journal menu
research papers
 | JOURNAL OF SYNCHROTRON RADIATION |
Nitrogen K-edge XANES – an overview of reference compounds used to identify `unknown' organic nitrogen in environmental samples
aInstitute for Land Use, University of Rostock, Justus von Liebig Weg 6, 18057 Rostock, Germany, bDepartment of Soil Science, University of Saskatchewan, Saskatoon, Canada S7N 5A8, and cCanadian Light Source Inc., University of Saskatchewan, Saskatoon, Canada S7N 0X4
*Correspondence e-mail: peter.leinweber@uni-rostock.de
The chemical nature of soil organic nitrogen (N) is still poorly understood and one-third to one-half of it is typically classified as `unknown N'. Nitrogen K-edge XANES spectroscopy has been used to develop a systematic overview on spectral features of all major N functions in soil and environmental samples. The absolute calibration of the photon energy was completed using the 1s → π* transitions of pure gas-phase N2. On this basis a library of spectral features is provided for mineral N, nitro N, amino acids, and substituted pyrroles, pyridines, imidazoles, pyrazoles, pyrazines, pyrimidines and Although N XANES was previously considered `non-destructive', effects of radiation damage were shown for two compound classes and an approach was proposed to minimize it. This new evidence is integrated into a proposal for the evaluation spectra from environmental samples with unknown composition. Thus a basis is laid to develop N K-edge XANES as a complementary standard research method to study the molecular composition and ecological functions of `unknown N' in soil and the environment.
Keywords: soil nitrogen; heterocycles; soft X-ray spectroscopy; radiation damage; NEXAFS.
1. Introduction
Organic nitrogen (N) compounds in soils are both a source and sink for N in the global N cycle. The soil organic N pool provides the substrate for microbial formation of mineral N (NH4 and NO3) essential for agricultural production and likewise is the source of serious environmental pollution (e.g. NO3 contamination of groundwater and emissions of N2O, an important greenhouse gas). Despite its importance, the organic N pool contributing to N mineralization has not been adequately defined at the molecular level (Curtin & Wen, 1999![[Curtin, D. & Wen, G. (1999). Soil Sci. Soc. Am. J. 63, 410-415.]](../../../../../../logos/arrows/s_arr.gif "Curtin, D. & Wen, G. (1999). Soil Sci. Soc. Am. J. 63, 410-415.")
) and the chemical nature of soil organic N remains poorly understood (Schulten & Schnitzer, 1998). Indeed, one-third to one-half of organic N in soil is typically classified as `unknown N' (Stevenson, 1994) and the chemistry of this N fraction remains one of the most controversial soil chemistry issues in the literature. The controversy exists, and remains unresolved, largely because research has been hampered by the absence of definitive analytical methods (Stankiewicz, 1998).
To date, four different analytical approaches have been used to identify organic N fractions: (i) wet chemistry, which has included hydrolysis in 6N HCl, and wet-chemical fractionation into hydrolysable N, NH3-N and unidentified N (Leinweber & Schulten, 1999), sometimes accomplished by the separation and quantification of individual amino acids in the hydrolyzate (Catroux & Schnitzer, 1987); (ii) Curie-point-pyrolysis gas chromatography/mass spectrometry (Cp Py-GC/MS) and pyrolysis-field ionization (Py-FIMS), often applied complimentarily (Schulten et al., 1997; Schulten & Schnitzer, 1998); (iii) 15N nuclear magnetic resonance spectroscopy (15N NMR) (Knicker et al., 1993; Knicker, 2000); and (iv) X-ray spectroscopic methods such as (Abe & Watanabe, 2004) and X-ray absorption near-edge structure (XANES) spectroscopy (Vairavamurthy & Wang, 2002; Jokic, Schulten et al., 2004; Jokic, Cutler et al., 2004; Zhuang et al., 2006).
There is general agreement that proteinaceous materials constitute the majority of soil organic N (Schulten & Schnitzer, 1998; Stevenson & Cole, 1999; Knicker, 2000); however, opinions diverge regarding the proportions and chemical nature of non-proteinaceous N compounds. For example, using wet chemical fractionation, Leinweber & Schulten (1999) concluded that heterocyclic N compounds comprise about 25% of organic N in soil. Similarly, others have suggested that heterocyclic N accounts for about 35% of organic N in soil on the basis of studies using GC/MS, Cp Py-GC/MS and Py-FIMS (Schulten et al., 1997; Schulten & Schnitzer, 1998). These observations are consistent with the wide range of heterocyclic N compounds documented in plant materials (Pedras et al., 2003; Somei & Yamada, 2004). However, Py-MS estimates have been criticized because heterocyclic N can form as an artifact of dissociation or rearrangement of the original structures during heating (Burdon, 2001; Vairavamurthy & Wang, 2002). This argument was supported by 15N NMR studies that sometimes failed to detect heterocyclic N (Knicker, 2000; DiCosty et al., 2003; Sjöberg et al., 2004) or only detected it in soils affected by wildfires (Knicker et al., 2005). However, others have suggested that 15N NMR spectroscopy is simply too insensitive to detect some soil organic N compounds because of the low 15N natural abundance (∼0.37%) in soil (Vairavamurthy & Wang, 2002) and overlap of spectral features of pyrrole-, indole- and purine-N with those of peptide-, amide- and pyrimidinic N (Abe & Watanabe, 2004; Kelemen et al., 2002).
Synchrotron-based XANES is a molecular-scale spectroscopy technique that yields electronic and structural information on the element of interest. An XANES spectrum is usually characterized by intense resonance features, arising from 1s electron transitions to unoccupied orbital and continuum levels and from multiple scattering of the emitted photoelectrons by the geometrical arrangement of neighbouring atoms around the absorbing atom. The energy position of the resonance features is related to the chemical environment of the absorber, such as (Stöhr, 1992; Schulze & Bertsch, 1995; Fendorf & Sparks, 1996). In the early 1990s, XANES at the N K-edge was used to determine the chemical structure of organically bound N in fossil fuels such as petroleum asphaltenes (Mitra-Kirtley et al., 1993). More recently, N K-edge XANES was successfully applied to characterize structures of N atoms in samples from biomaterials (e.g. Fujii et al., 2003; Samuel et al., 2006), interstellar matter (e.g. Feser et al., 2003; Wirick et al., 2004), material sciences (e.g. Watanabe et al., 2002; Choi et al., 2005) and carbon combustion technology (e.g. Xiao et al., 2005). In the first application of N K-edge XANES spectroscopy to geomacromolecules in environmental samples, Vairavamurthy & Wang (2002) noted that heterocyclic N compounds constituted up to 30% of total N in chemically fractionated humic substances and sediments, and postulated that abiotic formation of heterocyclic N compounds may occur during early organic matter transformations. More recently, Jokic, Schulten et al. (2004) used N K-edge XANES to demonstrate a possible mechanism for abiotic formation of heterocyclic N. Jokic, Cutler et al. (2004) subsequently used N K-edge XANES to detect the presence of heterocyclic N compounds in chemically unaltered whole soil samples. These recent studies provide compelling evidence for heterocyclic N, and also demonstrate that synchrotron-based spectroscopy offers a powerful chemical and structural probe for the study of soil organic N in whole soils. Because XANES is based on electron excitation, the method is considered to be non-destructive as the structure of the molecules remains unchanged and N speciation is unaltered (Vairavamurthy & Wang, 2002). Moreover, N XANES detects all N present, and not just 15N as with 15N NMR. However, as for other sophisticated analytical-chemical methods, the quality of results obtained by N XANES will depend on technical details of facility and equipment, and choice of appropriate reference standard substances for comparison with spectra from samples of unknown chemical composition.
Vairavamurthy & Wang (2002), and subsequently Jokic, Schulten et al. (2004) and Jokic, Cutler et al. (2004), interpreted N K-edge XANES spectra of environmental samples of different origins on the basis of analyses of a limited set of reference compounds including 2,6-di-p-tolylpyridine, 6-hydroxyquinoline, quinoline 6-carboxylic acid and acridine (pyridines), pyridine 2,5-dicarboxylic acid, pyridoxal 5′-phosphate (oxidized pyridine derivatives), 4,6-dihydroxyl pyridine (pyridone), carbazole, poly(9-vinylcarbazole) (pyrrole), albumin, N-acetyl-D-glucosamine, glycine anhydride (amide/peptide) and 4-nitrophenyl acetic acid (nitro compound). However, considering the findings of gel-chromatography acetylation GC/MS and Py-MS (Schulten & Schnitzer, 1998), the N K-edge XANES spectra of reference compounds including imidazoles, pyrazoles, pyrazines and pyrimidines are essential as some of them are DNA/RNA building blocks, which were identified in soils (Anderson, 1961; Cortez & Schnitzer, 1979). Recent quantification of DNA in soils (Bürgmann et al., 2001) supports the importance of these molecules for a complete picture of soil organic N. Therefore, our objective was to provide a systematic overview of the spectral features of all major N functions that potentially occur in soils and soil-related materials. Our strategic aim is to develop N K-edge XANES as a complementary standard research method in soil organic chemistry.
2. Experimental
The chemicals used as reference compounds were purchased from Sigma Aldrich, and the CAS (Chemical Abstracts Service) numbers are given in Table 1. All reference compounds were used without further purification. Powdered samples were applied to double-sided carbon tape (G3939, Plano GmbH, D-35578 Wetzlar, Germany) attached to a 10 mm-diameter stainless steel sample holder before insertion into the X-ray absorption chamber. To avoid problems associated with self-absorption, all reference compounds were spread as a very thin film onto the sample holder.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The N K-edge XANES spectra were measured on beamline 11ID-1 at the Canadian Light Source synchrotron. This beamline is equipped with a spherical-grating monochromator and is capable of providing 1011 photons s−1 at the N K-edge with a (E/ΔE) better than 10000 with an exit slit setting of 5 µm (Regier et al., 2006). For most measurements the exit slit was set at 50 µm. Only a few of the amino acids were measured using the 25 µm exit slit. The spectra were recorded in total using a two-stage multichannel plate detector.
The absolute calibration of the photon energy was completed using the 1s → π* transitions of pure gas-phase N2 and monitored over time using the same transition in 6-hydroxyquinoline, guanine, adenine and 2,3-pyrazinedicarboxylic acid. An for N2 was measured using a chamber mounted just upstream of the solid sample analysis chamber (Regier et al., 2007). The well studied 1s → π* vibrational manifold has an energy of 400.8 eV for the ν = 0 transition (Schwarzkopf et al., 1999). The N K-edge XANES measurement of the above reference standards exhibit a sharp 1s → π* peak which was calibrated using the N2 data. As the experiment took place over several different measurement periods, the beamline calibration could not be considered static, so one or more of the above reference standards was measured in two or more experimental runs and the energy calibration adjusted accordingly.
Data reduction was carried out using Athena (Version 0.8.05) (Ravel & Newville, 2005). The fluorescence data of at least two scans were averaged and background corrected by a linear regression fit through the pre-edge region. Then all spectra were normalized to an edge jump of 1. The spectra presented were smoothed by interpolation with two iterations.
3. Results and discussion
3.1. Spectra of reference compounds
Figs. 1 to 11 show the stacked N K-edge XANES spectra of standard reference compounds arranged into groups with similar structural properties. All spectral features in the energy range 398.6 to 405 eV reflect 1s → π* transitions, and the broader spectral features at higher energies >405 eV reflect 1s → σ* transitions. The energy positions of the most intensive resonance in the normalized spectra are compiled in Table 1. The peak position and shape varied significantly for the different model compounds. Generally, this was more pronounced for the 1s → π* than for the 1s → σ* transitions. In our analysis, we compared spectra that we measured in common with those from other studies. In many cases, systematic shifts in peak positions are observed, indicating a difference in beamline calibration. Since previous studies have not indicated methods of photon energy calibration, we have applied these systematic shifts to compare with our own spectra as outlined in the sections below.
Fig. 1 shows the spectra of common mineral N compounds in soil. All have a π* resonance around 401 eV. The peak height of this resonance increased in the order (NH4)2SO4 > NH4NO3 > (NH4)3PO4. Additional high-resolution scans (boxed spectra on the left in Fig. 1) resolved the resonance at around 401 eV in a series of π* resonances. The position and shape of these resonances were the same as reported for N2 gas by Schwarzkopf et al. (1999). Therefore, we assign this peak to N2 gas which was either trapped in the mineral N compounds or generated during X-ray exposure. The spectra of KNO3 and Ca(NO3)2 had two sharp π* resonances, one at 401.7 eV and the other at 405.4 eV. The first peak showed a small shoulder on the lower-energy side. This shoulder was more pronounced in the spectrum of Ca(NO3)2 than in the spectrum of KNO3. This shoulder coincided with the assigned N2 peak at 401 eV in the spectra of ammonium salts. In the high-resolution spectrum of KNO3 some of the distinctive π* resonance of N2 gas were visible. Therefore the shoulder on the lower-energy side of the first π* resonance at 401.7 eV can also be assigned to N2 gas released during the measurement. However, the majority of this intensive resonance is difficult to assign. Since the nitrate-N has the diagnostic resonance at 405.4 eV in KNO3 (Rodrigues et al., 2007) and Ca(NO3)2 (Fig. 1), we interpret the resonance at 401.7 eV as a decomposition product. If this explanation is correct, formation of this nitrate decomposition product decreased in the order KNO3 > Ca(NO3)2 > NH4NO3.
![[Figure 1]](ot5576fig1thm.gif) | Figure 1 Nitrogen K-edge XANES spectra of mineral N compounds. Inset boxes show sections of high-resolution spectra acquired at 0.01 eV step size and 0.5 s dwell time between 401 and 403 eV as indicated by the arrows in the lower resolved spectra (0.1 eV step size and 1.0 s dwell time). |
The high-energy peak of (NH4)2SO4 and (NH4)3PO4 at around 406 eV was assigned to the 1s → σ* transition. In the spectrum of NH4NO3 this transition was shifted by −0.6 eV. This can be explained by the overlap with the NO3 peak at 405.4 eV. This indicates that peaks >405 eV can be assigned to 1s → σ* transitions preferably if the sample is free of NO3-N. For NO3-N the 1s → σ* transition appeared at around 412 eV in all three nitrate salts.
Spectra of compounds containing N only in a nitro group showed a sharp π* peak at 403.8 to 403.9 eV and a broad high-energy peak around 413.7 eV (Fig. 2). This is explained by the higher positive charge at the N atom and, consequently, higher energy required for the 1s → π* transition. The energy position of the π* resonance agreed with Pavlychev et al. (1995) who reported a peak at 403.3 eV for the nitro group. Furthermore, our spectroscopic features for the nitro group confirm previous data by Vairavamurthy & Wang (2002) if we shift their energy scale by −1.1 eV. Since Vairavamurthy & Wang (2002) did not report photon energy calibration, Jokic, Cutler et al. (2004) applied an energy shift of −1.2 eV, and subsequently obtained good agreement in the energy positions of spectral features.
![[Figure 2]](ot5576fig2thm.gif) | Figure 2 Nitrogen K-edge XANES spectra of nitro-N-containing compounds. |
Amino acids form the largest group among the reference compounds under study (Fig. 3). The majority of amino acid spectra were dominated by a relatively broad σ* resonance peak at around 406 eV. All spectra showed a more or less pronounced peak at around 399 eV, followed by peaks or shoulders at around 401 eV, except for DL-aspartic acid, D-methionine, L-proline and L-histidine. The first three amino acids showed no spectral feature around 401 eV. The two sharp peaks at 400.0 and 401.6 eV in the L-histidine spectrum are assigned to the two N atoms in the heterocyclic molecule. The broad σ* resonance peak at around 406 eV confirms previous data by Zubavichus et al. (2005) and Vairavamurthy & Wang (2002). Furthermore, the spectra of L-arginine and L-histidine also agreed with Zubavichus et al. (2005), who showed similar spectra with peaks at a slightly higher energy (+0.2 to –0.3 eV). The first peaks at around 399 eV in most spectra and at 401.1 in the spectrum of threonine were not seen by Zubavichus et al. (2005). This may be explained by the difference in incident on the sample, which might have led to radiation damage of our amino acids (Zubavichus et al., 2005; 1011 photons s−1 at the spot size 1.2 mm × 0.5 mm versus 1011 photons s−1 in a 0.8 mm × 0.2 mm beam spot size). In a previous paper, Zubavichus, Zharnikov et al. (2004) discussed the sensitivity of amino acids to radiation damage. To check this, we compared the spectra obtained from the first scans with those of the second scans (Fig. 4). It appears that the peak at 398.8 eV (a) slightly increased in intensity and a new peak (b) became obvious at 400 eV. Furthermore, the shoulder at 401.6 eV (c) in the D-methionine spectrum became more pronounced towards a broad peak. The spectrum of D-arginine showed no difference to that of L-arginine (Fig. 3, Table 1). The sharp peak (c), assigned to a π* transition of the guanidine structure (Zubavichus et al., 2005), became less intense and broader. We explain these transformations in the spectra as chemical reactions in the sample under the impact of X-rays. Bozack et al. (1994) and Zubavichus, Zharnikov et al. (2004) described deamination and decarboxylation as possible pathways for the decomposition of amino acids under the impact of soft X-rays. Zubavichus, Fuchs et al. (2004) found a series of new peaks in the range 399 to 403 eV after 4 to 14 min X-ray impact to cysteine. The peaks (a), (b) and (c) (D-methionine) and (a) and (b) (D-arginine) in Fig. 4, not observed in the spectra of Zubavichus et al. (2005), suggest that such decomposition reactions may have already occurred during our first scans.
![[Figure 3]](ot5576fig3thm.gif) | Figure 3 Nitrogen K-edge XANES spectra of amino acids. |
![[Figure 4]](ot5576fig4thm.gif) | Figure 4 Nitrogen K-edge XANES spectra of amino acids: comparison of spectra merged from two first and second scans recorded at two different spots showing stability and damage of compounds under the beam impact. |
The spectra of amide and amine compounds showed distinct π* resonances around 401.4 eV (Fig. 5). This agrees well with Mitra-Kirtley et al. (1992) and Vairavamurthy & Wang (2002) (after calibration adjustments) and is explained by the π-electron cloud of the C=O bond closely coupled with the N. However, in contrast to Mitra-Kirtley et al. (1992) and Vairavamurthy & Wang (2002), our spectra of albumin and N-acetyl-D-glucosamine showed two low-energy resonances at around 399 eV and 399.9 eV. These resonances were more pronounced in the albumin than in the N-acetyl-D-glucosamine spectrum. Whereas both resonances were apparent in the first scan of albumin, these resonances were smaller in the first as compared with the second scans of N-acetyl-D-glucosamine (data not shown). Given the discrepancies between our results and those previously reported, together with observed spectral changes with subsequent scans, it is suggested that these resonances reflect N species formed by the impact of X-rays. It is possible that radiation damaged the amino acid polymer comprising albumin in a manner similar to that of pure amino acids. We conclude that the low-energy resonance at 398.9 eV in the albumin spectrum is due to X-ray decomposition whereas the second low-energy resonance at 399.9 eV can be attributed to both decomposition products and the first 1s → π* transition of L-histidine, which is an important constituent of albumin.
![[Figure 5]](ot5576fig5thm.gif) | Figure 5 Nitrogen K-edge XANES spectra of and amides. |
In the spectrum of chitin, a polymer of N-acetyl-D-glucosamine, the major peak occurred at 401.5 eV, and was more pronounced than in the monomer. This peak was missing in the spectra of D-glucosamine hydrochloride and sulfanilamide. These two compounds showed a broad low-intensity spectral feature around 400 eV (D-glucosamine hydrochloride) and 398.5 eV (sulfanilamide) and the broad spectral features of 1s → σ* transitions in the range 405.5 to 406.4 eV.
Pyrrole-2-carboxylic acid showed a distinct π* peak at 401.1 eV, carbazole a broader peak at 402.4 eV, 4-hydroxy-2-pyridone a relatively narrow peak at 401 eV, and 3-indoleacetonitrile a broad but less intensive peak at 402.2 eV (Fig. 6). We assign these spectral features to the N atom in the pyrrole and pyridone structure. The energy positions of the π* resonances of our pyrrole derivatives varied by approximately 1.3 eV. This is in agreement with others who also found that π* peak energy positions varied up to +0.5 to −1 eV for different pyrrole derivatives (Mitra-Kirtley et al., 1993; Mullins et al., 1993; Henning et al., 1996; Bach et al., 1997; Vairavamurthy & Wang, 2002). This divergence can be explained by differences in the chemical structures of these compounds. Our carbazole spectrum confirms previous data by Mullins et al. (1993) and Mitra-Kirtley et al. (1993) if we shift their energy scales by −1.1 eV and −0.9 eV, respectively. The resonance at 401 eV in the spectrum of 4-hydroxy-2-pyridone agreed with Vairavamurthy & Wang (2002) who called this substance 4,6-dihydroxypyridine. The designation as a pyridine can be misleading in this respect, because solid 2,4-dihydroxypyridine (synonym `3-deazaurazil') is stable in the tautomeric form 4-hydroxy-2-pyridone (Wilson, 1994). Pyridones generally had π* resonances at 400.7 ± 0.2 eV (Mitra-Kirtley et al., 1993; Vairavamurthy & Wang, 2002). The N in the pyridone structure is pyrrole-type as discussed below in more detail.
![[Figure 6]](ot5576fig6thm.gif) | Figure 6 Nitrogen K-edge XANES spectra of five-ring heterocycles. |
The spectrum of 3-indoleacetonitrile was largely dominated by another 1s → π* resonance at 399.9 eV. In agreement with previous studies of other nitrile compounds (Hitchcock et al., 1989; Apen et al., 1993; Carniato et al., 2005; Ray et al., 2005), we assign this peak to the nitrile structure. Since the nitrile bond has predominantly π character, an intensive peak is expected in the π* region. This is also confirmed by nitrile in the spectrum for 2-pyrimidinecarbonitrile (see below). For all pyrrole derivatives the 1s → σ* transition was observed as a broad feature at around 407 eV.
The 6-hydroxyquinoline and 2,3-pyridinedicarboxylic acid spectra both showed 1s → π* resonances at similar energy position of 398.7 eV and 398.8 eV, respectively (Fig. 7), despite their different chemical structures. This energy position is in agreement with previous studies of other pyridine derivatives (Mitra-Kirtley et al., 1992, 1993; Kelemen et al., 1994; Vairavamurthy & Wang, 2002). All pyridine analogues showed a broad feature in the range 407.5 to 409 eV.
![[Figure 7]](ot5576fig7thm.gif) | Figure 7 Nitrogen K-edge XANES spectra of 6-hydroxyquinoline and 2,3-pyridinedicarboxylic acid. |
Fig. 8 compares the spectrum of 3-(pyrrol-1-ylmethyl) pyridine, which contains both a pyridine and a pyrrole structure, with the spectra of carbazole (pyrrole) and 2,3-pyridinedicarboxylic acid (pyridine). The spectrum of 3-(pyrrol-1-ylmethyl) pyridine showed two distinct and clearly separated π* resonance features at 398.8 eV and 402.2 eV, respectively. The position of the lower-energy feature matches well with the π* resonance of the pyridine derivative 2,3-pyridinedicarboxylic acid. The feature at 402.2 eV matches well with the π* resonance of the pyrrole derivative carbazole or with pure pyrrole in the gas phase (Hennig et al., 1996). Therefore, we assign the first π* resonance in the spectrum of 3-(pyrrol-1-ylmethyl) pyridine to pyridinic N and the second higher-energy π* resonance to pyrrolic N. The difference in the energy position of the 1s → π* transition between pyridinic and pyrrolic N results from the difference in the effective nuclear charge of the absorbing N atom. It is known that the energy position of the π* transition increases with increasing effective nuclear charge or decreasing electron density on the absorbing atom. The N atom in the pyridine structure has its lone pair unshared in the sp2 orbital. The N atom in the pyrrole structure has the lone pair in the pz orbital, which is involved in the π cloud of the aromatic system. Therefore, pyrrole carries a lower electron density and consequently the π* resonance is at a higher energy position.
![[Figure 8]](ot5576fig8thm.gif) | Figure 8 Nitrogen K-edge XANES spectra of a five- and six-membered heterocyclic compound compared with 2,3-pyridinedicarboxylic acid and carbazole. |
All reference compounds having an imidazole structure showed at least two resonances, one around 400 eV and the other in the range 401.4 to 402.3 eV (Fig. 9). The imidazole structure contains two N atoms, which differ in the orbital description of the lone pair. Therefore, we assign the first resonance to the 1s → π* transition of the pyridinic and the second to the 1s → π* transition of the pyrrolic N. This assignment agrees with Apen et al. (1993) who measured imidazole in the solid and gas phase, and with Mitra-Kirtley et al. (1993) who described these two resonances in a 2-methylbenzimidazole spectrum. In the spectra of imidazole-4-acetic acid sodium salt and L-β-imidazolelactic acid the energy positions of both π* resonances were almost identical. The same resonances were observed in the L-histidine spectrum in Fig. 3. For imidazole-4-acetic acid the first π* resonance was slightly higher than the second resonance. For L-β-imidazolelactic acid (Fig. 9) and L-histidine (Fig. 3) the second resonance was higher in intensity than the first one. This difference in peak intensity was much more pronounced in the spectrum of 4-imidazoleacrylic acid. In this spectrum the peak for pyridinic N was reduced and formed a shoulder, and the peak for pyrrolic N appeared at 401.4 eV which is the lower end of the above range. The spectrum of 2-methyl-4-nitroimidazole was more complicated. We assign the peak at 399.9 eV to pyridinic N, the lower peak at 402.2 eV to pyrrolic N and the third peak at 403.9 eV to N of the nitro group. The assignment of the nitro-N agrees with the spectra in Fig. 2. 3.5-Pyrazoledicarboxylic acid monohydrate also exhibited two π* resonances of pyridinic (400 eV) and pyrrolic N (401.5 eV). This agrees with the above imidazole spectra and is interpreted by the similarity in electronic structure. In the spectrum of 1-H-pyrazole-1-carboxamidine the peaks at 399.8 eV and 402.3 are assigned to pyridinic and pyrrolic N, respectively. The peaks were slightly shifted if compared with the above ranges for pyridinic and pyrrolic N in the imidazole and pyrazole structures. This shift, as well as the two other spectral features at 400.6 and 404.4 eV, is probably due to the amidine structure. The spectral features of 1s → σ* transitions were in the range 406.7 to 407.3 eV.
![[Figure 9]](ot5576fig9thm.gif) | Figure 9 Nitrogen K-edge XANES spectra of five-ring heterocycles with at least two N atoms. |
All spectra of the differently substituted pyrazines showed one sharp intensive peak at 398.7 to 398.8 eV (Fig. 10). This is explained by the cumulative effect of 1s → π* transitions of the two pyridinic N atoms. Less intensive small features at 400 to 400.2 eV and 402.6 to 402.9 eV seem to vary with substitution but cannot yet be assigned in detail. The pyrimidine example 2-pyrimidinecarbonitrile also showed the intensive major peak at 398.8 eV because of the two pyridinic N atoms. In this spectrum the second peak at 400 eV is assigned to nitrile-N, in agreement with the nitrile peak in the spectrum of 3-indole-acetonitrile (Fig. 6). The σ* resonances were relatively uniform in the range 407.7 to 408.2 eV.
![[Figure 10]](ot5576fig10thm.gif) | Figure 10 Nitrogen K-edge XANES spectra of six-ring heterocycles with at least two N atoms. |
The spectra of O-substituted pyrimidines cytosine, uracil and thymine showed π* resonances at 399.2, 400.6 and 403 eV (cytosine), 401.1 to 401.4 eV and 402.2 eV (uracil and thymine). The double peak of the cytosine spectrum indicated pyridinic N (399.2 eV) and pyrrolic N (400.6 eV). The first peak was not observed in the spectra of uracil and thymine because they contain two pyrrolic N atoms. Comparison of the spectra of pyrimidines with, and without increasing O-substitution showed shifts of the π* resonances towards a higher-energy position. This was true for the comparison of pyridinic N in 2-pyrimidinecarbonitrile and cytosine (+0.4 eV), and of pyrrolic N in cytosine compared with uracil and thymine (+0.5 eV). Obviously the O-substitution affected the two pyrrolic N atoms differently, which was reflected by the double peak in the uracil and the shoulder in the thymine spectrum. The methyl group in the spectrum of thymine did not alter the energy position of the π* resonances but modified the shape of the spectrum. The spectra of showed several structural features in the σ* resonance region. In general, these three spectra agreed well with Mitra-Kirtley et al. (1992) if we apply an energy shift of −1.4 eV. Furthermore, the spectra of cytosine and thymine (first peak only) agreed well with Samuel et al. (2006) if we apply an energy shift of 0.2 eV (π1*) and +0.4 eV (π2*) to their spectra.
The π* resonances in the spectra of 4,5,6-triaminopyrimidinesulphate (399.6 eV) and sulfadimethoxine (399.8 eV) are slightly higher than in the above pyrazine and pyrimidine spectra.
The spectra of three compounds containing the purine structure all showed two π* resonances in the range 399.5 to 399.9 eV, and 401.1 to 401.6 eV (Fig. 11). The first peaks were always more intensive than the second peaks. This is explained by the ratio of pyridinic N to pyrrolic N of 4:1. In the spectrum of guanine the larger proportion of pyrrolic N was reflected by an increased intensity of the second peak. The exocyclic O-substitution at the purine ring structure resulted in shifts towards higher energy of the pyridinic N (+0.2 eV) and pyrrolic N (+0.6 eV) peaks. The spectral features of 1s → σ* transitions were in the range 407 to 407.6 eV. Again the spectra of adenine and guanine agreed with Mitra-Kirtley et al. (1992) provided an energy shift of −1.4 eV is applied, and with Samuel et al. (2006) provided an energy shift of +0.4 eV is applied.
![[Figure 11]](ot5576fig11thm.gif) | Figure 11 Nitrogen K-edge XANES spectra of five-ring and six-ring heterocycles which are ADP-, ATP-, RNA- and DNA-building blocks. |
The spectrum of DNA showed peaks at 399.4 and 401.6 eV, confirming DNA spectra published by Mitra-Kirtley et al. (1992) and Fujii et al. (2003). Since the above pyrimidines and purine structures are the N-containing building blocks of DNA, the peaks are assigned to pyridinic N and pyrrolic N. Surprisingly, the spectrum of RNA showed only one π* resonance at 401.3 eV. To the best of our knowledge, spectra of RNA have not been published before. The RNA differs from DNA in the sugar (ribose instead of desoxyribose), nucleobases (uracil instead of thymine) and the single-stranded conformation. Mitra-Kirtley et al. (1992) have shown that the glycosidic C—N linkage had an influence on the N K-edge XANES spectra of cytosine, uracil and guanine. More recently, Samuel et al. (2006) reported that addition of the DNA sugar and phosphate components resulted in noticeable, unexpected and still unexplained shifts of the π* resonances such as the appearance and disappearance of resolved doublet π*-peaks. Therefore, it can be hypothesized that the modification in the nucleobase and sugar possibly explains the completely different RNA spectrum. However, these preliminary results must be confirmed and RNA building blocks systematically studied in detail as recently shown for DNA (Samuel et al., 2006).
For the evaluation of N K-edge XANES spectra from samples with unknown composition, the following points become crucial for the compilation of the present spectra and comparison with literature data.
(i) The π* resonances at lowest energy (398.7 to 400 eV) can be assigned to pyridinic N in pyridines, imidazoles, pyrazoles, pyrazines, pyrimidines, cytosine and purine structures. Around 400 eV the π* resonances of pyridinic N in pyrazoles and purine structures can interfere with those of nitrile N. have been found in soil organic matter, and it was hypothesized that long-chain alkyl are soil specific (Schulten & Schnitzer, 1997; Leinweber & Schulten, 1998). Another relevant interference may originate from radiation damage products of amino acids, and labile which showed up with peaks at 398.9 and 399.9 eV.
(ii) Pyrrolic N in pyrroles, imidazoles, pyrazoles, pyrimidines and purine structures showed the π* resonance in the range 400.6 to 402.3 eV. Since this range was wider than previously reported by Mitra-Kirtley et al. (1993), Mullins et al. (1993), Hennig et al. (1996), Bach et al. (1997) and Vairavamurthy & Wang (2002), the present data confirm that pyrrolic N can clearly be differentiated from pyridinic N but may interfere with amide N.
(iii) Three spectra of different indicated 401.4 eV as the characteristic energy position of the amide N–π* resonance. This overlaps with the energy range found for pyrrolic N. Furthermore, the N2 gas peak (401 eV), unexpectedly released from mineral N, and a NO3 peak at 401.6 eV also interfered with pyrrolic N and amide N. However, since mineral N generally constitutes less than 1% of total N in arable soils, this interference is likely not to be relevant for the interpretation of N K-edge XANES of soil samples. It would be relevant, for instance, for the interpretation of N K-edge XANES of soil solutions or leachates that contain proportionately more mineral N.
(iv) Accordingly, broad π* resonance peaks at >400 eV may originate from amide N and pyrrolic N. The present data and the cited references showed a wide range of π* resonances for pyrrolic N. Therefore it seems difficult to resolve such broad peaks without complementary information on the abundance of pyrrolic N compounds and in the sample. Since amino acids and were shown to be susceptible to radiation damage, we deemed it essential to test whether radiation damage is likely to occur for soils and related samples.
3.2. Does radiation damage affect the evaluation of spectra from soil samples?
Fig. 12 shows that the first scan and second scan at two different spots of a Melfort clay sample differed. The peak (a) around 398.9 eV was unchanged. In both of the second scans a shoulder appeared at around 400 eV (b) and a peak at 401.5 eV (c) seemed to be slightly lower in intensity. The position of this shoulder agreed with a new peak observed at 400 eV in the second scan of D-methionine and D-arginine in Fig. 4. Since the smaller peak (c) agreed in energy position with the characteristic π* resonance of (Fig. 5), it is likely that and amino acids in the soil clay were decomposed under the beam impact. Radiation damage of amino acids observed by Zubavichus, Fuchs et al. (2004) was attributed to the formation of imino (C=NH) and nitrile (C≡N) derivatives. This was largely derived from combinations of C 1s, O 1s and N 1s XPS (X-ray photoelectron spectroscopy) spectra and electron impact-ionization The nitrile function in our spectra also had a π* resonance at 399.9 eV (Figs. 6 and 10). Therefore, it is possible that were formed by radiation damage of and amino acids in our samples although the formation of other N functions cannot be excluded. This partly disagrees with Vairavamurthy & Wang (2002) who concluded that X-ray analysis did not alter their sample spectra; however, their experimental approach to studying radiation damage and explanation of Fig. 3 remained unclear. Cherezov et al. (2002) reported damage of soft condensed biomaterials such as and membranes in bright X-ray synchrotron radiation. To avoid over-exposure they proposed to continuously move the sample during a measurement.
![[Figure 12]](ot5576fig12thm.gif) | Figure 12 Nitrogen K-edge XANES spectra of a Melfort soil clay sample: comparison of spectra merged from two first and second scans recorded at two different spots (upper) and single scans recorded at four different spots and the corresponding merged spectrum. |
Movement of the Melfort clay sample after each scan and recording the spectra at four different spots avoided radiation damage owing to shorter exposure time (Fig. 12, lower). There were no differences in peak positions, and the merged spectrum agreed with the first scans of a separate sample of the Melfort coarse clay. Lower signal-to-noise ratios for the four individual scans pointed to the importance of scan replications, which should be recorded at different spots in a homogeneous sample. In this way, possible effects of radiation damage on the evaluation of spectra from soil samples can be minimized although we cannot exclude the possibility that the X-ray radiation during the first scan damaged compounds in the sample.
4. Conclusions
Identification of organic N compounds in soil and environmental samples remains an important challenge for soil and environmental scientists. Much of the current controversy regarding soil organic N relates to methodological constraints and limitations that have hampered efforts to unequivocally resolve its molecular nature. We have presented a systematic overview of the spectral features of all major N functions likely to occur in soils and soil-related environmental materials. This study provides a more comprehensive library of spectral features than previously reported. To the best of our knowledge, this is the first time that N K-edge XANES spectra of imidazoles, pyrazoles and purine structures have been reported and discussed in the context of naturally occurring organic N in soils. Since shifts in the peak position according to the electronic structure of N atoms in various molecules were more pronounced for the 1s → π* than for the 1s → σ* transitions, it is suggested that the diagnostic value of a spectral feature is greater for the π* than for the σ* region in the XANES spectra. At this time, N K-edge XANES can be employed to reliably distinguish two π*-regions in spectra of unknown environmental samples: ≤400 eV (pyridinic N and some nitrile N) and >400 eV (amide N, nitro N and pyrrolic N). A quantitative estimate of the proportions of amide N and pyrrolic N requires a separate determination of one of the compounds, which remains a challenge for forthcoming studies. Furthermore, spiking experiments with multiple mixtures are necessary to evaluate to what extent pyridinic N and nitrile N can be distinguished in the low-energy range, and amide N and pyrrolic N in the higher-energy range. Radiation damage was observed for individual amino acids, and to some extent also for organic matter in a soil clay sample. Data suggest that radiation damage can be largely avoided by moving the sample holder after each scan. Given the subtle variations in spectral features, even for compounds sharing very similar structures, it is unlikely that the current N K-edge XANES can be used exclusively to quantitatively determine the chemical composition of an environmental sample. However, used in conjunction with existing complementary soil chemical analyses, N K-edge XANES can be developed as a powerful tool for elucidating the nature of `unknown soil N'.
Acknowledgements
The N K-edge XANES measurements were performed at the Canadian Light Source, which is supported by NSERC, NRC, CIHR and the University of Saskatchewan. This work was supported by a travel grant of the German Academic Exchange Service (D/05/50492). We thank Professor M. S. C. Pedras, Department of Chemistry, University of Saskatchewan, for providing some of the reference compounds. Thanks also to two anonymous referees for helpful comments which improved the manuscript.
References
Abe, T. & Watanabe, A. (2004). Soil Sci. 169, 35–43. Web of Science CrossRef CAS Google Scholar
Anderson, G. (1961). Soil Sci. 91, 156–161. CrossRef CAS Google Scholar
Apen, E., Hitchcock, A. P. & Gland, J. E. (1993). J. Phys. Chem. 97, 6859–6866. CrossRef CAS Web of Science Google Scholar
Bach, A., Beyer, L., Hennig, C., Hallmeier, K. H. & Sieler, J. (1997). Chem. Mon. 128, 217–224. CrossRef CAS Google Scholar
Bozack, M. J., Zhou, Y. & Worley, S. D. (1994). J. Chem. Phys. 100, 8392–8398. CrossRef Web of Science Google Scholar
Burdon, J. (2001). Soil Sci. 166, 752–769. Web of Science CrossRef CAS Google Scholar
Bürgmann, H., Pesaro, M., Widmer, F. & Zeye, J. (2001). J. Microbiol. Methods, 45, 7–20. Web of Science PubMed Google Scholar
Carniato, S., Taïeb, R., Kukk, E., Luo, Y. & Brena, B. (2005). J. Chem. Phys. 123, 214301. Web of Science CrossRef PubMed Google Scholar
Catroux, G. & Schnitzer, M. (1987). Soil Sci. Soc. Am. J. 51, 1200–1207. CrossRef CAS Google Scholar
Cherezov, V., Riedl, K. M. & Caffrey, M. (2002). J. Synchrotron Rad. 9, 333–341. Web of Science CrossRef CAS IUCr Journals Google Scholar
Choi, C. H., Park, J. & Kim, B. (2005). J. Phys. Chem. B, 109, 4333–4340. Web of Science CrossRef PubMed CAS Google Scholar
Cortez, J. & Schnitzer, M. (1979). Soil Sci. Soc. Am. J. 43, 958–961. CrossRef CAS Google Scholar
Curtin, D. & Wen, G. (1999). Soil Sci. Soc. Am. J. 63, 410–415. CrossRef CAS Google Scholar
DiCosty, R. J., Weliky, D. P., Anderson, S. J. & Paul, E. A. (2003). Org. Geochem. 34, 1635–1650. Web of Science CrossRef CAS Google Scholar
Fendorf, S. E. & Sparks, D. L. (1996). Methods of Soil Analysis: Chemical Methods, edited by D. L. Sparks, pp. 377–416, Soil Science Society of America Book Series 5. Madison, WI: Soil Science Society of America. Google Scholar
Feser, M., Wirick, S., Flynn, G. J. & Keller, L. P. (2003). 34th Lunar and Planetary Science Conference, 17–21 March 2003, Texas, USA. Google Scholar
Fujii, K., Akamatsu, K., Muramatsu, Y. & Yokoya, A. (2003). Nucl. Instrum. Methods Phys. Res. B, 199, 249–254. Web of Science CrossRef CAS Google Scholar
Hennig, C., Hallmeier, K. H., Bach, A., Bender, S., Franke, R., Hormes, J. & Szargan, R. (1996). Spectrochim. Acta A, 52, 1079–1083. CrossRef Google Scholar
Hitchcock, A. P., Tronc, M. & Modelli, A. (1989). J. Phys. Chem. 93, 3068–3077. CrossRef CAS Web of Science Google Scholar
Jokic, A., Cutler, J. N., Anderson, D. W. & Walley, F. L. (2004). Can. J. Soil Sci. 84, 291–293. CrossRef CAS Google Scholar
Jokic, A., Schulten, H. R., Cutler, J. N. & Schnitzer, M. (2004). Geochim. Cosmochim. Acta, 67, 2585–2597. Web of Science CrossRef Google Scholar
Kelemen, S. R., Afeworki, M., Gorbaty, M. L., Kwiatek, P. J., Solum, M. S., Hu, J. Z. & Pugmire, R. J. (2002). Energy Fuels, 16, 1507–1515. Web of Science CrossRef CAS Google Scholar
Kelemen, S. R., Gorbaty, M. L. & Kwiatek, P. J. (1994). Energy Fuels, 8, 896–906. CrossRef CAS Web of Science Google Scholar
Knicker, H. (2000). J. Environ. Qual. 29, 715–723. CrossRef CAS Google Scholar
Knicker, H., Fründ, R. & Lüdemann, H.-D. (1993). Naturwissenschaften, 80, 219–221. CrossRef CAS Web of Science Google Scholar
Knicker, H., Gonzalez-Vila, F. J., Polvillo, O., Gonzalez, J. A. & Almendros, G. (2005). Soil Biol. Biochem. 37, 701–718. Web of Science CrossRef CAS Google Scholar
Leinweber, P. & Schulten, H.-R. (1998). Soil Sci. Soc. Am. J. 62, 383–393. CrossRef CAS Google Scholar
Mitra-Kirtley, S., Mullins, O. C., Chen, J., van Elp, J., George, S. J., Chen, C. T., O'Halloran, T. & Cramer, S. P. (1992). Biochim. Biophys. Acta, 1132, 249–254. PubMed Web of Science Google Scholar
Mitra-Kirtley, S., Mullins, O., van Elp, J., George, S. J., Chen, J. & Cramer, S. P. (1993). J. Am. Chem. Soc. 115, 252–258. CrossRef CAS Web of Science Google Scholar
Mullins, O. C., Mitra-Kirtley, S., van Elp, J. & Cramer, S. P. (1993). Appl. Spectrosc. 47, 1268–1275. CrossRef CAS Web of Science Google Scholar
Pavlychev, A. A., Hallmeier, K. H., Hennig, C., Hennig, L. & Szargan, R. (1995). Chem. Phys. 201, 547–555. CrossRef CAS Web of Science Google Scholar
Pedras, M. S. C., Jha, M. & Ahiahonu, P. W. K. (2003). Curr. Org. Chem. 7, 1635–1647. Web of Science CrossRef CAS Google Scholar
Ravel, B. & Newville, M. (2005). J. Synchrotron Rad. 12, 537–541. Web of Science CrossRef CAS IUCr Journals Google Scholar
Ray, S. C., Pao, C. W., Chiou, J. W., Tasi, H. M., Jan, J. C. & Pong, W. F. (2005). J. Appl. Phys. 98, 033708. Web of Science CrossRef Google Scholar
Regier, T., Krochak, J., Sham, T. K., Hu, Y. F., Thompson, J. & Blyth, R. I. R. (2007). Nucl. Instrum. Methods A. In the press. Google Scholar
Regier, T., Paulsen, J., Wright, G., Coulthard, I., Tan, K., Sham, T. K. & Blyth, R. I. R. (2006). AIP Conf. Proc. 879, 473–476. CrossRef Google Scholar
Rodrigues, F., do Nascimento, G. M. & Santos, P. S. (2007). J. Electron Spectrosc. Relat. Phenom. 155, 148–154. Web of Science CrossRef CAS Google Scholar
Samuel, N. T., Lee, C.-Y., Gamble, L., Fischer, D. A. & Castner, D. G. (2006). J. Electron Spectrosc. Relat. Phenom. 152, 134–142. Web of Science CrossRef CAS Google Scholar
Schulten, H.-R. & Schnitzer, M. (1997). Biol. Fertil. Soils, 26, 1–15. Web of Science CrossRef Google Scholar
Schulten, H.-R., Sorge-Lewin, C. & Schnitzer, M. (1997). Biol. Fertil. Soils, 24, 249–254. CrossRef CAS Web of Science Google Scholar
Schulze, D. G. & Bertsch, P. M. (1995). Adv. Agron. 55, 1–66. CrossRef CAS Web of Science Google Scholar
Schwarzkopf, O., Borchert, M., Eggenstein, F., Flechsieg, U., Kalus, C., Lammert, H., Menthel, U., Pietsch, M., Reicherdt, G., Rotter, P., Senf, F., Zeschke, T. & Peatman, W. B. (1999). J. Electron Spectrosc. Relat. Phenom. 100–103, 997–1001. Web of Science CrossRef Google Scholar
Sjöberg, G., Knicker, H., Nilsson, S. I. & Berggern, D. (2004). Soil Biol. Biochem. 36, 609–618. Google Scholar
Somei, M. & Yamada, F. (2004). Nat. Prod. Rep. 21, 278–311. Web of Science CrossRef PubMed CAS Google Scholar
Stankiewicz, B. A. (1998). Nitrogen-Containing Macromolecules in the Bio- and Geosphere, pp. 1–12, ACS Symposium Series 707, edited by B. A. Stankiewicz and P. F van Bergen. Washington: American Chemical Society. Google Scholar
Stevenson, F. J. (1994). Humus Chemistry, Genesis, Composition, Reactions. 2nd ed. New York: John Wiley and Sons. Google Scholar
Stevenson, F. J. & Cole, M. A. (1999). Cycles of Soil. Carbon, Nitrogen, Phosphorus, Sulfur, Micronutrients, 2nd ed. New York: John Wiley and Sons. Google Scholar
Stöhr, J. (1992). NEXAFS Spectroscopy. Berlin: Springer. Google Scholar
Vairavamurthy, A. & Wang, S. (2002). Environ. Sci. Technol. 36, 3050–3056. Web of Science CrossRef PubMed CAS Google Scholar
Watenabe, N., Morais, J. & Alves, M. C. M. (2002). J. Phys. Chem. B, 106, 11102–11107. Google Scholar
Wilson, C. C. (1994). J. Chem. Crystallogr. 24, 371–373. CSD CrossRef CAS Web of Science Google Scholar
Wirick, S., Flynn, G. J., Keller, L. P. & Jacobsen, C. (2005). 67th Annual Meteoritical Society Meeting, Rio de Janeiro, Brazil. [https://www.lpi.usra.edu/meetings/metsoc2004/pdf/5092.pdf.] Google Scholar
Xiao, B., Boudou, J. P. & Thomas, K. M. (2005). Langmuir, 21, 3400–3409. Web of Science CrossRef PubMed CAS Google Scholar
Zhuang, S., Xu, M. & Hu, Z. (2006). J. For. Res. 17, 189–192. CrossRef CAS Google Scholar
Zubavichus, Y., Fuchs, O., Weinhardt, L., Heske, C., Umbach, E., Denlinger, J. D. & Grunze, M. (2004). Radiat. Res. 161, 346–358. Web of Science CrossRef PubMed CAS Google Scholar
Zubavichus, Y., Shaporenko, A., Grunze, M. & Zharnikov, M. (2005). J. Phys. Chem. A, 105, 6998–7000. Web of Science CrossRef Google Scholar
Zubavichus, Y., Zharnikov, M., Shaporenko, A., Fuchs, O., Weinhardt, L., Heske, C., Umbach, E., Denlinger, J. D. & Grunze, M. (2004). J. Phys. Chem. A, 108, 4557–4565. Web of Science CrossRef CAS Google Scholar
© International Union of Crystallography. Prior permission is not required to reproduce short quotations, tables and figures from this article, provided the original authors and source are cited. For more information, click here.
| JOURNAL OF SYNCHROTRON RADIATION |
