| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Earth Sciences, University of Leeds, Leeds LS2 9JT, United Kingdom
b Department of Plant and Soil Sciences, University of Delaware, 152 Townsend Hall, Newark, DE 19711
* Corresponding author (s.hunger{at}earth.leeds.ac.uk)
Received for publication March 4, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CP, cross-polarization • MAS, magic angle spinning • NMR, nuclear magnetic resonance • PL, poultry litter
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Considering the pathways in which P enters the aquatic environment, part of the negative effects of the overapplication of animal manures can be remedied by lowering the water-soluble concentration of P and thus preventing loss of P by runoff. A promising approach is the addition of chemical amendments, such as lime, ferric chloride, or alum [Al2(SO4)3 · 18H2O] (Moore and Miller, 1994; Moore et al., 1995a; Shreve et al., 1995). More recently, Dou et al. (2003) reported the effectiveness of alum and coal combustion by-products in reducing soluble P in dairy, swine, and broiler manure.
Alum has received particular interest in the treatment of PL for its effectiveness in reducing the amount of water-soluble P, while at the same time reducing the litter pH. A lower litter pH has been shown to prevent the loss of N by ammonia volatilization, which otherwise diminishes air quality as well as the quality of the litter as fertilizer (Moore et al., 2000; Shreve et al., 1995). However, amending PL with alum does not reduce the total concentrations of P. Chemical amendments therefore do not prevent the accumulation of P in soils but merely reduce the possibility of loss by runoff and leaching. To our knowledge, no information is available on the long-term stability of P in chemically amended PL.
Although the effectiveness of alum as an amendment for PL has been proven on a macroscopic scale in poultry houses and in the field, the reactions involved have been speculative. Initially, it was assumed that phosphate in animal manures reacts with Al3+ to form solid Al phosphate (AlPO4) (Moore and Miller, 1994), which is insoluble in the pH range typical for agricultural soils (Lindsay, 1979). Although this assumption has been recognized as oversimplified [e.g., Duffy and van Loon (1995)], little, if any, information about the mechanisms of P fixation in PL is available. This is mainly due to the lack of methods sufficiently sensitive to directly determine the species, or chemical form of P, in heterogeneous materials such as PL and soils.
Advances in spectroscopic techniques have recently permitted the spectroscopic investigation of P speciation in poultry litter and increased our understanding of P chemistry in PL. An X-ray absorption near edge structure (XANES) spectroscopic study of PL showed the presence of a calcium phosphate phase in the unamended PL samples, whereas features in the XANES spectra of the alum-amended PL samples indicated the formation of a P species similar to phosphate adsorbed to Al hydroxide (Peak et al., 2002). These results were confirmed and complemented by solid-state 31P NMR spectroscopy (Hunger et al., 2004). In addition to the Ca- and Al-bound phosphate phases identified by XANES, a labile, physically sorbed phase was detected in both amended and unamended samples. The calcium phase detected by XANES, which was identified by 31P NMR as either an amorphous tribasic calcium phosphate or a surface precipitate on calcium carbonate, was also found in the alum-amended samples. Using 31P{Al} TRAPDOR (TRAnsfer of Polarization during DOuble Resonance), a technique sensitive for Al and P nuclei in close proximity to each other, the presence of direct bonds between Al and phosphate could be confirmed (Hunger et al., 2004). The chemical shift range of the Al-bound phosphate species indicated a mixture of phosphate bound to Al hydroxide and an amorphous, uncondensed aluminum hydroxyl-phosphate phase, possibly a poorly ordered wavellite [Al3(OH)3(PO4)2·5 H2O]. Both P-XANES and 31P NMR confirmed that P in PL is present as orthophosphate (PO4), rather than chains containing two or more phosphate groups or P species in a lower oxidation state.
These methods are not without limitations, however. Both P-XANES and 31P NMR spectroscopy failed to detect or quantify organic P species, mainly due to the fact that positive identification requires a unique chemical environment around the phosphate group, which is not given in a heterogeneous matrix such as poultry litter. This limitation is especially severe because up to 60% of the total P extractable by 0.5 M NaOH and 50 mM EDTA has been identified as phytic acid (myo-inositol hexakisphosphate) by solution 31P NMR spectroscopy (Turner, 2004). Furthermore, paramagnetic cations [e.g., Fe(II), Fe(III), or Mn(II)] in close proximity to phosphate groups distort the local magnetic field and cause a broadening of the NMR signal by facilitating the relaxation of the P nucleus (Blumberg, 1960; Smernik and Oades, 2002). This paramagnetic effect leads to signal loss, rendering phosphate groups in association with paramagnetic cations undetectable.
Nevertheless, spectroscopic techniques offer a much more detailed insight into P speciation than has been possible with established methods, such as sequential chemical extraction. Sequential chemical extraction schemes have been used to characterize P species in animal waste materials and soils by separating them into operationally defined groups according to their solubility in a series of extractants (Condron et al., 1985; Dou et al., 2000, 2002; Hedley et al., 1982; Leinweber et al., 1997; Sharpley and Moyer, 2000). Dou et al. (2000) developed a very thorough extraction sequence from the one first proposed by Hedley et al. (1982). Manure samples are exhaustively extracted using first H2O and NaHCO3 to assess the labile species, followed by NaOH to extract Fe- and Al-bound species and occluded organic phosphate species, and finally HCl to dissolve calcium phosphate phases.
These assignments are, however, operationally defined and no information about chemical speciation is available from a sequential extraction. A very promising approach is to combine extraction techniques with spectroscopic techniques to verify assumptions about assignments made on a molecular level from macroscopic information. Frossard et al. (1994) characterized urban sewage sludge by sequential chemical fractionation with consecutive solid-state 31P NMR analysis of the residual solids. They found that a complex mixture of phosphate solids was present in the sludge, including several calcium phosphates and Al phosphate. Corresponding to the proposed sequence of dissolution, the calcium phosphate phases were dissolved during extraction with HCl. The Al phosphate phase, which was tentatively assigned to a poorly ordered wavellite, was present in the residual P fraction, although it was expected to dissolve during extraction with NaOH. It was reasoned that the poorly ordered wavellite reprecipitated during the extraction.
Sequential chemical extraction is a well-established method to characterize animal manures and estimate the risk of P loss after field application. Consequently, an extensive body of data exists in the literature. We report here results from an investigation combining sequential chemical extraction of alum-amended poultry litter with solid-state 31P NMR spectroscopic analysis of the solid material before extraction and of the residues after each extraction step. Findings from both techniques are compared to verify assumptions made with regard to P species from chemical extraction.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Extractants were prepared by dissolving reagent-grade NaHCO3 and NaOH in deionized water (>16 M
). Standardized 1.0 M HCl was used as received. The chemical extraction sequence was adapted from Hedley et al. (1982). The extractants were in their order of application: deionized water, 0.5 M NaHCO3, 0.1 M NaOH, and 1.0 M HCl. A standard extraction procedure was established as follows: 0.6 g of freeze-dried sample was shaken with 30 mL of extractant for 20 h (solid to solution ratio 1:50). The solids were separated from the supernatant by centrifugation (15 min, 15000 rpm) and decanting. The supernatant was passed through 0.2-µm Supor filter paper (Gelman Laboratories, Ann Arbor, MI) and stored at 4°C until analysis. The solid to solution ratio employed in this work was higher than that published for manure analysis by Hedley et al. (1982) (1:60), Dou et al. (2000) (1:100), Sharpley and Moyer (2000), or Frossard et al. (1994) (both 1:200), but in accordance with the ratio recommended by Kuo (1996) for soils. This ratio was chosen to allow for the larger sample amount required for solid-state NMR analysis. Higher solid to solution ratios, however, can cause incomplete extraction due to saturation of the extractant with respect to the solid phase to be dissolved.
Concentrations of P, Al, Ca, and Fe in the extracts were determined by inductively coupled plasma atomic emission spectroscopy (ICP–AES). The total concentrations of these elements were measured after digestion of a dried, unextracted sample with concentrated HNO3 and 30% H2O2 for 24 h at 50°C. A total number of four subsamples of each of the coarse and fine size separate of sample PL 181 were extracted. After each extraction step, one of the four samples was freeze-dried and subsequently used for 31P NMR analysis. Elemental concentrations are therefore average values of four samples for the first (deionized water), three for the second (0.5 M NaHCO3), and two for the third extraction step (1.0 M HCl). The residual amounts were calculated as the difference between the total concentrations determined by digestion and the sum of the extracted amounts. Total concentrations of P, Ca, Al, and Fe, and the pH of sample PL 181 are presented in Table 1. All concentrations are calculated as g/kg of dried litter.
|
The CP–MAS experiments used a proton 
/2-pulse of 3.5 µs, a contact time of 1.3 ms, and a pulse delay of 2 s. Depending on the P concentration, 2000 to 4000 scans were accumulated to give a signal to noise ratio of the most intense peak of at least 30:1. Single-pulse, proton-decoupled spectra were recorded using a 31P /2-pulse of 3.7 µs and a relaxation delay of 60 s. For these experiments up to 512 scans were accumulated.
The spectra were processed and analyzed using the NUTS NMR utility transform software (Acorn NMR, 2000). A line broadening of 50 to 100 Hz was applied before Fourier transformation and phase correction. Isotropic chemical shifts are all reported in ppm relative to the signal of 85% H3PO4 as an external reference; positive values correspond to low-field or high-frequency shifts.
The spectra were deconvoluted using a minimal set of peaks as described by Hunger et al. (2004). The Lorentzian/Gaussian ratio was fixed at a value of 1 (Lorentzian peak) for the relatively sharp peak at 
= 3.0 ppm and allowed to float for all other peaks. No further constraints were employed. An error margin of ±5% of the combined signal intensity of all signals is assumed for the peak areas of the deconvoluted peaks.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The strong enrichment of Al in the fine size separate is probably caused by the manner in which alum is applied. Alum was spread out in the poultry houses on top of existing layers of litter and bedding material (wood chips and saw dust) and incorporated into the upper few inches of the remaining litter by a spiked-tooth harrow (Sims and Luka-McCafferty, 2002). Further mixing was achieved by the birds' movements. No attempt was made to accomplish complete homogenization or dissolution; alum and its hydrolysis product [Al(OH)3] therefore remained in the litter as small particles in the fine size separate. As a consequence, sorption reactions of phosphate and metal ions occur at the interfaces of the Al hydroxide particles with the surrounding litter matrix and are therefore limited and rather slow. Phosphate is enriched in the fine size separate due to its high affinity for Al hydroxide.
Sequential Extraction
The amounts of P, Ca, Al, and Fe extracted from the two size separates of sample PL 181 with each extraction step are presented in Fig. 1 and 2
, respectively. Most of the P is acid-extractable or recalcitrant to extraction, but P is released during each extraction step. Dou et al. (2003) reported lower amounts of acid-extractable and residual P in alum-amended broiler litter. This can be attributed to different reasons, particularly incubation time, litter origin and treatment, or diet. Dou et al. (2003) reacted fresh, moistened litter with approximately the same amount of alum as was used in our study for 3 d, whereas our sample was stored for 2 yr after sampling from a house that had received alum. Although Dou et al. (2003) and Dao et al. (2001) had found that a few days' difference in incubation time had little effect on the fraction of water-soluble P in alum-amended poultry manure, substantial changes can be expected over time scales of months and years. In a preliminary study of the effects of sample aging on the P speciation in alum-amended and unamended PL, Hunger et al. (unpublished data, 2005) were able to show that within 6 mo time the proportion of relatively insoluble calcium and Al phosphate phases increases in both amended and unamended litter.
|
|
This sample had received alum and a significant amount of P was therefore expected to be associated with Al. Indeed, 1.5 and 4.5 g/kg of P (8.4 and 19%) are extracted from the coarse and fine size separates, respectively, by NaOH, concomitant with 7.2 and 29 g/kg of Al, respectively. More Al is present in the fine size separate than in the coarse size separate, which indicates that most of the Al hydroxide formed by hydrolysis of alum is present as particles smaller than 125 µm. The fact that much more Al than P is released by NaOH indicates that the fraction of Al associated with P makes up only a small part of the total amount of Al in the samples. Furthermore, Sharpley and Moyer (2000) demonstrated that a significant amount of organic P was also extracted with 0.1 M NaOH; indeed, NaOH is the extractant of choice to quantitatively extract organic phosphate species from soils (Turner et al., 2003). The NaOH fraction therefore contains a heterogeneous mixture of P species, of which P formerly associated with Al is merely one part.
Significant amounts of Al are also extracted by HCl (2.5 and 6.5 g/kg in the coarse and fine size separates, respectively). This acid-extractable Al fraction is either a different Al phase or Al hydroxide that was not extracted by NaOH due to supersaturation of the extractant. No attempt was made to exhaustively extract each fraction by repeated treatment with the extractant, a very time-consuming procedure developed by Dou and coworkers (Dou et al., 2000). Also, the samples were extracted at a higher solid to solution ratio than reported in the literature. It cannot, therefore, be excluded that the extract was saturated with respect to amorphous Al(OH)3. Using solid-state 31P NMR spectroscopy, Frossard et al. (1994) observed an Al phosphate phase in samples of biosolids after a similar extraction sequence. This phase, identified tentatively as a poorly ordered wavellite, was not observed in the residues after extraction with 0.1 M NaOH but rather after extraction with 1.0 M HCl. It was therefore concluded that it precipitated during extraction.
Most of the Ca in the samples is acid-extractable; however, some is present in the residual fraction. Similar results are seen for Fe, with an even higher proportion being recalcitrant to extraction. This has been attributed to insoluble or only slowly soluble crystalline calcium and iron phosphate minerals (Dou et al., 2000; Hedley et al., 1982). The major part of acid-extractable Ca can be attributed to calcium carbonate, which is contained in poultry feed. Calcium not absorbed during digestion precipitates as calcium carbonate under open atmosphere conditions and acts as a sorbent for phosphate (Griffin and Jurinak, 1973, 1974; Hinedi et al., 1992; Kuo and Lotse, 1972). Consequently, 37 and 39% of P are concomitantly extracted by 1.0 M HCl. The insoluble P not associated with metals has been mainly attributed to occluded phosphate species (Hedley et al., 1982).
Phosphorus-31 Nuclear Magnetic Resonance Spectroscopy
The 31P NMR spectra of the sequentially extracted coarse and fine size separates of sample PL 181 are presented in Fig. 3 and 4
, respectively. The P concentrations of the samples after extraction with HCl were too low to give spectra of good quality in an acceptable experiment time. Of these spectra only the CP–MAS NMR spectrum of the fine size separate after acid extraction is therefore presented in Fig. 5
.
|
|
|
There are several possible explanations for this behavior, the most obvious one being that the assignment is incorrect. Another explanation is that this P species is occluded, either in remnants of biological compartments, such as plant or microbial cells, or in humified organic matter or complexes of humified organic matter with inorganic polymers. Complexes between humified organic material and inorganic polymers, such as aluminosilicates and Al or iron hydroxide polymers, have been proposed for soils (Sparks, 2003; Stevenson, 1994). Myneni et al. (1999) demonstrated that changes in ionic strength and ionic composition of the background electrolyte cause changes in the shape and association of dissolved humic acid. Extracting PL with 0.5 M NaHCO3 probably causes the organic–inorganic aggregations to break up and release any occluded phosphate. The different behavior of litter samples reported here and by Hunger et al. (2004) illustrates differences in litter compositions. The sample used in this study had received alum; more and stronger associations between humified material and the added Al are expected to form. The litter sample used in the study reported by Hunger et al. (2004) was unamended.
In general, the NMR spectra of the extracted samples confirm some of the molecular assignments of the macroscopic P fractionation. Phosphate associated with Al in the fine size separate, which can be detected as a shoulder on the upfield side of the broad peak in the right column of Fig. 4 (Hunger et al., 2004), is removed by NaOH. More phosphate is associated with Al in the fine size separate than in the coarse size separate, because the NMR spectra of the coarse size separate are lacking the peak in the corresponding chemical shift region (–4 to –10 ppm). This agrees with the observation made in the macroscopic fractionation that less Al and P are extracted by NaOH from the coarse size separate.
Furthermore, the calcium phosphate phase is present in all extracted samples (left column of Fig. 3 and 4) and is only removed by HCl (Fig. 5). The peak of the calcium phosphate phase is suppressed in the CP–MAS NMR spectra and not as sharp as in the single-pulse spectra, but the shift to more negative values is clearly visible in the bottom spectra of Fig. 5. The CP–MAS NMR spectrum of the sample after acid extraction further reveals that the residual P fraction contains a rather uniform chemical species, as evidenced by the narrow peak. The spectrum further contains a shoulder on the upfield side of the peak in the chemical shift range –5 to –20 ppm. This indicates the presence of phosphate species bound to Al (Hunger et al., 2004) that were not completely removed during the extraction sequence or precipitated due to supersaturation. The concentration of this phase is, however, very low and it is quite possible that the signal observed in the spectrum on the bottom of Fig. 5 is hidden in the complex signal in the spectrum at the top. All spectra were normalized to the most intensive peak and do therefore not allow direct quantitative comparison. This phosphate phase associated with Al is therefore most probably a residue that was not completely extracted. This is also a possible explanation for the observations by Frossard et al. (1994), who found an Al hydroxyl-phosphate phase in the residual fraction of urban biosolids after a similar extraction sequence. The need to repeat each extraction step until the phosphate concentration in the extract is below detection, as proposed by Dou et al. (2000), is evident.
Extraction of the samples with deionized water narrows the complex peaks in the NMR spectra (Fig. 3 and 4). Because only a small fraction of P was removed during this step, this indicates that removal of soluble metal cations has simplified the chemical environment of both inorganic and organic phosphate. The shift to more negative chemical shift values with extraction further indicates that inorganic phosphate also reacted with Al hydroxide to form surface complexes and Al hydroxyl-phosphate precipitates (Hunger et al., 2004). While the reaction mixture was quite heterogeneous during storage in the chicken houses, it is more homogeneous during extraction, facilitating adsorption of soluble phosphate to Al hydroxide and/or calcium carbonate phases. Further extraction with NaHCO3 and especially so with NaOH shifts the peaks to more positive values, centering them on the calcium phosphate peak at 3.0 ppm in the single-pulse MAS spectra and slightly lower values in the CP–MAS spectra. These extractants efficiently remove organic phosphate compounds and inorganic phosphate not bound in a solid phase (Cade-Menun and Preston, 1996; Turner, 2004), leaving mainly the calcium phosphate phase (3.0 ppm).
Comparison of the Quantification by Extraction and Phosphorus-31 Nuclear Magnetic Resonance
Deconvolution of the signals in the single-pulse spectra was not possible to the extent reported previously (Hunger et al., 2004). Only the peak at = 3.0 ppm was reproducible over the course of the deconvolution. The main reason for this is possibly the limited number of peaks used for deconvolution, each of which consists of resonance lines of many different species that only differ in conformation and substitution (for organic phosphate) or the number and nature of cations complexed (for both organic and inorganic phosphate species). Extraction changes the proportions of these species, moving the apparent peak positions. The calcium phosphate phase, being the most homogeneous, did not undergo such profound changes. Therefore, only the fraction of calcium phosphate was determined from the single-pulse 31P NMR spectra of the residues of the coarse and fine size separates after each extraction step (Table 2). The expected enrichment of the calcium phosphate phase during the course of the extraction is only observed in the coarse size separate. The changes in both size separates, however, are well inside the error margin (±5%) and therefore not significant. The acid-extractable phosphate fraction, which presumably contains predominantly calcium phosphate, was 39 and 37% in the coarse and fine size separates, respectively. The calcium phosphate content determined by 31P NMR, however, is 4.3 and 3.6%, respectively. Clearly, sequential chemical extraction overestimates the calcium phosphate content by one order of magnitude by dissolving species other than calcium phosphate. As discussed previously, phosphate associated with Al is incompletely extracted with NaOH and makes up part of the phosphate fraction extracted by HCl. Furthermore, possible acid-extractable species that were not accounted for by 31P NMR are complexes of phytic acid (inositol hexakisphosphate) with a variety of metal cations.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The results of this study are, however, limited by the fact that only one sample was extracted and subsequently analyzed by solid-state 31P NMR. This sample had furthermore been stored for 2 yr and had probably undergone major changes toward more stable phosphate species (Hunger et al., unpublished data, 2005). It is therefore questionable whether these results are applicable to fresh poultry litter, particularly considering the variability of possible litter compositions. The general conclusion, however, remains that sequential chemical extraction schemes employing nonspecific reagents have to be interpreted with caution. Further research comparing speciation using different techniques is clearly needed to validate chemical extraction sequences.
Comparing these results with solution 31P NMR data of PL extracts (Turner, 2004) demonstrates that a combination of solid-state and solution 31P NMR spectroscopy will offer a much deeper insight into P speciation than each technique on its own. While solid-state 31P NMR spectroscopy is an invaluable tool to quantitatively assess the proportions of Ca and Al phases, solution 31P NMR spectroscopy of alkaline extracts of poultry litter allows the quantification of organic P species that are not detected by solid-state NMR spectroscopy due to the heterogeneity of the material. Both techniques, however, require sophisticated analytical equipment that is not available for routine measurements of animal manures in general. A combination of both techniques with established methods such as sequential chemical extraction, on the other hand, will allow the validation of results gained from those routinely employed methods. This will also allow the in-depth study of the effects of storage and chemical amendments on the phosphate speciation in poultry litter in particular and animal manures in general.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  J. G. Warren, C. J. Penn, J. M. McGrath, and K. Sistani The Impact of Alum Addition on Organic P Transformations in Poultry Litter and Litter-Amended Soil J. Environ. Qual., March 1, 2008; 37(2): 469 - 476. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. G. Warren, S. B. Phillips, G. L. Mullins, and L. W. Zelazny Impact of Alum-Treated Poultry Litter Applications on Fescue Production and Soil Phosphorus Fractions Soil Sci. Soc. Am. J., September 20, 2006; 70(6): 1957 - 1966. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. McGrath, J. T. Sims, R. O. Maguire, W. W. Saylor, C. R. Angel, and B. L. Turner Broiler Diet Modification and Litter Storage: Impacts on Phosphorus in Litters, Soils, and Runoff J. Environ. Qual., September 8, 2005; 34(5): 1896 - 1909. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
before and after fractionation of poultry litter sample PL 181.
