| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Geological and Environmental Sciences, Stanford University, Stanford, CA 94305-2115
b Department of Geosciences, San Francisco State University, San Francisco, CA 94132-4163
* Corresponding author (fendorf{at}stanford.edu)
Received for publication May 22, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: PBET, physiologically based extraction test
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Arsenic is a redox active element that generally persists in either the +3 or +5 oxidation states within soils. Arsenate 
is often considered less toxic than its counterpart arsenite 
(Ferguson and Gavis, 1972; Cullen and Reimer, 1989; Smith et al., 1998). Furthermore, the mobility and availability of arsenic will be controlled by reactions with soil solids. Arsenate generally adsorbs strongly to a host of solids (Oscarson et al., 1983; Pierce and Moore, 1982; Goldberg and Glaubig, 1988; Manning and Goldberg, 1997) while arsenite is more discriminate and tends to bind strongly only to ferric (hydr)oxides (Manning et al., 1998; Dixt and Hering, 2003). Moreover, the availability of arsenic to organisms is controlled by the stability of the solid-phase complex (Freeman et al., 1993, 1995; Rodriguez et al., 1999; Ruby et al., 1996), and iron oxide content along with pH appear to be principal soil factors controlling arsenate bioaccessibility (Yang et al., 2002).
Similar to arsenic, chromium is a redox active soil contaminant but one that has dramatic alterations in toxicity and mobility with changes in oxidation state. Trivalent Cr is rather benign to most plants and animals and binds strongly to soil solids (Fendorf, 1995). In contrast, hexavalent Cr is toxic to living cells, being a Class A human carcinogen (Kargacin et al., 1993); because of its anionic nature and inability to form strong chemical complexes with most soil materials, chromate is also highly mobile within the surface environment (Fendorf, 1995; Ball and Nordstrom, 1998). Recent work has demonstrated that Cr bioaccessibility is a function of soil type and retention time (Stewart et al., 2003).
Lead, in contrast to arsenic and chromium, is a redox-stable divalent cation that has a high affinity for numerous soil materials (Bargar et al., 1997), the specific complexes of which impart an important control on its dispersal (transport) and bioavailability (Ruby et al., 1996). And although regulatory agencies assume a fixed bioavailability of approximately 30% for Pb, it appears that Pb bioavailability will vary depending on its structural state within soils (Ruby et al., 1992, 1993).
While soils tend to bind the various species of As, Cr, and Pb to some degree, the specific mechanisms, and thus both the degree and strength of retention, may vary greatly and may change temporally. The binding mechanism will therefore influence the extent to which contaminant transport is attenuated and the availability of contaminants for uptake by biological organisms. With regard to the health hazards imposed by a contaminant, a number of steps or processes actually define the bioavailability of a contaminant to a given organism. But, in general, an initial step in the bioavailability process is the release of a contaminant from the solid phase into either the aqueous or gas phase, making this process dependent on the binding mode of the contaminant within soils (Ruby et al., 1992; Davis et al., 1993). Bioaccessibility is a term encompassing the release step in the bioavailability process and is defined as the fraction of a contaminant available for absorption in the gastrointestinal system of an organism (Ruby et al., 1996). Because bioaccessibility addresses the release of contaminants from the solid phase, an appreciation of retention mechanisms within soils is thus crucial for evaluating contaminant risk.
A number of mechanisms may be operational in trace element retention, and the mode of retention may change with (incubation) time. Both cations and anions will typically adsorb rapidly to soil surfaces forming outer-sphere (electrostatic or physical) complexes. Following the initial physi-sorption, a secondary (slower) step will follow leading to the development of an inner-sphere (chemical) complex (Hayes and Leckie, 1986; Zhang and Sparks, 1989, 1990a, 1990b). Further aging may lead to more extensive changes in the surface phase (for example, see Aharoni and Sparks, 1991). Surface diffusion within micropores can result in a virtually absorbed phase (Ainsworth et al., 1994; Axe and Trivedi, 2002); it may also give rise to nucleation and the development of a surface precipitate (for example, see Ford and Sparks, 2000). Finally, deposition of organic or inorganic material may occlude the contaminant (again the development of an absorbed phase).
At high(er) concentrations, trace elements may form either surface or discrete precipitates. Kinetic factors usually govern the phase that forms over a short period of time, which is primarily dictated by the activation energy or the energy barrier of a reaction. Generally, large well-crystallized particles are less soluble but have a higher activation energy. Consequently, amorphous particles are frequently found in soils and sediments due to their meta-stable conditions. Given sufficient time, these amorphous phases will transform into more crystalline solids (ripening), which are thermodynamically more stable (i.e., they have a lower solubility) (Ainsworth et al., 1994). Additionally, existing surfaces often provide a catalytic role in precipitation and lead to surface (or heterogeneous) precipitates.
Recent evidence has revealed the potential for mixed metal phases to form as precipitates on mineral surfaces, providing the resulting phase has a lower solubility than the parent substrate. Association of transition metals with unstable aluminosilicate clay minerals, such as pyrophyllite, may lead to the release of Al from the clay and incorporation of the transition ion in a solid having a hydrotalcite structure; such phases have been noted recently for Co (Thompson et al., 1999), Ni (Scheidegger et al., 1996), and Zn (Ford and Sparks, 2000). Upon aging, silicon appears to be reincorporated into the precipitate leading to the neoformation of a transition metal-bearing clay mineral (Ford and Sparks, 2000).
A universal theme of the aging processes described above is a decrease in the potential for contaminant release. That is, the availability of the contaminant should diminish with soil incubation time (assuming dramatic changes do not ensue that may destabilize the substrate). In this study, we investigated the changes in arsenic, chromium, and lead on exposure and incubation to soils obtained from Oak Ridge, TN. We evaluated the bioaccessibility of the three contaminants for human receptors using a physiologically based extraction test (PBET) coupled with selective sequential extractions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Experiments were initiated by placing quintuplicate samples of 20 g of soil into a 250-mL Nalgene reaction vessel (Nalge Nunc International, Rochester, NY). We spiked each soil with 200 mL of 5 mg L–1 As(III) as NaAsO2, Cr(III) as CrCl3, and Pb(II) as PbCl2 in a 0.001 M CaCl2 matrix at pH 3. In addition, two unspiked control subsamples per soil horizon were prepared. The soil slurries were placed on an orbital shaker and allowed to react for 10 h at which time supernatants were collected and analyzed for As, Cr, and Pb. A total of four spikes transpired, which resulted in final contaminant (As, Cr, and Pb) loadings of approximately 200 mg kg–1. Reaction vessels were maintained at field capacity (approximately 33% moisture content) for the duration of the experiment. To assure homogeneity at the time of sampling, each reaction vessel was homogenized.
Analytical Procedures
Trace elements were measured in the aqueous phase using inductively coupled plasma–optical emission spectrophotometry (ICP–OES) (IRIS model; Thermo Jarrell Ash, Franklin, MA) with a 10% accuracy range and quality control was checked every 15 samples. Detection limits were defined by 3 x the standard deviation of 7 blanks. Detection limits were 0.03 mg L–1 for As, 0.03 mg L–1 for Cr, and 0.04 mg L–1 for Pb. All reaction-ware used in the experiment was rinsed in 0.5 M HCl before use.
Physiologically Based Extraction Tests
The PBET is an in vitro leaching procedure that is used to determine metal concentrations that could be absorbed through digestion in the human upper GI tract (Ruby et al., 1992, 1993, 1996). The test involves simulating conditions in the stomach and small intestine, using realistic values for soil-to-solution ratios, stomach mixing, stomach emptying rates, and small intestine pH and chemistry (Ruby et al., 1993). The PBET is a screening-level test designed around the GI tract parameters of a 2- to 3-yr-old child, considered to be at greatest risk to metal exposure from accidental soil ingestion (Ruby et al., 1996). The intent of the PBET is to develop data that correlate well with measures of As and Pb bioavailability in animal models (Ruby et al., 1996). A strong, positive correlation between PBET and in vivo Pb values was demonstrated for the method used here (Ruby et al., 1996) and a comparable extraction, differing only in having a pH of 1.8 as opposed to 1.5, illustrated a strong correlation of PBET and in vitro gastrointestinal As levels (Rodriguez et al., 1999). The PBET analyses were conducted on soils at 0, 14, 30, 60, and 400 d after incubation.
The PBET procedure was modified after Ruby et al. (1996) and used 0.5 g of soil reacted with 50 mL of 1 M glycine at pH 3 for 1 h in a 35.6°C (96°F) water bath; a pH value of 3 was used to simulate less aggressive digestive conditions. The centrifuge tubes were affixed and submerged on a rotating horizontal armature (12 rpm) to simulate stomach mixing. The reaction vessels for each sample were centrifuged, and the supernatant collected and analyzed by ICP–OES.
Chemical Extractions
Selective sequential extractions (SSEs), using a series of increasingly aggressive chemical reagents, allow for general elemental partitioning patterns to be approximated (for example, see reviews by Pickering, 1981; Chao, 1984; and Martin et al., 1987). Soil extractions are operationally defined and refer to a reactive fraction rather than a chemically or mineralogically specific target (Tessier and Cambell, 1991; Kim and Fergusson, 1991).
Chemical extractions were initiated on soils using approximately 1 g (dry weight equivalent) of homogenized, composite samples. Following each extraction, the soil residue was washed with double-deionized water. Supernatants were filtered through a 0.2-µm membrane filter and acidified with trace element–grade HCl before ICP–OES analysis. Extractions were performed in quintuplicate along with a standard reference material (SRM 2709; National Institute of Standards and Technology, Gaithersburg, MD) and appropriate blanks. We used common extractions and a series that seeks to define three reactive pools: exchangeable, ligand dissolvable, and acid-extractable.
The first extraction involved using 1 M MgSO4 (pH 7) and shaking for 1 h; it is designed to release water-soluble and exchangeable As, Cr, and Pb. Next, the soil was shaken for 4 h with 20 mL of 0.2 M ammonium oxalate (pH 3) in the dark (AOD); metal (hydr)oxides with less stability (principally short-range order materials) are particularly susceptible to AOD extraction (Schwertmann, 1973; Jackson et al., 1986). The third extraction, 12 M HCl, shaken for 12 h, is designed to release constituents from more recalcitrant phases such as crystalline metal (hydr)oxides (Huerta-Diaz and Morse, 1992).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
Reactive Partitioning
To evaluate the partitioning of the trace elements and their potential relation to temporal changes in PBET levels, we used a set of common extractions on the soils. The extractions are not meant to provide definitive information on the solid phases but rather to (i) give information on the reactive nature of the contaminants (defined as reactive pools that are operationally based) and (ii) allow a relation between common procedures and bioaccessibility to be developed. The easily exchangeable fraction was estimated using an extractant of MgSO4; oxalate was employed as a common, natural chelate that may promote contaminant release. Finally, the fraction released by strong acid was assessed using HCl.
A discernable trend in the oxalate and acidic decomposition pools with increased incubation time is not apparent for either the A or B horizons (Table 3). Moreover, a large fraction of the trace elements remains even after treatment with these three reactants, constituting a residual pool that is resistant to release by ligand- or proton-promoted dissolution over the duration of the procedure. In contrast with oxalate- and HCl-extractable pools, the exchangeable fraction (MgSO4–extractable), although a small pool, does decrease with increased incubation time (Fig. 2) . In fact, the change in the exchangeable pool is very substantial and appears exponential with time, similar to that observed for the PBET levels.
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Alterations in bioaccessibility with time are not, however, restricted to simply a shift from outer- to inner-sphere complexes. Both the diffusion of ions within solids having internal porosity, such as hydrated, short-range order iron oxides (Axe and Trivedi, 2002), and the development of coprecipitates (e.g., Ford et al., 1999) with constituents of the substrate enhances the stability of contaminants within the solid phase, thus restricting their bioaccessibility. Precipitates inclusive of the contaminant may also become more crystalline with time and thus have a diminished solubility and bioavailability (Ruby et al., 1993). Such processes would not be expected to affect ion retention within the first few weeks of incubation but may account for an increasing resistance to release with incubation times of >30 d. A final process that could potentially alter the susceptibility of contaminants to extraction is occlusion by organic or inorganic materials. We suspected that under the unsaturated conditions of this study in combination with the low levels of dissolved organic carbon, that such a factor would not be appreciable on the time scale of these incubations. Moreover, the higher levels of organic matter within the A horizon would lead one to expect occlusion to have a greater effect than in the B horizon. However, PBET levels do not decline beyond 14 d of incubation and thus are consistent with the expectation that occlusion is not a factor in this study.
Examining differences in soil chemical properties is instructive for elucidating possible retention mechanisms and temporal trends in bioaccessibility. The dominant differences between the soil horizons are (i) higher pH, (ii) greater organic matter content, (iii) lower clay (size faction) content, and (iv) lower Fe (and Al) oxide content within the A horizon relative to the B horizon (Table 1), factors consistent with soil development and diagnostic for horizon designation. Cationic metals such as Cr(III) and Pb(II) would be expected to partition strongly on organic matter, forming either electrostatic or chemical bonds with surface functional groups, and to have greater retention at higher pH (Schindler and Stumm, 1987). Iron oxides also have a demonstrated affinity for both Cr(III) and Pb(II), with higher pH again favoring sorption (Dzomback and Morel, 1990). In contrast, arsenic does not bind appreciably to organic matter; and of the two potential oxidation states, arsenate favors adsorption at lower pH while arsenite has an adsorption maximum on iron oxides near neutrality (Manning et al., 1998; Dixt and Hering, 2003). Given the strong adsorption of both arsenate and arsenite on iron oxides, arsenic retention should be extensive across the pH range of these horizons given the loading levels promoted here.
Arsenic behaves in a manner consistent with chemical considerations of their reactivity and the soil properties. Given the levels of ferric iron within both the A and B horizons of the Melton Valley soil (Table 1), one would expect appreciable chemical retention and thus resistance to PBET-induced release. There is, in fact, a sizeable limitation in bioaccessibility and the magnitude is inversely proportional to the ferric oxide content; PBET levels of the B horizon are about half those of the A horizon and contain about twice the iron oxide content.
In contrast to arsenic, the retention of Cr(III) and Pb should be governed dominantly by pH. The sorption edge for hydrolyzable cations, such as Cr(III) and Pb(II), is directly proportional to their hydrolysis constants with the specific pH based on the sorbent. Thus, given the high pH of the A horizon (nearly 2.5 units higher) relative to the B horizon, in combination with appreciable organic matter and ferric oxide content, one would expect greater chemical retention within the upper horizon. This is indeed observed for Cr(III): the PBET levels are significantly lower in the A horizon than the B. Chemical binding, and thus diminished bioaccessibility, for Pb(II) within the B horizon is consistent with chemical considerations. Lead has a very high chemical affinity for numerous soil solids (organic matter, iron oxides, etc.) and generally demonstrates even greater retention than Cr(III) (Dzomback and Morel, 1990). Lead, however, is anomalous in its behavior within the A horizon (it has the greatest release with the PBET treatment). While we do not have definitive evidence for the behavior of lead within the A horizon, there are at least two possibilities to explain the increase in bioaccessibility with time. First, it is possible that organic matter is mineralizing during the incubation and liberating Pb. We find this possibility unlikely given the degree of carbon mineralization one would expect in these experiments coupled with the binding affinity of Pb for residual OM. A second possibility is the formation of lead carbonate species. Given a pH of 6.9, cerussite (PbCO3) and hydrocerrusite [Pb3(OH)2(CO3)2] are oversaturated (saturation indices of 0.12 and 2.43, respectively) within the A horizon assuming equilibrium with exchangeable and water-soluble Pb at atmospheric carbon dioxide levels. It is therefore possible that such phases, and in particular hydrocerrusite, are developing with an increased incubation period and are readily dissolved in the acidic solution of the PBET treatment. Furthermore, even the slightly acidic solutions of the oxalate extraction (pH 3) would lead to dissolution of these phases, as would the strong acid treatment (12 M HCl). What remains troubling, however, is that neither the oxalate nor the HCl (nor the sum of these two) levels increase with incubation time. We would expect an acid oxalate and HCl extract to correlate well, and positively, with the PBET levels if a lead carbonate phase was controlling extractability. A positive correlation of PBET with acid-extractable levels was in fact noted for alkaline soils contaminated with Pb from smelting activity (Basta and Gradwohl, 2000).
The decrease in bioaccessibility of As and Cr is also more rapid in the A horizon than within the B. We speculate that the finer texture of the lower horizon may decrease chemical reaction rates through diffusion limitations. Independent of the mechanism by which it results, one clearly notes that a steady state in bioaccessibility is achieved more rapidly within the A than the B horizon. Additionally, only a limited fraction of the contaminants is generally released by the PBET treatment; Pb within the A horizon is an exception in which nearly all of the added contaminant is released by the bioaccessibility treatment. Therefore, rather than using single default values for a bioaccessible fraction within bioavailability models, it is clear from this study that short-term (100 d of aging) temporal dependence must be considered along with, and more importantly, specific soil properties, both chemical and physical. Our results also illustrate the importance of incubation time in laboratory studies being conducted on contaminant-spiked soils or sediments; short incubation times may not represent the reactivity (either in regards to transport or bioavailability) that one would observed with increased incubation time. Decreases in bioaccessible As and Cr, but not Pb, were correlated with exchangeable (e.g., MgSO4–extractable) As and Cr. Poor correlation was found between bioaccessible As and Cr and acid ammonium oxalate–extractable forms of these contaminants. Rodriguez et al. (2003) showed that extraction of As-contaminated soils with acidic hydroxylamine hydrochloride, but not water-extractable As, was correlated with bioavailable As (in vivo swine). The ability of soil extractants to predict bioaccessible and bioavailable contaminants will depend on the soil type and waste media. More As than that extracted by MgSO4 (but less than acid ammonium oxalate–extractable) will be required to describe bioaccessibility and bioavailability in contaminated soils.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  Y. W. Lowney, R. C. Wester, R. A. Schoof, C. A. Cushing, M. Edwards, and M. V. Ruby Dermal Absorption of Arsenic from Soils as Measured in the Rhesus Monkey Toxicol. Sci., December 1, 2007; 100(2): 381 - 392. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Roberts, J. W. Munson, Y. W. Lowney, and M. V. Ruby Relative Oral Bioavailability of Arsenic from Contaminated Soils Measured in the Cynomolgus Monkey Toxicol. Sci., January 1, 2007; 95(1): 281 - 288. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
