On the thermodynamics and kinetics of scorodite dissolution

Journal Pre-proofs On the thermodynamics and kinetics of scorodite dissolution Xiangyu Zhu, D. Kirk Nordstrom, R. Blaine McCleskey, Rucheng Wang, Xiancai Lu, Siliang Li, H. Henry Teng PII: DOI: Reference: S0016-7037(19)30579-4 https://doi.org/10.1016/j.gca.2019.09.012 GCA 11435 To appear in: Geochimica et Cosmochimica Acta Received Date: Accepted Date: 8 May 2019 6 September 2019 Please cite this article as: Zhu, X., Kirk Nordstrom, D., Blaine McCleskey, R., Wang, R., Lu, X., Li, S., Henry Teng, H., On the thermodynamics and kinetics of scorodite dissolution, Geochimica et Cosmochimica Acta (2019), doi: https://doi.org/10.1016/j.gca.2019.09.012 This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Elsevier Ltd. All rights reserved. On the thermodynamics and kinetics of scorodite dissolution Xiangyu Zhua, D. Kirk Nordstromb, R. Blaine McCleskeyb, Rucheng Wangc, Xiancai Lud, Siliang Lia, H. Henry Tenga* a Institute of Surface-Earth System Science, Tianjin University, Tianjin 300072, China b US Geological Survey, 3215 Marine St., Suite E. 127, Boulder, CO 80303, USA c State Key Laboratory for Mineral Deposits Research, School of Earth Sciences and Engineering, Nanjing University, Nanjing 210023, China d Key Laboratory of Surficial Geochemistry, Ministry of Education, School of Earth Sciences and Engineering, Nanjing University, Nanjing, Jiangsu, 210023, China * To whom correspondence may be addressed. Email: huihenry.teng@tju.edu.cn 1 Abstract Scorodite (FeAsO4·2H2O)water interaction is critical to As distribution and storage in surface environment but is inadequately understood due to ambiguities in the mineral’s stability and weathering rate at atmospheric conditions. In the present study we attempted to experimentally determine the thermodynamic and kinetic parameters needed to compute solubility and dissolution rate of scorodite at 25 oC. Experiments were carried out in low pH (1.15) solutions using specially synthesized large scorodite crystals (with Mean Diameter of 27.9 μm). Such experimental conditions ensured the results were not subject to the influence of secondary Fe-bearing phase precipitation and grain size effect. Measured equilibrium concentrations of Fe and As, along with newly published Fe-As complexes association constants, were first used to determine the solubility products Ksp and dissolution rates rn at 50 – 90 °C. The obtained Ksp ~ T and rn ~ T dependence was then used to derive the Gibbs free energy, enthalpy, entropy, heat capacity, and activation energy for the dissolution reaction. Finally, we extrapolated the measurements to 25 °C and obtained room temperature solubility and dissolution rate, scrutinized the pH effect on dissolution, and analyzed the G ~ rn relation of the dissolution reaction. Our results show that literature data are likely overestimated scorodite solubility at pH > 4 ~ 4.5 due to neglecting the effect of ferric iron hydrolysis. Estimated ambient condition dissolution rate is an order magnitude lower than the earlier report, implicating the importance of size effect, but is one to two orders of magnitude higher than that of common rock-forming minerals, cautioning the proposed use of scorodite for As fixation and storage. The determined rn ~ G relation cannot be fully fitted by the transition state model, particularly at near equilibrium, suggesting dissolution in this study may be controlled by defect-assisted surface reactions. 2 1 Introduction Arsenic is a highly toxic metalloid and a world-wide concern as a human carcinogen. Indiscriminate use of arsenical pesticides during the early to mid-1900s has led to extensive contamination of soils worldwide (Smith et al., 1998). The continuous use of groundwaters with As levels above 10ppb (current drinking water limit sanctioned by the World Health Organization) for irrigation purpose further aggravated the pollution in many regions of the globe (Berg et al., 2001; Nordstrom, 2002; Fendorf et al., 2010; Smedley and Kinniburgh, 2013; Maghakyan et al., 2017). Geochemical cycling of arsenic at surface conditions is largely controlled by As-forming and As-containing minerals through dissolution, precipitation, incorporation, sorption, and desorption (Fendorf et al., 1997; Ford, 2002; Bostick and Fendorf, 2003; Amirbahman et al., 2006; Benzerara et al., 2008; Deditius et al., 2008; Catalano et al., 2011; Li et al., 2011; Zhu et al., 2014; Karimian et al., 2017; Sowers et al., 2017; Catelani et al., 2018). Of those minerals, scorodite deserves special attention because it is one of the least soluble As phases and a pivotal secondary As-bearing mineral in acidic Fe(III)-As(V)-H2O systems, such as arsenic-contaminated soil and mine wastes (Drahota and Filippi, 2009). The low solubility is in fact the basis of the suggested arsenic fixation strategy in mining industry (Riveros et al., 2001; Langmuir et al., 2006). The dissolution behavior of scorodite has been studied extensively for the past three decades (Dove and Rimstidt, 1985; Nordstrom and Parks, 1987; Robins, 1987; Krause and Ettel, 1988; Bluteau and Demopoulos, 2007; Majzlan et al., 2012). While the results laid the groundwork for understanding scorodite stability in aqueous environment, critical discrepancies exist concerning the measured mineral solubility (Ksp，reaction 2) that varies by five orders of magnitude (10-20.24±0.58 to 10-25.68±0.52) from case to case (Chukhlantsev, 1956; Nordstrom et al., 2014). A number of issues may be factored into the literature Ksp’s inconsistency for scorodite. (1) The effect of crystal size on solubility. It is well-known that, due to the Gibbs-Thompson effect, solubility of small particles increases with increasing curvature (reciprocal of particle radius) of the dissolving grains (Cabrera et al., 1954; Lasaga and Blum, 1986; Teng, 2004; Navrotsky, 2004; Fan et al., 2006; Navrotsky et al., 2008). Natural scorodite is usually small (mostly nano-sized) and, despite repeated laboratory attempts, synthesized 3 scorodite crystals rarely exceeded ~1 μm in diameter (Dutrizac and Jambor, 1988; Baghurst et al., 1995; Demopoulos et al., 1995; Bluteau and Demopoulos, 2007; Gonzalez-Contreras et al., 2010; Majzlan et al., 2012; Okamura et al., 2013; Kitamura et al., 2015; Okibe et al., 2017). The only known method to produce large (up to tens of micrometers) scorodite crystals was not developed until 2008 (Fujita et al., 2008), restricting earlier work to the use of submicron sized particles. As such, it is conceivable that the wide scatter in the reported Ksp values may have at least partially originated from the crystal size effect. (2) The formation of secondary Fe-containing minerals during dissolution. Due to the very low water solubility of ferric iron hydroxide (~10-39), incongruent dissolution (i.e. precipitation of Fe(OH)3) will take place when scorodite dissolves at pH > 2 – 3. For those studies where pH was not maintained sufficiently low (see section 4.3 for details), significant errors may have occurred in computing the thermodynamic driving force of the overall reaction (Demopoulos, 2005). (3) Overlook of certain solution chemistry effect. This concern stems from the observation that activity coefficients and ion pairs such as Fe-As complexes and Fe-Cl or Fe-SO4 complexes were rarely considered in previous studies that acquired Ksp measurements. On a related note, while significant progress was made in the past three decades to deduce the mechanisms and surface processes of mineral dissolution through quantifying relations between dissolution kinetics and solution free energy change on the basis of Transition State Theory (Eyring, 1935a, b; Glasstone et al., 1941; Connor et al., 1979; Lasaga, 1981; Aagaard and Helgeson, 1982; Lasaga, 1998; Lüttge, 2006), step wave model (Lasaga and Luttge, 2001), and dislocation theory (Brantley et al., 1986; Lasaga and Blum, 1986; Teng, 2004), little is known in the applicability of such advance to scorodite dissolution. To the best of our knowledge, only two studies touched upon the aspects of dissolution kinetics. The first (Harvey et al., 2006) reported an empirical rate law while the second (Bluteau and Demopoulos, 2007) examined the pH and temperature effect on apparent dissolution rate (i.e. without normalization by the sample’s surface area), and none of them considered the thermodynamic effect (i.e. saturation state) on dissolution kinetics, nor the actual physical mechanisms of scorodite dissolution. In the present study, we attempt to quantify scorodite dissolution using specially synthesized large (~28 μm on average, Fig. 2 and 3) crystals and low pH (1.15) solutions to eliminate particle-size effect and secondary 4 Fe-containing phase precipitation. The experimental results are used to extract activation energy, enthalpy, entropy, and heat capacity of scorodite dissolution reaction, and to provide mechanistic insight into the dissolution process based upon the relationship between dissolution kinetics and free energy change. 2 Methods 2.1 Scorodite synthesis and characterization Ambient condition synthesis of scorodite in the Fe(III)-As(V)-H2O system is proven difficult (Dutrizac and Jambor, 1988). We used a modified Fujita et al. (2008) method where experimental solutions of 0.3 mol/L As (V) and 0.3 mol/L Fe (II) (prepared using analytical grade Na2HAsO4·7H2O and FeCl2·4H2O and adjusted to pH = 1.5 using12 M HCl) were first heated to boiling (T = 96.1–96.3 °C) state, followed by slow Fe (II) oxidization via air (instead of oxygen) bubbling. The setup (a 1-litre conical flask attached to a condenser pipe) was sealed completely for 7 days, and the resulting precipitates were subsequently collected (via filtration using a 0.45 μm filter) and thoroughly washed (by dilute HCl solution at pH = 1.5). The final product was air dried and stored in a sealed opaque glass bottle. The synthesized products were characterized by scanning electronic microscope (SEM) (ZEISS SUPPA 55) with an oxford INCA energy dispersive spectroscopy (EDS), x-ray diffractometer (Rigaku D/max III, Cu kα radiation, 40kV and 20mA at 0.02°/0.3s per thermogravimetric analysis (PerkinElmer Pyris 1 TGA). step from 3° to 70° 2Ɵ), and The surface area of the crystals (Se) was experimentally determined by the BET method using N2 adsorption (Quantachrome Autosorb-iQ2-XR with outgassing T of 105 °C) and further computed using particle size distribution measured by laser diffraction (Malvern Mastersizer 3000) via 𝑉𝑖 6∑𝑑 𝑖 6 𝑆𝑒 = 𝜌∑𝑉 = 𝜌𝐷𝑆 𝑖 (1) where Vi is the relative volume by particle size class di, ρ the material density (3.27 g/cm3 for scorodite), and Ds the Sauter diameter (mean particle size assuming spherical geometry). 5 2.2 Dissolution experiments The dissolution chamber consisted of two 500 ml bottles each attached onto a submersible stirrer (Supplementary material, Fig. S1). The whole set-up was immersed into a water bath where the temperature was maintained at desired settings with a variability of ± 0.1 °C via a Lauda-Brinkmann thermostats (Model RCS 6). Ten grams of scorodite were enclosed in a double-layered 18 μm polyester mesh bag and suspended in diluted HCl solution to prevent potential physical abrasion to the crystals from the magnetic bars spinning that at pre-set rates of 400–500 rpm. The pH of the dissolution solutions was maintained at 1.15 (± 0.05) throughout the experimental duration. Two sets of data were obtained in this study, a short term one collected in the initial 4-5h of the experiments and the other at long-term upon the establishment of a steady state. Dissolution experiments ran parallel in the two 500 ml bottles at 50, 70, and 90 °C, respectively. The scorodite bags were immersed in the solution only after the desired temperature was reached and stabilized. An aliquot of 1 ml experimental solution was sampled every 15 min to 1 h in the first 6 hours and every 1 or 2 days after until a steady-state was reached. The sampled solution was filtered through a 0.45 μm membrane and diluted 100% by 1% HNO3 solution for subsequent Fe and As analyses. 2.3 Solution chemistry analyses Solution pH was measured using a Thermo Scientific 815600 Orion Ross combination electrode and an Orion STAR A325 pH meter. The electrode was calibrated using the pH 1.00 and 1.68 standards (Geotech, NIST traceable) thermally equilibrated with the experimental solutions. The As and Fe contents in the solutions were routinely determined by Inductively-coupled plasma-optical emission spectrometry (ICP-OES, Perkin-Elmer 7300 DV) using the characteristic wavelengths of 188.979 and 193.696 nm for As (detection limit 0.03 mg L − 1) and 238.204 and 259.939 nm for Fe (detection limit 0.002 mg L − 1). Typical relative standard deviation of the measurements is ~5% for As and ~2% for Fe. For solutions with As in the range of 0.001 mg L-1 (below the ICP-OES’s detection limit), a hydride generation atomic absorption spectrometer (HG-AAS, Perkin–Elmer AAnalyst 300 equipped with a FIAS–100 flow injection system and a quartz cell) 6 capable of detecting As at 0.0001 mg L-1 levels was employed. The relative standard deviation of the HG-AAS analyses in this study was ~ 2%. 3 Experimental results 3.1 Mineral characterization The synthesized crystals show a pale-yellowish green color with a garlic odor. Sharp peaks in the XRD spectra match well with the standard pattern in the PDF database (NO.70-0825) with no identifiable minor phases found in background, confirming the crystallization of scorodite (Fig. 1). SEM-EDS analysis further reveal the absence of any meaningful chemical impurities in the crystals. Multi-point fitting BET surface area of the crystals is approximately 0.43 m2/g, comparable to the calculated value of 0.11 m2/g using particle size distribution data (Fig. 2). SEM microphotography of the harvested crystals show mostly pyramidal or pseudo-octahedral shapes with sizes ranging from 10 to 30 μm (Fig. 3), well consistent with the direct measurement (with mean particle size of 27.9 μm) and the BET surface area-based estimation (Ds = 17.3 μm). EDS analyses further confirm that the Fe to As ratio is close to the ideal stoichiometric value of unity. TGA curves indicate that the onset of weight loss in the synthesized scorodite occurs at 168 °C (Fig. 4), nearly identical to that reported by previous studies (Bluteau and Demopoulos, 2007; Le Berre et al., 2008; Gonzalez-Contreras et al., 2010). The total weight loss amounts to 15.5%, very close to the theoretical water content of 15.6 % in scorodite. 7 Fig. 1 An XRD spectrum of the synthesized scorodite in comparison to the standard pattern in the PDF database Fig.2 Size distribution of the synthesized scorodite crystals. Listed values indicate the sizes in 10, 50, and 90 percentile along with the mean and the Sauter diameter. 8 Fig. 3 SEM images showing the pyramidal or pseudo-octahedral crystal morphology of the synthesized scorodite Fig. 4 TGA analysis showing the thermally driven weight loss in the synthesized scorodite 3.2 Mineral dissolution behavior As expected, the accumulated As and Fe concentrations increased with T and over time for both short (Fig. 5a) and long term (Fig. 5b). Steady states were reached within 30 days for all the temperatures. The experiment at 70 °C failed to reach completion due to apparatus malfunction at day 15 and the steady state was extrapolated from the trend shown in the 50 and 90 °C experiments. This method may introduce an estimated 9 1% uncertainty in the final solubility product obtained at 70 °C. Evaporation encountered at 50 and 70 °C was insignificant but became non-negligible at 90 °C. Accordingly, the data collected at 90 °C was corrected using an evaporation model (See supplementary S-1). Fig. 5 Time dependent variations of arsenic concentration in the first 4 or 5 h of scorodite dissolution experiment (a) and long-term dataset showing both arsenic and iron concentrations up to 38 days (b). Solid line is second order polynomial fittings for eye guidance. 4. Discussion 10 4.1 Thermodynamic analysis The Fe to As ratio maintained close to unity during the experiments indicating that the dissolution proceeded stoichiometrically. Assuming the reaction 𝐹𝑒𝐴𝑠𝑂4·2𝐻2𝑂 = 𝐹𝑒3 + + 𝐴𝑠𝑂34 ― +2𝐻2𝑂 (2) held true and reached steady states in our experiments, the solubility product of scorodite was computed using the Fe and As concentrations averaged over the last 3 measurements of the experiments (121.3, 149.6, and 345.5 μmol/L for 50, 70, and 90 °C respectively), corrected by activity coefficients estimated from the Hückel equation (also known as the B dot equation. Hückel, 1925; Helgeson, 1969. Supplementary material, S-2) via the computational code of PHREEQCI after taking into consideration the recently published association constants for Fe (III)-As(V) and Fe (III)-Cl- complexes (Zhu et al., 2017). Such obtained values of Ksp fit nicely to the following equation (Fig. 6): (3) where A1, A2, and A3 are constants and T the absolute temperature. Given the resemblance of Equation (3) to the thermodynamic relation (4) where ΔrG, ΔrCp, ΔrS,* ΔrH* are the Gibbs free energy, heat capacity, entropy (0 K), and enthalpy (0 K) of the reaction, and R the gas constant (see derivation in supplementary material, S-3), one can readily see that the heat capacity of the dissolution reaction can be regarded constant between 50 and 90 °C. In addition, a value of Ksp = 10-25.81±0.1 at 25 °C can be extrapolated from the fit. This value agrees well with that (10-25.83±0.07) given by Langmuir et al. (2006) and the recent evaluation (10-25.68 ± 0.52) by Nordstrom et al. (2014). Moreover, equation (4) can be used to derive Gibbs free energy, enthalpy, entropy, and heat capacity (Table 1) for reaction (2). 11 Fig. 6 Dependence of calculated scorodite solubility product (log Ksp) on temperature (1/T). The solid line is a fit to Eq. (3) and the open circle is the extrapolated value at 25 °C Table 1 Estimated Gibbs free energy, enthalpy, entropy, heat capacity, and solubility product of scorodite dissolution shown as reaction (2) at 298.15 K and in the temperature range of 298.15 K < T < 363.15 K 298.15 K T (298.15-363.15 K) ∆r G (kJ mol-1) 147.32 ± 0.57 –495.43+10.258×T–1.4218×TlnT ∆r H (kJ mol-1) -71.53 ± 1.94 -495.43+1.4218×T ∆r S (J K-1 mol-1) -735.3 ± 6.9 -8836.0+1421.8×lnT 1421.8 ± 67.3 1421.8 ± 67.3 -25.81 ± 0.1 -535.8+25878/T+171×logT ∆r Cp (J K-1 mol-1) log Ksp 4.2 Kinetic analysis 12 Scorodite dissolution rates in present study were calculated following the methods reported by Chermak and Rimstidt (1990), Lasaga (1998), and Harvey et al. (2006). Briefly, the measured concentration (C) versus time (t) are first fitted to a second order polynomial equation of the form (5), where a, b, and c are fitting constants. Dissolution rate (r) is then estimated by the differential form of Equation (5): (6). Final, the normalized rate rn is obtained by the following equation (7), where Vsol (0.00055 m3 in present study) is the volume of solution, Asp ( BET surface area, 0.43 m2 g-1) the specific surface area of the scorodite, and mscor (10 g) the mass of scorodite. The initial (i.e. when the solutions were free of As and Fe) dissolution rates in the experimental temperature range (50-90 °C) increased from 10-10.3 to 10-8.9 mol m-2 s-1 with increasing temperature. The linear relation of the initial rates (lnrn, derived from the dataset shown in Fig. 5a) to 1/T (Fig. 7) suggests an Arrhenius behavior ( ,where A is a pre-exponential factor in mol m-2s-1 and Ea the activation energy in kJ mol-1) for scorodite dissolution, consistent with previous findings (Baghurst et al., 1995; Bluteau and Demopoulos, 2007). Furthermore, the value of Ea = 77.2 kJ mol-1 extracted from Fig. 7 closely matches that (77 kJ mol-1) reported at a similar pH (1.1) (Baghurst et al., 1995), but significantly differs from those acquired at high pHs (61 kJ mol-1 at pH= 6 and 100 kJ mol-1 at pH=7-9) (Bluteau and Demopoulos, 2007), presumably indicative of the effect of secondary Fe-containing phase formed at neutral to alkaline conditions. The room temperature (25 °C) dissolution rate extrapolated from the experimental data using the Ea value comes to be logrn = -11.3 mol m-2 s-1, more than one magnitude lower than that (log rn = -9.86 mol m-2 s-1) acquired by Harvey et al. (2006). 13 Given the similar temperature (22 vs. 25 °C) and milder pH (2 vs. 1.15) but more than 20-fold greater (9.5 vs. 0.43 m2/g) mineral surface area in Harvey et al. (2006) experiment, we argue that the rate in our dataset is likely free of the grain-size effect widely observed during crystal dissolution (Briese et al., 2017). Fig. 7 Plot of scorodite dissolution rate (lnrn) vs. temperature (1/T) showing the Arrhenius behavior Thermodynamic effect on scorodite dissolution was examined by considering the dependence of rate on the free energy state ( ) of the experimental solutions. non-linear dependence for all experiments (Fig. 8). Plots of rn vs. ΔG revealed Far from equilibrium, for example, when ΔG < -15 kJ mol-1 for the experiment at 50 °C, the rn vs. ΔG curve was dominated by a plateau (Nagy and Lasaga, 1992; Hellmann and Tisserand, 2006) where the rate varied little with respect to solution saturation state; as the solution evolved to intermediate undersaturation states (-15 < ΔG < -5 ~ -2 kJ mol-1), a sharp decrease in dissolution rate (one and two orders of magnitude for 50/70 °C and 90 °C experiments, respectively) occurred. Close to and very-near equilibrium (0 > ΔG > -2 ~ -5 kJ mol-1), the rate decrease slowed down and exhibited a somewhat linear dependence on ΔG. The relation between dissolution rate and solution ΔG is in general described by: 14 (8) where at equilibrium and at far-from equilibrium ( , rate stabilizes at k). The forms of f (ΔG) vary but can be mathematically derived from three models: (1) the transition state theory (TST, equation 9) (Lasaga, 1981; Aagaard and Helgeson, 1982; Helgeson et al., 1984); (2) TST modified by surface defects (equation 10) (Morse, 1983; Lasaga, 1998; Teng, 2004), and (3) a sigmoidal formula (a combination of (1) and (2), equation 11) (Burch et al., 1993; Hellmann and Tisserand, 2006; Xu et al., 2012): (9) (10) (11) The parameter n is an apparent reaction order, k the rate constants, and m adjustable factors, all obtained as fitting constants in practice. Applying the two mechanistic models (equations 9 and 10) to our data (Fig. 8) shows that the TST-based rate law significantly overestimates the measurements when ΔG < ~ 10 kJ mol-1, whereas the surface-defect equation gives a much more accurate description of the experimental results even very-near equilibrium. The empirical sigmoidal model, as expected, fits the data with high fidelity within the entire experimental saturation range. The finding that the defect-based model gives a more satisfactory approximation to the experimental measurements can be interpreted in the context of dislocation theory (Burch et al., 1993; Bandstra and Brantley, 2008; Xu et al., 2012). It is well understood that mineral dissolution preferentially occurs at high energy sites such as defects and dislocations where lattice strain occurs (Holdren and Berner, 1979; Helgeson et al., 1984). Far-from equilibrium, when |ΔG| is greater than the melting enthalpy of crystals, dissolution is controlled by spontaneous 2D surface nucleation of etch pits (Brantley et al., 1986; Lasaga and Blum, 1986; Teng, 2004) regardless further changes in ΔG. At intermediate equilibrium when |ΔG| is comparable to the strain energy, 15 etch pits can only form at the high energy sites, rendering a strong negative dependence of dissolution rate on ΔG. Finally, when close to and very-near equilibrium and |ΔG| is smaller than the energy barrier for defect-assisted pit nucleation, the dissolution only has a weak dependence on ΔG as the surface reaction is limited at pre-existing step sites. It appears that the TST model’s overestimation in dissolution rate increases with increasing temperature. For example, at |ΔG| = ~ 2 kJ mol-1, the discrepancy between TST-predicted and measured rates is a factor of ~3 at 50 oC (Fig. 8a) but stretches to over an order of magnitude when temperature increases to 90 oC (Fig. 8c). A clear interpretation for this observation is not readily available. One explanation may be the limitation of the TST model when applied to particle detachment/attachment to solid surfaces in solution systems (Joswiak et al., 2018). The original TST theory was developed for gas particle collision reactions where the rate constant is expressed as the product of the thermal vibrational frequency v = kBT/ħ (ħ is the Planck’s constant) and the partition function of the activation complex (Lasaga, 1998). For heterogeneous reactions in solution, however, it is shown that the frequency term needs to incorporate solute diffusion and the geometry of the activation complex (Hill, 1975; Hill, 1976). Hence it is possible that the actual dependence of dissolution rate on T is weaker than that depicted by the linear relation of v = kBT/ħ. 16 17 Fig. 8 Scorodite dissolution rate (rn) vs. the corresponding Gibbs free energy difference (ΔG) at 50 (a), 70 (b), and 90 (c) °C. Solid circles: measured rates; dotted lines: TST fitting; dashed lines: surface-defect fitting; solid lines: sigmoidal fitting. 4.3 Predicted pH effect on scorodite dissolution While scorodite dissolution may proceed as described by Reaction 2, the subsequent Fe3+ hydration (e.g. Fe3+ + H2O → Fe (OH)2+ + H+) becomes non-negligible when pH > 3. The removal of ferric iron not only changes the solution supersaturation but affects aqueous As concentration due to the adsorption of arsenate on the resultant iron hydroxide. Such pH effect can be quantitatively assessed using the Ksp value and other thermodynamic parameters obtained in this study. Plotting literature data of scorodite dissolution on our calculated solubility to pH relation shows a reasonable agreement at pH < 3, but reveals one to two order of magnitude difference at high pHs (Curve A, Fig. 9). When ferrihydrite precipitation is taken into consideration, however, the reported measurements become significantly closer to the prediction (Curve B, Fig. 9), indicating that majority of the early lab data were likely subject to the effect of ferrihydrite precipitation. If goethite instead of ferrihydrite is considered, the predicted release of As overestimates the reported values when pH > 2.5 (Curve C, Fig. 9) but closely matches our data of As in a field collected acid mining drainage (AMD) solutions (pH ~ 3, the chemical analysis of AMD was provided in the supplementary material, S-4). Demopoulos (2005) detected the presence of highly metastable nano-sized 2-line ferrihydrite phase associated with scorodite dissolution in laboratory. In long term, however, ferrihydrite will give rise to the more stable phase goethite, a common occurrence in natural environment (Cornell et al., 1989; Hamzaoui et al., 2002; Cudennec and Lecerf, 2006). Thus, it appears that the scorodite+goethite system may be more fitting for field settings. Finally, we estimated As adsorption using surface complexation model (see supplementary material for detailed parameter, S-5) and found that approximately 10% and 2% of the total arsenic may be surface-bound on ferrihydrite and goethite, respectively. These values are noticeable but likely insignificant if iron hydroxide is derived solely from scorodite dissolution. 18 A Fig. 9 Scorodite solubility (shown as arsenic concentration) at different pH (25 °C): comparison among congruent dissolution calculation, incongruent dissolution calculation, and experimental measurement; A: scorodite congruent dissolution curve; B: reaching equilibrium with ferrihydrite during scorodite dissolution; C: reaching equilibrium with goethite during scorodite dissolution. Open and cross symbols are literature data and solid circles show the field collected acid mine drainage data in this study. The solubility products of ferrihydrite and goethite are derived from the thermodynamic properties reported by Navrotsky et al. (2008, see supplementary material S-6). 4.4 Implications for environmental consequences Scorodite dissolution rate at Earth’s surface conditions estimated from the dataset reported in this study is on the order of 10-11 ~ 10-12 mol/m2/s. This rate is nearly 1 - 2 order of magnitude higher than those reported for common silicate minerals (e.g. K-feldspar, kaolinite) in ambient environment (Brantley et al., 2007). Assuming scorodite (grain size > 5 m) is the main storage of As in a porous geological medium (e.g. aquifer) that has an average water content of 15% (w/w) and a background As concentration of 15 ppm (Smith et al., 19 1998), dissolution at this rate (log rn = -11.3 mol m-2 s-1 at 25°C) would render initially As-free water to one with arsenic concentration surpassing the 10 ppb threshold value within ~17 hours. In reality this during could be significantly shorter as scorodite dissolution in natural settings is likely much (up to an order of magnitude) faster due to the small crystal sizes. This understanding, even taking into consideration of As adsorption by existing ferric oxyhydroxide or/and other minerals in nature, suggests that cautions need to be exercised when scorodite is used for As immobilization as natural waters in contact with the arsenic storage media can quickly (within days) become polluted. 5 Summary Scorodite dissolution was investigated at pH=1.15 in the temperature range of 50–90 °C. Assuming equilibria were established at the end of experiments, we extrapolated the measurements to 25 °C and obtained thermodynamic properties of the dissolution reaction at room temperature. Based upon the temperature dependence of dissolution kinetics, we also derived the activation energy of the dissolution reaction and subsequently estimated the rate of scorodite dissolution at ambient conditions. Estimated scorodite solubility at 25 oC is in good agreement with literature data at low pH conditions but is one to two orders of magnitude smaller when pH > 4 ~ 4.5. However, when ferrihydrite is allowed to precipitate, the match between estimation and the reported measurements are significantly improved, suggesting that previous studies may be subject to the effect of incongruent dissolution. On the other hand, calculated ambient condition scorodite dissolution rate is one to two orders of magnitude higher than that of common rock-forming minerals, indicating that previously assumed low solubility may not be a solid rationale for treating scorodite as a safe storage for As in natural environments or industrial settings. Finally, an analysis of the rn ~ G relation in the experimental duration suggests that scorodite dissolution in this study is largely a dislocation-controlled process. 20 Acknowledgments We would like to thank Kate Campbell and Tyler Kane of the USGS, and Huan Liu of Nanjing University for their assistance on many occasions during this study. We thank three anonymous reviewers whose comments and suggestions helped to improve the quality of this article. Xiangyu Zhu is grateful to the China Scholarship Council. This work was supported by the National Natural Science Foundation of China (Grant No. 41802032, 41830859, 41861144026) and the National Research Program of the USGS. Appendix A. Supplementary Material S-1, S-2, S-3, S-4, S-5, S-6 and Figure S1 Reference Aagaard, P.; Helgeson, H.C. (1982) Thermodynamic and kinetic constraints on reaction rates among minerals and aqueous solutions; I, Theoretical considerations. Am. J. Sci. 282, 65-77. Alexandratos, V.G., Elzinga, E.J. and Reeder, R.J. (2007) Arsenate uptake by calcite: macroscopic and spectroscopic characterization of adsorption and incorporation mechanisms. Geochim. Cosmochim. Acta 71, 4172-4187. Amirbahman, A., Kent, D.B., Curtis, G.P. and Davis, J.A. (2006) Kinetics of sorption and abiotic oxidation of arsenic (III) by aquifer materials. Geochim. Cosmochim. Acta 70, 533-547. Baghurst, D.R., Barrett, J. and Mingos, D.M.P. (1995) The hydrothermal microwave synthesis of scorodite: iron (III) arsenate (V) dihydrate, FeAsO4· 2H2O. J. Chem. Soc., Chem. Commun., 323-324. Bandstra, J.Z. and Brantley, S.L. (2008) Surface evolution of dissolving minerals investigated with a kinetic Ising model. Geochim. Cosmochim. Acta 72, 2587-2600. Benzerara, K., Morin, G., Yoon, T.H., Miot, J., Tyliszczak, T., Casiot, C., Bruneel, O., Farges, F. and Brown, G.E. (2008) Nanoscale study of As biomineralization in an acid mine drainage system. Geochim. Cosmochim. Acta 72, 3949-3963. Berg, M., Tran, H.C., Nguyen, T.C., Pham, H.V., Schertenleib, R. and Giger, W. (2001) Arsenic contamination of groundwater and drinking water in Vietnam: A human health threat. Environ. Sci. Technol. 35, 2621-2626. Bluteau, M.C. and Demopoulos, G.P. (2007) The incongruent dissolution of scorodite — Solubility, kinetics 21 and mechanism. Hydrometallurgy 87, 163-177. Bostick, B.C. and Fendorf, S. (2003) Arsenite sorption on troilite (FeS) and pyrite (FeS2). Geochim. Cosmochim. Acta 67, 909-921. Brantley, S.L., Crane, S.R., Crerar, D.A., Hellmann, R. and Stallard, R. (1986) Dissolution at dislocation etch pits in quartz. Geochim. Cosmochim. Acta 50, 2349-2361. Brantley, S., Kubicki, J. and White, A. (2007) Kinetics of Water-Rock Interaction. Springer, New York. Briese, L., Arvidson, R.S. and Lüttge, A. (2017) The effect of crystal size variation on the rate of dissolution – A kinetic Monte Carlo study. Geochim. Cosmochim. Acta 212, 167-175. Burch, T.E., Nagy, K.L. and Lasaga, A.C. (1993) Free energy dependence of albite dissolution kinetics at 80°C and pH 8.8. Chem. Geol. 105, 137-162. Cabrera, N., Levine, M.M. and Plaskett, J.S. (1954) Hollow dislocations and etch pits. Physical Review 96, 1153-1153. Catalano, J.G., Luo, Y. and Otemuyiwa, B. (2011) Effect of aqueous Fe(II) on arsenate sorption on goethite and hematite. Environ Sci Technol 45, 8826-8833. Catelani, T., Perito, B., Bellucci, F., Lee, S.S., Fenter, P., Newville, M., Rimondi, V., Pratesi, G. and Costagliola, P. (2018) Arsenic uptake in bacterial calcite. Geochim. Cosmochim. Acta 222, 642-654. Chermak, J.A. and Rimstidt, J.D. (1990) The hydrothermal transformation rate of kaolinite to muscovite/illite. Geochim. Cosmochim. Acta 54, 2979-2990. Chukhlantsev, V. (1956) Solubility products of arsenates. Journal of Inorganic Chemistry (USSR) 1, 1975-1982. Connor, J.N.L., Jakubetz, W. and Lagana, A. (1979) Comparison of quasi-classical, transition state theory, and quantum calculations of rate constants and activation energies for the collinear reaction X + F2 → XF + F (X = Mu, H, D, T). J. Phys. Chem. 83, 73-78. Cornell, R.M., Giovanoli, R. and Schneider, W. (1989) Review of the hydrolysis of iron(III) and the crystallization of amorphous iron(III) hydroxide hydrate. J. Chem. Technol. Biotechnol.46, 115-134. Cudennec, Y. and Lecerf, A. (2006) The transformation of ferrihydrite into goethite or hematite, revisited. J. Solid State Chem. 179, 716-722. Deditius, A.P., Utsunomiya, S., Renock, D., Ewing, R.C., Ramana, C.V., Becker, U. and Kesler, S.E. (2008) A proposed new type of arsenian pyrite: Composition, nanostructure and geological significance. Geochim. Cosmochim. Acta 72, 2919-2933. Demopoulos, G., Droppert, D. and Van Weert, G. (1995) Precipitation of crystalline scorodite (FeAsO4·2H2O) from chloride solutions. Hydrometallurgy 38, 245-261. Demopoulos G. P. (2005) On the preparation and stability of scorodite. In: Arsenic Metallurgy (eds. R. G. Reddy and V. Ramachandran). TMS Press, San Francisco. pp. 25-50. Dove, P.M. and Rimstidt, J.D. (1985) The solubility and stability of scorodite, FeAsO4·2H2O. Am. Mineral. 70, 838-844. 22 Drahota, P. and Filippi, M. (2009) Secondary arsenic minerals in the environment: a review. Environ. Int. 35, 1243-1255. Dutrizac, J. and Jambor, J. (1988) The synthesis of crystalline scorodite, FeAsO4·2H2O. Hydrometallurgy 19, 377-384. Eyring, H. (1935a) The activated complex and the absolute rate of chemical reactions. Chem. Rev. 17, 65-77. Eyring, H. (1935b) The activated complex in chemical reactions. J. Chem. Phys. 3, 107-115. Fan, C., Chen, J., Chen, Y., Ji, J. and Teng, H.H. (2006) Relationship between solubility and solubility product: The roles of crystal sizes and crystallographic directions. Geochim Cosmochim Ac 70, 3820-3829. Fendorf, S., Eick, M.J., Grossl, P. and Sparks, D.L. (1997) Arsenate and chromate retention mechanisms on goethite. 1. Surface structure. Environ. Sci. Technol. 31, 315-320. Fendorf, S., Michael, H.A. and van Geen, A. (2010) Spatial and temporal variations of groundwater arsenic in south and southeast asia. Science 328, 1123-1127. Ford, R.G. (2002) Rates of hydrous ferric oxide crystallization and the influence on coprecipitated arsenate. Environ. Sci. Technol. 36, 2459-2463. Fujita, T., Taguchi, R., Abumiya, M., Matsumoto, M., Shibata, E. and Nakamura, T. (2008) Novel atmospheric scorodite synthesis by oxidation of ferrous sulfate solution. Part II. Effect of temperature and air. Hydrometallurgy 90, 85-91. Glasstone, S., J. Laidler, K. and Eyring, H. (1941) The theory of rate processes: the kinetics of chemical reactions, viscosity, diffusion and electrochemical phenomena. McGraw-Hill Press, London. Gonzalez-Contreras, P., Weijma, J., Weijden, R.v.d. and Buisman, C.J.N. (2010) Biogenic scorodite crystallization by Acidianus sulfidivorans for arsenic removal. Environ. Sci. Technol. 44, 675-680. Hamzaoui, A., Mgaidi, A., Megriche, A. and El Maaoui, M. (2002) Kinetic study of goethite formation from ferrihydrite in alkaline medium. Ind. Eng. Chem. Res. 41, 5226-5231. Harvey, M.C., Schreiber, M.E., Rimstidt, J.D. and Griffith, M.M. (2006) Scorodite dissolution kinetics: implications for arsenic release. Environ. Sci. Technol. 40, 6709-6714. Helgeson, H.C. (1969) Thermodynamics of hydrothermal systems at elevated temperatures and pressures. Am. J. Sci. 267, 729-804. Helgeson, H.C., Murphy, W.M. and Aagaard, P. (1984) Thermodynamic and kinetic constraints on reaction rates among minerals and aqueous solutions II. Rate constants, effective surface area, and the hydrolysis of feldspar. Geochim. Cosmochim. Acta 48, 2405-2432. Hellmann, R. and Tisserand, D. (2006) Dissolution kinetics as a function of the Gibbs free energy of reaction: An experimental study based on albite feldspar. Geochim. Cosmochim. Acta 70, 364-383. Hill T. L. (1975) Effect of rotation on diffusion-controlled rate of ligand-protein association. Proc. Natl. Acad. Sci. USA 72, 4918–4922. Hill T. L. (1976) Diffusion frequency factors in some simple examples of transition-state rate theory. Proc. Natl. Acad. Sci. USA 73, 679–683. 23 Holdren, G.R. and Berner, R.A. (1979) Mechanism of feldspar weathering—I. Experimental studies. Geochim. Cosmochim. Acta 43, 1161-1171. Hückel, E. (1925) On the theory of concentrated aqueous solutions with strong electrolytes. Phys. Z. 26, 93–147. Joswiak, M.N., Doherty, M.F. and Peters, B. (2018) Ion dissolution mechanism and kinetics at kink sites on NaCl surfaces. Proc. Natl. Acad. Sci. 115, 656-661. Karimian, N., Johnston, S.G. and Burton, E.D. (2017) Antimony and arsenic behavior during Fe(II)-induced transformation of jarosite. Environ. Sci. Technol. 51, 4259-4268. Kitamura, Y., Okawa, H. and Sugawara, K. (2015) Synthesis of large scorodite particles using short period time sonication to enhance agglomeration of precursor. Jpn. J. Appl. Phys. 54, 1-6. Krause, E. and Ettel, V. (1988) Solubility and stability of scorodite, FeAsO4·2H2O: new data and further discussion. Am. Mineral.73, 850-854. Langmuir, D., Mahoney, J. and Rowson, J. (2006) Solubility products of amorphous ferric arsenate and crystalline scorodite (FeAsO4·2H2O) and their application to arsenic behavior in buried mine tailings. Geochim. Cosmochim. Acta 70, 2942-2956. Lasaga, A.C. (1981) Transition state theory. Rev. Mineral. 8, 135-170. Lasaga, A.C. (1998) Transition State Theory. In Kinetic Theory in the Earth Sciences (Ed. A.C. Lasaga). Princeton University Press, pp. 152-219. Lasaga, A.C. and Blum, A.E. (1986) Surface chemistry, etch pits and mineral-water reactions. Geochim. Cosmochim. Acta 50, 2363-2379. Lasaga, A.C. and Luttge, A. (2001) Variation of crystal dissolution rate based on a dissolution stepwave model. Science 291, 2400-2404. Le Berre, J., Gauvin, R. and Demopoulos, G. (2008) A study of the crystallization kinetics of scorodite via the transformation of poorly crystalline ferric arsenate in weakly acidic solution. Colloids Surf., A. 315, 117-129. Li, W., Harrington, R., Tang, Y., Kubicki, J.D., Aryanpour, M., Reeder, R.J., Parise, J.B. and Phillips, B.L. (2011) Differential pair distribution function study of the structure of arsenate adsorbed on nanocrystalline γ-Alumina. Environ. Sci. Technol. 45, 9687-9692. Lüttge, A. (2006) Crystal dissolution kinetics and Gibbs free energy. J. Electron Spectrosc. 150, 248-259. Maghakyan, N., Tepanosyan, G., Belyaeva, O., Sahakyan, L. and Saghatelyan, A. (2017) Assessment of pollution levels and human health risk of heavy metals in dust deposited on Yerevan’s tree leaves (Armenia). Acta Geochimica 36, 16-26. Majzlan, J., Drahota, P., Filippi, M., Grevel, K.D., Kahl, W.A., Plášil, J., Boerio-Goates, J. and Woodfield, B.F. (2012) Thermodynamic properties of scorodite and parascorodite (FeAsO4·2H2O), kaňkite (FeAsO4·3.5H2O), and FeAsO4. Hydrometallurgy 117-118, 47-56. Morse, J.W. (1983) The kinetics of calcium carbonate dissolution and precipitation. Rev. Mineral. Geochem. 11, 24 227-264. Nagy, K.L. and Lasaga, A.C. (1992) Dissolution and precipitation kinetics of gibbsite at 80°C and pH 3: The dependence on solution saturation state. Geochim. Cosmochim. Acta 56, 3093-3111. Navrotsky, A. (2004) Energetic clues to pathways to biomineralization: Precursors, clusters, and nanoparticles. Proc. Natl. Acad. Sci. 101, 12096-12101. Navrotsky, A., Mazeina, L. and Majzlan, J. (2008) Size-driven structural and thermodynamic complexity in iron oxides. Science 319, 1635-1638. Nordstrom, D. and Parks, G.A. (1987) Solubility and stability of scorodite, FeAsO4· 2H2O: discussion. Am. Mineral. 72, 849-851. Nordstrom, D.K. (2002) Public health - Worldwide occurrences of arsenic in ground water. Science 296, 2143-2145. Nordstrom, D.K., Majzlan, J. and Königsberger, E. (2014) Thermodynamic properties for arsenic minerals and aqueous species. Rev. Mineral. Geochem. 79, 217–255. Okamura, H., Mimura, M., Nagai, M., Itoh, H. and Komatsu, R. (2013) Growth and characterization of scorodite crystals from an aqueous solution by photo catalysis method. Trans. Mat. Res. Soc. Japan 38, 341-344. Okibe, N., Morishita, S., Tanaka, M., Sasaki, K., Hirajima, T., Hatano, K. and Ohata, A. (2017) Bioscorodite crystallization using Acidianus brierleyi: Effects caused by Cu(II) present in As(III)-bearing copper refinery wastewaters. Hydrometallurgy 168, 121-126. Riveros, P.A., Dutrizac, J.E. and Spencer, P. (2001) Arsenic disposal practices in the metallurgical industry. Can. Metal. Quart. 40, 395-420. Robins, R.G. (1987) Solubility and stability of scorodite, FeAsO4. 2H2O: Discussion. Am. Mineral. 72, 842-844. Robins, R.G., 1990. The stability and solubility of ferric arsenate: an update. In The Extraction and Processing Dvision (EPD) Congress '90 ( ed. D. R. Gaskell). The Minerals, Metals & Materials Society (TMS), Anaheim, California. pp. 93–104. Smedley, P.L. and Kinniburgh, D.G. (2013) Arsenic in Groundwater and the Environment. In Essentials of Medical Geology (Ed. O. Selinus.). Springer Netherlands, pp. 279-310. Smith, E., Naidu, R. and Alston, A.M. (1998) Arsenic in the soil environment: A review. In Advances in Agronomy (ed. D. L. Sparks). Elsevier Academic Press Inc, San Diego, pp. 149-195. Sowers, T.D., Harrington, J.M., Polizzotto, M.L. and Duckworth, O.W. (2017) Sorption of arsenic to biogenic iron (oxyhydr)oxides produced in circumneutral environments. Geochim. Cosmochim. Acta 198, 194-207. Teng, H.H. (2004) Controls by saturation state on etch pit formation during calcite dissolution. Geochim. Cosmochim. Acta 68, 253-262. Xu, J., Fan, C. and Teng, H.H. (2012) Calcite dissolution kinetics in view of Gibbs free energy, dislocation 25 density, and pCO2. Chem. Geol. 322-323, 11-18. Zhu, X., Nordstrom, D.K., McCleskey, R.B., Wang, R. and Lu, X. (2017) Thermodynamic properties in the Fe(II)-Fe(III)-As(V)-HClO4–H2O and Fe(II)-Fe(III)-As(V)-HCl–H2O systems from 5 to 90 °C. Chem. Geol. 460, 37-45. Zhu, X., Wang, R., Lu, X., Liu, H., Li, J., Ouyang, B. and Lu, J. (2014) Secondary minerals of weathered orpiment-realgar-bearing tailings in Shimen carbonate-type realgar mine, Changde, Central China. Miner. Petrol., 1-15. 26