| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Chemistry, Queen Mary, University of London, London E1 4NS, UK
b International Rice Research Institute, DAPO Box 7777, Metro Manila, Philippines
* Corresponding author (n.mahieu{at}qmul.ac.uk)
Received for publication June 2, 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: BIARC, Bicol Integrated Agricultural Research Center • CaHA, calcium humate • CPMAS, cross polarization with magic angle spinning • HA, humic acid • IRRI, International Rice Research Institute • LTCCE, Long-Term Continuous Cropping Experiment • LTFE, Long-Term Fertility Experiment • MHA, mobile humic acid • NMR, nuclear magnetic resonance • SOM, soil organic matter • TOSS, total suppression of sidebands
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In comparison with single-crop systems, annual production of two or three rice crops with high rates of fertilizer application increases both the time that soil remains submerged and the total amount of recycled organic inputs from roots and stubble. Thus, incorporated plant material that is more resistant to microbial degradation under anaerobic conditions might gradually accumulate in the soil organic matter (SOM) of paddy soils, with unknown consequences for nutrient cycling.
To identify any changes in the chemical nature of SOM of potential relevance to soil nutrient supply, we have applied nuclear magnetic resonance (NMR) spectroscopy to the mobile humic acid (MHA) and calcium humate (CaHA) fractions extracted from 12 soils with varying numbers of annual irrigated rice crops, with different fertilizer treatments, and at different sites. In this study we synthesize results from solid-state 13C and 15N NMR spectroscopy and solution 31P NMR spectroscopy that were published separately (Olk et al., 1996, 1998; Mahieu et al., 2000a,b) and some additional spectra not previously published. In order to evaluate the precision of the NMR methods and to enhance our understanding of SOM formation, we present correlations of spectral proportions with each other and with previously measured humic acid (HA) properties: visible light absorption and concentrations of free radicals and hydrogen (H). Dipolar dephased 13C NMR spectra of these samples are also shown for the first time.
When the HA extractant is NaOH, as was the case here, polyvalent cations are conventionally first removed from the soil through HCl washing before HA extraction (Schnitzer, 1982). This step increases the amount of recovered HA by enabling the extraction of older, more recalcitrant HA that had been stabilized by the polyvalent cations (Campbell et al., 1967). However, such stabilized material should be less involved in nutrient cycling and less responsive to crop management. To better depict the effects of recent crop intensification on interactions between nutrient cycling and SOM quality, we extracted both polyvalent cation-bound HA (CaHA) and also HA not bound to polyvalent cations (MHA), presumed to be composed of younger, less humified materials.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In a second set of samples designed to study site-specific effects on SOM properties, we sampled two treatments of the Long-Term Fertility Experiment (LTFE) at three sites: the IRRI farm, the Philippines Rice Research Institute (PhilRice) in Central Luzon, and the Bicol Integrated Agricultural Research Center (BIARC) in Southern Luzon (Olk et al., 1998). The most conspicuous difference between the three LTFE sites is the degree to which the soil dries during the fallow between harvest of the wet season crop and transplanting of the dry season crop: the surface layer of the PhilRice soil dries thoroughly while at IRRI or BIARC it remains saturated or wet. Soil types were Andaqueptic Haplaquoll (IRRI), Vertic Tropaquept (PhilRice), and Typic Pelludert (BIARC). All treatments sampled at the IRRI and the LTFE sites had continuous unchanged management for at least 30 years prior to sampling except for the prilled urea (11 years), Sesbania (11 years), and rice–soybean (10 years) treatments.
For further comparison of soil hydrology regimes, we also sampled the rice–upland crop rotation experiment, begun in 1993 on the IRRI farm. Its treatments and crop management are described in detail by Witt et al. (2000). Here we sampled soil in April 1995 from treatments cropped to either rice–rice or rice–maize annual rotations. Samples were collected shortly before harvest of either a rice crop (submerged conditions) or maize crop (aerated conditions).
All soil samples were taken between flowering and physiological maturity crop growth stages in the plow layer. Sampled soils were stored under refrigeration at their field moisture level until extraction of the HA fractions.
Humic Acid Extraction
Olk et al. (1996)(1998) described the HA extraction in detail. In brief, the MHA was extracted by initially shaking fresh, undried soil (36 g oven-dried basis) overnight in 360 mL 0.25 M NaOH under N2 atmosphere and isolated by centrifuging and acidification of the supernatant to pH 2. The remaining soil was mixed and suspended twice in 0.0025 M CaCl2 solutions and then centrifuged. The supernatant was discarded. Except for the LTCCE soils, the amounts of SOM removed in these alkaline CaCl2 solutions were negligible, suggesting that the NaOH extraction of the MHA was exhaustive. For the LTCCE soils, solubilized MHA appeared to interact with fine clay particles that remained in suspension. Following the washes with CaCl2, we had therefore to wash the soil four times with water to remove these suspended particles in order to cleanly isolate the CaHA (Olk et al., 1996).
The soil was then decalcified through repeated washes in 0.1 M HCl until the pH of the supernatant remained below 1.3, followed by two to three washes with H2O until the supernatant pH was greater than 2.5. The soil was again shaken overnight in 0.25 M NaOH under N2 atmosphere to extract the CaHA by centrifuging and acidification to pH 2. Humic acids were then shaken in HF–HCl for 3 d to dissolve inorganic soil particles, dialyzed for 3 d against successively weaker HCl or water solutions to remove salts and excess acid, frozen, and lyophilized.
When the HA extraction was repeated on different days on subsamples of the same soils, the coefficients of variation of the HA masses ranged from 3 to 21% and were generally around 10%. Means and standard deviations of some samples were reported by Olk et al. (1996)( 1998).
Humic Acid Characterization
Gross Chemical Analysis
The C and N concentrations of the MHA and CaHA were determined on a CHN automated elemental analyzer. Total soil organic C was determined before and after HA extraction by a modified Walkley–Black method (Nelson and Sommers, 1975), total soil N by Kjeldahl digestion and distillation, and soil organic P through ignition and extraction with 0.5 M H2SO4 (Olsen and Sommers, 1982). Total concentrations of P in the HA and soil were determined by perchloric acid digestion and spectrophotometric analysis (Olsen and Sommers, 1982). Free ferric oxide (free Fe) was determined following a modified procedure of Asami and Kumada (1959). Hydrogen concentration was measured by a Fisons Instruments (Crawley, UK) EA 1108 elemental analyzer. Visible light absorption was measured on solutions of 3.6 to 7.2 mg HA solubilized in 30 cm3 0.05 M NaHCO3 (Chen et al., 1977). Organic free radical concentration was calculated from electron spin resonance spectra generated by a Bruker (Rheinstetten, Germany) ER-200D SRC ESR spectrophotometer operating at X-band frequency with 100-kHz magnetic field modulation. Measurements were collected over a small field range (10 mT) centered at about the resonance field of the free electron (g = 2.0023), using a microwave frequency of 9.52 GHz, microwave attenuation of 13 dB, and a modulation amplitude of 0.63 mT. Further details were provided by Olk et al. (2000). All analyses of soil and HA were done on composites bulked from different replicates of the same treatments.
Nuclear Magnetic Resonance Spectroscopy
Solid-state 13C and 15N NMR spectra were obtained from field composite HA samples on a Bruker MSL 300 spectrometer operating at 77.5 MHz for 13C and 30.4 MHz for 15N. Solution 31P NMR spectra were obtained using a Bruker AM 400 spectrometer operating at 162 MHz and a Bruker AMX 600 spectrometer operating at 243 MHz for 31P.
The 13C spectra were obtained using cross polarization and magic angle spinning (CPMAS) with total suppression of side bands (TOSS). Spinning rates were 4.2 to 4.9 kHz, contact time was 1 or 1.5 ms, relaxation delay was 1 s, and the number of scans ranged from 3339 (0.9 h) to 25200 (7 h). Chemical shift values were measured with respect to tetramethylsilane. We used the Bruker WinNMR software to measure peak areas for the following chemical shift regions: 0 to 45 ppm (alkyl), 45 to 65 ppm (N-alkyl and methoxy), 65 to 95 ppm (O-alkyl), 95 to 108 ppm (acetal), 108 to 140 ppm (unsubstituted and alkyl-substituted aromatic or aromatic), 140 to 160 ppm (O-substituted aromatic or phenolic), and 160 to 220 ppm (carboxyl, amide, ester, ketone, and aldehyde or carboxyl). Actual spectral boundaries were the natural "valleys" closest to the indicated chemical shift values. We performed dipolar dephasing (DD) with TOSS on some HAs, using an interruption of the decoupling for 40 µs before acquisition to identify nonprotonated C plus CH3 and mobile C (Preston, 1996). Variable contact time experiments showed greater loss of carbon visibility for a contact time of 1.5 ms than for 1 ms, but the difference was similar for all functional groups.
The 15N CPMAS spectra were obtained at magic angle spinning rates of 4.0 to 5.0 kHz, with 1-ms contact time and 200- or 220-ms relaxation delay. Between 271000 (15 h) and 629000 (38.4 h) scans were accumulated. All spectra were referenced to external nitromethane (= 0 ppm). We used the Bruker WinNMR software to measure peak areas for the following chemical shift regions (Beyer et al., 1997; Clinton et al., 1995; Knicker et al., 1997; Skene et al., 1997): -180 to -239 (heterocyclic such as indoles, pyrroles and imidazoles), -239 to -283 (amide or peptide structures), -283 to -330 (amine and aniline), and -330 to -370 (free aliphatic amino groups of peptides, amino acids and amino sugars, and terminal amino groups of peptides or amino).
For solution 31P NMR, 50 to 100 mg of humic acid was dissolved in 1 mL of 0.5 M NaOH–D2O. Drops of 1 M NaOH were added to bring the pH to above 12 if necessary. A total acquisition time of 1.5 s was used on the AM 400 spectrometer and 2 s on the AMX 600 spectrometer. On both spectrometers, a 30° pulse width and broadband proton decoupling were used. Depending on the samples, between 6000 and 12000 scans were added. After convolution with a Lorentzian of width 10 Hz, the areas were measured either electronically using the Bruker WinNMR software (spectra from AMX 600 spectrometer), or by cutting and weighing (spectra from AM 400 spectrometer). Estimated peak areas differed negligibly between the two methods for tested spectra from the AMX 600 spectrometer. Chemical shifts were measured relative to 85% orthophosphoric acid. The chemical shift regions were assigned to monoester P (3 to 6 ppm), sugar diester P (0.7 to 3 ppm), diester P (-1 to 0.7 ppm), unknown P (-2 to -1 ppm), and inorganic P (peak around 6 ppm).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Mobile Humic Acid and Calcium Humate Fractions
The MHA had higher N and H concentrations and lower free radical concentration and visible light absorption than did the CaHA in every field treatment (Table 1). Hydrogen concentration increased with increasing cropping intensity, whereas free radical concentration and visible light absorption decreased. Nitrogen concentration decreased in the MHA with increasing cropping intensity. Both fractions were dated by 14C dating as modern age (Olk et al., 1996), the CaHA fraction being of slightly older age than the MHA (data not shown). Combined, the MHA and CaHA accounted for 13 to 24% of total soil C and 11 to 24% of total soil N. These proportions were slightly greater in the triple-cropped soils (means 21% C and 19% N) than in the aerated soils (means 18% C and 16% N) and greater in optimal fertilizer treatments than in control treatments, especially for the MHA. The MHA and CaHA had similar concentrations of C, ranging from 482 to 536 g kg-1 (MHA) and 498 to 545 g kg-1 (CaHA), but MHA had a greater concentration of N than did CaHA (Table 1). Consequently, the C to N ratio was larger for the CaHA (mean 14.0) than for the MHA (mean 10.9).
|
For the two fractions combined, N concentration was strongly correlated (all P < 0.001, N = 24) with visible light absorption (r = -0.83), free radical concentration (r = -0.77), and H concentration (r = 0.79). For the MHA alone, free Fe was negatively correlated with C content (r = -0.80, P < 0.01, N = 12) and positively with P content (r = 0.60, P < 0.05, N = 12).
Nuclear Magnetic Resonance Spectroscopy
Figure 1
shows the 13C CPMAS spectra of MHA and CaHA from selected soils with increasing number of irrigated crops per year ranging from 0 to 3 from the maize sample to the LTCCE control sample. As reported by Olk et al. (1996)(1998), the MHA differed structurally from the CaHA in all four soils. The resonances at 56 ppm (N-alkyl or methoxy C) and around 150 ppm (O-substituted aromatic or phenolic C) were more prominent in the MHA than the CaHA. The aromatic region (110–140 ppm) was more pronounced in the CaHA, especially in the aerated maize and rice–soybean fields. These trends were reflected in the spectral area proportions for each fraction when averaged across soils having the same number of irrigated rice crops (Fig. 2)
.
|
|
The shapes and distinctness of some peaks changed with cropping intensity and between the two fractions in manners that were not reflected in spectral proportions. The intensively cropped soils had a more clearly defined peak at 115 ppm and a more prominent sharp resonance at 30 ppm in the alkyl region.
Figure 3 shows examples of the dipolar dephased spectra for the fractions from two soils. The alkyl peaks decreased greatly, indicating that there were very few CH3 groups and that most of the CH and CH2 groups were in rigid structures (Baldock et al., 1990; Golchin et al., 1996). The peak at 56 ppm appears to represent mostly N-alkyl C, and not methoxy C, as it almost disappeared in the dipolar dephased spectra. The peak at 103 ppm disappeared in the dipolar dephased spectra, indicating that the fractions contained very few or no tannins (Preston, 1996). The peak at 115 pm also disappeared, suggesting that it represented H-substituted aromatic C in guaiacyl units (Preston et al., 1990).
|
|
|
Figure 6 shows the solution 31P NMR spectra of the fractions from two selected soils, and Fig. 7 shows spectral area proportions for fractions when averaged across soils with the same number of irrigated rice crops. As reported by Mahieu et al. (2000a), monoester P (3 to 6 ppm) dominated the spectra, accounting for 40 to 70% of the total peak area, generally less in the submerged soils than in the aerated ones and less in the MHA than the CaHA. Diester P (-1 to 0.7 ppm) was the second most important species with 8 to 25% of the total peak area. The region between 0.7 and 3 ppm (sugar diester P) was nearly as prominent as that of diester P in some samples. The combined proportions of diester P and sugar diester P increased with submergence (from 26–28% for the soils with one or no irrigated rice crop to 42–44% for the soil with three irrigated crops per year). They were much greater in the MHA than in the CaHA. The sharp signal at about 6 ppm (inorganic P) constituted about 10% of total spectral area. The region between -2 and -1 ppm (unknown P; Mahieu et al., 2000a) contributed to about 11% of the spectral area of the fractions from the aerated soils. There was little or no signal in this region for the submerged soils. The ratio of monoester to diester and sugar diester, an index of humification, was in the MHA about half that of the CaHA for most soils and was greater for the aerated soils (data not shown). Fertilizer treatment appeared to have little influence on P forms. Only three spectra had signals from phosphonate and none had signals from pyrophosphates.
|
|
For all HA samples, visible light absorption and free radical concentration were strongly positively correlated (all but one case P < 0.001, N = 24) with spectral proportions of aromatic C (r > 0.81), carboxyl C (r > 0.61), heterocyclic N (r > 0.85), monoester P (r > 0.78), and the mono- to diester P ratio (r > 0.75) (Table 2). They were strongly negatively correlated (all P < 0.001, N = 24) with amide N (r < -0.78) and N-alkyl C (r < -0.84). Hydrogen concentration had strong correlations of the opposite signs with the same spectral proportions (all P < 0.001, N = 24). For the MHAs alone, visible light absorption and free radical concentration were strongly positively correlated with proportions of carboxyl C (r > 0.72, P < 0.01, N = 12), monoester P (r > 0.67, P < 0.05, N = 12), and mono- to diester P (r > 0.60, P < 0.05, N = 12), and negatively with proportions of amide N (r < -0.71, P < 0.01, N = 12), and N-alkyl C (r < 0.72, P < 0.01, N = 12). Hydrogen concentration was inversely correlated with the same spectral proportions (|r| > 0.85, P < 0.001, N = 12).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Spectral proportions of functional groups and humic properties conventionally associated with humification showed that the MHA and CaHA were made of less humified material with increasing intensity of irrigated rice cropping. Visible light absorption and the concentration of free radicals decreased and H concentration and labile forms of P increased. Both fractions became poorer in aromatic and carboxyl C, inorganic and monoester P, and heterocyclic and amino N, and richer in N-alkyl, O-alkyl, and phenolic C and amide N.
More recent analyses of these same HA samples plus the associated whole soils by pyrolysis–gas chromatography–mass spectrometry demonstrated that depicted changes in the chemical nature of the MHA and CaHA with increasing cropping intensity are indicative of changes in bulk SOM (D.C. Olk, unpublished data, 2000). Specifically, for the MHA, CaHA, and whole soil, intensified cropping was associated with increased relative abundance of lignin residues and decreased relative abundance of nitrogenous compounds, mostly heterocyclic N.
These results illustrate in reverse order the concept of humification as the depletion of labile compounds over time in favor of more recalcitrant compounds through various biotic and abiotic processes (Stevenson, 1994; Zech et al., 1997). With increasing cropping intensity, the SOM originally present in the soil appears to become diluted by additions of incompletely decomposed crop residues. Consequently, light absorption decreases as the SOM becomes increasingly plantlike. Carbon forms shift from microbially reworked aromatics toward lignin residues, especially phenols. Levels of H, amide N, and diester P increase as their input rates exceed degradation rates, and they dilute humified compounds of the original SOM, such as heterocyclic N, monoester P, and free radicals.
The vast majority of NMR spectral proportions for each nucleus (13C, 15N, and 31P) were highly correlated with NMR spectral proportions for other nuclei and with non-NMR measurements of humic properties. Thus, all NMR spectroscopy methods detected parallel changes in the chemical state of HA, whether between the MHA and CaHA or with increasing degree of soil submergence. All NMR spectroscopy methods were correlated to comparable degrees with visible light absorption, H concentration, and organic free radical concentration: for each nucleus there were at least two spectral regions having correlation coefficients of absolute value greater than 0.80 (Table 2). Of all spectral regions, the N-alkyl C region had the strongest correlation coefficients with these non-NMR measurements. The weakest correlations were for amine N, alkyl C, acetal C, and unknown P and were for the most part statistically nonsignificant.
Heterocyclic N also had some of the strongest correlation coefficients with the non-NMR properties. Together with the high values for amide N, these correlations provide some of the best evidence currently available for the capacity of 15N NMR spectroscopy to reproducibly detect heterocyclic N and other organic N forms. By demonstrating regular variation in the occurrence of heterocyclic N and amide N, our results may allow new insights into the factors that influence organic N forms.
Many spectral proportions were correlated with each other (Table 3). Amide N, heterocyclic N, and N-alkyl C had the most correlations, but phenolic C, carboxylic C, monoester P, sugar diester P, and diester P were also correlated with many others. Correlations of amide N, N-alkyl C, phenolic C, sugar diester P, and diester P were of the same sign whereas correlations of heterocyclic N, aromatic C, carboxylic C, and monoester P ran opposite to them.
The 13C CPMAS spectra of our fractions resembled the published spectra of other humic acids, except for the stronger N-alkyl and phenolic C peaks in the intensively cropped fields. These peaks were prominent in HA of submerged soils for different soil types across the LTFE sites and with varying quantities of amended inorganic fertilizers and organic residues. Together with a decrease in visible light absorption, the quantities of phenolic products released through CuO oxidation of whole soils (Olk et al., 1996), and large C to N ratios, these results suggest the presence of partially degraded lignin residues, including phenols. Our results represent strong evidence that this accumulation is a general characteristic of intensive irrigated rice cropping in the lowland tropics.
Previously, we speculated (Olk et al., 1996) that decreased availability of native N in the soils of double- and triple-cropped rice fields (Cassman et al., 1995) might be explained by the stabilization of organic N through bonding with the accumulating phenolic compounds. Stabilization products might resemble aniline N or perhaps aromatic amines (Flaig et al., 1975). Aniline N may further transform into heterocyclic N. Our 15N NMR study found, however, that the proportions of amine N, aniline N, and heterocyclic N did not increase with increasing degree of soil submergence for either HA fraction (Fig. 5). Therefore, formation of heterocyclic N in these soils appears to be associated exclusively with gradual humification and does not appear to be responsible for the massive stabilization of soil N that has occurred under recent intensive irrigated cropping. At the same time, other potential binding reactions between phenolic compounds and nitrogenous compounds have been suggested in the literature that would allow preservation of the amide N form (Theis, 1945; Biederbeck and Paul, 1973).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. C. Olk A Chemical Fractionation for Structure-Function Relations of Soil Organic Matter in Nutrient Cycling Soil Sci. Soc. Am. J., April 19, 2006; 70(3): 1013 - 1022. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. K. Ganjegunte, G. F. Vance, C. M. Preston, G. E. Schuman, L. J. Ingram, P. D. Stahl, and J. M. Welker Soil Organic Carbon Composition in a Northern Mixed-Grass Prairie: Effects of Grazing Soil Sci. Soc. Am. J., September 29, 2005; 69(6): 1746 - 1756. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. B. Ve, D. C. Olk, and K. G. Cassman Characterization of Humic Acid Fractions Improves Estimates of Nitrogen Mineralization Kinetics for Lowland Rice Soils Soil Sci. Soc. Am. J., July 1, 2004; 68(4): 1266 - 1277. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. B. Ve, D. C. Olk, and K. G. Cassman Nitrogen Mineralization from Humic Acid Fractions in Rice Soils Depends on Degree of Humification Soil Sci. Soc. Am. J., July 1, 2004; 68(4): 1278 - 1284. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
.
