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


     

Published online 1 May 2008
Published in J Environ Qual 37:1190-1200 (2008)
DOI: 10.2134/jeq2007.0326
© 2008 American Society of Agronomy, Crop Science Society of America, and 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
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Voegelin, A.
Right arrow Articles by Kretzschmar, R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Voegelin, A.
Right arrow Articles by Kretzschmar, R.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Voegelin, A.
Right arrow Articles by Kretzschmar, R.
Related Collections
Right arrow Batch Studies
Right arrow Toxic Trace Metals
Right arrow Sorption/Exchange
Right arrow Heavy Metals
Right arrow Soil Pollution

TECHNICAL REPORTS

Heavy Metals in the Environment

Zinc Fractionation in Contaminated Soils by Sequential and Single Extractions: Influence of Soil Properties and Zinc Content

Andreas Voegelin*, Gerome Tokpa, Olivier Jacquat, Kurt Barmettler and Ruben Kretzschmar

Soil Chemistry Group, Institute of Biogeochemistry and Pollutant Dynamics, ETH Zurich, ETH Zentrum CHN, CH-8092 Zürich, Switzerland

* Corresponding author (voegelin{at}env.ethz.ch).

Received for publication June 19, 2007.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions
 REFERENCES
 
We studied the fractionation of zinc (Zn) in 49 contaminated soils as influenced by Zn content and soil properties using a seven-step sequential extraction procedure (F1: NH4NO3; F2: NH4–acetate, pH 6; F3: NH3OHCl, pH 6; F4: NH4–EDTA, pH 4.6; F5: NH4–oxalate, pH 3; F6: NH4–oxalate/ascorbic acid, pH 3; F7: residual). The soils had developed from different geologic materials and covered a wide range in soil pH (4.0–7.3), organic C content (9.3–102 g kg–1), and clay content (38–451 g kg–1). Input of aqueous Zn with runoff water from electricity towers during 26 to 74 yr resulted in total soil Zn contents of 3.8 to 460 mmol kg–1. In acidic soils (n = 24; pH <6.0), Zn was mainly found in the mobile fraction (F1) and the last two fractions (F6 and F7). In neutral soils (n = 25; pH ≥6.0), most Zn was extracted in the mobilizable fraction (F2) and the intermediate fractions (F4 and F5). The extractability of Zn increased with increasing Zn contamination of the soils. The sum of mobile (F1) and mobilizable (F2) Zn was independent of soil pH, the ratio of Zn in F1 over F1+F2 plotted against soil pH, exhibited the typical shape of a pH sorption edge and markedly increased from pH 6 to pH 5, reflecting the increasing lability of mobilizable Zn with decreasing soil pH. In conclusion, the extractability of Zn from soils contaminated with aqueous Zn after decades of aging under field conditions systematically varied with soil pH and Zn content. The same trends are expected to apply to aqueous Zn released from decomposing Zn-bearing contaminants, such as sewage sludge or smelter slag. The systematic trends in Zn fractionation with varying soil pH and Zn content indicate the paramount effect of these two factors on molecular scale Zn speciation. Further research is required to characterize the link between the fractionation and speciation of Zn and to determine how Zn loading and soil physicochemical properties affect Zn speciation in soils.

Abbreviations: ECEC, effective cation exchange capacity • SEP, sequential extraction procedures • SSR, solution-to-soil ratio • XRF, X-ray fluorescence • Zn-HIM, Zn bound in the Al-hydroxy interlayers of clay minerals


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
Materials and Methods
 Results and Discussion
 Conclusions
 REFERENCES
 
SOIL contamination with zinc (Zn) may result from the application of sewage sludge or fertilizers and from the emissions of mining, smelting, industry, and traffic. At elevated concentrations, Zn is toxic to soil microorganisms and plants and may adversely affect soil fertility and crop yield. The bioavailability of Zn is strongly affected by its speciation (i.e., the chemical forms in which Zn is present in soil). In acidic soils, Zn solubility is mostly controlled by cation exchange and adsorption processes on clay minerals and soil organic matter, and the bioavailability of Zn is relatively high. In near-neutral to alkaline soils, specific adsorption reactions on soil organic matter, clay minerals, Fe- and Al-oxides, and precipitation processes lower the solubility and bioavailability of Zn (Adriano, 2001; Alloway, 1995).

Sequential extraction procedures (SEPs) are widely used to characterize the fractionation of Zn and other trace elements in soils and sediments (Ahnstrom and Parker, 1999; Gleyzes et al., 2002; Shuman, 1985; Tessier et al., 1979; Young et al., 2006; Zeien and Brümmer, 1989). Fractionation data obtained from SEPs are often interpreted in terms of Zn speciation ("carbonate bound," "organically bound," "bound to Mn-oxides," etc.). However, limited selectivity of the reagents as well as readsorption and reprecipitation processes during sequential extraction limit the validity of such interpretations (Gleyzes et al., 2002; Young et al., 2006). Furthermore, a number of Zn species not included in classical interpretations of sequential extraction schemes were recently shown to be quantitatively relevant in contaminated soils, such as Zn-bearing layered double hydroxides, phyllosilicates, and Al-hydroxy interlayered clay minerals (Juillot et al., 2003; Manceau et al., 2000; Manceau et al., 2005; Scheinost et al., 2002; Voegelin et al., 2005). Nevertheless, single and sequential extractions represent an economic, efficient, and important means to characterize the lability and fractionation of Zn in contaminated soils and to study their changes in response to experimental treatments or changing environmental factors (Nolan et al., 2003; Voegelin et al., 2003). Even though Zn fractionation data cannot unequivocally be interpreted in terms of Zn speciation, changes or differences in Zn fractionation clearly indicate changes or differences in Zn speciation. Furthermore, single salt extractions are also part of many national regulations on permissible levels of trace metals in soils (McLaughlin et al., 2000; Young et al., 2006).

Studies on the adsorption of Zn to whole soils and on the solubility and extractability of Zn in pristine and contaminated soils showed that soil pH is the most important factor affecting Zn reactivity in soils. With increasing soil pH, Zn adsorption at constant Zn concentration in solution increases and Zn solubility and extractability at constant soil Zn content decrease (Elzinga et al., 1999; Hornburg et al., 1995; McBride et al., 1997; Voegelin and Kretzschmar, 2003). Statistical analyses further suggested that soil organic matter and Zn loading are major factors affecting Zn fractionation and solubility in contaminated soils (Hornburg et al., 1995; McBride et al., 1997; Tye et al., 2003). Because slow "aging" reactions, such as precipitate formation and ripening or intraparticle diffusion, may gradually reduce the solubility of Zn in soils contaminated with Zn in readily available form, long-term Zn solubility and bioavailability may be overestimated by short-term laboratory fractionation studies with soils spiked with dissolved Zn2+ (Smolders et al., 2003; Zhang et al., 2004). In studies with field soils contaminated over decades, on the other hand, the slow release of Zn from primary Zn-bearing contaminants such as metal ores, slag, or sewage sludge may mask the effect of soil properties and Zn loading on Zn fractionation and reactivity.

Around galvanized power line towers, corrosion of protective Zn coatings leads to localized soil contamination by input of dissolved Zn from runoff water (Jones and Burgess, 1984; Karlén et al., 2001). Therefore, soils contaminated from the runoff of galvanized electricity towers represent an ideal system to study the influence of soil properties and Zn content on the fractionation and extractability of Zn in soils, avoiding effects caused by short incubation times or Zn-bearing contaminants with slow dissolution kinetics. The objectives of this study therefore were (i) to investigate the fractionation of Zn in a wide range of soils contaminated from the runoff of galvanized power line towers using the seven-step SEP from Zeien and Brümmer (1989), (ii) to analyze the influence of soil physicochemical parameters and Zn content on Zn fractionation by multiple linear regressions, and (iii) to study the relation between different extracts used for the characterization of the labile Zn pool in soils.


   Materials and Methods
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results and Discussion
 Conclusions
 REFERENCES
 
Soil Sampling and Characterization
We collected topsoil material (0–5 cm) close to the foundations of 49 power line pylons (26–74 yr old) across different geologic and climatic regions of Switzerland. The soil samples were air-dried at 25°C and sieved to <2 mm. The dry samples were stored in plastic containers in the dark at room temperature.

Soil pH was determined by suspending 1 g of soil in 10 mL of 0.01 mol L–1 CaCl2 solution. Suspensions were shaken for 10 min and allowed to settle for 30 min before pH measurements were taken with a glass electrode. The total inorganic carbon content of soils with pH >5.9 was determined by reacting 0.3 to 0.9 g of soil material with 1 mol L–1 sulfuric acid (H2SO4) in a reaction flask. The evolving CO2 was adsorbed in a Nesbitt bulb containing NaOH-coated sorbent and was quantified gravimetrically. The analysis was performed in duplicate or triplicate. The total carbon content of the soils was determined on ball-milled samples (<50 µm) using a CHNS elemental analyzer (CHNS-932; LECO, St. Joseph, MI). The samples were incinerated at 730°C, and the CO2 was measured using infrared absorption spectrometry. Total organic carbon content was determined by subtracting the total inorganic C content from the total carbon content. After pretreatment of the soils with H2O2 for the removal of organic matter, the texture of the soil samples was determined by the pipette method (Gee and Or, 2002).

The contents of exchangeable Ca2+, Mg2+, K+, Na+, Al3+, and Mn2+ were determined in duplicate by extracting 7 g of soil in 210 mL of 0.1 mol L–1 BaCl2 for 2 h (Hendershot and Duquette, 1986). The effective cation exchange capacity (ECEC) was calculated from the charge equivalents of the extracted cations. Extractable Zn2+ was also analyzed in 0.01 mol L–1 CaCl2 equilibrated for 24 h (1 g of soil in 50 mL of solution, triplicates). For the quantification of Ca, Mg, K, Na, Al, Fe, Mn, and Zn, we used an inductively coupled plasma–optical emission spectrometer (Liberty 200; Varian, Palo Alto, CA) equipped with an ultrasonic nebulizer. The extracts were measured in 10- and 100-fold dilution depending on element concentration and the calibration range. For all analyses, standards were prepared in the background electrolyte of the (diluted) extracts to minimize matrix effects. Standards were run at least every 20 samples, and blanks were measured at regular intervals.

For the analysis of the total element contents, 4 g of soil were ball-milled to <50 µm, mixed with 0.9 g of wax, homogenized, and pressed into pellets. The pellets were analyzed using an energy-dispersive X-ray fluorescence (XRF) spectrometer (Spectro X-Lab 2000; Spectro, Kleve, Germany).

Sequential Extraction Procedure
The soils were sequentially extracted following the extraction scheme of Zeien and Brümmer (Zeien and Brümmer, 1989). This SEP consists of seven extraction steps. The extractants used for each step and the hypothetical interpretation of the seven fractions are listed in Table 1 . For the extraction, 2 g of soil were weighed into 50-mL centrifuge tubes. Extractions were performed on an overhead shaker unless stated otherwise. After each extraction, the samples were centrifuged at 3500 g (15 min at 20°C), and the supernatant was removed, filtered (0.45-µm nylon filters), and acidified (1% v/v concentrated HNO3). For fractions 2 to 6, the extraction step was followed by a washing step to collect the remaining extractant solution. The extractant and washing solution were combined for analysis. The pH of the extractant solutions was adjusted using dilute NH3 or HCl. All soils were extracted in duplicate or triplicate. One blank was extracted with each batch of soil samples. Briefly, the extraction sequence was as follows (extractant composition, solution-to-soil ratio [SSR] in mL g–1, extraction time): Fraction 1 (F1): 1 mol L–1 NH4NO3 (SSR = 25; time = 24 h); Fraction 2 (F2): 1 mol L–1 NH4–acetate, pH 6.0 (extracts from calcareous soils were titrated to pH 6.0 using 0.1 mol L–1 HCl) (SSR = 25; time = 24 h), washing with 1 mol L–1 NH4NO3 (SSR = 12.5; time = 10 min); Fraction 3 (F3): 0.1 mol L–1 NH2OH-HCl + 1 mol L–1 NH4–acetate, pH 6.0 (SSR = 25; time = 30 min), washing twice with 1 mol L–1 NH4–acetate, pH 6.0 (SSR = 12.5; time = 10 min); Fraction 4 (F4): 0.025 mol L–1 NH4–EDTA, pH 4.6 (SSR = 25; time = 90 min), washing with 1 mol L–1 NH4–acetate, pH 4.6 (SSR = 12.5; time = 10 min); Fraction 5 (F5): 0.2 mol L–1 NH4–oxalate, pH 3.25 (SSR = 25; time = 2 h, dark), washing with 0.2 mol L–1 NH4–oxalate, pH 3.25 (SSR = 12.5; time = 10 min, dark); Fraction 6 (F6): 0.1 mol L–1 ascorbic acid + 0.2 mol L–1 NH4–oxalate, pH 3.25 (SSR = 25; time = 2 h, in water bath at 96°C), washing with 0.2 mol L–1 NH4–oxalate, pH 3.25 (SSR = 12.5; time = 10 min, dark). The solutions from the extraction steps F1 to F6 were analyzed by inductively coupled plasma–optical emission spectrometry (see section on soil characterization for details). Fraction 7 (F7): The residual fraction was determined by XRF (see section on soil characterization for details). To have sufficient material for the preparation of wax pellets, the residual soil material from the duplicate or triplicate extractions of individual soils was combined for XRF analysis. The relative deviation of the total Zn extracted in the sequential extraction procedure from the total Zn content determined by XRF was usually <10%, except for GER (–10.1%), DUR (–10.2%), TRE (+11.2%), GLO (+14.9%), LAUS (+16.9%), and TAL (+19%). The average of the relative deviations was –0.4%, and the average of the absolute relative deviations was 5.6%. Thus, the deviation of the total Zn extracted by SEP from the total Zn determined by XRF was in general small and was not systematically biased to higher or lower Zn amounts.


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

  		Table 1. Sequential extraction procedure according to Zeien and Brümmer (1989). Extractants used and hypothetical interpretation of the seven fractions. The last column reports the parameters tested by multiple linear regression.

 
Multiple Linear Regression Analysis
The sequentially extracted fractions of Zn were analyzed by multiple linear regression, taking soil chemical parameters and total and extractable element contents into consideration. All data except soil pH were log transformed for statistical analysis. For all seven fractions, linear regression was first performed using soil pH and total Zn content as the only parameters. Subsequently, additional parameters were tested for each fraction. These parameters were selected based on the hypothetical interpretations of the individual fractions (Table 1). Additional parameters were considered to significantly improve the fit if their addition led to an increase in the adjusted r2. Only parameters with p < 0.05 were considered. Further linear regressions were calculated for the amounts of Zn extracted with 1 mol L–1 NH4NO3 (first fraction of the SEP), 0.1 mol L–1 BaCl2, and 0.01 mol L–1 CaCl2. All statistical calculations were performed with Systat Version 11 (Systat Software Inc., San Jose, CA).

Calculation of Zinc Speciation in 1 mol L–1 NH4NO3, 0.01 mol L–1 CaCl2, and 0.1 mol L–1 BaCl2
The speciation of 0.5 mmol L–1 total dissolved Zn in 1 mol L–1 NH4NO3, 0.01 mol L–1 CaCl2, and 0.1 mol L–1 BaCl2 at different pH values was calculated using PhreeqC Version 2 (Parkhurst and Appelo, 1999). Activity coefficients were obtained form the Davies equation. Speciation calculations included the formation of ZnClx2–x and Zn(OH)x2–x complexes (x = 1–4) with stability constants from the PhreeqC database and the formation of Zn(NH3)x2+ complexes (x = 1 to 4) with stability constants from the NIST 46 database (Martell et al., 1997). For the deprotonation of ammonium (pK = 9.25) and hydrolysis of Ca2+ and Ba2+, stability constants from the PhreeqC database were used.


   Results and Discussion
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
Conclusions
 REFERENCES
 
Soil Properties and Zinc Contents
For data interpretation, the soils were split in two groups according to their pH value: "acidic soils" with pH <6.0 and "neutral soils" with pH ≥6.0. Selected soil physicochemical parameters of acidic and neutral soils sorted by increasing total Zn contents are provided in Tables 2 and 3 , respectively. The soils in both groups covered a large range of soil properties (pH, clay and organic carbon content, texture) and had developed from a wide variety of geologic parent materials (calcareous and noncalcareous alluvium, limestone, sandstone, dolomite, calcareous and noncalcareous conglomerate, glacial till, paragneiss, orthogneiss, and granite) under different climatic conditions (altitude between 300 and 2200 m, all expositions). Fifteen of the 25 neutral soils contained ≥2.4 g kg–1 inorganic C (i.e., ≥2% [w/w] CaCO3 if inorganic C mainly CaCO3), characterizing them as calcaric materials according to the World Reference Base for Soil Resources (FAO, 2006). The Zn contents of all soils exceeded the upper limit of normal geogenic background concentrations of Zn in Swiss soils, which ranges from 0.9 mmol kg–1 Zn (57 mg kg–1) in acidic soils to 2.0 mmol kg–1 Zn (132 mg kg–1) in neutral soils (Keller and Desaules, 2001). According to the Swiss ordinance relating to impacts on the soil (VBBo, 1998), soils with Zn contents >30.6 mmol kg–1 (2000 mg kg–1) are considered to be heavily contaminated, which applied to 13 out of 25 acidic and 16 out of 24 neutral soils. Most soil samples for this study were collected next to the foundations of the power line towers at locations where Zn levels were expected to be highest. Therefore, the Zn contents reported in this work do not allow judging the extent of Zn contamination around power line towers.


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

  		Table 2. Characterization of acidic soils with pH <6.0 (n = 24). Subscripts for Zn, Mn, Al, and Fe indicate the method, extraction, or fraction of the sequential extraction procedures by which the reported concentrations were determined.

 

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

  		Table 3. Characterization of neutral soils with pH ≥6.0 (n = 25). Subscripts for Zn, Mn, Al, and Fe indicate the method, extraction, or fraction of the sequential extraction procedures by which the reported concentrations were determined.

 
Statistical information on the physicochemical properties and Zn contents of the acidic soils, the neutral soils, and all soil samples is provided in Table 4 . Both soil groups exhibited similar variation in the contents of organic C and clay. The median and mean of the ECEC of neutral soils were larger than of acidic soils, reflecting the pH dependence of the ECEC (Curtin and Rostad, 1997). In Fig. 1 , the total Zn contents of all soils are plotted against their soil pH values. Total Zn contents showed considerable variation even within narrow pH ranges. This variation was likely related to differences in Zn input, which was controlled by the type of power line tower (e.g., age, thickness and corrodibility of Zn coating, application of protective paint) and climatic factors (e.g., temperature, precipitation) (Odnevall Wallinder et al., 2000; Odnevall Wallinder et al., 2001). On the other hand, Zn retention in soils is controlled by soil properties, most importantly soil pH (Alloway, 1995). Therefore, the increase in Zn contents with increasing soil pH within the group of acidic soils was likely due to increased Zn retention in soil. Within the group of neutral soils, no clear trend with soil pH was observed. This suggests that the soil Zn contents in neutral soils with high Zn retention capacity were mainly related to differences in Zn input. Higher Zn retention at higher soil pH also explains the higher average and 95th percentile of Zn contents of neutral than of acidic soils (Table 2).


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

  		Table 4. Statistical characterization of the physicochemical properties and Zn contents of acidic soils (pH <6.0), neutral soils (pH ≥6.0), and all soils.

 

Figure 1
View larger version (16K):
[in this window]
[in a new window]

  		Fig. 1. Total Zn contents versus soil pH for acidic (pH <6; Table 2) and neutral (pH ≥6; Table 3) soils (separated by vertical dashed line). Soils with total Zn contents of <23 mmol kg–1, 23 to 78 mmol kg–1, and >78 mmol kg–1 are separated by horizontal dashed lines.

 
Effect of Soil pH and Total Zinc Content on Zinc Fractionation in the SEP
The total amounts and relative fractions of Zn sequentially extracted from acidic and neutral soils are shown in Fig. 2 . The soils are arranged by increasing total Zn content from left to right. In both sets of soils, the percentage of readily extractable forms of Zn tended to increase with increasing total Zn content. Comparing the two groups of soils, it is evident that fraction F1 was more relevant in soils with lower pH and fraction F2 in soils with higher pH. In Fig. 3 , the relative amounts of Zn extracted from acidic and neutral soils are summarized in box plots. These plots confirmed the tendency that soils with lower pH contained a higher fraction of Zn in F1, whereas soils with higher pH contained a larger fraction in F2. In addition, the box plots indicated a tendency for Zn to be extracted in the fractions F6 and F7 at lower soil pH and in the fractions F4 and F5 at higher soil pH. Thus, from Fig. 2 and 3, we concluded that soil pH and total Zn content were the two major factors affecting the fractionation of Zn as determined by the SEP.


Figure 2
View larger version (76K):
[in this window]
[in a new window]

  		Fig. 2. Results from the sequential extraction of acidic soils (A, C) and neutral soils (B, D). Absolute amounts are shown in the upper panels (A, B) and relative amounts (normalized by the sum of extracted Zn) in the lower panels (C, D). Soils are arranged by increasing total Zn content (X-ray fluorescence analyses) from left to right. The hypothetical interpretation of the sequential extraction is provided in Table 1.

 

Figure 3
View larger version (21K):
[in this window]
[in a new window]

  		Fig. 3. Summarizing comparison of the sequentially extracted fractions of Zn in acidic soils (open boxes) and neutral soils (hatched boxes). Boxes cover the 25th to 75th percentile range (including the median [lines dividing boxes] and the arithmetic mean [squares]), bars cover the 5th to 95th percentile range, and crosses indicate the maximum and minimum values. The hypothetical interpretation of the extracted fractions is provided in Table 1.

 
To further explore the effect of soil pH and total Zn content on the fractionation of Zn, the sequential extraction data were analyzed by multiple linear regressions. For these analyses, all data except soil pH were log transformed to obtain normally distributed data. The respective regression equations 1 to 7 are provided in Table 5 . Soil pH and total Zn content were significant parameters for all fractions. The sign of the coefficients for soil pH reflected the trends observed in Fig. 3, i.e., a tendency for increasing extraction of Zn with the fractions F1, F6, and F7 at lower soil pH)(Table 5, Eq. 1, 6, 7), and vice versa for increasing extraction of Zn with the fractions F2, F3, F4, and F5 at higher soil pH (Table 5, Eq. 2–5). In acidic topsoils, leaching of Zn from the mobile fraction F1 likely caused the trend toward lower total Zn contents and higher percentages of Zn in the two most strongly retained fractions (F6 and F7). In neutral and calcareous topsoils, retention of Zn in the mobilizable fraction F2 and intermediate fractions F4 and F5 may effectively prevent Zn leaching to deeper soil horizons, resulting in higher accumulated levels of Zn than in more acidic topsoils. The coefficient B related to total Zn content steadily decreased from 1.512 in F1 to 0.332 in F7 (Table 5). This showed that increasing amounts of Zn introduced into the soils increasingly accumulated in the more readily extractable fractions. Because geogenic Zn is extracted in later extraction steps (Wilcke et al., 2005), the shift toward more readily extractable Zn with increasing Zn loading may have partly been caused by the decreasing influence of geogenic Zn on fractionation results. However, most soils were contaminated with Zn at concentrations far above the upper limits of normal geogenic Zn contents in Swiss soils (ranging from 0.9 mmol kg–1 Zn in acidic to 2.0 mmol kg–1 Zn in neutral soils [Keller and Desaules, 2001]). Linear regressions restricted to soils containing at least 20 mmol kg–1 Zn (i.e., at least 10 times above the upper limit of normal geogenic Zn contents, 14 of 24 acidic and 21 of 25 neutral soils) qualitatively confirmed that increasing Zn loading resulted in increasing Zn fractionation into earlier extraction steps (regressions including soil pH and total Zn content; coefficients B for total Zn content of 1.41, 1.12, 1.07, 1.19, 0.79, 0.35, and 0.18 for F1 to F7, respectively). Thus, although geogenic Zn may have affected Zn fractionation in the soils with the lowest contamination levels, the trends in Zn fractionation revealed by multiple linear regression analysis including all soils mainly described the fractionation behavior of Zn introduced with the runoff from power line towers.


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

  		Table 5. Multiple linear regression analysis of the amounts of Zn extracted in the fractions F1 to F7 (ZnFi) of all studied soils (n = 49).{dagger}

 
Linear regression analysis was also performed on the log-transformed percentages of extracted Zn. For this purpose, the regression equation was adapted by replacing the coefficient B assigned to the total Zn content by (B – 1). The respective linear regression equations for percentages of extracted Zn resulted in the same values for the coefficients A, B, and C; their standard errors; and the standard error of estimates as reported for linear regression equations based on absolute amounts of extracted Zn (Table 5). However, the adjusted r2 of regression equations based on percentages of extracted Zn were in general considerably lower than for respective equations based on absolute fractions (Table 5), reflecting the inherent correlation between absolute Zn fractions and total Zn content. Despite this correlation, including the total Zn content significantly improved the regression equations for the percentages of extracted Zn for all fractions except F5 (for which the lowest adjusted r2 was obtained), confirming that total Zn content not only affected the amounts of Zn in individual extraction steps, but also the distribution of Zn over the individual fractions.

Effect of Soil Physicochemical Parameters on Zinc Fractionation in the SEP
In a second step, the effects of additional parameters on the quality of the regressions were tested (Table 5, Eq. [8–13]). The additional parameters were selected based on the hypothetical interpretation of the sequentially extracted Zn fractions (Table 1). Note that the hypothetical interpretations for the fractions F1, F2, and F7 are rather operational (exchangeable, mobilizable, and residual, respectively; Table 1), whereas the fractions F3, F4, F5, and F6 are hypothesized to correspond to Zn bound to different types of sorbent phases (Mn-oxides, organic substances, weakly crystalline Fe-oxides, and crystalline Fe-oxides, respectively; Table 1). Regarding the interpretation of the regression results, the extractants vary in the mode by which Zn is extracted from soil. In the fraction F1, the extractant releases Zn by competitive adsorption and exchange mechanisms, whereas in later extraction steps, the extractant mainly serves to dissolve the target sorbent phase (F2, F3, F5, F6) and/or to lower the free Zn activity in solution by complexation (F2, F4).

Because fraction F1 was considered to represent exchangeably adsorbed Zn, parameters tested were clay content and organic C content as major factors controlling the concentration of exchange sites and the experimentally determined ECEC. From the tested parameters, only the organic C content led to a significant improvement of the linear regression (Table 5, Eq. [8]). The negative coefficient for organic C content suggested that increasing organic C contents mainly caused stronger Zn retention rather than a higher fraction of exchangeably bound Zn. According to Zeien and Brümmer (1989), the fraction F2 corresponds to specifically adsorbed and carbonate-bound Zn (Table 1). Therefore, we tested the amount of Ca extracted in fraction F2 (as an indirect measure of CaCO3), the content of inorganic C (only for soils with pH ≥6.0), the content of organic C, the clay content, and the ECEC as additional regression parameters. The only parameter resulting in a higher adjusted r2 was the ECEC (Table 5, Eq. [9]). Considering that the ECEC increases with increasing soil clay content, soil organic carbon content, and soil pH (Meyer et al., 1994), the positive coefficient found for the ECEC might indicate that the fraction of Zn in F2 increased as the potential for specific adsorption of Zn to organic matter and clay increased with increasing pH.

For the fractions F3 ("Mn-oxide bound"), F4 ("organically bound"), and F5 ("bound to amorphous Fe-oxides"), the linear regressions were improved by the addition of parameters representing the target sorbent phase (Mn extracted in F3, organic C content, and Fe extracted in F5, respectively; Table 5, Eq. [10–12]). Because the coefficients assigned to the respective parameters (Mn in F3, organic C in F4, Fe in F5) were all positive (Table 5, Eq. [10–12]), the regressions seemingly supported the hypothetical interpretation of the different fractions. However, in 27 of the 49 studied soils, the molar ratio of Zn over Mn extracted in F3 was greater than unity, indicating that Mn-oxides could not be the main sorbent from which Zn was released. For F4 and F5, the corresponding molar ratios (Zn in F4 over organic C; Zn in F5 over Fe in F5) did not allow judging the validity of the hypothetical interpretation because all soils contained much more organic C and oxalate-extractable Fe than Zn. However, the results for Zn in F3 demonstrated that improvements in the linear regressions for the fractions F3, F4, and F5 by adding an indicator for the hypothetic target phase must be interpreted with care. The results may suggest that some of the extracted Zn has been associated with the hypothesized target phase, but not necessarily the largest fraction.

The extractants used for F5 and F6 ("weakly crystalline" and "crystalline Fe-oxides," Table 1) were considered to dissolve amorphous and crystalline Al-oxides, respectively (McKeague and Day, 1966). Therefore, parameters tested for the fractions F5 and F6 included the amount of Fe, Al, and Fe+Al extracted in the respective extraction step. In the case of the fraction F5 discussed previously, only the extracted Fe led to an improvement in the linear regression, but not the extracted Al or the sum of extracted Fe and Al (Table 5, Eq. [12]). In contrast, and in contradiction to the hypothetical interpretation of F6, only the amount of extractable Al in F6 (but not the amount of Fe or the sum of Fe and Al) led to a substantial improvement of the linear regression (Table 5, Eq. [13]). Based on X-ray absorption spectroscopy, Zn bound in the Al-hydroxy interlayers of clay minerals (Zn-HIM) was recently postulated to be a major Zn species in clayey acidic soils (Manceau et al., 2005). Using the same SEP as in the present study, Zn-HIM was shown to be mainly extracted in the fractions F6 and F7 (Scheinost et al., 2002). Thus, the improvements in the linear regression for F6 achieved by the amount of Al extracted in this fraction might be due to the extraction of Zn-HIM.

Characterization of Mobile and Mobilizable Zinc
For the assessment of risks associated with trace metals in soils, knowledge on the mobilizable metal pool (quantity) and the mobile fraction (intensity) are of primary importance (McLaughlin et al., 2000). Extractants like EDTA or DTPA are considered to probe the mobilizable metal fraction, whereas unbuffered salt extracts (NH4NO3, CaCl2, BaCl2) are interpreted as indicators for the mobile fraction (Gupta et al., 1996; Hornburg and Brümmer, 1993; McLaughlin et al., 2000). In two previous studies, we found that the sum of Zn contained in the first two fractions of the SEP used here represented a good estimate for the pool of Zn mobilizable via competitive cation adsorption reactions and that the ratio of Zn in F1 over the sum extracted in F1+F2 indicated the lability of that pool toward mobilization by competitive sorption processes (Voegelin and Kretzschmar, 2003; Voegelin et al., 2003). In Fig. 4A , the sum of Zn extracted in F1 and F2 is plotted against the total Zn content of the soils. Comparison of the acidic and neutral soils revealed no systematic differences. Likewise, the linear regression of F1+F2 over total Zn (Table 6 , Eq. [14]) was not significantly improved if soil pH was added, as judged by radj2 (not shown). This suggested that the increase of the fraction F1 and the simultaneous decrease of the fraction F2 with decreasing soil pH (Table 5) nearly cancelled out in the sum of the two fractions. The coefficient B > 1 of Eq. [14] (Table 6) further indicated that not only the absolute amount but also the percentage of Zn extracted in F1+F2 increased with increasing total Zn content of the soils, as observed in Fig. 2C and 2D.


Figure 4
View larger version (31K):
[in this window]
[in a new window]

  		Fig. 4. (A) Sum of Zn extracted in the fractions F1 and F2 versus total Zn (linear regression Eq. [16] from Table 6). (B) Ratio of Zn in F1 over sum of Zn extracted in F1 and F2 (line: F1/(F1+F2) calculated using linear regression Eq. [17] from Table 6). (C) Ratio of Zn extracted by 0.01 mol L–1 CaCl2 (SSR = 50 L kg–1) over Zn in F1 (1 mol L–1 NH4NO3; SSR = 25 L kg–1). (D) Ratio of Zn extracted by 0.1 mol L–1 BaCl2 (SSR = 30 L kg–1) over Zn in F1 (1 mol L–1 NH4NO3; SSR = 25 L kg–1).

 

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

  		Table 6. Multiple linear regression analyses for mobilizable and mobile metal fractions.{dagger}

 
In Fig. 4B, the ratio of Zn in F1 over Zn in F1+F2 is plotted against soil pH. The plot exhibits the typical shape of a desorption pH-edge with a marked decrease between pH 5 to 6. This indicated that the lability of mobilizable Zn decreased with increasing soil pH. The observed trend compared with the results from an extraction study on 158 soils from northern Germany, in which exchangeable Zn was observed to markedly increase at soil pH below 5.3 (Hornburg and Brümmer, 1993). However, the position of the pH-edge observed in Fig. 4B depends on the extractant used to determine mobile/exchangeable Zn in F1 (SSR and extractant strength). The observed increase in Zn lability with decreasing soil pH is in line with the relatively low total Zn contents of the most acidic soils (Fig. 1), supporting our earlier interpretation that total Zn contents were determined by Zn input with runoff water and extent of Zn leaching from soils with low retention capacity. Interpreting F1 as the mobile Zn fraction in an adsorption/cation exchange equilibrium with the remaining mobilizable Zn in F2, F1 was expressed as a linear regression over F2 and soil pH (Table 6, Eq. [15]). This resulted in a markedly higher radj2 than obtained for the regression based on total Zn content (Table 5, Eq. [1]). To avoid data where fraction F1 may have been overestimated due to Zn complexation by NH3 (see next paragraph), the same regression was also calculated for all soils with pH <6.7 (n = 42), which further improved the regression (Table 6, Eq. [16]). The coefficient B of Eq. [16] was near unity, indicating a nearly constant ratio between Zn in F1 and F2 at constant pH. Therefore, Eq. [16] was further simplified by fixing the coefficient for the fraction F2 to unity (Table 6, Eq. [17]), which is equivalent to calculating the regression equation for log(F1/F2) as a function of pH. The radj2 of Eq. [17] was nearly identical to the one of Eq. [16], indicating that an additional coefficient for logF2 was not required. Based on Eq. [17], the line representing F1/(F1+F2) in Fig. 4B was calculated. Considering the SSR of 25 L kg–1 used to extract the fraction F1, Eq. [17] was reformulated as a pH-dependent Kd value linking the concentration in the fraction F1 to the amount of the remaining mobilizable Zn in F2:

Formula 18[18]
The coefficient 0.827 describing the pH dependence of the logKd closely compares to a coefficient of 0.792 reported from a short-term laboratory adsorption study with 38 soils (Anderson and Christensen, 1988). Although the Kd calculated from the fractions F1 and F2 is independent of the SSR (25 mL g–1) used to determine the fraction F1, it is still conditional for the extractant used (1 mol L–1 NH4NO3) and thus cannot be used to estimate Zn concentrations in real soil solution. However, its pH dependence clearly indicated that the sorption affinity of mobilizable Zn decreased with decreasing soil pH.

The ratios of Zn extracted with CaCl2 (0.01 mol L–1; SSR 50 L kg–1) and BaCl2 (0.1 mol L–1; SSR 30 L kg–1) over Zn extracted by NH4NO3 (1 mol L–1; SSR 25 L kg–1) as a function of soil pH are plotted in Fig. 4C and 4D, respectively. At low pH values, less Zn was extracted with 0.01 mol L–1 CaCl2 than with 1 mol L–1 NH4NO3; this result is attributable to the much lower CaCl2 concentration. In contrast, the amounts of Zn extracted with 0.1 mol L–1 BaCl2 were similar to those extracted with 1 mol L–1 NH4NO3, indicating the higher extraction efficiency of bivalent Ba2+ than monovalent NH4+. With soil pH increasing from 4 to 6.7, CaCl2–extractable and especially BaCl2–extractable Zn slightly increased relative to NH4NO3–extractable Zn. These trends were also reflected in the respective regression equations [19] and [20] (Table 6), suggesting an increase of the relative extraction efficiency of the bivalent cations Ca2+ and Ba2+ over the monovalent cation NH4+ as soil pH values increased and Zn adsorption shifted from cation exchange to specific adsorption reactions. In calcareous soils, Ca2+ and Ba2+ may also be more effective than NH4+ in mobilizing Zn adsorbed or co-precipitated on calcite surfaces. At pH values >6.7, the fractions of CaCl2–extractable and BaCl2–extractable over NH4NO3–extractable Zn decreased abruptly (Fig. 4C and 4D). At the same time, some of the samples with pH >6.7 also exhibited a rather high F1/(F1+F2) ratio (Fig. 4B). These trends were attributable to increasing Zn complexation by NH3 at increasing solution pH in the 1 mol L–1 NH4NO3 extracts (Lebourg et al., 1998). Speciation calculations listed in Table 7 demonstrate that the fraction of Zn in Zn(NH3)x2+ complexes in NH4NO3 extracts strongly increases between pH 6.5 and 7. At pH 8, almost all Zn in 1 mol L–1 NH4NO3 is complexed by NH3. In contrast, Zn speciation in 0.1 mol L–1 BaCl2 extracts is dominated by free Zn2+ and to a lesser extent by pH-independent complex formation with Cl–. In 0.01 mol L–1 CaCl2 extracts, Cl– complexation is almost negligible due to the lower Cl– concentration and because most dissolved Zn is free Zn2+, with a minor fraction of hydrolyzed Zn at pH 8. Thus, BaCl2 or CaCl2 extracts are more appropriate than NH4NO3 extracts to characterize mobile Zn in soils with pH >6.7. Therefore, the regression equations [16–17] and [19–21] in Table 6 were limited to soil samples with pH <6.7. Regression Eq. [21] expresses the fraction of Zn extracted with CaCl2 as a function of soil pH and Zn extracted with BaCl2. This regression had a higher radj2 than the respective equation based on NH4NO3–extractable Zn (Eq. [19]), reflecting the relatively similar chemistry of CaCl2 and BaCl2 compared with NH4NO3.


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

  		Table 7. Speciation of 0.5 mmol L–1 Zn in unbuffered salt extracts using 1 mol L–1 NH4NO3, 0.1 mol L–1 BaCl2, and 0.01 mol L–1 CaCl2 as a function of solution pH. Numbers indicate fractions of total dissolved Zn.

 

   Conclusions
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions
REFERENCES
 
The results from single and sequential batch extractions exhibited systematic trends in Zn extractability with soil pH and total Zn loading. The interpretation of fractionation results in terms of molecular-scale Zn speciation is hampered by limited extractant selectivity, element redistribution during extraction (Gleyzes et al., 2002; Young et al., 2006), and the presence of Zn species not explicitly considered in classical interpretations of extraction schemes. However, assuming that operational Zn fractionation is related to chemical Zn speciation, the trends observed in single and sequential extractions allow us to conclude that soil pH and Zn loading are the dominant factors affecting Zn speciation. Improvements in individual regression equations achieved by additional parameters (organic carbon content [F1, F4], ECEC [F2], NH2OH-HCl–extractable Mn [F3], acid oxalate extractable Fe [F5], and ascorbic acid/oxalate extractable Al [F6]) further point to the influence of those parameters on Zn speciation. Future studies on the influence of soil properties and Zn content on the lability and speciation of Zn in contaminated soils based on of advanced analytical techniques such as isotope dilution or synchrotron spectroscopy need to take these factors into account.

Total, mobilizable, and mobile soil Zn contents were closely related to each other and reflected the paramount influence of soil pH on Zn retention and mobility. The percentage of mobilizable Zn (F1+F2) increased with total Zn content but did not depend on soil pH (Fig. 4A). This lack of pH dependence could be attributed to the counteracting effect of pH on the amount of Zn in F1 and F2. The ratio F1/(F1+F2), on the other hand, was highly pH dependent and substantially increased with decreasing pH (Fig. 4B, Eq. [17]), indicating that the lability of mobilizable Zn increased with decreasing pH. Increasing mobility and leaching of Zn from soils with decreasing soil pH in turn explained the trend toward lower soil Zn contents at lower soil pH values (Fig. 1).

The soils studied in this work were contaminated with aqueous Zn from the corrosion of galvanized power line towers and had been equilibrated over decades under field conditions. Similar trends in Zn fractionation are expected for soils contaminated with Zn in readily soluble form (e.g., ZnO [Voegelin et al., 2005]). In cases where soil contamination originates from sewage sludge, emissions of mining and smelting, or other sources with Zn in solid or complexed form, the chemical properties of the contaminants may determine Zn speciation and extractability over long periods. However, the slow decomposition or weathering of these primary contaminants results in the release of aqueous Zn, which is then expected to follow the fractionation trends described in the present study.


   ACKNOWLEDGMENTS
 
Stefan Egli and Christian Bitterli (both ETH Zurich) are acknowledged for their help with the characterization of the soil samples. This project was financially supported by the Swiss National Science Foundation under contracts no. 200021-101876 and 200020-116592.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and 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
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Voegelin, A.
Right arrow Articles by Kretzschmar, R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Voegelin, A.
Right arrow Articles by Kretzschmar, R.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Voegelin, A.
Right arrow Articles by Kretzschmar, R.
Related Collections
Right arrow Batch Studies
Right arrow Toxic Trace Metals
Right arrow Sorption/Exchange
Right arrow Heavy Metals
Right arrow Soil Pollution

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
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome