| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Institute of Soil Science and Soil Geography, University of Bayreuth, D-95440 Bayreuth, Germany
klaus.kaiser{at}uni-bayreuth.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
0.1 M) of competing anions like SO2-4 or H2PO-4 increased the NOM desorption. Though H2PO-4 was most efficient in desorbing NOM, the extractability was only 60% at the highest H2PO-4 concentration. The most significant desorption occurred when solution pH was raised. For goethite, NOM desorption reached a maximum at a pH above the point of zero charge (PZC) of the mineral. With increasing surface coverage of the sorbent by NOM, the proportion of desorbable NOM decreased for all extractants. Increased sorption hysteresis was also observed with an increasing time period between sorption and desorption. The desorption was more pronounced for NOM compounds that exhibit hydrophilic properties and have low contents of aromatic structures and carboxyl groups. The irreversible binding of NOM, especially of the lignin-derived portion, to soil minerals seems to result from its polyelectrolytic nature. This may favor the formation of multi-site coordinative bonds and effective shielding of the binding ligands by other parts of the sorbed molecule.
Abbreviations: DOC, dissolved organic carbon • NOM, natural organic matter • OC, organic carbon • PZC, point of zero charge
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
Sorption of NOM to mineral surfaces is not only the major control for NOM and NOM-assisted transport, it also alters the surface properties of the sorbent. Surface coverage with NOM increases the binding capacity for metals and organic pollutants (Davis and Leckie, 1978; Murphy et al., 1992) and improves the resistance against weathering (Smeck and Novak, 1994). In addition, sorption to mineral surfaces is assumed to contribute to the stabilization and accumulation of NOM in soils (Hedges and Oades, 1997). Thus, knowledge of the reversibility of NOM sorption to mineral phases is important for predicting transport of pollutants and element cycling in soils.
While the sorption of NOM to minerals and soils is well investigated, little work has focused on desorption of NOM. The limited studies on this subject indicate that there is little desorption of NOM from soil material or Fe oxides under solution conditions similar to those during sorption (Qualls and Haines, 1992; Gu et al., 1994). No information on the release of sorbed NOM due to the changing composition of the soil solution is available yet. It is also unknown whether fractionation of sorbed NOM compounds occurs during the desorption. Sorption of inorganic anions such as phosphate and sulfate by surface complexation to hydrous oxide surfaces exhibits increasing hysteresis with time (Barrow and Shaw, 1975, 1977). It is reasonable to assume that NOM sorption to hydrous oxides may exhibit similar characteristics as this process also involves surface complexation via ligand exchange (e.g., Parfitt et al., 1977; Tipping, 1981; Wershaw et al., 1995).
The objective of our study was to examine the reversibility of NOM sorption to soils and soil minerals and its controlling factors. We conducted batch experiments that determined NOM release from amorphous Al(OH)3, goethite, and subsoil material low in organic carbon (OC) with solutions of varying composition. Surface soil materials were not included in this study because they contain not only sorbed but also particulate NOM. This material contributes to NOM release in desorption experiments, thus making proper interpretation of results in terms of reversing sorption processes impossible. We also investigated the effects of aging on the reversibility of NOM sorption. Fractionation, according to Aiken and Leenheer (1993), and 13C-NMR spectroscopy were used to characterize the functional group distribution and structure of desorbed NOM functionally and structurally.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
|
-FeOOH) was prepared by slow oxidation of a 0.5 M FeSO4 solution adjusted to pH 7 (Schwertmann and Cornell, 1991). Instead of using NaHCO3, the solution was kept at pH 7 by a microprocessor-controlled automatic burette containing 1 M NaOH. The precipitate was washed with deionized water and dialyzed against deionized water for 14 d. Finally, it was freeze-dried and sieved to particles <0.63 mm. X-ray diffraction showed that the sample was pure and of low crystallinity. Amorphous Al(OH)3 was precipitated by slowly neutralizing a solution of 1 M Al(NO3)3 with NaOH (Huang et al., 1977). The fresh precipitate was washed with deionized water, dialyzed against deionized water for 7 d and freeze-dried. Then it was sieved to particles <0.63 mm. The product was X-ray amorphous. The two minerals were stored at 5°C. Soil material was taken from the 3Bw horizon of an Oxyaquic Haplumbrept (Soil Survey Staff, 1994) derived from a carbonate-free Jurassic claystone. The sampling site was located near Bamberg, NE Bavaria, Germany. It was covered by a 70-year old Scotch pine (Pinus sylvestris L.) forest. The sample was air-dried and sieved to <2 mm prior to the sorption experiments.
Selected properties of the two minerals and the soil material are given in Table 2 .
|
Due to the release of indigenous NOM from the soils, the results of the sorption experiments did not conform to Freundlich nor Langmuir isotherms. For systems in which native adsorbed substances need to be considered, the initial mass approach (Nodvin et al., 1986) is useful. This model has described the sorption of DOC to soils (Vance and David, 1992; Guggenberger and Zech, 1992; Donald et al., 1993). In the initial mass approach, the concentration of a substance adsorbed or released (normalized to soil mass) is plotted as a function of the initial concentration of the substance (normalized to soil mass). The amount of NOM sorbed was calculated by the difference between DOC concentrations in the initial and the equilibrated solutions.
Desorption experiments were carried out within 24 h after the sorption experiment. The experimental conditions were the same as those used for the sorption experiments (5°C, 24 h shaking, solid:solution ratio = 1:5 for the soil material and 1:80 for the minerals). Solutions used for the desorption were solutions free of DOC but of similar inorganic composition to the initial solutions (i.e., artificial soil solution), distilled H2O, 0.1 M Na2SO4, 0.1 M NaH2PO4, and 0.1 M NaOH.
The influence of time on NOM desorption by 0.1 M NaH2PO4 and 0.1 M NaOH was investigated using subsamples of the hydrous oxides (highest surface loadings) that were extracted 24, 48, 72, and 120 h after the sorption experiment.
The effects of increasing concentrations of NaH2PO4, Na2SO4, and NaCl were examined. The concentrations were 0.05, 0.1, 0.5, 1, 5, 10, 50, and 100 mM. Only sorbents equilibrated with the highest DOC concentration were used.
An additional set of desorption experiments was carried out with distilled H2O adjusted to pH values ranging from 4.0 to 9.8 by addition of either 1 M HCl or 1 M NaOH. Again, only sorbents equilibrated with the highest DOC concentration were used. After 4 and 8 h of shaking, the pH values were readjusted. Amorphous Al(OH)3 was excluded from this experiment because of its dissolution at low and high pH values (<4.5 and >8.5).
After the 24-h shaking period, the suspensions were filtered through 0.45-µm polyethersulfone membrane filters and the DOC concentrations and the DOC fractions were determined in the filtrates. From the water content (see above) and the concentrations of DOC and DOC fractions in the equilibrated supernatants, the amount of DOC remaining in the sorbent material from the sorption experiment was calculated and used to correct the amounts of desorbed OC.
Electrophoresis with Mineral Particles
The electrophoretic mobility of goethite and Al(OH)3 particles with and without sorbed NOM was measured at 20°C in a cubic cell with a microelectrophoresis unit (Zetasizer 3000, Malvern Instruments, Malvern, UK). A minimum of 20 particles were timed in each direction and a mean velocity was calculated from the measurements. Electrophoretic mobilities were transformed into zeta potentials (
). Mineral samples for electrophoretic measurements were prepared by suspending 50 mg of the mineral phases in 1 L of 0.002 M NaCl at pH 4. The suspensions were shaken for 24 h at 5°C prior to the measurements.
Carbon-13 Nuclear Magnetic Resonance Spectroscopy of Sorbing and Desorbing Natural Organic Matter
For the 13C-NMR spectroscopic characterization of solution and desorbed NOM, 5 L of the solution with the highest DOC concentration were added to 25 g of goethite, shaken for 24 h at 5°C, and centrifuged at 2000 g. The supernatant was filtered through a 0.45-µm polyethersulfone membrane filter. The filtrate was acidified to pH 2 with 1 M HCl, then pumped through a column of XAD-8 resin. The effluent of the XAD-8 column, which contained the hydrophilic NOM fraction, was pumped through a column packed with a strongly acidic cation exchange resin (AG MP-50, BioRad Laboratories, Richmond, CA) to remove cations other than H+. As hydrochloric acid forms a low-boiling azeotrope with acetonitrile, Cl- was removed from the sample by repeated evaporation with acetonitrile (Aiken and Leenheer, 1993). Finally, the sample was dissolved in distilled water and freeze-dried. After the passage of the hydrophilic fraction, the XAD-8 column was rinsed with 0.01 M HCl. The NOM adsorbed to the XAD-8 (hydrophobic acidic fraction) was eluted with 0.1 M NaOH. The eluate was then pumped through an AG MP-50 column for protonation and freeze-dried.
The sorbent material was extracted with 0.1 M NaH2PO4. Experimental conditions and treatment of the desorbed NOM were the same as for the sorption experiments. Phosphate was precipitated from the cation exchanger–treated total and hydrophilic NOM using 1 M Mg acetate/NH4 acetate buffered at pH 8 (Aiken and Leenheer, 1993). Excess Mg2+ and NH+4 were removed using an AG MP-50 column; then the solution was evaporated in order to remove acetic acid. The whole procedure caused a loss of organic C of <3%.
In the initial solutions and the equilibrium solutions of the sorption and desorption experiments, the hydrophobic acidic and hydrophilic fractions accounted for 94 to 98% of the total DOC. These values suggest that only a small proportion of the DOC (2–6%) was within the hydrophobic neutral fraction. The hydrophobic acidic fraction, therefore, represented nearly the entire hydrophobic NOM fraction.
Liquid-state 13C-NMR spectra of the hydrophilic and hydrophobic NOM fractions were recorded on an Avance DRX 500 spectrometer (Bruker GmbH, Karlsruhe, Germany) at a resonance frequency of 125.77 MHz for 13C. Samples of 100 to 150 mg were dissolved in 3 mL of 0.5 M NaOD. Using a 2.0-s pulse delay and inverse-gated decoupling, 
30 000 scans were recorded for each sample. Resonance areas were calculated by electronic integration. In general, the 13C-NMR spectra of NOM consisted of four regions: (i) the alkyl region (0–50 ppm), mainly representing C atoms bonded to other C atoms (methyl, methylene, and methine groups); (ii) the O–alkyl region (50–110 ppm), mainly representing C bonded to O (carbohydrates, alcohols, and ethers); (iii) the aromatic region (110–160 ppm), representing C in aromatic systems and olefins, and (iv) the carbonyl region (160–210 ppm), including carboxyl C (160–190 ppm) and ketone C (190–210 ppm). Further information on the assignment of 13C-NMR regions is given by Wilson (1987) and McKnight et al. (1992).
|
Results and discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
|
|
5 mmol kg-1. Less than 3% of the readily sorbed OC was released from goethite and Al(OH)3 by the artificial soil solution (Fig. 1). The desorption was also low for the 3Bw horizon, but it was not possible to quantify the exact amount because the soil material released indigenous OC. Thus, the sorption of NOM to the three mineral phases showed the same strong hysteresis as observed for hematite (Gu et al., 1994) and for soil material from an Ultisol AB horizon (Qualls and Haines, 1991).
Natural Organic Matter Desorption under the Influence of the Solution Composition
Figure 2
shows that the desorption of OC from the minerals and the 3Bw horizon was not a function of increasing concentrations of NaCl. Hence, the desorption of NOM sorbed to mineral surfaces seems to be independent of the solution ionic strength. The same independence was found for the sorption of NOM on soils and hematite (Jardine et al., 1989; Gu et al., 1994).
|
60% of the DOC sorbed to goethite or Al(OH)3 was recovered. One reason for this finding may be multiple-site binding between NOM and the sorbents, so that desorption requires simultaneous detachment of all binding ligands (Stumm, 1992; Gu et al., 1994). Another explanation may involve the shielding of binding ligands against exchange with competing anions by other moieties of the macromolecule. Some of the nonbinding, negatively charged carboxylate groups may repel other anions and thus prevent them from approaching the binding sites. Such a consideration agrees well with the finding that NOM sorbed to the surface of hydrous Al and Fe oxides shifts the PZC of the sorbent to lower pH and, therefore, increases the sorption of cations (Davis and Leckie, 1978; Tipping, 1981). This assumption was justified by negative zeta potentials () that were measured for mineral particles with large surface loading (Fig. 6)
.
|
|
|
Desorption of Natural Organic Matter at Increasing Surface Coverages
For all extractants, the desorbed portion of OC decreased with increasing surface coverage (Fig. 5)
. In contrast, synthetic polymers exhibited the lowest desorption at low surface coverage; i.e., when multisite binding of the sorbing molecules is most likely (Podoll et al., 1987). At higher surface coverage, the surface area available for a sorbing macromolecule is presumably smaller. The sorbing molecules, therefore, tend to sorb to mineral surfaces in a configuration that results in fewer attachment points (Podoll et al., 1987; Stumm, 1992). For NOM sorbed to the surfaces of Al and Fe hydrous oxides, it has been shown by IR spectroscopy that with increasing surface coverage fewer carboxyl groups per molecule are involved in complexation with surface metals (Parfitt et al., 1977; Kaiser et al., 1997). Thus, higher surface coverage normally should favor desorption of the sorbed NOM. Differences in the relationship between surface coverage and the reversibility of the sorption of synthetic polymers and NOM may result from their different chemical compositions and structures. Generally, the synthetic polymers were linear (e.g., Podoll et al., 1987), whereas NOM contains complex, three-dimensional polyelectrolytes. Sorption of NOM at higher surface coverage may reduce the number of bindings per molecule, but at the same time it may enhance the repulsion of competing anions caused by nonbinding ligands. The change of the surface charge, expressed as zeta potentials () from positive values for the uncovered mineral phases to negative values at higher NOM surface coverages confirms this assumption (Fig. 6).
|
A third explanation for the reduced desorption at higher surface coverage may be that with the closer package of molecules, intramolecular interactions (i.e., metal bridging) became more prominent and led to the formation of greater NOM units that were less easily desorbable. Because the major cations in the NOM solution used for the sorption experiments were H+, NH+4, and K+, bridging effects were unlikely.
Sorption of NOM is known to increase the sorption capacity of mineral surfaces for hydrophobic organic molecules, such as polycyclic aromatic hydrocarbons and chlorobenzenes (Murphy et al., 1992; Barber, 1994). Thus, a fourth possible explanation of decreasing NOM desorption at high surface loadings may involve interactions between hydrophobic moieties of the sorbed molecules. Yet, unlike uncharged hydrophobic contaminants, the NOM used in this study contained numerous acidic groups that were at least partly ionized. The so-called hydrophobic NOM fraction was particularly rich in carboxylic acidity (Table 1). According to 13C-NMR spectroscopy, each sixth C atom in this fraction belonged to a carboxyl group. In addition, the prominent resonances in the O–alkyl region indicate that a major portion of NOM were carbohydrates. (Table 4) . The predominance of hydrophilic moieties within the NOM used here suggests that direct NOM x NOM interactions were limited. Nevertheless, they cannot be ruled out.
|
88% for goethite and 97% for Al(OH)3 at the first two extraction times. Natural organic matter desorption from the goethite decreased for the two later extraction times to 80 and 71%, whereas the release from Al(OH)3 did not decrease as a function of time. The reason for the different NOM release from the two minerals involved the complete dissolution of Al(OH)3 in the strongly alkaline extractant.
|
Structural Aspects of Natural Organic Matter Desorption
In the sorption of NOM to minerals and the 3Bw horizon, the hydrophobic (or so-called humic) NOM fraction was preferentially removed from the solution (Fig. 8)
. This fractionation is a well documented feature of the NOM sorption to soils and minerals (e.g., Jardine et al., 1989; Kaiser et al., 1997). It occurs with a preferential sorption of the high-molecular-weight NOM (e.g., Gu et al., 1995; Wang et al., 1997). These results are comparable, as the hydrophobic fraction represents the high-molecular-weight components within the NOM (Table 1; Donald et al., 1993; Gu et al., 1995). Some authors considered favorable entropy changes and hydrophobic interactions to be responsible for the high-affinity sorption of the hydrophobic fraction (Jardine et al., 1989; Baham and Sposito, 1994). On the other hand, there is evidence that the sorption of NOM results from strong-binding ligands, e.g., of the phthalic and salicylic acid type (Jekel, 1986; Gu et al., 1994; 1995; Edwards et al., 1996), which are more abundant in the hydrophobic fraction that derives from lignin degradation (Guggenberger and Zech, 1994). Carbon 13–NMR spectroscopy suggests that aromatic acid ligands occur in the hydrophobic acid fraction, whereas the hydrophilic fraction completely lacks aromatic C (Table 4). Even within the hydrophobic acid fraction of NOM, preferential removal of carboxyl and aromatic C moieties has been noticed in the sorption to soils and minerals (Table 4; McKnight et al., 1992; Kaiser et al., 1997). Because these moieties seem to form strong complexes with metals on the surface of Al and Fe oxides (Gu et al., 1994; 1995; Kaiser et al., 1997), low desorption of the hydrophobic fraction is to be expected.
|
-4. The hydrophilic NOM fraction was completely desorbed from goethite and Al(OH)3 by 0.1 M NaH2PO4, as well as by 0.1 M NaOH. The release by 0.1 M Na2SO4 amounted to 13 to 20%. Overall, the effects of SO2-4 and H2PO-4 on desorption of NOM fractions were similar to their effects during the NOM sorption (Gu et al., 1995; Kaiser and Zech, 1997). The desorption of hydrophilic NOM by NaH2PO4 and NaOH was complete for all surface coverages investigated (Fig. 8). The desorption of hydrophobic NOM, in contrast, decreased with increasing loading of the sorbent. One explanation of the reduced desorption of total OC at high surface coverage was that the preferential sorption of the hydrophobic fraction formed strong complexes with surface metals. Because hydrophobic NOM contains more acidic functional groups than the hydrophilic fraction (Table 1 and 4), more ligands per molecule may form strong surface complexes.
The sorption of the hydrophilic DOC fraction on goethite showed no preferential removal of any C moieties from the solution (Table 4). Sorption of this fraction seems to be weak but uniform. In the sorption of the hydrophobic fraction, there was preferential accumulation of aromatic and carboxyl C on the sorbent, leaving more alkyl C in the aqueous phase. The O–alkyl C content of the sorbed NOM did not differ from the initial content. These findings are in agreement with already published 13C-NMR spectroscopic data on the sorption of the hydrophobic NOM fraction onto hydrous oxides (McKnight et al., 1992; Kaiser et al., 1997). The low retention of the hydrophilic NOM fraction may be caused by high proportions of weakly sorbing structural elements such as alkyl and O–alkyl C. The fact that no preferential sorption of carboxyl groups occurred within the hydrophilic NOM fraction and that the preferential sorption of hydrophobic, NOM carboxyl C was accompanied by a preferential sorption of aromatic C indicates that acidic aromatic structures account for the strong sorbing properties of the hydrophobic NOM.
Due to the complete desorption of hydrophilic NOM by 0.1 M NaH2PO4, the composition of the desorbed DOC did not differ from that of the sorbed material (Table 4). Desorption of hydrophobic NOM again led to a fractionation among structural elements. Alkyl and O–alkyl C were desorbed to a higher extent than carboxylic and aromatic C. It seems that molecules containing aromatic acid structures that form strong complexes on the goethite surface were more difficult to remove than molecules with higher contents of O–alkyl and alkyl C.
|
Summary and conclusions |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
The results agree well with the suggestion that the sorption of NOM to minerals involves the formation of strong complexes to surface metals by ligand exchange between acidic organic ligands and OH groups at the surface of the sorbent. Multisite binding, changes in the conformation, and mutual shielding of the NOM macromolecules may contribute to the low desorption of sorbed NOM. Thus, sorption favors storage and preservation of NOM in soils.
The sorption–desorption features of Al and Fe hydrous oxides are quite similar to those of the studied subsoil. We conclude, therefore, that these substances are a major control on NOM sorption and storage in forest soils.
|
ACKNOWLEDGMENTS |
|---|
Received for publication April 1, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. M. Uselman, R. G. Qualls, and J. Lilienfein Contribution of Root vs. Leaf Litter to Dissolved Organic Carbon Leaching through Soil Soil Sci. Soc. Am. J., August 9, 2007; 71(5): 1555 - 1563. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X.-L. Huang, Y. Chen, and M. Shenker Solid Phosphorus Phase in Aluminum- and Iron-Treated Biosolids J. Environ. Qual., March 1, 2007; 36(2): 549 - 556. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Mikutta, J. Kruger, F. Lang, and M. Kaupenjohann Acid Polysaccharide Coatings on Microporous Goethites: Controls of Slow Phosphate Sorption Soil Sci. Soc. Am. J., August 3, 2006; 70(5): 1547 - 1555. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Mikutta, F. Lang, and M. Kaupenjohann Soil Organic Matter Clogs Mineral Pores: Evidence from 1H-NMR and N2 Adsorption Soil Sci. Soc. Am. J., November 1, 2004; 68(6): 1853 - 1862. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. C. Arbestain, M. E. Barreal, and F. Macias Phosphate and Sulfate Sorption in Spodosols with Albic Horizon from Northern Spain Soil Sci. Soc. Am. J., March 1, 2002; 66(2): 464 - 473. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
