The influence of humus fractionation on the chemical composition of soil organic matter studied by solid-state 13C NMR

Four samples of soil organic matter and their humic acids, fulvic acids and humin were studied with solid-state ¹³CP MAS NMR. The whole soil samples were fractionated using NaOH and HCl in order to extract humic acids, fulvic acids and humin. This investigation indicates that conventional humus fractionation does not significantly change the content of different functional groups in soil.     

Quantitative solid-state 13C NMR spectroscopy of organic matter fractions in lowland rice soils

Spin counting on solid‐state ¹³C cross‐polarization (CP) nuclear magnetic resonance (NMR) spectra of two humic fractions isolated from tropical lowland soils showed that only 32–81% of potential ¹³C NMR signal was detected. The observability of ¹³C NMR signal (C_obs) was higher in the mobile humic acid (MHA) than in the calcium humate (CaHA) fraction, and increased with increasing intensity of irrigated rice cropping. NMR observability appeared to be related to the nature of the organic carbon, with phenol‐ and methoxyl‐rich samples having the higher values of C_obs. The Bloch decay (BD) technique provided more quantitatively reliable ¹³C NMR spectra, as evidenced by values of C_obs in the range 91–100% for seven of the eight humic fractions studied. The BD spectra contained considerably more aryl and carbonyl signal, and less O–alkyl and alkyl signal, with the greatest differences between CP and BD spectra observed for the samples with low C_obs(CP). The causes of low CP observability were investigated using the spectral editing technique RESTORE ( RE storation of S pectra via T _CH and T O ne R ho (T_1ρH) E diting). Rapid T_1ρH relaxation was found to be primarily responsible for the under‐representation of carbonyl carbon, whereas inefficient cross‐polarization was primarily responsible for the under‐representation of aryl carbon in CP spectra. Proton NMR relaxation rates T_1H and T_1ρH were found to correlate with other NMR properties and also with cropping management. Non‐uniform rates of T_1H relaxation in two of the CaHA fractions enabled the generation of proton spin relaxation editing subspectra.     

13C solid-state NMR assessment of decomposition pattern during co-composting of sewage sludge and green wastes

The fate of organic matter during composting is poorly understood. Therefore, we analysed composts of sewage sludges and green wastes (44 samples representative of 11 stages of biodegradation) by conventional chemical methods: pH, humic (HA) and fulvic acid (FA) content, C, N and organic matter (OM) content, and by ¹³C CPMAS NMR to assess the decomposition process of the organic matter. Chemical changes clearly occurred in two phases: first, decomposition of OM during the first 2 months was characterized by decreased C/N ratios, OM content and increased pH; and second, a humification process with increased HA/FA ratios. NMR spectrum changes confirmed this pattern, with an increase in aromaticity and a decrease in alkyl C. A decrease of syringyl to guaiacyl ratio (S/G), a sign of lignin transformation, also indicated humification during composting. NMR spectroscopic properties of composts were also studied by means of principal components analysis (PCA) and revealed changes according to the degree of compost maturation. The factorial map presents a chronological distribution of composts on the two first principal components. The influences of eight chemical factors on the PCA ordination of composts as monitored by their evolution by NMR were also studied by multivariate analyses. PCA clearly indicated two phases: the rapid decomposition of organic matter followed by the formation of humic‐like substances. The first phase, that is ‘new’ composts, was strongly correlated with OM contents, pH and C/N ratios whereas the second phase, corresponding to ‘old’ compost, was correlated with pH, HA content and HA/FA ratio. These results confirm that knowledge of the formation of humic substances is indispensable to suitable monitoring of the composting process.     

Following the decomposition of ryegrass labelled with 13C and 15N in soil by solid-state nuclear magnetic resonance spectroscopy

Investigating the biogeochemistry of plant material decomposition in soil has been restricted by difficulties extracting and identifying organic compounds. In this study the decomposition of ¹³C- and ¹⁵N-labelled Lolium perenne leaves mixed with mineral soil has been investigated over 224 days of incubation under laboratory conditions. Decomposition was followed using short-term rates of CO₂ evolution, the amounts of ¹³C and ¹⁵N remaining were determined by mass spectrometry, and ¹³C and ¹⁵N solid-state nuclear magnetic resonance (NMR) spectroscopy was used to characterize chemically the plant material as it decomposed. After 224 days 48% of the added ¹³C had been lost with a rapid period of C0₂ evolution over the first 56 days. The fraction of cross-polarization magic angle spinning (CP MAS) ¹³C NMR spectra represented by O-alkyl-C signal probably in carbohydrates (chemical shift, 60–90 p.p.m.) declined from 60 to 20% of the spectrum (chemical shift, 0–200 p.p.m.) over 224 days. The rate of decline of the total ¹³C exceeded that of the 60–90 p.p.m. signal during the first 56 days and was similar thereafter. The fraction of the CP MAS ¹³C NMR spectra represented by the alkyl- and methyl-C (chemical shift, 10–45 p.p.m.) signal increased from 5 to 14% over the first 14 days and was 19% after 224 days. CP MAS ¹³C NMR of ¹³C- and ¹⁵N-L. perenne contained in 100-μm aperture mesh bags incubated in the soil for 56 days indicated that the remaining material was mainly carbohydrate but there was an increase in the alkyl- and methyl-C associated with the bag's contents. After 224 days incubation of the labelled ¹³C- and ¹⁵N-L. perenne mixed with the soil, 40% of the added N had been lost. Throughout the incubation there was only one signal centred around 100 p.p.m. detectable in the CP MAS ¹⁵N NMR spectra. This signal corresponded to amide ¹⁵N in peptides and may have been of plant or microbial origin or both. Although there had been substantial interaction between the added ¹⁵N and the soil microorganisms, the associated redistribution of ¹⁵N from plant to microbial tissues occurred within the amide region. The feasibility of following some of the component processes of plant material decomposition in soil using NMR has been demonstrated in this study and evidence that microbial synthesis contributes to the increase in alkyl- and methyl-C content of soil during decomposition has been represented.     

13C NMR characterization of humic acids from soils of a development sequence

Humic acids were isolated from nine topsoils in New Zealand tussock grasslands. Cross-polarization ¹³C NMR spectra of solid samples were used to estimate fractions of carbon contained in different types of chemical functional groups. The degree of oxygen substitution of aromatic rings showed a strong negative correlation with soil development. Aromaticities greater than 0.25 were found in humic acids from only the two least-developed soils.     

13C NMR of humic substances: pH and solvent effects

Since the concentration of free radicals in humic subtances increases at high pH the use of basic solutions for ¹³C NMR spectroscopy may cause broadening and loss of aromatic signals, with distortion of intensity distributions. No such effects were found in ¹³C spectra of soil humic and fulvic acid, an aquatic fulvic acid, and two phenolic polymers run in aqueous solutions at different pH values, and in dimethylsulphoxide. With increasing pH, the peak in the carboxyl region shifted in a manner consistent with greater dissociation of carboxyl and phenolic groups, and also certain features in the aliphatic and carboxyl regions were enhanced under some solution conditions. Elevated solution temperature (70°C) caused only slight improvement in the resolution of some lines. Chemical shifts were determined for some known phenolic and benzenecarboxylic acid compounds in DMSO and NaOD. The range for phenolic carbons extended to 173 ppm in NaOD, while some aromatic carbons occurred around 105 ppm, in the same region as anomeric carbons. Thus, even under quantitative acquisition conditions, relative areas may be used only to estimate proportions of different types of carbons and functional groups.     

The effect of water content on solid-state 13C NMR quantitation and relaxation rates of soil organic matter

Two hydrofluoric acid‐treated soils were prepared with water contents ranging up to 22% by exposing them to a range of atmospheric humidities. There was no effect of water content on the chemical shift distribution of nuclear magnetic resonance (NMR) signal in ¹³C cross‐polarization (CP) NMR spectra. The sensitivity of the ¹³C CP NMR spectra decreased slightly with increasing water content. Much of this decrease could be attributed to decreases in T_1ρH relaxation rates, caused by enhanced molecular mobility of the organic matter in the presence of absorbed water. Rates of T_1H relaxation were very sensitive to water content, and average T_1H relaxation rates decreased four‐ to five‐fold from the smallest to the largest water content. Rates of T_1H relaxation were non‐uniform, and were better modelled by two‐T_1H component fits than one‐T_1H component fits. The ratio of rapidly to slowly relaxing components increased with increasing water content. Proton spin relaxation editing (PSRE) subspectra revealed substantial changes in the nature of these two components with increasing water content. These results indicate the presence of an organic matter component that is very sensitive to water content, transforming from slowly relaxing at a small water content to rapidly relaxing at a greater water content. This component was shown to be rich in O–alkyl and carbonyl C, and may be hemicellulosic root exudates and microbial mucilages. The slowly relaxing PSRE component was a mixture of ligno‐cellulose and alkyl biopolymers, whereas the rapidly relaxing component was primarily charcoal for one of the soils, and was reminiscent of dissolved organic carbon for the other soil. These findings show that care must be taken in controlling water contents when using PSRE to study organic matter.     

13C and 15N natural abundances of urban soils and herbaceous vegetation in Karlsruhe, Germany

We undertook what we believe to be a unique survey of the natural abundances of ¹³C and ¹⁵N in urban soils and plants in Karlsruhe (Germany), a European city of average size. We found broad patterns of these abundances in both soils and plants, which reflected geology and land use. In contrast with studies on smaller areas (showing the direct effect of human activities), our study first determined the extent to which the abundances correlated with land use or underlying geology and then assessed how we could further test such relationships. The spatial pattern of δ¹³C in surface soil correlated with that of the underlying parent material; construction activities superimposed a secondary signal. Maize cultivation was a source of less negative soil δ¹³C, whereas the C₃ vegetation is a source of more negative soil δ¹³C. There was a footprint of less negative plant δ¹³C in the industrial and port areas; plant δ¹³C downwind of the city was less negative than upwind, which might relate to atmospheric pollution from the port area or to differences in soil properties. There was no significant effect of wind direction or geology on soil or plant δ¹⁵N, which was correlated mainly with land use. The largest soil δ¹⁵N was under agriculture and the smallest under woodland. The abundance of ¹⁵N in inner-urban soil and plants was intermediate between those of agriculture and forests. This study represents a major advance in the use of stable isotope geochemistry in understanding urban environments.     

The dynamics of organic matter in rock fragments in soil investigated by 14C dating and measurements of 13C

Rock fragments in soil can contain significant amounts of organic carbon. We investigated the nature and dynamics of organic matter in rock fragments in the upper horizons of a forest soil derived from sandstone and compared them with the fine earth fraction (<2 mm). The organic C content and its distribution among humic, humin and non‐humic fractions, as well as the isotopic signatures (Δ¹⁴C and δ¹³C) of organic carbon and of CO₂ produced during incubation of samples, all show that altered rock fragments contain a dynamic component of the carbon cycle. Rock fragments, especially the highly altered ones, contributed 4.5% to the total organic C content in the soil. The bulk organic matter in both fine earth and highly altered rock fragments in the A1 horizon contained significant amounts of recent C (bomb ¹⁴C), indicating that most of this C is cycled quickly in both fractions. In the A horizons, the mean residence times of humic substances from highly altered rock fragments were shorter than those of the humic substances isolated in the fine earth. Values of Δ¹⁴C of the CO₂ produced during basal respiration confirmed the heterogeneity, complexity and dynamic nature of the organic matter of these rock fragments. The weak ¹⁴C signatures of humic substances from the slightly altered rock fragments confirmed the importance of weathering in establishing and improving the interactions between rock fragments and surrounding soil. The progressive enrichment in ¹³C from components with high‐¹⁴C (more recent) to low‐¹⁴C (older) indicated that biological activity occurred in both the fine and the coarse fractions. Hence the microflora utilizes energy sources contained in all the soil compartments, and rock fragments are chemically and biologically active in soil, where they form a continuum with the fine earth.     

The influence of vegetational origin and degree of humification of organic soils on their chemical composition, determined by solid-state 13C NMR

Organic soil samples with different vegetational background and others with variation in the degree of humification, were investigated with solid-state ¹³C NMR. This work indicates that the vegetational origin and degree of humification of the organic matter appear to influence the distribution of functional groups in organic soils considerably, but one year of decomposition under controlled laboratory conditions gave only small changes in the chemical composition.     

13C NMR studies of organic matter in whole soils: I. Quantitation possibilities

A detailed discussion of the quantitative nature of ¹³C CPMAS NMR spectra as applied to solid samples, such as soil, is given. In particular, the influence of the cross-polarization (CP) time constant (T_CH), the relaxation time constant of protons in the rotating frame (T_1pH) and the contact time (t_c) in the CPMAS experiment are considered. Three distinct quantitation regimes are numerically identified according to sample parameters t_CH and T_1PH > and the experimental choice of t_c: (i) quantitation obtainable from a single CPMAS spectrum; (ii) quantitation obtainable from a series of CPMAS spectra; and (iii) quantitation not possible using CPMAS. Strategies for the measurement of sample parameters T_CH and Ti_pH are reviewed. When quantitation is not possible using CPMAS it is necessary to regress to the direct polarization (DP) of ¹³C nuclei. The sensitivity problems of DPMAS are discussed, as too are general factors that affect the quantitation of ¹³C data such as spinning sidebands. More specifically in relation to soil samples, the effects on quantitation arising from the presence of paramagnetics and the actual methods for the measurement of signal intensities are covered.     

Spin accounting and RESTORE – two new methods to improve quantitation in solid-state 13C NMR analysis of soil organic matter

Rapid T_1ρH relaxation and inefficient cross‐polarization have long been known to affect quantitation in solid‐state ¹³C cross‐polarization (CP) NMR spectra of soil organic matter. We have developed two new techniques to overcome these problems. The first, spin accounting, enables accurate gauging of how quantitative a spectrum is likely to be. The result is expressed as the percentage of potential NMR signal that can be accounted for (C_obs). Spin accounting improves on the established spin counting technique by correcting for rapid T_1ρH relaxation and inefficient cross‐polarization. Spin accounting identifies three components: one that is well represented in CP spectra, one that is under‐represented in CP spectra due to rapid T_1ρH relaxation, and one that is under‐represented in CP spectra due to inefficient cross‐polarization. For a range of eight de‐ashed soils, C_obs was in the range 83–106%, indicating that virtually all potential signal could be accounted for after correcting for rapid T_1ρH relaxation and inefficient cross‐polarization. The second new technique, RESTORE (RE storation of S pectra via T _CH and T O ne R ho (T_1ρH) E diting), generates subspectra for the three components identified in spin accounting. The sum of the three RESTORE subspectra is essentially a corrected CP spectrum. The RESTORE spectra of all eight soils more closely resembled the corresponding, and presumably quantitative, Bloch decay spectra than did the CP spectra. RESTORE identifies the types of structures underestimated by CP, and the cause of their underestimation. Rapid T_1ρH relaxation most affected carbonyl and carbohydrate carbons, whereas inefficient cross‐polarization most affected aromatic carbons.     

Interactions between peat and sodium acetate, ammonium sulphate urea or wheat straw during incubation studied by 13C and 15N NMR spectroscopy

Transformations of sodium acetate, ammonium sulphate, urea and wheat straw in peat have been studied by determining the distribution of ¹⁵N-labelled material, and by ¹³C and ¹⁵N nuclear magnetic resonance spectroscopy (NMR) using cross polarization (CP) and magic-angle spinning (MAS). Samples of an oligotrophic blanket peat were incubated for 6 months at 15°C with ¹⁵N ammonium sulphate, ¹⁵N urea, ¹³C¹⁵N urea, ¹⁵N-labelled wheat straw or ¹³C sodium acetate. The incubated samples were separated into fractions of >1 mm, 1–0.5 mm, 0.5–0.25 mm, 0.25–0.15 mm, 0.15–0.05 mm, 0.05–0.005 mm and a water-soluble fraction by wet sieving, and were then freeze-dried. The distribution of ¹⁵N between the fractions was obtained after isotope-ratio analysis by mass spectrometry, and the 0.5–0.25 mm, 0.05–0.005 mm and water-soluble fractions from the incubations were examined by ¹³C and ¹⁵N NMR. ¹³C-labelled acetate increased carbohydrate resonances in the 0.05–0.005 mm and soluble material, but an organic acid derived from the substrate was still present 6 months later. Incorporation of ¹⁵N from ammonium sulphate into the peat was low, and more than 50% of the added N was detected in the soluble fraction still present as ¹⁵NH+₄. As carbohydrate and soluble organic matter were detected in the peat, it was concluded that microbial activity and N immobilization were restricted by poor aeration and low pH. Urea, in contrast, interacted with all the fractions examined, with some ¹⁵N being incorporated into a range of compounds that included protein, peptides, amides, amino acids and carbamates or lactam derivatives. A small proportion of labelled ¹⁵N from wheat straw, orginally present in the > 1 mm and 14.5 mm fractions, had moved into the 0.05–0.005 fraction during incubation and sieving. The ¹³C spectra suggested that the presence of the straw may have stimulated decomposition of the peat components.     

Improvement of 13C and 15N CPMAS NMR spectra of bulk soils, particle size fractions and organic material by treatment with 10% hydrofluoric acid

The small organic matter content of mineral soils makes it difficult to obtain ¹³C and ¹⁵N nuclear magnetic resonance (NMR) spectra with acceptable signal-to-noise ratios. Subjecting such samples to hydrofluoric acid removes mineral matter and leads to a relative increase in organic material. The effect of treatment with 10% hydrofluoric acid on bulk chemical composition and resolution of solid-state ¹³C NMR spectra was investigated with six soils, some associated particle size fractions, plant litter and compost. The treatment enhanced the signal-to-noise ratio of the solid-state ¹³C NMR spectra. The improvement in spectrum quality was greatest in the clay fraction of soil contaminated with coal ash. The removal of paramagnetic compounds associated with the ash may be the main reason for the improvement. Based on total C, total N, C/N ratio and intensity distribution of the solid-state ¹³C NMR spectra, no changes in organic matter composition could be detected, except for a possible loss of carbohydrates. After treatment with HF, solid-state ¹⁵N NMR spectra of particle size fractions were obtained and indicated that the observable nitrogen is present mostly as peptides and free amino groups. Extraction with hydrofluoric acid is recommended as a routine treatment prior to solid-state ¹³C and ¹⁵N NMR on soil containing little C or N and soil samples containing paramagnetic compounds from natural or anthropogenic sources.     

Characteristic alterations of quantity and quality of soil organic matter caused by forest fires in continental Mediterranean ecosystems: a solid-state 13C NMR study

The variable effect of different types of forest fires on the quantity and quality of soil organic matter (SOM) was analysed by comparing burnt and unburnt soils from six forest ecosystems in central Spain by organic elemental analysis and solid‐state ¹³C nuclear magnetic resonance (NMR) spectroscopy. Whole soil samples were collected 1 to 2 years after the fires and included one site affected by two fires within 2 years. The fire‐affected soils showed no common pattern with respect to the amount of additional carbon (C_add) but at all sites, the fire enhanced the aromatic‐C content. The weakest fire intensity resulted in the greatest aromatic‐C enrichment factor, EF_{I(aromatic C)} indicating the greatest local accumulation of char. The respective C_add disclosed an EF_{I(aromatic C)} to EF_{I(alkyl C)} ratio, B_char, of c.1, which supports a small degree of charring. Extensive combustion and volatilization at stronger fire intensities yielded a decrease of EF_{I(aromatic C)} and an increase in B_char. These trends are in good agreement with fire intensity and forest fuel combustibility in the various sites and therefore these indices could be used to elucidate the quality and quantity of char input that occurs during and after forest fires. No ¹³C NMR evidence for substantial inputs from non‐charred plant necromass was found for any of the single‐burn soils. The large carboxyl‐C content of C_add is evidence of the occurrence of oxidation reactions very shortly after the fire. In comparing the single and double‐burn sites, no additional char input was observed for the double‐burn site, possibly because of complete combustion of young shrubs and char remains during the second fire. The large O‐alkyl‐C portion found in C_add of the double‐burn soil is best explained by decreased litter degradation.     

Sludge-derived organic carbon in an agricultural soil estimated by 13C abundance measurements

The objective of this study is to develop a method to follow the dynamics of sludge‐derived organic carbon, which will allow us to understand the behaviour of trace metals in the sludge‐treated soils. We studied, in a sandy agricultural soil of southwest France, cultivated with maize and amended with sewage‐sludge over 20 years, the dynamics of different sources of organic matter and compared this with a control, which had never received any treatment. For the first time, a method is proposed that will distinguish and quantify sludge‐derived organic carbon, maize‐derived organic carbon, and native organic carbon. This method is based on the mean differences in δ¹³C abundances between native (−26.5‰), maize (−12.5‰) and sludge (−25.4‰) organic carbon. Three hypotheses on the dynamics of soil organic matter sources are proposed: (i) isotopic differences observed between control and sludge‐treated soils are due only to the incorporation of sludge C, whereas in the others, the control was used to model the incorporation of (ii) maize C or (iii) native C in the sludge‐treated soils. The comparison of the stocks of each source (native C, maize C and sludge C) found in the bulk soil with the sum of corresponding stocks found in particle‐size fractions allowed us to reject the two first hypotheses and to validate the last one. Repeated applications of sewage‐sludge induced accumulation of sludge‐derived organic carbon in the topsoil, and simultaneously contributed to the preservation of maize‐derived organic carbon. When sludge applications ceased, the rapid decrease in soil organic matter stocks was mostly caused by the degradation of the sludge‐derived organic carbon sources. At the same time, the maize‐derived organic carbon shifted from the coarsest fraction (200–2000 μm) to the finest fraction (0–50 μm). Therefore, this study has shown that repeated applications of sewage‐sludge induced changes in soil organic matter dynamics over time.     

Transformation and binding of 13C and 14C-labelled atrazine in relation to straw decomposition in soil

As a source of organic matter, crop residues affect the behaviour of pesticides in agricultural soils. The fate of [U‐ring‐¹³C] and [U‐ring‐¹⁴C] atrazine (6‐chloro‐N‐ethyl‐N‐isopropyl‐1,3,5‐triazine‐2,4‐diamine) was investigated during laboratory incubation under controlled conditions in a loamy soil amended with wheat straw at two different states of decomposition: no preliminary decomposition or 6 months’ preliminary decomposition. After 3 months, non‐extractable, so‐called ‘bound’, ¹³C‐atrazine residues were recovered in three particle‐size fractions (> 200, 50–200 and < 50 μm), and investigated with solid‐state ¹³C‐NMR spectroscopy. Parallel incubations with [U‐ring‐¹⁴C] atrazine were carried out to quantify the bound residues as well as the extractable and mineralized fractions. The effect of straw residues on atrazine behaviour depended on whether they had been previously decomposed or not. When straw was decomposed for 6 months prior to incubation, atrazine mineralization was enhanced to 50% of the initial ¹⁴C in contrast to 15% of the initial ¹⁴C in soil alone and soil amended with fresh straw. In parallel, atrazine bound residues were formed in greater amount representing up to 20% of the initial ¹⁴C. CP/MAS ¹³C‐NMR on soil size fractions of soil–straw mixtures after incubation with ¹³C‐atrazine showed that bound residues contained mostly triazinic C, corresponding to atrazine or primary metabolites. Non‐humified organic materials recovered in size fractions > 200 and 50–200 μm contained significant amounts of bound residues, especially when straw was added to the soil. CP/MAS ¹³C‐NMR analysis of humic acids obtained from < 50‐μm fractions was difficult due to overlapping of the native carboxyl ¹³C signal with the ¹³C‐atrazine signal.     

PROTON-DECOUPLED PULSE FOURIER-TRANSFORM 13C MAGNETIC RESONANCE OF SOIL ORGANIC MATTER

The structure of some humic acids, fulvic acids and lipid fractions of soil organic matter have been investigated by ¹³C-n.m.r. spectroscopy. Signals from aromatic groupings were normally absent in humic and fulvic acids. There were differences in spectra from extracts prepared from different sources or by different techniques. ^{1 3}C-n.m.r. spectra of lipid fractions were generally similar to each other but there were some distinct differences that may be used to a ‘fingerprint’ the lipids.     

Phosphate sorption by synthetic amorphous aluminium hydroxides: a 27Al and 31P solid-state MAS NMR spectroscopy study

The sorption of phosphate on amorphous aluminium hydroxides was investigated using ²⁷Al and ⁷¹P solid-state magic-angle spinning nuclear magnetic resonance (MAS NMR) spectroscopy, following the effect of different exposures to soluble phosphate. The spectra obtained were compared with the spectrum of amorphous aluminium phosphate. Aluminium in the unreacted hydroxide had a 100% octahedral co-ordination. When dried at 200°C and exposed to soluble phosphate, very little (maximum 0.1%) amorphous aluminium hydroxide transformed to a tetrahedral co-ordination (A1 bound by oxygen bridges to four P atoms), even after 120d. The tetrahedral co-ordination exists in aluminium phosphate gel, although most of its A1 atoms exhibit an octahedral co-ordination. For the aluminium hydroxide dried at 200°C, no formation of aluminium phosphate in which aluminium is in octahedral co-ordination could be detected, not even when the aluminium hydroxide was exposed to a phosphate solution for 120 d. We concluded that the formation of aluminium phosphate is restricted to the surface of the hydroxide. Most of the phosphate which is bound to the aluminium oxide however may not have formed a ‘bulk solid’ aluminium phosphate, but is adsorbed on the internal and external surface of the oxide. The same amorphous aluminium hydroxide, dried at 70°C instead of 200°C, is converted much more rapidly to aluminium phosphate when exposed to soluble phosphate. We propose a P-induced weathering mechanism to describe P sorption on amorphous aluminium hydroxides at high P concentrations. In addition to NMR, phosphate adsorption experiments conducted on aluminium hydroxides dried at different temperatures produced evidence that the porosity of the aluminium hydroxide aggregated particles can also be a factor controlling the rate of phosphate uptake from solution, if the aggregate is stable (is not resuspended) in solution.