Cation Exchange Capacity and Exchange Coefficients INTRODUCTION

Transcription

1 Cation Exchange Capacity and Exchange Coefficients M. E. SUMNER AND W. P. MILLER, University of Georgia, Athens, Georgia Chapter 40 INTRODUCTION Origin of Charges on Soil Particles The origin of cation exchange capacity (CEC) lies in the negative charges carried by soil particles, usually clay, organic matter and sesquioxides. A full discussion of the origin and nature of these charges is presented in Chapter 41 (Zelazny et ai., 1996). Basically these charges fall into two distinct categories, being either permanent or variable (i.e., ph dependent) depending on whether or not ambient conditions (ph or salts) in the soil solution affect their magnitude. Much confusion in the literature concerning the measurement and interpretation of CEC has stemmed from the lack of recognition that these charges fall into two distinct categories exhibiting different behavior. These problems will be addressed in the discussion of the methods for the determination of CEC. Operational Definition of Cation Exchange Capacity Until relatively recently, CEC was defined in rather simple terms in elementary texts as "the sum total of the exchangeable cations that a soil can absorb" (sic) (Brady, 1984) or "the capacity of a soil to hold cations" (Jones, 1982), which does not recognize the importance of the procedure used in determining the outcome of the CEC measurement. The negative charges on soil particles are neutralized by an excess of cations and a deficit of anions (negative adsorption or anion repulsion) over that which would be present in the absence of the solid phase. This essentially means that the major part of the negative charge attracts cations for neutralization and the remainder is involved in repelling anions, which is equivalent to attracting cations. Thus CEC should be defined as CEC = Mt~cess + Ab-;;ficit [1] Copyright 1996 Soil Science Society of America and American Society of Agronomy, 677 S. Segoe Rd., Madison, WI 53711, USA. Methods of Soil Analysis. Part 3. Chemical Methods-SSSA Book Series no

2 1202 SUMNER & MILLER where MX+ and AX- are the cations and anions in the system, expressed on a charge basis (Le., cmole kg-1 soil for cations retained, cmol a kg- 1 for anions). Fortunately the magnitude of the anion deficit is quite small in many cases, varying from 5 to 20% of the CEC at an equilibrium electrolyte concentration of 0.1 M to 0.5 to 2% at M, and thus can be and often is ignored. In addition, all soils contain salts in the soil solution in equilibrium with the exchange sites. During the determination of CEC, these salts are extracted with the exchangeable cations and can result in errors. Thus M~teess in the system would be equal to the total MX+ extracted less the product of the moisture content and equilibrium concentration of the soil solution with respect to MX+. In systems low in salt, this correction is very small and often can be neglected. To be unambiguous in the definition of CEC, one should state that it is the total charge excess of cations over anions in the system. Although this appears simple, many complications arise during measurement. The discussion so far has assumed that the charges on the soil particles are not affected by conditions in the ambient solution. From Chapter 41 (Zelazny et ai., 1996) on charging, the variable charge component is very definitely altered by both the concentration and valence of the ions in the equilibrium solution. In addition, specific adsorption of both cations and anions can have marked effects on CEC. It therefore becomes clear that the CEC of a soil depends rather critically on the manner and conditions in which it is determined. Consequently, CEC always must be operationally defined in terms of variables such as ph, concentration, nature and valence of the ions and buffer capacity of the extracting salt solution, the nature of the washing liquid to remove entrained salt solution and the temperature. In addition, the purpose for which the CEC is being determined must be taken into account. For example, the method selected depends on whether one seeks a value at the "field" ph and electrolyte concentration of the soil, or a value reflecting the conditions after the soil has been limed to ph 7. Historical Review of Cation Exchange Capacity Methods Following Thompson's (1850) discovery of the cation exchange phenomenon, Way (1850, 1852) developed methodology which demonstrated both the permanent and variable charge character of soils. But it was not until 1887 that the first quantitative method for its measurement was developed (Kelley, 1948) involving extraction with saturated NH 4CI by repeated decantation and estimation of the CEC as the loss of NUt from the solution; this method proved to be subject to substantial error. In nonsaline soils, the sum of cations in the leachate would give the CEC. This value, including exchangeable Al3+ determined by an unbuffered salt solution, later became known as the Effective Cation Exchange Capacity (ECEC). Subsequent improvements featured leaching with 1 M NH 4 CI and washing, initially with water and later with methyl (CHzOH) or ethyl alcohol (CzHsOH) to remove the excess NH 4CI entrained in the soil prior to estimation of NHt adsorbed (Kelley & Brown, 1924, p. 1-34). Although not realized at the time, the use of an unbuffered salt in this methodology was appropriate for the estimation of CEC at "field" ph, which is particularly important in highly weathered acid soils as we shall see later. At this stage, the problems presented by the

3 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1203 presence of salts, both soluble and insoluble, in the stoichiometry of cation exchange were known. Schollenberger & Dreibelbis (1930) drew attention to the so-called advantages of using a buffered salt such as CH 3 COONH 4 (ammonium acetate, NH40Ac) at ph 7 and washing with neutral ethanol rather than water to prevent "hydrolysis" of part of the NHt adsorbed (Kelley, 1948; Schollenberger & Simons, 1945). In hindsight, this was a particularly unfortunate selection of extracting solution for acid, highly weathered soils because the buffered salt automatically raises the ph of the soil to 7, greatly inflating the CEC likely to exist in the soil under field conditions. The reason for the selection of ph 7 and a buffered salt was probably an accident of location, as at this time most of the work on CEC was being conducted in western Europe, Russia and the western and midwestern USA, which were the most important areas of agricultural production and where the soils are generally neutral to alkaline in reaction. In addition, the nature of soil acidity was not clearly understood even as late as 1948, as illustrated by the statements of Kelley (1948): "the exchange capacity of a soil depends on the ph of the solution used in the determination," and "is now widely used in America as a measure of the CEC of acid, neutral and alkaline soils," and "is especially well adapted to this determination." Unfortunately, many of these sentiments persist even to this day, and the 1 M NH 4 0Ac (ph 7) method is still widely used even on highly weathered, variable charge, acid soils, under which conditions, it gives highly inflated values for CEC (Grove et ai., 1982; Gillman & Hallman, 1988). Despite the fact that it was selected as a standard method for the classification of soils (Soil Survey Staff, 1975), the continuing use of this method on acid soils is inappropriate for a clear understanding of their exchange chemistry. An excellent parallel discussion of how the views on soil acidity developed during this period is presented by Thomas (1977), which should be consulted for an appreciation of their impact on CEC estimation. It is interesting to speculate what course CEC method development might have taken had the work been conducted in the humic tropical areas of Latin America, Mrica and Asia where the soils are generally highly acid. At this stage in the development of CEC methods, there seemed to be a greater realization of the potential for underestimation of the CEC using 1 M NH 4 0Ac (ph 7) on alkaline soils than for its gross overestimation in the case of acid soils (Kelley, 1948). Again, this seems to be a quirk of location in that Kelley conducted most of his work in the western USA where many soils are alkaline. During this period, many workers developed variations of the methodology for use on calcareous soils to reduce the effect of CaC0 3 solubility in the extracting solution on the estimation of exchangeable Ca2+ with the most successful and widely adopted technique being developed by Mehlich (1942). This involved the use of 0.1 M BaCl 2 buffered to ph 8.2 with triethanolamine (C 6H 1SN0 3 ) (TEA), in which CaC0 3 is practically insoluble. Because the Ba2+ adsorbed by the soil is a measure of its CEC, the presence of SOa- in the soil results in inflated values. Bascomb (1964) modified this approach to incorporate the "compulsive exchange" principle to replace the adsorbed Ba2+ efficiently with MgS04. Unfortunately Mehlich (1945) recommended the use of the BaCITTEA method on acid soils using a back titration of the leachate to estimate exchangeable H+, and this

4 1204 SUMNER & MILLER has lead to much confusion among users not entirely familiar with the principles involved. Most of the W replaced by BaCI2-TEA do not exhibit acid character at the "field" soil ph, and are merely a reflection of the magnitude of the potential variable charge on the soil. During this period, base saturation ([ exchangeable cations/ceq x 100) became a much-used parameter for assessing the acid character of soils. With our current knowledge of the variable charge character of many acid soils, it becomes clear why the use of CEC values obtained by the 1 M N~OAc (PH 7) and BaCI2-TEA methods resulted in unrealistically low values for base saturation. In the case of soils containing soluble salts, no special problems were presented in terms of measuring CEC by a replacing cation provided that the salts were removed in the process. However, it was usually impossible to arrive unequivocally at the CEC of such soils by summation of exchangeable cations (Kelley, 1948). The next step forward was taken by Schofield (1939), who revisited variable charges on soil particles and developed a method for the measurement of positive and negative charges at different ph values using a saturating solution of 0.2 M NH 4 CI, weighing the entrained excess for which correction was made and replacing with 0.2 M KN03 (Schofield, 1949). Subsequently, Sumner (1963) compared the effects of correcting for the entrained solution and its removal by washing with ethyl alcohol (CH30H)/water mixtures on the CEC obtained. The acid soils studied lost adsorbed cation and anion in equivalent amounts as washing proceeded, indicating that as the ionic strength decreased, positive and negative charges were mutually neutralizing each other. Because a low electrolyte concentration is likely to be the normal situation in soils under field conditions, this charge neutralization will lead to a more realistic measure of the actual CEC of the soil. Unfortunately, this washing procedure results in a final electrolyte concentration which is unknown and may vary considerably. In addition, there is always the possibility that some of the exchangeable cations might be hydrolyzed from the exchange sites as the electrolyte concentration approaches zero. Nevertheless, this washing step with water was shown to measure the net charge on the soil which, for most soils, corresponds to the ECEC as originally proposed by Coleman et al. (1959) as the sum of exchangeable Ca2+, Mg2+, K+, Na+, and Al3+ (Grove et ai., 1982). The ECEC approach is not valid for soils containing appreciable levels of salts. In such cases it is impossible to distinguish unambiguously between the cations of the salt and those which are exchangeable. The above work set the stage for the development of cation exchange methods which, in addition to considering the effects of buffering and ph on the results, also recognized the profound effect of the equilibrium electrolyte concentration. The first method was developed by Gillman (1979) and considerably improved by Gillman & Sumpter (1986a). It is based on the "compulsive exchange" principle originally enunciated by Bascomb (1964). Basically, the soil is saturated with a 0.2 M BaCl2 solution followed by reduction of the electrolyte concentration to M BaCl2 (I = M), estimation of the amount of entrained solution and then compulsive exchange with M MgS0 4, which results in Ba2+ being quantitatively replaced by Mg2+ on the exchange sites. The

5 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1205 CEC is calculated by difference after measuring Mg2+ remaining in solution after the compulsive exchange and correcting for the Ba2+ in the entrained solution. The value obtained is called the CEC CE and is equal to the capacity of the soil to retain basic cations (CEC B ) measured by Ca2+ in soils not containing Al3+ (Gillman & Hallman, 1988). Under acid conditions when Al3+ is present, the total CEC (CEC T) is calculated as CECB + Al3+. It also is possible to measure corresponding values for anion exchange capacities (AEC B andaecce) by measuring the adsorption of Cl. In principle, this compulsive exchange method should prove to be universally applicable to all soils provided an appropriate equilibrium ionic strength is selected. Criteria for Selection of Cation Exchange Capacity Methodology Today it is clearly understood that CEC is operationally defined and tailored to the purpose to which the results are to be applied and to the peculiarities of various catagories of soils (Bache, 1976; Rhoades, 1982; Uehara & Gillman, 1981). To facilitate the following discussion, the criteria for methodology selection will be divided on a final use basis. Characterization of Soils in "Field" State From the point of view of describing processes in field soils that may be affected by CEC, there is little doubt that the best approach is to use methods involving unbuffered extracting solutions so that the CEC is measured at the "field" ph value of the soil. Methods involving water/alcohol washes to remove entrained electrolyte are acceptable, but those which control and correct for the ionic strength of the final equilibrium solution are likely to give more meaningful results. Depending on the soil, the ionic strength of the final equilibrium solution should be adjusted to approximate that of the natural soil solution. For soils which are at or above ph 7 or contain little variable charge material, appropriately buffered extracting solutions can be used with water/alcohol washing. In soils which do not contain appreciable levels of soluble salts or free CaC03, CEC can be estimated as the sum of exchangeable Ca2+, Mg2+, K+, Na+, and Al3+, expressed in moles of cation charge per soil mass (e.g., cmoi.: kg-1 soil). Soil Classification If the objective is to compare CEC values of soils as an aid to soil series classification, a strong case can be made for conducting the measurements at a standard ph value. If this is not done, anthropogenic effects such as liming and fertilization could cause the same soil to have widely different CEC values, particularly if it contains appreciable amounts of variable charge surfaces. It should be realized that the measurement of the CEC values of acid variable charge soils at a standard neutral to alkaline ph value will result in values very different from those extant in the field. In such circumstances, a strong case could be made for measuring the CEC at both the standard and field ph values. Having opted to measure CEC at a standard ph, its value must be decided. Good arguments can be made in favor of both the commonly used ph 7 and 8.2

6 1206 SUMNER & MILLER standards which were selected based on considerations of the equilibria between CO 2 and free CaC0 3. The ph of a fully base saturated soil in equilibrium with atmospheric CO 2 and excess CaC0 3 is 8.2, whereas that of a biologically active topsoil in equilibrium with a higher CO 2 content (1%) and free CaC03 is ph 7.3 (Bache, 1976). Cation Exchange Selectivity Coefficients The study of cation exchange reactions has held a special place of interest within soil chemistry dating from the early part of the twentieth century. Some of the first applications of chemical thermodynamics to soil systems were made to cation exchange reactions, in an effort to understand the equilibrium relationships between ionic composition of the solution and the surface (adsorbed, exchangeable) cations (Sposito, 1981a). Interest has continued to this day, as evidenced by the continued publication of papers on this topic in the scientific journals. In the context of this discussion, the term "cation exchange reaction" is to be understood as the relatively rapid, reversible replacement of one cation held by the negative charge of a soil colloid by another cation. "Selectivity" results from some combination of properties of the particular cation and/or the colloid exchanger that may enhance adsorption of one ion in preference to another. The study of exchangeable selectivity is central to the description of the cationic composition of the soil exchange phase, which itself is a key property in the management of soils. The study of soil K reactions, including the quantity/intensity (Q/l) studies of Beckett (1964) which were based in cation exchange theory, have attempted to understand the K fertility status of soils. The continued interest in Na exchange stems from the importance of Na in determining the physical properties of semiarid and irrigated soils with respect to dispersion and hydraulic conductivity (Shainberg & Letey, 1984). And more recently, studies of trace metal adsorption and solute transport have used cation exchange reactions to predict the solubility and transport of contaminants in an environmental context (Gaston & Selim,1990). The continued interest in cation exchange selectivity, both in practical use and in theoretical studies, has been hampered to some degree by varying procedural approaches to data collection and manipulation used in the literature. This variety of methods is based in historical differences in conceptualization of the exchange process, and while none is necessarily superior a priori to another, clear delineations need to be made between them. This section will attempt to define these different approaches, and suggest a critical methodology for data collection applicable to each. Formulation of Exchange Reactions Binary cation exchange reactions may be formulated in a number of ways; for the case of monovalent/divalent exchange, the replacement of one cation held on an exchange surface (symbolized as X- a monovalent exchange site) by another cation from the solution phase may be formulated as

7 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS NaX + Ca 2 +? cax 2 + 2Na+ [2] where X- represents 1 mole of surface negative charge. If we assume this is a reversible reaction that can be treated thermodynamically, an equilibrium expression for the constant of exchange can be written as [3] If we specify in this equation that ( ) are to indicate activities of the various solution and surface species, then this expression represents the true thermodynamic equilibrium constant for the reaction. Activities of soluble ions are readily calculated using the Davies approximation of the Debye-Huckel theory as where [4] log y, = -1).51 z,[ 1 ~'I [5] and 1= 0.5 Li(ZrCi) [6] where ai = activity of species i (moles L-1), Ci = concentration of i (moles L-1), Yi = activity coefficient of species i, Zi = valence of cation i, and I = ionic strength of the solution. Activities for adsorbed species may be expressed analagously as as.i = ~/;, where surface quantities are represented as mole fractions (Mi> moles/total adsorbed moles) and/is the activity coefficient ofthe adsorbed phase. The use of molar quantities and mole fractions for adsorbed species has been termed the Vanselow convention by Sposito (1977), after the early pioneer in exchange chemistry, while if equivalents of charge and equivalent fractions (Ei) are employed, such formulations may be referred to as using the Gapon convention. Values of/can be computed from exchange data (Sposito, 1981a), but in practice surface activity coefficients are often neglected by factoring them out of the equilibrium expression. Thus, Eq. [3] may be written where K eq =(~lk E 2 c J(NaX) [7] [8]

8 1208 SUMNER & MILLER ' ' ~ " 0.7 U) c:: 0.6 ~ u <1) "",. '5 0- ~ UJ/j ::;:--; f:},7 /,"... "-///,/,,/... -"'. / /" //.,././ /',/./ / /,/,/ /' / / / // / / / // / / / / I / /. / / I / / I / /... / / /.' / I / /1... /. Total Normality 1/ /... / 0.02SM.Na Ca- / I 1.'-/ D.OSOM,Na Ca-- / 1 /... / 0.100M. Na.Ca -- I I... / 0.250M. Na Ca --_. (I I./ 1/ )' I/./ (/ M, Na Ca All cone., Mg-Ca ECa (equiv. fraction, soln) Fig "Ideal" (nonpreference exchange isotherms for Na+-Ca2+ and Mg2+_Ca2+ exchange (after Sposito, 1981b). Kc is a conditional equilibrium constant, in that it may vary with the composition of the exchange phase (Le., as Mj varies from 0 to 1); this variation is due to the fact that the surface activity coefficients (h) may change with different surface coverages (Mj ) of the cations involved. The expression for Kc in Eq. [8] was first proposed by Vanselow (1932), and is referred to as the Vanselow selectivity coefficient, Kv' The term "coefficient" is used to reflect the conditional, or variable nature of this value with composition of the exchange phase. A plot of Kv vs. Ej (equivalent fraction of component i) or M j may show variation in Kv with exchanger composition, indicating that the ratio of activity coefficients (fna}l!cax) is not constant as the composition of the exchange phase varies. This "non-ideal" behavior, defined as /; i; t i; 1 (Sposito & Mattigod, 1979), is commonly observed for exchange on both reference minerals and soil clays; Sposito (1981b), in fact, points out that mixtures of minerals present in soil clays will never behave ideally due to the differential selectivity of the components over a range of cation compositions. A value of Keq may nevertheless be evaluated as the integral of In Kv over the range of Ej or Mj from 0 to 1 (that is, the area under the curve of a plot of Kv vs. Ej). Free energies of exchange (~G) and other thermodynamic parameters may then be computed using this value of Keq. Preferential adsorption of one cation relative to another is indicated when the value of Keq 1= 1. A plot of Ej (equivalent fraction of cation i in solution) vs. E j (equivalent fraction of i on surface) also may be used to evaluate cation preference. For mono-monovalent or di-divalent exchange, the result should be a straight line of slope = 1 when Keq = 1 (Fig. 4~ 1); for the mono-divalent case, the equilibrium expression and cation mass balance equations can be rearranged to give

9 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS where E =1- ( [ 1, Ca A TN (1 - ECa)2 2 A = 'YNa 'YCa 1 ]] (1 -Eca) 1209 [9] [10] and TN = total salt normality (Le., total anion charge in molarity of the solution). Thus, for exchange reactions involving unequal valence cations, the nonpreference isotherm is actually exponential in shape, and varies with the total ionic strength of the equilibrium solution, approaching a slope of one only at high salt concentration (Fig. 40-1; Sposito, 1981b). Exchange data must therefore be compared to the nonpreference isotherm computed at a similar salt level in order to evaluate the preference of the exchanger for one cation over another. Other Exchange Coefficients Other coefficients have been proposed to represent the cation exchange process. Gapon (cited in Sposito, 1981a) wrote an expression for the equation Cao.sX + Na+ ~ NaX Ca+2 [11] using equivalent fractions, Ei = equivalents/total adsorbed charge, as [12] This Gapon coefficient, KG, is commonly used in the literature for expressing cation selectivity, and has been used by the Salinity Laboratory in work on irrigation water quality (Salinity Lab. Staff, 1954). Defining sodium adsorption ratio (SAR) = (Na)/(Ca)l!2 and exchangeable sodium ratio (ESR) = (ENaX/ECaX)' Eq. [12] becomes ESR KG = SAR' ESR = KG SAR [13] If exchangeable sodium percentage, ESP = 100 X ENa, is used, then [ ESP ) KG = SAR (100 - ESP) [14] While SAR was originally formulated as an activity ratio by Gapon, field application of this equation has often reverted to the use of total cation concentrations determined in saturated pastes, often including Mg2+ in the denominator (Le., [Na+]/([Ca2+] + [Mg2+])1!2, where brackets indicate total concentrations in mmol L-l). The average value of KG = determined for a range of Califor-

10 1210 SUMNER & MILLER nia soils by the Salinity Laboratory Staff (1954) using this approach has been criticized as potentially in error due to the effects of ionic strength and complexation on activities of the ions (Sposito & Mattigod, 1977). Complete anion and cation analyses and computer speciation of solutions is needed to accurately calculate activity-based SAR values, which may not always be justified in practice, given that SAR values based on concentration are often within 10% of the activitybased values, and considering the rather wide range of KG values reported in the literature for various soil materials (Levy et ai., 1988). The use of equivalent, rather than mole fractions in the Gapon equation also has been criticized as nonthermodynamic, since an activity coefficient (f;) cannot be meaningfully assigned to a "formal" quantity such as an equivalent (Sposito, 1977; Ogwada & Sparks, 1986). Sposito (1977), using the reverse reaction of Eq. [11], derived a new expression for the Gapon equation that is mathematically consistent with the Vanselow equation, and thus thermodynamically correct, as [15] At ENaX <0.2, the expression in parentheses on the right hand side of eq. [15] is nearly equal to one, and E NaX also is approximately equal to ESR [ENaX = ESR/(1 + ESR)]. Thus, this equation reduces to KG = SAR/(2 ESR) at low exchangeable Na, which is a linear relation differing only by a factor of two from the "traditional" Gapon equation of Eq. [13]. The question of the "correct" units for expressing exchangeable cations has introduced some confusion in the literature, as other exchange expressions (notably that of Gaines and Thomas, described by Sposito, 1981a) also have used equivalent fractions, and derived thermodynamic parameters (e.g., IlG values for exchange) based on them. While any correctly formulated exchange coefficient may be used to describe a given cation exchange reaction, it might be argued that the use of moles and mole fractions has greater thermodynamic significance, and any standardization would allow much more ready comparison of coefficients published in the literature. On the other hand, the utility of the equh'alent, or mole of charge, concept in CEC has been proven practically and historically, and is not likely to soon disappear. PROBLEMS IN MEASUREMENT OF CATION EXCHANGE CAPACITY Presence of Soluble Salts and Carbonates The first step in CEC determinations typically involves the replacement of exchangeable cations by a saturating solution. Cations other than that of the saturating solution may be present, arising from the solubility of the salts and CaC0 3 in the extracting solution. Soluble salts are readily removed during the extraction, but sparingly soluble salts such as CaS04 and CaC0 3 will continue to dissolve for prolonged periods. Polemio & Rhoades (1977) have shown that such

11 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1211 dissolution can be appreciable in commonly used saturating solutions. In addition, such salts can be soluble in the washing solution designed to remove the entrained solution. As a result, not all the exchange sites will be occupied by the saturating cation, and this will lead to an underestimation of the CEC. The accurate determination of exchangeable cations in saline and calcareous soils is clearly compromised because of this problem of quantitatively separating soluble or sparingly soluble from exchangeable cations during the extraction procedure by any method. As a result in such cases, CEC cannot be estimated by summing exchangeable cations, and often not by simple saturation techniques. Soluble cations arising from dissolution reactions also cause errors in selectivity coefficient determinations. Moderately soluble compounds such as gypsum and calcite must either be removed prior to initial cation saturation, or such dissolution accounted for during the measurement (Amrhein & Suarez, 1990). Dissolution of the clay itself also may be a concern if very low ionic strengths are used during washing or equilibration steps, releasing Mg2+, K+ and other cations to the solution phase (Frenkel & Suarez, 1977). Effect of Cation and Anion Type In soils containing micaceous minerals, problems may arise when cations such as K+ and NHt are used in the saturation process due to their fixation in the interlayer positions of the minerals. This can result in underestimation of CEC (Bower, 1950). In addition, some monovalent ions used for extraction are somewhat less efficient than divalent ions in removing Al3+. Many polyvalent ions such as Al3+ and trace metals form hydroxy ions, which may result in an overestimation of ECEC if they are assumed to be present as the simple ions. However, for Ba2+ at ph values below nine, only very small amounts of hydroxy ions are formed, and thus errors in CEC estimation would be minimized (Bache, 1976). When monovalent ions such as NHt and Na+ are used to saturate the CEC, hydrolysis may occur to a considerable extent as the electrolyte concentration is reduced, resulting in a loss of cation and underestimation of CEC. In selectivity coefficient determinations, choice of cations is dictated by the study objectives; it should be noted, however, that the theory of cation exchange assumes exchange reactions are reversible, and that there are a fixed number of exchange sites accessible to both cations (assuming fixed ionic strength and ph). Trace metals (Cu, Zn, Mn, etc.) may violate these assumptions on clays with oxidic or organic components, where metals may exchange with protons on uncharged surface functional groups (Sposito & Fletcher, 1985; Sposito, 1981a). As noted above, K+ and NHt also are strongly and irreversibly adsorbed by some clay minerals. Such "specific adsorption," by involving proton exchange and limited reversibility, is better modeled using site-binding models which account for the stoichiometry and competing ions (i.e., H+) involved in such reactions. Anions such as SO~- and pog- are specifically adsorbed by variable charge surfaces, which can result in an increase in CEC; hence the preference for Cl- or Cl04 salts in CEC measurements (Matsue & Wad a, 1985; Hendershot & Duquette, 1986). In addition, formation of soluble complexes with anions may result in errors in both CEC and selectivity coefficient determinations. Sposito et al.

12 1212 SUMNER & MILLER (1983) have suggested that their observed increases in CEC at higher Ca coverages on bentonite may be due to CaCI+ complex formation, also which influenced the selectivity of the surface for Ca vs. Na. Suarez & Zahov (1989), however, found no anion effect on CEC using Cl-, SO~-, and CI04" on montmorillonite, and little difference in Ca-Mg selectivity. ph EtTects and the Use of Buttered Solutions Almost all topsoils and many subsoils contain variable charge surfaces (organic matter, sesquioxides and clay mineral edges) which can associate and dissociate H+ depending on the ambient ph value (Uehara & Gillman, 1981). Thus when a solution buffered at a particular ph value such as 1 M NH 4 0AC (ph 7) or BaClz-TEA (PH 8.2) is used during the extraction process, the ph of the soil is brought to the ph value of the solution. The magnitude of the error incurred in the CEC measurement depends on the difference in ph values between the soil and extracting solution. Such buffered solutions cannot be used to estimate exchangeable Al3+. Thus if estimates of CEC under field conditions are required, methods involving unbuffered salt solutions should be used. Variation in ph also may affect cation selectivity of soil clays, as functional groups may show differences in selectivity as they dissociate with increasing ph (Miller et ai., 1990). Some studies, however, show little effect of ph on Ky (Sposito & Fletcher, 1985). It is probably desirable to measure selectivity coefficients at near field ph, and to avoid ph levels below 4.5 to 5, where clay dissolution releases Al3+ to solution to greatly complicate the situation. Buffers containing phosphate or acetate must be avoided due to complex formation with divalent cations. Ionic Strength EtTects and Removal of Entrained Electrolyte The magnitude of the charge on variable charge surfaces also is a function of the concentration of the equilibrium solution (Uehara & Gillman, 1981). Therefore it is important to fix or measure the ionic strength of the solution at the end of the extraction process during CEC measurement. Usually the ionic strength of the solution is selected to approximate that of the soil solution under field conditions. In methods which involve a water/alcohol wash to remove excess saturating solution, errors in the CEC measured can arise if all the electrolyte is not removed. Such washing also results in an unknown value for EC (electrical conductivity) of the final equilibrating solution. Furthermore, prolonged washing can result in the hydrolysis of adsorbed saturating ions, giving rise to low values for CEC. Variation in ionic strength has an effect on the distribution of cations between solution and surface for mono-divalent exchange, as predicted by theory (Sposito, 1981b); however, there is no effect on the value of Ky obtained, provided that activity correction of soluble ions is performed. Many reported KG values prior to 1980 did not employ such corrections in mono-divalent exchange at appreciable ionic strengths, and thus may be in error (Sposito & Mattigod, 1977). Most measurements of exchange coefficients are made at 0.01 to 0.05 M ionic

13 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1213 strength, largely in order to provide sufficient solution cations to effect the desired exchange with the solid phase. CATION EXCHANGE CAPACITY OF SOILS CONTAINING SALTS, CARBONATES OR ZEOLITES Introduction Arid region soils often contain carbonates and other soluble salts, resulting in complications with respect to the quantities of exchangeable cations extracted. A number of methods have been proposed to overcome these difficulties involving the use of double extractions and LiEDTA (Begheyn, 1987), BaClz-TEA (Mehlich, 1939), CaCl 2 (Papanicolaou, 1976), and NaCI-NaOAc (Gupta et ai., 1985). The method of Amrhein & Suarez (1990) which has been selected here, was developed to facilitate the measurement of CEC and the exchangeable cations in calcareous and gypsiferous soils by taking into account the dissolution of calcite and gypsum during the saturation and extraction steps. In addition, it accounts for anion exclusion, but does not correct for primary weathering which is assumed to be negligible in comparison to that of calcite and gypsum. The soil is first saturated with 0.2 M CaCl 2 solution adjusted to ph 8.2 and then extracted with 0.5 M Mg(N03)2, correcting for the entrained CaCl2 solution. Corrections for gypsum and calcite dissolution are made from the SO~- and HCO) contents of the soil solution prior to extraction, and the saturating and extracting solutions. Apparatus Reagents Method 1. Atomic absorption spectrometer. 2. Centrifuge. 3. Centrifuge tubes (30 ml). 4. Reciprocating or end-over-end shaker. 5. Vortex stirrer. 1. Saturating solution, 0.2 M CaCI 2 /O.0125 M CaS04' Dissolve g CaCl 2 2H 2 0 and 2.15 g CaS04 2H 20 in approximately 900 ml deionized water and adjust to ph 8.2 using saturated Ca(OHh solution. Make up to 1 L with deionized water. 2. Dilute saturating solution, M CaCI 2 Dissolve 3.68 g CaCl 2 2H 2 0 and make up to 1 L with deionized water. 3. Extracting solution, 0.5 M Mg(N03h- Dissolve g Mg(N03h 2H 2 0 and make up to 1 L with deionized water M KH(I03h solution. Dissolve g KH(I03h and make up to 1 L with deionized water.

14 1214 SUMNER & MILLER Procedure Weigh 5 g soil into a preweighed 30-mL centrifuge tube and add 20 ml of the saturating solution. Shake for 5 min, centrifuge, decant and save supernatant. The Vortex stirrer is used to resuspend the soil. Repeat this process four times, combining supernatants in a 100-mL volumetric flask for determinaion of exchangeable Mg, K and Na. Add 20 ml of the dilute saturating solution, shake for 5 min, centrifuge and decant supernatant. Repeat this process twice. Decant and save the last supernatant for determination of Ca, S04, CI and alkalinity (HC03), and then reweigh the tube plus contents to obtain the weight of entrained solution. Add 20 ml of the extracting solution, shake for 5 min, centrifuge and save supernatant. Repeat this process a further two times. Combine all the supernatants in a 100-mL volumetric flask and determine Ca, S04, CI and alkalinity (HC0 3 ). The cations Ca, Mg, K and Na are determined by atomic absorption spectrophotometry [Chapters 19 (Helmke & Sparks, 1996),21 (Suarez, 1996)], CI by Chloridometer (LABCONCO, Kansas City, MO) [Chapter 31 (Frankenberger et ai., 1996)], S04 by turbidimetry [Chapter 33 (Tabatabai, 1996)] and alkalinity by titration to ph 4.40 using the 0.01 M KH(I03)2' Calculations CEC = 10 x {Tea - THcOJ - Ts04 + V([HC03] + [S04] - [Ca]) - (To - V[Cl]) - [S04] (To - V[Cl]/[CI] - [HC0 3](To - V[Cl])/[Cl]} where T denotes ions in the Mg(N03h extract in millimoles of cation charge per kilogram, [ ] denotes ion concentration in the fmal rinse with dilute saturating solution in millimoles of cation charge per liter, V is the volume of entrained solution/weight of soil in liters per kilogram, CEC is the cation exchange capacity in centinioles of cation charge per kilogram. Comments The primary advantage of this method is that it permits the estimation of exchangeable cations and CEC simultaneously while correcting for the presence of soluble salts and the dissolution of calcite and gypsum. This method also enables selectivity coefficients to be calculated. The correction for calcite and gypsum is based on the assumption that any HC03" and SOI- found in the extracting solution in excess of that in the entrained solution is due to dissolution. The saturating solution contains saturated gypsum to reduce gypsum dissolution during the saturation process, so that when traces are present they are not removed prior to extraction. The method avoids the use of barium salts which increase calcite dissolution resulting from the precipitation of BaC0 3 A much simpler method, requiring fewer analyses, but which measures the CEC only of soils containing carbonates, gypsum and zeolites was presented in the previous edition of this volume (Rhoades, 1982).

15 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1215 Introduction CATION EXCHANGE CAPACITY OF ALL orner SOILS Compulsive Exchange Method This method was originally proposed by Gillman (1979) and subsequently modified by Gillman & Sumpter (1986a) to measure CECcE (compulsive exchange CEC). It has been shown to be a measure of the basic cation exchange capacity (CECB), which is defined as the capacity of the soil to retain basic cations under field conditions (simulated by a standard ionic strength of M) (Gillman & Hallman, 1988; Gillman & Sumpter, 1986b). For over 200 soils of widely varying origin and for 22 Andisols, the relationship between CEC B and CEC CE was very similar for the two cases CEC B = CEC CE,.z = 0.82 (200 soils) CEC B = CEC CE,.z = 0.88 (22 Andisols) [14] [15] The total cation exchange capacity (CECr) is obtained by adding exchangeable AI obtained by extraction with 1 M KCl (Bertsch & Bloom, 1996; see Chapter 18) to the value for CEC B (Gillman & Sumpter, 1986b). In addition, the anion exchange capacity (AEC) also can be estimated. This method has been successfully applied to all types of soil including saline and calcareous versions although originally developed for use on highly weathered, variable charge soils. The soil is initially saturated with Ba2+ and then brought to an equilibrium solution ionic strength similar to that of the original soil solution. The Ba2+ is then exchanged by Mg2+ by addition of MgS04, which precipitates BaS04(S). After readjustment of the ionic strength to a value comparable to that of the soil solution, the quantity of Mg2+ adsorbed (= CEC) is estimated as the loss of Mg2+ from the MgS04 solution added. Apparatus 1. Bench-top centrifuge equipped for 30-mL polypropylene centrifuge tubes with caps. 2. ph meter with combination electrode. 3. Conductivity meter, preferably with the facility to operate in "ratio mode" using two electrodes. 4. Vortex stirrer. 5. End-over-end shaker or reciprocating shaker. 6. Top loading balance reading to 0.01 g. 7. Dispensers and micropipettes. Reagents 1. Saturating solution, 0.2 M BaClzlO.2 M NH4Cl. Dissolve 48.9 g BaCl2 2H20 and 10.7 g NH4CI and make up to 1 L with deionized water.

16 1216 SUMNER & MILLER Procedure M BaCl2 solution. Dissolve 12.2 g BaCl2 2H20 and make up to 1 L with deionized water. 3. Equilibrating solution, M BaCI2. Dissolve g BaCl2 2H20 and make up to 1 L with deionized water. 4. Reactant solution, M MgS04. Dissolve g MgS04 7H20 and make up to 1 L with deionized water M MgS04 solution. Dissolve g MgS04 7H20 and make up to 1 L with deionized water. 6. Ionic strength reference solution, M MgS04. Dissolve g MgS04 7H20 and make up to 1 L with deionized water. 7. Sulfuric acid (H2S04), 0.1 M. Weigh a 30-mL centrifuge tube (Column A, Table 40--1), add approximately 2 g soil, and reweigh to determine the exact soil mass (Column B). Add 10 ml deionized water and shake for 1 h. Measure the suspension electrical conductivity (EC) and ph. If the soil contains soluble salts as indicated by EC, wash with 20 ml 70% ethanol in water and then 20 ml 10% ethylene glycol (C2H 6 0 2) in water and discard solutions. Add 10 ml 0.2 M BaCIz/0.2 M NH4Cl solution, shake for a further 2 h, centrifuge and retain supernatant for estimation of exchangeable cations [Chapter 19 (Helmke & Sparks, 1996) and 20 (Suarez, 1996)]. Add 20 ml 0.05 M BaCl2 to the tube, mix thoroughly with Vortex stirrer, centrifuge and discard supernatant. Care should be taken to avoid loss of soil material, which can be effected by removal of the supernatant by suction. To bring the soil to the standard M ionic strength, wash three times with 20- ml portions of M BaClz solution. During the last washing after thorough mixing, measure the suspension ph (phbaci2) prior to centrifugation. If AEC is to be determined, retain the supernatant for Cl- determination (Variable C 2 ). Weigh the tube and contents (Column C) to estimate the volume of entrained BaClz solution. Add 10 ml M MgS04 solution to begin the compulsive exchange of Table Worksheet for computation of CEC by the compulsive exchange method. Weighings Number ofo.5-ml increments Volume MgS04 Tube + of expressed Tube + soil + Entrained 0.05M as Final Sample phbaclz Tube soilt BaCI 2 Final volume MgS M volume CECCE+ A B C E VI = D V2 = 10 + V3 = C-B 5xD E -B t If contents have been transferred to a beaker, substitute "wi. beaker + soil" for "wi. tube + soil." AEC (cmola kg-i) = 50(CI V3 - C2VI); where CI = concentration of CI- in final solution, and C 2 = concentration of CI- in entrained solution millimoles per milliliter. + CEC CE (cmole kg-i) = 100(Mg added - Mg remaining)/weight soil = (l00/wl. soil) (0.01 V V 3)

17 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1217 Mg for Ba. Mix thoroughly and allow to stand for 1 h. Compare the EC of the suspension with that of the reference M MgS04 solution. If the conductivity ratio (CR = ECsusr!ECref) < 1.0, add O.5-mL increments of 0.05 M MgS04 and record the number of such additions. If initial CR > 1.0, take no further action. Measure ph of the suspension. If phsusp > phbacl2 by more than 0.2 to 0.3 units, add 0.1 M H2S04 dropwise until ph = phbac12 and allow to stand for at least 1 h. Reduce CR to 1.0 by adding deionized water and allow to stand overnight. Recheck phsusp and CR. If necessary, readjust to phbac12 with 0.1 M H2S04, and CR to 1.0 ± 0.05 with deionized water. When satisfied that the appropriate conditions of ph and ionic strength have been established, reweigh the tube (Column E). If AEC is to be determined, centrifuge and determine Cl- concentration in the supernatant (Variable C t ) and in the solution retained above (Variable C 2 ) [Chapter 31 (Frankenberger et ai., 1996)]. Calculation The worksheet shown in Table gives example calculations for this method, and its use considerably simplifies the record-keeping necessary with the procedure. Such a table might readily be constructed in a computer spreadsheet application, greatly facilitating computations. Comments Matsue & Wada (1985) criticized this method on the grounds that it could not be applied to soils which specifically adsorb SOa-. They proposed using 0.01 M SrCl2 instead of BaCl2 and extracting with 0.5 M HCI or 1 M NH40Ac. Hendershot & Duquette (1986) found that the CEC of predominantly variable charge soils measured by compulsive exchange with MgS04 was higher than when the Ba was replaced by MgCI2, suggesting that specific adsorption of SO~- may be responsible. Subsequently, Wada & Matsue (1987) questioned Gillman & Sumpter's (1986a) use of H2S04 to reduce the ph of the MgS04 soil suspension to its value in BaCI2, and calculated that the CEC may be overestimated by 10 to 15% using this method. Gillman & Hallman (1988) addressed these criticisms by using CaCl2 instead of BaCl2 and extracting with 1 M N~OAc; their results showed that on a range of Andisols, specific adsorption of SOi- was not a problem. As far as the use of H2S04 is concerned, they indicated that the magnitude of the error was small and would only be significant when a soil had >5 cmol+ kg-1 of both CEC and AEC. None of the soils they studied showed any of these characteristics even remotely. If this method is to be used to estimate exchangeable cations, care should be taken to prepare standards in the same matrix as the unknown solutions; Ca2+ in particular suffers some interference from soi- with air-acetylene atomic absorption determination. When soils contain soluble salts, pretreatment should be used to remove them, but this can cause errors in the estimation of exchangeable cations. However the presence of soluble salts will have little or no effect on the measurement of CEC CE The reason for the use of an equilibrating solution of divalent cations of ionic strength equal to M is because it approximates that of the "soil solu-

18 1218 SUMNER & MILLER tion" of many highly weathered soils (Gillman & Bell, 1978). In soils where the ionic strength of the soil solution differs substantially from M, an appropriate value should be used. Although BaCl2 is not usually used as an electrolyte to measure soil ph, phbaci2 should be a reliable estimate except for saline soils where the removal of soluble salts generally results in an increase in ph. For greater convenience, a conductivity meter which has the capability of being operated in a ratio mode is preferred, although a meter with a single electrode will suffice. With experience, it becomes possible with highly buffered soils to overadjust phbaci2 with 0.1 M H2S04, knowing that an upward drift will occur. Similarly, adjustment of CR to 1.0 by the addition of deionized water is soil-dependent. Should the capacity of the centrifuge tube be exceeded, simply transfer the contents to a weighed beaker and continue. A modification to the procedure which makes it less time-consuming has been proposed by Sumner et ai. (1994), in which instead of dilution with water to bring the CR to 1.0 in the final step, the sample is centrifuged prior to water addition and the EC of the supernatant is measured. From a calibration curve relating EC of the supernatant to water added, the amount of water that would have been added can be estimated. Cation exchange capacity values obtained by the compulsive exchange method are similar to those obtained by the silver thiourea method (Searle, 1986; Gillman et ai., 1983), sum of exchangeable cations (ECEC) (Gillman et ai., 1983; Grove et ai., 1982), and 0.2 M NH4CI (Grove et ai., 1982) methods. Introduction Unbuffered Salt Extraction Method This method is based on the original proposal of Schofield (1949) which enabled the measurement of the CEC of a soil at its "field ph" value. An unbuffered salt solution is used in place of the buffered solutions such as NH40Ac and BaCl2-TEA which were in vogue at that time for saturating the exchange complex. The method presented here is a modification of the procedure described by Grove et ai. (1982). It involves the saturation of the exchange sites with NHt using an unbuffered NH4CI solution, reducing the ionic strength to an appropriate value (or removing the entrained salt with water), assessing the volume of the solution which is entrained and then displacing NH4 + with a solution of KN0 3. The quantities of NHt and CI- in the final extract are corrected for the amounts in the entrained solution. If the volume of entrained solution is measured, this method also permits the estimation of the anion exchange capacity from the quantity of Cl- adsorbed. Apparatus 1. Bench-top centrifuge mL centrifuge tubes with caps. 3. Vortex stirrer. 4. End-over-end or reciprocating shaker.

19 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1219 Reagents Procedure 5. Top-loading balance reading to 0.01 g. 6. Dispensers mL volumetric flasks. 1. Saturating solution, 0.2 M NH4CI. Dissolve 10.7 g NH4Cl and make up to 1L with deionized water. 2. Equilibrating solution, 0.04 M NH4CI. Dissolve 2.1 g NH4Cl and make up to 1 L with deionized water. 3. Extracting solution, 0.2 M KN0 3. Dissolve 20.2 g KN0 3 and make up to 1 L with deionized water. Weigh 5 g of soil into a preweighed 50-mL centrifuge tube. Add 30 ml of 0.2 M NH4CI and shake for 5 min, centrifuge and decant supernatant into a 250- ml volumetric flask, being careful to avoid loss of soil. Add 30 ml of 0.2 M NH4CI, resuspend the soil using the Vortex stirrer (Scientific Industries, Bohemia, NY), shake for 5 min, centrifuge and decant supernatant into volumetric flask. Repeat this process three more times, combining supernatants prior to making up to volume with 0.2 M NH4CI. Save this solution for the determination of exchangeable Na, K, Ca, Mg and AI [Chapter 18 (Bertsch & Bloom, 1996), 19 (Helmke & Sparks, 1996), and 20 (Suarez, 1996)]. Two options are possible at this point: EITHER wash three times with deionized water and discard the supernatant, OR add 3 x 30-mL portions of 0.04 M NH4CI, resuspend, shake for 5 min, centrifuge and discard the supernatant each time, and then weigh tube to determine volume of entrained solution. Add 30 ml 0.2 M KN0 3, resuspend, shake for 5 min, centrifuge and collect supernatant in a 250-mL volumetric flask. Repeat this process a further four times, combining the supernatants. Analyze this solution for NHt [Chapter 38 (Mulvaney, 1996)] and, if entrained solution was measured and AEC is desired, for Cl- [Chapter 31 (Frankenberger et ai., 1996)]. Calculations With water wash, CEC = (NHt x 5)/18 where NHt = NHt in KN0 3 extract in milligrams per liter. CEC = cation exchange capacity in centimoles of cation charge per kilogram. With correction for entrained solution, CEC = x NHt X VEn where NHt = NHt in KN0 3 extract in milligrams per liter. V En = volume of entrained solution in milliliters

20 1220 SUMNER & MILLER AEC = 0.14 x Cl X YEn where CI- = CI- in KN0 3 extract in milligrams per liter AEC = anion exchange capacity in centimoles of anion charge per kilogram Comments When water washes are used instead of estimating the volume and concentration of the entrained solution, the soil may begin -to disperse, and a higher speed centrifuge may be necessary to separate the phases. The concentration of 0.04 M NH 4 CI was selected as that which would prevent the deflocculation of the clay in most soils. It is in the middle of the range used by Matsue & Wada (1985). The values of CEC obtained at this concentration are similar to those obtained in the compulsive exchange method at a concentration of M BaCI 2 Grove et al. (1982) suggested that the net charge on the soil was given by the value for CEC obtained using the water wash, which was essentially equal to the value obtained when the AEC was subtracted from the CEC. This value also was equal to the sum of exchangeable cations (ECEC). Introduction Ammonium Acetate (ph 7) Method Although this method has been used for many years, it overestimates the "field" CEC of soils with a ph <7. Nevertheless, it is a standard method used in the classification of soils (Soil Surv. Lab. Staff, 1992) and consequently warrants citation here. Because the NH 4 0Ac used during the procedure is buffered at ph 7, the method causes variable charge sites in acid soils not active at the field ph to become ionized and consequently measured. There are a number of variants of this method using both batch and continuous leaching techniques. The methodology selected here is a leaching tube method proposed by the Soil Survey Laboratory Staff (1992). Apparatus Reagents 1. Mechanical vacuum extractor (Centurion International, Inc., Lincoln, NE Model 24). 2. loo-ml volumetric flasks. 3. Top-loading balance weighing to 0.01 g. 1. Saturating solution, 1 M NH 40Ac ph 7.0. Mix 68 ml N~OH (sp gr. 0.90) and 57 ml 99.5% CH 3 COOH per liter of solution desired. Cool, adjust to ph 7.0 with CH 3COOH or NH 4 0H and dilute to 1 L with deionized water. 2. Ethanol, 95%. 3. Replacing solution, 1 M KCI. Dissolve 74.5 g KCI, dilute to 1 L of deionized water.

21 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1221 Procedure Prepare the leaching tubes by placing either filter paper or filter paper pulp into the syringe barrels and compressing it with the plunger. Weigh 5.0 g soil into the tube and place on mechanical vacuum extractor. Add 25 ml 1 M NH40Ac, stir and leach. Add an additional 25 ml 1 M NH40Ac and allow to stand overnight by using a pinch clamp or by stoppering the leaching tube. Leach and make up to volume and save for determination of exchangeable cations if required [Chapter 19 (Helmke & Sparks, 1996) and 20 (Suarez, 1996)]. Add about 10 ml ethanol to the soil pad, stir and leach. Leach with 100 ml ethanol and check for NHi in leachate. If NHi is present, leach with an additional 100 ml ethanol. Discard leachate. Now leach with a total of 60 ml M KCl and make up to 100 ml. Determine NH4 in the leachate by an appropriate method [Chapter 38 (Mulvaney, 1996)]. Calculations where NH4 is the concentration in the leachate in milligrams per liter Introduction Summation of Cations (Effective Cation Exchange Capacity) The concept of effective cation exchange capacity (ECEC) was first formalized by Coleman et al. (1959) as the sum of the exchangeable Ca, Mg and Al displaced from the soil using 1 M KCl but has evolved to include Na and K. There is considerable evidence to show that the quantities of exchangeable cations extracted from nonsaline noncalcareous soils by any of the common extracting solutions are very similar (Grove et ai., 1982). Thus it is possible to use this technique to estimat>! cation exchange capacity in soils which do not contain salts and carbonates. Apparatus As listed under "Preparations" for "Cation Exchange Capacity of All Other Soils," or "Apparatus" for "Unbuffered Salt Extraction Method." Reagents As listed for "Reagents" for "Cation Exchange Capacity of All Other Soils," or "Reagents" for "Unbuffered Salt Extraction Method." Procedure Determine Ca, Mg, K, Na and Al by atomic absorption spectrometry as outlined in Chapters 19 (Helmke & Sparks, 1996), 20 (Suarez, 1996), and 18 (Bertsch & Bloom, 1996).

22 1222 SUMNER & MILLER Calculations Calculate exchangeable ions (W+) in centimoles of cation charge per kilogram as M"+ = (M"+ x Vx n)/(wxa) where M"+ = concentration of cation in extract in milligrams per liter V = volume of extract (ml) n = valence of cation W = weight of soil (g) A = atomic weight of cation ECEC = Ca + Mg + K + Na + Al where Ca = exchangeable Ca in centimoles of cation charge per kilogram Mg = exchangeable Mg in centimoles of cation charge per kilogram K = exchangeable K in centimoles of cation charge per kilogram Na = exchangeable Na in centimoles of cation charge per kilogram AI = exchangeable AI in centimoles of cation charge per kilogram Comments For soils containing salts and carbonates, this procedure results in highly inflated values for cation exchange capacity because of the appreciable solubility of these materials in the extracting solutions. For all other soils, the values obtained are very similar to those measured by the methods designed to determine the CEC at "field ph" as described in "Unbuffered Salt Extraction Method" and "Compulsive Exchange Method" above. This agreement is to be expected on theoretical grounds. Over a wide variety of soils, values for the sum of basic cations extracted by a variety of extractants such as 1 M NH 4 0Ac, 0.2 M NH 4 CI, 0.2 M BaCI2> 0.2 M CaClz, 1 M BaClz-TEA and 0.01 M SrClz (Bache, 1976; Grove et ai., 1982; Gillman & Hallman, 1988; Hendershot & Duquette, 1986; Matsue & Wada, 1985) plus AI extracted with 1 M KCI or 0.2 M NH 4 CI [Skeen & Sumner, 1967a,b; Chapter 18 (Bertsch & Bloom, 1996)] were essentially the same, indicating that almost any extractant is suitable for estimation of ECEC. MEASUREMENT OF SELECTIVITY COEFFICIENTS Introduction A wide range of methods and computational approaches have been used in measurements of selectivity coefficients, despite a common objective: to bring solutions containing varying ratios of two competing cations into equilibrium with an exchanger phase at fixed ph and ionic strength, and to measure both solution and exchanger compositions. In arriving at a method to suit a particular objective, several choices must be made:

23 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1223 Exchanger Phase In using reference clays or standard exchanger materials, size-fractionated suspensions are typically prepared for use in the exchange experiments. With soils, whole soil material may be used, or the colloid fraction separated and handled as a suspension. Saturating Technique Homoionic clays or soils may be used directly in the exchange experiment by adding mixed salt solutions and allowing exchange to occur. Or, the exchanger may be preadjusted to a fractional coverage of one cation by washing with concentrated solutions containing activity ratios (AR) equivalent to desired cation equivalent ratios on the clay. Using a homoionic exchanger is more convenient, but requires careful selection of solution ionic strength and solution/soil ratio, in order to provide sufficient soluble cations to exchange with the homoionic clay. Range of Surface Coverage Most studies will attempt to span the entire range of 0 to 100% coverage of the exchanger with each cation; however, choice of a more limited range of interest is acceptable, although will not allow computation of Keq (see "Formulation of Exchange Reaction"). Determination of Exchangeable Cations After equilibrium is attained (2-12 h), cations held on the exchanger may be measured by displacement with an added salt, with or without a washing step to remove entrained solution. Or, total dissolution of the clay may be performed, after accounting for entrained solution and the presence of cations in the clay matrix (Sposito et at, 1981). Alternately, at low surface coverages, the disappearance of one cation from solution may be used as a measure of adsorption if a homoionic exchanger saturated with the competing ion is used (Milberg et ai., 1978). Computational Method Any or all of the exchange coefficients mentioned earlier (see "Formulation of Exchange Reactions") may be computed with the soluble and exchangeable cation data. Activity corrections should be made if the competing ions are of dissimilar valance, or appreciable S04 or other complexing anions are present in solution. The following method is an example of a combination of the above choices used in published coefficient determinations. Possible modifications are discussed at the end of the method description. Use of Homoionic Clay Fractions to Determine the Vanselow Selectivity Coefficient This method, modeled after Sposito et ai. (1981) and Sposito & Fletcher (1985), uses <2 Ilm clay (either separated from soil or reference geologic mater-

24 1224 SUMNER & MILLER ial) to compute Kv. This clay fraction should be separated by standard techniques (Jackson, 1974), and stored as a concentrated (>20 g L- 1 ) homo ionic suspension containing at least 0.1 M salt to prevent hydrolysis and clay dissolution. Procedure Ten or more mixed solutions of the cations spanning the range of 1 to 100% exchanger coverage are suggested, with at least three replicates of each. In formulating solution concentrations, assume nonpreferent adsorption and use Fig. 4~1 to choose solution equivalent ratios (E) that yield roughly evenly spaced exchanger equivalent ratios (E). The formulas in Table 4~2 may then be used to calculate concentrations for the individual cations at each t and a given ionic strength (1). Sufficient solution cations must be added to the homoionic clays to ensure that exchange does not overly deplete the solution of one cation. Choice of mass of added clay and concentration and volume of salt added are the important variables. For example, in equilibrating a calcium-saturated clay with a 100% Na solution, the total equivalents of Na added should be 6 to 10 times the equivalents of exchangeable Ca. Thus, 0.25 g of clay with a CEC of 50 cmol e kg-1 contains cmol e Ca2 +, 50 ml of M NaCI contains 1.25 cmol e, providing a 10:1 ratio of added solution cations over those initially exchangeable. Before use, the clay suspension should be washed to a lower solution I «0.01 M), and the clay concentration carefully measured by drying an aliquot at 105 C; the salt content also should be measured in the solution phase. Pipetting from a vigorously stirred suspension may be used to dispense the clay into weighed centrifuge tubes (50 or 100 ml) after computing the volume needed to deliver the needed mass of clay. Water and concentrated (i.e., 0.5 M) salt solution of the individual cations are then added to achieve the required concentrations computed from Table 4~ 1. Remember to account for the cation transferred with the clay suspension in the calculations. After equilibration (typically with shaking, 8-24 h), centrifuge the tubes and fully decant the solution phase and save for cation analysis. Weigh the tubes to determine the volume of entrained solution, then displace the exchangeable cations by adding a volume of concentrated ( M) salt containing a different cation than the two of interest (see "Problems in Measurement of Cation Exchange Capacity"). Corrections should be made for the contribution of entrained solution to the cation levels in the displacement solution (see "Problems in Measurement of Cation Exchange Capacity"). Table Solution equivalence conditions, ionic strength, and concentration formulas for exchange reactions of Cations i at a given solution equivalent ratio (E) and ionic strength (l).:f: Type of exchange Equation mono)-monoz Equiv. condition Ionic strength Solve for Cj, C m Ej = c;i( c) + cz)t I = c) + Cz cj=ei t Cj is expressed in moleslliter. Ej = 2cd(2c) + 2cz) = Cj /(c) + cz) 1= 3c) + 3cz Cj =E;I/3 Em = cm/(cm + 2cd) 1= cm + 3Cd cm = zeml/(3 - Em) :f: c = concentration, d = divalent cation, m = monovalent cation; anions are assumed monovalent.

25 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1225 Determine cations in the equilibrium and displacement solutions by atomic absorption spectroscopy or other suitable method [see Chapters 19 (Helmke & Sparks, 1996) and 20 (Suarez, 1996)]. Extreme care should be exercised in these analyses, including the use of standards made in matrices to match samples, in order to obtain the most accurate results possible. Calculations Solution concentrations may be expressed in moles per liter or millimoles per liter (mmol L-1); activity ratios and coefficients for mono-divalent exchange will differ by a factor of 31.6 (i.e., -V1000) when computed using these two sets of units. Exchangeable ions may be computed as mole fractions, using the sum of the moles of exchangeable ions as denominators for M;, or as equivalent fractions. For mono-monovalent or di-divalent exchange, Ky may be calculated directly from [17] using free concentrations (i.e., corrected for complexation using the approach of Sposito & Mattigod, 1977) of cations i andj rather than activities, since the activity coefficients of the two cations are equal. For mono-divalent exchange (i.e., Na-Ca), [18] individual'y values can be calculated directly using eqs. 4-6, or solutions may be speciated using geochemical speciation programs such as GEOCHEM (Sposito & Mattigod, 1977) or MINTEQ (Allison et ai., 1991). The Ky values obtained over a range of M; may be plotted as In Ky vs. M; to examine the dependence of Ky on exchanger composition. Integrals for the curve will ~ield Keq. Plots similar to Fig also may be made after computing E; and E; and compared with the nonpreference isotherms shown. Comments Rieu et al. (1991) have pointed out the large analytical errors in determination of selectivity coefficients, stemming largely from uncertainty in exchangeable cation measurements at low M; or E;. They recommend an empirical approach to relating E; and E; rather than a thermodynamic one due to their estimates of 15 to 25% error in computed K values; adequate replication (3-5), conscientious measurements, and careful choice of initial solution concentrations will minimize these problems. Overall error should be computed for each set of replicates as a standard deviation in K, and reported or used in statistical testing. Many variations on the proposed procedure are possible, depending on personal preference, experimental objectives, etc. Whole soils may readily be employed, with soil mass chosen based on the expected CEC of the clay fraction.

26 1226 SUMNER & MILLER Sampling from soil containers should be done carefully to avoid sampling error, and salt release should be checked by analyses of selected equilibrium solutions for release of cations (K, Mg, AI) from the solid phase. It is possible to preadjust M j to a value close to that expected at equilibrium for each solution composition by presaturating the solid phase with concentrated salt solutions. This eliminates substantial change in composition of the solution phase after equilibration. The concentrated solutions (0.5 M), chosen to span the range of Mj of interest, are subsequently replaced with more dilute solutions of the same activity ratio [i.e., (Na)/(Ca)l/2 for Na-Ca exchange using the Gapon equation] until the final desired ionic strength (typically M) is reached. This approach has advantages in that a wider and more predictable range of E j is obtained, compared to beginning the exchange process with a homoionic clay. However, the process is considerably more time-consuming in that each tube must go through numerous washing steps with the progressively more dilute solutions. Details of this method are described by Levy et al. (1988) and Miller et al. (1990). Any of the other exchange coefficients discussed (Gapon, Gaines-Thomas) may be computed using the data obtained, in order to compare results with other such values in the literature. However, it should be noted that much early work did 'not correct solution concentrations to activities, and may in some cases thereby be in error. Anion exclusion during selectivity coefficient measurements has the potential to reduce CEC, particularly with high-charge clays (>50 cmol e kg-i) at equilibrium ionic strengths <0.05 to 0.10 M (Amrhein & Suarez, 1990). Anion concentrations in equilibrium solutions will increase above initial levels if this occurs, but this was not observed in one study on montmorillonite (Sposito & Fletcher, 1985). Binary exchange coefficients have been used to model tertiary-level exchange systems, where three or more cations are present in the system; while there is some question as to the theoretical basis of this practice, the resulting model predictions appear to be consistent (Chu & Sposito, 1981). ACKNOWLEDGMENTS Contribution from the Department of Crop and Soil Sciences, University of Georgia, Athens, GA REFERENCES Allison, J.D., D.S. Brown, and K. Novo-Gradac MINTEQA2-PRODEF, a geochemical assessment model for environmental systems: Ver. 3 users manual. USEPA 600/3-91/021. USEPA, Athens, GA. Amrhein, c., and D.L. Suarez Procedure for determining sodium-calcium selectivity in calcareous and gypsiferous soils. Soil Sci. Soc. Am. J. 54: Bache, B.W The measurement of cation exchange capacity of soils. J. Sci. Food Agric. 27: Bascomb, c.l Rapid method for the determination of cation exchange capacity of calcareous and non-calcareous soils. J. Sci. Food Agric. 15:

27 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1227 Beckett, P.H.T Studies on soil potassium: II. The immediate Q/I relations of labile potassium in the soil. 1. Soil Sci. 15:9-23. Begheyn, L.Th A rapid method to determine cation exchange capacity and exchangeable bases in calcareous, gypsiferous, saline and sodic soils. Comm. Soil Sci. Plant Anal. 19: Bertsch, P.M., and P.R. Bloom Aluminum. p In D.L. Spark et al. (ed.) Methods of soil analysis. Part 3. Chemical methods. SSSA Book Ser. 5. SSSA and ASA, Madison, WI. Bower, c.a Fixation of ammonium in difficultly exchangeable forms under moist conditions by some semi-arid region soils. Soil Sci. 70: Brady, N.C Nature and Properties of Soils. MacMillan Co., New York. Chu, S.-Y. and G. Sposito The thermodynamics of ternary cation exchange systems and the subregular model. Soil Sci. Soc. Am. I. 45: Coleman, N.T., S.B. Weed, and R.I. McCracken Cation-exchange capacity and exchangeable cations in Piedmont soils of North Carolina. Soil Sci. Soc. Am. Proc. 23: Frankenberger, Ir., W.T., M.A. Tabatabai, D.C. Adriano, and H.E. Doner Bromide, chlorine, and fluorine. p In D.L. Sparks et al. (ed.) Methods of soil analysis. Part 3. Chemical methods. SSSA Book Ser. 5. SSSA and ASA, Madison, WI. Frenkel, H., and D.L. Suarez Hydrolysis and decomposition of calcium montmorillonite. Soil Sci. Soc. Am :887-89l. Gaston, L.A., and H.M. Selim Predicting cation mobility in montmorillonitic media based on exchange selectivities of montmorillonite. Soil Sci. Soc. Am : Gillman, G.P A proposed method for the measurement of exchange properties of highly weathered soils. Aust. I. Soil Res. 17: Gillman, G.P., and L.c. Bell Soil solution studies on weathered soils from tropical North Queensland. Aust. 1. Soil Res. 16: Gillman, G.P., R.C. Bruce, B.C. Davey, I.M. Kimble, P.L. Searle, and 1.0. Skjemstad A comparison of methods used for determination of cation exchange capacity. Commun. Soil Sci. Plant Anal. 14: Gillman, G.P., and M.I. Hallman Measurement of exchange properties of Andisols by the compulsive exchange method. Soil Sci. Soc. Am : Gillman, G.P., and E.A. Sumpter. 1986a. Modification to the compulsive exchange method for measuring exchange characteristics of soils. Aust. 1. Soil Res. 24: Gillman, G.P., and E.A. Sumpter. 1986b. Surface charge characteristics and lime requirements of soils derived from basaltic, granitic and metamorphic rocks in high-rainfall tropical Queensland. Aust. 1. Soil Res. 24: Grove, 1.H., C.S. Fowler, and M.E. Sumner Determination of the charge character of selected acid soils. Soil Sci. Soc. Am. J. 46: Gupta, R.K., C.P. Singh, and I.P. Abro! Determining cation exchange capacity and exchangeable sodium in alkali soils. Soil Sci. 139: Helmke, P.A., and D.L. Sparks Lithium, sodium, potassium, rubidium, and cesium. p In D.L. Sparks et a!. (ed.) Methods of soil analysis. Part 3. Chemical methods. SSSA Book Ser. 5. SSSA and ASA, Madison, WI. Hendershot, W.H., and M. Duquette A simple barium chloride method for determining cation exchange capacity and exchangeable cations. Soil Sci. Soc. Am. I. 50: Jackson, M.L Soil chemical analysis--advanced course. 2nd ed. Dep. Soil Sci., Univ. Wisconsin, Madison, WI. lones, U.S Soil fertility and fertilizers. Reston Pub!. Co., Reston, VA. Kelley, w.p Cation exchange in soils, Reinhold Pub!. Corp., New York. Kelley, w.p., and S.M. Brown Replaceable bases in soils. California Agric. Exp. Stn. Tech. Pap. 15:1-39. Levy, GJ., H.v.H. van der Watt, I. Shainberg, and H.M. du Plessis Potassium-calcium and sodium-calcium exchange on kaolinite and kaolinitic soils. Soil Sci. Soc. Am. I. 52: Matsue, N., and K. Wada A new equilibrium method for cation exchange capacity measurement. Soil Sci. Soc. Am. I. 49: Mehlich, A Use of triethanolamine acetate-barium hydroxide buffer for the determination of some base-exchange properties and lime requirements of soil. Soil Sci. Soc. Am. Proc. 3: Mehlich, A Rapid estimation of base-exchange properties of soils. Soil Sci. 53:1-14. Mehlich, A Effect of type of colloid on calcium adsorption capacity and on exchangeable hydrogen and calcium as measured by different methods. Soil Sci. 60:

28 1228 SUMNER & MILLER Milberg, R.P., D.L. Brower, and J.V. Lagerwerf Exchange adsorption of trace quantities of cadmium in soils treated with calcium and sodium: A reappraisal. Soil Sci. Soc. Am. J. 42: Miller, W.P., H. Frenkel, and KD. Newman Flocculation concentration and sodium/calcium exchange of kaolinitic soil clays. Soil Sci. Soc. Am : Mulvaney, R.L Nitrogen-inorganic forms. p In D.L. Sparks et al. (ed.) Methods of soil analysis. Part 3. Chemical methods. SSSA Book Ser. 5. SSSA and ASA, Madison, WI. Ogwada, R.A, and D.L. Sparks Use of mole or equivalent fractions in determining thermodynamic parameters for potassium exchange in soils. Soil Sci. 141: Oster, J.D., and G. Sposito The Gapon coefficient and the exchangeable sodium percentagesodium adsorption ratio relation. Soil Sci. Soc. Am. J. 44: Papanicolaou, E.P Determination of cation exchange capacity of calcareous soils and their percent base saturation. Soil Sci. 121: Polemio, M., and 1.0. Rhoades Determining cation exchange capacity: A new procedure for calcareous and gypsiferous soils. Soil Sci. Soc. Am : Rhoades, Cation exchange capacity. p In AL. Page et al. (ed.) Methods of soil analysis. Part 2. Agron. Monogr. 9. ASA and SSSA, Madison, WI. Rhoades, Salinity: Electrical conductivity and total dissolved solids. p In D.L. Sparks et al. (ed.) Methods of soil analysis. Part 3. Chemical methods. SSSA Book Ser. 5. SSSA and ASA, Madison, WI. Rieu, M., J. Touma, and H.R. Gheyi Sodium-calcium exchange on Brazilian soils: Modeling the variation of selectivity coefficients. Soil Sci. Soc. Am. J. 55: Salinity Laboratory Staff Diagnosis and improvement of saline and alkali soils. USDA Agric. Handb. 60. U.S. Gov. Print. Office, Washington, DC. Schofield, R.K The electrical charges on clay particles. Soils Fert. 2:1-5. Schofield, R.K Effect of ph on electric charges carried by clay particles. J. Soil Sci. 1:1-8. Schollenberger, C.J., and F.R. Dreibelbis Analytical methods in base-exchange investigations in soils. Soil Sci. 30: Schollenberger, C.J., and R.H. Simons Determination of exchange capacity and exchangeable bases in soils. Soil Sci. 50: Searle, P.L The measurement of soil cation exchange properties using the single extraction silver thiourea method: An evaluation using a range of New Zealand soils. Aus!. J. Soil Res. 24: Shainberg, I., and J. Letey Response of soils to sodic and saline conditions. Hilgardia 52:2. Skeen, J.B., and M.E. Sumner. 1967a. Exchangeable aluminum: I. The efficiency of various electrolytes for extracting aluminum from acid soils. S. Afr. J. Agric. Sci. 10:3-10. Skeen, J.B., and M.E. Sumner. 1967b. Exchangeable aluminum: II. The effect of concentration and ph value of the extractant on the extraction of aluminium from acid soils. S. Afr. J. Agric. Sci. 10: Soil Survey Staff Soil taxonomy: A basic system of soil classification for making and interpreting soil surveys. USDA-SCS Agric. Handb U.S. Gov. Print. Office, Washington, DC. Soil Survey Laboratory Staff Soil survey laboratory methods manual. Soil Surv. Invest. Reps. 42. USDA-SCS, Washington, DC. Sposito, G The Gapon and the Vanselow selectivity coefficients. Soil Sci. Soc. Am. J. 41: Sposito, G. 1981a. Cation exchange in soils: An historical and theoretical perspective. p In M. Stelly (ed.) Chemistry in the soil environment. ASA, Madison, WI. Sposito, G. 1981b. The thermodynamics of soil solutions. Oxford Univ. Press, Oxford, England. Sposito, G., and P. Fletcher Sodium-calcium-magnesium exchange reactions on a montmorillonitic soil: III. Calcium-magnesium selectivity. Soil Sci. Soc. Am. J. 49: Sposito, G., KM. Holtzclaw, L. Charlet, C. Jourany, and AL. Page Sodium-calcium and sodium-magnesium exchange on Wyoming bentonite in perchlorate and chloride background ionic media. Soil Sc. Soc. Am. J. 47: Sposito, G., K.H. Holtzclaw, C.T. Johnston, and C.S. LeVesque-Madore Thermodynamics of sodium-copper exchange on Wyoming bentonite at 298 K Soil Sci. Soc. Am. J. 45: Sposito, G., and S. V. Mattigod On the chemical foundation of the sodium adsorption ratio. Soil Sci. Soc. Am. J. 41: Sposito, G., and S.V. Mattigod Ideal behavior in Na+-trace metal cation exchange on Camp Berteau montmorillonite. Clays Clay Miner. 27:

29 CATION EXCHANGE CAPACITY & EXCHANGE COEFFICIENTS 1229 Suarez, D.L., and M.E Zahow Calcium-magnesium exchange selectivity of Wyoming montmorillonite in chloride, sulfate, and perchlorate solutions. Soil Sci. Soc. Am : Suarez, D.L Beryllium, magnesium, calcium, strontium, and barium. p In D.L. Sparks et al. (ed.) Methods of soil analysis. Part 3. Chemical methods. SSSA Book Ser. 5. SSSA and ASA, Madison, WI. Sumner, M.E Effect of alcohol washing and ph value of leaching solution on positive and negative charges in ferruginous soils. Nature (London) 198: Sumner, M.E., L. de Ramos, and U. Kukier Modification to compulsive exchange method for determining cation exchange capacity of soils. Comm. Soil Sci. Plant Anal. 25: Tabatabai, MA Sulfur. p In D.L. Sparks et al. (ed.) Methods of soil analysis. Part 3. Chemical methods. SSSA Book Ser. 5. SSSA and ASA, Madison, WI. Thomas, G.w Historical developments in soil chemistry: Ion exchange. Soil Sci. Soc. Am. J. 41: Thompson, H.S On the absorbent power of soils. J. R. Agric. Soc. England 11: Uehara, G., and G. Gillman The mineralogy, chemistry, and physics of tropical soils with variable charge clays. Westview Press Inc., Boulder, CO. Vanselow, A.P Equilibria of the base-exchange reactions of bentonites, permutites, soil colloids, and zeolites. Soil Sci. 33: Wada, K., and N. Matsue Comments on "Modification of the compulsive exchange method for cation exchange capacity determination". Soil Sci. Soc. Am. J. 51:841. Way, J.T On the power of soils to absorb manure. J. R. Agric. Soc. England 11: Way, J.T On the power of soils to absorb manure. J. R. Agric. Soc. England 13: Zelazny, L.w., L. He, and A. Vanwomhaudt Charge analysis of soils and anion exchange. p In D.L. Sparks et al. (ed.) Methods of soil analysis. Part e. Chemical methods. SSSA Book Ser. 5. SSSA and ASA, Madison, WI.