JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

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 Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bang, J.
Right arrow Articles by Hesterberg, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bang, J.
Right arrow Articles by Hesterberg, D.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Bang, J.
Right arrow Articles by Hesterberg, D.
Related Collections
Right arrow Laboratory Column Studies
Right arrow Toxic Trace Metals
Right arrow Heavy Metals
Right arrow Soil Chemistry
Published in J. Environ. Qual. 33:891-901 (2004).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

TECHNICAL REPORTS

Heavy Metals in the Environment

Dissolution of Trace Element Contaminants from Two Coastal Plain Soils as Affected by pH

JiSu Bang and Dean Hesterberg*

Department of Soil Science, North Carolina State University, Box 7619, Raleigh, NC 27695-7619

* Corresponding author (dean_hesterberg{at}ncsu.edu).

Received for publication May 23, 2003.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Trace element mobility in soils depends on contaminant concentration, chemical speciation, water movement, and soil matrix properties such as mineralogy, pH, and redox potential. Our objective was to characterize trace element dissolution in response to acidification of soil samples from two abandoned incinerators in the North Carolina Coastal Plain. Trace element concentrations in 11 soil samples from both sites ranged from 2 to 46 mg Cu kg–1, 3 to 105 mg Pb kg–1, 1 to 102 mg Zn kg–1, 3 to 11 mg Cr kg–1, <0.1 to 10 mg As kg–1, and <0.01 to 0.9 mg Cd kg–1. Acidified CaCl2 solutions were passed through soil columns to bring the effluent solution to approximately pH 4 during a 280-h flow period. Maximum concentrations of dissolved Cu, Pb, and Zn at the lowest pH of an experiment (pH 3.8–4.1) were 0.32 mg Cu L–1, 0.11 mg Pb L–1, and 1.3 mg Zn L–1 for samples from the site with well-drained soils, and 0.25 mg Cu L–1, 1.2 mg Pb L–1, and 1.4 mg Zn L–1 for samples from the site with more poorly drained soils. Dissolved Cu concentration at pH 4 increased linearly with increasing soil Cu concentration, but no such relationship was found for Zn. Dissolved concentrations of other trace elements were below our analytical detection limits. Synchrotron X-ray absorption near edge structure (XANES) spectroscopy showed that Cr and As were in their less mobile Cr(III) and As(V) oxidation states. XANES analysis of Cu and Zn on selected samples indicated an association of Cu(II) with soil organic matter and Zn(II) with Al- and Fe-oxides or franklinite.

Abbreviations: At, soil samples from Atlantic, NC • Bo, soil samples from Bogue, NC • CBD, citrate–bicarbonate–dithionite • XANES, X-ray absorption near edge structure


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE DEVELOPMENT OF effective in situ remediation technologies requires a sound understanding of the long-term environmental impacts of potentially toxic trace elements (heavy metals and metalloids) in soils or subsurface systems. Total metal concentrations in soils may not necessarily reflect the environmental threat of these metals because increased hazard is usually associated with increased metal solubility, mobility, and bioavailability (McGrath, 1994; Schmidt, 1997). Therefore, information on the solubility and mobility of trace elements in soils, in addition to total metal concentrations, is vital for assessing the hazard of contaminants and for guiding the approach to soil remediation.

The solubility and mobility of trace element contaminants in soils depend on chemical properties such as chemical speciation of the trace element, soil pH and redox potential, and water movement. For example, metal cation solubility typically increases with decreasing pH, and is related to pH-dependent sorption on various soil minerals and organic matter (McBride, 1989; Sauvé et al., 2000; Sukreeyapongse et al., 2002) or pH-dependent changes in solubility of minerals (Lindsay, 1979). On the contrary, the solubility of oxyanions (e.g., arsenate, molybdate, and selenite) typically decreases with decreasing pH, and may be attributed to oxyanion association with surfaces of minerals carrying pH-dependent charge such as Fe- and Al-oxides (e.g., Goldberg, 1992; Zhang and Sparks, 1989; Neal et al., 1987; Hesterberg, 1998). In addition, the formation of soluble complexes (e.g., with dissolved organic matter) increases dissolved metal concentrations in soil solution, and the solubility of dissolved organic matter and metal–organic complexation depends on pH (see Schnitzer and Hansen, 1970; Stevenson, 1976; Sposito, 1986; Boyle and Fuller, 1987; Amrhein et al., 1992; Buffle, 1988). Such interrelated soil chemical processes make it challenging to develop mechanistic models that can predict the pH- or redox-dependent dissolution of potentially toxic trace elements in soils of varying mineralogical properties.

In an attempt to develop a unified set of soil–water partitioning coefficients (Kd) to be used in environmental risk assessments for soil metals, Sauvé et al. (2000) compiled data from more than 70 independent studies. Variations in Kd were up to five orders of magnitude. Despite this high variability, significant relationships were found between Kd and soil solution pH for Cd, Cu, Ni, Pb, and Zn, with total (acid-digestable) soil trace element and soil organic matter concentrations accounting for some of the additional variability (Sauvé et al., 2000). In a more recent study, Sukreeyapongse et al. (2002) used a stirred-flow reactor to determine the pH-dependent kinetics of Cd, Cu, and Pb dissolution from four samples of surface soils from an Ultisol and an Alfisol. Two of the samples were from fields that had received sludge amendments, and all samples had total soil trace element concentrations of <50 mg kg–1. Metal dissolution increased with decreasing pH, and kinetic measurements showed that the relative metal dissolution rates (dissolution rate/total soil metal concentration) had a similar relationship with pH for all three metals. In general, relative dissolution rates of Cd, Cu, and Pb increased markedly when pH decreased below pH 4, with some differences observed between the soils (Sukreeyapongse et al., 2002).

Other studies revealed that the solubility (and hence mobility potential) of organic and metallic contaminants in soils tended to decrease over time (Eccles, 1998; Alexander, 2000; Yin and Allen, 2000). This phenomenon, termed "aging effect," has been attributed to the conversion of more soluble solid-phase metal species into less soluble (chemically more stable) species (Sauvé et al., 2000). For example, mineral forms of contaminants may recrystallize over time into more stable minerals, or adsorbed species may diffuse into the host mineral or be incorporated during growth of the mineral surface (McBride, 1994, Chapter 4; Ford et al., 2001). Metal sulfide precipitation may also decrease solubility of trace elements under reducing conditions (Gambrell, 1994). To assess the potential long-term risks of soil trace element contaminants that are left in place, it is important to also understand the effects of transient soil properties such as pH, which may increase contaminant mobility over time.

In this research, we addressed the effect of long-term soil acidification on dissolution and potential mobility of selected trace elements. Analogous to the study of Sukreeyapongse et al. (2002), our soil samples contained low concentrations (<100 mg kg–1) of trace element contaminants. However, our research included samples from a range of depths in well-drained and poorly drained sandy soils. Also, we used a column flow setup to more directly determine trace element dissolution during the course of declining pH. Furthermore, we applied molecular spectroscopy (XANES) analysis to give insight on the relationship between pH-dependent dissolution and speciation of trace elements. Soils in humid areas tend to acidify over time because of a net influx of protons from natural or anthropogenic sources such as atmospheric deposition. The rate of acidification depends on the rate of proton inputs versus the pH buffering (proton consumption) capacity of a given soil (Van Breemen et al., 1983). Johnston et al. (1986) showed that the surface horizon of a reforested agricultural soil in Rothamsted, UK was acidified from pH 7 to pH 4.2 during a 100-yr period after liming was stopped. Many sandy soils in the Coastal Plain of the southeastern USA have been limed as agricultural land, and when abandoned, are potentially susceptible to more rapid acidification because of the typically lower pH buffering capacity of sandy soils (Pierre, 1927; Alabi et al., 1986). Therefore, pH may be a principal determinant of long-term trace element dissolution and mobility in these soils.

Soils of sandy texture surrounding two abandoned waste incinerator sites on the eastern Coastal Plain of North Carolina were used for the studies. One of the sites contains well-drained soils and the other contains poorly drained soils. The objective of this research was to measure the effect of acidification on dissolution of selected trace element contaminants in soils from these sites, and to establish the range of soil pH that must be maintained to prevent excessive metal dissolution.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soil Sampling and Characterization
Samples were collected in October 2001 from soils surrounding abandoned waste incinerators at two sites associated with the U.S. Marine Corps Air Station (MCAS) at Cherry Point, NC: (i) the Marine Corps Auxiliary Landing Field (MCALF) in Bogue, NC, and (ii) the Marine Corps Outlying Landing Field (MCOLF) near Atlantic, NC. An 8-cm-diameter soil auger was used to collect samples of surface soil near the entranceways and windows of the concrete incinerators, where activities such as transfer of waste and ashes into and out of the incinerator would potentially contaminate the soils. We also collected some additional soil samples at locations and depths (to approximately 1 m) that were expected to contain greater concentrations of As, Cr, Cu, Zn, or Pb based on a previous survey of contaminant concentrations in 44 soil samples taken at various locations and depths at the sites (Hesterberg, 2001). Soil samples that were not affected by the incinerators were collected about 150 m to the northwest of the Bogue incinerator site and about 500 m to the northeast of the Atlantic incinerator site. Trace element concentrations in these samples were assumed to be representative of noncontaminated soils in the area, although historical inputs of contaminants are not known. Maps showing the sampling locations relative to the incinerators are shown in Bang (2002).

All sample handling in the laboratory was done in a glove box under an Ar(g) atmosphere, and field-moist samples were stored in amber-colored borosilicate glass bottles with Teflon-lined rubber seal inserts and aluminum crimp caps to try to minimize sample changes during the course of the study. The soil samples were placed on ice shortly after collection, and otherwise stored at 4°C.

To determine whether the soils were alkaline (calcareous) or acidic, soil pH was measured using a 1:1 soil to deionized water ratio (Thomas, 1996). Moist soil color was determined in the laboratory on sieved samples using Munsell Soil Color Charts (Munsell Color, 1998). Extractable Fe, Al, and Mn were determined using citrate–bicarbonate–dithionite (CBD) solution and on separate samples using acid ammonium oxalate (0.2 mol L–1, pH 3) in the dark (Jackson et al., 1986). To estimate organic verse inorganic (carbonate) soil carbon, organic carbon was measured using the Walkley–Black method (Nelson and Sommers, 1996), and total carbon was determined by combustion on a PerkinElmer (Wellesley, MA) Model 2400 CHN elemental analyzer.

Soil sample concentrations of USEPA priority metals and other selected trace elements, including As, Cu, Cd, Cr, Pb, Ag, Zn, Fe, Ba, Se, and Mn, were determined by acid digestion using the USEPA SW-846 3050 analysis series. Extracts were analyzed using inductively coupled plasma–optical emission spectrometry (ICP–OES) for As, Ba, Cd, Se, and Ag, or flame atomic absorption spectrometry (FAAS) for Cr, Cu, Zn, Fe, and Mn (Wagner, 1996, p. 65–74). Method detection limits were determined for each instrument and respective method. For quality control of the digestion procedure, triplicate blanks for digestion and analysis methods, matrix spike duplicates, and triplicate samples of a standard reference material (2710 Montana Soil; National Institute of Standards and Technology, Gaithersburg, MD) were carried along with each set of soil sample digests. Digestions were done using "trace metal"-grade HCl and HNO3 (Fisher Scientific, Pittsburgh, PA). Solutions from digested soil samples were stored in 125-mL high-density polyethylene (HDPE) sample bottles at 4°C until analysis.

Laboratory Column Flow Study—Acidification
Column flow experiments were conducted on seven columns of soil samples from the Bogue site (Bo) (designated Bo-1, Bo-2, Bo-3, Bo-4, Bo-5, Bo-6, and Bo-7) and four columns of soil samples from the Atlantic site (At) (designated At-1, At-2, At-3, and At-4). The experiments were conducted to determine the concentrations of dissolved metals eluted from soil samples as a function of pH. The experimental system comprised a flask of 0.01 mol CaCl2 L–1 solution purged with either water-saturated N2 gas or compressed air, a peristaltic pump, a 3-cm-diameter x 1-cm-long soil column, 60-mL Teflon bottles to collect effluent samples, a potentiometer for pH and Eh measurement, and a PC controller to continuously collect pH and Eh data. In-line microsized pH and Eh (redox) electrodes were connected into the Teflon flow line between a soil column and effluent sample bottle.

Whereas other laboratory column studies used columns of 10 to 16 cm in length (Boyle and Fuller, 1987; Dunnivant et al., 1992; Alesii et al., 1980), we used a 1-cm-long column to yield a shorter transport distance for dissolved trace elements and to decrease the pH gradient across the length of the column. A diameter of 3 cm accommodated enough soil to ensure that the trace elements were not fully depleted from the column during a flow experiment. Each 1-cm-long soil sample was encased in a column made of 3.8-cm-long sections of Teflon (FEP) tubing (3.2-cm i.d.) fitted with two 2.5-cm-long chromatography column end plugs with 20-µm Teflon bed support screens. Approximately 13.7 g of the moist soil were packed at a bulk density of 1.7 g cm–3 into a 1-cm-long column (assuming a particle density of 2.65 g cm–3), yielding a porosity of 0.36 (1 pore volume = 2.88 cm3). To pack a column, the soil material was mixed well with deoxygenated, deionized water to make a slurry in a Teflon beaker. The suspension was poured very rapidly and quantitatively into a Teflon sleeve set in the bottom end plug while simultaneously drawing a vacuum to remove free water. The sandy soil settled uniformly in 2 to 3 s and the excess water was quickly removed by the vacuum pump. Columns for the more reduced (redox) soil samples from the Atlantic site were packed in a glove box under an N2(g) atmosphere and a red-filtered safelight.

Deoxygenated (N2 purged) or oxygenated (aerated with normal air) electrolyte solutions were passed through each column at a rate of 0.1 mL per minute during an approximately 300-h flow period. Experiments were conducted at ambient temperature of approximately 22°C. After an initial 24-h application of deoxygenated 0.01 mol CaCl2 L–1 solutions, aerated 0.01 mol CaCl2 L–1 solutions were applied to the more oxidized soil samples from the Bogue site, and deoxygenated 0.01 mol CaCl2 L–1 solutions were pumped through the soil samples from the Atlantic site. A solution of 0.01 mol CaCl2 L–1 was chosen because this Ca concentration is close to that predicted by thermodynamics for an alkaline (pH 7.8) soil in equilibrium with atmospheric CO2 (Lindsay, 1979). The mean pH of alkaline soils used in our column flow experiments was 7.6 ± 0.2 (Table 1, discussed below). To determine how dissolved metal concentrations were affected by pH, 0.01 mol CaCl2 L–1 solutions containing 0.00005 to 0.002 mol HCl L–1 were passed through the columns (450–490 column pore volumes) to decrease pH at a fairly constant rate (0.02–0.86 mmol H+ kg–1 soil h–1) from the initial (near-neutral or alkaline) effluent pH to about pH 4 during approximately 280 h of flow. Because of the long time period of each column-flow experiment (10 d of flow), most soil samples were analyzed only once, whereas one sample (Bo3) was analyzed in triplicate to assess variability. In this latter experiment, two of the three column samples (Reps 1 and 2) were acidified to approximately pH 4 during 150 h of flow, and the third sample (Rep 3) was acidified over 280 h as was done in all other experiments. Results (discussed below) were reproducible despite a nearly twofold difference in the rate of acidification.


View this table:
[in this window]
[in a new window]
  		Table 1. Selected characteristics of soil samples collected from the Bogue (Bo) and Atlantic (At) incinerator sites and of reference soil samples (Bo-R1, Bo-R2, At-R1, and At-R2) collected approximately 150 or 500 m from incinerators.

 
Effluent samples were collected continuously for approximately 10-h intervals (approximately 16 column pore volumes) to obtain enough sample to analyze the trace elements of concern without dilution. The samples were acidified to pH 2 to 2.5 with 1 mol HCl L–1 and stored in 60-mL Teflon bottles at 4°C until analyzed. Dissolved As, Ba, Cd, Se, and Ag were measured using inductively coupled plasma–optical emission spectrometry, and concentrations of Cu, Zn, Pb, Cr, Fe, and Mn were analyzed using flame atomic absorption spectrometry.

X-Ray Absorption Near Edge Structure Spectroscopy Analysis
The oxidation states of Cr and As and speciation of Cu and Zn were analyzed using synchrotron XANES spectroscopy on selected soil samples taken at various locations and depths from the Bogue or Atlantic site. These sampling locations were near those of the present study and the extent of XANES analyses reported here was contained by the feasibility of collecting data based on trace element concentrations. The XANES data collection was done at Beamline X-11A of the National Synchrotron Light Source (NSLS), Brookhaven National Laboratory (BNL). The principles and procedures for collecting and analyzing XANES data were described by Sayers and Bunker (1988), Fendorf and Sparks (1996), and Fendorf (1999). Approaches and applications of X-ray absorption spectroscopy used to determine the speciation of heavy metals in geochemical systems (soils and ground water aquifer matrices) were based on those given by Wang et al. (1998) and Hesterberg et al. (1997).

Moist soil samples and standard mineral samples were mounted in polyacrylamide or thin Teflon sample holders for XANES analysis. Mounting of soil samples was done under an N2(g) atmosphere, and samples were held in the sample holders using KAPTON (Budnick Converting, Columbia, IL) tape. To calibrate the XANES spectra with respect to proportions of Cr(VI) and Cr(III), two mineral standards, K2Cr2O7 (Allied Chemical, Morristown, NJ) and Cr2O3 (Aldrich, Milwaukee, WI), were mixed in various proportions between 0 and 100 mol % Cr(VI) and diluted in boron nitride to 350 mmol Cr kg–1. These standards were analyzed in transmission mode, while soil samples, of much lower Cr concentrations, were analyzed in fluorescence mode using a 13-element, solid-state detector. Multiple scans were collected for each sample and assemble-averaged to improve the data quality. The Si(111) monochromator at Beamline X-11A was detuned 40% during Cr XANES data collection. The XANES spectra at the As K-edge (11870 eV) were also collected using transmission mode for standards of As2O3 [As(III)] and As2O5 [As(V)] (Alfa Aesar, Ward Hill, MA) and fluorescence mode for multiple scans on soil samples of lower As content. The monochromator was detuned 30% for As data collection. All Cr and As XANES spectra were linear baseline and single-point background corrected at a consistent energy using Kaleidagraph (Synergy Software, 2000), following procedures of Sayers and Bunker (1988).

To characterize solid-phase chemical speciation of Cu and Zn in selected soil samples containing greater concentrations of these elements, XANES spectra were collected at the Cu and Zn K-edges. Analogous to Cr and As data collection, data for soil samples and standards of adsorbed species were collected in fluorescence mode. Mineral standards were diluted in boron nitride to a concentration yielding approximately unit absorption (approximately 600 mmol kg–1) and data were collected in transmission mode. Monochromator detuning was 40% for Cu and 30% for Zn. Ensemble-averaged spectra were normalized using the computer program MACXAFS (Bouldin et al., 1995). We estimated the dominant solid-phase species in soil samples using nonlinear, least-squares fitting (linear combination fitting; e.g., Hesterberg et al., 1997) using the program developed for SciLab 2.6 (Scilab Group, 2002) as reported by Beauchemin et al. (2003). The XANES spectra for all binary and ternary combinations of available Zn and Cu standards were fit to the sample spectra. Using ternary combinations gave only minor improvement in the overall best fit, so the best-fit results reported here are for binary combinations of standards.

The following standards were used in the fitting analysis of Zn K-XANES data: Zn adsorbed on gibbsite at 30 mmol kg–1 (analog for Zn on Al-oxides), Zn(II) adsorbed on ferrihydrite at 160 mmol kg–1 (analog for Zn on Fe-oxides), Zn(II) coprecipitated with ferrihydrite at 30 mmol kg–1, franklinite (ZnFe2O4), wurtzite (ZnS), smithsonite (ZnCO3), zinc carbonate hydroxide monohydrate [ZnCO3·2Zn(OH2)2·H2O], zincite (ZnO), zinc hydroxide [Zn(OH)2], zinkosite (ZnSO4), and hopeite [Zn3(PO4)2·4H2O]. Standards used for fitting of Cu K-XANES data were Cu(II) complexed with soil humic acid (data from Alcacio et al., 2001), Cu(II) adsorbed on goethite (data from Alcacio et al., 2001), tenorite (CuO), and covellite (CuS). Mineral standards were purchased from chemical suppliers and characterized using X-ray diffraction analysis. Linear combination fitting analysis was done over the energy range 9655 to 9740 eV for Zn XANES spectra and from 8980 to 9050 eV (sample Bo4*) or 9100 eV (Bo1*) for Cu spectra. The shorter range was used for the Bo4* sample to avoid excessive contribution of a sloping background at higher energies in the spectrum for this sample.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soil Properties
Selected chemical and mineralogical properties of the 15 samples collected around and away from the incinerators are given in Tables 1 and 2. Soil samples from the Bogue site (Bo) were typically brown or yellowish brown in color, while samples from the Atlantic site (At) were all very dark gray or black in color (Table 1). The concentrations of CBD-extractable Fe were typically lower in samples from the Atlantic site (102–2850 mg Fe kg–1) compared with those from the Bogue site (4400–9900 mg Fe kg–1), indicating lower concentrations of Fe-oxide minerals at the Atlantic site (Table 1). The greater oxalate- (Feo) to CBD-extractable Fe (Fed) ratio for samples from the Atlantic site (Feo/Fed = 0.2–0.4) compared with samples from the Bogue site (Feo/Fed = 0.07–0.1) indicated a greater proportion of less stable, poorly crystalline Fe-oxides at the Atlantic site (Schwertmann, 1993). At the Bogue site, Feo to Fed ratios indicated that more crystalline Fe-oxides were present, and the soil colors suggested that goethite was the dominant Fe-oxide mineral. The yellowish to brown color, high concentration of CBD-extractable Fe, and low oxalate- to CBD-extractable Fe ratio at the Bogue site were all indicative of well-drained soil and more oxidizing redox conditions. The low-chroma soil colors, lower concentrations of CBD-extractable Fe, and higher ratios of oxalate- to CBD-extractable Fe at the Atlantic site were indicative of poorly drained soil conditions and more active redox processes (Schwertmann, 1993). However, Eh values measured in effluent samples at the beginning of column flow studies were not significantly ({alpha} = 0. 05) different between samples from the Atlantic site (mean = 319 ± 6 mV) and samples from the Bogue site (mean = 387 ± 83 mV). Soil samples from the Bogue site had CBD-extractable Al ranging from 740 to 1350 mg Al kg–1, suggesting that Al was present in Al-oxide minerals or Al-substituted Fe-oxides (Jackson et al., 1986). The CBD-extractable Al in soil samples from the Atlantic site mostly remained below analytical detection limits of 110 mg Al kg–1. The low concentration of the CBD-extractable Mn at both sites indicated the low contents of reducible Mn in the soils. A trend of higher organic carbon contents in samples from the Atlantic site (2–31 g kg–1) compared with those from the Bogue site (0.7–7 g kg–1) (Table 1) suggests that organic carbon contents may have accumulated in the former due to more reducing redox conditions. Despite the apparent trend, a statistical t test showed that organic carbon contents of soil samples from the two sites were not significantly ({alpha} = 0.05) different because of high variation in organic carbon contents.


View this table:
[in this window]
[in a new window]
  		Table 2. Concentrations of selected elements measured in soil samples from the Bogue (Bo) and Atlantic (At) incinerator sites and in reference soil samples (Bo-R1, Bo-R2, At-R1, and At-R2) collected approximately 150 or 500 m from incinerators.{dagger}

 
The pH values of soil samples from the Bogue site ranged from pH 5.3 to 8.1 (mean soil pH = 7.2 ± 0.9). The pH levels of Wando soils (thermic, coated Typic Quartzipsamments) and Seabrook soils (mixed, thermic Aquic Udipsamments) mapped in the vicinity of the Bogue site range from 4.5 to 7.3 (Goodwin, 1978). Soil samples from the Atlantic site had pH values ranging from 5.8 to 7.6 (mean soil pH = 6.3 ± 0.8). Poorly drained Leon soils (sandy, siliceous, thermic Aeric Alaquods) and Murville soil (sandy, siliceous, thermic Umbric Endoaquods) mapped in the vicinity of the Atlantic site (Goodwin, 1978) are classified as extremely acid to slightly acid, meaning that pH ranges from pH 3.6 to 5.5.

Soil Trace Element Concentrations
Concentrations of trace elements in the 15 soil samples collected for this study are given in Table 2. Concentrations of trace elements in seven soil samples taken from the vicinity of the incinerator at the Bogue site ranged from 4 to 46 mg Cu kg–1, 3 to 55 mg Pb kg–1, 10 to 98 mg Zn kg–1, 3 to 11 mg Cr kg–1, 0.3 to 10 mg As kg–1, and <0.01 to 0.9 mg Cd kg–1. Silver was not detected in any of the samples from the Bogue site at the detection limit of 0.59 mg Ag kg–1. Acid-soluble Fe in samples from the Bogue site was comparable with CBD-extractable Fe (Table 1 and 2). Concentrations of selected trace elements in soil samples taken from around the incinerator at the Atlantic site ranged from 2 to 25 mg Cu kg–1, 29 to 105 mg Pb kg–1, 6 to 102 mg Zn kg–1, 1 to 9 mg Cr kg–1, <0.1 to 0.9 mg As kg–1, and <0.01 to 0.6 mg Cd kg–1. Silver was not detected in any of the four samples from the Atlantic site at the detection limit of 0.59 mg Ag kg–1. In general, soils around the two incinerator sites contain potentially toxic trace elements occurring at concentrations of ≤100 mg kg–1.

Column Flow Studies
Metal Dissolution Trends
Figure 1 shows temporal trends in pH and dissolved Cu, Pb, and Zn for the column flow experiment on soil sample Bo-2. These trends were generally representative of trends observed for all soil samples. With increasing throughput of acidified 0.01 mol CaCl2 L–1 solution over the course of 300 h, pH declined and dissolved Cu, Zn, and Pb concentrations increased, particularly at low pH levels (Fig. 1). The column flow results on triplicate subsamples of soil sample Bo3 indicated that trends in dissolved Cu or Zn concentration as a function of pH were reproducible (range of SD = 0–0.11), despite differences in the rate of acidification between sample replicates (data not shown). Data for dissolved Pb and other trace elements remained below our analytical detection limits.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 1. Typical trends of dissolved Cu (•), Pb ({diamondsuit}), and Zn ({blacktriangleup}) concentrations and pH (—) in effluent solutions collected during a column flow experiment on a soil sample from the Bogue site (Bo-2) as a function of the throughput volume of acidified, 0.01 mol CaCl2 L–1 solution.

 
Figure 2 shows dissolved Cu, Pb, and Zn as a function of pH for column flow experiments on all 11 soil samples from both the Bogue and Atlantic sites. Dissolved concentrations of these trace elements increased with decreasing pH when a column was acidified below some threshold pH for metal dissolution (discussed in more detail below), regardless of whether the soil was more reduced (Atlantic site) or more oxidized (Bogue site). Samples from the reference sites were not included in column flow studies. The maximum dissolved concentrations of Cu, Pb, or Zn observed, usually at the lowest pH achieved, varied between soil samples and between the three metals (Fig. 2).



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 2. Dissolved concentrations of Cu (A), Pb (B), and Zn (C) in effluent solutions for column flow experiments on eleven soil samples from the Bogue (Bo) and Atlantic (At) sites. For reference, ground water standards adopted in North Carolina were 1.0 mg Cu L–1, 0.015 mg Pb L–1, and 2.1 mg Zn L–1.

 
Using mass balance, we determined the amount of each soil trace element that was dissolved and removed from each column during a flow experiment to ensure that complete removal of trace elements did not occur. The proportion of each trace element removed was calculated as the difference between the approximate amount of trace element in the column (based on trace element concentrations in Table 2 multiplied by the mass of soil in a column) and the cumulative amount of trace element discharged in column effluent. Based on these calculations, the proportions of Cu, Pb, and Zn discharged from each column during the entire flow experiment were always less than 50% of soil Cu, less than 6% of soil Pb, and less than 65% of soil Zn.

Despite measurable quantities of Cr, As, Cd, and Ag in soil samples (Table 2), aqueous concentrations of these elements measured in selected effluent samples from all eleven soil columns were all below our analytical detection limits of 0.17 mg Cr L–1, 0.03 mg As L–1, 0.008 mg Cd L–1, and 0.01 mg Ag L–1. The soil samples analyzed contained Ba and Se at concentrations of <80 mg Ba kg–1 or <0.4 mg Se kg–1 (data not shown), yet no dissolved Ba and Se were detected in any effluent samples. Thus, unlike Cu, Pb, and Zn, we observed no detectable increase in dissolved concentrations of these elements as a function of pH. Arsenic and chromium may occur in either of two oxidation states in soils, with Cr(III) being typically less mobile and less hazardous than Cr(VI), and As(V) being typically less mobile than As(III) (Alloway, 1995). Oxidation states of Cr and As can be analyzed using XANES spectroscopy. Synchrotron XANES data for selected soil samples from a preliminary survey of the sites (Hesterberg, 2001) showed that Cr(III) was the dominant oxidation state of Cr, with no Cr(VI) detected (Fig. 3) . The absence of Cr(VI) is indicated by the absence of the pre-edge spike at 5994 eV as shown in the 10% Cr(VI) standard in Fig. 3. Likewise, XANES analysis indicated that As in selected samples was predominantly in the typically less mobile As(V) oxidation state, although minor proportions of As(III) may be present (Fig. 4) . The proportions of As(III) and As(V) are indicated by the intensities of peaks in the As K-XANES at energies corresponding to those of the As(III) and As(V) standards. A shoulder corresponding with the energy of the white-line peak for As(III) is discernable on the low energy side of the main As(V) white-line peak in spectra for some soil samples, suggesting that minor amounts of As(III) were present. However, note that the relative intensity of the normalized white-line peak for the As(III) standard is lower than that of the As(V) standard. Thus, relative peak intensities for As in the soil samples are not directly related to the proportion of total As in each oxidation state.



View larger version (37K):
[in this window]
[in a new window]
  		Fig. 3. Stacked, chromium K-XANES (X-ray absorption near edge structure) spectra for soil samples collected for a preliminary study (Hesterberg, 2001) at various depths from sampling locations around the incinerator at the Bogue site. A comparison with standards containing 0 or 10 mol % of total Cr as Cr(VI) indicated that <10% of the Cr in the soil samples was Cr(VI). Samples designated as Bo4*, Bo7*, and Bo8* correspond to samples listed in Tables 1 and 2 that were in closest proximity.

 


View larger version (40K):
[in this window]
[in a new window]
  		Fig. 4. Arsenic K-XANES (X-ray absorption near edge structure) spectra for soil samples collected for a preliminary study (Hesterberg, 2001) from sampling locations around the incinerator at the Bogue site as compared with As(III) and As(V) standards. Spectra for soil samples showed that As(V) was dominant. The energy scale is normalized to the K edge of elemental As at 11867 eV. See Fig. 3 caption for sample designations.

 
Copper K-XANES spectra for two soil samples from the Bogue site and Zn K-XANES spectra for nine samples from both sites are shown in Fig. 5 and 6 along with selected standards. Linear combination fitting results are given in Table 3. To the extent that the standards used in the fitting are representative of the soil metal species, the similarity of spectra for soils and Cu on humic acid in Fig. 5 and the fitting results in Table 3 suggested that Cu(II) in the two surface soil samples analyzed was predominantly bounded with soil organic matter. Zinc K-XANES spectra in Fig. 6 were consistently fit with the standard of Zn(II) adsorbed on gibbsite, usually in combination with either Zn(II) adsorbed on ferrihydrite or franklinite as a minor species. Fitting results suggested that two of the surface soil samples (Bo2 and At 2) contained minor amounts of Zn sulfide. The one unique sample (Bo4) was fit with a combination of Zn-hydroxide and Zn(II) adsorbed on ferrihydrite (Table 3). This soil sample had the greatest pH (8.1) of those analyzed (Table 1).



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 5. Copper K-XANES (X-ray absorption near edge structure) spectra for soil samples Bo1* and Bo4* (collected in the vicinity of samples Bo1 and Bo4 from this study) and Cu standards used for linear combination fitting. Dashed lines represent the best-fit combination of standards reported in Table 3 for soil samples.

 


View larger version (43K):
[in this window]
[in a new window]
  		Fig. 6. Zinc K-XANES (X-ray absorption near edge structure) spectra for selected soil samples used in column flow experiments or collected in the vicinity of column samples (Bo1*, At5*), and selected standards used in linear combination fitting. Dashed lines represent the best-fit combination of standards reported in Table 3 for soil samples.

 

View this table:
[in this window]
[in a new window]
  		Table 3. X-ray absorption near-edge structure fitting results of Cu and Zn showing the relative proportion (percentage, normalized to sum = 100%) of each Cu and Zn standard that yielded the best fit to the soil K-XANES (X-ray absorption near edge structure) data in linear combination fitting.

 
Concentrations of oxalate- and CBD-extractable Fe and CBD-extractable Al reported in Table 1 were mostly greater than concentrations of the trace elements being studied (Table 2), so sorption of Cu, Pb, and Zn on Fe- and Al-oxide minerals is a possibility at the Bogue site. Soil organic matter was reported to have a greater affinity for Cu(II) [and Cd(II)] than oxide minerals, especially under acid conditions (McLaren et al., 1983; Zachara et al., 1992). Mineral–organic matter complexes may also be important metal sorbents in soils, with the possibility of ternary complexes being formed as shown for Cu(II) by Alcacio et al. (2001).

Threshold pH for Metal Dissolution
Because we could not detect Cr, As, Cd, Ag, Se, or Ba in column effluent samples, the following discussions focus on Cu, Pb, and Zn dissolution. As shown in Fig. 2, key characteristics of pH-dependent metal dissolution were that (i) column effluent data often exhibited a threshold pH below which concentrations of Cu, Pb, or Zn increased exponentially, and (ii) the maximum concentration of dissolved metal was typically found at the lowest pH levels. We quantified the threshold pH for a given column as the pH below which dissolved metal concentrations were greater than five times the standard deviation of dissolved concentrations measured at a pH of >4.5 for Cu and Pb (when detectable), and at a pH of >5.5 for Zn. Dissolved Cu and Pb were below our flame atomic absorption spectrometry detection limits (0.008 mg Cu L–1 and 0.018 mg Pb L–1) in the majority of effluent samples collected when pH was >4. In cases where Cu or Pb could not be detected at pH > 4.5, the threshold pH was taken as the pH below which dissolved metal first increased to a sustained detectable level.

The threshold pH at which dissolved Cu increased substantially ranged from 3.9 to 4.4 (mean threshold pH = 4.1 ± 0.2) for the seven soil column flow experiments on samples from the Bogue site. Threshold pH levels at which dissolved Pb showed a substantial increase occurred in only two of the seven soil columns, at pH 4.2 and 4.0 (mean threshold pH = 4.1 ± 0.1). For comparison, Sukreeyapongse et al. (2002) found that the relative release rates of Cu and Pb from soils from Denmark and Thailand showed similar trends with decreasing pH. In two of our other columns, dissolved Pb was detectable, but only increased to levels of 0.005 and 0.022 mg L–1 at the lowest pH. The threshold pH values for Zn dissolution ranged from 4.2 to 5.3 for seven samples (mean threshold pH = 4.6 ± 0.4). Based on data for Zn (which showed the highest threshold pH for dissolution), dissolution of Cu, Pb, and Zn can be minimized at the Bogue site by maintaining soil pH of >5.3 (by liming as needed). At pH levels below the threshold pH for a given element (particularly Cu, Pb, or Zn), the mobility of that metal could potentially increase if water flow is sufficient, assuming that long-term metal behavior in the field can be assessed from our short-term column-flow experiments.

Consistent with observations on soil samples from the Bogue site, the threshold pH levels for Cu and Pb dissolution during soil acidification were lower than those for Zn on samples from the Atlantic site (Fig. 2). The threshold pH levels for Cu dissolution observed on two of the four soil columns were 3.90 and 4.03 (mean threshold pH = 3.97 ± 0.1). A threshold pH for Pb dissolution was found for only one column sample at pH 4.0. Threshold pH values for Zn dissolution were found on three of the four samples, and ranged from pH 5.1 to 5.4 (mean threshold pH = 5.2 ± 0.2). These results indicate that dissolution of heavy metals in soils at the Atlantic site can be minimized by maintaining a pH of >5.2, analogous to our findings for the Bogue site.

Maximum Dissolved Metal Concentrations
The maximum concentrations of dissolved metals observed in column experiments on soil samples from the Bogue site were 0.32 mg Cu L–1, 0.11 mg Pb L–1, and 1.25 mg Zn L–1. In the sample taken in the vicinity of the incinerator window (Bo-3), the maximum concentrations were observed at the lowest pH achieved in this column flow experiment, consistent with the general trend of soil metal cation solubility increasing with decreasing pH (McBride, 1989). For reference, these maximum observed Cu and Zn concentrations of 0.32 mg Cu L–1 and 1.25 mg Zn L–1 were less than the ground water quality standard of 1.0 mg Cu L–1 and 2.1 mg Zn L–1 adopted in North Carolina for Class GA ground water (North Carolina Department of Environment, Health, and Natural Resources, 1993). The greatest dissolved Pb concentration measured in these column effluents (0.11 mg Pb L–1) was nearly 10-fold greater than the ground water standard of 0.015 mg Pb L–1 adopted in North Carolina (North Carolina Department of Environment, Health, and Natural Resources, 1993). A lack of sensitivity of flame atomic absorption spectrometry for Pb (detection limit of 0.017 mg L–1) did not allow accurate evaluation of the concentrations of dissolved Pb in most effluent samples relative to the ground water standard.

The maximum dissolved concentrations of metals observed for any of the four columns from the Atlantic site were 0.25 mg Cu L–1, 1.2 mg Pb L–1, and 1.4 mg Zn L–1, all in the sample taken near the incinerator entranceway (At-1). These maximum concentrations of Cu and Zn are also less than the Class GA ground water standards adopted in North Carolina (North Carolina Department of Environment, Health, and Natural Resources, 1993). The maximum dissolved Pb concentration of 1.2 mg L–1 was 78-fold greater than the ground water standard of 0.015 mg Pb L–1. However, Pb was only detectable in samples from one of the four soil columns from the Atlantic site and only at a pH of <4.0. Nevertheless, these results are consistent with the notion that trace element dissolution in soils is affected by chemical conditions (e.g., pH) in addition to total soil metal concentration.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Overall pH Dependence of Metal Dissolution
The pH-dependent concentration of dissolved Zn in column effluent samples from both soils varied from <0.08 mg Zn L–1 at near-neutral pH to 1.25 mg Zn L–1 at pH of <4 (Fig. 2). A nearly constant or perhaps slightly increasing dissolved Zn concentration was observed between pH 6.5 and 5.0. Dissolved Zn was typically greater than dissolved Cu and Pb over the entire pH range represented by our data (pH 3.8–8.1; Fig. 2). This difference could perhaps be attributable to the speciation of these metals in the soils studied. For example, XANES speciation results (Fig. 5 and 6, Table 3) suggested that Cu(II) and Zn(II) in the samples analyzed occurred mainly in adsorbed forms (to organic matter or Al and Fe oxides). Stevenson (1976) indicates that Cu2+ (and Pb2+) are more strongly adsorbed than Zn2+ by soil organic matter and Fe-oxide minerals. The affinity sequence of organic matter for trace elements reported by Stevenson (1976) and Schnitzer and Skinner (1965) was in the order Cu2+ > Pb2+ > Zn2+. The results of several studies showed that the intrinsic affinity of heavy metal cations on goethite ({alpha}-FeOOH) decreased in the order Cu2+ > Pb2+ > Zn2+ (Forbes et al., 1976; McKenzie, 1980; Pickering, 1988). Adsorption edges reported by Kinniburgh et al. (1976) indicated that the pH level where 50% of each metal cation was retained by Fe-oxide minerals increased in the order (Pb2+ at pH 3.1) < (Cu2+ at pH 4.1) < (Zn2+ at pH 5.4) < (Cd2+ at pH 5.8); for noncrystalline Al-oxide gels, the pH of 50% retention increased in the order (Cu2+ at pH 4.8) < (Pb2+ at pH 5.2) < (Zn2+ at pH 5.6). These data indicate that Zn2+ would generally be released from Fe- and Al- oxides at a higher pH level than Cu2+ or Pb2+, consistent with our observation of greater threshold pH levels for soil Zn release (pH 4.6 ± 0.4) compared with Cu or Pb release (pH 4.1 ± 0.2 and pH 4.1 ± 0.1, respectively).

Dissolution of Metals in Relation to Total Soil Metal Concentrations
To compare metal dissolution between different soil samples, we evaluated dissolved metal concentrations at pH 4.0. Under the conditions of our column acidification study, dissolved Cu and Zn concentrations at pH 4.0 were not significantly different among soil samples from each of the two sites ({alpha} = 0.05), despite differences in soil properties and redox conditions. Therefore, effluent data for the 11 columns from both sites were pooled to determine whether there was a relationship between dissolved metal concentrations at pH 4 and the initial, acid extractable soil metal concentrations (reported in Table 2). This analysis was not done for Pb (or other metals) because dissolved Pb was detected at pH 4 in only 5 of the 11 soil samples. However, it is noteworthy that the soil sample with the greatest concentrations of dissolved Pb in column effluent samples at pH of ≤4 was the sample containing the greatest concentration of soil Pb (Fig. 2, Table 2).

Dissolved concentrations of Cu and Zn at pH 4.0 in column effluent samples are shown as a function of initial soil Cu and Zn concentrations in Fig. 6. Effluent concentrations of Cu at pH 4 could be related to soil Cu using a linear model with a significant coefficient of determination (r2 = 0.74, p < 0.001), indicating that soil Cu concentration significantly influenced Cu dissolution (Fig. 7) . Wu et al. (2000) also found that Cu (and Cd) solubility was related to total metal concentration in their samples. In contrast, dissolved Zn at pH 4 was not linearly related to soil Zn, but remained at <0.2 mg L–1 for initial soil Zn concentrations between 8 and 70 mg Zn kg–1, and was 5- to 6-fold greater in two samples containing 90 and 98 mg Zn kg–1. Multiple linear regression analysis showed no significant relationship between dissolved Cu or Zn concentration at pH 4 and soil organic matter or CBD-extractable Fe concentrations.



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 7. The relationship between acid-digestable soil Cu or Zn concentrations and dissolved Cu (•) and Zn ({blacktriangleup}) at pH 4.0 in column effluent samples.

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
Column flow results showed that acidification of columns to an effluent pH of approximately 4 had no apparent effect on dissolved Cr, As, Cd, and Ag concentrations, which remained below our analytical detection limits. Synchrotron XANES spectra of soil samples from one site showed that Cr and As were in their less mobile and potentially less toxic Cr(III) and As(V) forms, which may explain the lack of detectable dissolution. However, dissolved Cu, Pb, and Zn concentrations increased as pH declined below threshold levels of pH 4.0 to 4.1 for Cu and Pb and pH 4.6 to 5.2 for Zn. These results for Cu and Zn were consistent with XANES data suggesting that Cu was mainly bound with soil organic matter in two surface soil samples analyzed, and Zn was usually associated with Al- and Fe-oxide type minerals. Maximum dissolved concentrations of Cu, Pb, and Zn observed under the lowest pH conditions achieved (pH ≤ 4.0) in samples from both sites were ≤0.32 mg Cu L–1, ≤1.38 mg Zn L–1, and ≤1.17 mg Pb L–1. For all 11 columns studied, effluent concentrations of Cu at pH 4 were linearly related to soil Cu concentrations, but Zn showed no similar trend. Although nondetectable concentrations of trace elements in column effluent samples at the current soil pH suggested that metal mobility would be very low, acidification of the soils at these sites over time to values as low as pH 3.6, as reported for nearby naturally occurring soils, could potentially mobilize Cu, Pb, and Zn. Our column data indicated that the soils at these sites could be maintained at a pH of >5 to minimize the solubility and mobility of these trace elements.


   ACKNOWLEDGMENTS
 
Funding for this research was provided by the U.S. Department of the Navy through the U.S. Marine Corps Air Station at Cherry Point, NC. We thank Ken Cobb and Kim Hutchison for assistance with sample collection and laboratory efforts, and we are grateful to Dr. Kumi Pandya, Beamline scientist at NSLS Beamline X-11A, for providing assistance with XANES data collection. This research was carried out (in part) at the National Synchrotron Light Source, Brookhaven National Laboratory, which is supported by the U.S. Department of Energy, Division of Materials Sciences and Division of Chemical Sciences.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2004 33: 799-804. [Full Text]  




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 Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bang, J.
Right arrow Articles by Hesterberg, D.
Right arrow Search for Related Content