Poly(methylene) Crystallites in Humic Substances Detected by

The NMR spectra of various different humic substances show a sharp peak at a 13C ...... Michal Shechter, Baoshan Xing, Frank-Dieter Kopinke, and Benny...
0 downloads 0 Views 95KB Size
Download PDF
Environ. Sci. Technol. 2000, 34, 530-534

Poly(methylene) Crystallites in Humic Substances Detected by Nuclear Magnetic Resonance

characterization of poly(methylene) crystallites in various humic substances by variable-temperature magic-angle spinning and spin diffusion NMR experiments and by wideangle X-ray scattering. Implications of these crystallites and the surrounding mobile amorphous poly(methylene) regions for soil properties are briefly discussed.

WEI-GUO HU,† JINGDONG MAO,‡ BAOSHAN XING,‡ AND K L A U S S C H M I D T - R O H R * ,† Polymer Science and Engineering Department and Plant and Soil Science Department, University of Massachusetts, Amherst, Massachusetts 01003

Experimental Section

Crystalline domains composed of poly(methylene) chains have been detected by solid-state NMR and wide-angle X-ray scattering (WAXS) in several samples of soil organic matter, including humins, surface soil (peat), and humic acids extracted from surface soil and young coal. From the melting range of 60->80 °C and 1H spin diffusion experiments, a crystallite thickness of ca. 3 nm or 25 CH2 units is deduced. The overall fraction of (CH2)n carbons in these materials is up to 9%. Nearly half of this poly(methylene) is crystalline, while the rest is noncrystalline and more isotropically mobile. In humin, several crystalline and noncrystalline poly(methylene) domains form larger aggregates. The crystallites are expected to be resistant to environmental attack and thus inert in the soil and have long residence times, while the mobile amorphous regions may play a role in the sorption of nonpolar molecules in soil.

Introduction Soil organic matter (SOM) represents a major component of the world’s surface carbon reserves. It exerts a profound influence on many aspects of the nature of soil and on many environmental processes, such as fertility, ion-exchange capacity, water-holding capacity, and sorption of metals and organics. The structures of individual SOM fractions regulate their reactivity, property, and functions but are poorly understood. A better understanding of SOM structures, particularly of humic substances, would help to determine their origin and genesis, reactivity, and roles in environmental processes. Significant fractions of SOM are poorly soluble or insoluble, being mainly constituted of macromolecules. Therefore, solid-state nuclear magnetic resonance (NMR) is the method of choice for investigating their chemical and physicochemical structures. Modern solid-state NMR can provide information on composition, segmental dynamics, domain sizes, and local ordering of macromolecules (1, 2). However, because of the highly diverse and irregular chemical structure of SOM, solid-state NMR of SOM has been limited mainly to composition characterization (3-5). So far, SOM has generally been considered to be amorphous. In this paper, we describe the identification and * Corresponding author e-mail: [email protected]; phone: (413)577-1417; fax: (515)294-0105. Current address: Chemistry Department, Iowa State University, Ames, IA 50011. † Polymer Science and Engineering Department. ‡ Plant and Soil Science Department. 530

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 3, 2000

To represent a range of different origins and processing procedures of SOM, the following samples were chosen: (i) a humic acid and a humin extracted from a peat at Amherst, MA (AMH-HA and AMH-humin); (ii) an International Humic Substance Society reference material, Florida Pahokee peat (FLA-peat), and a humin extracted from it (FLA-humin); (iii) a commercial humic acid purchased from Aldrich, extracted from brown coal (ALD-HA). As references to the SOM samples, high-density polyethylene (HDPE), (CH2)n (∼70% crystalline), and an ethylene/ vinyl acetate copolymer with 9% vinyl acetate by weight (∼35% crystalline) were also measured. The NMR experiments were performed on a Bruker MSL300 spectrometer at a 1H frequency of 300.13 MHz and a 13C frequency of 75.47 MHz. Four kinds of experiments were conducted: (i) cross-polarization (CP) from 1H to 13C; (ii) direct polarization (DP) by a 13C 90° pulse; (iii) a GoldmanShen 1H spin-diffusion experiment (6) with 13C detection, consisting of a 30-µs 1H T2 filter period flanked by 90° pulses, a mixing time tm for spin diffusion, followed by a CP pulse sequence (7); the magnetization was stored along (z in alternate scans to reduce effects of 1H T1 relaxation (1, 8, 9). The same pulse sequence but without the 1H T2 filter time was also used to measure 1H T1 relaxation during tm; (iv) a 1H inversion recovery experiment (180°-pulse-t recov-90°pulse) probing 1H T1 relaxation, with 13C detection using total suppression of sidebands (TOSS). All experiments were performed under magic angle spinning (MAS), with proton decoupling during detection. The spinning speed was 3 kHz in the variable-temperature experiments, 5 kHz for the spin diffusion and inversion-recovery, and 4 kHz otherwise. At these spinning speeds, the side-band intensity of (CH2)n carbons is less than 5% of the center band, and side bands of other major components are much less than the (CH2)n signal. The CP time was 0.8 ms (0.1 ms in the spin-diffusion experiment), the recycle delay was 1 s, and the 1H decoupling power was 70-90 kHz. Simple CP spectra were typically taken in 2-4 h, the DP spectrum of ALD-HA was taken in 4 h, and that of AMH-HA was taken in 12 h. The wide-angle X-ray scattering (WAXS) experiments were conducted on a Siemens D500 diffractometer. The source of the radiation is the Cu KR line, with a wavelength of 1.5418 Å. For AMH-humin, the intensity was averaged for 51 h, and for ALD-HA, it was averaged for 22 h.

Results and Discussion NMR Identification of Rigid all-trans-Poly(methylene) Domains. Figure 1a shows the CP/MAS 13C NMR spectrum of AMH-humin. The signal spreads over a wide range and can be assigned to aliphatic, aromatic, and other chemical groups (3-5). There is a sharp peak at 32.9 ppm, which is shown expanded in Figure 1b. In the DP/MAS spectrum (lower traces), the 32.9 ppm peak is relatively lower and the 31 ppm peak becomes prominent. The 32.9 and 31 ppm peaks are assigned to rigid all-trans and mobile gauchecontaining (CH2)n units (poly(m)ethylene-like chains), respectively. This assignment is possible by comparing with 10.1021/es990506l CCC: $19.00

 2000 American Chemical Society Published on Web 12/29/1999

FIGURE 2. CP/MAS 13C spectra of FLA-humin (top) and ALD-HA (bottom), showing the characteristic peak-and-shoulder pattern of semicrystalline (CH2)n near 32 ppm. Spinning speed, 4 kHz.

FIGURE 1. (a) MAS 13C NMR spectra of AMH-humin: (top) crosspolarization (CP); (bottom) direct polarization (DP) with a recycle delay of 2 s. (b) Expanded methylene region of the spectra in panel a. (c) CP and DP (recycle delay ) 2 s) 13C spectra of polyethylene (PE) with 3 mol % of vinyl acetate comonomer. (d) CP and DP (recycle delay ) 10 s) 13C spectra of high-density PE. Spinning speed, 4 kHz. the CP and DP spectra of polyethylene (PE) samples, as shown in Figure 1, panels c and d. The spectrum of Figure 1c is that of a material with 3 mol % of polar vinyl acetate comonomers, while Figure 1d is the signal of a pure high-density polyethylene (HDPE) material. In all the spectra, the sharp crystalline signal near 32.8 ppm and the broader amorphous signal at 31 ppm can be clearly distinguished. Even though the amorphous and crystalline regions have a very similar chemical structure, they exhibit slightly different chemical shifts due to their different conformations. In the crystalline state, (CH2)n chains are all-trans, while in the amorphous state, the chains have both trans and gauche conformations. Due to the γ-gauche effect (10), the crystalline and amorphous signals appear at different chemical shifts. The DP spectra in Figure 1, panels c and d, were acquired with recycle delays of 2 and 10 s, respectively, which are short as compared to 13C T relaxation times of the crystalline component (>100 1 s). The crystalline signal is underrepresented due to the incomplete relaxation, while the amorphous signal is shown much better since its 13C T1 is much shorter than 10 s as a result of chain mobility. The CP and DP spectra of AMH-humin (Figure 1a,b) display a similar feature: in the DP spectrum the all-trans signal is suppressed while the gauche-containing signal is better represented. This indicates the nearly immobile, crystalline nature of the 32.9 ppm signal. We have also found the all-trans and mobile gauchecontaining poly(methylene) components in other humic substances. Figure 2 shows the CP spectra of FLA-humin and ALD-HA. In the latter, the sharp peak at 32.9 ppm and the shoulder centered at 31 ppm are remarkable. The signals of the other chemical groups, which are hardly visible here, can be made clearly observable by line broadening; they show the characteristics of an old humic acid (5). Figure 3, panels a and b, show the CP and DP spectra of AMH-HA. Similar to Figure 1, panels a and c, in the DP spectrum of AMH-HA the all-trans signal at 32.9 ppm is suppressed. It is clear that the fraction of the all-trans component in AMHHA is much smaller than in AMH-humin.

FIGURE 3. (a) CP/MAS and (b) DP/MAS (recycle delay ) 2 s) 13C spectra of AMH-HA. In the CP spectrum, the crystalline signal at 33 ppm is clearly seen (indicated by the arrow). In the DP spectrum, the crystalline signal is suppressed because of incomplete 13C T1 relaxation during the short recycle delay. A quantification of the poly(methylene) components is achieved through DP with fully relaxed all-trans signal, which shows that at 20 °C the all-trans and the amorphous (CH2)n in AMH-humin together make up (7 ( 2)% of the total material, with (3 ( 1)% being crystalline. The percentage of (CH2)n in ALD-HA is about (9 ( 2)%, (4 ( 1)% being all-trans. In FLA-humin, the total (CH2)n fraction is (6 ( 2)%, and (2.5 ( 1)% are all-trans. In the peat humic acid samples, the (CH2)n fractions are smaller. Signal intensity at 30-33 ppm in various SOM samples and its assignment to (CH2)n groups has been documented in the literature (4, 11-16). It has also been reported that part of it has a short 1H-1H dipolar dephasing time and long 13 C T1 (15, 17). However, the semicrystalline nature of this component was not recognized since the characteristic twopeak structure was not resolved in the spectra. It is easily made invisible by the line broadenings introduced by short acquisition times, data smoothing, or weak decoupling. The peak at 32.9 ppm typically merges with its surrounding if the acquisition time is less than 10 ms at B0 ) 7 T, or less than 30 ms at B0 ) 2.3 T, or if the time data are multiplied with a corresponding window function to smooth the spectrum. Investigating forest soil organic matter by solid-state NMR, Ko¨gel-Knabner et al. (15) found that the (CH2)n signal in the neighborhood of 30 ppm can be decomposed into more rigid and more mobile parts and that the rigid fraction increases with depth of extraction. In view of our results, it must be expected that the more rigid part described in ref 15 is crystalline; since it is more resistant to biodegradation, it has a longer residence time than the amorphous chains. The high concentration of (CH2)n crystallites in the humic acid from brown coal (ALD-HA) supports this view. Wide-Angle X-ray Scattering. To further confirm the crystalline nature of the 32.9 ppm peak, wide-angle X-ray VOL. 34, NO. 3, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

531

FIGURE 4. Wide-angle X-ray scattering patterns of AMH-humin, ALD-HA, and HDPE. Reflections characteristic of crystalline orthorhombic PE are seen in all three traces. The small peaks at 2θ ) 9° and 19.7° in AMH-humin scattering are from unidentified crystals. scattering (WAXS) was performed. Figure 4 shows the WAXS patterns of ALD-HA, AMH-humin, and HDPE (as a reference). The strongest Bragg reflections of orthorhombic PE are indexed as (110), with a lattice spacing of 4.11 Å, and as (200), with a spacing of 3.71 Å, which appear at 2θ ) 21.6° and 24.0°, respectively, for Cu KR radiation (18). AMH-humin shows a small (110) reflection at 2θ ) 21.6°, and ALD-HA shows a more intense (110) reflection as well as a (200) reflection at 2θ ) 24.0°. Due to the small quantity and finite size (see below) of the crystallites, some reflections will be too wide to be distinguished from the background. Therefore, the reflection intensities do not necessarily reflect the quantity shown by NMR. These scattering experiments confirm that crystalline PE-like domains exist in the samples studied and have the common orthorhombic crystal modification. Visser et al. (19) have reported evidence for hexagonal crystalline domains in microbial humic acids, but these are different from the orthorhombic crystallites found here, and their chemical structure is not clear. Schnitzer et al. (20) have performed systematic X-ray studies of various humic acids. They found that one of three broad bands observed correlates with the aliphatic content. This, however, is not the signal of poly(methylene) crystallites reported here; it is nearly an order of magnitude wider than the Bragg peaks that we have observed. The broad band actually corresponds to the “amorphous halo” observed for noncrystalline PE and other hydrocarbons, which reflects the “liquidlike” short-range order. Comparison with our X-ray scattering patterns (Figure 4) shows that the signal-to-noise ratio in the diffractograms of ref 20 is insufficient to detect the small Bragg reflections of the poly(methylene) crystallites in humic acids. Size of the Crystallites. The melting point Tm of a PE crystallite is a function of its thickness d, (Tm - Tm∞) ∼ 1/d (Gibbs-Thomson relation), where Tm∞ (∼145 °C) is the melting point for infinitely thick crystals (21). For example, the melting point is ∼135 °C for common HDPE (thickness ∼20 nm). Therefore, the thickness of a crystallite can be estimated by its melting point. Figure 5 shows the temperature dependence of the (CH2)n NMR signals in AMH-humin and ALD-HA as compared with those of the ethylene/vinyl acetate copolymer. As the temperature increases, the alltrans peaks in all three samples become smaller, indicating crystallite melting. The signals of the gauche-containing chains become sharper, indicating higher mobility. This also explains why the intensity of the amorphous-phase signals does not increase more strongly as the conformations change from trans to gauche: the cross-polarization efficiency is reduced by the higher mobility. 532

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 3, 2000

FIGURE 5. Variable-temperature cross-polarization 13C NMR spectra of (a) AMH-humin, (b) ALD-HA, and (c) P(E/VAc) copolymer. The decrease of the left peak with increasing temperature is due to crystal melting.

FIGURE 6. Variable-temperature cross-polarization 13C NMR spectra of FLA-peat. Melting of the crystallites is clearly observed. In the copolymer, the vinyl acetate side groups disrupt crystallization of (CH2)n units, so the crystallites are only ∼4 nm thick, as determined by small-angle X-ray scattering; as a result, the melting point is depressed to ∼90 °C, Figure 5c. In Figure 5a, the crystallites in AMH-humin are found to have a melting point of ∼75 °C, and those of ALD-HA have an average melting point of 70 °C. From comparison with Figure 5c, with the melting temperatures of long-chain paraffins (Table VIII.4 in ref 21), and with the melting point depression according to the Gibbs-Thomson equation in Figure VIII.10 in ref 21, this corresponds to a thickness of ∼3 nm. If, as is usual in PE and n-alkanes, the shortest dimension of the crystal is approximately along the chain axis, 3 nm corresponds to ∼25-30 CH2 units (depending on the tilt of the chains relative to the crystal surface). If the crystals were nearly cubic blocks, the sidelength would be larger than 3 nm. The crystallites in FLA-peat show a similar melting behavior, see Figure 6: again, the crystalline signal decreases and the noncrystalline band increases at high temperatures. The melting of the crystallites in these SOM samples is proof of the crystalline nature of the 32.9 ppm peak. The actual crystallite thickness may deviate from the above estimate since end groups or chemical defects in the (CH2)n chains may shift the melting point. To estimate the crystal

FIGURE 7. (a) 13C NMR detected 1H spin diffusion in AMH-humin, after a 30-µs 1H T2 filter selecting the mobile components. The diffusion times tm are indicated; the top spectrum is the regular CP spectrum without selection. The fact that the signal of the amorphous regions survives the 1H T2 filter proves that they possess significant segmental mobility. The crystalline signal reappears due to spin diffusion from the mobile amorphous regions. (b) 13C NMR detected 1H inversion recovery, after the indicated recovery delays. At 70 ms, only the poly(methylene) remains inverted, showing that 1H spin diffusion from the other, faster-relaxing protons is slow. Spinning speed, 5 kHz. size in a different way, a Goldman-Shen 1H spin diffusion experiment was performed on AMH-humin, see Figure 7. After selection of 1H magnetization in the most mobile regions using a 30-µs 1H T2 filter, diffusion of the magnetization into the surrounding regions occurs during a variable mixing time tm and is observed by the reappearance of the signal of the rigid regions. The smaller the domain sizes, the faster the reequilibration process. For better structural resolution, the 1H magnetization distribution is detected indirectly on 13C, after 1H-13C cross polarization. With longer tm, the increase of the crystalline peak height and the decrease of the amorphous signal are evident in Figure 7. The equilibrium is reached within 20-50 ms, which corresponds to a crystallite thickness of ∼4 nm (7). Similar results were obtained for ALD-HA (not shown). The simultaneous decrease of the amorphous and increase of the crystalline poly(methylene) signal show that the mobile amorphous and all-trans crystalline domains are in close proximity, forming larger poly(methylene) regions; this is confirmed by their long 1H T1 relaxation time, which is distinct from that of the rest of the material (see next section). In thin PE crystallites, fast chain flips occur (22, 23) that produce an increased MAS line width, decreased 1H line width (24), and shorter 13C T1, 13C T2, and 1H T1F relaxation times (25, 26). In addition, some dynamic disorder may be present that has similar spectral effects (7). In Figures 1 and 2, the crystalline signal (∆ν ∼120 Hz) is not as sharp as that of HDPE (∆ν ∼25 Hz). The 13C T1 of the ALD-HA crystalline peak (32.8 ppm) is ∼5 s [defined by the time of signal decay to 1/e of its initial height in a CP/T1 experiment (8)], which is much shorter than in HDPE samples (26). The T1F,Η of 5 ms is also shorter than that of HDPE (23). These short relaxation times support the notion of thin (CH2)n crystallites in SOM. Large Poly(methylene) Regions. Strong evidence that several crystalline and amorphous (CH2)n domains are aggregated in larger poly(methylene) regions is provided by the 1H T1 relaxation times in AMH-humin. T1 values of 200 and 170 ms were found for the crystalline and amorphous poly(methylene), respectively, as compared to 110 ( 10 ms for all other residues. Figure 7b demonstrates this difference in 13C-detected 1H inversion-recovery spectra, where a

recovery delay of 70 ms produces a pure poly(methylene) spectrum, while the other signals are passing through zero. This means that on a time scale of ∼200 ms, 1H spin diffusion does not equilibrate the magnetization of the poly(methylene) with the rest of the material. By comparing with the spin diffusion into the crystallites (Figure 7a), the diameter of a poly(methylene) region is estimated to be >3 times the crystallite thickness, i.e., it exceeds 12 nm. Such poly(methylene) domains appear to be a general feature of humins and soils (27, 28). Evidence for Crystallinity. The notion of a crystalline component in soil organic matter may be difficult to accept. Therefore, we will briefly discuss the strong evidence gathered in this paper and refute potential alternatives. The NMR spectra of various different humic substances show a sharp peak at a 13C chemical shift of 32.9 ppm, which is characteristic of all-trans (CH2)n segments packed similar to the common orthorhombic crystal modification of PE. The line width is unusually small as compared to that of other signals in the spectra of SOM, indicating a relatively high degree of uniformity in terms of composition, conformation, and packing as well as large distances from paramagnetic species. Relaxation times show that the all-trans segments are much less mobile than their gauche-containing counterparts. The all-trans signal disappears relatively suddenly around 75 °C, which can only be interpreted as melting. We have also documented the corresponding crystalline peaks in wideangle X-ray diffraction, with strong broadening due to the small crystallite size (3-4 nm), fully consistent with the NMR results. Our NMR data exclude a model of extended all-trans chains with a simple thermal activation of higher energy conformers. If the activation of gauche conformers occurred continuously, the sharp 13C line would shift gradually to the right (10) without changing much in intensity. We do not observe this but rather find that the rigid trans signal disappears, over a relatively narrow temperature range, exactly as expected for the melting of thin crystals. If the crystal disintegrated with increased temperature but the chain conformation remained mostly trans, the signal would broaden but remain centered near 33 ppm, contrary to our observation of vanishing signal in that region at high temperature. Models with extended mostly trans-poly(methylene) chains dispersed in an amorphous environment are not only inconsistent with the spin diffusion and the scattering data but also thermodynamically untenable: Restricting the chain conformation to nearly exclusively all-trans entails a high entropic penalty T∆S, which can be overcome only by the enthalpy reduction ∆H due to the favorable intermolecular interactions in a sufficiently large aggregate, or bundle, of parallel chains. From crystal nucleation theory, it is known that in order for the Gibbs free energy (∆G ) ∆H - T∆S) of the aggregate to be lower (more stable) than that of the supercooled melt, at least 20 chains must come together in the bundle (21). For such a large bundle, the term crystallite seems appropriate. While ordered single chains in an amorphous environment are highly unlikely, inside a polymer crystal significant degrees of disorder and dynamics are often observed: Helical jumps, conformational disorder that is often dynamic, and even crystalline polymer “rotator phases” are known (29). This supports our view that the all-trans aggregates observed in SOM, even if they are partly disordered, are nevertheless crystalline. Possible Origin and Length of Poly(methylene) Chains and Relation with Soil Properties. The (CH2)n component may derive from aliphatic biopolymers such as cutan or suberan in the protective layers of higher terrestrial plants or from similar polymers in algal cell walls (13, 30-32). VOL. 34, NO. 3, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

533

Despite their initial low concentrations in organisms, the high preservation potential of such structures will result in their selective enrichment during diagenesis (4, 11-16), which is thought to be related to the formation of kerogen (11, 12). We propose that this resistance to degradation is at least partly due to the semicrystalline nature of the poly(methylene). Being inaccessible to chemicals (33) or enzymes, crystalline domains in polymers are much harder to biodegrade than their amorphous counterparts. For example, in poly(-caprolactone), [-(CH2)5-COO-]n, the amorphous regions are degraded prior to the crystallites (34). The (CH2)n fraction in AMH-HA is smaller than in AMH-humin (16), which indicates that only a small portion of the (CH2)n chains in SOM is bonded to polar groups that make them soluble. The length of the poly(methylene) chains is uncertain at this point. It must exceed the minimum of 25 CH2 units found to span a crystal. Linear aliphatic chains, including n-alkanes, up to C101 have been detected in humic substances by pyrolysis-soft ionization mass spectrometry (35). The large poly(methylene) domains found in our study would be consistent with long chains traversing several crystallites and amorphous regions, with branching in the amorphous regions. On the other hand, models of cutan and suberan suggest structures containing (CH2)nCH3 chains with n g 35 attached to a polysaccharide or aromatic core (31, 32). The presence of the poly(methylene) crystallites may contribute specific properties to soil. Since crystalline regions are much less permissive to small molecules than the amorphous material, they must have very different sorption capacities. On the other hand, the mobile amorphous poly(methylene) regions may provide sorption sites for nonpolar molecules; in fact, humins have higher sorption affinities for dichlorobenzenes and naphthalene than do the corresponding humic acids (36), which have smaller poly(methylene) fractions. The crystallites may also influence swelling and solubility properties. This is consistent with the enrichment of the (CH2)n components in the insoluble humin fraction of the AMH-peat material. In an environment where the soil temperature is high, the physical properties of the soil may change because many of the crystallites have already melted at 70 °C. Due to its exceptional biological stability, the crystalline component might serve as an “internal standard” of the evolution of SOM. In short, the semicrystalline poly(methylene) component could bring a new perspective to the understanding of soil organic matter.

Acknowledgments B.X. thanks the U.S. Department of Agriculture and National Research Initiative Grants Program (97-35102-4201 and 9835107-6319) for financial support.

Literature Cited (1) Schmidt-Rohr, K.; Spiess, H. W. Multidimensional Solid-State NMR and Polymers; Academic Press: London, 1994. (2) Schmidt-Rohr, K.; Hu, W.; Zumbulyadis, N. Science 1998, 280, 714. (3) Preston, C. M. Soil Sci. 1996, 161, 144. (4) Baldock, J. A.; Oades, J. M.; Waters, A. G.; Peng, X.; Vassallo, A. M.; Wilson, M. A. Biogeochemistry 1992, 16, 1.

534

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 3, 2000

(5) Mao, J.; Hu, W.-G.; Schmidt-Rohr, K.; Davies, G.; Ghabbour, E. A.; Xing, B. Soil Sci. Soc. Am. J., accepted. (6) Goldman, M.; Shen, L. Phys. Rev. 1966, 144, 321. (7) Clauss, J.; Schmidt-Rohr, K.; Adam, A.; Boeffel, C.; Spiess, H. W. Macromolecules 1992, 25, 5208. (8) Torchia, D. A. J. Magn. Reson. 1978, 30, 613. (9) Clauss, J.; Schmidt-Rohr, K.; Spiess, H. W. Acta Polym. 1993, 44, 1. (10) Tonelli, A. E. NMR Spectroscopy and Polymer Microstructure: the Conformational Connection. In Methods in Stereochemical Analysis; Marchand, A. P., Ed.; VCH: New York, 1989. (11) Hatcher, P. G.; Spiker, E. C.; Szeverenyi, N. M.; Maciel, G. E. Nature 1983, 305, 498. (12) Tegelaar, E. W.; De Leeuw, J. W.; Derenne, S.; Largeau, C. Geochim. Cosmochim. Acta 1989, 53, 3103. (13) Tegelaar, E. W.; de Leeuw, J. W.; Saiz-Jimenez, C. Sci. Total Environ. 1989, 81/82, 1. (14) Tegelaar, E. W.; Kerp, H.; Visscher, H.; Schenck, P. A.; de Leeuw, J. W. Paleobiology 1991, 17, 133. (15) Ko¨gel-Knabner, I.; Hatcher, P. G.; Tegelaar, E. W.; de Leeuw, J. W. Sci. Total Environ. 1992, 113, 89. (16) Almendros, G.; Guadalix, M. E.; Gonzalez-Vila, F. J.; Martin, F. Org. Geochem. 1996, 24, 651. (17) Newman, R. H.; Condron, L. M. Solid State Nucl. Magn. Reson. 1995, 4, 259. (18) Brandrup, J., Immergut, E. H., Eds. Polymer Handbook; John Wiley: New York, 1989. (19) Visser, S. A.; Mendel, H. Soil Biol. Biochem. 1971, 3, 259. (20) Schnitzer, M.; Kodama, H.; Ripmeester, J. A. Soil Sci. Soc. Am. J. 1991, 55, 745. (21) Wunderlich, B. Macromolecular Physics; Academic Press: New York, 1980; Vol. 3. (22) Mansfield, M.; Boyd, R. H. J. Polym. Sci.: Polym. Phys. Ed. 1978, 16, 1227. (23) Hu, W.-G.; Boeffel, C.; Schmidt-Rohr, K. Macromolecules 1999, 32, 1611. (24) Olf, H. G.; Peterlin, A. J. Polym. Sci.: A-2 1970, 8, 753. (25) McCall, D. W. Acc. Chem. Res. 1971, 4, 223. (26) Axelson, D. E.; Mandelkern, L.; Popli, R.; Mathieu, P. J. Polym. Sci.: Polym. Phys. Ed. 1983, 21, 2319. (27) Preston, C. M.; Newman, R. H. Can. J. Soil. Sci. 1992, 72, 13. (28) Smernik, R. J.; Oades, J. M. Geoderma 1999, 89, 219. (29) Hu, W.-G.; Schmidt-Rohr, K. Acta Polym. 1999, 50, 271. (30) Baker, E. A. In The Plant Cuticle; Cutler, D. F., Alvin, K. L., Price, C. E., Eds.; Academic Press: New York, 1982; p 139. (31) Tegelaar, E. W.; Leeuw, J. W. D.; Largeau, C.; Derenne, S.; Schulten, H. R.; Mu ¨ ller, R.; Boon, J. J.; Nip, M.; Sprenkels, J. C. M. J. Anal. Appl. Pyrol. 1989, 15, 29. (32) McKinney, D. E.; Bortiatynski, J. M.; Carson, D. M.; Clifford, D. J.; de Leeuw, J. W.; Hatcher, P. G. Org. Geochem 1996, 24, 641. (33) Blundell, D. J.; Keller, A.; Ward, I. M.; Grant, I. J. J. Polym. Sci., Polym. Lett. Ed. 1966, 4, 781. (34) Huang, S. J. In Encyclopedia of Polymer Science and Engineering; Mark, H. F., Bikales, N. M., Overberger, C. G., Menges, G., Eds.; John Wiley: New York, 1985; Vol. 2, p 220. (35) Schnitzer, M.; Schulten, H.-R. Sci. Total Environ. 1989, 81/82, 19. (36) Xing, B. in Humic Substance: Structure, Properties and Uses; Davies, G., Ghabbour, E. A., Eds.; Royal Societ of Chemistry: Cambridge, 1998; p 173.

Received for review May 3, 1999. Revised manuscript received November 1, 1999. Accepted November 12, 1999. ES990506L