SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 1 January 2007
Published in Soil Sci Soc Am J 71:101-107 (2007)
DOI: 10.2136/sssaj2006.0023
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Johnston, R. B.
Right arrow Articles by Singer, P. C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Johnston, R. B.
Right arrow Articles by Singer, P. C.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Johnston, R. B.
Right arrow Articles by Singer, P. C.
Related Collections
Right arrow Soil Mineralogy
Right arrow Toxic Trace Metals
Right arrow Geochemical Processes

SOIL CHEMISTRY

Solubility of Symplesite (Ferrous Arsenate): Implications for Reduced Groundwaters and Other Geochemical Environments

Richard B. Johnston*

United Nations Children's Fund (UNICEF), Water & Environmental Sanitation Section Dhaka, Bangladesh

Philip C. Singer

Dep. of Environ. Sciences and Eng. Univ. of North Carolina Chapel Hill, NC 27599

* Corresponding author (rjohnston{at}unicef.org).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Arsenic in drinking water represents a major public health concern globally, with tens of millions of people thought to be consuming unsafe amounts of As daily. Groundwater is particularly vulnerable to As contamination, and one geochemical environment conducive to natural mobilization of As is found in shallow alluvial Holocene aquifers. Bacterial consumption of organic matter in these aquifers creates anoxic conditions, leading to dissimulatory reduction of Fe oxides. In these environments, dissolved As and Fe often co-occur. While As(III) is thermodynamically favored under these conditions, As(V) often is found as well. Concentrations of dissolved As(V) and Fe(II) can be constrained by the precipitation of ferrous arsenate solid phases such as symplesite [Fe(II)3(As(V)O4)2·8H2O]. We present a new solubility product for symplesite (pKso = 33.25) based on controlled laboratory precipitation experiments. Using this solubility product, we conducted equilibrium geochemical modeling using one of the most extensive groundwater data sets from Bangladesh to show that a number of water samples are oversaturated with a variety of solid phases, including siderite, calcite, rhodochrosite, vivianite, and symplesite. Symplesite or other ferrous arsenate solid phases could be a significant sink for As(V) and Fe(II) in Bangladesh and other countries with similar geochemical environments. Symplesite could also control As(V) solubility in extreme environments such as alkaline lakes or in brines used for regeneration of anion exchange resins.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
During the past decade, it has become evident that naturally occurring As in groundwater is much more widespread than previously thought. Following the recognition in the mid-1990s that much of the alluvial groundwater of Bangladesh was contaminated with high levels of As—exceeding 1 mg L–1 in cases—water quality surveys have revealed the presence of geogenic As in other Asian countries, including Vietnam, Cambodia, Nepal, Myanmar, and Afghanistan. India and China have longstanding recognized zones of As contamination, but new contaminated areas within these countries are still being identified (e.g., Bihar state in India). Tens of millions of people are exposed to unsafe levels of As in drinking water in these countries, and tens of thousands of arsenicosis patients have already been identified.

Arsenic mobility in the subsurface is thought to be controlled by adsorption onto aquifer materials, particularly oxide minerals and clays. High levels of dissolved As can occur in anoxic aquifers, where partial or complete reductive dissolution of Fe oxides is driven by bacterial consumption of organic matter (Aggett and Obrien, 1985; Lovley, 1991; Nickson et al., 1998; Nickson et al., 2000). Oxic aquifers may also contain significant levels of dissolved As, especially under alkaline conditions where desorption of anions is favored (Smedley and Kinniburgh, 2002).

Under the strongly reducing conditions typical of young, organic-rich aquifers, As is typically found as As(III), while under oxidizing conditions the As(V) species are thermodynamically favored. However, redox transformations of As(III) to As(V) and vice versa are kinetically limited (Cherry et al., 1979) so that nonequilibrium distributions of As(III) and As(V) are often encountered in natural waters. For example, in Bangladesh the National Hydrochemical Survey measured arsenite and total As, along with other analytes, in 271 wells; 123 of these had measurable levels of dissolved As but dissolved O2 and NO3 were below detection limits (in most cases 0.1 and 0.3 mg L–1, respectively) (Kinniburgh and Smedley, 2001). In 54% of these anoxic wells, the As(III)/total As ratio was <0.5; in 17% of the wells the ratio was below 0.1. Elsewhere, in a study of As cycling in lakes, Aurillo et al. (1994) reported that in late summer and fall, oxic surface waters from Upper Mystic Lake had an As(III)/total As ratio of approximately 0.5 (Aurillo et al., 1994). Both of these studies illustrate As(III) levels far from thermodynamic equilibrium.

If the high levels of dissolved As in reducing groundwaters are caused by reductive dissolution of Fe oxides, one would expect a strong correlation between dissolved Fe and As levels in groundwater. While most high-As waters under these conditions have relatively high Fe contents, there is no clear correlation between the two elements in groundwater (Kinniburgh and Smedley, 2001; Nickson et al., 2000; Ahmed et al., 2004). Furthermore, dissolved Fe levels, while high, are not as high as would be expected from complete dissolution of Fe(III) phases. Some invoke reoxidation of Fe(II) to account for the missing Fe (Zheng et al., 2004) while others point to the possibility that solid Fe(II) phases such as ferrous sulfide, pyrite, siderite, vivianite, or green rusts may sequester significant amounts of Fe(II) in reducing groundwaters (Horneman et al., 2004). Geochemical modeling has shown that shallow groundwater in Bangladesh is often saturated with respect to calcite, siderite, and vivianite (Nickson et al., 2000; Ahmed et al., 2004), and a number of ferrous solids including pyrite (Nickson et al., 2000), ferrous phosphate, and ferrous silicate (Harvey et al., 2002) have been identified in core sediments from As-affected areas in Bangladesh.

In principle, solid phases could also limit arsenate solubility. Scorodite (Fe(III)As(V)O4·2H2O) is a well-known mineral in As-bearing ore deposits, but is only formed under strongly acidic conditions (Dove and Rimstidt, 1985; Rochette et al., 1998). Arsenates of various divalent metals (Ba, Ca, Cu, Mg, Mn, Ni, and Zn) exist as analogs of vivianite [Fe(II)3(PO4)2·8H2O] but most are too soluble to limit arsenate concentrations in natural systems (Essington, 1988; Voigt et al., 1996; Bothe and Brown, 1999a, 1999b). Isomorphic substitution of divalent metals, or of PO4 for AsO4, occurs readily in these minerals and solid solutions among the vivianite analogs can be expected (Frost et al., 2003). Mineral identification is complicated by the fact that x-ray diffraction patterns are identical for many of the vivianite analogs (Frost et al., 2003).

Symplesite and parasymplesite are arsenate analogs of vivianite with the formula Fe(II)3(As(V)O4)2·8H2O from triclinic and monoclinic crystal systems, respectively (Mori and Ito, 1950; Roberts et al., 1990). Relatively few studies have examined the properties of these minerals. A recent study explored the Raman and infrared spectra of various arsenate minerals including symplesite (Frost et al., 2003), and a few studies have reported thermodynamic data for ferrous arsenate species (Hess and Blanchar, 1976; Khoe et al., 1991; Sadiq, 1997; Gonzalez and Monhemius, 1998). In the metallurgical literature, Khoe et al. (1991) have reported a solubility product for ferrous arsenate of 4 ± 1 x 10–41. In laboratory microcosms, bacterial reduction of Fe(III) in scorodite reportedly produces an acidic ferrous arsenate phase (Fe(II)HAs(V)O4·xH2O) (Cummings et al., 1999), and traces of symplesite along with other unknown ferrous arsenate phases (Papassiopi et al., 2003). But to our knowledge, precipitation of authigenic symplesite has not been documented in natural systems.

In this study we determined a solubility constant for symplesite based on controlled laboratory experimental data, and used geochemical modeling to suggest that this phase might occur in natural systems, as well as under more extreme conditions.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Chemicals
All chemicals were of certified ACS grade, and were used without further purification. Stock solutions of Fe(II) and As(V) were made by dissolving Fe(II)(NH4)2(SO4)2 and NaH2As(V)O4 in deionized water. Sodium nitrate was used for ionic strength control. All solutions were prepared inside an anaerobic glove box (Coy Laboratory Products, Grass Lake, MI, 2% H2 atmosphere) using deionized water that had been deaerated by boiling under a N2 atmosphere outside the glove box for 30 min. Strong acid (HNO3) and base (KOH) were used for pH adjustment.

Experimental and Analytical Methods
An acidic solution containing 1.00 mM Fe(II) and 0.50 mM As(V) in dilute HNO3 (pH ~4.5) was prepared from stock solutions of Fe(II)(NH4)2(SO4)2·6H2O and NaH2AsO4 in a background electrolyte of 100 mM NaNO3. The solution was continuously mixed with a magnetic stirrer in a 500-mL reactor inside the anaerobic chamber. Strong base (100 mM KOH) was added dropwise, resulting in the steady increase in pH (measured with an Orion probe, Thermo Electron Corp., Waltham, MA) until the onset of precipitation at approximately pH 7.0, which was marked by a drop in pH and the appearance of a white precipitate. Precipitation was allowed to proceed for 1 h, after which 10-mL aliquots were transferred in duplicate to 15-mL polypropylene centrifuge tubes (Falcon, BD Biosciences, Bedford, MA). Variable amounts of KOH were then added to the precipitation reactor, and pH was recorded after each addition. After an increase of approximately 0.2 to 0.3 pH units, duplicate aliquots were transferred to centrifuge tubes. The tubes were placed on an end-over-end shaker in the glove box, and equilibrated for 6 to 8 h. Temperature in the anaerobic chamber ranged from 26.2 to 27.0°C during the experiment.

After equilibration, samples were filtered through prerinsed 0.2-µm nylon syringe filters for analysis of Fe(II) and As, and the final pH was recorded. Iron (II) was measured in the glove box using a modified Ferrozine method (Stookey, 1970), while As samples were acidified and measured within 2 d using graphite furnace atomic absorption spectroscopy (Perkin Elmer 5100, Wellesley, MA, with Zeeman correction).

In a separate experiment, a larger quantity of precipitate was prepared by mixing 5 mM Fe(II) with 3 mM As(V), and raising the pH with KOH to ~7.5, which calculations indicated to be below the saturation point for Fe(OH)2. This experiment was done with no inert electrolyte to simplify precipitate identification. After stirring magnetically for 1 h, the resulting precipitate was allowed to settle overnight in the glove box. A small amount of precipitate was dissolved in dilute HNO3 to allow measurement of the Fe(II)/As(V) ratio. Approximately 20 mL of slurry was collected, dewatered using a nylon syringe filter (0.2 µm), and dried in a dessication chamber under anoxic conditions. The crystal structure of the precipitate was analyzed using x-ray powder diffractometry (Rigaku Multiflex. Rigaku Corp., The Woodlands, TX) with a Cu-K{alpha} radiation source ({lambda} = 1.5418 Å, 40-mA current). The precipitate was mounted on a glass slide and immediately covered with a drop of glycerol to prevent oxidation. The scan range was from 5 to 60° (2{theta}) with a speed of 1° s–1.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
In all samples, a white precipitate was clearly visible, which was subsequently identified by x-ray diffraction analysis as ferrous arsenate [symplesite, Fe(II)3(As(V)O4)2]. Iron (II) and As(V) concentrations dropped dramatically during equilibration, with the greatest reductions occurring at higher pH (see Fig. 1 ). The solid lines shown are model predictions (see below). The pH also decreased by as much as 1.4 units. Arsenic (V) precipitated in accordance with

Formula 1[1]


Figure 1
View larger version (14K):
[in this window]
[in a new window]

  		Fig. 1. Precipitation of ferrous arsenate. Initial conditions: 1000 uM Fe(II), 500 uM As(V), 100 mM NaNO3. Solid lines represent model predictions.

 
This precipitation reaction liberates protons from the dissolved As species, causing the pH to fall. At pH 8.5, dissolved As levels were reduced by >99.9%, from an initial concentration of 500 µM (37.5 mg L–1) to 0.32 µM (24 µg L–1).

Above pH 8.5, precipitation of Fe(OH)2 was observed (see Eq. [2]), which resulted in a further drop in dissolved Fe(II) levels and production of a bluish-green precipitate. Equilibrium As(V) levels began to rise above pH 8.5, as Fe(II) was increasingly sequestered in ferrous hydroxide.

Formula 2[2]

Figure 1 shows good agreement between replicate samples, although the pH in some cases varied by as much as 0.2 units. This variation is not surprising given the lack of a pH buffer in the system, apart from the As(V) system itself.

The stoichiometry of the precipitate was assessed by measuring the amounts of dissolved Fe(II) and As(V) removed from solution. Linear regression (Fig. 2 ) shows that for every mole of Fe(II) removed, 0.666 mole of As(V) was lost, consistent with Eq. [1]. This stoichiometry was confirmed by precipitate dissolution and elemental analysis of the resulting solution, which yielded an Fe(II)/As(V) ratio of 0.62. X-ray diffraction patterns showed clear peaks (d = 6.602, 2.329, 7.823, 3.206, and 1.665 Å, in order of decreasing intensity) that matched the main peaks of the reference spectrum of symplesite.


Figure 2
View larger version (12K):
[in this window]
[in a new window]

  		Fig. 2. Ferrous arsenate stoichiometry. Open symbols represent conditions of oversaturation with respect to Fe(OH)2(s).

 
The program PHREEQC-2 (Parkhurst and Appelo, 1999) was used to determine the ion activity product (IAP) for samples taken from the precipitation experiment. Activity coefficients were calculated using the Davies equation. The relevant equilibrium constants used in the calculations are shown in Table 1. Reactions involving ferrous sulfate species are included since the Fe(II) stock was prepared from Fe(II)(NH4)2(SO4)2, and significant formation of ferrous sulfate species could affect formation of ferrous arsenate complexes. For modeling purposes, redox transformations of Fe(II) and As(V) were blocked.


View this table:
[in this window]
[in a new window]

  		Table 1. Equilibrium constants (log K) for dissolved species used in geochemical modeling.

 
The IAPs calculated for symplesite and Fe(II)(OH)2 are shown in Fig. 3 . The eight samples with the lowest initial pH values (6.40–7.10 at the beginning of equilibration) have significantly higher IAPs for symplesite than other samples; we consider that these samples remained oversaturated because of relatively slow precipitation kinetics under these conditions. At higher pH, precipitation was more rapid and an equilibrium value seemed to be reached within the equilibration period of 6 to 8 h. Averaging the 16 data points that were considered to be at equilibrium with ferrous arsenate (closed symbols in Fig. 3), we calculated a solubility product, pKso, of 33.25 (±2{sigma} = 0.46) for symplesite. Figure 3 also shows Fe(II)(OH)2 precipitation occurring at approximately pH 8.5. The average IAP above this point was 10–15.2, which is in excellent agreement with the reported pKso of 15.1 for Fe(II)(OH)2 (Stumm and Morgan, 1996). This serves as a validation of the quality of the Fe(II) and pH measurements, as well as the modeling analysis.


Figure 3
View larger version (12K):
[in this window]
[in a new window]

  		Fig. 3. Log ion activity products (IAP) for ferrous arsenate (squares, left axis) and ferrous hydroxide (triangles, right axis). Dashed and solid lines represent equilibrium solubility products for ferrous arsenate and ferrous hydroxide, respectively. Filled symbols denote samples considered to be at equilibrium with ferrous arsenate or ferrous hydroxide phases.

 
In Fig. 1, the solid lines show model predictions of dissolved Fe(II) and As(V) given the initial conditions of 1000 uM Fe(II) and 500 uM As(V) and the experimentally determined solubility product. The model matches observed concentrations well, and correctly predicts the increase in As(V) and drop in Fe(II) as Fe(II)(OH)2 precipitates above pH 8.5.

The solubility constant we report is several orders of magnitude larger than that reported by Khoe et al. (1991), of approximately 10–41.2. The reasons for this large difference are not clear. The difference could be explained by nonequilibrium conditions in our experiments, or by the possible passage of colloidal symplesite through the 0.2-µm nylon filters used in our system. We do not believe either of these potential explanations to be true. It is also possible that different ferrous arsenate solids, having the same Fe(II)/As(V) ratio, are formed under acidic and neutral conditions. The Khoe et al. (1991) experiments were based on precisely measuring the pH at the onset of precipitation (near pH 2) while our experiments measured soluble Fe(II) and As(V) in the presence of precipitate from pH 6 to 9. Further work is required to resolve the discrepancy between these reported solubility constants; ideally the solubility constant should be calculated from both oversaturated and undersaturated conditions.

It was hypothesized that a linear free energy relationship might exist between divalent metal precipitates of phosphate and arsenate of the form Me(II)3(YO4)2·8(H2O), where Y represents As(V) or P, since the two oxyanions are structurally similar. Published constants were found for many solids, but in only four cases were both arsenate and phosphate solubility products available (Table 2). No clear relationship is evident from these four sets of data, but in the case of Cd(II) and Fe(II), the solubility products for both oxyanions are very close.


View this table:
[in this window]
[in a new window]

  		Table 2. Thermodynamic solubility products (pKso) for Me(II)3(YO4)2·8(H2O).

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Bangladesh
Using the PHREEQC-2 model and its associated database, modified to include As compounds, equilibrium speciation can be computed for complex mixtures. Figure 4 is a simulated stability diagram for symplesite, indicating total dissolved As(V) concentrations in equilibrium with symplesite for various Fe(II) concentrations. The calculations were made for two ionic strengths representative of those in Bangladesh groundwaters. The figure shows that drinking water meeting the World Health Organization guideline value for As of 10 µg L–1 (World Health Organization, 2004) is unlikely to be saturated with respect to symplesite unless the Fe concentrations are unrealistically high. In reduced groundwaters, however, such as those found in Bangladesh, it is not uncommon to find water containing 100 to 1000 µg L–1 As and significant amounts of dissolved Fe (10% of samples in the National Hydrochemical Survey exceeded 10 mg L–1) at circumneutral pH (Kinniburgh and Smedley, 2001). Symplesite might represent a significant sink for arsenate and Fe(II) in these systems.


Figure 4
View larger version (16K):
[in this window]
[in a new window]

  		Fig. 4. Stability field for symplesite. Isopleths represent total arsenate concentration in micrograms per liter. Solid lines were calculated with an ionic strength of 0.01, while dotted lines have an ionic strength of 0.05. Areas to the right of the isopleths are oversaturated with respect to symplesite.

 
In Bangladesh groundwaters, dissolved As is predominantly present as As(III) (Bhattacharya et al., 2002), but As(V) is dominant in sediments near the water table, where highly enriched bands of Fe(III) are found (Breit et al., 2004). These are thought to be caused by repeated fluctuations of the water table: when submerged, As and Fe(II) adsorb to ferric oxide surfaces, which may also partially dissolve due to dissimilatory reduction by bacteria. Bacterial reduction of Fe(III)(OH)3 would liberate As to solution, without necessarily reducing As(V) to As(III) (Cummings et al., 1999). When water levels drop again, the adsorbed Fe(II) is oxidized, creating fresh Fe(III)(OH)3 adsorption sites. Adsorbed Fe(II) is more easily oxidized than the free species (Stumm and Lee, 1961), and reactive intermediary species produced during the oxygenation of Fe(II) can oxidize As(III) to produce As(V) (Johnston, unpublished data, 2005; Roberts et al., 2004).

In Bangladesh, a number of researchers have conducted geochemical surveys measuring As and other parameters in groundwater. One of the most extensive is the National Hydrochemical Survey (NHS) (Kinniburgh and Smedley, 2001), which analyzed 271 wells in three special study areas for a broad suite of parameters. Samples were filtered at collection through 0.2-µm acetate filters. Iron was not speciated, but is assumed to be completely present as ferrous species, since the solubility of Fe(III) at neutral pH is negligible. Other researchers have reported very high Fe(II)/total Fe ratios in Bangladesh groundwater (Horneman et al., 2004).

Concentrations of major ions (Ca2+, Cl–, F–, Fe2+, HCO3–, K+, Mg2+, Mn2+, Na+, NH4+, NO3–, PO43–, H3SiO3–, and SO4–2) from this data set were used along with As(V), As(III), pH, and temperature to model geochemical speciation. Arsenic and Fe redox transformations were blocked by decoupling the species in PHREEQC-2. Only the 256 records where pH data were available were used in the analysis. Solubility products for solid phases used in the model are listed in Table 3.


View this table:
[in this window]
[in a new window]

  		Table 3. Solubility products (log K) used for geochemical modeling.

 
As a check on the input data, a charge balance was performed for each sample. A consistent positive bias was noted, with a median value of 8.7%. This suggests that cations are systematically overestimated, or that anions are missing or underestimated in the NHS database. The bias is not gross, however: 97% of samples had an error of <20%.

Saturation indices (SIs) for various solid phases were calculated for each sample in the NHS dataset as the log of the ratio of the IAP to the equilibrium constant. A SI of zero represents equilibrium conditions, while negative and positive SIs represent conditions of under- and oversaturation, respectively. To graphically present the SIs for each mineral, the SIs were sorted and plotted against the percentage of the total sample having values greater than or equal to that SI value. The resulting curve is called a cumulative frequency distribution (CFD). The CFDs of SIs for the different waters with respect to the various minerals are plotted in Fig. 5 .


Figure 5
View larger version (23K):
[in this window]
[in a new window]

  		Fig. 5. Cumulative frequency distributions of saturation indices for various minerals. Geochemical data from Kinniburgh and Smedley (2001).

 
More than half of the samples in the database (57%) were found to be oversaturated with respect to calcite, and approximately two-thirds were oversaturated with respect to siderite and rhodochrosite. The steep slope of the CFDs for these three carbonate minerals just above the equilibrium solubility (SI = 0) is consistent with the presence of these minerals in the subsurface and the control they exert on the chemical composition of the respective waters. Precipitation of these carbonate minerals may account for the low concentrations of Fe(II) and Mn(II) in a number of these waters. In contrast, about half of the waters are oversaturated with respect to hydroxyapatite [Ca5(PO4)3OH], but the CFD showed no change in slope above the saturation point. This may reflect a kinetic limit on precipitation for this mineral. All of the waters were highly undersaturated (SI < –2) with respect to Fe(OH)2 and manganous arsenate (not shown).

About 4% of the samples were found to be oversaturated with respect to symplesite, and 5% with respect to vivianite (Ii the Khoe et al. [1991] solubility constant is used, >90% of the samples were oversaturated with respect to symplesite).However, only about a third of the samples saturated with respect to symplesite were also saturated with respect to vivianite. The chemical composition of the samples oversaturated with respect to symplesite is summarized in Table 4. These samples are typical of circumneutral reduced groundwater in Bangladesh, where O2, NO3, and SO4 have been reduced by microbial consumption of organic matter, resulting in high alkalinity and dissolved Fe. Even under these conditions, significant amounts of dissolved As can be found in the oxidized As(v) state.


View this table:
[in this window]
[in a new window]

  		Table 4. Geochemistry of Bangladesh groundwater samples oversaturated with respect to symplesite (n = 12).

 
Ferrous phosphate phases have been identified in sediments from As-contaminated aquifers in Bangladesh (Harvey et al., 2002). Since the CFD for symplesite saturation is very similar to that of vivianite and the solubility constants are similar (see Table 3), it seems likely that ferrous arsenate phases may also exist in the same aquifers. Furthermore, since a solid solution is expected between vivianite and symplesite, arsenate may be sequestered in ferrous phosphate phases even when the water is undersaturated with respect to symplesite.

If arsenate were present in phosphate phases, it should be released during acid dissolution of the solid phase. Several studies of sequential dissolution experiments have been conducted on sediments from As-affected aquifers in Bangladesh, though acid-soluble As is generally interpreted as being carbonate bound. Most find little to moderate amounts of As in the acid-soluble fraction (e.g., Harvey et al., 2002; Akai et al., 2004), suggesting that phosphate minerals would be a minor rather than a major sink for As in these sediments.

Other Environments
Apart from reduced groundwater, there are a number of extreme environments in which symplesite could be stable. For example, the high pH (9.8) and elevated concentration of As(V) (200 uM) in Mono Lake (Oremland et al., 2004) would limit dissolved Fe(II) to 8 µg L–1. Symplesite has been identified in industrial waste sites where As-rich smelter slag has been used as landfill (USEPA, 1998). Finally, ion exchange resins are used to remove arsenate from drinking water. Resin regenerant is highly enriched in arsenate, and typically must be processed before disposal. Addition of ferrous salts to precipitate symplesite could present an attractive alternative to conventional coprecipitation and adsorption of arsenate using ferric or aluminum coagulants, which generates large volumes of waste sludge.

In arid alkaline groundwaters, As is typically present as As(V). For example, drinking water in the U.S. city of Fallon, NV, is alkaline (pH 9.1) and contains from 70 to 120 µg L–1 As, predominantly as As(V) (Welch et al., 2003). Under these conditions, modeling suggests that symplesite would limit As(V) concentration to <10 µg L–1 on addition of approximately 10 to 15 mg L–1 Fe(II). Ferrous iron would quickly react with the low levels of dissolved O2 in these waters (~1 mg L–1), producing small amounts of hydrous ferric oxide, which would also remove As(V) through adsorption and coprecipitation. While conventional treatment would require pH reduction to favor adsorption of the arsenate anions onto hydrous ferric oxide, the high pH of these waters would actually improve As(V) removal through precipitation of ferrous arsenate.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
We have presented a new experimentally derived solubility constant for the ferrous arsenate mineral symplesite. Geochemical modeling using the new constant suggests that some reduced groundwaters of Bangladesh are oversaturated with respect to symplesite. Precipitation of authigenic symplesite could be a significant sink for both As(V) and Fe(II) under these conditions. Our modeling results confirm the work of others, indicating that Bangladesh groundwaters are frequently oversaturated with respect to siderite, vivianite, and rhodochrosite and that these minerals could represent important sinks for Fe(II) and Mn(II); however, previous studies have not considered the possibility of precipitation of ferrous arsenate phases such as symplesite. Precipitation of ferrous arsenate could be an effective control measure to reduce arsenate levels in As(V)-rich alkaline waters used to produce drinking water, or in specialized applications such as management of ion exchange regenerant brines.


   ACKNOWLEDGMENTS
 
This work was made possible by support from a Royster fellowship at the University of North Carolina and a National Science Foundation Graduate Research Fellowship. Thanks to the British Geological Survey for making the National Hydrochemical Survey data freely available, and to Dr. Peter White (UNC Chemistry Department) for assistance in the x-ray diffraction analyses.

Received for publication January 16, 2006.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Johnston, R. B.
Right arrow Articles by Singer, P. C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Johnston, R. B.
Right arrow Articles by Singer, P. C.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Johnston, R. B.
Right arrow Articles by Singer, P. C.
Related Collections
Right arrow Soil Mineralogy
Right arrow Toxic Trace Metals
Right arrow Geochemical Processes

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome