SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 11 January 2008
Published in Soil Sci Soc Am J 72:63-72 (2008)
DOI: 10.2136/sssaj2006.0079
© 2008 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by López-Periago, J. E.
Right arrow Articles by Simal-Gándara, J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by López-Periago, J. E.
Right arrow Articles by Simal-Gándara, J.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by López-Periago, J. E.
Right arrow Articles by Simal-Gándara, J.
Related Collections
Right arrow Soil Kinetics
Right arrow Toxic Trace Metals
Right arrow Sorption/Exchange

SOIL CHEMISTRY

Copper Retention Kinetics in Acid Soils

José Eugenio López-Periago*, Manuel Arias-Estévez, Juan Carlos Nóvoa-Muñoz, David Fernández-Calviño, Benedicto Soto, Cristina Pèrez-Novo and Jesus Simal-Gándara

Faculty of Sciences, Univ. of Vigo, As Logoas Ourense, Ourense 32004, Spain

* Corresponding author (edelperi{at}uvigo.es).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Retention and release kinetics of Cu on four acid Typic haplumbrepts developed on two different types of parent rock material (granite and amphibolite) were studied with a stirred-flow chamber (SFC) method. The granitic soils were lower in organic material and lower in Fe and Al oxides than the soils formed in amphibolite. The kinetic parameters were assessed in four consecutive Cu retention–release cycles by alternately applying pulses of solutions with and without Cu. Granite soils showed lower total Cu retention (7–12 mmol kg–1) than amphibolite soils (16–21 mmol kg–1) after one single pulse application of 0.0787 mmol Cu L–1 at pH 5.5, which may be due to differences in their organic and oxides compositions. The amount of Cu retained diminished to 40 to 25% in the third retention cycle relative to the first, suggesting that the soils' Cu retention depends on the previous metal loading. Conversely, the released Cu was approximately 20% of that retained in the first cycle, but the amounts released were similar for all cycles and all soils. The results obtained were fitted using a first-order equation for both retention and release of Cu. In the first cycle, first order rate coefficients of retention ranged from 0.084 to 0.56 min–1, and increased by about a factor of two in the next cycle. Release rate coefficients were more than 10 times lower than those of retention, and less dependent on the previous metal loading. The process of retention and release of Cu was found to be hysteretic, which suggest that the desorption mechanism or path is not an exact opposite of the adsorption mechanism or path.

Abbreviations: BTC, breakthrough curve • CV, chamber volume • LDF, linear driving force • SFC, stirred-flow chamber


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The Cu concentration in soil depends on the nature of its parent material and is especially high in soils developed on mafic and intermediate rocks (Kabata-Pendias and Pendias, 2001). In recent years, the Cu content of surface horizons in agricultural soils in the vicinity of some industries has risen significantly world wide. The increased levels of Cu have resulted largely from the use of organic and inorganic fertilizers, nearness to metal processing industries and repeated application of Cu-based pesticides (Kabata-Pendias and Pendias, 2001).

The increased levels of Cu in soils have promoted studies on the way the metal partitions between soil compartments (Shuman, 1991; Arias et al., 2004). Most studies have revealed that the Cu is strongly bound to the organic fraction or to amorphous inorganic oxides. The adsorption and desorption of Cu in soils depend primarily on soil pH and organic matter (McLaren et al., 1983; Msaky and Calvet, 1990; Harter, 1992; Silveira, 2002).

The retention and release of Cu in soil can be time-dependent if its kinetics is slow enough relative to water motion. Therefore, studying the retention kinetics of this metal may help elucidate its dynamics in soil. Kinetic methods can also be useful to examine the interaction of Cu with soil. For example, Harter and Lehmann (1983) successfully resolved diffusion and ion exchange of Cu via kinetic experiments in soil horizons of a Typic Paleudult and a Typic Fragiochrept. Also in the same soils, Lehmann and Harter (1984) used the Cu release kinetics to discriminate retention energies for this metal, and Harter (1991, 1992) examined its competition with other metals. Other studies on the adsorption kinetics of Cu on different soil types (Mattigod et al., 1981; Sposito, 1982) and Histosols (Aringhieri et al., 1985) have revealed that the process involves a rapid initial step by which the metal is adsorbed at readily accessed sites and a second, slower step typical of adsorption on modified surfaces, coprecipitation reactions and diffusion into inner surfaces.

Much research on reaction kinetics in soils has relied on batch techniques, the efficiency of which is limited by the slowness of phase separation processes, or on miscible displacement in soil columns, which is diffusion-controlled. The SFC technique (Heyse et al., 1997) alleviates the shortcomings of batch method by virtue of the adsorbate being continuously injected, which avoids the problems posed by a decreased concentration in the liquid phase in batch processes. Also, during the desorption phase the metal released into the liquid phase is removed from the chamber, which accelerates further desorption, especially if the flow rate is fast enough to preclude readsorption of the released metal. The advantage of SFC with respect to the column experiments is that the film diffusion resistance, which can be the rate limiting step in the sorption kinetics can be more effectively reduced than by the stirring of the soil suspension. These features favor the SFC technique, which allows one to examine the kinetics of the retention process over a short time scale.

Desorption of metals in soils can often be much slower than adsorption, and sorption reactions are often pseudo-irreversible (Strawn and Sparks, 1999). Due to slow desorption, successive Cu loads may take place before reaching equilibrium with the soil, causing the sorption to become dependent on the metal loading history. This may be a reason for the commonly observed hysteresis in adsorption isotherms (Verburg and Baveye, 1994).

We hypothesize that the history of Cu application can also affect sorption dynamics. Therefore, when Cu is repeatedly applied to soil, its previous retention-release history may influence future Cu dynamics. Understanding the dynamics of Cu in soil should consider its Cu loading history. The purpose of this work was to examine the behavior of Cu retention and release on the minute scale with a view to identifying and quantifying the effects of successive adsorption–desorption cycles on the dynamics of Cu retention by acid soils.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soils
Four soil samples were collected from two different locations in Galicia (NW Spain). Two of the soils (Amph1 and Amph2) developed from amphibolite and the other two (Gra1 and Gra2) from granite; all four were Typic haplumbrepts (Soil Survey Staff, 1997). The major minerals in the amphibolite soils were halloysite, allophane and gibbsite (Macías and Calvo, 1992). The granite soils had lower Fe contents and a coarser texture than the amphibolite soils. The clay fraction in the surface horizons of the granite soils was dominated by mica strongly weathered to vermiculite and contained abundant hydroxyl-Al interlayers (Macías and Calvo, 1992). The studied soil samples were collected from the A horizon, air-dried, and sieved through a 2-mm screen. The <2-mm fraction was then used for characterization analyses.

Soil particle density was measured by water picnometry (Blake and Hartge, 1986). Soil texture was determined by using the standard pipette method following sieving (Gee and Bauder, 1986), and organic C on a Finnigan Flash EA-1113-NC catalyzed dry combustion analyzer from Thermo Electron Corp. (Waltham, MA). Soil pH was measured in H2O and 0.1 M KCl, using a 1:2.5 soil/solution ratio. Exchangeable Al was determined by displacement with 1M KCl by atomic absorption spectrophotometry. Effective cation exchange capacity (CECe) was determined as the sum of bases (Na, K, Ca and Mg) extracted with 0.2 M NH4Cl and exchangeable Al (Bertsch and Bloom, 1996). All samples were extracted for Fe and Al (Fe-OX, Fe-PP, Al-OX, and Al-PP) with ammonium oxalate (Blakemore, 1978) and sodium pyrophosphate (Bascomb, 1968). Dithionite–citrate extractable Fe and Al (Fe-DC and Al-DC) (Holmgren, 1967) were also determined. The pyrophosphate reagent is an effective extractant for humus complexes of Al and Fe, while the oxalate reagent additionally extracts non-crystalline Fe and Al hydrous oxides, allophane, imogolite and allophane-like constituents. Exchangeable cations were extracted with 1 M ammonium acetate and determined by atomic absorption (Ca and Mg) or atomic emission spectrophotometry (Na and K). Table 1 summarizes the properties of the soils.


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

  		Table 1. Selected physical and chemical properties of the soils.

 
Batch Experiments
To determine the maximum number of adsorption sites at equilibrium, the four soils were also subjected to batch isotherm tests. To this end, 1 g of soil was pre-equilibrated in a 0.01 M Ca(NO3)2 background solution at pH 5.5 for 24 h in each run. A pH of 5.5 was chosen because at lower (more acid) pH values, the soil buffering capacity would facilitate the release of Al to the solution, thus critically affecting the reactions of Cu with soil. The pH 5.5 is a usual value for acid soils and ensures mostly that complexation effects will be kept constant.

To account for the time required to reach equilibrium, Cu adsorption was measured at five different incubation times (1, 4, 6, 24, and 48 h) for Amph1 and Gra1 soils.

For the batch tests, in each sample 1 g of soil was suspended in 5 mL of each of the following Cu(NO3)2 solutions (0.16, 0.32, 0.79, 1.2, 1.6, 2.4, and 3.2 mM). The pH sample was adjusted to pH 5.5 following addition of the Cu solution and re-adjusted after 24 h. The solid concentration in the samples after the addition of water was 100 g L–1. Samples were incubated on an end-over-end shaker at 18 rpm for 48 h and then centrifuged at 800 x g for 15 min. The supernatant thus obtained was filtered and diluted with 1 M HNO3 to determine total Cu by atomic absorption spectrometry with an error of only 2%.

The Cu adsorption results were used to determine the parameters of the Langmuir equation:

Formula 1[1]
where q is the adsorbed concentration of Cu at equilibrium (mmol kg–1), C the solution concentration (mmol L–1), kL the Langmuir adsorption constant (L kg–1) and qLmax the maximum adsorbed concentration at equilibrium (mmol L–1).

Kinetic Tests
Reactor and Operating Conditions
The SFC (Fig. 1 ) was made of polypropylene and had an inlet side port at the bottom and a cover with an outlet port at the top. Two polytetrafluoroethylene (PTFE) filters 10 mm in diameter and 0.45 µm in pore size were fitted immediately below the outlet port and over the inlet port to retain soil samples in the chamber. The chamber volume was 1.2 cm3. Both the influent and effluent were carried through 0.5 mm i.d. PTFE tubing connected to a Gilson Minipuls 3 peristaltic pump (Gilson S.A.S., Villiers Le Bel, France). The temperature was kept at 25 ± 0.1°C by placing the reactor chamber in a thermostated cell connected to a circulating water bath. The flow rate was monitored throughout and found to oscillate by <3%. Stirring was provided by a PTFE-coated magnetic bar that was spun at 400 rpm. Effluent fractions were collected into 1.5-mL polypropylene Eppendorf vials by using a Gilson FC 203 G automatic fraction collector (Gilson S.A.S., Villiers Le Bel, France).


Figure 1
View larger version (17K):
[in this window]
[in a new window]

  		Fig. 1. Schematic depiction of the stirred-flow chamber.

 
Two unbuffered solutions containing no Cu and another two containing different Cu concentrations were used. The former two were 0.0787 and 0.157 mM Ca(NO3)2, respectively, and the latter 0.0787 and 0.157 mM Cu(NO3)2, respectively. All were adjusted to pH 5.5 with HNO3. Stirred-flow tests were conducted on the soil fraction finer than 0.5 mm. To this end, an amount of about 0.2 g of soil was placed together with a magnetic stirring bar in the reaction chamber, which was then filled up with Ca(NO3)2 solution and carefully closed to avoid bubble formation. The Ca(NO3)2 solution was circulated at a flow rate (Jw) of 0.35 mL min–1. The Cu concentration and pH of the effluent were measured on a periodic basis.

Once the pH leveled off at 5.4 to 5.6 and no native Cu was detected in the effluent, the system was deemed ready for application of Cu pulses. In the retention step, a solution of Cu(NO3)2 was circulated by using a Rheodyne 5001 switching valve (Rheodyne Europe GmbH, Bensheim, Germany) to select the mobile phase; for the release step, the valve was actuated in the opposite direction to circulate Ca(NO3)2 solution again.

The Cu concentrations in the different fractions obtained during each kinetic run were used to construct breakthrough curves (BTCs) to examine the variation of Cu retention and release.

The adsorption of Cu by the chamber walls was checked in a experiment where there was an absence of soil in the chamber using a 200-min-pulse of 0.0787 mM Cu at Jw = 0.56 mL min–1.

To determine if the retention and release were limited by the rate of adsorption instead of transport in the SFC, two tests proposed by Bar-Tal et al. (1990) were used. The former test, involving altering the flow rate, was performed at 0.35, 0.56, and 0.7 mL min–1 using a Cu concentration of 0.0787 mM and a soil-loaded chamber. The latter, involving stopping the flow, was performed at the same Cu concentration at a flow rate of 0.56 mL min–1; following application of 23 to 27 mL (20–23 chamber volumes, CV).

Retention-Release Tests
The kinetics of retention and release were studied in consecutive Cu retention–release cycles that were established by alternately applying pulses of solutions containing and excluding Cu at a flow rate of 0.56 mL min–1. The two steps were repeated four times at each Cu concentration (0.0787 and 0.157 mM).

The amount of Cu retained or released in the retention tests was calculated from the concentrations in the effluent fractions from the reactor obtained in two tests (with and without soil in the reactor), using the equation of Yin et al. (1997):

Formula 2[2]
where q(i) (mmol kg –1) is the concentration of Cu retained by the soil over the discrete interval i, which denotes the position in the sequence of the effluent fraction from the start of the run; C(i), the Cu concentration in the effluent fraction (mmol mL–1) at time Step i; and C(i+1), the concentration in the chamber, which was assumed to coincide with that in the next effluent fraction (i+1) and is approximately 0.9–1 CV. Subscripts 1 and 2 in the equation denote the concentrations in the effluent fractions corresponding to the absence and presence of soil in the chamber, respectively; {Delta}t is the time difference between each effluent fraction; Jw is flow rate; Ve is the effective solution volume in the chamber (L); and m the mass of soil placed in it (kg).

Parameter Ve was estimated by fitting the expression c(t) = cin [1– exp(–Jw t/Ve)] to match the BTC resulting from the application of a pulse of Cu with an influent concentration (cin) containing 0.0787 mmol Cu(NO3)2 at Jw = 56 mL min–1 SFC without soil.

Modeling Kinetic Data
The equation used for the rate of retention or release in a SFC, which was proposed by Bar-Tal et al. (1990), was as follows:

Formula 3[3]
where C is the Cu concentration in the chamber (mol L–1), t time (min), Cout the Cu concentration leaving the chamber (mol L–1) and q the amount of Cu adsorbed in it (mol kg–1). Since the solution was well mixed in the chamber, C = Cout. All other variables in the equation were defined above. The term dq/dt describes the kinetics of Cu retention or release by the soil; therefore Eq. [3] is in fact a system of ordinary differential equations that allow the Cu concentrations retained and released by the soil to be calculated with provision for Cu dilution in the chamber.

Breakthrough curves were constructed by numerically solving Eq. [3] and the kinetic law, using the Runge-Kutta explicit formula (Dormand and Prince, 1980) as implemented in the ODE45 routine of MatLab v. 6.5.0R14 (Mathworks, Inc., Natick, MA). The initial conditions for the SFC were set at dC/dt = dq/dt = C = q = 0. The initial boundary condition was Cin = Ci(t), where Ci(t) is the function describing the application of a Cu pulse to the influent. In the retention step t = (0–200) min, Ci(t) equaled the concentration in the Cu solution; in the release step, Ci(t) = 0 at t = (200–400) min. The kinetic variables in the Eq. [3] were optimized by using a nonlinear fitting subroutine which provided the best fit in terms of r2 and the root mean square error statistic as corrected for the number of data pairs and variables: RMSE = [rss/(n – p)]0.5, where rss is the summation of residuals (viz. the differences between the measured and estimated concentrations in each time interval) squared, n the number of data pairs and p that of parameters.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Batch Experiments
Kinetics of Cu adsorption on two soils (Amph1 and Gra1) at the added concentration of 0.787 mM is shown in Fig. 2 . After 24 h, 99 and 86% of the Cu initially added to the suspension had been removed in Amph1 and Gra1, respectively. Since no significant changes in Cu retention occurred between 24 and 48 h of incubation, 24 h was long enough to ensure equilibrium in both soils.


Figure 2
View larger version (11K):
[in this window]
[in a new window]

  		Fig. 2. Kinetics of Cu adsorption on soils at an added concentration of 0.787mM. Soils: (A) Amph1, (B) Gra1.

 
Figure 3 illustrates the adsorption of Cu by the studied soils after 24 h of incubation. As it can be seen, the four soils exhibited a typical L-shaped sorption isotherm (Sparks, 1989). When added at low concentrations, most Cu was retained by the soils. This indicates a strong affinity for this metal. As the amount of adsorbed Cu increased, the slope of the sorption isotherm tended to level off, which suggests that the soil was approaching its maximum adsorption capacity. Table 2 shows the results obtained by nonlinear least squares regression fitting the adsorption data to the equation for the Langmuir isotherm (Eq. [1]), minimizing r2 between the observed and predicted data (Schulthess and Dey, 1996). The calculated isotherms are shown as lines in Fig. 3. Fits were acceptable for all soils; therefore, the Langmuir model provides an accurate description of Cu adsorption after 24 h of incubation.


Figure 3
View larger version (15K):
[in this window]
[in a new window]

  		Fig. 3. Equilibrium adsorption isotherms. Symbols represent experimental data and the line the Langmuir model. Soils: (A) Amph1, (B) Amph2, (C) Gra1, and (D) Gra2.

 

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

  		Table 2. Parameters fitted to the Langmuir equation (Eq. [1]) and q estimates obtained at equilibrium with the Cu solutions used in the SFC tests. r2 is the degree of correlation between the observed and calculated q values for the batch experiments.

 
Based on KL and qLmax values (Table 2), the amphibolite soils exhibited higher Cu adsorption capacity than did the granite soils. This may be a result of higher contents in organic carbon, clay, extractable Fe and Al, and also increased CECe, in the amphibolite soils. These soil parameters play a key role in Cu adsorption by acid soils (Arias et al., 2004). The two amphibolite soils differed little in their organic C concentrations, so the greater KL and qLmax values for Amph1 may be a result of it containing a higher concentration of Fe oxides (Table 1). With the granite soils, Gra2 had an increased qLmax value relative to Gra1, presumably due to the higher concentrations of organic C and Fe oxides in the former. Although KL for Gra2 was half of Gra1, its difference in terms of Cu interaction energy with colloidal soil particles was smaller with respect to the amphibolite soils.

Stirred-Flow Chamber: Flow Rate and Stopped Flow Experiments
Variable flow rate and stopped-flow tests were conducted to check whether the reactor operating conditions were appropriate for studying the Cu retention kinetics. In addition, these tests can help to elucidate the mechanisms controlling the retention process.

The X-shaped symbols in Fig. 4a represent the observed Cu concentration in the outflow without soil (Jw 0.56 mL min–1), while the line represents a theoretical nonadsorbing tracer. Based on the mass balance obtained in the absence of soil, the mass of Cu retained by the chamber fell within the range of experimental error (2%). During the retention step in Amph1 (Jw 0.56 mL min–1) the pH in the outflow fractions decreased from 5.5 to 4.4 when CV < 20. This decrease of pH was due to a reduction in Ca release rate (Fig. 4b) from soil samples which were previously treated with Ca(NO3)2. Then at CV higher than 20, pH values increase slowly to 5.0 (Fig. 4b); this increase of pH suggests that Cu retention mechanisms are related to proton release.


Figure 4
View larger version (19K):
[in this window]
[in a new window]

  		Fig. 4. Results of the retention tests in the stirred-flow chamber. Observed breakthrough curves (BTCs) (symbols) obtained: (a) with no soil; (b) displacement of Ca+2 and pH variation during Cu+2 retention in Amph1; and (c) and (d), respectively, Amph1 and Gra1 at a variable flow rate under the same conditions. Lines represent the BTCs calculated by using the first-order model.

 
Flow Rate
Tests with soil in the chamber showed that the increase in Cu concentration in the outflow was delayed relative to the theoretical tracer. With soil Amph1 at Jw = 0.56 mL min–1, no Cu was detected in solution until 20 chamber volumes (CV) were circulated (Fig. 4c). This suggests that retention was faster than the mass flux of Cu entering the chamber. The amount of Cu initially retained during this period accounted for 18 mmol kg–1 (65% of that retained throughout the test). At Jw = 0.7 mL min–1, however, it accounted for <2% of the total amount of Cu retained, showing that increasing mass flux of Cu exceed the retention rate. These results were similar for all soils, albeit less marked in the granite soils as shown in Fig. 4d.

Comparing Cu concentration in experiments with soil against without soil (Fig. 4a, c-d), it was observed a slower increase in Cu concentration for all studied flow rates was observed when the experiment was performed with soil. For a flow rate of 0.56 mL min–1, the increase in Cu concentration at the effluent took place approximately after 18 CV, whereas for 0.35 mL min–1 this increase was observed after 50 CV (Fig. 4c). Following Strawn and Sparks (2000), this behavior can have two different origins: one is the effect of a gradual decrease in the number of readily available adsorption sites, and the other the simultaneous occurrence of secondary slower retention mechanisms that might gain prominence once the faster process has finished. It is possible that both mechanisms can occur simultaneously, however we cannot confirm either of these two possibilities.

The retention rates for the granite soils (Fig. 4d) were lower than those for the amphibolite soils. In the release step, Cu concentration decreased with increasing flow-rate. Accordingly, Cu release must also be influenced by the rate of the process.

Stopped Flow
Figure 5 shows the effect of stopping the flow at Jw = 0.56 mL min–1 through soils Amph1 and Gra1. With Amph1, halting the flow during the retention step decreased the Cu concentration in the effluent from 0.039 to 0.031 mM (i.e. by 20.5%). With Gra1, halting the flow after circulating 20 CV reduced the Cu concentration from 0.075 to 0.061 mM (18.7%). The decrease of dissolved Cu concentration in the chamber was detected after the flow was resumed, indicating that the retention continued when the flow was stopped. In the release step (data not shown), the flow was stopped for 15 min following passage of 23 mL (20 CV). Restarting the flow caused the Cu concentration in soil Amph1 to rise from 0.0205 to 0.0239 mM (16.6%), and in Gra1 from 0.0120 to 0.0195 mM (62.5%). The effects of stopping the flow or altering the flow rate were similar for all soils and allow one to conclude that, at Jw = 0.56 mL min–1, retention and release of Cu in both soils are time-dependent processes.


Figure 5
View larger version (13K):
[in this window]
[in a new window]

  		Fig. 5. Breakthrough curves obtained in the stopped-flow tests. Jw = 0.56 mL min–1. The symbols represent observed data and lines the concentrations calculated from Eq. [4].

 
The BTCs in Fig. 6 compare the results for the retention step following application of a pulse of 0.0787 mM Cu at Jw = 0.56 mL min–1 to all soils. Based on the calculated mass balance after the retention step (80 CV), the amounts of Cu retained by soils Amph1, Amph2, Gra1 and Gra2 were 20.6, 12.2, 6.3 and 8.3 mmol kg–1, respectively. These concentrations can tentatively be ascribed to the ability of soil organic matter to form strong complexes with Cu, in which respect the granite soils depart from the amphibolite soils. Additionally, the greater concentration of extractable Fe oxides in Amph1 may have also caused the granite soil to retain more Cu than the Amph2 soil. There was good consistency with their respective KL and qLmax values as calculated from the adsorption isotherms. As regards the granite soils, Gra1 retained less Cu than did Gra2.


Figure 6
View larger version (17K):
[in this window]
[in a new window]

  		Fig. 6. BTCs for the four studied soils after one retention cycle. Jw = 0.56 mL min–1. Cin = 0.0787 mM. Symbols represent experimental data and lines the first-order model.

 
Retention Cycles
Figure 7 shows the first BTCs for each retention step as obtained in an experiment involving four consecutive retention–release cycles on soil Amph2. Each cycle includes a pulse application of Cu and metal releases with background electrolyte. The results for Cycle 4 are not shown because they overlapped with those for Cycle 3. As it can be seen, the curve shift decreased and the Cu concentration in the outflow increased with increasing number of cycles. An identical behavior was observed in all soils. This indicates that the retention rate depends on the previous history of retention–release cycles. In other words, the soil retention capacity and the retention rate are adversely influenced by the previous incomplete release of Cu. The incomplete release of Cu affects the adsorption of new Cu ions. The proportions of Cu retained by soil Amph2 in each cycle with respect to the amount of Cu circulated through the chamber were 62, 56, 52, and 50%, respectively. The difference between Cycles 3 and 4 was within the range of experimental error. Figure 8 shows the amount of Cu retained as function of CV for all soils as calculated from Eq. [2], and Table 3 summarizes the amounts retained by the soils in each retention cycle.


Figure 7
View larger version (21K):
[in this window]
[in a new window]

  		Fig. 7. Experimental BTCs (symbols) obtained following application of three successive pulses to soil Amph2 in the retention step. Jw = 0.56 mL min–1. Cin = 0.0787 mM. Lines represent the first-order model. The BTC for the fourth pulse has been omitted for clarity as it overlapped with that for the third.

 

Figure 8
View larger version (20K):
[in this window]
[in a new window]

  		Fig. 8. Copper concentration retained over three (A, C, and D) and four (B) consecutive retention–release cycles in soils Amph1 (A), Amph2 (B), Gra1 (C) and Gra2 (D). Symbols represent data calculated from Eq. [2] and lines the first-order kinetic model. Jw = 0.56 mL min–1. Cin = 0.0787 mM.

 

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

  		Table 3. Amount of Cu retained by the soil (mmol kg–1) after each retention cycle during the stir-flowed chamber (SFC) tests with pulses of [Cu] = 0.0787 mM.

 
No curve shifts were observed during the release steps in any soil (see release BTCs for Amph2 in Fig. 9 ), which indicates that the release rate decreased from the start of the experiment. The amount of Cu released in each step was smaller than that retained in the corresponding retention step. Therefore, a fraction of Cu was retained by the soil. The amount of Cu released increased after each cycle (from 2.65 to 3.04 to 3.43 to 3.57 mmol kg–1, after each pulse of [Cu] = 0.0787 mM). Similar results were found using [Cu] = 0.159 mM (viz. 3.68 to 3.84 to 4.43 to 5.02 mmol kg–1). Increasing the release with cycling suggests that the sorption was not at equilibrium in low reversible sites. These sites should be progressively occupied as the number of Cu pulses passed by the chamber.


Figure 9
View larger version (15K):
[in this window]
[in a new window]

  		Fig. 9. Breakthrough curves for Cu release from soil Amph2 in four successive cycles (1–4). Jw = 0.56 mL min–1. Cin = 0.

 
Based on the high slope of the sorption isotherm at low equilibrium concentrations, there should be long tails in the release process, especially if Cu release from the soil is slow. Since the desorption behavior departs markedly from the sorption behavior, and the desorption BTCs are nearly identical, one can conclude that the kinetics of the slow desorption process is too slow to be measured with the stirred-flow reactor at flow rates around 0.56 mL min–1.

Stirred-Flow Chamber: Kinetic Models
The retention and release of Cu in soils are two complex processes because soil is heterogeneous in nature. The rates of these processes in acid soils involve species that may be influenced by both transport and chemical reactions. In the SFC experiments, the transport in the bulk solution and the transport across the liquid film at the solid-liquid interface is favored by stirring. However, surface diffusion as defined in Aharoni and Sparks (1991) and the transport into intraparticle voids can still be the rate determining steps.

Retention–release rates, dq/dt, have been described in terms of various models for first-, pseudo-first and fractional-order reactions (Amacher et al., 1988; Selim et al., 1990; Amacher, 1991), pseudo-second order reactions (Bhattacharyya and Gupta, 2005), and second-order kinetic exchange reactions with competition from protons (Shi et al., 2005).

The simplest kinetic retention model most accurately describing Cu retention and release in all studied soils was found to be the following first-order function:

Formula 4[4]
where k1 and k-1 are the retention and release rate constant, respectively; qmax, the retained concentration at infinite time; and qf, the retained concentration that can be fast released. This equation is first-order in available sites (qmax– q) for retention and first-order in available fast desorbing Cu for release. Expressions similar to the first term on the right side of Eq. [4] have been used to describe chemical kinetics (Strawn and Sparks, 2000). This expression is the same as the linear driving force (LDF) model of sorption kinetics that, because of its simplicity, it is frequently used in the mathematical simulations of diffusion into a spherical adsorbing particle (Sircar and Hufton, 2000), as well as the mass transfer between the solid and solution phase (Parker and Jardine, 1986). Mathematically the LDF approximation is indistinguishable from first-order adsorption kinetics.

The parameters of the release term in Eq. [4] (second on the right-hand side) were modeled by using the release data. For most of the experiments, the flux controlling solution residence time into the chamber volume (Jw/Ve) was more than two times higher than the retention coefficient of Cu in the soil suspension (k1). This means that the stirred flow system was measuring the sorption rates.

Assuming that the retention rate was negligible during the release step, the first term on the right-hand side of Eq. [4] was zero. The boundary condition for Eq. [3] was Cin = 0. At t > 0, Eq. [4] simplified to

Formula 5[5]
with the initial conditions, dq/dt = 0 and qf = q0 (where q0 is the initial Cu concentration in the soil as calculated from the total mass released in the corresponding release step). Note that qf is the fraction of Cu that is rapidly desorbed. Therefore, Eq. [5] only describes those release processes that are fast enough to be measured within the time frame of the experiments.

The curves in Fig. 4 to 7 are the BTCs calculated from the system of Eq. [3] and [4]. As it can be seen, the model fits the experimental curves, the differences in shape possibly being the result of experimental errors, particularly of partial clogging of the filters or imperfect mixing in the chamber.

Initially, the calculated retention rate exceeded that of Cu in the influent and caused a shift in the BTCs. The retention rate of Cu decreased and its concentration in the effluent increased as q was raised (i.e., as the amount of available adsorption sites, qmax– q, becomes smaller).

The parameter values used (k1 = 0.08 ± 0.04 and qmax = 28 ± 7 for Amph1, and k1 = 0.247 ± 0.05 and qmax = 6.5 ± 0.4 for Gra1) allowed the BTCs to be simulated at variable flow rates (see lines in Fig. 4c and 4d). Table 4 shows the parameter values for the first-order kinetic model (Eq. [4]) in all retention experiments. The range of rate constants was much wider than the millisecond time scale typical of ion exchange kinetics in oxide surfaces (Sparks, 2001), and falls into the range of sorption on humic acids (Bunzl et al., 1976) which suggests that both complexation reactions and transport can be the rate limiting steps.


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

  		Table 4. Fitting of the results to a first-order retention model (Eq. [4]).

 
The optimized values for qmax in the first retention cycle equaled the simulated equilibrium concentration in the isotherms for all soils. In consecutive cycles, better fits were found by gradually reducing qmax at the start of each cycle (qmax values in Table 4). These reductions of qmax decreased the theoretical maximum adsorption rate at the start of each cycle [ (dq/dt)max = k1 x qmax] due to a progressive decrease of vacant adsorption sites.

The k1 estimates were subject to large standard errors. This suggests that mass transfer is diffusion controlled because it is more strongly dependent on fluctuations in the SFC system. The k1 estimates increased with increasing number of cycles. Since Eq. [4] represents an "average" retention process, this suggests that faster processes prevail after a number of cycles. Thus, the sites with the lowest adsorption rates could also be those with the lowest desorption rates. On the other hand, the sites exhibiting faster retention of Cu also showed faster desorption rates. The fitted k–1 values (see Fig. 10 and Table 5 ) were similar in all cases, which suggest that Cu was released via the same mechanism in the four soils studied. In addition, the small differences in released mass (q0) between cycles suggest that the fast release was not dependent on the previous metal additions. We believe that the amount of Cu released may result from exchange with Ca2+. The ratio of q0 to effective cation exchange capacity (CECe) (expressed as a percentage of cmolc kg–1) was 45, 45, 20, and 22% for Gra1, Gra2, Amph1 and Amph2, respectively. These values show that only a fraction of the effective cation exchange capacity is related to the fast desorbed Cu, and that it is twice greater for the granite soils than for the amphibolite soils. We believe that the amphibolite soils can take more Cu from exchange sites to form slowly reversible bonds than can the granite soils. The amphibolite soils contain more reactive surfaces forming complexes with Cu, presumably inner sphere complexes (McBride, 1989), than do the granite soils. On the other hand, the granite soils contain more permanently charged clay minerals where Cu is preferably bound to silanol groups of the lattice face resulting in weak bonds (McBride, 1989). Instead, in amphibolite soils with dominant variable charge, Cu mainly interact with aluminol groups forming stronger bonds.


Figure 10
View larger version (16K):
[in this window]
[in a new window]

  		Fig. 10. Copper concentration retained over three (A, C, and D) and four (B) consecutive release cycles in soils Amph1 (A), Amph2 (B), Gra1 (C) and Gra2 (D). Symbols represent data calculated from Eq. [2] and lines the first-order kinetic model. Jw = 0.56 mL min–1. Cin = 0.0787 mM.

 

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

  		Table 5. Fitting of the results to a first-order release model(Eq. [5]).

 
The fraction fhys = [(qmax1– qmax4)/qmax1], where 1 and 4 denote cycle number (see Table 4), provides an approximation to the ratio of Cu retained by highly hysteretic retention relative to the total sorption capacity as estimated from the first-order kinetics. fhys values ranged from 0.60 to 0.73, which suggests that the fraction of tightly adsorbed forms of Cu was similar in all soils.

These results show that the history of Cu loading can affect the experimental determination of kinetic parameters when desorption is not complete. The hysteretic behavior can be explained by incomplete desorption. Therefore, it affects the determination of kinetic coefficients in soil sorption experiments. In view of these findings, slow desorption release rates should be considered in further SFC experiments.

The heterogeneity of the binding sites in soils can result in both qmax and the sorption rate coefficient being affected by competing cations, pH and Cu concentration in feed solution loading. Using buffered soil suspensions facilitates the study of the chemistry of adsorption with proton exchange. However, we chose to perform tests under conditions closer to those prevailing in the field and allow the soil buffering mechanism to act freely.

Our tests were performed only at pH of 5.5, which is the most usual value for acid soils. At lower (more acid) pH values, the soil buffering capacity would facilitate the release of Al to the solution, thus critically affecting the behavior of Cu. Under these experimental conditions, one cannot ascertain whether the effect on Cu is due to the solution pH or to dissolution of cations to keep the soil pH constant. On the other hand, at higher (more alkaline) pH levels, part of the soil organic matter is dissolved, which also alters the behavior of Cu. Again, this precludes distinguishing between the effects of pH and dissolved organic matter (DOM) on Cu (Shi et al., 2005), but the pH fixed to 5.5 ensures mostly that complexation effects will be kept constant.

The amphibolite soils possess a higher total retention capacity and exhibit more hysteretic retention. In this respect, the disparate behavior of the granite and amphibolite soils can be ascribed to the increased organic C contents in the latter. On the other hand, the greater contents in extractable Fe oxides of Amph1 may have also caused it to retain more Cu than Amph2 did. Despite the differences in their respective properties, the effect of slow desorption acts in the same way in all the soils studied.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Kinetic retention and release tests conducted in a SFC revealed that Cu retention–release dynamics in the studied acid soils is hysteretic. Based on the experimental time-frame used for each cycle in the retention–release cycling tests (170 min), a fraction of the sorption sites exhibit less hysteretic dynamics in the overall sorption process. The retention step is reproduced by a first-order kinetics in available adsorption sites. The factor most strongly influencing the retention kinetics is the previous metal addition. Based on the SFC release tests, the fast release process is less hysteretic than the overall sorption process.

The capacity of the soils for Cu retention depends on previous metal addition cycles whereas its fast release do not. This is consistent with hysteretic behavior of a fraction of the adsorbing Cu. The history of Cu loading can affect the experimental determination of the kinetic parameters when desorption is not complete. Hysteretic behavior can be explained by incomplete desorption. Therefore, the slow desorption release mechanisms should be modeled on further SFC studies.


   ACKNOWLEDGMENTS
 
This work was supported by the Xunta de Galicia under contract PGIDIT06RAG38301PR, and by the Spanish Ministry of Science and Technology under contract AGL2006-04231/AGR, together with Ramón y Cajal contracts awarded to M. A.-E. and E. L.-P. and with a Parga Pondal contract awarded to J.C. N.-M.


   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. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication February 20, 2006.


   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
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by López-Periago, J. E.
Right arrow Articles by Simal-Gándara, J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by López-Periago, J. E.
Right arrow Articles by Simal-Gándara, J.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by López-Periago, J. E.
Right arrow Articles by Simal-Gándara, J.
Related Collections
Right arrow Soil Kinetics
Right arrow Toxic Trace Metals
Right arrow Sorption/Exchange

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