JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (6)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Smernik, R. J.
Right arrow Articles by Merrington, G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Smernik, R. J.
Right arrow Articles by Merrington, G.
Agricola
Right arrow Articles by Smernik, R. J.
Right arrow Articles by Merrington, G.
Related Collections
Right arrow Soil Organic Matter
Right arrow Other Environmental Contamination
Right arrow Municipal Waste
Journal of Environmental Quality 32:1523-1533 (2003)
© 2003 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America

TECHNICAL REPORTS
Waste Management

Advanced Solid-State Carbon-13 Nuclear Magnetic Resonance Spectroscopic Studies of Sewage Sludge Organic Matter

Detection of Organic "Domains"

Ronald J. Smernik*,a, Ian W. Olivera and Graham Merringtona,b

a Dep. of Soil and Water, Waite Agricultural Research Institute, Univ. of Adelaide, Glen Osmond, South Australia 5064, Australia
b Environment Agency, National Centre for Ecotoxicology & Hazardous Substances, Wallingford, Oxfordshire, OX10 8BD, UK

* Corresponding author (ronald.smernik{at}adelaide.edu.au)

Received for publication December 29, 2002.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Two novel solid-state 13C nuclear magnetic resonance (NMR) spectroscopic techniques, PSRE (proton spin relaxation editing) and RESTORE [Restoration of Spectra via TCH and T1{rho}H (T One Rho H) Editing], were used to provide detailed chemical characterization of the organic matter from six Australian sewage sludges. These methods were used to probe the submicrometer heterogeneity of sludge organic matter, and identify and quantify spatially distinct components. Analysis of the T1H relaxation behavior of the sludges indicated that each sludge contained two types of organic domains. Carbon-13 PSRE NMR subspectra were generated to determine the chemical nature of these domains. The rapidly relaxing component of each sludge was rich in protein and alkyl carbon, and was identified as dead bacterial material. The slowly relaxing component of each sludge was rich in carbohydrate and lignin structures, and was identified as partly degraded plant material. The bacterial domains were shown, using the RESTORE technique, to also have characteristically rapid T1{rho}H relaxation rates. This rapid T1{rho}H relaxation was identified as the main cause of underrepresentation of these domains in standard 13C cross polarization (CP) NMR spectra of sludges. The heterogeneous nature of sewage sludge organic matter has implications for land application of sewage sludge, since the two components are likely to have different capacities for sorbing organic and inorganic toxicants present in sewage sludge, and will decompose at different rates.

Abbreviations: BD, Bloch decay • CP, cross polarization • NMR, nuclear magnetic resonance • PSRE, proton spin relaxation editing • RESTORE, Restoration of Spectra via TCH and T1{rho}H (T One Rho H) Editing • T1H, proton spin-lattice relaxation rate in the static frame • T1{rho}H, proton spin-lattice relaxation rate in the rotating frame • TCH, rate of transfer of polarization from 1H to 13C nuclei during the contact time of the cross polarization experiment • VCT, variable contact time • VSL, variable spin lock


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE AMOUNT OF SEWAGE SLUDGE produced worldwide is increasing, and, with the phasing out of disposal of sewage to the ocean, the proportion applied to agricultural land is also increasing. The USA and the European Union will each produce more than 8 million Mg (dry weight) of sewage sludge by 2005, and more than 50% of this will be applied to agricultural land (Düring and Gäth, 2002). Addition of sewage sludge can improve soil fertility (Düring and Gäth, 2002, and references therein), as it is usually rich in nutrients, particularly nitrogen (Binder et al., 2002) and phosphorus (Hogan et al., 2001). The high content of organic matter in sewage sludge also can improve soil physical properties such as texture and water holding capacity (see, for example, Joshua et al., 1998). However, sewage sludge can contain high concentrations of potentially toxic organic chemicals (Jones and Sewart, 1997; Wang and Jones, 1994), heavy metals (Walter et al., 2002; Oudeh et al., 2002), and/or pathogens (Dumontet et al., 2001), which can adversely affect the health of crops, livestock, humans, and the environment.

Sewage sludge application to land has, until now, been regulated almost exclusively according to total contents of toxicants (Düring and Gäth, 2002). It has been suggested that better regulations would take into account the effects of sludge properties on the bioavailability of toxicants (McLaughlin et al., 2000). Such an approach would require a better understanding of how the sludge matrix holds toxicants in unavailable forms. The amount and the nature of organic matter in the sewage sludge are of crucial importance in this regard. Sorption to sludge organic matter has been implicated in the reduced bioavailability of both heavy metals (see, for example, Hooda and Alloway, 1994; McBride, 1995), and organic toxicants (see, for example, O'Connor, 1996). However, there is debate as to the permanence of this effect in the face of decomposition of the organic matter (Hooda and Alloway, 1994; McBride, 1995). The nature of sludge organic matter also influences the fertilizer properties of sewage sludge, as the release of mineral nitrogen from sewage sludge is a consequence of organic matter decomposition.

An improved understanding of the role of organic matter in controlling the long-term effects of sewage sludge applied to agricultural land requires the development and application of analytical techniques that can determine the chemical nature or "quality" of sludge organic matter: how it differs from native soil organic matter, the influence of the source and level of sewage treatment, and the changes that occur on decomposition in the field.

Solid-state 13C NMR spectroscopy is recognized as the best analytical method for determining the gross chemical composition of complex organic matrices (Kögel-Knabner, 2000). Solid-state 13C CPMAS (cross polarization with magic angle spinning) NMR spectroscopy has been used to characterize sewage sludge in a number of studies (Piotrowski et al., 1984; Leinweber et al., 1996; Ayuso et al., 1997; Hsiao and Lo, 2001; Stacey et al., 2001; Rowell et al., 2001). However, in a recent study we showed that the cross polarization (CP) technique substantially underestimates the proportion of alkyl (saturated) carbon in sludge organic matter (Smernik et al., 2003a). Furthermore, we showed that the alternative Bloch decay (BD) technique, otherwise known as direct polarization (DP) or single pulse excitation (SPE), does provide accurate quantitative distributions of structural units. Underestimation of alkyl carbon in CP spectra was attributed to the effect of molecular motion on reducing the efficiency of cross polarization (Hu et al., 2000; Smernik and Oades, 2000a,b; Preston, 2001). Underestimation of alkyl carbon has been reported for CP spectra of soil organic matter (Hu et al., 2000; Smernik and Oades, 2000a, b), but the degree of underestimation was much higher for the sludge organic matter (Smernik et al., 2003a).

In this study, we follow up our initial solid-state 13C NMR study of six Australian sewage sludges (Smernik et al., 2003a), by applying novel solid-state NMR techniques to probe the submicrometer heterogeneity of sewage sludge organic matter. The aim is to enhance our understanding of sludge organic matter structure beyond the "average" chemical structures provided by conventional NMR spectra. The concept of soil organic matter consisting of spatially distinct pools, each with varying physical properties, is well established. For example, modern carbon cycling models contain several organic "pools" that turnover at different rates on account of their different chemical composition (Falloon and Smith, 2000), and the existence of highly sorptive organic matter "domains" has been invoked to account for nonlinear aspects of the sorption of hydrophobic organic chemicals to organic matter (see, for example, Huang and Weber, 1997; Lueking et al., 2000). We also investigate the reasons for underestimation of alkyl carbon in 13C CP NMR spectra of sewage sludges.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Sample Collection and Preparation
The six sewage sludges were sourced from five Australian sewage treatment plants. Details of sludge sources and composition, including solid-state 13C CP (cross polarization) and BD (Bloch decay) NMR spectra, can be found in our earlier paper (Smernik et al., 2003a).

Air-dried sewage sludge samples were ground to <2 mm and were treated with hydrofluoric acid (HF) to isolate the organic matter, according to the method of Skjemstad et al. (1994). Briefly, this consisted of nine successive treatments with 2% hydrofluoric acid solution. Each treatment involved shaking 3-g portions of whole sludge in 50 mL of 2% hydrofluoric acid solution, end-over-end, for periods of 1 h (five times), 16 h (three times), and 64 h (once). Between treatments, samples were centrifuged and the supernatant was discarded and replaced with fresh 2% hydrofluoric acid solution. Following the final treatment, the residue was rinsed three times with deionized water, then freeze-dried. Carbon recoveries on HF treatment ranged between 79 and 87% (Smernik et al., 2003a). Note that great care must be taken when working with HF to ensure against contact with the skin or inhalation of vapors. Even relatively small exposures can result in serious injury or death.

Nuclear Magnetic Resonance Spectroscopy
Solid-state 13C magic angle spinning (MAS) NMR spectra were obtained at a 13C frequency of 50.3 MHz on a Varian (Palo Alto, CA) Unity 200 spectrometer. Samples were packed in a 7-mm-diameter cylindrical zirconia rotor with Kel-F end-caps and spun at 5000 ± 100 Hz in a Doty Scientific (Columbia, SC) MAS probe.

Inversion–recovery experiments (Smernik et al., 2000) were performed to determine the rate and uniformity of T1H relaxation, where T1H is the proton spin-lattice relaxation rate in the static frame. Thirteen recovery delays of between 0.1 ms and 1 s were used. Recycle delays of between 1 and 4 s were used and experiments were run in an interleaved fashion, with blocks of 32 scans acquired in turn, to a total of 1200 to 4000.

Variable contact time (VCT) and variable spin lock (VSL) experiments were performed as part of the RESTORE procedure (Smernik and Oades, 2003) for determining rates of T1{rho}H relaxation and TCH, where T1{rho}H is the proton spin-lattice relaxation rate in the rotating frame and TCH is the rate of transfer of polarization from 1H to 13C nuclei during the contact time of the cross polarization experiment. Variable contact time experiments consisted of an array of eight contact times (2, 2.5, 3, 4, 5, 6, 8, and 10 ms). The experiments were run in an interleaved fashion, with 32 scans acquired for each contact time, in turn. This was repeated until a total of 3000 to 4000 scans was acquired. A 1-s recycle delay was employed for all samples.

Variable spin lock experiments (otherwise known as delayed contact time experiments) were performed with three different contact times: 200 µs, 1 ms, and 2 ms. For the 200-µs contact time VSL experiments, 10 spin lock times were used (0, 0.3, 0.8, 1.3, 1.8, 2.3, 2.8, 3.8, 4.8, and 5.8 ms); for the 1-ms contact time VSL experiments, 10 spin lock times were used (0, 0.5, 1, 1.5, 2, 3, 4, 5, 7, and 9 ms); and for the 2-ms contact time VSL experiments, 8 spin lock times were used (0, 0.5, 1, 2, 3, 4, 6, and 8 ms). The VSL experiments were run in an interleaved fashion, with blocks of 32 scans acquired in turn, to a total of 3000 to 5000, with a 1-s recycle delay between scans.

Three spectra were acquired as input spectra for generating RESTORE subspectra: a 1-ms contact time with 0 spin lock spectrum, a 5- or 6-ms contact time with 0 spin lock spectrum, and a 1-ms contact time with 1-, 2- or 3-ms spin lock spectrum. These spectra were acquired in an interleaved fashion, with blocks of 32 scans acquired in turn, to a total of 10000 to 25000, with a 1-s recycle delay between scans.

Free induction decays were acquired with a sweep width of 40 kHz; 1216 data points were collected over an acquisition time of 15 ms. All inversion–recovery, VCT, VSL, and RESTORE spectra were zero-filled to 8192 data points and processed with a 50-Hz Lorentzian line broadening and a 0.010-s Gaussian broadening. Chemical shifts were externally referenced to the methyl resonance of hexamethylbenzene at 17.36 ppm. After linear baseline correction between 300 and -100 ppm, spectra were integrated between 300 and -10 ppm. These integrals make up the data sets used in the fitting procedures.

The inversion–recovery experiments were analyzed by statistically comparing one- and two-T1H component fits to the data, using the method of Smernik et al. (2000). Two T1{rho}H component fits were performed by the method of Smernik et al. (2002), using data from both the 200-µs and 1-ms contact time VSL experiments. RESTORE analysis of combined VSL and VCT experiments and generation of RESTORE subspectra were performed according to the method of Smernik and Oades (2003).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Proton Spin Relaxation Editing
Proton spin relaxation editing is a technique for determining the chemical nature of components of composite materials that are heterogeneous at the submicron scale. Proton spin relaxation editing has been described as "virtual fractionation" (Smernik et al., 2000) because it enables the generation of separate NMR spectra for each component, without the need for physical separation and isolation of these components. Proton spin relaxation editing has been used to identify components of wood (Newman and Hemmingson, 1990; Newman, 1992), soil organic matter (Preston and Newman, 1992; Preston and Newman, 1995; Smernik et al., 2000), and kerogen (Petsch et al., 2001). Proton spin relaxation editing relies on the different components having different characteristic 1H relaxation rates (either T1{rho}H or T1H). A process called spin diffusion homogenizes 1H relaxation rates so long as the 1H nuclei are in close contact. Thus, all 1H nuclei in small organic molecules generally have the same T1{rho}H and T1H relaxation rates. Spin diffusion is not efficient over longer scales, and materials that are heterogeneous may exhibit nonuniform rates of 1H relaxation. The smallest "domain size" that can be probed by PSRE is given by (DT)1/2, where D is the spin diffusion coefficient (typically 2 x 10-11 to 10-10 cm2/s) and T is the 1H relaxation rate (Zumbulyadis, 1983). Therefore, PSRE based on nonuniform T1{rho}H relaxation rates, which are typically in the range 0.5 to 5 ms for organic matter (Smernik and Oades, 2003), can be used to probe heterogeneity at scales of a few nanometers and above, whereas PSRE based on nonuniform T1H relaxation rates, which are typically in the range 10 to 200 ms for organic matter (Smernik et al., 2000), can be used to probe heterogeneity at scales of around 10 nm and above.

The uniformity of T1H relaxation in the sewage sludge organic matter was tested by comparing one- and two-component fits to inversion–recovery data using an F ratio test (Smernik et al., 2000). The one-component model assumes a single, homogeneous T1H relaxation rate, while the two-component model allows for varying proportions of two components characterized by separate, discrete values of T1H. The two-component fit was found to be superior for each of the six sludges at a confidence limit of >99.99%. The one- and two-component fits to the inversion–recovery data for Bolivar 95 sludge are shown in Fig. 1 as an example. The superiority of two-component fit to the data is clear. All of the other sludge samples showed similar nonuniform T1H behavior (data not shown).



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 1. Comparison of one-component and two-component fits to inversion–recovery data for Bolivar 95 sludge.

 
The results of one- and two-component fits are summarized in Table 1 . Average T1H values, as determined from the one-component fit, ranged from 74 ms (West Hornsby sludge) to 160 ms (Chelsea 96 sludge). These values are generally longer than those reported for a series of HF-treated soils (22.1–119 ms, Smernik et al., 2000) and soil humic acids (5.9–32.8, Smernik et al., unpublished data). The two-component fits to the inversion–recovery data provide two characteristic T1H values, T1Hfast and T1Hslow, and the proportion of carbon atoms associated with the rapidly and slowly relaxing pools of 1H nuclei (Table 1). The value of T1Hfast ranged between 44 ms (West Hornsby sludge) and 97 ms (Port Kembla sludge). The value of T1Hslow ranged between 228 ms (West Hornsby) and 385 ms (Port Kembla). Again, these values are generally longer than those reported for HF-treated soils (Smernik et al., 2000) and soil humic acids (Smernik et al., unpublished data). The ratio of rapidly to slowly relaxing components ranged from 35:65 for Werribee 97 to 66:34 for West Hornsby. Four of the sludges (Bolivar 95, Bolivar 97, West Hornsby, and Port Kembla) contained more rapidly relaxing component than slowly relaxing component, whereas the other two sludges (Chelsea 96 and Werribee 97) contained less rapidly relaxing component than slowly relaxing component.


View this table:
[in this window]
[in a new window]
  		Table 1. Results of one- and two-component fits to inversion–recovery data for hydrofluoric acid (HF)–treated sewage sludges.

 
Nonuniformity in T1H relaxation shows that sewage sludge is heterogeneous, and the characteristic T1H values give a lower limit for the size of the separate component domains of around 10 nm. The real value of PSRE is in identifying the nature of the rapidly and slowly relaxing components via the generation of PSRE subspectra. In this way the relaxation information is related to the chemical shift or frequency information usually associated with NMR spectroscopy. The principle of spectral editing on the basis of differential proton spin relaxation rates is most easily understood by considering the recovery delays that produce a null signal. From Fig. 2 , it is clear that the rapidly relaxing component of Bolivar 95 (T1H = 65 ms) produces no signal at a recovery delay of around 35 ms. Therefore, the NMR spectrum obtained with this recovery delay will contain signal only from the slowly relaxing component (although the spectrum will appear inverted). Similarly, at a recovery delay of around 180 ms, no signal will be produced by the slowly relaxing component, and the NMR spectrum obtained with this recovery delay will only contain signal from the rapidly relaxing component. Component subspectra can be obtained in this way (Zumbulyadis, 1983), but it is in fact possible to generate subspectra for the rapidly and slowly relaxing components from linear combinations of any two inversion–recovery spectra. This can be achieved either through a trial-and-error process (Preston and Newman, 1992), or, if the characteristic T1H values (i.e., T1Hfast and T1Hslow) are known, as is the case here, the appropriate linear combinations can be calculated (Smernik et al., 2000). The latter method is nonsubjective, and on this basis would appear to be superior.



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 2. Inversion–recovery curves for rapidly relaxing (T1H = 65 ms) and slowly relaxing (T1H = 338 ms) components of Bolivar 95 sludge.

 
Proton spin relaxation editing subspectra for the sewage sludge organic matter samples were generated using the method of Smernik et al. (2000), and are presented in Fig. 3 . Two features are immediately evident. First, the differences between the rapidly and slowly relaxing component subspectra are large, and second, the differences between the sludges are small. This implies that sewage sludge organic matter is composed of two very different, spatially distinct components, and that the nature of these components varies little among a range of sludges derived from five different treatment works that employ different treatment processes and service different catchments. Differences in the "bulk chemistry" of the sludges, evidenced by variation in their 13C CP and BD NMR spectra (Smernik et al., 2003a), are therefore the result of differing proportions of the two components identified here.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 3. Solid-state 13C (T1H) proton spin relaxation editing (PSRE) nuclear magnetic resonance (NMR) subspectra of hydrofluoric acid (HF)–treated sewage sludges.

 
The rapidly relaxing component (Fig. 3) is dominated by alkyl carbon (0–45 ppm), but also contains a substantial quantity of carbonyl carbon (170–180 ppm), and a prominent broad resonance at 50 to 60 ppm. The slowly relaxing subspectra (Fig. 3) generally contain sharper resonances. The slowly relaxing component is rich in carbohydrate carbon, with strong, sharp resonances in the range 60 to 90 ppm (C2–C6 carbons of sugar residues) and also at 100 to 105 ppm (anomeric C1 sugar carbon).

The features evident in Fig. 3 are confirmed in Table 2 , which shows the distribution of signal in the PSRE subspectra between four chemical shift regions, 190 to 165 ppm (carbonyl), 165 to 110 ppm (aromatic), 110 to 45 ppm (O-alkyl), and 45 to 0 ppm (alkyl). Carbonyl carbon accounts for roughly twice as much of the NMR signal in the rapidly relaxing subspectra (12.0–13.5%) as it does in the slowly relaxing subspectra (3.8–7.5%). There is also slightly more aromatic carbon in the rapidly than the slowly relaxing subspectrum, for all samples except Werribee 97, which is also the most aromatic of the sludges overall. The much higher carbohydrate content of the slowly relaxing component is reflected in its higher proportion of O-alkyl carbon (53.8–58.4% versus 27.6–36.2% for the rapidly relaxing component). Conversely, the rapidly relaxing component contains nearly twice as much alkyl carbon (36.4–42.5%) than does the slowly relaxing component (22.7–23.5%).


View this table:
[in this window]
[in a new window]
  		Table 2. Percent of total nuclear magnetic resonance (NMR) signal contained in four chemical shift regions in 13C proton spin relaxation editing (PSRE) NMR spectra of hydrofluoric acid (HF)–treated sludges.

 
The rapidly relaxing PSRE subspectra of the sewage sludge organic matter samples closely resemble the 13C CP NMR spectrum of bacterial culture reported by Baldock et al. (1990), and shown in Fig. 4a . Bacterial cell membranes, which consist mainly of glycerophospholipids, account for the highly alkyl nature of bacterial biomass and residues. Bacteria are also rich in protein, which would account for the strong carbonyl resonance, as well as the broad resonance at 50 to 60 ppm, which can be assigned to carbon adjacent to the amino N in amino acid residues of protein. Much of the aromatic carbon can be attributed to aromatic amino acids.



View larger version (13K):
[in this window]
[in a new window]
  		Fig. 4. Solid-state 13C cross polarization (CP) nuclear magnetic resonance (NMR) spectra of (a) bacterial biomass (reproduced from Baldock et al., 1990, by permission of CSIRO Publishing), and (b) light (<1 Mg m-3) fraction of soil organic matter (reproduced from Smernik and Oades, 2000a, with permission from Elsevier Science).

 
The slowly relaxing PSRE subspectra of the sewage sludge organic matter samples appear similar to the 13C CP NMR spectra of slightly decomposed plant material, such as the soil light (<1 Mg m-3) fraction, shown in Fig. 4b (Smernik and Oades, 2000a). The 13C CP spectra of forest floor litter (see, for example, Preston and Trofymow, 2000) also exhibit similar features to those of the slowly relaxing PSRE subspectra shown in Fig. 3. The strong and sharp O-alkyl resonances indicate the presence of cellulose. The aromatic regions of the slowly relaxing PSRE subspectra generally contain prominent resonances at around 150 ppm, which can be assigned to O-substituted aromatic carbon. This is indicative of the presence of lignin or structures derived from lignin, and is again consistent with a plant source for this component.

Previous applications of the PSRE technique to soil organic matter detected components quite different to those detected here for sewage sludge organic matter. Preston and Newman (1992)(1995) detected an alkyl-rich slowly relaxing component in soil humins. Similar alkyl-rich slowly relaxing components were reported by Newman and Condron (1995) for dairy pond sludge, Smernik et al. (unpublished data) for soil humic acids, and Smernik and Oades (1999) for an HF-treated soil treated with Cu2+. Smernik et al. (2000), in a study utilizing PSRE on eight HF-treated soils, found that the rapidly relaxing component contained most of the aromatic carbon, especially for soil with high charcoal contents, and that the slowly relaxing component contained most of the O-alkyl and alkyl carbon. In all of these cases, the main factor controlling relaxation rates was the concentration of free radicals: components with higher concentrations of free radicals (organic free radicals or paramagnetic metal ions) relaxed more rapidly. All of these reports are in contrast to the findings here, where the rapidly relaxing PSRE component is alkyl-rich and unlikely to harbor high concentrations of organic free radicals or paramagnetic cations. For the sludges it would appear that molecular mobility controls relaxation rates; the rapid relaxation of the alkyl-rich component indicating a higher degree of molecular mobility. This demonstrates that the alkyl carbon in sewage sludge is quite different in nature to the alkyl carbon in soils, a conclusion supported by the anomalously low CP observability of sewage sludge alkyl carbon reported in our earlier paper (Smernik et al., 2003a).

RESTORE Technique
The RESTORE [Restoration of Spectra via TCH and T1{rho}H (T One Rho H) Editing] technique was developed to improve the quantitative reliability of the cross polarization (CP) technique for acquiring solid state 13C NMR spectra of soil organic matter and related materials. It is closely related to PSRE, the main difference being that spectral editing is based not only on differential proton spin relaxation rates (in this case, T1{rho}H), but also on rates of cross polarization (TCH). RESTORE generates subspectra of sample components, but because there is no equivalent process to spin diffusion that homogenizes TCH, not all RESTORE components represent "domains" that are spatially distinct.

Quantitation in CP spectra is compromised by two processes: (i) rapid 1H relaxation during the contact time and (ii) slow rates of magnetization transfer from 1H to 13C nuclei for carbons remote from nearest hydrogen neighbors and in regions of high molecular mobility. The dependence of CP signal intensity (I) on rates of cross polarization (1/TCH) and 1H relaxation (1/T1{rho}H) can be approximated by Eq. [1] (Smernik and Oades, 2003):

[1]
where {alpha} = 1 - (TCH/T1{rho}H) and t is the contact time.

Cross polarization spectra of soil organic matter are generally acquired with a contact time of around 1 ms. At this contact time most protonated (typically TCH = approximately 0.05 ms) and nonprotonated (typically TCH = approximately 0.35 ms) carbons produce very similar amounts of CP signal, so long as T1{rho}H is >2.5 ms (Smernik and Oades, 2000a). However, as can be seen in Fig. 5 , carbons associated with rapidly relaxing 1H nuclei (T1{rho}H = 1 ms), or that cross polarize slowly (TCH = 4 ms), produce much less signal at a contact time of 1 ms, and are therefore routinely underestimated by CP. The RESTORE technique employs variable spin lock (VSL) and variable contact time (VCT) experiments to identify these underestimated pools of carbon, and to produce subspectra of these carbon pools from a linear combination of three spectra with different spin lock and contact times. The subspectra can then be combined, with each component scaled according to its specific observability, to produce a corrected or "restored" CP spectrum. For example, for the system shown in Fig. 5, RESTORE would scale up the rapidly relaxing (T1{rho}H = 1 ms) component and the slowly cross polarizing (TCH = 4 ms) component by factors of 1.7 and 4.1, respectively, relative to the other component.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 5. Plots of signal intensity versus contact time for various combinations of TCH and T1{rho}H, calculated from Eq. [1].

 
The results of the RESTORE fitting procedure are shown in Table 3 for the six sewage sludges. The RESTORE model consists of three components, characterized by different combinations of TCH and T1{rho}H. Component SS (CSS) has a short TCH and a short T1{rho}H, component SL (CSL) has the same short TCH as component SS, but a long T1{rho}H, and component LL (CLL) has a long TCH and a long T1{rho}H (equal to T1{rho}H for component SL). The two T1{rho}H values, T1{rho}H (short) and T1{rho}H (long), are determined from a two-component fit to data from variable spin lock (VSL) experiments (see Materials and Methods). Simultaneous fits to data from VSL and variable contact time (VCT) experiments determine the values of TCH (long) (TCH for component LL), and the relative proportions of the three components.


View this table:
[in this window]
[in a new window]
  		Table 3. Results of RESTORE [Restoration of Spectra via TCH and T1{rho}H (T One Rho H) Editing] fit for hydrofluoric acid (HF)–treated sewage sludges.

 
Each sample exhibited nonuniform T1{rho}H relaxation, with T1{rho}H for the rapidly relaxing component, T1{rho}H (short), ranging between 0.62 ms and 1.28 ms, and T1{rho}H for the slowly relaxing component, T1{rho}H (long), ranging between 3.86 ms and 5.23 ms. Three of the sludges exhibited TCH (long) values of 5 ms, the maximum value allowed in the RESTORE analysis (Smernik and Oades, 2003), while TCH (long) was in the range 1.01 to 2.66 ms for the other three sludges. For each sludge, the largest proportion of 13C nuclei was in component SL (45–69%). This is the RESTORE component that is most easily observed in CP spectra, as indicated by the larger specific observabilities (0.81–0.85) compared with those of components SS and LL. A substantial proportion (26–44%) of 13C nuclei was in component SS. This RESTORE component is underrepresented in CP spectra due to rapid T1{rho}H relaxation during the contact time. Specific observabilities for component SS ranged between 0.28 and 0.53. A much smaller proportion (5–11%) of 13C nuclei was in component LL. This RESTORE component is underrepresented in CP spectra due to slow cross polarization. The specific observability for component LL varied greatly between the samples, reflecting the large variation in TCH (long). For those samples where TCH (long) was 5 ms, the specific observability was low (0.16), signifying that the CP spectra of these samples severely underrepresent component LL. For the other three samples, the specific observability for component LL was much higher (0.28–0.56), and for these samples component LL is not as severely underestimated in the CP spectra.

As with PSRE, the most useful feature of the RESTORE technique is the generation of component subspectra. RESTORE component 13C NMR subspectra of the six sewage sludge organic matter samples are shown in Fig. 6 . The corresponding chemical shift range distributions are in shown in Table 4 . As was the case for the PSRE subspectra, the RESTORE subspectra are very different for each sample, but differ little between sludge samples.



View larger version (30K):
[in this window]
[in a new window]
  		Fig. 6. Solid-state 13C RESTORE [Restoration of Spectra via TCH and T1{rho}H (T One Rho H) Editing] nuclear magnetic resonance (NMR) subspectra of hydrofluoric acid (HF)–treated sewage sludges.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Percent of 13C nuclei contained in four chemical shift regions for 13C RESTORE [Restoration of Spectra via TCH and T1{rho}H (T One Rho H) Editing] nuclear magnetic resonance (NMR) subspectra of hydrofluoric acid (HF)–treated sewage sludges.{dagger}

 
The subspectra for component SS (Fig. 6) are very similar in appearance to the rapidly relaxing PSRE subspectra (Fig. 3), with alkyl carbon accounting for the largest proportion of signal (37.6–50.6%, Table 4). The proportions of carbonyl and aromatic carbon are also very similar for corresponding rapidly relaxing PSRE subspectra (Table 2) and RESTORE component SS subspectra. The similarity of these two sets of subspectra indicates that the pool of 1H nuclei that have short T1H relaxation rates also have short T1{rho}H relaxation rates. As discussed above, this rapidly relaxing component closely resembles bacterial biomass. The cause of rapid T1{rho}H relaxation is again most likely to be relatively rapid molecular motion.

The subspectra for component SL (Fig. 6) are also quite similar in appearance to the slowly relaxing PSRE subspectra (Fig. 3). In both sets of subspectra, O-alkyl carbon provides the largest proportion of signal (45.0–57.1% in SL spectra compared with 53.8–58.4% for slowly relaxing PSRE subspectra). The SL subspectra (Fig. 6, Table 4) generally contain more alkyl and carbonyl carbon than corresponding slowly relaxing PSRE subspectra (Fig. 3, Table 2).

Of the three different RESTORE subspectra, component LL exhibits the greatest differences between the sludges. Nonetheless, there are some features common to all of the LL subspectra. The alkyl region is prominent and distinctive for each LL subspectrum, with three narrow resonances at around 30, 23, and 15 ppm. The resonance at 15 ppm can be assigned to methyl groups, while that at 23 ppm is most likely to be that of methylene carbons adjacent to methyl groups, or the methyl carbon of acetyl esters. It would thus appear that the ends of alkyl chains have characteristically slow rates of polarization. These regions would be expected to have the highest degrees of molecular mobility. However, the main cause of underrepresentation of alkyl carbon in the CP spectra of the sewage sludge organic matter is rapid T1{rho}H relaxation, since component SS, which is also alkyl rich, represents a much greater proportion of carbon than does LL. The LL subspectra are richer in aromatic carbon, especially O-aryl carbon (145–165 ppm), than either of the other two RESTORE components. The cause of slow cross polarization of aromatic carbon is most likely remoteness from nearest 1H neighbors, and not high molecular mobility.

One further aspect of the RESTORE subspectra deserves comment: the presence of small inverted resonances in most of the component SS subspectra. The model of CP spin dynamics described by the RESTORE technique is a simplified one, in that there are only three distinct pools of 13C nuclei, each characterized by discrete values of TCH and T1{rho}H. The inverted resonances expose the limitations of this model and indicate that there are some 13C nuclei with T1{rho}H values outside the range T1{rho}H (short) to T1{rho}H (long) (Smernik et al., 2000). In particular, inverted resonances in the rapidly relaxing subspectrum (SS) indicate the presence of a sludge component with T1{rho}H > T1{rho}H (long). Such a component was detected in the Port Kembla sludge through the acquisition and analysis of an additional 1-ms contact time VSL experiment, in which the maximum spin lock was increased from 10 to 25 ms. Employing a relatively long (1 ms) contact time diminishes signal from rapidly relaxing components, and including longer spin lock times emphasizes slowly relaxing components. A two-T1{rho}H fit to data from this VSL experiment yielded values of T1{rho}H (short) and T1{rho}H (long) of 2.85 and 8.82 ms, respectively. Proton spin relaxation editing subspectra were generated based on these T1{rho}H values and are shown in Fig. 7 . The rapidly relaxing subspectrum (Fig. 7) has features in common with both SS and SL RESTORE subspectra. The slowly relaxing subspectrum (Fig. 7) contains very little signal outside the O-alkyl region and is very similar to that reported by Newman and Hemmingson (1990) for crystalline cellulose. This slowly relaxing component represents 24% of total NMR signal for a 1-ms contact time CP spectrum. However, given the bias of CP against more rapidly relaxing components at this contact time, this component constitutes less than 8% of total carbon. Note that the inverted resonances evident in the SS subspectra (Fig. 6) correspond exactly with the resonances of the slowly relaxing component in Fig. 7. The sharp O-alkyl resonances in the LL subspectra (Fig. 6) are also artifacts, a further consequence of the presence of crystalline cellulose. Pichler et al. (2000), using PSRE based on T1H relaxation rates, found evidence for slowly relaxing crystalline cellulose in municipal solid waste. The slowly relaxing component in their samples also appeared to contain plastic. We find no evidence for the presence of plastic in our sewage sludges.



View larger version (17K):
[in this window]
[in a new window]
  		Fig. 7. Solid-state 13C T1{rho}H proton spin relaxation editing (PSRE) nuclear magnetic resonance (NMR) subspectra for Port Kembla sewage sludge.

 
Carbon-13 RESTORE NMR spectra, that is, the weighted sum of the three RESTORE component subspectra, are shown in Fig. 8 . The RESTORE spectra exhibit signal distributions intermediate between those of the corresponding CP and BD spectra (Smernik et al., 2003a), which differed substantially from each other. We reported that around 30% of NMR signal was "missing" from CP spectra, based on spin counting measurements. Clearly, RESTORE has partially corrected the biases inherent in the CP technique, producing spectra more closely resembling the quantitative BD spectra.



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 8. Solid-state 13C RESTORE [Restoration of Spectra via TCH and T1{rho}H (T One Rho H) Editing] nuclear magnetic resonance (NMR) spectra (weighted average of RESTORE subspectra) of hydrofluoric acid (HF)–treated sewage sludges.

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Sewage sludge organic matter was found to have a heterogeneous structure at the submicrometer scale, as evidenced by nonuniform rates of 1H NMR relaxation (T1H and T1{rho}H). Proton spin relaxation editing (PSRE) was used to generate 13C NMR subspectra of sludge components characterized by different T1H values. The rapidly relaxing component of each sludge sample was rich in lipids and protein, and appeared to have a similar composition to bacterial biomass. The slowly relaxing component of each sample was rich in cellulose and appeared to have a similar composition to partly degraded plant material. A related spectral editing technique, RESTORE, was used to generate 13C NMR subspectra of sludge components characterized by different combinations of T1{rho}H relaxation rate and cross polarization rate (TCH). The RESTORE subspectra characterized by rapid rates of T1{rho}H relaxation (component SS) were very similar to the rapidly relaxing PSRE subspectra, while subspectra for another RESTORE component, SL, were similar to the slowly relaxing PSRE subspectra. The third RESTORE component (LL), characterized by slow rates of cross polarization, represented the smallest proportion of total carbon and had a distinctive alkyl region containing three sharp resonances.

The cause of poor quantitation in 13C CP NMR spectra of sewage sludge organic matter, noted in an earlier study (Smernik et al., 2003a), is clear from the RESTORE results. The rapid T1{rho}H relaxation rate of RESTORE component SS means that a large proportion of signal for this component decays away during the 1-ms contact time usually used for 13C CP NMR spectra. The underrepresentation of component SS results in CP spectra that contain much less alkyl carbon signal than the quantitative Bloch decay (BD) spectra (Smernik et al., 2003a).


   ACKNOWLEDGMENTS
 
This work was, in part, funded by an Australian Research Council (ARC) grant.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2003 32: 1167-1172. [Full Text]  



This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
R. J. Smernik
A New Way to Use Solid-State Carbon-13 Nuclear Magnetic Resonance Spectroscopy to Study the Sorption of Organic Compounds to Soil Organic Matter
J. Environ. Qual., June 7, 2005; 34(4): 1194 - 1204.
[Abstract] [Full Text] [PDF]

Home page
 J. Environ. Qual.Home page
R. J. Smernik, I. W. Oliver, and M. J. McLaughlin
Changes in the Nature of Sewage Sludge Organic Matter During a Twenty-One-Month Incubation
J. Environ. Qual., September 1, 2004; 33(5): 1924 - 1929.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (6)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Smernik, R. J.
Right arrow Articles by Merrington, G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Smernik, R. J.
Right arrow Articles by Merrington, G.
Agricola
Right arrow Articles by Smernik, R. J.
Right arrow Articles by Merrington, G.
Related Collections
Right arrow Soil Organic Matter
Right arrow Other Environmental Contamination
Right arrow Municipal Waste

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS