Environ. Sci. Technol. 2002, 36, 2459-2463
Rates of Hydrous Ferric Oxide Crystallization and the Influence on Coprecipitated Arsenate ROBERT G. FORD* U. S. Environmental Protection Agency, Office of Research and Development, National Risk Management Research Laboratory, 919 Kerr Research Drive, Ada, Oklahoma 74820
Arsenate coprecipitated with hydrous ferric oxide (HFO) was stabilized against dissolution during transformation of HFO to more crystalline iron (hydr)oxides. The rate of arsenate stabilization approximately coincided with the rate of HFO transformation at pH 6 and 40 °C. Comparison of extraction data and X-ray diffraction results confirmed that hematite and goethite were the primary transformation products. HFO transformation was significantly retarded at or above an arsenate solid loading of 29 455 mg As/kg HFO. However, HFO transformation proceeded at a significant rate for arsenate solid loadings of 4208 and 8416 mg As/kg HFO. At a solid loading of 8416 mg As/kg HFO, XRD results suggested arsenate primarily partitioned to hematite. Comparison of HFO transformation rates observed in this research to rates obtained from the literature at pH 6 and temperatures ranging from 24 to 70 °C suggests that arsenate stabilization could be realized in oxic environments with a significant fraction of iron (hydr)oxides. While this process has not been documented in natural systems, the predicted half-life for transformation of an arsenic-bearing HFO is approximately 300 days at 25 °C at solid loading < 8415 mg As/kg HFO. The projected time frame for arsenate stabilization indicates this process should be considered during development of conceptual and analytical models describing arsenic fate and transport in oxic systems containing reactive iron (hydr)oxides. The likelihood of this process would depend on the chemical dynamics of the soil or sediment system relative to iron (hydr)oxide precipitation-dissolution reactions and the potential retarding/competing influence of ions such as silicate and organic matter.
Introduction Field and laboratory research has shown that partitioning of arsenic between solution and soil or sediment solids in aqueous systems is often tied to the cycling of iron (1-4). Arsenic is commonly associated with iron (hydr)oxide minerals in oxic systems (5, 6). Arsenic may be sequestered from solution during the precipitation of soluble iron resulting in the formation of a poorly crystalline hydrous ferric oxide (HFO) containing coprecipitated arsenic (7, 8). These solids are thermodynamically metastable and may transform to more crystalline products such as goethite or hematite (9, 10). This type of process is commonly observed in terrestrial systems where precipitation of a metastable phase precedes * Corresponding author phone: (580)436-8872; fax: (580)436-8703; e-mail:
[email protected]. 10.1021/es015768d Not subject to U.S. Copyright. Publ. 2002 Am. Chem. Soc. Published on Web 04/25/2002
the formation of thermodynamically stable phase(s) due to the energetics of solid-phase nucleation-precipitation (11, 12). The fate of arsenic during this transformation process is uncertain. Inorganic arsenic exists as a tetrahedral anion, arsenate or HAsO42-, under oxic conditions in natural waters (13, 14). Spectroscopic investigations indicate that arsenate forms inner-sphere complexes at surface bonding sites on iron (hydr)oxides (15, 16). Arsenate adsorption onto HFO is characterized by the formation of monodentate and bidentate corner-sharing bonds with surface iron octahedra (17, 18). Paige et al. (19) have examined the fate of arsenate coprecipitated with HFO during aging at pH 12 and temperatures of 60 and 70 °C. These authors observed arsenate desorption and retardation of HFO transformation in the presence of arsenate. Arsenate desorption correlated with the extent of transformation, suggesting arsenate immiscibility within the transformed iron (hydr)oxide phase(s). However, desorption would also be facilitated at the high pH of the experimental system. Several changes in arsenate partitioning may occur during HFO transformation to more crystalline forms: (1) arsenate may remain partitioned to the surface sites of HFO or the transformation product, (2) arsenate may desorb from the iron (hydr)oxide due to a decrease in surface area and subsequent loss of surface site density, or (3) arsenate may be incorporated or occluded within the structure of a more crystalline iron (hydr)oxide. The latter process is typically not considered, since common HFO transformation products, goethite and hematite, do not possess tetrahedral structural sites that could readily accommodate arsenate. However, there is evidence to suggest that the structure of goethite is flexible enough to accommodate variants to the ideal structure. It has been proposed that carbonate may be incorporated into the goethite structure during crystal growth in low-temperature systems resulting in the formation of an FeOOH-Fe(CO3)OH solid solution (20-22). It is postulated that carbonate forms an inner-sphere surface complex at bonding sites within structural grooves along the c crystallographic axis and is subsequently trapped during particle growth along the a and b crystallographic axes. In addition, there is evidence that a nonstoichiometric form of hematite, hydrohematite, precipitates in low-temperature environments (23, 24). This observation suggests that the hematite structure may also be capable of accommodating deviations from the ideal structure. These observations support the contention that the formation of nonideal crystal structures in near-surface environments provides an avenue for the stabilization of trace elements incorporated during crystal growth (e.g. ref 25). The research described herein was initiated to examine the influence of aging on the stability of arsenate coprecipitated with HFO under controlled laboratory conditions. Data were collected to assess (1) the transformation rate of HFO to more stable products, (2) the extent to which arsenate was stabilized within more crystalline iron (hydr)oxides, and (3) the rate of arsenate stabilization. These results, in conjunction with HFO transformation rate data available for a range of pH and temperature, were evaluated to provide a baseline assessment of the time scale of solid recrystallization processes that may stabilize arsenate coprecipitated with iron (hydr)oxides.
Methods and Materials Precipitation and Aging. Coprecipitates were generated by addition of 1 M NaOH to acidified solutions of ferric iron and VOL. 36, NO. 11, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2459
arsenate using an automatic titration system. Throughout this manuscript, coprecipitation refers to the partitioning of arsenate to HFO concurrent with hydrolysis and precipitation of ferric iron with no implied mechanism. Mixed ferric iron and arsenate solutions were prepared from stock solutions to achieve final concentrations of 0.01 M Fe(III) and 0-0.7 mM As(V). The ferric iron stock solution consisted of 0.1 M Fe(NO3)3 in 0.1 M HNO3, and the arsenate stock solution consisted of 7 mM Na2HAsO4‚7H2O in 0.1 M HCl. In all cases, starting titration solutions consisted of 20 mL of 0.1 M Fe(NO3)3, 20 mL of arsenate solution, and 140 mL of deionized water (Millipore 18 MΩ). The starting titration pH ranged between 1.5 and 1.7, and approximately 10 mL of 1 M NaOH was required to reach the pH ) 6 endpoint. Precipitate slurries were allowed to equilibrate for 1 h on the titrator followed by the addition of 10 mL of a 0.1 M solution of 2-(4morpholino)ethane sulfonic acid (MES) that had been preadjusted to pH ) 6. MES was employed to buffer the pH over long aging periods due to its demonstrated low trace element binding capacity and negligible impact on the rate of HFO transformation to crystalline products (26-28). The slurries were then allowed to age in a shaker bath at 40 °C. Based on the concentrations of Na+, NO3-, Cl-, and the deprotonated form of MES, the calculated ionic strength in each experiment was approximately 0.05 M. Arsenic Partitioning and Transformation Rate. The transformation of HFO to more crystalline forms was monitored via periodic extraction with 0.4 M HCl (29, 30). Dissolved iron was measured by reduction to Fe2+ with dithionite and spectrophotometric detection of its complex with 1,10-phenanthroline at 510 nm. Dissolved arsenic in solution prior to and following extraction was measured by inductively coupled plasma emission spectroscopy or graphite furnace atomic absorption spectroscopy (0.022 mg/L quantitation limit after dilution). The degree of arsenate partitioning to HFO was assessed by difference between dissolved and total arsenic prior to extraction. The fraction of nonextractable iron and arsenate was determined by difference between the total added mass and the extracted mass for either element. For all experiments, solid phase arsenate was greater than 99.4% during the entire period of observation (at least 112 d). Thus, arsenate desorption was insignificant under these experimental conditions in contrast to previously published results (7, 19). The rate of HFO transformation was calculated from extraction data assuming a first-order process (30). Solids collected at each sampling time were characterized by X-ray diffraction (XRD) using a thin film sample preparation method (31). XRD data was collected using Cu KR radiation via continuous scan in 0.1 °2θ steps at a rate of 0.01 °2θ min-1. No phases other than HFO (2-line ferrihydrite;), goethite (PDF 29-713), or hematite (PDF 33-664) were detected by XRD (32, 33). The relative quantitative distribution of crystalline iron (hydr)oxides was assessed via reference to a known mass of an internal standard (1.5 µm R-Al2O3, corundum; Johnson Matthey). Relative intensities were calculated with respect to neighboring corundum peaks for goethite and hematite, respectively.
Results and Discussion Fe and As Partitioning. Extraction data showing the transformation of HFO to more crystalline products are shown in Figure 1. The fraction of extractable iron is shown as a function of aging time for systems with a range of arsenate loading on the solid (Figure 1A). Transformation of HFO is retarded significantly at loading levels of 29 455 and 42 079 mg As/kg HFO. However, significant transformation occurs at arsenic loadings of 4208 and 8416 mg As/kg HFO. The natural logarithm was calculated for these data and pseudo-firstorder initial transformation rates were determined (Figure 2460
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 11, 2002
FIGURE 1. Extraction-based rate of HFO transformation over a range of arsenate surface loading at pH 6. (A) The fraction of extractable iron as a function of aging time. (B) Transformed extraction data assuming a first-order process. Duplicate analyses are shown for each aging time. 1B). These results demonstrate a strong dependence on arsenate loading, with an approximate order-of-magnitude reduction in the pseudo-first-order rate constant between systems with no arsenate and 42 079 mg As/kg HFO. Desorption of arsenate was not observed during aging for all arsenate solid loadings investigated. Thus, arsenate remained partitioned to the solid phase throughout the aging period. Data to assess the concurrent stabilization of arsenate are shown in Figure 2. The fraction of extractable arsenate is shown as a function of aging time for systems containing arsenate (Figure 2A). At time zero, >95% of the coprecipitated arsenate was extracted with 0.4 M HCl for all experiments (note results for 42 079 mg As/kg HFO system). Thus, arsenate associated with HFO was clearly distinguished from more stable forms. More than 80% of the solid phase arsenate is associated with an extractable form after aging for systems with 29 455 and 42 079 mg As/kg HFO. In contrast, less than 20% of the solid phase arsenate is extractable after aging for systems at lower loading levels. The extractable fraction of arsenate for aged systems with 4208 and 8416 mg As/kg HFO is less than the extractable iron fraction. This is shown more clearly by plotting the extractable As:Fe ratio normalized to the total As:Fe ratio in the coprecipitates (Figure 2B). Data that plot below a value of one indicate that arsenate is preferentially partitioned to a nonextractable phase relative to iron. This suggests that arsenate is being concentrated within the more crystalline iron (hydr)oxide fraction. Similar
FIGURE 2. Extraction-based rate of arsenate transformation at surface loading of 4208, 8416, 29 455, and 42 079 mg of As/kg HFO. (A) The fraction of extractable arsenate as a function of aging time. (B) The ratio of extracted As:Fe as a function of aging time. All values were normalized to the total As:Fe ratio in the coprecipitate for each arsenate solid loading. The solid line at a value of one indicates congruent partitioning of As and Fe into a nonextractable phase. behavior has been observed during crystallization of Mnand Ni-HFO coprecipitates (26). Independent assessment of HFO transformation to more crystalline phases was provided by XRD analysis of solid subsamples collected during aging. Relative XRD peak intensities for goethite and hematite as a function of aging time are shown in Figure 3. The slope of a linear regression to the relative intensity data was calculated to assess the relative predominance of goethite or hematite formation. Comparison of extraction data and XRD results indicated that while the overall rate of HFO transformation to more stable phases decreased with increasing arsenate solid loading (up to 8416 mg As/kg HFO), the apparent rate of hematite formation increased. At a loading of 8416 mg As/kg HFO, goethite was not detected in the solids until after 690 h, whereas hematite formed throughout all stages of the aging process (Figure 3C). This observation suggests that arsenate was partitioned to hematite, since stabilization of arsenate also occurred throughout the entire aging period for this system. However, this must be considered speculative without spectroscopic confirmation of the arsenic structural environment in aged samples. There was no evidence of peak shifts for either goethite or hematite that would suggest substitution of arsenate into crystallographic structural positions. Mineral-specific transformation data could not be derived for systems with higher arsenate loading due to the inability to detect crystalline iron (hydr)oxides by XRD. Numerous studies have documented the preferential formation of hematite from HFO in the presence of sorbed ions, presumably due to inhibition of HFO dissolution or
FIGURE 3. Mineral-specific transformation data for goethite and hematite based on XRD peak intensity ratios referenced to the internal standard, corundum. Peak intensity ratios for goethite and hematite are shown as a function of aging time (A) in the absence of arsenate and at an arsenic solid concentration of (B) 4208 mg As/kg HFO and (C) 8416 mg As/kg HFO. Data are shown for the goethite (111) and hematite (110) peak intensities relative to the corundum (104) peak intensity. Error bars are based on an assessment of the variability associated with triplicate preparation and analysis of one sample. The lines represent a linear regression through the data. G ) goethite, H ) hematite. nucleation of goethite (9, 34, 35). It has generally been observed that sorbed ions retard both goethite and hematite formation. However, Ford et al. (30) observed a promotion of the overall HFO transformation rate in the presence of sorbed Pb at pH 6 and 70 °C. Several studies have demonstrated a specific hematite-promoting effect in the presence of Cd or oxalate under certain aging conditions (36, 37). Insufficient molecular-level information is currently available to identify the specific mechanisms governing hematite/ goethite formation in the presence of sorbed ions. However, observed association of trace elements with goethite/ hematite in soils and sediments indicate that processes are active in natural systems that result in the partitioning during formation of these more crystalline solid phases (38-41). Numerous studies have clearly documented the association of arsenic with hydrous iron oxides in terrestrial systems (5, 8, 42-44). These results in combination with experimental studies indicate that the sorption capacity of HFO (or ferrihydrite) can approach values as high as approximately 500 000 mg As/kg HFO (7, 15, 18, 45, Table A.1 in Supporting Information)! Incorporation or occlusion of arsenate within the structure of goethite or hematite has previously not been observed in natural or experimental systems. It is difficult to envision substitution of arsenate into structural sites of goethite or hematite. HFO primary particles may serve as the building blocks for goethite or hematite nucleation and growth. Defect structures may develop during aggregationVOL. 36, NO. 11, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2461
FIGURE 4. Summary of first-order rates (left axis) and half-lives (right axis) for HFO transformation as a function of temperature at pH 6. A notation is included in the legend for systems not carried out at pH 6. (A) Estimated rates and transformation half-lives for Fe-only experiments conducted in the absence and presence of soil minerals. The regression line was fit to initial rates estimated for experiments conducted in the absence of soil minerals and aged at pH 6 (10-200 mM Fe). (B) Initial rates derived for all experiments in this study are shown as open triangles. Rates derived from published extraction data for HFO aging in the presence of L-cysteine are also shown. Details of reported experimental conditions and rate derivations for published data are provided in the Supporting Information. based crystal growth, and these defect sites may serve as the host for sorbed ions (46-48). Spectroscopic data supporting incorporation of hematite-like nuclei within the dissimilar structure of diaspore (AlOOH) provides compelling evidence for this mode of foreign ion inclusion (25). Waychunas et al. (15) indicated that coprecipitated arsenate disrupts the growth of HFO nuclei into more extended structures at solid loading of 100 000 mg As/kg HFO or greater. Both Waychunas et al. (49) and Manceau (17) observed an increase in Fe-Fe coordination number with aging (pH 8, 25 °C,
