| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Lab. of Soil Science and Geology, Dep. of Environmental Sciences, Wageningen Univ., P.O. Box 37, 6700 AA Wageningen, The Netherlands
* Corresponding author (wattel{at}hotmail.com).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Abbreviations: AMS, accelerator mass spectroscopy • ECEC, effective cation-exchange capacity • MRT, mean residence time • SOM, soil organic matter • XRD, X-ray diffraction
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
As far as we know, there is no literature available in which the effect of clay mineralogy on the MRT of organic matter is studied, except for Wattel-Koekkoek et al. (2003). In that study, we determined the MRT of kaolinite- and smectite-associated SOM (equals all SOM present in the clay-size fraction) in soils taken from the ISRIC collection, sampled from seven countries, using 14C (from now on also referred to as ‘the ISRIC experiment’). Kaolinite-associated SOM had a fast turnover (360 yr on average). Smectite-associated SOM had a relatively slow turnover, with an average MRT for the whole clay-size fraction of 1100 yr. Differences in turnover times between organic matter associated with kaolinite and smectite were significant. Multiple linear regression indicated that clay mineralogy, described by the ECEC or specific surface area, was the main factor explaining differences in the MRT of the extracted SOM.
However, in the ISRIC experiment the variation within each group was large. There are two possible factors that may cause this variation. First, the samples originated from seven different countries, implying differences in local climate, vegetation, etc. Second, for the ISRIC experiment we studied the whole clay-size fraction, including its ‘free’ organic matter. To better understand the effect of clay mineralogy on the MRT of organic matter, in this paper we focus on clay-bound organic matter and exclude the free organic fraction.
The aim of this research is first, to determine the MRT of organic matter that is bound to different clay mineral surfaces, and second, to explain the variance in the measured MRTs using multilinear regression. We especially studied organic matter that is bound to kaolinite or smectite. Kaolinite is present in a range of weathered tropical soil, such as Acrisols, Nitisols, and Oxisols (FAO, 1990). Smectite is particularly abundant in Vertisols.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
-radiation at 40 kV and 30 mA, with a focusing monochromator. The divergence slit was set at 1°, the receiving slit at 0.2 mm, and the anti-scatter slit at 1°. Peak areas of the clay minerals were measured to compare the (semi-quantitative) XRD diffractograms and reported in percentages of total peak area (Table 1). The clay fractions did not contain phyllosilicates other than kaolinite and smectite.
|
Physical Fractionation
The fractionation procedure has been described by Wattel-Koekkoek and Buurman (unpublished data, 2002). The procedure is as follows.
We collected the clay-size (<2 µm) fractions of all ten samples, using ultrasonic (full) dispersion and sedimentation in water. To ensure complete ultrasonic disruption, we measured mass recovery of clay using different values of applied energy, using the method described by Roscoe et al. (2000). To reach maximum recovery of the clay-size fraction, 430 J mL–1 was needed for RED1, RED1A, RED2, RED3, RED4, and RED5; and 365 J mL–1 for VER2, VER3A, VER3B, and PHAE3, using a soil/solution ratio of 1:10.
The clay-size fractions were siphoned off after a certain sedimentation time and then freeze-dried. In total 100 g of clay-size fraction was accumulated for all ten samples. We applied density fractionation to eliminate the possible effect of particulate organic matter. The dried clay-size fractions were separated into a light and a heavy fraction using a NaI solution with a density of 1.7 g cm–3 (Roscoe et al., 2001). Of each clay-size fraction, 50 g was placed in a 500-mL centrifuge tube with 500 mL of NaI solution, shaken end-over-end for 60 min, and left standing at room temperature for 15 min. After centrifugation (15 min, 3500 rpm), the supernatant was filtered (Whatman, GF/A, Whatman Ltd., Kent, UK) into a millipore vacuum unit. The fraction recovered on the filter was washed with 0.01 M CaCl2 solution (100 mL) and 200 mL of distilled water. The sediment was resuspended in NaI and centrifuged two more times as described above. The three subfractions recovered from the filter were joined, oven-dried at 50°C and stored for analysis. This fraction was called the "light clay-size fraction," or "free" organic matter. The heavy fraction was flocculated once with 0.01 M CaCl2 and washed about 10 times with distilled water until the clay fraction remained in suspension after 24 h, after which it was freeze-dried. This procedure was repeated twice for each soil, to separate all 100 g of clay per soil sample (Fig. 1) .
|
 | [1] |
In-Depth Analysis of a Subset
From the 10 soils, two sets of two samples with the highest content of kaolinite and smectite were selected for further analysis. We selected only two soils per clay type due to limited analytical capacity: RED3 and RED5 (kaolinite clays) and VER3A and VER3B (smectite clays). All clay-size fractions contained small quantities of quartz.
The four samples were air-dried and passed through a 2-mm sieve. The pH, organic C content, CECsoil, and particle-size distribution were measured according to Buurman et al. (1996) (Table 2). To characterize the clay minerals, the ECEC of the whole clay-size fraction was measured and the ECEC of the mineral phase was calculated (see Table 3 and the section under Effective Cation Exchange Capacity). Furthermore, the four soils were fractionated not only physically, but also chemically (see Chemical Fractionation).
|
|
 | [2] |
 | [3] |
Chemical Fractionation
The heavy clay-size fractions of RED3, RED5, VER3A, and VER3B, were shaken in 0.5 M NaOH under N2 for 24 h (0.5 L, soil/solution = 1:10). We centrifuged the four solutions, decanted the supernatants, and shook the four residues with deionized water for 2 h. After centrifugation, we acidified the combined supernatants of each sample (= NaOH extract) to pH = 1 using concentrated HCl, and added concentrated HF until a concentration of 0.3 M HF was reached. The four supernatants were dialyzed to pH 6 against deionized water, and freeze-dried (NaOH extract). Next, the NaOH-residue was shaken for 24 h with 0.1 M Na4P2O7 under N (soil/solution = 1:10). The solution was centrifuged and decanted, and the residue was shaken with deionized water for 2 h. After centrifugation, we acidified the combined supernatants to pH 1 with concentrated HCl, and added concentrated HF until a concentration of 0.3 M HF was reached. Both the supernatant (pyrophosphate extract) and the pyrophosphate-residue were dialyzed and freeze-dried (Fig. 1).
We chose to extract first with NaOH because we expect it to extract mainly free (Choudri and Stevenson, 1957) and kaolinite-complexed organic components. Sodium hydroxide deprotonates the aluminum-hydroxide edges of kaolinites, and part of the organic matter, thereby liberating organic molecules. We chose Na4P2O7 for subsequent extraction, because we expect it to form complexes with (exchangeable) polyvalent cations present at smectite surfaces, thereby breaking down the cation bridges between the exchangeable cations and organic matter (Choudri and Stevenson, 1957; Schnitzer and Schuppli, 1989). We measured the amount of C in the total and heavy clay-size fraction, the freeze-dried extracts, and the residue, using a Carlo Erba elemental analyzer, model EA 1108 (Carlo Erba, Milan, Italy). To determine the C content of the total and the heavy clay-size fractions three replicates of RED3, RED5, VER3A, and VER3B were taken. The C content of the NaOH and Na4P2O7 extracts was determined only in duplicate, because of lack of material. Standard deviations were not calculated for these extracts.
Furthermore the 14C age of the NaOH-extracted, Na4P2O7–extracted and residual organic material was measured.
Density Fractionation Method Test
To test the density fractionation method, we filtered the NaOH and Na4P2O7 extracts using a glass filter (Schleicher and Schuell, Dassel, Germany, GF 55, 
1 µm). If the NaI method failed, the filters should catch ‘free’ organic matter >1 µm such as plant remains, and the filtrate should be older of age than the non-filtered extracts, assuming that relatively large plant remains have a young 14C age.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
The smectitic whole clay-size fractions of VER3A and VER3B had an ECEC of 37 and 43 cmol kg–1 respectively, mostly due to the mineral phase, and 2.4 g Fe 100 g clay–1 of dithionite extractable Fe.
Carbon Distribution
The total clay-size fractions had C contents varying between 2 and 4% (mass fraction) (Table 4). Compared with the total mass and C content of the whole clay-size fraction, the yield of the light fraction was negligible. After NaI treatment, all the C present in the clay-size fraction ended up in the heavy clay-size fraction. Therefore, the C content of heavy and total clay-size fractions were not significantly different.
|
Carbon-14 Age
Table 5 shows the measured 14C activity, the calculated MRT and corrected 14C activity for the whole and the heavy clay-size fractions. In the kaolinitic soils, the average MRT of organic matter was 523 ± 137 yr in the clay-size fractions and 536 ± 136 in the heavy clay-size fractions. The average MRT of organic matter in smectite-dominated soils was 438 ± 227 yr in the clay-size fractions and 456 ± 256 in the heavy clay-size fractions. T-tests (Sokal and Rohlf, 1995) indicate that in both kaolinitic and smectitic soils, the 14C ages of whole and heavy clay-size fraction were not significantly different from each other. Also the MRT of the organic matter in the whole and heavy clay-size fraction did not differ between kaolinite-dominated and smectite-dominated soils.
|
|
The organic matter in the pyrophosphate-residues had a MRT of 381 ± 60 yr for RED3, 452 ± 58 yr for RED5, 328 ± 61 yr for VER3A, and 412 ± 60 yr for VER3B. The results seem to indicate that the turnover time of SOM in residues also does not differ between kaolinite- and smectite-dominated soils.
The measurements suggest that the four fractions of the RED3, RED5, VER3A, and VER3B (heavy clay-size fraction, NaOH extract, Na4P2O7 extract, residues) were not significantly different from each other. However, of the four fractions, the pyrophosphate extracts seem to be slightly older than the other ones.
Did Density Fractionation Work?
There was no significant difference in 14C activity between filtered and non-filtered extracts (Table 7). With the naked eye, we only saw a small amount of mineral material on the filter. Both indicate that the density fractionation did work. However, we cannot be completely certain because it also possible that the free organic fraction was smaller than 1 µm.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
During the ISRIC experiment, we did find ‘free’ plant remains in the clay-size fraction using a scanning electron microscope (Wattel-Koekkoek et al., 2003). In the present case, the mass of light organic matter separated by NaI was negligible (Table 4). The 14C measurements support these findings, as the whole and the heavy clay-size fractions are not significantly different in 14C activity (Table 5). This may indicate that all organic matter in the clay-size fraction is bound or that density fractionation is not effective to separate fine organic materials. Previous studies also show very low yields of the light fraction of the clay-size fraction and suggest that in clay-size fractions a continuum exists between density extremes of organic and mineral properties (Christensen, 1992).
Kaolinite-Dominated Clay-Size Fractions
RED3 and RED5 have a high content of dithionite-extractable Fe and little oxalate-extractable Fe, which indicates that their clay-size fractions contain a lot of crystalline Fe oxides. Furthermore, both soils had a red color, suggesting that the crystalline Fe is hematite.
The SOM in the whole and heavy clay-size fractions of all six kaolinite-dominated soils had a fast turnover (Table 5). Apparently the kaolinite and the Fe-oxides present, bind organic matter weakly, hardly affecting SOM turnover. These results agree with our findings in the ISRIC experiment (Wattel-Koekkoek et al., 2003), and with those of Shang and Tiessen (1997).
Of all organic C present in the extracts of the heavy clay-size fractions of RED3 and RED5, 50% was extractable by NaOH and 15% by Na4P2O7. Similar results of extractability were found by Wattel-Koekkoek et al. (2001).
Wattel-Koekkoek et al. (2001)(Wattel-Koekkoek and Buurman, unpublished data, 2002) showed that NaOH-extracted SOM from kaolinite is rich in polysaccharides. Polysaccharides are relatively easy to decompose, so their presence suggests that kaolinite-bound SOM has a relatively fast turnover. This is confirmed by the short MRT of the NaOH extracts of RED3 and RED5 (Table 6).
Although not statistically significant, the SOM in the pyrophosphate extracts of RED3 and especially RED5 seem to have a slightly longer MRT compared with the other fractions of these soils and compared with the pyrophosphate extracted SOM of the kaolinitic soils of the ISRIC experiment. This could be related to the chemical composition of the pyrophosphate-extracted SOM. According to Wattel-Koekkoek and Buurman (unpublished data, 2002), the pyrophosphate extracts of RED3 and RED5 contain relatively many aromatic compounds. These compounds have a larger recalcitrance toward decomposition than polysaccharides (Baldock et al., 1992).
The organic C in the pyrophosphate-residues of RED3 and RED5, which represents about 35% of the C in the heavy clay-size fraction (Table 4), had a relatively fast turnover (Table 6). This suggests that the organic matter in the residues is only weakly bound to the mineral surfaces.
Smectite-Dominated Clay-Size Fractions
The organic matter in the whole and heavy clay-size fractions of the four smectite-dominated soils had a relatively fast turnover: 438 yr for the whole clay-size fraction and 456 yr for the heavy clay-size fraction on average (Table 5). These results contrast with those of the ISRIC experiment, where smectite-associated SOM had a relatively slow turnover (average MRT for the whole clay-size fraction of 1100 yr). Theng et al. (1992) even found a MRT of 5680 yr for the organic matter in a smectitic clay-size fraction from New Zealand.
About 50% of all organic C present in the heavy clay-size fraction of the smectite-dominated soils was extractable by NaOH (Table 4). This contrasts with Wattel-Koekkoek et al. (2001), who found that only a small part (15%) of the organic matter in the clay-size fraction of smectite-dominated soils was extractable by NaOH, and that the extractability by subsequent pyrophosphate was much higher.
The SOM in the NaOH and Na4P2O7 extracts of VER3A and VER3B have a relatively short MRT of about 400 to 600 yr. This agrees with Wattel-Koekkoek and Buurman (unpublished data, 2002), who found that the NaOH and Na4P2O7 extracts from the heavy clays of VER3A and VER3B, were dominated by easily decomposable O-alkyl (polysaccharide) C, suggesting fast turnover. The fast turnover, however, seems to disagree with the ISRIC experiment, where we found that the organic matter extracted by NaOH and Na4P2O7 had an average MRT of 730 and 2000 yr respectively. Also Arai et al. (1996) analyzed the 14C activity of the C in a combined NaOH/Na4P2O7 extract of a Vertisol from India and found a relatively long MRT of 4650 yr.
Pyrophosphate-extracted organic matter of smectitic soils VER3A and VER3B seems to have the longest MRT compared with the NaOH extract and residue. This agrees with the ISRIC experiment, where pyrophosphate extracted SOM of smectitic clays was significantly older than the other fractions. It was then suggested that smectitic clays can retain SOM by forming cationic bridges with the SOM, thereby restricting decomposition.
The organic C in the pyrophosphate-residues of VER3A and VER3B had a relatively fast turnover (Table 6). This indicates that the organic matter in the residues was only weakly bound to the mineral surfaces, or that density fractionation did not work and the residue contained particulate organic matter (<1 µm).
In summary, the organic matter bound to the smectite-dominated clays from Mozambique was largely extractable by NaOH, was probably rich in polysaccharides (Wattel-Koekkoek and Buurman, unpublished data, 2002), and had a relatively fast turnover. The organic matter of the smectitic soils from the ISRIC experiment was extractable by Na4P2O7 rather than NaOH, relatively rich in aromatic C, and had a long MRT. How can we reconcile these seemingly contrasting results?
Factors that Determine the Mean Residence Time
Multi-linear regression of the data of the ISRIC experiment (Wattel-Koekkoek et al., 2003) revealed that the main factors explaining differences in the 14C activity of the organic matter in the different fractions, were the ECECmin and the specific surface area of the clay-size fraction, each on their own explaining 43 and 50% of the variance. This showed that charge and the amount of mineral surface available play a crucial role in the retention of organic matter. Furthermore, we then saw a negative regression coefficient between the ECECmin and the 14C activity of a fraction, and suggested that clays with a high ECEC can stabilize SOM via for example, cationic bridges, thereby limiting the decomposability, resulting in a high MRT, and thus a low 14C activity.
The smectitic soils used during the ISRIC experiment had an average ECECmin of 67 ± 22 cmol kg–1. The smectitic soil from Mozambique however, only had an ECECmin of about 29 cmol kg–1 (Table 3). In the ISRIC-experiment, SOM from smectite-dominated samples was hardly extractable by NaOH and largely extractable by pyrophosphate, supposedly because pyrophosphate can dissolve SOM that is bound via exchangeable cations. The relatively low number of polyvalent cations in the smectitic soils from Mozambique may explain why a large part was extractable by NaOH. Furthermore, the low number of polyvalent cations in smectitic soils from Mozambique may have limited its ability to bind SOM via cation bridges, resulting in a short MRT and thus a high 14C activity. In short, we hypothesize that the seemingly contrasting behavior by the smectitic soils from Mozambique is caused by their low ECEC and that in actuality they do behave similar.
To check the role of the ECECmin, and other possible explaining factors for variation in 14C age, we calculated (multiple) linear regressions (n = 22; 14 from the ISRIC experiment and eight from Mozambique), using 14acorr of the extracts as dependent variable and a selection of independent variables. The selection was made based on literature and the available data.
We used
Part of the independent variables, for example, the NMR data, was measured earlier (Wattel-Koekkoek et al., 2001; Wattel-Koekkoek and Buurman, unpublished data, 2002).
The single factor best explaining the variance in 14C activity was the ECECmin (R2adj = 0.43) (Table 8), which correlated negatively. This agrees with results from the multiple regression from the ISRIC experiment (Wattel-Koekkoek et al., 2003). Furthermore, this finding supports our hypothesis that the apparently deviating behavior by the smectitic soils from Mozambique is caused by their low ECEC, and that in fact they behave similar to the soils from the ISRIC experiment.
|
When using only one of the NMR components as independent variable (O-alkyl, alkyl, carbonyl, or alkyl-C), Aromatic C best explains the variance in 14C activity (R2adj = 0.19). Aromatic C correlates negatively, presumably because aromatic C is relatively resistant to decomposition. Carbonyl C also shows a negative regression coefficient. Carbonyl groups may lower decomposition rates by forming bonds with exchangeable cations at the clay surface. Alkyl C did not explain variance in 14C activity.
The type of extract is positively correlated with 14C activity, indicating that the NaOH extract contains young, easily decomposable SOM and the pyrophosphate extract contains old, recalcitrant SOM. This suggests that pyrophosphate-extracted SOM was relatively strongly bound (e.g., via exchangeable cations) and hydroxide-extracted SOM rather loosely bound to the mineral surface. Temperature shows a positive regression coefficient with 14C activity of the extracts, explaining 27% of the variance. The positive regression coefficient was expected because microbial activity and thus decomposition increases with temperature.
The optimal fit (R2adj = 0.70), was reached when using ECECmin, alkyl-C %, temperature, and extract type as input variables. This agrees with the ISRIC experiment, where the same factors explained 67% of the variance, indicating again that the soils from Mozambique behave similar to the soils from the ISRIC. The second best set of variables was smectite percentage, alkyl-C percentage, temperature, and extract type, explaining 68% of the variance (Table 8).
When ECEC or percentage of smectite is used as independent variables, the percentage of alkyl-C, and not O-alkyl, is the best predicting factor of the chemical data available. Addition of any other NMR-based variable lowers the coefficient of determination. This suggests a relationship between clay mineralogy and percentage of alkyl associated to the clay. Theng et al. (1992) found that aliphatic components are intercalated in the interlayers of smectites of a Spodosol. However, it seems improbable that the aliphatic components that we found in the extracts were intercalated, because intercalated matter is bound very strongly and is unlikely to be dissolved by NaOH or Na4P2O7.
It is remarkable that the percentage of alkyl-C shows a positive regression coefficient with 14C activity. We expected a negative regression coefficient with 14C activity as aliphatic components are relatively resistant to decomposition (Lichtfouse et al., 1995; Huang et al., 1999). Our results however agree with Meredith (1997), who analyzed soils of several ages with solid state 13C NMR, and found that an old stagnopodzol soil had a lower proportion of alkyl C than younger brown earth soils. Although his study concerns completely different soils than ours, it is remarkable in both studies that a high content of alkyl-C is related to young organic matter. Possibly the alkyl-C represents relatively easily decomposable microbial lipids.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Received for publication February 26, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
13C distribution in particle-size fractions. Eur. J. Soil Sci. 51:445–454.This article has been cited by other articles:
|
|
 |
|
  Y. L. Zinn, R. Lal, J. M. Bigham, and D. V. S. Resck Edaphic Controls on Soil Organic Carbon Retention in the Brazilian Cerrado: Texture and Mineralogy Soil Sci. Soc. Am. J., June 8, 2007; 71(4): 1204 - 1214. [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 |
Journal of Environmental Quality |
||||
