| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Environmental Sciences, Policy and Management, Univ. of California at Berkeley
b College of Chemistry, Univ. of California at Berkeley
c Stanford Magnetic Resonance Lab., Stanford Univ., Stanford, CA 94305-2115
* Corresponding author (bjcm{at}pangea.stanford.edu)
Received for publication June 2, 2000.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONS AND RECOMMENDATIONSREFERENCES |
|---|
Abbreviations: ICP, inductively coupled plasma • NMR, nuclear magnetic resonance
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONS AND RECOMMENDATIONSREFERENCES |
|---|
Because natural soil P concentrations are low, concentrated extracts are used for soil 31P NMR spectroscopy. There is no general agreement on the best extractant. Those in use include: 0.5 M NaOH (Newman and Tate, 1980; Hawkes et al., 1984); the cation exchange resin Chelex in water (Adams and Byrne 1989; Condron et al., 1996; Taranto et al., 2000); 0.5 M NaOH plus Chelex (Gressel et al., 1996; Cade-Menun and Preston 1996; Escudey et al., 1997); 0.1 to 0.5 M Bu4NOH (Emsley and Niazi, 1983); 0.5 M NaOH plus NaF (Sumann et al., 1998; Miltner et al., 1998); 0.5 M NaOH plus 0.1 M EDTA (Cade-Menun and Preston, 1996; Dai et al., 1996); and 0.25 M NaOH plus 0.05 M EDTA plus Chelex (Robinson et al., 1998). These extractants solubilize different concentrations of P and cations from soils (Cade-Menun and Preston, 1996). Chelex is used to remove paramagnetic ions from solution, which can improve spectra by reducing line broadening (Gressel et al., 1996). In contrast, NaOH-EDTA extracts high P concentrations, but Fe and Mn remain in solution complexed with EDTA (Cade-Menun and Preston, 1996; Dai et al., 1996). By solubilizing different metal concentrations, these extractants may have differing effects on T1 values for P forms in solution.
There has been only one published study of relaxation times for soil P (Newman and Tate, 1980), using 0.5 M NaOH as the extractant. This study reported T1 values of 2.0 to 3.0 s for orthophosphate and orthophosphate monoesters and diesters. A general rule is that delay times between scans should be five times T1 (Knicker and Nanny, 1997). However, delay times in use for soil 31P NMR spectroscopy include 20 s (Newman and Tate, 1980; Tate and Newman, 1982), 5 s (Condron et al., 1985), 2.2 s (Adams and Byrne, 1989), 2.0 s (Gressel et al., 1996), 1.5 s (Cade-Menun and Preston, 1996; Dai et al., 1996), 1.0 s (Zech et al., 1985), and 0.2 s (Guggenberger et al., 1996a,b; Sumann et al., 1998; Rubæk et al., 1999). With the use of different extractants, resulting in different paramagnetic ion concentrations, it is possible that some of these delay times may not be long enough for spectra to be quantitative.
Therefore, the objectives of this study were to: (i) measure spin-lattice (T1) relaxation times for a soil and a litter extracted with three commonly used extractants for soil 31P NMR spectroscopy; (ii) determine differences in concentrations of P and cations extracted by these methods; and (iii) examine relationships between cation concentrations and T1 values for the soil and litter extracts.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONS AND RECOMMENDATIONSREFERENCES |
|---|
|
For each extractant, 100 mL of liquid was used with either 10 g of soil or 6 g of litter. All samples were extracted in 125-mL Erlenmeyer flasks at room temperature overnight, with occasional stirring. Samples were then centrifuged at 15 000 rpm for 20 min. The Chelex samples also required filtration with Buchner funnels and Whatman (Maidstone, UK) 41 filter paper. A subsample of each was analyzed by inductively coupled plasma (ICP) spectroscopy (Thermo Jarrell Ash Corp., Franklin, MA) to determine the concentrations of P, Fe, Mn, Al, and Ca extracted by each method. The remainder of each sample was lyophilized.
To prepare the samples for NMR spectroscopy, 0.3 to 1.0 g of the lyophilized extract was weighed into a 50-mL plastic centrifuge tube with 2.6 mL of D2O and 0.4 mL of 10 M NaOH. The NaOH was added here to increase and standardize the pH, for optimal peak separation (Crouse et al., 2000). Samples were vortexed for 2 min and were then left to stand at room temperature for 2 h, with occasional vortexing. Samples were then centrifuged for 20 min at 15 000 rpm, and decanted into 10-mm NMR tubes. Samples were prepared no more than one hour prior to analysis by NMR spectroscopy. After NMR analysis, the contents of each NMR tube were analyzed by ICP spectroscopy for concentrations of P, Fe, Mn, Al, and Ca.
Nuclear Magnetic Resonance Spectroscopy
Phosphorus-31 NMR spectra were acquired at 202.45 MHz on a Bruker DRX-500 spectrometer running XWIN-NMR v2.1 and equipped with a 10-mm broadband observe probe. Samples were maintained slightly above room temperature (32°C) over the course of the experiments utilizing the BVT-3000 variable temperature unit.
All spectra were acquired with 32 768 total data points over a 20 325.20 Hz spectral window centered at 0 ppm (referenced to the phosphate singlet in an external standard of phosphoric acid in D2O). The P 90° pulse width was determined to be 16.5 ms on a trimethyl phosphate standard sample.
Saturation–Recovery Experiment
A saturation recovery sequence was used for the T1 determinations (saturation–recovery–delay–read pulse–acquisition). Saturation was achieved by an aperiodic sequence with a train of 90° pulses separated by decreasing delays. The initial delay was 1.024 s decreasing by a factor of two until 0.001 s was reached. The recovery delays utilized for the T1 determinations were 0.1, 0.2, 0.4, 0.8, 1.6, and 3.2 s, plus 6.4 s for Chelex-extracted samples. Composite 90° pulses were used throughout the saturation recovery sequence to compensate for any pulse width variability between samples. Data for T1 determination were typically acquired over a period of 12 to 18 h per sample depending on P content. Variations in instrument and sample stability were averaged over time by interleaving recovery delay acquisitions. This was done by cycling through the recovery delay list 7 to 14 times at 232 to 256 scans per free induction decay (FID). The FID data were stored in a serial file to be separated out and co-added prior to processing. Total number of scans collected per recovery delay varied from 1848 to 3584 scans depending on sample.
One-dimensional spectra were acquired (4000 scans, 35° read pulse, 1-s recycle delay) on the samples before and after the saturation recovery experiments to note sample changes which may have occurred over the extended runs.
Data Analysis
All spectral data analysis was done with XWIN-NMR (Bruker Analytik, Germany), v.2.1 and v.2.6. Interleaved FIDs were separated and co-added with standard automation programs. Spectra were processed with one times zero-filling and a line-broadening of 14 Hz. Peak-picking, peak-intensity determination, and spectral integration were performed with automated peak analysis tools. Peak integrals were further analyzed by visual inspection, especially in cases where baseline separation of adjacent peaks was not evident. Relaxation times were determined by fitting to peak intensity data with T1 analysis tools within XWIN-NMR.
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONS AND RECOMMENDATIONSREFERENCES |
|---|
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONS AND RECOMMENDATIONSREFERENCES |
|---|
The greatest diversity of peaks was seen in the litter sample extracted with Chelex; the least in the soil sample extracted with NaOH. A typical range of P forms was observed (Cade-Menun and Preston, 1996; Sumann et al., 1998), although no peaks for polyphosphate were visible. Polyphosphate presumably was not present in our litter or soil samples, because polyphosphate peaks have been seen in other studies using NaOH-EDTA as an extractant (Cade-Menun and Preston, 1996; Dai et al., 1996). The teichoic acid peak at 1.4 ppm was not present in the NaOH extract for either litter or soil, but was present in the EDTA and Chelex extracts. Because the rate of P recovery was highest in the NaOH extracts, and because a peak in this region was visible in other studies using NaOH as an extractant (Hawkes et al., 1984; Guggenberger et al., 1996a), the teichoic acid peak was probably not visible due to line broadening from the higher concentration paramagnetic ions, rather than non-extraction by NaOH. Reduced line broadening from the removal of paramagnetic ions would also explain the higher resolution of peaks in the orthophosphate monoester region (4.0–5.8 ppm) in the Chelex litter extracts. There does not, however, appear to be any reduction in line broadening by Chelex in the soil sample with respect to the EDTA sample. This is probably due to the smaller differences in Fe extracted from the soil sample by these methods, relative to the differences in Fe concentrations extracted from litter by the various methods.
In addition to the presence or absence of peaks seen with the different extractants, there is also variation in peak areas, which are used to determine the relative distribution of P within each P form. There were no obvious differences in P distribution within each class of P compounds among the extractants for the litter sample.
However, the spectrum for the soil Chelex extract (Fig. 1) shows a larger orthophosphate peak and smaller orthophosphate monoester peaks, relative to the spectra for NaOH or EDTA soil extracts. The calculated distribution of P within each P form (Fig. 3) also shows this. Additionally, the longest delay time, 6.4 s, was not long enough for the peak areas in the Chelex sample to have stabilized (Fig. 3). The apparent trend in Fig. 3 suggests that the Chelex spectrum would show more P as orthophosphate relative to spectra for the other two extractants, given a sufficiently long delay time. This clearly demonstrates the importance of adequate delay times, if the results obtained by 31P NMR spectroscopy on soil extracts are to be considered quantitative. The differences seen in P forms among the extractants may result from differences in the solubilization of P forms by each extractant. For example, the apparent lower concentration of orthophosphate monoesters in the soil Chelex extract relative to NaOH or EDTA may be a consequence of the lower percentage of total P extracted with the Chelex method. It has also been suggested that the removal of bridging metals such as Fe may cause changes in the relative distribution of P species (Shand et al., 1999), because P atoms close to Fe atoms may not be detectable by NMR (Bedrock et al., 1995). This might account for the apparently higher concentration of orthophosphate in the soil Chelex extract, which had lower Fe concentrations than did the NaOH or EDTA extracts.
The T1 values determined in this study are generally lower than the 2 to 3 s reported by Newman and Tate (1980). Possible explanations for this difference include: lower concentrations of paramagnetic ions in the samples prepared by Newman and Tate (1980); temperature differences during analysis, which will increase relaxation times (Ramarajan et al., 1981); and pH differences, which will affect T1 values in samples containing paramagnetic ions (Granot et al., 1979; McCain and Markley, 1980). The T1 values from our experiment were similar to values reported for pure P compounds with EDTA or Chelex at various Fe concentrations (Elgavish and Granot, 1979). There appears to be a weak relationship of T1 to metal concentrations within the NMR tube. Paramagnetic ions such as Fe3+, Mn2+, Cr3+, and Cu2+ would enable more efficient relaxation (Smernick and Oades, 1999), but Cr3+ and Cu2+ are not found in significant concentrations in the samples used in our study. Calcium was measured because it is strongly affected by the chelators EDTA and Chelex. The soil used in this study was an acid forest soil in which P adsorption is controlled more by Al and Fe than by Ca. In an alkaline soil, these chelators may increase the amount of P extracted by complexing the Ca from Ca-phosphates. However, because Ca is not paramagnetic, it is not expected to influence the NMR spectra in any way (Smernick and Oades, 1999), and no relationship was observed here. The apparent relationship of Al to T1 is not real, because Al is not a paramagnetic ion and cannot affect T1. Instead, this probably reflects the close link between P cycling and Fe and Al in this type of soil.
It may be possible to infer the relationship of these P species to metals, on the basis of changes in T1 for each P form in the various extractants (Elgavish and Granot, 1979). In the NaOH extracts, we may assume that any P forms that are complexed to metals (e.g. Fe or Al oxides) via ligand exchange remain complexed in a similar fashion to the original soil or litter sample. However, in the EDTA extracts, the metals are present but are chelated by EDTA (Owens et al., 2000), while in the Chelex extracts the metals were removed with the Chelex resin during filtration. Therefore, we can infer that the P species most associated with metals in the soil are those for which T1 increases from NaOH to EDTA to Chelex. That is, the time required for spin-lattice relaxation will increase as the P species are no longer as closely associated with the metals that are absorbing energy. This relationship can be seen for orthophosphate (6.3 ppm), some orthophosphate monoester peaks (5.8 and 4.7 ppm) and pyrophosphate (-4.0 ppm). These are the P species thought to be bound to metals in soil (Miltner et al., 1998, Celi et al., 1999). In contrast, the relationship of T1 versus extractants is not seen for orthophosphate diesters. These are probably adsorbed to metals by their organic moieties (Hawkes et al., 1984), and thus should be less intimately associated with any metals affected by these chelators. However, it should be stressed that only two samples were analyzed in this study, and that confirmation of these apparent relationships requires the analysis of a range of soil samples with differing P forms and a range of metal concentrations.
We conclude that changes in the sample can occur after 12 h in the magnet, and are due to the incubation of the sample at 32°C and are not due to the effects of the magnet itself, as suggested by Pant et al. (1999). Newman and Tate (1980) report the same changes occurring over weeks rather than hours. They suggest that this shift is consistent with hydrolysis of the ester linkage of phosphonolipid, and assign the peak at 19.8 ppm to alkyl phosphonic acid, and that at 18.3 ppm to phosphonolipid. Several other researchers report two peaks in this region, including Hawkes et al. (1984), Makarov (1998), and Carman et al. (2000). Others report only one peak (e.g. Bedrock et al., 1994; Makarov et al., 1997; Rubæk et al., 1999). Most of the experiments showing only one peak for phosphonates were run at 20 to 25°C, so temperature may be the controlling factor. Crouse et al. (2000) report that the optimal temperature for 31P NMR of poultry manure extracts is 20°C. However, they only looked at peaks in the orthophosphate monoester region.
|
CONCLUSIONS AND RECOMMENDATIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONS AND RECOMMENDATIONSREFERENCES |
|---|
Our general recommendations for other researchers studying soil P with 31P NMR spectroscopy are as follows: first, choose an extractant carefully, because it will influence the concentration of P extracted and the P forms visible with NMR spectroscopy. For our study, NaOH-EDTA seemed to be the best soil extractant: it extracted more P than Chelex, needed shorter delay times than Chelex, and had less line broadening than NaOH. However, for the litter sample, from which higher Fe concentrations were extracted, the Chelex extract had the least line broadening. Thus, while consistency of extractants would facilitate comparisons among studies, there may not be a single extractant suitable for all samples. Addition of EDTA to the Chelex-treated solution could further decrease line-broadening in samples with high Fe concentrations (Elgavish and Granot, 1979), but would require longer delay times. These longer delay times could be reduced by adding paramagnetic ions such as Cr3+ or relaxation agents such as chromium acetylacetonate (Nanny and Minear, 1994) to the sample. However, adding these agents may cause other changes, and would need to be carefully monitored. Treating EDTA extracts with Chelex, as Robinson et al. (1998) and Crouse et al. (2000) did, would probably have no effect, because EDTA has several orders of magnitude higher affinities toward metal ions than Chelex (Granot et al., 1979). The only metal ions which could be removed from the EDTA extracts by Chelex would be those not complexed by EDTA; this could be avoided by the use of a higher concentration of EDTA in the original extracting solution.
A second recommendation is to measure T1 values wherever possible, and to publish them with other NMR data. This is particularly aimed at research groups that regularly work with the same soil type and the same NMR parameters, and that have access to the blocks of machine time needed for T1 measurements. This would ensure that appropriate delay times are used. It would also allow research groups with more limited NMR access to predict the delay times needed for their analyses. Additionally, knowing T1 values for each P form in a variety of soils, especially if coupled with paramagnetic ion concentrations, could give valuable information on the relationship of P forms to paramagnetic ions.
Third, Fe and Mn should be measured in extracts prior to concentration or freeze–drying. This can be used to control the quality of spectra by controlling the final Fe and Mn concentrations in the NMR tube. The goal would be to maximize the P concentration in the tube without putting excess Fe and Mn in the tube. On the other hand, too little Fe and Mn will increase relaxation times, and thus will require longer delay times, or the addition of relaxation agents, as discussed above. We cannot as yet establish an ideal concentration for paramagnetic ions.
A fourth recommendation is to use appropriate delay times if the results from 31P NMR spectroscopy are to be considered quantitative. Our results suggest that delay times of 1 to 2 s should be adequate for quantitative analysis for all but soil Chelex extracts; none of the T1 values measured in our study were low enough to warrant the use of delay times as short as 0.2 s. However, this may not be true for other soil types. Using shorter pulse angles will also shorten the delay times needed.
Finally, we recommend acquiring the minimum scans necessary for adequate resolution, and keeping the temperature close to room temperature. This should help to prevent sample degradation.
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONS AND RECOMMENDATIONSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONS AND RECOMMENDATIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. A. Pasek Rethinking early Earth phosphorus geochemistry PNAS, January 22, 2008; 105(3): 853 - 858. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. He, T. Ohno, B. J. Cade-Menun, M. S. Erich, and C. W. Honeycutt Spectral and Chemical Characterization of Phosphates Associated with Humic Substances Soil Sci. Soc. Am. J., August 22, 2006; 70(5): 1741 - 1751. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. W. McDowell, I. Stewart, and B. J. Cade-Menun An Examination of Spin-Lattice Relaxation Times for Analysis of Soil and Manure Extracts by Liquid State Phosphorus-31 Nuclear Magnetic Resonance Spectroscopy J. Environ. Qual., January 5, 2006; 35(1): 293 - 302. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. S. Toor, B. J. Cade-Menun, and J. T. Sims Establishing a Linkage between Phosphorus Forms in Dairy Diets, Feces, and Manures J. Environ. Qual., July 5, 2005; 34(4): 1380 - 1391. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. O. Maguire, J. T. Sims, W. W. Saylor, B. L. Turner, R. Angel, and T. J. Applegate Influence of Phytase Addition to Poultry Diets on Phosphorus Forms and Solubility in Litters and Amended Soils J. Environ. Qual., November 1, 2004; 33(6): 2306 - 2316. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. C. Hansen, B. J. Cade-Menun, and D. G. Strawn Phosphorus Speciation in Manure-Amended Alkaline Soils J. Environ. Qual., July 1, 2004; 33(4): 1521 - 1527. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. H. Dao Ligands and Phytase Hydrolysis of Organic Phosphorus in Soils Amended with Dairy Manure Agron. J., July 1, 2004; 96(4): 1188 - 1195. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. L. Turner and A. E. Richardson Identification of scyllo-Inositol Phosphates in Soil by Solution Phosphorus-31 Nuclear Magnetic Resonance Spectroscopy Soil Sci. Soc. Am. J., May 1, 2004; 68(3): 802 - 808. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. L. Turner Optimizing Phosphorus Characterization in Animal Manures by Solution Phosphorus-31 Nuclear Magnetic Resonance Spectroscopy J. Environ. Qual., March 1, 2004; 33(2): 757 - 766. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. L. Turner, N. Mahieu, and L. M. Condron Phosphorus-31 Nuclear Magnetic Resonance Spectral Assignments of Phosphorus Compounds in Soil NaOH-EDTA Extracts Soil Sci. Soc. Am. J., March 1, 2003; 67(2): 497 - 510. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
