Knowledge of the chemical structure of soil organic matter and other polymers in the environment is rather limited, but would be required for a true understanding of soil formation, diagenesis, or retention of water and contaminants. Solid-state NMR provides unique opportunities for analyzing these otherwise nearly intractable systems.
We have been pursuing a worldwide unique program of developing solid-state NMR methods for characterizing the composition of such complex organic solids. By a suite of new NMR techniques, we can identify more than 36 functional groups (see Fig. 3), and determine their concentrations based on quantitative direct-polarization NMR. Information on the nanometer-scale structure is obtained by various spin-diffusion and relaxation experiments. The focus of these investigations is natural organic matter (NOM), as found in soils, swamps, rivers, oceans, or sediments, but the NMR methods are similarly applicable to many other organic solids. On the basis of greatly increased knowledge on the structure of soil organic matter, we are addressing questions of soil fertility, formation of natural organic matter, and the sorption of nonpolar contaminants in soil. We have used advanced NMR to study:
Figure. Top: Typical 13C MAS NMR spectrum of a humic acid, and the limited traditional peak assignments, with significant uncertainties due to overlap with other functional groups (dashed lines). Bottom: More than 30 functional groups that we can identify by advanced solid-state NMR techniques. Ovals refer to identification based on the 13C chemical-shift-anisotropy, dashed circles to dipolar dephasing.
Identification of functional groups. We have developed a set of spectral editing techniques for 13C NMR spectra of complex organic matter. With detection efficiencies sufficient for complex natural organic matter, we can selectively observe 13C NMR signals of:
In addition, we can select signals of CH vs. OH and NH protons (by 1H CSA recoupling). Pulse programs for these methods can be found on this website. Much of this work was supported by the National Science Foundation.
In combination with the 13C isotropic chemical shift, these advanced spectral-editing filters often enable unique identification of functional groups, see Fig. 3. Most of these signals can also be quantified, for instance by means of optimized direct-polarization experiments. Furthermore, we can identify esters, COOH groups, phenols, and aromatic ethers by means of various two-dimensional techniques. Various applications of these methods are briefly discussed in the following.
A likely molecular origin of a rice-yield decline. Intensive cropping of irrigated lowland rice for 20-30 years has led to significant declines in grain yield for high-yielding field trials. This yield decline has been attributed to decreased availability of soil nitrogen (N), which is mostly held in the soil organic matter. In collaboration with Dr. Dan Olk of the National Soil Tilth Laboratory in Ames, and formerly of the International Rice Research Institute (IRRI) in the Philippines, we have performed a detailed structural analysis of a humic acid fraction extracted from a continually submerged, triple-cropped rice soil, and a corresponding humic acid fraction extracted from an aerobic single-cropped rice soil for reference. In particular, we applied our recently developed SPIDER method for selecting the signals of carbons bonded to nitrogen. We have found significant amounts of N directly bonded to aromatic rings in the intensively cropped soil. Since N bonded to aromatics is not readily plant available, this observation may help explain the observed yield decline.
Our quantitative 13C NMR combined with spectral editing have shown that the triple-cropped soil humic acid is rich in lignin derivatives (>45% of all carbon). The chemical shift of the N-bonded aromatic carbons and the peak intensities indicate that these signals are due to amide groups bonded to lignin aromatic rings. In contrast, the single-cropped humic acid has a lower lignin content and showed peak intensities more characteristic of easily degradable peptides, and less N bonded to aromatic carbons. [Proc. Nat. Acad. Sci. 101, 6351-6354 (2004)]
Nearly quantitative 13C NMR of untreated mineral soil. Do the amounts of recalcitrant components of soil organic matter (SOM) vary across the landscape position? To address this question, Dr. Michael Thompson in the Agronomy department at ISU selected four Mollisols in central Iowa for study by NMR. Spin counting by correlation of the integral NMR intensity with the C concentration by elemental analysis showed that our NMR spectra were ≥ 85% quantitative for the majority of the samples studied. For untreated whole-soil samples with <2.5 wt% C, which is considerably less than in most previous quantitative NMR analyses of SOM, useful spectra that reflected ≥65% of all C were obtained. The NMR analyses allowed us to conclude (1) that the HF treatment (with or without heat) had low impact on the organic C composition in the samples, except for protonating carboxylate anions to carboxylic acids, (2) that most organic C was observable by NMR even in untreated soil materials, (3) that esters were likely to compose only a minor fraction of SOM in these Mollisols, and (4) that the aromatic components of SOM were enriched to ~53% in the poorly drained soils, compared with ~48% in the well-drained soils; in plant tissue and particulate organic matter (POM) the aromaticities were ~18% and ~ 32%, respectively. Nonpolar, non-protonated aromatic C, interpreted as a proxy for charcoal C, dominated the aromatic C in all soil samples, composing 69 - 78% of aromatic C and 27 - 36% of total organic C in the whole soil and clay-fraction samples. [Geochim, Cosmochim. Acta, accepted for publication]
Biochar. Pyrolysis or gasification of biomass may economically produce energy and chemicals from a wide range of biorenewable resources. These processes yield some amount of char, typically between 5 and 20% of feedstock mass, which may be used as biochars, i.e. applied as a soil amendment and/or a carbon sequestration agent, as in the dark, fertile terra preta soils in the Amazon, which have been shown to contain man-made charcoal functioning as soil organic matter. The link between char properties and their efficacy in soils, however, is not well understood, much less how to engineer the process conditions to produce desired biochar properties. This is especially true for chars from gasification and fast pyrolysis.
A key aspect of determining char quality for biochar and other applications is the ability to quantitatively characterize the forms of carbon present. In particular, concern has been expressed about "incompletely" pyrolyzed biomass as it may provide too much bio-available carbon to the soil without enough simultaneous nitrogen, resulting in nitrogen immobilization and therefore, negative short-term effects on plant yield. In collaboration with Catherine Brewer and Dr. Robert Brown on campus, we are studying biochars by NMR. Direct-polarization magic-angle spinning is the only available method for reliably determining the aromaticity and composition of chars. Further, it can be combined with dipolar dephasing to quantify the fraction of non-protonated aromatic C. The size of clusters of fused aromatic rings in chars can also be determined based on complementary NMR methods of spectral analysis and 1H-13C dipolar distance estimates. [Environ. Progr. Sustain. Energy 28, 386, 2009]
Products of the Maillard reaction. The Maillard reaction between reducing sugars and amino acids is important in food science (like baking cookies: reaction of sugar and egg protein) and has also been considered as a potential humification pathway in soils. In the melanoidin polymers formed, the sugar is completely transformed into complex structures resembling humic acids. The exciting aspect about this "artificial" complex organic matter is the possibility of introducing specific 13C and 15N isotopic label, which enable NMR structural studies in unprecedented detail. In particular, the fate of the amino acid in the model Maillard reaction between glucose and glycine in a 1:1 molar ratio has been investigated by applying advanced 13C and 15N solid-state nuclear magnetic resonance (NMR) techniques to 13C- and 15N-labeled melanoidins formed in dry and solution reactions. Quantitative 13C NMR shows that ~23% of carbon is from glycine; the ~2% loss compared to the 25% glycine C in the reactants is due to the COO moiety being liberated as CO2 (Strecker degradation). 13C J-modulation experiments on melanoidins made from doubly13C-labeled glycine show that the C-C backbone bond of ~2/3 of the incorporated amino acid stays intact, and about half of all glycine is incorporated as N-CH2-COO without fragmentation. Degradation processes without CO2 loss affect ~1/8 of glycine in dry reaction, and ~1/4 in solution. These results indicate that Strecker degradation affects ~1/4 (dry reaction) to 1/3 (in solution) of all glycine but is not the main pathway of glycine incorporation. Spectra of Strecker degradation products show that C2 of glycine reacts to form N-CH3, C-CHn-C, or aromatic units. The gycine-C1 carbon incorporated into the melanoidins remains ≥95% part of COO moieties; ~5% of amides are also detected. The C2-N bond stays intact for ~70% of the incorporated glycine. The 15N spectra show many peaks, over a 200-ppm range, documenting a multitude of different chemical environments of nitrogen. The majority (>78 %) of nitrogen, in particular most pyrrolic N, is not protonated. Since N-H predominates in amino acids and proteins, nonprotonated nitrogen may be a characteristic marker of Maillard reaction products. [J. Agri. Food Chem., accepted for publication]