.; White, A. F. Kinetics of mineral dissolution. In Kinetics of water-rock interaction; Springer Science +Business Media, LLC: New York, 2007; p 151. (41) Welch, S. A.; Vandevivere, P. Effect of microbial and other naturally occurring polymers on mineral dissolution. Geomicrobiol. J. 1994, 12 (4), 227−238. (42) Yang, Y.; Ronzio, C.; Jun, Y.-S. The effects of initial acetate concentration on CO2-brine-anorthite interactions under geologic CO2 sequestration conditions. Energy Environ. Sci. 2011, 4 (11), 4596−4606. (43) Cole, D. R.; Chialvo, A. A.; Rother, G.; Vlcek, L.; Cummings, P. T. Supercritical fluid behavior at nanoscale interfaces: Implications for CO2 sequestration in geologic formations. Philos. Mag. 2010, 90 (17), 2339−2363. (44) Bethke, C. M.; Yeakel, S. Reference Manual: The Geochemist’s Workbench (Release 8.0); RockWare, Inc.: Golden, 2009. (45) Makreski, P.; Jovanovski, G.; Kaitner, B. Minerals from Macedonia. XXIV. Spectra-structure characterization of tectosilicates. J. Mol. Struct. 2009, 924−926 (0), 413−419. (46) Zhang, M.; Wruck, B.; Barber, A. G.; Salje, E. K. H.; Carpenter, M. A. Phonon-spectroscopy on alkali-feldspars: Phase transitions and solid solutions. Am. Mineral. 1996, 81.

(47) Zhang, M.; Salje, E. K. H.; Carpenter, M. A.; Parsons, I.; Kroll, H.; Reed, S. J. B.; Graeme-Barber, A. Exsolution and Al-Si disorder in alkali feldspars: Their analysis by infrared spectroscopy. Am. Mineral. 1997, 82, 849−857. (48) Hu, Y.; Ray, J. R.; Jun, Y.-S. Biotite−brine interactions under acidic hydrothermal conditions: Fibrous illite, goethite, and kaolinite formation and biotite surface cracking. Environ. Sci. Technol. 2011, 45 (14), 6175−6180. (49) Shao, H.; Ray, J. R.; Jun, Y.-S. Effects of salinity and the extent of water on supercritical CO2-induced phlogopite dissolution and secondary mineral formation. Environ. Sci. Technol. 2011, 45 (4), 1737−1743. (50) Shao, H.; Ray, J. R.; Jun, Y.-S. Dissolution and precipitation of clay minerals under geologic CO2 sequestration conditions: CO2− brine−phlogopite interactions. Environ. Sci. Technol. 2010, 44 (15), 5999−6005. (51) Kharaka, Y. K., Reactive transport modeling to study changes in water chemistry induced by CO2 injection at the Frio-I brine pilot. In Lawrence Berkeley National Laboratory: Lawrence Berkeley National Laboratory; LBNL Paper LBNL-3056E, 2010. (52) Lu, J.; Kharaka, Y. K.; Thordsen, J. J.; Horita, J.; Karamalidis, A.; Griffith, C.; Hakala, J. A.; Ambats, G.; Cole, D. R.; Phelps, T. J.; Manning, M. A.; Cook, P. J.; Hovorka, S. D. CO2−rock−brine interactions in Lower Tuscaloosa Formation at Cranfield CO2 sequestration site, Mississippi, U.S.A. Chem. Geol. 2012, 291 (0), 269−277. (53) Stumm, W.; Wollast, R. Coordination chemistry of weathering: Kinetics of the surface-controlled dissolution of oxide minerals. Rev. Geophys. 1990, 28 (1), 53−69. (54) Goldsmith, J. R.; Laves, F. The microcline-sanidine stability relations. Geochim. Cosmochim. Acta 1954, 5 (1), 1−19. (55) Lasaga, A. C. Fundamental Approaches in Describing Mineral Dissolution and Precipitation Rates. In Chemical Weathering Rates of Silicate Minerals; White, A. F., Brantley, S. L., Eds.; Mineralogical Society of America: Washington, DC, 1995; Vol. 31, pp 23−86. (56) Morrow, C. P.; Nangia, S.; Garrison, B. J. Ab initio investigation of dissolution mechanisms in aluminosilicate minerals. J. Phys. Chem. A 2009, 113 (7), 1343−1352. (57) Brady, P. V.; Walther, J. V. Kinetics of quartz dissolution at low temperatures. Chem. Geol. 1990, 82 (0), 253−264. (58) Blum, A. E.; Stillings, L. L. Feldspar dissolution kinetics. In Chemical Weathering Rates of Silicate Minerals; 1995; Vol. 31, pp 291− 351. (59) Ribbe, P. H. The crystal structure of the aluminum-silicate feldspars. In Feldspars and their reactions; Parson, I., Ed.; Springer: 1993. (60) Dobrowolski, R.; Stefaniak, E.; Bilinski, B.; Staszczuk, P. Study of the surface adsorption properties of feldspar. Adsorpt. Sci. Technol. 1999, 17 (2), 73−83. (61) Hellmann, R.; Penisson, J. M.; Hervig, R. L.; Thomassin, J. H.; Abrioux, M. F. An EFTEM/HRTEM high-resolution study of the near surface of labradorite feldspar altered at acid pH: evidence for interfacial dissolution-reprecipitation. Phys. Chem. Miner. 2003, 30 (4), 192−197. (62) Tsomaia, N.; Brantley, S. L.; Hamilton, J. P.; Pantano, C. G.; Mueller, K. T. NMR evidence for formation of octahedral and tetrahedral Al and repolymerization of the Si network during dissolution of aluminosilicate glass and crystal. Am. Mineral. 2003, 88 (1), 54−67. (63) Nangia, S.; Garrison, B. J. Ab initio study of dissolution and precipitation reactions from the edge, kink, and terrace sites of quartz as a function of pH. Mol. Phys. 2009, 107 (8−12), 831−843. (64) Furrer, G.; Stumm, W. The coordination chemistry of weathering: I. Dissolution kinetics of δ-Al2O3 and BeO. Geochim. Cosmochim. Acta 1986, 50 (9), 1847−1860. (65) Dove, P. M. The dissolution kinetics of quartz in sodium chloride solutions at 25 degrees to 300 degrees C. Am. J. Sci. 1994, 294 (6), 665−712. 157

dx.doi.org/10.1021/es302324m | Environ. Sci. Technol. 2013, 47, 150−158

Environmental Science & Technology

Article

(66) Dove, M. T.; Keen, D. A.; Hannon, A. C.; Swainson, I. P. Direct measurement of the Si−O bond length and orientational disorder in the high-temperature phase of cristobalite. Phys. Chem. Miner. 1997, 24 (4), 311−317. (67) Kroll, H.; Ribbe, P. H. Determining (Al,Si) distribution and strain in alkali feldspars using lattice parameters and diffraction-peak positions; a review. Am. Mineral. 1987, 72 (5−6), 491−506.

158

dx.doi.org/10.1021/es302324m | Environ. Sci. Technol. 2013, 47, 150−158