NONEXCHANGEABLE K is released to the exchangeable

Kinetics of Nonexchangeable Potassium Release from Two Coastal Plain Soils 1 ABSTRACT The kinetics of nonexchangeable-k release using H-saturated resin were investigated on Kalmia (fine-loamy, siliceous, thermic Typic Hapludults) and Kennansville (loamy, siliceous, thermic Arenic Hapludults) soil profiles from the Coastal Plain of Delaware. Calciumsaturated soil samples were equilibrated with H-saturated resin from 0.5 to 960 h. Equilibrium in K release in both soil profiles was attained in about 960 h. The kinetics of K release were evaluated using the Elovich, parabolic diffusion law, first-order diffusion, and zero-order equations. The first-order diffusion equation described the K-release kinetics best as evidenced by the highest correlation coefficient (r) and the lowest value of the standard error of the estimate (SE). The parabolic diffusion law also described the data satisfactorily indicating diffusion-controlled exchange. The zero-order and Elovich equations did not describe the data well as shown by higher SE values than those found with the first-order diffusion and parabolic diffusion law equations. Nonexchangeable-K release rate coefficients (k 2 ) ranged from 1.20 to 2.2 X 10~ 3 h~' in the Kalmia soil and from to 2.9 X 10 3 h ' in the Kennansville soil. The magnitude of the /c 2 values suggested low rates of nonexchangeable-k release from the two soils. Additional Index Words: first-order diffusion kinetics, diffusioncontrolled exchange, Elovich equation. Martin, H.W., and D.L. Sparks. 1983. Kinetics of npnexchangeable potassium release from two Coastal Plain soils. Soil Sci. Soc. Am. J. 47:883-887. NONEXCHANGEABLE K is released to the exchangeable form when levels of exchangeable and soil solution K are decreased by plant uptake and leaching (Jackson, 1964; Sparks et al, 1979; Sparks, 1980). Schmitz and Pratt (1953) found that while 47% of the variation in corn (Zea mays L.) yield percentage was due to exchangeable soil K levels, 88% was ascribed to the quantity of nonexchangeable and exchangeable K. 1 Published with the approval of the Director of the Delaware Agric. Exp. Stn. as Misc. Paper no. 1018. Contribution no. 150 of the Dep. of Plant Science, Univ. of Delaware, Newark, DE 19711. Received 11 Oct. 1982. Approved 12 May 1983. 2 Former Graduate Research Assistant and Associate Professor of Soil Chemistry, respectively. The address of the Senior Author is Soil Science Dep., Univ. of Florida, Gainesville, FL 32611. H. W. MARTIN AND D. L. SPARKS 2 The release rate of nonexchangeable K from micas (Mortland, 1961; Reed and Scott, 1962; Scott, 1968; Feigenbaum et al., 1981) and from vermiculite (Mortland and Ellis, 1959) is diffusion controlled. Mortland (1958) investigated the kinetics of K release from biotite and found that the release rate was first-order using a batch technique and zero-order with a miscible displacement method. Mortland and Ellis (1959), using 0.1./V NaCl as an extractant, found the release of nonexchangeable K from vermiculite was first order. Using HNO 3 extraction at 301 and 31 IK, Huang et al. (1968) showed that K release was first order for biotite, microcline, muscovite, and phlogopite. The rate coefficients were in the order biotite > phlogopite > muscovite > microcline. Using a H-saturated resin, Feigenbaum et al. (1981) found that relative rate coefficients based on the parabolic diffusion equation averaged 0.40 for muscovite particles. While there are data on the kinetic reactions between solution and exchangeable forms of K in soils (Sparks et al., 1980a and 1980b; Sparks and Jardine, 1981; Sparks and Rechcigl, 1982), there are few reports in the soil chemistry literature on the kinetics of nonexchangeable-k release from soils. Accordingly, the objectives of this study were: to investigate the kinetics of nonexchangeable-k release using a resin technique, and to employ various kinetic equations to describe nonexchangeable-k release from two Atlantic Coastal Plain soils. MATERIALS AND METHODS Soil samples were taken at 0.15-m depth increments from 0 to 0.90 m of a (fine-loamy, siliceous, thermic Typic Hapludults) and a (loamy, siliceous, thermic Arenic Hapludults) from Delaware. These soils were chosen because anomalous crop responses to applied K were previously noted (Sparks and Liebhardt, 1982). The soil samples were air-dried and gently ground to pass through a 2-mm sieve in preparation for laboratory analyses. Soil Characterization Analyses Soil ph was determined using a 1:1 water-to-soil ratio, and organic matter was estimated using a modified Walk-

884 SOIL SCI. SOC. AM. J., VOL. 47, 1983 ley-black procedure (Allison, 1965). Cation exchange capacity (CEC) was measured by a MgCl 2 saturation with subsequent displacement by CaCl 2 (Rich, 1962; Okazaki et al., 1963). Particle-size analysis was determined by the hydrometer method (Day, 1965). Before mineralogical analysis, samples were treated with NaOCl adjusted to ph 9.5 to remove organic matter (Anderson, 1963). Iron oxides were removed using a Na dithionite-citrate-bicarbonate procedure (Mehra and Jackson, 1960). The clay and silt fractions were passed wet through a 300-mesh sieve and separated by centrifugation. Oriented mounts of the clay fraction were prepared by depositing ~ 250 mg of clay on a ceramic tile, saturating with K or Mg under suction, washing free of salts, and glycolating the Mgsaturated samples. X-ray dim-action patterns of Mg glycerolsaturated samples at 298 and 383K and K-saturated samples at 298, 383, 573, and 823K were obtained using a Diano XRD 8300 AD x-ray diffractometer equipped with a graphite monochromator, PDP-8 computer, and a printout. The samples were scanned at 2 (26) per min using CuKa radiation. Soil samples from each depth of the two soils were characterized as to their K status. Exchangeable K was extracted using \M NH 4 C1 and 0.5M CaCl 2, nonexchangeable K with boiling \M HNO 3 (Pratt, 1965), and total K with HF digestion (Bernas, 1968; Buckley and Cranston, 1971). The potassium in each of the extracts was measured using a Perkin- Elmer 5000 atomic absorption spectrophotometer. Mineral K was estimated by subtracting the sum of CaCl 2 and HNO 3 - extractable K from total K. Kinetics of Nonexchangeable-K Release Before initiating the kinetic studies, each of the soil samples was Ca-saturated with 0.5M CaCl 2 to remove native exchangeable K. The samples were then washed with deionized water until a negative test for Cl~ was obtained with AgNO 3. Duplicate 2-g samples of Ca-saturated soil were added to 80-mL polypropylene centrifuge tubes with 4 g of moist Bio-Rad AG 50WX H-saturated resin and 50 ml of O.OOlAf HC1. The resin had a CEC of 54.1 mol(h + ) kg-'. Homoionic H-resin was prepared by leaching the resin with l.oa/ HC1 solution and washing out the salt with deionized Table 1 Selected chemical and physical properties and mineralogy of the < 2-/un clay fraction of Kalmia and Kennansville soils. Particle-size analysis Depth Sand Silt Clay ph Organic matter CEC m % % mol (>/2Mg») kg-'xlo-! Mineral suite of < 2-fOn clay fractiont 69.221.2 9.6 5.0 2.2 2.3 Ka,J, Cv a, Q 3, Gi«, Fe s 67.521.411.1 5.3 1.4 2.0 Ka,, Cv 2, Ve,, Q,, Gi s, Fe, 70.1 14.9 15.0 5.2 0.5 1.8 Ka Q,, Ve,, Cv 4, Gi., Fe, 74.5 11.3 14.2 4.9 0.3 1.7 Ka,,Q,,Cv 3, Gi., Mi., Fee 81.6 84.6 7.2 6.9 11.2 4.6 8.5 4.7 0.1 1.4 1.0 Ka,,Q,,Cv 3 Ka,. Q,. Fe 3 84.6 82.7 78.7 75.1 81.0 83.2 12.0 12.6 15.2 15.0 8.3 5.4 3.4 4.7 6.1 9.9 10.7 11.4 5.2 5.6 5.8 5.8 5.6 5.5 1.1 0.1 1,4 1.1 1.4 Q,, Cv,, Ka,, Gi. Cv Q,, Ka 3, Ve., Gi,, Fe. Cv,,K a!,q 3, Gi., Fe s Ve Ka 2, Q s, Gi., Fe s, Cv. Ka,,Cv 2,Q 3, Ve., Gi.. Fe. Ka.,Q,.Ve,, Cv., Gi 5, Fe, t Ka = kaolinite; Cv = chloritized vermiculite, Q = quartz; Gi = gibbsite; Fe = feldspars; Ve = vermiculite. t Subscript 1 = most abundant; 6 = least abundant. water. The samples were equilibrated at 298 K ± 1 for 30 min to 40 d on a reciprocating shaker. Forty days was a time when an apparent equilibrium in nonexchangeable-k release was obtained in all the soil samples. To minimize weathering and abrading of the soils, the shaker was turned off every other hour during the equilibration period. After equilibration, the soil was separated from the resin on a 60-mesh sieve and the resin was leached with 80 ml of \M NH 4 C1 to remove the nonexchangeable K. The leachate was brought to the 100-mL volume and analyzed for K as before. Nonexchangeable-K release conformed to first-order kinetics which for this study is described as follows: dkjdt = k 2 (K 0 - K t ), [I] where K, = nonexchangeable K released at time t, K 0 = nonexchangeable K released at 40 d, k-i = nonexchangeable-k release rate coefficient, t = time. Integrating, with appropriate boundary conditions of t = 0; K, = 0, Eq. [1] becomes ln(k 0 -K t ) = lnk 0 - k 2 t. [2] RESULTS AND DISCUSSION Selected chemical and physical properties and mineralogy of the clay fraction of the Kalmia and Kennansville soils are given in Table 1. The CEC and organic matter contents of the two soils were low, which is typical of Atlantic Coastal Plain soils. The Kalmia soil contained higher amounts of clay at all depths except for the 0.75- to 0.90-m increment. Considerable quantities of kaolinite, chloritized vermiculite, and quartz were present throughout the profiles of the two soils. Lesser quantities of vermiculite and feldspars were also found in both soils. Potassium Chemistry of the Soils The amounts of CaCl 2 - and NH 4 Cl-extractable K were generally higher in the Kalmia than in the Kennansville soil (Table 2). At the three lowest increments in the two soils the amounts of NH 4 Cl-extractable K were about the same, whereas more CaCl 2 -extractable K was found in the Kennansville soil than in the Kal- Depth m Table 2 Potassium chemistry of the Kalmia and Kennansville soils. NH.C1 CaCU HNO 3 H-saturated resin <# 0 )T Mineral Kt Total K mmol (K*) kg~' 2.25 1.35 1.30 1.45 0 1.65 2.10 6 0.95 1.38 1.48 1.60 1.72 0.95 1.05 0.97 1.23 1.15 2.20 1.92 1.86 1.75 1.89 2.04 2.18 1.80 1.64 2.13 7 1.60 1.15 0.69 0.67 1.23 1.31 1.49 2.09 2.22 2.14 2.79 2.52 2.23 1 1.39 1.12 6 1.92 1.86 37.58 41.23 41.99 41.68 46.58 49.71 31.26 34.79 34.29 33.58 29.47 35.28 t Represents the quantity of K extracted at 40 d with H-saturated resin. j Mineral K = [(total K) - (CaCl, Ext. K + HNO, Ext. K)]. 40 44.10 44.90 44.40 49.70 52.90 34.50 37.70 37.10 37.60 33.80 39.00

MARTIN & SPARKS: KINETICS OF NONEXCHANGEABLE K RELEASE FROM TWO COASTAL PLAIN SOILS 885 23.00 17.90 Q- 15.40 823K 7.67 823K CONTROL < 5.10 I o o KALMIA SANDY LOAM KENNANSVILLE LOAMY SAND 0.00 100 200 300 400 500 600 700 800 900 1000 1100 1200 TIME, h Fig. 1 Amount of nonexchangeable K released vs. time in the 0.45- to 0.60-m depth of Kalmia and Kennansville soils. 573K 573K CONTROL mia soil. Although clay content is important in the extractable-k status of these soils, the K-containing mineral content may be more important. For example, at the depth of 0.45 to 0.60 m, the clay contents were 14.2 and 9.9% in the Kalmia and Kennansville soils, respectively, whereas the amounts of CaQ 2 -extractable K were 0.97 and 1.23 mmol kg" 1, respectively. This can probably be ascribed to the higher content of vermiculite in the Kennansville soil than in the Kalmia soil (Table 1). Ammonium chloride extracted more K than CaCl 2 in the two soils. Both soils contained considerable quantities of vermiculitic clay minerals (Table 1) which usually contain "wedge zones" or specific sites for K adsorption (Rich, 1964; Sparks and Liebhardt, 1981; Sparks and Liebhardt, 1982). Ammonium, with a crystalline radius of 0.143 nm (Rich, 1968), could displace K + with a crystalline radius of 0.133 nm (Rich, 1968) from the sites, whereas the wedge zones would selectively screen out the larger Ca 2+ ion. This would result in higher levels of extractable K with NH 4 C1 than with CaCl 2. Thus, in soils of this type, NH 4 C1 or NH 4 OAc would tend to overpredict exchangeable K. The total K levels averaged 46.3 and 36.6 mmol kg" 1 in the Kalmia and Kennansville soils, respectively. These are comparable to total K levels found in Florida Coastal Plain soils (Yuan et al., 1976) but considerably lower than quantities noted for Virginia Coastal Plain soils (Sparks et al., 1980). Greater than 88% of the total K in the two soils was in the mineral phase. The high portion of total K in these primary mineral forms suggests that parent material was the origin of most of the K at each depth in the two soils. Nitric acid-extractable K, which is used as an index of nonexchangeable K (Pratt, 1965), averaged 1.94 and 2.33 mmol kg" 1 in the Kalmia and Kennansville soils, respectively. These relatively low levels would be expected where much of the total K was in the mineral form (Sparks et al., 1980). Kinetics of Nonexchangeable-K Release Before initiating the kinetic studies, a number of measurements were made on the resin-soil suspen- 383K 383K CONTROL 298K 298K CONTROL 1.4 1.2.72.48.43.36.33 d-spacing. nm Fig. 2 X-ray diffractogram tracing of K-saturated < 2-ian clay fraction of the 0.45- to 0.60-m depth of soil. sions. Levels of K in the leachate after washing the separated resin were measured and found to be extremely low. Samples containing H-resin and the HC1 solution without soil were run in quadruplicate for each of the times investigated. The amount of K released averaged 0.10 X 10~ 5 mol L" 1. This quantity was subtracted from the total amount of nonexchangeable K released from the soils at each time increment. The level of solution K markedly affects the release of nonexchangeable K from clays. The K concentration in the solution phase must be kept very low, or K release will be inhibited (Reed and Scott, 1962; Rausell-Colom et al., 1965; Fanning and Karamidas, 1979; Feigenbaum et al., 1981). The concentration of soluble K in the soil-resin suspension of this study ranged from 1.00 to 0 X 10" 3 mmol L" 1. Rausell-Colom et al. (1965), using a leaching technique, found that concentrations of solution K up to 1.0 mmol L" 1 did not retard K release from trioctahedral mica whereas concentrations above 0.10 mmol L~" inhibited K release from muscovites. The total amount of nonexchangeable K released from the 0.45- to 0.60-m depth of the two soils to the homoionic H-resin is presented in Fig. 1. Equilibrium in K release was attained in about 40 d for the two soils. Although not shown, a similar trend was observed in the other depths of the two soils. With the exception of the two lowest depths of the Kennansville soil, more nonexchangeable K was released at 40 d

886 SOIL SCI. SOC. AM. J., VOL. 47, 1983 (K p ) from the Kalmia soil than from the Kennansville soil (Table 2). This would be expected since the Kalmia soil profile generally contained more clay than the Kennansville soil profile (Table 1). The higher level of nonexchangeable K released from the 0.60- to 0.75- m depth of the Kennansville soil can be ascribed to the large quantity of vermiculite present (Table 1) which is a major source of nonexchangeable K (Sparks, 1980). Cation exchange resins of various saturations have been employed to investigate K release from soil clay minerals. Calcium- and Na-saturated resins were found to be unsatisfactory when used with any minerals more stable than trioctahedral micas (Arnold, 1958; Stahlberg, 1959; Haagsma and Miller, 1963; Feigenbaum et al., 1981). Arnold (1958) found muscovite and hydrous mica were comparatively resistant to H-resin attack. Haagsma and Miller (1963) showed that little acid decomposition of soil minerals took place above ph 2.5 in a soil-resin mixture. However, Talibudeen et al. (1978) argued that H-saturated resin may be destructive to soil minerals and they recommended Casaturated resin. In this study we found that the H- saturated resin did not appear to cause destruction of the soil minerals. The < 2-pm clay fractions of the soil samples after a 40-d equilibration with the H-saturated resin were compared to those without H-resin treatment using x-ray diffraction (Fig. 2). There was no alteration of the mineral suite after the 40-d equilibration with H-resin. This is also reflected in the K 0 values. In each of the depths of the two soils except for the 0.45- to 0.60-m depth of the Kalmia soil, less K was removed at 40 d with the resin than with the HNO 3 extraction which is an index for nonexchangeable K. The higher values for the HNO 3 -extractable K would suggest some extraction of mineral K as suggested by Barber and Matthews (1962). Kinetic Equations to Describe Nonexchangeable-K Release Mathematical equations including the Elovich, parabolic diffusion, first-order diffusion, and zero-order were tested by least square regression analysis for nonexchangeable-k release from the Kalmia and Kennansville soils (Table 3) to determine which equation best described the data. The correlation coefficient (r) Equation Table 3 Correlation coefficient (r) and standard error of estimate (SE) of various kinetic equations for nonexchangeable-k release from the Kalmia and Kennansville soils, t 1. Elovich: K t = a + b In t 2. Parabolic diffusion law: K t /K 0 = a + bt>" 3. First-order diffusion: \n(k 0 -K t ) = a-bt 4. Zero order: (K 0 -K t ) = a-bt Kalmia sandy loam SEJ x 10-4 3.30 5.49 1.35 9.71 r 12 0.980-0.990-0.985 Kennansville loamy sand SEt 2.30 1.26 1.40 6.63 r 71 0.984-0.986-0.977 t The r and SE values represent averages for the six depths of each soil. j SE is in mol kg-'. 1.0 0.6 0.4 0.0 - -0.4-0.6 - I- TIME, h 100 200 300 400 500 600 o KALMIA SANDY LOAM KENNANSVILLE LOAMY SAND Fig. 3 First-order kinetics of nonexchangeable-k release from the 0.45- to 0.60-m depth of Kalmia and Kennansville soils. and the standard error of the estimate (SE) were calculated for each equation. The standard error of the estimate is defined as SE = - K t *)*/(n-2)]w, [3] where K t and K t * are the measured and calculated concentrations of K released at time t and n is the number of measurements. The first-order diffusion equation was the best of the various kinetic equations studied to describe the reaction rates of K release from the two soils, as evidenced by the highest value of r and the lowest value of SE (Table 3). The parabolic diffusion law also described the data satisfactorily indicating diffusion-controlled exchange. This was also found in pure minerals by others (Mortland and Ellis, 1959; Mortland, 1961; Reed and Scott, 1962; Scott, 1968; Feigenbaum et al., 1981). The relationship showing the good fit of the data for the 0.45- to 0.60-m depth of the two soils to the first-order equation is shown in Figure 3. The zeroorder equation is not suitable to describe the kinetic Table 4 First-order nonexchangeable-k release rate coefficients (k?) of Kaimia and Kennansville soils. Depth k, x 10- ] h- 1.9 1.9 2.1 1.8 2.2 1.8 1.6 1.7 2.3 2.9 2.5

MARTIN & SPARKS: KINETICS OF NONEXCHANOEABLE K RELEASE FROM TWO COASTAL PLAIN SOILS 887 data as can be seen from the large values of SE, despite the fact that the values of r are quite high (Table 3). The Elovich equation satisfactorily described the rate of K exchange between solution and exchangeable phases in soils (Sparks et al., 1980b) and the kinetics of P release and sorption in soils (Chien and Clayton, 1980). However, it did not satisfactorily describe the kinetics of nonexchangeable-k release from the soils we studied as evidenced from the low r values and high SE values (Table 3). Nonexchangeable-K release rate coefficients (k 2 ) were calculated for all depths of the two soils (Table 4) using the first-order equation since it best described the data. The k 2 values ranged from 1.2 to 2.2 X 10~ 3 hr 1 in the Kalmia soil and from to 2.9 X 10~ 3 h~' in the Kennansville soil. The magnitude of the k 2 values suggests low rates of nonexchangeable-k release from the two soils. The magnitude of the k 2 values differed little between depths in the two soils as would be expected from the similar clay mineral suites and clay content (Table 1). ACKNOWLEDGMENTS The authors wish to express their appreciation to the Potash and Phosphate Institute and to the University of Delaware Research Foundation for partial funding of this research.