594

Published May, 1977

Determining Cation Exchange Capacity: A New Procedure for Calcareous and Gypsiferous Soils1
MARIO POLEMIC AND J. D. RnoADES2
ABSTRACT The new method presented and the errors it minimizes are
A new, two-step procedure is presented for determining the cation
discussed by contrasting it with the CEC method probably
exchange capacity (CEC) of calcareous and gypsifcrous soils. The used most frequently for salt-affected soils, the method of
method eliminates most of the errors inherent in conventional meth- Bower et al., (1952).
ods. The cation exchange sites are saturated with sodium by four suc-
cessive "equilibrations" of the soil (4-5 g) with 33-ml increments of a SOURCES OF ERROR IN CONVENTIONAL CEC
pH 8.2, 60% ethanol solution of 0.4N NaO Ac - 0.1/V NaCI (saturating METHODS
solution). The saturated sample then is extracted with three 33-ml
increments of l.ttV, pH 7 magnesium nitrate. Total sodium (Na,) and Each of the three steps commonly used in the conven-
chloride (Cl,) are subsequently determined in the extracted solution. tional methods of determining the CEC of soils can in-
Chloride (Cl,) is determined so that the soluble sodium (Na^,), from troduce errors. Some of these errors are enhanced if the
the excess saturating solution carried over from the saturation step to soils contain calcium carbonate and/or gypsum. While the
the extraction step, may be deducted from the total sodium (Na,) to ob- method of Bower et al. (1952) is the standard used in many
tain exchangeable sodium (Na,,xc.h), which is equivalent to the CEC. laboratories, it cannot be used with gypsiferous soils with-
Thus, out first removing the gypsum. The basic error sources in
most conventional methods can be pointed out in discussing
CEC = Naexch = (Na, - Nasol) = Na, - [Cl,(Na/Cl)saniraUn« solution] the Bower method. This discussion also serves to contrast
our method with that of Bower.
where (Na/Cl) sanirating miMimis the ratio of Na to Cl w the
saturating The steps of the "Bower" method are:
solution.
1) Saturation of the cation exchange sites by four succes-
Additional Index Words: cation exchange capacity, calcareous soils, sive "equilibrations" of the soil with normal, NaOAc of pH
gypsiferous soils. 8.2.
2) Removal of excess saturating solution by three wash-
ings with 95% ethanol; each washing solution is decanted
ETHODS OF determining cation exchange capacities after centrifugation.
M (CEC) of soils generally consist of saturating the soil
with saline solutions of an index cation, eliminating the
3) Replacement of the adsorbed Na by three extractions
with neutral, normal NH4OAc for subsequent quantitative
excess saturating solution, and then determining the amount determination.
of extracted index cation per unit dry weight of soil. The Possible sources of error in this method are:
results obtained are often different with different methods. 1) In the saturation step, the exchange sites may not
Variations in results are not surprising in view of the many be completely saturated with Na because other cations in
possible complicating interactions between saturating, the "saturating solution" compete with Na for adsorption
washing, and extracting solutions and soil constituents dur- sites. These other cations derive from the dissolution of
ing the analytical operations and the fact that CEC is not a CaCO3, gypsum, and silicate minerals in the saturating
conservative property of the soil (Frink, 1964; Okazaki et solution. Gypsum is quite soluble in the NaO Ac solution
al., 1962, 1963, 1964; Rich, 1962; Smith et al., 1966). The (31 meq/liter)3; CaCO3 is less soluble (1.5 meq/liter)3; only
complications arising from the dissolution of calcium car- a few silicates are sufficiently "soluble" to produce signifi-
bonate and gypsum in the soil during CEC determinations cant errors with this method (Rhoades and Krueger, 1968).
are particularly troublesome. Although methods have been This error would result in an underestimate of CEC.
proposed for calcareous and gypsiferous soils (Carpena et 2) In the washing step there are four potential sources of
al., 1972; Garman and Hesse, 1975; Papanicolaou, 1976; error:
Yaalon et al., 1962), they either are not generally applicable a) Adsorbed Na may be removed by hydrolysis;
to soils containing both calcium carbonate and gypsum or b) Adsorbed Na may be replaced by cations—
they are too cumbersome and demanding for routine deter- especially Ca—brought into solution in the washing solvent
minations. For this reason we evaluated the sources of error from the dissolution of CaCO3, gypsum, and silicates;
inherent in conventional methods of CEC determinations c) Cation exchangers—especially fine-clay particles
and developed a new method for calcareous and gypsiferous and organic matter—may be lost during decantation be-
soils that eliminates most of these errors while retaining the cause they tend to disperse as the excess electrolyte is
advantages of accuracy and simplicity. removed during washing;
d) Some of the original saturating cation may be re-
'Contribution from the U.S. Salinity Laboratory, USDA-ARS, River- tained in the sediment, and subsequently extracted as an
side, CA 92502. Presented at Am. Soc. Agron. Meeting, Houston, Texas,
282 Nov. - 3 Dec. 1976. Received 1 Oct. 1976. Approved 4 Jan. 1977. exchangeable cation if the washing is incomplete or salt is
Visiting Scientist and Supervisory Soil Scientist, respectively, U.S. retained.
Salinity Laboratory, P.p. Box 672, Riverside, CA 92502. Permanent All but (d) result in underestimates of CEC.
address of Mario Polemic is Universita Degli Studi Di Ban, Istituto Di
Chimica Agraria, Via G. Amendola, 165/A, 70100 BARI, Italy. This stay 3) In the replacement step there are two potential sources
in Riverside was sponsored with a "Consiglio Nazionale Delle Ricerche" of error:
Fellowship.
a) Adsorbed Na may be trapped between interlayers
Experimentally determined values. by contraction of expandable 2:1 layer silicates, especially
524
 POLEMIC & RHOADES: DETERMINING CATION EXCHANGE CAPACITY 525

vermiculites and weathered micas, due to ammonium fixa- fate and carbonate before determining CEC was
tion and not removed during extraction. This is a common recommended by Mehlich (1942) when the method is to be
problem in many arid-land soils (Bower, 1950). This ad- used for gypsiferous and calcareous soils. Such leaching is
sorption would result in an underestimate of CEC. not generally advisable, however, because exchangers tend
b) Nonexchangeable Na may be extracted from zeo- to be lost concurrently through dispersion, which is en-
litic- and feldspathoid-minerals by the ammonium ion hanced by the lowered electrolyte level produced with
(Rhoades and Krueger, 1968). This would result in high leaching.
CEC values. In the methods of Yaalon et al. (1962), Okazaki et al.
Soils containing appreciable gypsum cannot be used with (1962), and Papanicolaou (1976), the errors involved in the
this method because Ca would be brought into solution dur- washing step are avoided by eliminating this step. Okazaki
ing saturation and washing. et al. did not recommend their method for calcareous or
gypsiferous soils; we conclude that it cannot be used be-
LIMITATIONS OF PRESENT METHODS FOR cause NH4OAc extracts not only adsorbed Ba but also Ba
CALCAREOUS AND GYPSIFEROUS SOILS precipitated as BaCO3 and BaSO4. In the Yaalon et al.
Two CEC methods have been proposed for gypsiferous method, the exchange complex is saturated with pH 8.2,
soils (Tucker, 1974; Garman and Hesse, 1975), but each LiOAc-LiCl solution, then adsorbed and soluble Li are
replaced with pH 8.2, Ca(OAc)2 solution. The necessity of
suffers some limitations. The method of Tucker (1974) uses
NH4+ in 60% ethanol as the saturating cation and KNO3-Ca washing the soil is eliminated by assaying for Cl to deter-
mine the remaining excess saturating salts in the replacing
(NO3)2 as the replacing solution. As shown by Bower solution and deducting this amount from the total Li found
(1950), many arid-land soils contain mica-type minerals
that fix NH 4 and K within interlayers. Such reactions pre- in the used replacing solution. Errors due to incomplete sat-
uration because of CaCO3 dissolution are minimized by
clude complete saturation and extraction of the saturating
keping the pH at 8.2; however, gypsum dissolution would
cation and, hence, result in decreased CEC values. In addi-
tion to this limitation, exchangers may disperse and be lost not be prevented. A potential error exists in the method
during the final saturation step in which 0.05M NH4C1 is even for calcareous soils, though discounted by Yaalon, et
used. With the method of Garman and Hesse (1975), BaCl2 al. Cation exchangers have very low affinity for Li adsorp-
is used as the saturating solution to coat gypsum with tion; hence, during saturation even small amounts of Ca in
BaSO4 to prevent gypsum from dissolving during saturation solution will preclude complete saturation. For example,
Papanicolaou (1976) showed that 12 leachings with IN LiCl
of the exchange sites with Ba. Magnesium sulfate (Q.Q5N)
is used as the replacing solution (100 ml); the Mg lost from solution were insufficient to remove all adsorbed cations
this solution, per unit weight of soil, is taken as the CEC. from soils. In the Papanicolaou CEC method, 1A7 CaCl2 (pH
7.0) is used as the saturating solution; adsorbed Ca is
However, appreciable soluble Ca was also observed in the replaced by extraction with IN NaNO3 (pH 7.0). Total
replacing solution and we conclude that gypsum dissolves
(Ca2+ + Mg2+), Cl- and (CO32- + HCO3-), are determined
in the replacing solution. Under such conditions, Ca would in the combined extract. The CEC is calculated as (Ca +
be competing with Mg for exchange sites; to this extent, Mg
Mg) - (Cl- + CO32- + HCO3-). This method could be
loss from the replacing solution would be reduced and, used for soils containing gypsum if SO4 were determined.
hence, calculation of CEC with this method would be low.
The method, however, requires numerous analyses.
Additional weaknesses of the method lie in the uncertainty For the preceding reasons it is apparent that still lacking
of the extent to which BaCl2 prevents gypsum dissolution in is a reliable, generally applicable, and simple method for
the saturating step and how much exchangeable barium is routinely determining CEC of calcareous and gypsiferous
lost in the washing step (200 ml of distilled water). The reli-
soils, especially the latter.
ability of this method seems questionable.
Several methods have been proposed for calcareous soils PROPOSED METHOD
(Mehlich, 1938, 1942; Bower, et al., 1952; Yaalon, et al.,
1962; Bascomb, 1964; Tucker, 1974; Papanicolaou, 1976). The following two-step procedure for CEC was developed to
The limitations of the Bower method have been discussed help meet this need:
1) Saturation of the cation-exchange sites of the soil (4-5 g) by
and, as will be shown later, CaCO3 does produce apprecia- four successive "equilibrations" with 33-ml increments of a pH
ble errors in CEC determinations with this method. 8.2, 60% ethanol solution of OAN NaOAc-0. IN NaCl. Equilibra-
Mehlich (1938, 1942) used aqueous solutions of buffered tions are achieved by shaking the soil-solution mixtures for 5 min.
(pH 8.2) BaCl2 -triethanolamine (TEA) to saturate the ex- The "exhausted" solutions are decanted after centrifugation. If
change complex of calcareous soils, taking advantage of the the soil is initially high in salts (ECe > 2 mmhos/cm), wash the
ability of Ba to be adsorbed on CaCO3 to minimize its soil with one 33-ml increment of water before saturation. Exces-
sive washing should be avoided to prevent loss of particles during
subsequent dissolution during exchange saturation. Others decantation.
have used this method, with slight modifications, for cal- 2) Extraction of the saturated sample with three 33-ml incre-
careous soils with variable success (Hanna and Reed, 1948; ments of I.ON, pH 7 solution of MgNO3. Both sodium (Na,) and
Pratt and Holowaychuk, 1954; Bascomb, 1964; and Car- chloride (Clt) are subsequently determined in this solution. Chlo-
pena et al., 1972). Most of these methods have the limita- ride is determined so that the soluble sodium (Na^) from the
tion that Ba loss from the saturating solution is taken as excess saturating solution carried over from the saturation step to
the extraction step may be deducted from the total sodium to ob-
equivalent to the CEC; this ignores the loss of Ba by adsorp- tain exchangeable sodium (NaeXCh). Thus,
tion on CaCO3 and includes the errors inherent in "differ-
ences" methods. Extensive leaching to remove soluble sul- CEC = (Na, - Nasol) = Na, - [Cl,(Na/Cl) salurating solutiim]
 526 SOIL SCI. SOC. AM. J . , VOL. 41, 1977

Table 1—CEC of Pachappa soil to which various increments of Table 2—Properties of soils and CEC's as determined by the Bower et
calcium carbonate and gypsum were added, as determined al. (1952) (A) and newly developed (B) methods.
by the Bower et al. (1952) method (A) and
newly developed method (B). CEC by method
Soil no. SPf pHst ECe§ Pwf CaCO3 CaS04-2H20 A B
Amount of minerals
addedf CEC by method % mmhos/cm ——— %" meq/100 g
Difference
Sample CaC03 CaS04 A B (A-B) 3565 29 7.6 2.37 1.53 2.2 „ 15.0 16.0
3571 38 7.4 1.09 2.08 3.2 - 18.2 19.0
————— 9'o —————— meq/100 g meq/100 g 29 9.2 4.96 1.06 5.6 - 18.0 13.7
3582
1 „ „ 9.7 9.7 3576 82 8.1 7.88 4.14 1.4 0.15 33.5 36.5
2 0.5 - 9.4 9.7 -0.3 3564 40 7.6 79.5 4.30 7.7 0.84 31'.3 20.8
3 1 - 9.4 9.7 -0.3 3572 53 8.3 12.0 1.73 12.9 0.06 20.5 19.5
4 2 _ 9.3 9.7 -0.4 3580 46 7.6 5.13 2.33 3.8 1.46 22.0 24.4
5 4 _ 9.2 9.7 -0.5 3581 48 7.0 7.28 2.93 0.05 _ 25.5 27.9
6 8 .. 8.8 9.7 -0.8 C3/575 54 7.8 12.9 7.40 25.0 26.5 15.6 16.2
7 12 - 8.7 9.7 -1.0 C4/142 49 7.2 3.1 4.51 24.0 14.4 19.9 21.1
8 16 - 8.6 9.6 -1.0 C4/155 45 7.7 2.9 2.52 28.0 7.3 12.7 12.7
11 0.5 9.3 9.7 -0.4
12 1 9.3 9.7 -0.4 Saturation paste water content as percentage.
13 2 9.2 9.7 -0.5 : pH of soil saturation paste.
14 4 9.1 9.7 -0.6 ! Electrical conductivity of saturation extract.
15 8 8.5 9.5 -1.0 [ Air-dry water content (hygroscopic water).
16 - 12 8.4 9.6 -1.2
17 .. 16 8.2 9.6 -1.4
18 0.5 0.5 9.4 9.7 -0.3
19 1 1 9.4 9.7 -0.3 negligible reductions, were obtained with the newly devel-
20 2 2 9.2 9.6 -0.4
21 4 4 8.9 9.7 -0.8 oped method.
22 8 8 8.8 9.4 -0.6 Table 2 shows the CEC's of 11 different soils, which
23 12 12 8.7 9.5 -0.8
24 16 16 8.5 9.4 -0.9 contain various amounts of CaCO3, gypsum and salinity.
t As added to 4 g of soil.
With two exceptions, which are explained later, the CEC's
were higher when determined with the newly developed
method; this was as expected since the major errors in the
Bower method result in reduced CEC values. Papanicolaou
where (Na/Cl)saturatitlg „,„,„,„, is the ratio of Na to Cl in the saturat-
ing solution. (1976) also found low CEC values for calcareous soils when
The solubilities of gypsum and CaCO3 in the saturating solution he used the Bower method.
are sufficiently low (5.8 and 4.0 meq/liter for gypsum and calcite, Lower CEC values were obtained with the newly devel-
respectively)3 and the exchange affinity for Na vs. Ca is suf- oped method compared to the Bower method for two soils.
ficiently high to allow essentially complete saturation of the CEC To decide which method was the more accurate, especially
with Na. All the errors associated with washing are avoided by for these two soils, we made an indirect test, since there
omitting this step. The errors associated with the extraction step exists no absolute reference method for CEC. CEC values
are minimized by replacing NH4OAc with MgNO3, since Mg is
not fixed in soils nor does it extract many nonexchangeable cations were correlated with saturation percentage and air-dry (hy-
(Rhoades and Krueger, 1968). groscopic) water content, two soil properties known to be
proportional to CEC (Banin and Amiel, 1969). In correla-
EXPERIMENTAL PROCEDURE tions involving saturation percentage, soils with > 10%
CaCO3 were excluded because their relation between SP
To test the developed method, we conducted two kinds of ex-
periments. In the first experiment various increments of CaCO3 and CEC would be expected to differ from that of noncal-
and/or gypsum were added to 4 g of Pachappa soil, which is free of careous soils. Similarly, in correlations involving hygro-
these minerals in its natural state. The CEC's of these mixtures scopic water content, soils with > 5% gypsum and very
were then determined using the newly developed method and that high salinity (Soil 3564) were excluded because the method
of Bower et al. (1952). In the second experiment, the CEC's of used to determine air-dry water content (loss of weight upon
soils with different CaCO3 and gypsum contents and other proper- heating to 105°C) would be inaccurate due to loss of water
ties were determined with the two methods. of crystallization and hydration, respectively.
Sodium and chloride in the extracted solutions were determined
by atomic absorption spectrophotometry and potentiometric titra- The coefficients of linear correlation (r) between CEC
tion, respectively. Calcium carbonate in the soils was determined and SP were 0.78 and 0.96 where CEC was determined
by the manometric method of Williams (1949). Gypsum, satura- with the Bower and newly developed methods, respec-
tion percentage (SP), saturation extract electrical conductivity tively. When the CEC values of the two problem soils, as
(ECe), air-dry water content (Pw), and saturation paste pH (pHs) determined by the Bower method, were excluded from the
were determined by methods described by U.S. Salinity Laborato- regression, r was increased to 0.97. Thus, the problem soil
ry Staff (1954). CEC values determined by the Bower method seem ques-
tionable, rather than those determined by the newly devel-
RESULTS AND DISCUSSION
oped method. The r values determined between CEC and
The CEC's of Pachappa soil to which various amounts of hygroscopic water were 0.93 and 0.99 where CEC was de-
CaCO3 and gypsum were added as determined by the newly termined with Bower and newly developed methods, re-
developed method (B) and the Bower et al. (1952) method spectively. The relations between hygroscopic water con-
(A) are given in Table 1. Each value is the average of two tent and CEC, as determined with the new method, is
replicates. In all cases, increasing increments of CaCO3 shown in Fig. 1. The better correspondence between CEC
and/or gypsum resulted in reduced CEC values with the and hygroscopic water content when the CEC values ob-
"Bower" method, whereas nearly constant CEC values, or tained with the new method are used, while not proof in it-
 POLEMIC & RHOADES: DETERMINING CATION EXCHANGE CAPACITY 527

40
• / 3564 was washed once with water before extraction to
e / remove most of the soluble sodium salts. No other pretreat-
o 35' / ments (saturation, etc.) were made. Soil 3582 released 9.5
T!° rm099 / and 8.0 meq Na/100 g upon extraction with NH4OAc and
£ | so- ' / MgNO3, respectively. The analogous values for Soil 3564
E
* were 16.6 and 8.0 meq/100 g, respectively. This evidence
| as- a/ supports our assumption that these soils release non-
~ / exchangeable Na when extracted with ammonium acetate in
•| 20- A,/ the Bower method.
I 7 * The CEC values obtained with the two methods are com-
° 15- /* pared in Fig. 2. Linear correlation is excellent (/• = 0.99), if
I / the two CEC values concluded to be incorrectly determined
I 10- using the Bower method are excluded. The CEC values de-
^ termined with the new method are higher for the rest of the
J 5. soils, as expected, because steps are eliminated that produce
S errors in conventional methods.
0_ __^_^_^_^_^_^_^_^_^_^ We conclude that the new method presented here is prac-
0.0 0.5 1.0 1.5 2.0 2.5 3.o 3.5 4.0 4.5 5.0 tical, simple, and reliable for determining CEC. We recom-
Air-Dry Water Content, percent mend its adoption, especially under arid-land conditions for
Fig. 1—Relations between cation exchange capacity and air-dry water calcareous and gypsiferous soils and for soils containing
content of soils. zeolites or feldspathoids. The method is not intended for
acidic soils or other soils having appreciable anion ex-
change capacities.
o>

9"! «l /
Is /
* £ 3 5-
s r • 0.99 f

i °- y
(excluding o points) /

s
=

D
O
25-

20-
/
B
/A
A/

n
°<3S64>

= 15 /
I / <Wi
UJ 10- /

§ 5
5 -
o- ——i——i——i——i——i——i——i——i
0 5 10 15 20 25 30 35 40
Cation Exchange Capacity (Bower Method),
meq/100 gms
Fig. 2—Comparison of cation exchange capacities of soils as deter-
mined by the new and Bower methods.

self, gives credence to our conclusion that it is the Bower

method CEC values that are questionable.
As mentioned earlier, high CEC values may be obtained
in two ways in the Bower method. Some of the Na in the
saturating solution may be carried over and subsequently
extracted if the washing is incomplete. Or, nonexchangea-
ble Na may be brought into solution during the extraction
with NH4 if the soils contain zeolitic or feldspathoidic min-
erals (Rhoades and Krueger, 1968). It seemed that the latter
case was more likely since one of the problem soils (3582)
has a high pH (9.2) and is high in bicarbonate content, the
two conditions required for zeolites to exist in soils (Baldar
and Whittig, 1968). The other problem soil (3564) was
dominated by Na salts. To ascertain if these soils did indeed
contain nonexchangeable Na which was extracted with
NH4OAc but not MgNO3, as speculated, we extracted these
soils using the extraction procedures of each method. Soil
528 SOIL sci. soc. AM. j., VOL. 41, 1977