| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Plant, Soil, and Entomological Sci., Univ. of Idaho, P.O. Box 442339, Moscow, ID 83844-2339 USA
b Dep. of Plant and Soil Sciences, Univ. of Delaware, Newark, DE 19717-1303 USA
dgstrawn{at}uidaho.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMethods and materialsResults and discussionStirred-Flow ExperimentsSummaryREFERENCES |
|---|
,
) of Pb in a Matapeake silt loam soil (Typic Hapludult) were studied by stirred-flow and batch experiments. In addition, we studied the effects of soil organic matter (SOM) on sorption and desorption behavior by treating the soil with sodium hypochlorite to remove the SOM fraction, and using a soil with six times as much SOM (St. Johns loamy sand [Typic Haplaquods]) as the Matapeake soil. Lead sorption consisted of a fast initial reaction in which all of the Pb added to the stirred-flow chamber was sorbed. Following this initial fast reaction, sorption continued and appears to be rate limited (indicated by a decrease in the outflow concentration when the flow rate was decreased, or when the flow was stopped). The total amount of Pb sorbed was 102, 44, and 27 mmol kg-1 for the St. Johns soil and the untreated and treated Matapeake soils, respectively. Desorption experiments were conducted on the soils with the background electrolyte as the eluent in the stirred-flow chamber. In the St. Johns soil only, 32% of the total sorbed Pb was desorbed, while 47 and 76% of the sorbed Pb was released from the untreated and treated Matapeake soil, respectively. The correlation between SOM in the soils, and the percentage Pb desorbed from the soils suggests that SOM plays an important role in slow desorption reactions of Pb from soil materials. Aging experiments in which sorbed Pb was incubated for 1, 10, and 32 d showed that sorption incubation time had no effect on Pb desorption behavior. Analysis of the treated and untreated Matapeake soils by x-ray absorption fine structure (XAFS) spectroscopy revealed that the local atomic structure of sorbed Pb is distinctly different in the two samples. In the soil treated to remove SOM, the data were well represented by theoretical models using O, Si, and Pb backscattering atoms. In the untreated soil, the XAFS data were best described by O and C backscatterers. These XAFS results confirm that the sorption mechanisms in the two systems are different.
Abbreviations: CV, chamber volume • SOM, soil organic matter • XAFS, x-ray absorption fine structure
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMethods and materialsResults and discussionStirred-Flow ExperimentsSummaryREFERENCES |
|---|
Lead sorption behavior is often initially fast, followed by a slow reaction (Benjamin et al., 1981; Hayes et al., 1986; Strawn et al., 1998). The fast reaction is most likely adsorption via electrostatic attraction, and/or inner-sphere complexation with functional groups present on the soil components. There are several possible reasons for the slow sorption steps, such as: slow interparticle diffusion in porous minerals and organic matter, formation of precipitates on surfaces which can sometimes be slower than typical sorption, and adsorption onto sites that have relatively large activation energies (Fuller et al., 1993; Papelis et al., 1995; Loehr et al., 1996; Scheidegger et al., 1998). It is possible that multiple slow reaction mechanisms are responsible for the slow sorption reactions in soils.
There are three main processes that control the fate and bioavailability of metals in soils: (i) removal of metals from the soil solution by sorption onto soil particles, (ii) release of the metal from the soil particle to the soil solution (desorption), and (iii) precipitation–dissolution of the metal as an independent phase in the soil matrix. Less is known about the desorption behavior of metals from soils than the other two processes. This is unfortunate since once a soil is contaminated desorption is an important process that controls the bioavailability of the contaminant. If accurate assessments of the fate of metals in soils are to be gained, it is critical that desorption behavior be studied as well as sorption behavior.
It is often observed that desorption reactions are slower than sorption reactions. Smith and Comans (1996) predicted that adsorption half-lives for Cs sorption on sediments were between 50 and 125 d, while desorption half-lives were on the order of 10 yr. Failure to include the slow desorption reaction in transport models severely underestimated the remobilization potential of Cs. Slow desorption reactions have also been observed for metals (Kuo and Mikkelsen, 1980; McKenzie, 1980; Schultz et al., 1987; Ainsworth et al., 1994; Scheidegger et al., 1996). A possible reason that desorption reactions are slower than sorption reactions is because the sorbate undergoes a transformation from one state to another, for example: conversion from an adsorbed species to a surface precipitate. In addition, desorption reactions often require large activation energies, which cause the reaction to proceed slowly (McBride, 1994).
The rates of Pb sorption and desorption on mineral surfaces have been observed to be similar (Ainsworth et al., 1994; Gunneriusson et al., 1994; Strawn et al., 1998). However, in some cases desorption from soils is much slower than sorption. This could be due to the presence of SOM and/or the formation of Pb multinuclear complexes with carbonates, phosphates, and sulfates present in the soil. Bunzl et al. (1976) found that the desorption rate of Pb from SOM was significantly slower than the adsorption rate. Similar trends were found for Pb adsorption and desorption on activated carbon (Wilczak et al., 1993).
Soil organic matter is an important component of the soil since it has a high surface area, and has functional groups that are Lewis bases (e.g., carboxyl and phenol groups) that metals can form chemical bonds with (Sparks, 1995). It has been observed that Pb forms strong complexes with SOM, and that it can out compete most other metals for adsorption sites on SOM (Kendorff and Schnitzer, 1980; Elliot et al., 1986; Jin et al., 1996; Suave et al., 1998). Thus, it is important that a better understanding of Pb desorption behavior from the SOM be gained.
Kinetic studies are most often conducted by means of batch reactors. While valuable information has been obtained from these experiments, the batch technique is not ideal in desorption studies since readsorption of the sorbate and accumulation in solution can hinder desorption (Sparks, 1989). One of the most beneficial uses of stirred-flow reactors is to study desorption reactions since the desorbed products are continuously removed from solution. There are several other reasons that stirred-flow reactors are advantageous for studying sorption and desorption reactions: relatively fast reactions can be measured compared with batch methods; carbon dioxide can be easily removed from the system; sorption and desorption kinetics can be measured in a single experiment; and the system is well mixed, which aids in elimination of film diffusion and calculation of the dilution occurring in the reaction chamber (Sparks, 1989; Amacher, 1991; Yin et al., 1997).
The objectives of this study are to characterize sorption and desorption behavior of Pb on soil, and to determine the importance of SOM on sorption and desorption. To accomplish these objectives, we used both batch and stirred-flow studies and soils with three different amounts of SOM. Information obtained from this study will give insights to scientists and engineers that may lead to improved remediation strategies, disposal practices, and risk assessments. Such information is critical, since Pb is used in a variety of industrial and manufacturing processes and is one of the most common contaminants found at hazardous waste sites (Reed et al., 1996).
|
Methods and materials |
|---|
|
TOPABSTRACTINTRODUCTIONMethods and materialsResults and discussionStirred-Flow ExperimentsSummaryREFERENCES |
|---|
0–20 cm) of a Matapeake silt loam (Typic Hapludult) and a St. Johns loamy sand (Typic Haplaquods) from Delaware were used in this study. The physicochemical and mineralogical properties of the soils were determined by standard methods (Sparks, 1996), and are reported in Table 1
. Studies were also conducted on a Matapeake soil that was treated to remove organic matter by the following procedure: 30 g of soil were suspended in 50 mL of sodium hypochlorite (
) (Na-hypochlorite is an effective oxidant for SOM, and is less invasive than other SOM removal procedures [Lavkulich and Wiens, 1970]); the suspension was heated to 338 K and allowed to react for 20 min; the samples were then centrifuged, decanted, and treated with additional Na-hypochlorite. This procedure was repeated four times. Following oxidation of the organic matter the soil was washed repeatedly with a 1 M NaNO3 solution to Na saturate. The Na-saturated soil was then washed with deionized water to remove excess NaNO3. The resulting soil had 0.1% SOM remaining, as determined by the Walkley Black method (Nelson and Sommers, 1996). Analysis of the soils for total carbon by loss on emissions resulted in 1.38% for the untreated Matapeake soil, and 0.18% for the treated Matapeake soil.
|
All solutions were made with ACS reagent grade chemicals and distilled deionized water. Carbon dioxide was eliminated from all experiments by working under an N2 atmosphere. The temperature of all experiments was maintained at 298 K.
Batch Experiments
To gain insight into the overall sorption behavior, and make comparisons to the results from the stirred-flow experiments, batch isotherm and kinetic studies were conducted on the Matapeake soil. One gram of soil was preequilibrated for 24 h in background solution adjusted to
. The background solution consisted of a mixture of NaNO3 and 0.05 M MES (buffer, [2-(N-morpholino) ethane sulfonic acid]). The buffer was necessary to control pH during the adsorption experiments since initial experiments revealed that Pb sorption and desorption involve proton exchange reactions. Titration experiments with this buffer in solutions of Pb confirmed the results of others (Good et al., 1966; Baeyens and Bradbury, 1997); no detectable complexation reactions occur between Pb and the MES buffer. The concentration of NaNO3 used in the experiments was adjusted to make the total
(including the MES and the Pb solution).
In the isotherm experiment, the initial Pb concentrations ranged from 1 to 12 mM. Samples with no soil present were prepared by identical procedures as the sorption samples to determine the exact initial concentration of the sorption samples. The solid concentration of the samples was 100 g L-1. Following the addition of the Pb solution the pH of the samples was adjusted to
. After 24 h, the samples incubated at the highest concentrations required minor pH adjustments. The samples were incubated for a total of 48 h on an end-over-end shaker operating at 20 rpm. At the end of the incubation period, the samples were centrifuged at 7800 g for 10 min. The supernatant was filtered through a 0.2-µm filter and diluted 2:3 with 1 M HNO3. The solution was then analyzed for total Pb with an ICP spectrometer.
At the start of the kinetic experiment, 3.5 mL of a Pb stock solution
was added to 1.5 mL of suspension (0.50 g soil). The pH of the suspension was maintained at 5.50 throughout the experiment with the MES buffer. Minor pH adjustments were made throughout the experiment by addition of NaOH. Periodically (8 min–800 h) samples were removed from incubation and processed by the same procedures as mentioned above. The total amount of Pb adsorbed in the kinetic and the isotherm experiments was calculated from the difference between the initial and final Pb concentrations.
Stirred-Flow Experiments
The reaction chamber used for these experiments was similar in design to the reaction chamber used in the experiments of Bar-Tal et al. (1990). The setup of the stirred-flow experiment is illustrated in Fig. 1
. The pump used in these experiments was a piston displacement pump designed for use in an HPLC system. The experiment was started by placing 0.75 g of pretreated soil in the reaction chamber and filling the chamber with background electrolyte (
and
). A 25-mm-diam filter membrane with a 0.2-µm-pore size was used in the reaction chamber. Upon sealing the reaction chamber, CO2 free background electrolyte solution was flowed through the chamber for 10 min at 1 mL min-1 and the suspension was allowed to sit (no flow) for 20 min to preequilibrate. The reactor had an approximate volume of 8 mL. The suspension in the reaction chamber was stirred by a magnetic stir bar (12.7 mm long and 3 mm in diameter) at 400 rpm. To initiate the experiment a Pb solution (
,
) was pumped into the chamber. The effluent was collected with a fraction collector set to collect 2 mL of solution per tube. The fraction collector was started when the first drop of solution exited the outflow tube. The flow rate in the experiments was either 0.4 mL min-1 or 4 mL min-1. The flow rate was monitored and found to be stable within ±3% throughout the experiment.
|
To determine if the sorption and desorption reactions were rate limited, the flow was stopped during the experiment for 30 min. If the reaction were not at equilibrium when the flow was stopped, then sorption would continue (Bar-Tal et al., 1990). This would result in a drop in the [Pb] in the chamber, which can be detected in the outflow when the flow resumes.
The effects of aging on desorption were measured on the Matapeake soil by first adsorbing Pb by the batch incubation procedures described above, and then transferring the soil to the stirred-flow chamber for desorption. In these experiments, the initial sorptive in the chamber was a 2.32 mM Pb solution, which is the same concentration present in the chamber at the start of the desorption experiments conducted directly after the stirred-flow sorption experiments. The outflow from all of the experiments was acidified and analyzed with an ICP spectrometer for total Pb.
The data from the stirred-flow experiments were plotted as a function of chamber volumes (CV), which was calculated by multiplying the flow rate and the time and dividing by the volume of the chamber. This was useful for comparisons of data obtained from experiments conducted at different flow rates since it normalized the data relative to the flow rate. Thus, any differences between the data obtained at the different flow rates were the result of changes in the extent of the surface reactions.
To analyze accurately the results from stirred-flow experiments, it was necessary to calculate the effective volume of the chamber (Vce). To accomplish this, tracer experiments at several different flow rates and no soil present in the chamber were conducted. From these data Vce was calculated by fitting the outflow concentration using Eq. [1] (
) and subtracting the volume of the soil. The resulting Vce for
, 2.0, and 4.0 mL min-1 are 7.98, 8.10, 8.32 mL, respectively (a maximum difference of 4%). Use of the effective volume of the chamber (Vce) instead of the actual volume to calculate CV allowed data obtained at different flow rates to be compared. Thus, in all of the adsorption and desorption experiments presented in this study the flow rate dependent values of Vce were used to calculate CV.
Data Analysis
One of the limitations of stirred-flow analysis is that the amount of sorbate being sorbed at any given time cannot be measured directly. Instead, it is necessary to use models to derive the adsorption behavior. A common approach to modeling reactions in a stirred-flow system utilizes a mass balance equation (Bar-Tal et al., 1990). The mass balance equation for a stirred-flow reactor is as follows:
 | (1) |
Where Cchamber is the [Pb] in the chamber (mmol L-1), t is time (min), Vc is the volume of solution in the chamber (L), Cin is the [Pb] that is being pumped into the chamber (mmol L-1), Cout is the [Pb] leaving the chamber (mmol L-1), Q is the flow rate (L min-1), q is the amount of Pb sorbed (mmol kg-1), and m is the mass of soil (kg). Since in the stirred-flow chamber the solution is well mixed, the [Pb] in the chamber equals [Pb] leaving the chamber
. If there is no soil in the chamber, or there is no reaction occurring, then
, and the equation can be integrated to find Cout as a function of time, which is equivalent to the outflow concentration of a non-sorbing tracer.
Within the mass balance equation the rate of the reaction on the surface is represented by the reaction term dq/dt. The mass balance equation can be solved for dq/dt only in limited cases (Skopp and McCallister, 1986; Sparks, 1989; Bar-Tal et al., 1990). Some of the equations that have been used to describe sorption behavior (dq/dt) in a stirred-flow reactor are first-order, fractional-order, elovich, and parabolic diffusion (Skopp and McCallister, 1986; Sparks, 1989; Bar-Tal et al., 1990). However, on the basis of the Pb sorption behavior observed in this study, these equations are not appropriate since the sorption behavior consist of a fast and slow reaction phase. A better model to describe the Pb sorption data would include a fast equilibrium reaction, and a slow rate-limited secondary reaction. Several different multireaction models have been developed to describe metal sorption and desorption in soils (Amacher et al., 1988; Selim et al., 1989; Smith et al., 1996). In this study, these equations were not used.
For the desorption study, to make relative comparisons, it is useful to normalize the data by the total amount of Pb sorbed. To do this, we estimated the amount of Pb desorbed at a given time from the difference in the areas under the curves of the tracer, and the outflow concentration when desorption is conducted. The total amount of Pb sorbed can be estimated by a similar approach, i.e., when the outflow concentration is the same as the inflow concentration the difference between the blank and the sorption curve represents the total amount of Pb removed from solution. The amount of Pb sorbed or desorbed is then calculated by multiplying the area by the flow rate and dividing by the mass of soil. To calculate the areas under the curves, the integral of a cubic-spline interpolated from Cout(t) was calculated. These calculations were made by the algorithms in the computer program Scientist (MicroMath Research, Salt Lake City, UT).
XAFS Experiments
Treated (SOM removed) and untreated soil samples were analyzed by XAFS to determine the local atomic structure surrounding the sorbed Pb in the presence and absence of SOM. The samples were incubated using the same procedures described above. For XAFS analysis, the wet pastes were loaded into Teflon sample cell holders in the glovebox and sealed with Kapton Tape.
XAFS data acquisition of the LIII-edge (13055 eV) was conducted at beamline X-11A at the National Synchrotron Light Source (NSLS), Brookhaven National Laboratory, Upton, NY. Details of the XAFS data collection procedures and data analysis are described in Strawn et al. (1998). In this study, three to five scans were merged, the data were normalized, and the background was removed by fitting three knots at unequal distances. The data were then converted to k-space, windowed, and fourier filtered to convert to R-space. Fitting of the RSF was attempted by theoretical paths for Pb, O, Si, and C backscatterer atoms generated from model compounds using FEFF 6.0. Estimations of the errors associated with the coordination numbers (N), bond distances (R) and Debye-Waller factors (
2) were made from the confidence limits of the least squares non-linear fitting procedure. Previous data analysis (Strawn et al., 1998) has shown that these values are usually larger than the absolute errors associated with fitting model compounds.
|
Results and discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMethods and materialsResults and discussionStirred-Flow ExperimentsSummaryREFERENCES |
|---|
 | (2) |
|
and
. When the initial concentration of Pb is low, most of the Pb is sorbed by the soil. This is characteristic of a strong affinity of the soil for Pb. As the initial concentration of Pb increases, the slope of the sorption isotherm levels off, indicating that the soil is reaching a sorption maximum. The sorption behavior displayed by the Matapeake soil is typical L-type sorption behavior which is commonly observed for metal sorption by soils (Sparks, 1995).
The results of the batch sorption kinetic experiments are presented in Fig. 3
. Within the first 8 min (first sampling time), a very fast reaction occurred, accounting for 78% of the total sorbed Pb. Following the initial fast reaction, the sorption reaction continued for 100 h. At times longer than 100 h, only a small amount of additional sorption occurred. After 800 h, 79% of the Pb initially added to the suspension had been removed. Since a majority of the Pb is sorbed prior to the first sample (78% of the total Pb sorbed occurred within 8 min), and the rest of the sorption occurs within 100 h, it is hypothesized that there are two distinct reactions occurring: one fast reaction and one slow reaction. The slow data were modeled using a reversible first-order reaction:
 | (3) |
|
. The overall apparent rate constant obtained from the fit of the data was
. As determined by this value, the half life for the slow reaction was calculated to be 58 h. The two stage time-dependent Pb sorption behavior is similar to the results observed on pure components and soils obtained by other researchers. Carriere et al. (1995) found that Pb sorption on soil was fast at low initial Pb concentrations, but at higher concentrations the reaction became much slower, taking several days to reach a steady state. From these results, Carriere et al. (1995) suggested that precipitation reactions were responsible for the slow sorption reaction since they were most noticeable at the higher [Pb]. Hayes and Leckie (1986) and Benjamin and Leckie (1981) found that sorption of Pb on iron hydroxides (common minerals in soils) was initially fast, followed by a much slower reaction. Similar results were found by Strawn et al. (1998) for Pb adsorption on aluminum oxide. Slow Pb adsorption has also been observed on pure organic materials (Wilczak et al., 1993). Since soil is a mixture of organic and inorganic components, and it contains several different types of sorption sites, it is likely that several mechanisms are responsible for the slow sorption reactions. These may include diffusion, precipitation, and/or sorption reactions on sites that have a higher activation energy than the fast sorption sites.
|
Stirred-Flow Experiments |
|---|
|
TOPABSTRACTINTRODUCTIONMethods and materialsResults and discussionStirred-Flow ExperimentsSummaryREFERENCES |
|---|
2 x Cout/Cin) of the breakthrough from 2.7 to 2.4 CVs, respectively. This shift is a result of changes in the reaction occurring in the chamber. The average time the solute spends in the reactor vessel (
) is a function of the CV and the flow rate
(Brezonik, 1994). For the experiment in which
,
, while for the experiment conducted at
,
. In a system where the reaction occurring is slower than the solute residence time (), the outflow concentration would be expected to increase as the flow rate increases. This is the behavior observed in Fig. 4 for Pb sorption on the Matapeake soil, suggesting that sorption kinetics are being measured.
|
6%. This is a significant drop giving additional evidence that the reaction is rate limited. Thus, it can be concluded from the experiments in which the flow rate is changed, and the stopped-flow experiment, that the reactions in the stirred-flow reactor at 4 mL min-1 are rate limited. Although not shown, data from 2 mL min-1 also show a shift in the outflow concentration that is intermediate between the 0.4 and 4 mL min-1 experiments, suggesting that at this flow rate the reaction is also rate limited.
The solid line in Fig. 4 is the theoretical tracer calculated from Eq. [1], when
. The midpoint of the breakthrough for this data is 0.75 CVs. The midpoint for the sorption experiment conducted at 4 mL min-1 occurs at 2.4 CV, this is shifted 1.65 CVs from the midpoint of the breakthrough in the tracer. This retardation shows that the soil is adsorbing significant amounts of Pb. One of the distinctive characteristics present in the stirred-flow experiments is that initially all of the Pb input into the chamber is sorbed, resulting in the outflow [Pb] being zero until 1.2 CV. The initial adsorption accounts for 65% of the total sorption that occurred on the soil. Since in this experiment
, the fast reaction occurring on the soil must have a half life <2 min. Fast adsorption reactions for Pb on oxides have been studied by pressure-jump relaxation (Hayes et al., 1986; Yasunaga and Ikeda, 1986). Results from these studies suggest that the fast adsorption reaction is primarily due to the formation of bonds with functional groups that are readily available on surfaces, and that the reaction occurs within seconds. These types of adsorption reactions are too fast to be measured by the stirred-flow reactor.
After 1.2 CV, the Pb concentration in the outflow becomes >0 (Fig. 4). This suggests that Pb sorption is either slowing down, or that the soil has reached its maximum sorption capacity. If the soil were to sorb no additional Pb from solution, then the breakthrough would follow the dashed line (a shifted theoretical tracer data set) in Fig. 4. However, the [Pb] in the outflow is retarded compared with the tracer, indicating that sorption is continuing at these longer times. The stopped-flow and variable flow rate experiments indicate that the sorption reaction occurring after 1.2 CV is rate limited. The slow reaction could be due to sorption on sites which are less readily available than those in which fast sorption occurred (diffusion limited), or a result of secondary sorption mechanisms such as precipitation or sorption on sites with larger activation energies.
Sorption on the Matapeake soil reached a steady state after about
. At this point, the total sorption on the soil was 44 mmol kg-1. This value is only 68% of the total sorption that occurred in the batch kinetic experiment that was allowed to equilibrate for 800 h. The difference between the total sorption measured by batch and stirred-flow methods is most likely due to the fact that the slow reaction is occurring too slowly (t1/2 = 58 h) to detect any noticeable change in the [Pb] in the outflow solution in this series of experiments.
The effect of SOM on Pb sorption is shown in Fig. 5
. In the treated Matapeake soil (0.1% SOM), the midpoint of the breakthrough arrives 0.95 CV earlier than in the untreated soil. The total amount of Pb sorbed in the treated soil was 27 mmol kg-1. This is a 40% decrease compared with the untreated Matapeake soil. If the difference in Pb sorption between the treated and untreated soil is due to the presence of SOM, then the amount of Pb sorbed on the SOM under these conditions is at least 810 mmol kg-1 (1620 mmol charge (kg SOM)-1). Since the very slow sorption reaction is not included in the stirred-flow measurement, the true sorption capacity may be higher. The measured SOM sorption capacity is very high, and demonstrates the importance of SOM for Pb sorption in soils. The CEC range for SOM that is reported in the literature is 1500 to 3000 mmol charge (kg SOM)-1, depending on the pH of the soil solution (Sparks, 1995).
|
The midpoint of the breakthrough for the St. Johns soil occurred after 5.1 CV (Fig. 5). This is 2.4 CV later than the breakthrough from the untreated Matapeake soil. In addition, the slope of Cout/Cin (CV) is significantly smaller than it is for the Matapeake soil, and much different than the tracer. These differences imply that sorption is continuing after the initial fast reaction. The total sorption on the St. Johns soil is 102 mmol kg-1. This is 2.3 times the sorption occurring in the Matapeake soil. Since the clay contents in the Matapeake and the St. Johns soils are similar (Table 1), but the St. Johns soil has 6 times more SOM, the most likely reason for the increased Pb sorption is the increase in sorption sites existing on the SOM. As determined by the sorption capacity calculated for the SOM in the Matapeake soil [810 mmol (kg SOM)-1], the predicted amount of Pb sorbed on the SOM fraction in the St. Johns soil (13% SOM) is 105 mmol (kg whole soil)-1, which is close to the observed value of total Pb sorbed in the St. Johns soil (102 mmol kg-1). Thus, from the stirred-flow experiments, it is clear that SOM is an important factor for Pb sorption in the environment.
The results obtained from the desorption experiments on the untreated Matapeake soil are shown in Fig. 6 . Since a much smaller amount of Pb is recovered in the desorption experiment than was removed from solution in the sorption experiment the desorption reaction is not complete. This can be considered as an apparent hysteresis (DiVincenzo and Sparks, 1997) since the reaction will most likely be reversible if given enough time. Changing the flow rate from 0.4 to 4 mL min-1 had insignificant effects on Cout/Cin (CV). If desorption was fast and reversible within the time frame of the experiment, then all of the sorbed Pb would be recovered from the soil, and there would be no change in the desorption breakthrough curves. However, only a fraction of the total sorbed Pb was recovered, indicating that the reaction is not at equilibrium. From the sorption isotherm, it was observed that at low equilibrium concentrations the slope of the isotherm is steep (high surface loadings and low solution concentration); this would cause a significant amount of tailing in the desorption measured in the stirred-flow reactor, particularly if the release of Pb from the soil is slow. Since the desorption behavior is clearly much different than sorption behavior, and the desorption breakthrough curves are nearly identical, it can be concluded that the desorption kinetics are too slow to measure by the stirred-flow reactor at flow rates of 0.4 mL min-1.
|
3 d. Similar results were found by Ainsworth et al. (1994) and Gunneriusson et al. (1994) for Pb desorption on iron oxides. Desorption experiments conducted on montmorillonite (Strawn, 1998) have revealed that the time scales for Pb sorption and desorption are similar. Thus, it appears that Pb sorption on oxides and clays is reversible. However, on the basis of the results of this study, desorption from whole soils is not completely reversible within the time scales of these experiments, and the amount of Pb desorbed is directly correlated to the percentage SOM present in the soils. From these observations, it can be concluded that Pb desorption from the SOM fraction of soils is slow. The most likely reason for this behavior is that the type of complex that forms between the Pb and the SOM is stronger than the complex forming on the surface of the mineral. On the surfaces of minerals, the only type of Lewis bases for Pb to sorb are hydroxyls, which are considered to be hard bases (low polarizability) (Sparks, 1995). The functional groups on SOM include carboxyls, phenols, amines, and several sulfur containing functional groups which are soft bases (Sparks, 1995). Since soft acids prefer soft bases, the complexation of Pb by the functional groups on the SOM would be preferential to the hard acid hydroxyl ligands present on the mineral surface (Sparks, 1995).
|
|
11% in adsorption. Despite increasing the sorption incubation time from 1 to 32 d, the percentage of Pb desorbed decreased only slightly. This result shows that the complexes formed during Pb sorption are stable within 32 d of incubation, and do not convert into phases that are less readily desorbed.
|
XAFS Experiments
Figure 8
shows the background subtracted k3 weighted 
functions for the treated and untreated Matapeake soil samples. The structures are distinctly different in phase, wavelength, and shape, particularly at higher k. The uniqueness of the structures is due to differences in backscattering from the atoms residing in the first and second coordination shells surrounding the sorbed Pb.
|
are presented in Fig. 9
. The first major peaks that occur in the RSF are located at 1.60 and 1.75 
in the treated and untreated Matapeake soil, respectively. In the untreated Matapeake soil, a large peak is observed at 2.4 . This peak is noticeably smaller and shifted in the treated sample. The identity of the atoms surrounding the sorbed Pb can be determined by fitting the data to the theoretical model created using the backscattering simulation program FEFF 6.0. As described in the Methods and Materials, the fits of the samples were made by means of theoretical pathways from Pb, O, Si, and C as the backscattering atoms. In SOM, there also exists N and S containing functional groups that can form complexes with Pb (Sparks, 1995). Analysis of the total N and S in the Matapeake soil revealed that in the untreated Matapeake soil there was only 0.016% total S, and 0.2% total N. In the treated Matapeake soil, S and N were not detected. The total C in the untreated Matapeake soil was 1.38% (7 and 87 times the total N and S, respectively). Thus, contributions to the XAFS from N and S are not likely to be large unless the Pb is only forming complexes with these functional groups. Manceau et al. (1996) used XAFS to speciate Pb contaminated soil. They found that the XAFS spectra from the soil were best modeled by a linear combination of Pb–salicylate and Pb–catechol reference compounds, suggesting that the Pb is complexed to these types of functional groups in the SOM. Xia et al. (1997) also found that Pb adsorption on humic acids isolated from SOM was best described by fitting first shell O atoms and second shell C atoms. Thus, in this study, S and N were not considered as second shell backscattering atoms (note: because of the similar atomic size of N and C, isolating the individual contributions from these atoms to the EXAFS is difficult). The best fits of the data
are represented by the dashed lines in Fig. 9, and the resulting values for N, R, and 2 are presented in Table 4
. In both soils, the fits represented the data very well.
|
|
The EXAFS of the untreated Matapeake soil was best described by fitting O and C atoms as the backscattering atoms present in the local atomic structure of sorbed Pb. Attempts to fit Si and or Pb in the second shell of the sorbed Pb were unsuccessful. The predicted bond distance between the Pb and O (2.29 ) is distinctly longer than the RPb-O in the treated Matapeake soil
. The bond distance observed between the Pb and C atoms is 3.05 . This distance is shorter than the distance predicted by Xia et al. (1997) (3.26 ) for Pb sorption on humic acid, but too long for Pb-O (O and C cannot be distinguished by XAFS because of their similar size). The shorter bond distance can result from a decrease in the Pb-O-C bond angle occurring in the sample, or coordination to different ligands than were present in the humic acid sample. Manceau et al. (1996) used XAFS to predict that the speciation of Pb in a contaminated soil was 60% Pb–salicylate complexes and 40% Pb–catechol complexes. These complexes involve the formation of a five-membered chelate complex between the phenol functional groups of the catechol, and a six membered chelate complex formed between the phenol and carboxyl functional groups of the salicylate.
The differences in the RSF reveal that the local atomic structure surrounding the sorbed Pb atom is different in the two samples. In the treated soil, C was not detected in the local atomic structure of the sorbed Pb, while in the untreated soil, the dominant backscatterers were O and C. Figure 10 illustrates the sorption mechanisms predicted from the XAFS results using a clay mineral surface and a phenol functional group from SOM as the sorbents for the Pb (these models are only examples of the bonding environments, the actual bonding environment in the soil involves several different types of sorbents). The XAFS results confirm that the sorption mechanisms in the two soils are different. Since in the untreated sample no Pb or Si backscatterers were detected, most of the Pb must be sorbed onto the SOM, suggesting that the SOM is outcompeting, or blocking, the mineral surfaces from sorbing available Pb.
|
|
Summary |
|---|
|
TOPABSTRACTINTRODUCTIONMethods and materialsResults and discussionStirred-Flow ExperimentsSummaryREFERENCES |
|---|
Measurement of sorption kinetics using the stirred-flow reactor revealed that the extent of the fast and slow reactions occurring in the soil is directly dependent on the amount of SOM present. Specifically, with increasing SOM content the rate of Pb sorption decreases. The Pb sorption capacity of the SOM is estimated to be at least 810 mmol kg-1. Desorption experiments conducted using the stirred-flow reactor revealed that Pb desorption is hysteretic within the time frame of these experiments. The fraction of Pb that was desorbed from the soil decreased as the amount of SOM present in the soil increased. This indicates that the slow desorption reaction is primarily related to the SOM fraction of the soil. XAFS results from this study confirm that the types of complexes forming in the soil with SOM and the soil without SOM are different, and help explain the macroscopic Pb sorption–desorption behavior.
The information presented in this study shows the importance of not only measuring sorption equilibrium to predict the behavior of Pb in soil, but also measuring time-dependent sorption and desorption reactions. This additional information will allow scientists and engineers to make better predictions about the transport, bioavailability, and sorption and desorption behavior of Pb in soil. Such information is critical for protecting natural resources, developing improved remediation strategies, and making better risk assessments.Jin Bailey Yu Lynch 1996; Loehr Webster 1996; Scheidegger Sparks 1996; Wilczak Keinath 1993
|
ACKNOWLEDGMENTS |
|---|
Received for publication January 14, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMethods and materialsResults and discussionStirred-Flow ExperimentsSummaryREFERENCES |
|---|
-FeOOH)/water interface: The influence of chloride. Geoch. Cosmochim. Acta 1994;58(22):4973-4983.This article has been cited by other articles:
|
|
 |
|
  Y. Hashimoto, H. Matsufuru, M. Takaoka, H. Tanida, and T. Sato Impacts of Chemical Amendment and Plant Growth on Lead Speciation and Enzyme Activities in a Shooting Range Soil: An X-ray Absorption Fine Structure Investigation J. Environ. Qual., May 20, 2009; 38(4): 1420 - 1428. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. E. Lopez-Periago, M. Arias-Estevez, J. C. Novoa-Munoz, D. Fernandez-Calvino, B. Soto, C. Perez-Novo, and J. Simal-Gandara Copper Retention Kinetics in Acid Soils Soil Sci. Soc. Am. J., January 11, 2008; 72(1): 63 - 72. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. C. W. Tsang, W. Zhang, and I. M. C. Lo Modeling Cadmium Transport in Soils Using Sequential Extraction, Batch, and Miscible Displacement Experiments Soil Sci. Soc. Am. J., April 5, 2007; 71(3): 674 - 681. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. G. Scheckel and J. A. Ryan Spectroscopic Speciation and Quantification of Lead in Phosphate-Amended Soils J. Environ. Qual., July 1, 2004; 33(4): 1288 - 1295. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. D. Yanai, D. G. Ray, and T. G. Siccama Lead Reduction and Redistribution in the Forest Floor in New Hampshire Northern Hardwoods J. Environ. Qual., January 1, 2004; 33(1): 141 - 148. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. M. Hettiarachchi, J. A. Ryan, R. L. Chaney, and C. M. La Fleur Sorption and Desorption of Cadmium by Different Fractions of Biosolids-Amended Soils J. Environ. Qual., September 1, 2003; 32(5): 1684 - 1693. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Brown, R. L. Chaney, J. G. Hallfrisch, and Q. Xue Effect of Biosolids Processing on Lead Bioavailability in an Urban Soil J. Environ. Qual., January 1, 2003; 32(1): 100 - 108. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. E. Brown Jr. and N. C. Sturchio An Overview of Synchrotron Radiation Applications to Low Temperature Geochemistry and Environmental Science Reviews in Mineralogy and Geochemistry, January 1, 2002; 49(1): 1 - 115. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
