A simple densimetric method for the separation of free and occluded particulate organic materials was developed and applied to five virgin soils. The free organic matter was isolated by suspending the soil in sodium polytungstate solution (d = 1.6 Mg m-3) and decanting the light material. The remaining soil was disaggregated by sonification for liberation of occluded organic materials. The free light fraction consisted of large, undecomposed or partly decomposed root and plant fragments. This fraction comprised 0.59-4.34% of soil dry weight and accounted for 6.9-31.3% and 5.9-22.1% of total soil carbon and nitrogen respectively. Identifiable components of the occluded fraction were small particles of incompletely decomposed organic residues, pollen grains, particles of plant tissue such as lignin coils and phytoliths. This fraction comprised 0.69-1.81% of soil dry weight and represented 9.2-17.5% and 6.2-14.1% of the total soil carbon and nitrogen. The proportion of soil organic carbon recovered as the occluded fraction was high in soils with high clay contents. The chemical composition of occluded and free organic materials was investigated by solid-state 13C CP/MAS NMR spectroscopy. Despite the differences in soils, environmental conditions and vegetation, the organic structure of the free light fraction was similar in four of the five soils. This fraction consisted of 55-63% O-alkyl C, 18-25% alkyl C, 14-18% aromatic C, and 5-7% carbonyl C. In the other soil, this fraction showed a higher proportion of alkyl C (31%) and lower O-alkyl C (46%). Most of the differences between soils were associated with organic materials contained in the occluded light fraction. The differences in chemical structure between the occluded light fraction and free light fractions were similar in all examined soils. The NMR data showed that the proportion of O-alkyl C was lower and alkyl C higher in the occluded light fractions than in the free light fractions. The proportion of aromatic and carbonyl carbon was higher in the occluded fractions of three soils while the percentage of these two types of carbon remained unchanged in the two other soils. It is considered that the occluded organic matter is an old pool of carbon that has been accreted within aggregates during decades of root growth and it is that pool which is lost due to cultivation.

Solid-state 13C nuclear magnetic resonance (NMR) spectroscopy has become an important tool for examining the chemical structure of natural organic materials and the chemical changes associated with decomposition. In this paper, solid-state 13C NMR data pertaining to changes in the chemical composition of a diverse range of natural organic materials, including wood, peat, composts, forest litter layers, and organic materials in surface layers of mineral soils, were reviewed with the objective of deriving an index of the extent of decomposition of such organic materials based on changes in chemical composition. Chemical changes associated with the decomposition of wood varied considerably and were dependent on a strong interaction between the species of wood examined and the species composition of the microbial decomposer community, making the derivation of a single general index applicable to wood decomposition unlikely. For the remaining forms of natural organic residues, decomposition was almost always associated with an increased content of alkyl C and a decreased content of O-alkyl C. The concomitant increase and decrease in alkyl and O-alkyl C contents, respectively, suggested that the ratio of alkyl to O-alkyl carbon (A/O-A ratio) may provide a sensitive index of the extent of decomposition. Contrary to the traditional view that humic substances with an aromatic core accumulate as decomposition proceeds, changes in the aromatic region were variable and suggested a relationship with the activity of lignin-degrading fungi. The A/O-A ratio did appear to provide a sensitive index of extent of decomposition provided that its use was restricted to situations where the organic materials were derived from a common starting material. In addition, the potential for adsorption of highly decomposable materials on mineral soil surfaces and the impacts which such an adsorption may have on bioavailability required consideration when the A/O-A ratio was used to assess the extent of decomposition of organic materials found in mineral soils.

The nature of organic carbon in the <2, 2-20, 20-53, 53-200, and 200-2000 pm fractions of four surface soils was determined using solid state 13C nuclear magnetic resonance (n.m.r.) spectroscopy with cross polarisation and magic angle spinning (CP/MAS). Analyses were repeated after high energy ultraviolet photo-oxidation was performed on the three finest fractions. All four soils studied contained appreciable amounts of physically protected carbon while three of the soils contained even higher amounts of charcoal. It was not possible to measure the charcoal content of soils directly, however, after photo-oxidation, charcoal remained and was identified by its wood-like morphology revealed by scanning electron microscopy (SEM) together with a highly aromatic chemistry determined by solid state 13C n.m.r. Charcoal appears to be the major contributor to the 130 ppm band seen in the n.m.r. spectra of many Australian soils. By using the aromatic region in the n.m.r. spectra, an approximate assessment of the charcoal distribution through the size fractions demonstrated that more than 88% of the charcoal present in two of the soils occurred in the <53 pm fractions. These soils contained up to 0.8 g C as charcoal per 100 g of soil and up to 30% of the soil carbon as charcoal. Humic acid extractions performed on soil fractions before and after photo-oxidation suggest that charcoal or charcoal-derived material may also contribute significantly to the aromatic signals found in the n.m.r. spectra of humic acids. Finely divided charcoal appears to be a major constituent of many Australian soils and probably contributes significantly to the inert or passive organic carbon pool recognised in carbon turnover models.

Samples from the surface horizons of six virgin soils were collected and separated into density fractions. Based on the spatial distribution of organic materials within the mineral matrix of soil, the soil organic matter (SOM) contained in various density fractions was classified as: (a) free particulate OM, (b) occluded particulate OM, and (c) colloidal or clay-associated OM. The compositional differences noted among these three components of SOM were used to describe the changes that OM undergoes during decomposition when it enters the soil, is enveloped in aggregates and eventually is incorporated into microbial biomass and metabolites and associated with clay minerals. The occluded organic materials, released as a result of aggregate disruption, were in various stages of decomposition and had different degrees of association with mineral particles. Changes in the degree of association of occluded organic materials and mineral particles with decomposition are discussed and form the basis of a model which illustrates the simultaneous dynamics of microaggregates and their organic cores. This model indicates a major role for carbohydrate-rich plant debris in formation and stabilization of microaggregates.

Five surface soils of differing chemical and mineralogical compositions were subjected to either a sequence of dithionitelcitrate extractions in which the soil: citrate ratio was varied or to a sequence of 1% or 2% HF extractions. The 2% HF treatment resulted in the removal of the highest Fe concentrations (79-95%) while the dithionitelcitrate extractions were less effective in removing Fe from the same soils (18-75%). The Fe remaining after HF treatment appeared to be mostly associated with ilmenite crystals which were only slowly attacked by the dilute acid. During the 2% HF treatments, some organic carbon was lost (8-17%). but this loss had no significant effect on the organic chemistry of the samples as determined by CP/MAS 13C n.m.r. spectroscopy.The total 13c signal recovered after the various treatments was found to be correlated, in order of decreasing significance, with the mineral fraction present in the sample, the organic carbon/Fe ratio and the mass magnetic susceptibility. The expression (organic carbon/Fe) -0.147(mineral fraction present in the sample) +0.043(l/mass magnetic susceptibility), accounted for 85.3% of the variation in the relative visibility of the 13C signal.Prior to solid state CP/MAS 13c n.m.r. analysis, the recommended pretreatment for surface soils containing Fe involves a sequence of 2% HF extractions in the order 5x1 h, 2x16 h and 1x64 h. For soils high in Fe, this procedure allows CP/MAS 13c n.m.r. spectra to be acquired that would otherwise be difficult to obtain. It also results in a significant increase in sensitivity and in resolution of the 13c n.m.r. spectra of soils with moderate Fe contents.