Soil Science Society of America Journal 65:239-243 (2001)
© 2001 Soil Science Society of America
DIVISION S-8-NUTRIENT MANAGEMENT & SOIL & PLANT ANALYSIS
Oxidation Rates of Commercial Elemental Sulfur Products Applied to an Alkaline Silt Loam from Arkansas
N.A. Slatona,
R.J. Normanb and
J.T. Gilmourb
a Univ. of Arkansas Rice Research & Extension Center, P.O. Box 351, Stuttgart, AR 72160
b Dep. of Crop, Soil and Environmental Sciences, Plant Science 115, Univ. of Arkansas, Fayetteville, AR 72701
Corresponding author (Nslaton{at}uaex.edu)
 |
ABSTRACT
|
|---|
Knowledge of elemental sulfur (S0) oxidation kinetics for commercial S0 sources is required before they can be recommended for use as an S fertilizer source or soil acidulent. A wide number of S0 products are manufactured and differ in physical traits that influence oxidation rate and their relative effectiveness to supply S or acidify the soil. The objectives of this research were to determine the oxidation rate of four commercial S0 sources and to measure soil pH and electrical conductivity (EC) changes due to S0 oxidation in an alkaline Hillemann (fine-silty, mixed, thermic Albic Glossic Natraqualf) silt loam. Water-extractable SO2-4–S, soil pH, and EC were measured at five sample dates after S0 application in three laboratory incubation studies. Products evaluated included wettable S (WS90), Tiger 90 (T90), Disper-Sul90 (DS90), and the experimental product S92. The proportion of SO2-4–S recovered was regressed over the 90- or 94-d incubation time using a straight-line model to determine oxidation rate constants (k). An individual k explained oxidation of each product, except WS90, which required two k values to model WS90 oxidation. The first k (0.0589 mg SO2-4–S mg S-1 d-1) for WS90 represented a rapid oxidation phase between application and the first sample date at 10 d. The second k (0.01359 mg SO2-4–S mg S-1 d-1) described the slower oxidation rate observed for the remainder of the study. Of the four products tested, WS90 had the highest k. Oxidation of S92 (0.00063 mg SO2-4–S mg S-1) tended to be more rapid than DS90 (0.00021 mg SO2-4–S mg S-1) or T90 (0.00032 mg SO2-4–S mg S-1), which were not different. Only WS90 resulted in agronomically significant reductions in soil pH. Compared with the control, soil pH was reduced from 8.1 to 6.7 by application of 1000 kg S ha-1. Oxidation of S0 followed zero-order kinetics. For each individual product, k was similar among application rates. These results suggest that commercial S0 products have different rates of oxidation. Knowledge of the oxidation kinetics of the different commercial S0 sources will aid in developing use recommendations to growers for acidification of alkaline soils.
Abbreviations: DS90, Disper-Sul 90% elemental S pastille • k, oxidation rate constant • S0, elemental sulfur • S92, experimental 92% elemental S granule • T90, Tiger 90% elemental S pastille • WS90, wettable 90% elemental S
|
INTRODUCTION
|
|---|
ELEMENTAL S exists in a reduced state and must be oxidized to SO2-4 and H+ to acidify the soil or provide plant-available S. Elemental S is the standard acidulent applied to soil for pH reduction. Acidification of alkaline soils with S0 has been demonstrated to increase rice (Oryza sativa L.) growth and yield (Chapman, 1980; Moore et al., 1994). However, the yield response of rice grown on alkaline soils amended with S0 immediately before seeding depends on the amount of pH reduction prior to flooding (Slaton et al., 1997). When applied at equal rates, S0 products that oxidize quickly reduce soil pH and provide immediate nutritional benefits that produce healthier seedlings and higher rice yields compared with slow oxidizing products.
Approximately 42% of the soil samples analyzed from fields used for rice and irrigated soybean [Glycine max (L.) Merr.] production in Arkansas have soil pH greater than 6.5 (DeLong et al., 1999). Long-term use of groundwater high in Ca, Mg, and bicarbonate for crop irrigation has resulted in alkaline soil pH, which aggravates P and Zn deficiencies of rice. Flood-irrigated rice can tolerate a wide range of soil pH, but management of soil pH for optimum nutrient availability is an important consideration in Arkansas rice production. Unlike liming practices for acid soils, recommendations for acidification of alkaline soils are not readily available to mid-south rice farmers.
A knowledge of general oxidation rates of commercial S0 products is essential for effective soil and crop management where soil acidification is recommended. Oxidation rates for available S0 sources often differ due to variations in particle size (Li and Caldwell, 1966), manufacturing processes, and additives (Lindemann et al., 1991). The oxidation rate also differs among soils (Nor and Tabatabai, 1977) and is reported to be dependent on soil fertility (Sholeh et al., 1997) and environmental conditions such as soil temperature and moisture (Janzen and Bettany, 1987a).
Few research studies have been conducted to establish predictive means of estimating oxidation rates of S0 products differing in size and shape. Janzen and Bettany (1987b) suggested that S0 oxidation rate should be expressed per unit of S0 surface area rather than by mass, since particle size has a significant influence on oxidation rate. Their model accounted for the continuous decrease in S0 particle size as oxidation progressed. McCaskill and Blair (1989) also proposed a model to predict S0 oxidation based on particle size, soil temperature, and moisture. Both models require a knowledge of the soil's capacity to oxidize S0. The S0 used in studies with the objective of modeling or comparing S0 oxidation rates among soils has generally been finely ground and sieved to a uniform particle size, and it is thus not representative of most commercial S0 products intended for wide-scale application to production fields. In a thorough review of the literature we could not find a comparison of the oxidation rates of such S0 products in their commercial retail, unmodified form. The objectives of this research were to characterize the oxidation rate of several commercial S0 sources and to measure soil pH and EC changes due to S0 oxidation in an alkaline silt loam soil.
|
MATERIALS AND METHODS
|
|---|
Three incubation studies were conducted to evaluate the kinetics of S0 oxidation for an alkaline Hillemann silt loam with a history of rice and soybean production. The soil was obtained from a grower field located in Cross County, Arkansas. Selected soil chemical properties are presented in Table 1. All application rates were calculated assuming a bulk density of 1.35 g cm-3 and a hectare soil slice of 10.16 cm for a weight of 1371600 kg soil ha-1. Oxidation of an experimental S0 product referred to as S92 (920 g S kg-1, Sulfer Works, Inc., Calgary, AB, Canada) was compared at five application rates equivalent to 250, 500, 750, 1000, and 2000 kg S ha-1 (Study 1). A wettable S0 (WS90, 900 g S kg-1, Martin Resources, Odessa, TX), sieved to have a particle diameter finer than 0.42 mm, was compared at three rates equivalent to 250, 500, and 1000 kg S ha-1 (Study 2). Lastly, S92, WS90, and two additional products, DS90 (900 g S kg-1, Martin Resources) and T90 (900 g S k-1, Tiger Resources Technology, Inc., Calgary, AB, Canada), were compared at two rates equivalent to 500 and 1000 kg S ha-1 (Study 3). The WS90 product used in Study 3 was passed through a 0.15-mm sieve, resulting in a particle diameter of <0.15 mm. The DS90 and T90 are manufactured as pastilles that are relatively large (>2-mm diam.) (Saik, 1995). A single pastille, small or large, was added to each incubation vessel for each of the two rates. The S92 was passed through a nest of sieves to determine particle-size distribution. Results showed that 0, 1.8, 94.3, and 3.9% of the S92 granules were retained on sieve sizes of 2.0, 1.7, 0.84, and 0.42 mm in diameter, respectively.
Elemental S was mixed with 25 g of air-dry soil in 250-mL polycarbonate centrifuge bottles. Double-deionized water was added to each incubation vessel to bring soil moisture to 0.3 kg water kg soil-1 (-0.01 MPa). Lids were loosely screwed to the top of each bottle. Samples were aerated and moisture was adjusted weekly by measuring weight loss and adding double-deionized water to each bottle to maintain the desired moisture content. Samples were incubated at 25°C and sampled 10, 20, 35, 66, and 94 d after S0 application for Studies 1 and 2. Measurements were taken at 10, 20, 35, 60, and 90 d during Study 3. The entire 25-g soil sample was used for analysis; thus a new sample (centrifuge bottle) was used at each sample date. At each sample date, the centrifuge bottles were removed from the incubator, moisture determined, and double-deionized water was added to bring the total volume to 50 mL for determination of EC and pH in a soil weight to water volume ratio of 1:2. Samples were brought to a final volume of 125 mL water for SO2-4–S determination. Samples were shaken for 1 h, centrifuged at 4771 g, filtered through Whatman no. 42 filter paper, centrifuged at 13000 g to remove suspended clays, and filtered through a 0.45-µm ion chromatography micropore filter. Sulfate-S was determined by ion chromatography (Dionex Model DX-300 with OmniPak Pax-100 column, Dionex Corp., Sunnyvale, CA). The eluent consisted of 2.5% methanol in water (20%), 80 mM NaOH (40%,), and deionized water (40%). The eluent flow rate was 1 mL min-1.
Elemental S products were checked for the presence of water-soluble S (SO2-4–S). Water extraction indicated that S92 contained an average of 6.9% of the total S as SO2-4–S. Therefore, 6.9% of the total S added was subtracted from the SO2-4–S recovered for each S92 sample before statistical analysis.
In each study, proportion data (mg SO2-4–S recovered/mg added S0–S) for all product and rate combinations were regressed vs. time using a zero-order kinetics model. Time zero data were not used in the statistical analysis. Soil receiving no S0 application was not included in the statistical analysis, but was used to account for organic S mineralization in S0 treatments for each sample date. Slopes (oxidation rate constants, k) and intercepts were tested for dependence on product and/or rate using standard analysis of covariance techniques. Single degree of freedom contrasts were used to determine differences among slopes where appropriate. When the intercept was significantly greater than zero, the mean proportion data collected at the first sample date (Day 10) was divided by the number of days incubated to determine a rapid fraction k. At each sampling time, four replicates of each S0 product and rate were used in Studies 1 and 3. Triplicate samples were used for Study 2. Statistical analyses were performed using SAS version 6.12 (SAS Institute, 1995).
In Study 3, pH and EC measurements were analyzed as a randomized complete block two by four factorial and an untreated control with four replications for each treatment. Statistical analyses were performed using SAS version 6.12. Where appropriate, differences among treatments were identified using Fisher's protected least significant difference test at the 0.05 significance level of probability.
|
RESULTS
|
|---|
Oxidation Rate
Recovery of SO2-4–S showed a positive linear relationship across S92 and WS90 rates (Fig. 1)
. Janzen and Bettany (1987b) also found a linear relationship for extractable S among multiple S0 application rates for a Weyburn loam (Typic Cryoboroll) 
and Sylvania sandy loam (Oxyaquic Haplohumult; 
). Neither application rate nor the interaction between application rate and S0 source had a significant effect on the proportion of SO2-4–S recovered in our three studies. Therefore, data for each S0 source were averaged across application rates to calculate k. The lack of an application rate effect on k suggests that S0 oxidation reactions in soil followed zero-order kinetics. The lack of an application rate effect on S0 oxidation rate in soil is well documented (Li and Caldwell, 1966; Janzen and Bettany, 1987b), but S0 oxidation kinetics in soils has not been clearly stated to follow zero-order kinetics.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 1. Influence of elemental S application rate on oxidation of S92 and WS90 at Sample Day 35 during Incubation Studies 1 and 2
|
|
Oxidation rate constants for the four S0 products compared at 500 and 1000 kg S ha-1 followed the general order: 
(Table 2). Although the slope or k value for S92 was not statistically different from DS90 
or T90 
, k for S92 was numerically two to three times greater than DS90 or T90. This suggested that S92 tended to oxidize faster than DS90 or T90. Oxidation rate constants for S92 were similar between Studies 1 and 3 (Table 2). The predicted amount of S0 oxidized for three S0 products, averaged over 500 and 1000 kg S ha-1 application rates, are shown in Fig. 2
.
View this table:
[in this window]
[in a new window]
|
Table 2. Zero-order rate constants (slopes) of elemental S products added to an alkaline Hilleman silt loam incubated at 25°C in three incubation studies
|
|
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 2. The predicted percentage of elemental S oxidized for three products in Study 3 using zero-order oxidation rate constants
|
|
A lag phase for SO2-4–S production as described by Sholeh et al. (1997), Janzen and Bettany (1986), and Chapman (1989) was not found for any product. Intercepts for WS90 were significantly greater than zero in Studies 2 and 3. The mean SO2-4–S recovery of the total S added for WS90 10 d after application averaged 30.4% in Study 2 and 58.9% in Study 3. This suggested that both a rapid and slow oxidation phase occurred for WS90. Mineralization of organic C has previously been divided into distinct phases of decomposition, which include a rapid and slow fraction, to explain different rates of decomposition for a single substrate (Clark and Gilmour, 1983). This concept could also be used to explain oxidation of some S0 sources. At least two rate constants are required to predict S0 oxidation for WS90 accurately. Nor and Tabatabai (1977) also observed an initial rapid oxidation of S0 in a number of Iowa soils (Table 3). They suggested that a rapid period of oxidation preceded a slower oxidation phase during the first 14 d of their incubation studies. Selected data from Nor and Tabatabai (1977) and Li and Caldwell (1966) were converted to proportions (mg SO2-4–S recovered per mg S0 added) and regressed over time assuming zero-order kinetics to determine the presence of rapid and slow oxidation phases (Table 3). Results from both studies supported the presence of two distinct phases of oxidation for S0 added to most soils and most S0 particle sizes. It is possible that the rapid phase as defined here may actually represent more than a single phase. However, more intensive sampling following S0 application would be needed to describe the rapid phase accurately.
Several possible explanations can be offered as reasons for a decrease in oxidation rate with time. Data from our study suggested that only WS90 contained both a rapid and slow oxidation phase. Additionally, the slow-phase k of WS90 was more rapid than the k of any other product. Differences in WS90 particle size may be partially responsible for the different apparent rate constants. The particle diameters of WS90 used in Studies 2 and 3 were less than 0.42 and 0.15 mm, respectively. Data for the different WS90 particle sizes showed that the rapid-phase k increased as particle diameter decreased (Table 2). Data from Li and Caldwell (1966) support this observation (Table 3). Very fine particles may have oxidized rapidly during the first 10 d of incubation then a slower oxidation of coarser particles followed. Studies that use a 6-d incubation period as recommended by Janzen and Bettany (1987a) to determine S0 k values for prediction may be in error if rapid and slow oxidation phases exist for the product or soil being tested. Chapman (1989) proposed that as S0 oxidized, the soil pH immediately surrounding the S0 particles may become very acidic and adversely affect heterotrophic S oxidizers that are primarily responsible for S0 oxidation (Lawrence and Germida, 1988).
In contrast to the initial rapid oxidation phase found in our studies and studies conducted by Li and Caldwell (1964) and Nor and Tabatabai (1977), Janzen and Bettany (1986), as well as Chapman (1989), observed a lag phase at the beginning of oxidation studies. The lag phase was followed by a more rapid period of oxidation that plateaued near 100% oxidation. These three phases produce a sigmoidal shaped curve for SO2-4–S production. Chapman (1989) noted that length of the lag phase increased as incubation temperature decreased. The lag phase was attributed to a time requirement for microbial population increases and the time required for microbial colonization of S0 particles.
Soil pH and Electrical Conductivity
The increase in soil EC and decrease in pH following S0 application reflected the relative amounts of S0 oxidized for the S0 products studied at rates of 500 and 1000 kg S ha-1 (Tables 4 and 5). Although statistically significant changes were measured for pH and EC within and among treatments, only the WS90 product at application rates of 500 or 1000 kg ha-1 produced agronomically significant changes in soil pH and EC. In general, EC increased within the untreated control and with application of all S0 products during the incubation. A gradual increase in EC occurred for all rates of S92, DS90, T90, and the untreated control. Compared with the control, EC did not increase within a sample date for either application rate of DS90 or T90. Electrical conductivity increased to a maximum and either remained constant or decreased for both WS90 rates. A field study comparing preplant application rates of WS90 and T90 ranging from 0 to 2000 kg ha-1 to rice on a Calhoun silt loam (fine-silty, mixed, thermic Typic Glossaqualfs) found similar results for soil pH and EC (Slaton et al., 1997). By 34 d after application, 2000 kg WS90 ha-1 significantly reduced soil pH from 7.6 (control) to 6.0 and increased EC from 222 to 1344 µS cm-1. In comparison, an equal rate of T90 reduced soil pH to 7.2 and failed to significantly increase soil EC. Application rates of 
350 kg ha-1 of S0 were required to neutralize the CaCO3 content (1.1 g CaCO3 kg soil-1) of this Hillemann soil. Based on equations in Table 2, only WS90 applied at the 500 and 1000 kg S ha-1 rates could have neutralized the CaCO3 content of the soil during the 90- to 94-d incubation periods.
View this table:
[in this window]
[in a new window]
|
Table 4. Influence of elemental S source and application rate (kg S ha-1) on soil electrical conductivity for incubation Study 3
|
|
|
CONCLUSIONS
|
|---|
Commercial S0 sources differed in the their rate of oxidation, but followed a linear relationship across application rates that suggests zero-order kinetics. The S92, DS90, and T90 products all exhibited a slow linear oxidation rate that could be explained using a zero-order rate constant that was independent of application rate. The WS90 product required two rate constants to describe oxidation since two distinct phases of oxidation occurred. A rapid oxidation phase occurred immediately after soil application and was followed by a slower oxidation phase. Data illustrate that a knowledge of product oxidation rate is needed before an S0 product can be properly selected to supply plant-available S or acidify an alkaline soil. Soil pH can be reduced very quickly by use of a product that oxidizes quickly. In contrast, application of S0 having a very slow rate of oxidation, like T90 or DS90, may be unsuitable for initial soil pH adjustment. These slow oxidizing products may be very useful in long-term management of pH after an alkaline soil is adjusted to a target pH value by use of a product having a rapid oxidation rate like WS90. Information is still needed on other commercial S0 products that differ in formulation from the ones investigated in these studies. Likewise, oxidation rates of each commercial S0 product need to be evaluated on an array of soils to determine if a mean rate of oxidation for each product can be established or if some S0 products are more suitable than others for some soils or soil pH adjustments.
|
NOTES
|
|---|
Published with the approval of the director, Arkansas Agric. Exp. Stn, manuscript No. 99134. Research partially funded by rice grower check-off contributions administered by the Arkansas Rice Research and Promotion Board.
Received for publication December 21, 1999.
|
REFERENCES
|
|---|
- Bundy, L.G., and J.M. Bremner. 1972. A simple titrimetric method for determination of inorganic carbon in soils. Soil Sci. Soc. Am. Proc. 36:273–275.
- Chapman, A.L. 1980. Effect of sulphur, sulphuric acid and gypsum on the yield of rice on the Cununurra soils of the Ord irrigation area, western Australia. Aust. J. Exp. Agric. Anim. Husb. 20:724–730.
- Chapman, S.J. 1989. Oxidation of micronized elemental sulphur in soil. Plant Soil 116:69–76.
- Clark, M.D., and J.T. Gilmour. 1983. The effect of temperature on decomposition at optimum and saturated soil water contents. Soil Sci. Soc. Am. J. 47:927–929.[Abstract/Free Full Text]
- DeLong, R.E., S.D. Carroll, S.L. Chapman, W.E. Sabbe, and W.H. Baker. 1999. Soil test and fertilizer sales data: Summary for the growing season 1998. p. 9–24. In W.E. Sabbe (ed.) Arkansas soil fertility studies 1998. Res. Ser. 463. Ark. Agric. Exp. Stn., Fayetteville, AR.
- Janzen, H.H., and J.R. Bettany. 1986. Release of available sulfur from fertilizers. Can. J. Soil Sci. 66:91–103.
- Janzen, H.H., and J.R. Bettany. 1987a. The effect of temperature and water potential on sulfur oxidation in soils. Soil Sci. 144:81–89.
- Janzen, H.H., and J.R. Bettany. 1987b. Measurement of sulfur oxidation. Soil Sci. 143:444–452.
- Lawrence, J.R., and J.J. Germida. 1988. Relationship between microbial biomass and elemental sulphur oxidation in agricultural soils. Soil Sci. Soc. Am. J. 52:672–677.[Abstract/Free Full Text]
- Li, P., and A.C. Caldwell. 1966. The oxidation of elemental sulfur in soil. Soil Sci. Soc. Am. Proc. 30:370–372.
- Lindemann, W.C., J.J. Aburto, W.M. Haffner, and A.A. Bono. 1991. Effect of sulfur source on sulfur oxidation. Soil Sci. Soc. Am. J. 55:85–90.[Abstract/Free Full Text]
- McCaskill, M.R., and G.J. Blair. 1989. A model for the release of sulfur from elemental S and superphosphate. Fert. Res. 19:77–84.
- Moore, P.A., Jr., Y. Sun, D.M. Miller, C.A. Beyrouty, and R.J. Norman. 1994. Phosphorus deficiency in rice on high-pH soils. p. 160–167. In W.E. Sabbe (ed.) Arkansas soil fertility studies 1993. Res. Ser. 436. Ark. Agric. Exp. Stn., Fayetteville, AR.
- Nor, Y.M., and M.A. Tabatabai. 1977. Oxidation of elemental sulfur in soils. Soil Sci. Soc. Am. J. 41:736–741.[Abstract/Free Full Text]
- Saik, R.D. 1995. The evolution of a sulphur bentonite fertilizer: One company's perspective. Sulphur Agric. 19:74–77.
- SAS Institute. 1995. SAS/STAT user's guide. Version 6.12. ed. SAS Inst. Cary, NC.
- Sholeh, R.D., B. Lefroy, and G.J. Blair. 1997. Effect of nutrients and elemental sulfur particle size on elemental sulfur oxidation and the growth of Thiobacillus thiooxidans. Aust. J. Agric. Res. 48:497–501.
- Slaton, N.A., S. Ntamatungiro, C.E. Wilson, Jr., and R.J. Norman. 1998. Influence of two elemental sulfur products applied to an alkaline silt loam on rice growth. p. 326–329. In R.J. Norman and T.H. Johnston (ed.) Rice Research Series 1997. Res. Ser. 460. Ark. Agric. Exp. Stn., Fayetteville, AR.
TOP
ABSTRACT
INTRODUCTION
