Geochemical and Hydrological Reactivity of Heavy Metals in Soils - Chapter 9 pps

Green CONTENTS 9.1 Introduction 9.1.1 Association of Trace Metals with Mn Oxides 9.1.2 Previous Studies on the Effect of Soil Reduction on Metal Solubility 9.1.3 Influence of Electrolyte

9 Reduction/Cation

Exchange Model of the Coincident Release of Manganese and Trace Metals following Soil Reduction

Dean M Heil, Grant E Cardon, and Colleen H Green

CONTENTS

9.1 Introduction

9.1.1 Association of Trace Metals with Mn Oxides

9.1.2 Previous Studies on the Effect of Soil Reduction

on Metal Solubility

9.1.3 Influence of Electrolyte Concentration and Cation Exchange Reactions

9.1.4 Processes and Reactions Controlling the Solubility of Mn and Trace Metals Following Reduction

9.2 Case Study

9.3 Reduction/Cation Exchange Model

9.3.1 Reduction Model

9.3.2 Cation Exchange Model

9.3.3 Calculation of Cation Exchange Coefficients

9.3.4 Comparison of Model Predictions to Experimental Data

9.3.4.1 Prediction of Ca, Mg, Mn, and Sr

9.3.4.2 Prediction of Ni and Zn

9.3.5 Model Limitations

9.3.6 Applications to Chemical Transport Modeling

9.4 Summary References

L1623_Frame_09.fm Page 229 Thursday, February 20, 2003 10:55 AM

9.1 INTRODUCTION

The effect of soil reduction on trace metal solubility has important implications to both plant availability and toxicity, and chemical transport The release of metals associated with Mn and Fe oxides following reductive dissolution is an important mechanism that can potentially increase the soluble concentrations of metals.1,2 The potential for the release of trace metals following soil reduction appears to be the greatest for slightly to moderately reduced soils, with redox potentials between 100 and 400 mV Under these conditions, redox potentials are sufficiently low to dissolve

Mn or Fe oxides, but not low enough to precipitate metal sulfides In highly reduced soils with redox potentials less than approximately 0 mV, the precipitation of metal sulfides limits the soluble concentration of trace metals.1,3 Dissolution of Mn oxides precedes Fe oxide dissolution because of the lower redox potential required to dissolve Fe oxides.4 Manganese (IV) oxides become unstable at a redox potential (EH) of approximately 300 mV, whereas Fe (III) oxides are stable until EH decreases

to less than 100 mV, with the exact values of EH required to initiate reductive dissolution dependent on pH Consequently, the dissolution of Mn oxides may play

a more important role in metal solubilization in the early stages of soil reduction when redox potential is low enough to dissolve Mn oxides, but Fe oxides may still

be stable

9.1.1 A SSOCIATION OF T RACE M ETALS WITH M N O XIDES

Manganese oxides have a high affinity for many of the trace metals.5,6 In addition

to surface adsorption, trace metals accumulate in Mn oxides by substitution and co-precipitation.7 The adsorptive properties of Mn oxides for metals observed in the laboratory are verified in soils, as Mn oxide nodules separated from soils contain concentrations of trace metals that are considerably greater than the metal concen-trations in the bulk soil.7,8 The potential for association of trace metals with Mn oxides via co-precipitation or substitution is high when soils are subject to alternate wetting and drying cycles,9 and Mn oxide crystals are forming

9.1.2 P REVIOUS S TUDIES ON THE E FFECT OF S OIL R EDUCTION

ON M ETAL S OLUBILITY

Several researchers have reported an increase in the soluble concentrations of trace metals under reducing conditions Chuan et al.10 found that the release of soluble

Pb, Cd, and Zn from a soil increased as EH was decreased from 325 to −100 mV at

a constant pH Davranche and Bollinger11 observed that Pb and Cd adsorbed to synthetic Mn or Fe oxide was released into solution as the solid phases were progressively dissolved by increasing concentrations of a reducing agent The desta-bilization of Fe and Mn oxides following the addition of a reducing agent to a contaminated soil caused an increase in the soluble concentrations of both Cd and

Pb.11 Soil adsorbents not dissolved by reductive dissolution were considered to have

a large effect on the solubility of the metals, as Cd concentrations did not increase substantially until pH was less than approximately 6, and Pb did not increase until

pH was less than 4 The authors noted that this difference in the behavior of Cd and

L1623_FrameBook.book Page 230 Thursday, February 20, 2003 9:36 AM

Pb is consistent with a 2-pH unit difference in the adsorption edges of these two metals for natural colloids Charlatchka and Cambier2 reported that soluble concen-trations of Pb and Cd increased with time in flooded soil cores, and concluded that

a decrease in pH caused by reduction processes played a critical role in elevating soluble metal concentrations Incubation of the soil in a pH stat-redox cell revealed that at fixed pH, soluble concentrations of Pb, Cd, and Zn increased with incubation time, coinciding with a decrease in redox potential.2 Destabilization of Mn and Fe oxides was considered to be an important mechanism for the release of trace metals under steady pH In cases where pH increases following reduction, trace metal solubilities have been observed to decrease Kashem and Singh12 reported that for all three soils that were studied, the soluble concentrations of Cd and Zn decreased following saturation, and Ni decreased in two of the three soils These decreases in metal solubility coincided with a pH increase in all three soils, and were attributed

to enhanced sorption and possibly greater stability of metal oxides or other minerals

at higher pH

9.1.3 I NFLUENCE OF E LECTROLYTE C ONCENTRATION AND C ATION

E XCHANGE R EACTIONS

Another mechanism that could influence the solubilization of trace metals under reducing conditions could be the displacement of exchangeable metals by high concentrations of dissolved Mn released following the dissolution of Mn oxides Soluble Ca and Mg have been observed to increase following soil reduction, and this has been attributed to the displacement of those cations from exchange sites by dissolved Mn and Fe.13,14 This process may be described by the reaction:

CaX2 + Mn2+ = MnX2 + Ca2+ (9.1) The increased concentration of divalent cations could also be expected to displace trace metals from cation exchange sites as well as Ca and Mg Although the exchangeable metal concentrations in many soils are low, exchangeable metal con-centrations are generally much greater than soluble metals concon-centrations, and could act as a source to the solution phase under reducing conditions when electrolyte concentration (EC) is increased

9.1.4 P ROCESSES AND R EACTIONS C ONTROLLING THE S OLUBILITY

OF M N AND T RACE M ETALS F OLLOWING R EDUCTION

The soluble concentration of Mn under reducing conditions will depend on the amount of Mn oxide dissolved and the extent of re-precipitation or adsorption of the released Mn(II) to soil colloids Xiang and Banin15 found that a significant fraction of the Mn released by Mn oxide dissolution within 3 days of saturation was redistributed to cation exchange sites Manganese can also be retained by specific adsorption to Fe oxides, organic matter, and layer silicate clay minerals.5,16,17 A review of the mechanisms controlling adsorption of Mn to soil constituents is provided by Khattack and Page.18

L1623_FrameBook.book Page 231 Thursday, February 20, 2003 9:36 AM

The solubility of trace metals under reducing conditions will depend on the amount of Mn oxide dissol ved, the concentration of the trace metals initially asso-ciated with the Mn oxide fraction, and the retention of the released metals to soil solids following dissolution of the Mn oxide Trace metals initially associated with the Mn oxide fraction may also be retained by specific adsorption reactions,19 involving surface hydroxyl sites on mineral and organic soil colloids The partition-ing of metals between these sites and the solution phase is highly dependent on pH,

as predicted by the following reaction20:

=SOH + Mz+ = =SOM(z-1)+ + H+, (9.2) where =SOH is a surface hydroxyl functional group In cases where pH changes significantly during soil reduction, we can expect that it will be necessary to include specific adsorption reactions to model the changes in solubility of both Mn and trace metals

9.2 CASE STUDY

The solubility of Mn, Zn, Ni, and Sr following saturation of soil columns was studied for two soils collected from the Alamosa River Basin, Colorado These two soils are classified as the LaJara (coarse-loamy, mixed (calcareous), frigid typic hapla-quolls) and Mogote (fine-loamy, mixed (calcareous), frigid aquic ustorthents) series Soils in this region have a history of irrigation with water impacted by acid mine drainage Basic soil chemical and physical properties are shown in Table 9.1 Sam-ples were collected from the base of the soil columns at 12-h intervals up to 84 h This time frame was chosen to simulate the period of saturation following flood irrigation of soils in this region Details of procedures are described by Green.21 Reduction experiments are often performed with the addition of a carbon source to accelerate a decrease in EH The data used in the present model are from treatments that did not receive an amendment with an additional carbon source

TABLE 9.1

Chemical and Physical Properties of LaJara and Mogote Soils

Soil pH a

CEC b (cmol kg

−−−−1 )

CCE c (g kg

−−−−1 )

Fe ox d (g kg

−−−−1 )

OC e (g kg

−−−−1 )

Sand (%) Silt (%)

Clay (%)

a 24-h 1:1 soil:water pH.

d Iron oxide content.

L1623_FrameBook.book Page 232 Thursday, February 20, 2003 9:36 AM

The results of these studies are described by Green et al.22 The major findings

are summarized here The soluble concentrations of Mn, Zn, Ni, and Sr increased

in both soils following reduction Total electrolyte concentration also increased

following reduction, and this change was due mainly to increases in soluble Ca and

Mg concentrations Redox potentials decreased to values that were sufficient to

initiate the dissolution of Mn oxides within 24 h after saturation, and remained

nearly constant through 84 h Iron oxides were apparently stable under the redox

conditions and time frame of our experiments, as increases in soluble Fe were not

observed Total electrolyte concentration (EC) also increased continuously

through-out the 84-h saturation period, with most of the change in EC associated with

increased concentrations of soluble Ca and Mg Soluble concentrations of Pb and

Cd were also measured, but were below the instrument detection limits for many

samples For these reasons, we chose to test the fit of the data from these experiments

to a cation exchange model including Mn, Ca, Mg, Sr, Ni, and Zn The data used

to test the cation exchange model were taken from the average of duplicate columns

for each of the two soils

9.3 REDUCTION/CATION EXCHANGE MODEL

9.3.1 R EDUCTION M ODEL

The amount of Mn oxide dissolved over an 84-h time period for each experiment

was calculated based on the assumption that the observed increase in EC was due

to displacement of exchangeable cations by dissolved Mn2+ Electrolyte

concentra-tion was calculated by summing the contribuconcentra-tions from the divalent caconcentra-tions:

EC = 2 ([Ca2+] + [Mg2+ ] + [Mn2+] + [Sr2+] + [Ni2+] + [Zn2+]), (9.3) where EC is in molc l−1 Although soluble concentrations of K and Na changed

slightly between 24 and 84 h, these cations were not included in the calculation of

EC as they were not included in the cation exchange model The exclusion of Na

and K from the cation exchange model was based on the observation that Na and

K accounted for only 9% and 7%, respectively, of the total increase in electrolyte

concentration in the LaJara and Mogote soils The total concentration of Mn

dis-solved between 24 and 84 h was calculated as

[Mn]d = (EC84− EC24)/2 (9.4) with [Mn]d expressed as mol l−1 A constant dissolution rate of Mn oxides was also

assumed by dividing the total Mn dissolved into five 12-h intervals, beginning with

24 h at which time dissolution of Mn(IV) oxides began This yielded a concentration

of Mn of 1.89 E-4 M for the LaJara soil and 4.10 E-4 M for the Mogote soil for

each 12-h interval In terms of the concentration of Mn in the soil, the amount of

Mn dissolved after 84 h was 5.41 E-4 mol kg−1 for the LaJara soil, and 1.17 E-3

mol kg−1 for the Mogote soil Compared to the total concentration of reducible Mn

(Table 9.2), this represents 30% of the reducible Mn oxide for the LaJara soil, and

L1623_FrameBook.book Page 233 Thursday, February 20, 2003 9:36 AM

51% for the Mogote soil For each time period, the total Mn concentration available

for cation exchange reactions, in units of molarity, was calculated as

[Mn]T(t) = [Mn]i + [Mn]d(t), (9.5) where [Mn]i is the initial total concentration of Mn available for cation exchange

reactions at the beginning of the experiment, and [Mn]d(t) is the concentration of

Mn(II) released by dissolution of Mn oxide after each time period The initial

concentration of Mn was calculated as

[Mn]i = [Mn]exch + [Mn]s (9.6) with [Mn]exch the initial concentration of exchangeable Mn, and [Mn]s the

concen-tration of soluble Mn from the saturated columns at the 24-h time period The

concentrations of exchangeable metals were converted from mol kg−1 to mol l−1 by

multiplying by the solid:solution ratio of the soil at saturation, which was 1.7 kg l−1

for both soils The total concentrations of Ca and Mg were fixed based on the initial

exchangeable and soluble concentrations, as in equation 9.6

For modeling the release of Sr, Ni, and Zn, two approaches were taken In the

first model, the total concentration of Sr, Ni, and Zn available for cation exchange

reactions was fixed as the sum of the initial concentrations of soluble and

exchange-able concentrations as in equation 9.6 In the second model, the concentrations of

Zn, Ni, and Sr released by dissolution of Mn oxide were included and were

consid-ered to be available for cation exchange reactions The corresponding equation for

the total concentration of each metal for the second model was

[M]T(t) = [M]exch + [M]s + fM,Mn-ox [Mn]d(t) (9.7) where fM,Mn-ox is the number of moles of metal M per mole of Mn in the Mn oxide

fraction The quantity of metals associated with the Mn oxide fraction used to

TABLE 9.2

Metal Concentrations in Exchangeable, Mn-Oxide, and Total Fractions of

LaJara and Mogote Soils a

LaJara (mol kg −−−−1 ) Mogote (mol kg −−−−1 ) Exchange

able Mn-oxide Total

Exchange able Mn Oxide Total

L1623_FrameBook.book Page 234 Thursday, February 20, 2003 9:36 AM

calculate fM,Mn-ox for each metal was measured by a sequential extraction procedure23 (Table 9.2)

9.3.2 C ATION E XCHANGE M ODEL

Cation exchange reactions involving Ca, Mg, Mn, Sr, Ni, and Zn were modeled based on the following reaction24:

CaX2 + M2+ = MX2 + Ca2+ (9.8) The cation exchange equation corresponding to this reaction is

(9.9)

where Kex is the cation exchange coefficient The cation exchange reaction may be separated into two component half-reactions to facilitate computer modeling25,26:

M2+ + 2 X− = MX2, (9.10) with the equilibrium constant for the formation of MX2 represented by Kf The corresponding mass balance equation for cation exchange sites as applied to this problem was

XT = 2 (CaX2 +MgX2 + MnX2 + SrX2 + NiX2 + ZnX2) (9.11) where XT is the total concentration of cation exchange sites in molc kg−1 In order

to represent fixed charge sites where the concentration of uncomplexed X− is essen-tially zero, the convention used by Stadler and Schindler26 was followed, with the log Kf for the formation of CaX2 in equation 9.10 set equal to 20.0 We verified that following this convention resulted in less than 0.1% of exchange sites unoccupied

by cations, and modeling results were not dependent on the value of log Kf for CaX2 for values between 10 and 20 The log Kf values for equation 9.10 for cations other than Ca were obtained by adding the value of log Kex to 20.0 The MINTEQA2 computer speciation program27 was used for modeling

9.3.3 C ALCULATION OF C ATION E XCHANGE C OEFFICIENTS

Values for cation exchange coefficients were calculated using exchangeable cation concentrations from 1-h 1-M KCl extraction (Table 9.2) The quantity of exchange-able cations were calculated based on the surface excess of each cation24:

qi = ni − Mw mi (9.12)

K ex= MX Ca++ CaX M 2 2

2 2

where qi is the surface excess in mol kg−1, ni is the total number of moles of the cation extracted per kilogram of dry soil, Mw is the gravimetric water content of the slurry, and mi is the molality of the cation in the supernatant solution We obtained values of mi by performing 1-h extractions with water at the same solid:solution ration as for the 1-M KCl extractions We found that correction for the concentration

of soluble cations significantly reduced the calculated exchangeable concentrations

of Zn and Ni, and this would have a substantial effect on the model predictions for these elements if uncorrected values were used Data for soluble metal and cation concentrations were taken from the 24-h time period of the reduction experiments The soluble cation and metal concentrations at 24 h in the soil columns were similar

to the 1-h batch, water-soluble concentrations Values for Kex for the overall cation exchange reactions, based on Ca as the initial cation occupying exchange sites corresponding to equation 9.8 were first calculated from experimental data (Table 9.3) For modeling purposes, values for log Kf for the half-reactions for each metal corresponding to equation 9.10 were then determined (Table 9.3)

9.3.4 C OMPARISON OF M ODEL P REDICTIONS TO E XPERIMENTAL D ATA

9.3.4.1 Prediction of Ca, Mg, Mn, and Sr

The concentrations of soluble Ca, Mg, and Sr were consistent with the cation exchange model for both soils studied (Figures 9.1 to 9.4) Experimental Ca concentrations were greater than model predictions between 36 and 60 h; this could be a result of deviation from linear dissolution of the Mn oxide as was assumed in the model Soluble Mn concentrations increased by a factor of 5 times between 24 and 84 h for the LaJara soil (Figure 9.2) and 28 times for the Mogote soil (Figure 9.4) The greater relative increase in soluble Mn concentration in the Mogote soil is a result of the increased amount of Mn oxide dissolved for that soil, as noted above This is consistent with the higher amount of reducible Mn in the Mogote versus LaJara soils (Table 9.2), as well as a slightly lower EH in the Mogote soil during reduction.21 The model predicted these changes in Mn solubility with fair accuracy, with the soluble Mn at

84 h underestimated by the model by 17% for the LaJara soil, and overestimated by the model by 20% for the Mogote soil Comparison of the soluble Mn concentrations

TABLE 9.3

Cation Exchange Coefficients and Equilibrium Constants for LaJara and Mogote Soils

K ex log K f K ex log K f

at 84 h to the total amount of Mn released by Mn-oxide dissolution reveals that approximately 6% and 3%, respectively, of the released Mn remained in solution for the LaJara and Mogote soils, with the balance retained by cation exchange sites The cation exchange capacity (CEC) of these two soils is very similar (Table 9.2) There-fore, the tendency for a smaller fraction of the dissolved Mn to be proportioned to exchange sites in the LaJara versus the Mogote soils is probably due to the higher amount of initial exchangeable Mn in the LaJara soil (Table 9.2) The addition of Zn,

FIGURE 9.1 Soluble concentrations of Ca and Mg from LaJara soil.

FIGURE 9.2 Soluble concentrations of Mn and Sr from LaJara soil.

0.0E+00

2.0E-04

4.0E-04

6.0E-04

8.0E-04

1.0E-03

1.2E-03

1.4E-03

1.6E-03

Hours

Ca model

Ca experiment

Mg model

Mg experiment

0.0E+00

1.0E-05

2.0E-05

3.0E-05

4.0E-05

5.0E-05

6.0E-05

7.0E-05

Hours

Mn model

Mn experiment

Sr model

Sr experiment

Ni, and Sr from the Mn oxide fraction caused only slight changes in the predicted concentrations of Ca, Mg, Mn, and Sr The model results shown in Figures 9.1 to 9.4 were plotted using the results of the model without the addition of Ni, Zn, and Sr initially associated with Mn oxides However, these results are representative of both models for those elements Although Sr is present in the Mn oxide fraction, the amount

of Sr released as a result of Mn oxide dissolution is small compared to the initial

FIGURE 9.3 Soluble concentrations of Ca and Mg from Mogote soil.

FIGURE 9.4 Soluble concentrations of Mn and Sr from Mogote soil.

0.0E+00

5.0E-04

1.0E-03

1.5E-03

2.0E-03

2.5E-03

Hours

Ca model

Ca experiment

Mg model

Mg experiment

0.0E+00

1.0E-05

2.0E-05

3.0E-05

4.0E-05

5.0E-05

6.0E-05

7.0E-05

8.0E-05

Hours

Mn model

Mn experiment

Sr model

Sr experiment