lable at ScienceDirect

Soil Biology & Biochemistry 67 (2013) 235e244

Contents lists avai

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lb io

Is turnover and development of organic matter controlled by mineralcomposition?

Geertje Johanna Pronk a, b, *, Katja Heister a, Ingrid Kögel-Knabner a, b

a Lehrstuhl für Bodenkunde, Technische Universität München, 85350 Freising-Weihenstephan, Germanyb Institute of Advanced Study, Technische Universität München, Lichtenbergstrasse 2a, D-85748 Garching, Germany

a r t i c l e i n f o

Article history:Received 10 May 2013Received in revised form1 September 2013Accepted 8 September 2013Available online 20 September 2013

Keywords:Clay mineralFerrihydriteCharcoalAcid hydrolysisOrganic nitrogenArtificial soil

* Corresponding author. Lehrstuhl für BodenkuMünchen, 85350 Freising-Weihenstephan, Germany.

E-mail address: pronk@wzw.tum.de (G.J. Pronk).

0038-0717/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.soilbio.2013.09.006

a b s t r a c t

Interactions between organic matter (OM), minerals and charcoal may play an important role in thedevelopment and stabilization of OM in soils. However, these interactions are difficult to characterize innatural soils, which are usually very complex systems with unknown initial conditions. We developedso-called ‘artificial soils’ with a texture and OM content similar to natural arable soils that were incu-bated up to 18 months. The aim was to determine the turnover and development of OM with incubationtime, and to establish the effect of mineral composition and charcoal presence on organic carbon (OC)and N distribution and properties. Artificial soils were composed of quartz, manure as OM source and amicrobial community extracted from a natural arable soil, with 8 different mixtures of montmorillonite,illite, ferrihydrite, boehmite and charcoal. We determined C and N particle size distribution with timeand used solid-state 13C nuclear magnetic resonance (NMR) spectroscopy and acid hydrolysis to deter-mine the development of OM composition. The CO2 respiration rate and distribution of OC and N withparticle size was similar for all artificial soil compositions. OC and N accumulated in the <20 mm fractionover time and approximately 50% of coarse (>200 mm) particulate OM was lost after 18 months of in-cubation. 13C NMR spectroscopy indicated accumulation of protein-rich OC in the <20 mm fraction, likelyin the form of microbially produced substances. Acid hydrolysis showed a higher content of non-hydrolysable N in the mixtures containing clay minerals, indicating that some of the nitrogen presentwas strongly bound to phylosilicate surfaces. Ferrihydrite did not have any effect on non-hydrolysable N.From this, it can be concluded that in the artificial soils, clay minerals were more important than metal-oxides for the binding of nitrogen and OC. Overall, the artificial soils developed similarly to incubationexperiments with natural soils, and were therefore a valuable model system where the effect of specificcomponents on the development and turnover of soil OM could be determined under simplifiedconditions.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Recent research indicates that minerals may play an importantrole in the stabilization of organic matter (OM) and especiallynitrogen-containing compounds in soil (e.g. Kleber et al., 2007;Nannipieri and Eldor, 2009; Knicker, 2011). However, it is difficultto determine the effect of mineral composition on soil organicmatter (SOM) development in natural soils, which are complexsystems where mineral composition is usually not well-definedand initial conditions are generally unknown. Incubation experi-ments with simplified systems of known composition and initial

nde, Technische UniversitätTel.: þ49 8161715381.

All rights reserved.

conditions can therefore be useful to gain understanding of thefactors determining the development of OM over time, and its in-teractions with minerals. Artificial mixtures have been used in thepast to simulate processes taking place in soils. Madhok (1937)already used so-called ‘synthetic soils’ composed of mineral mix-tures to test the decomposition of cellulose in 1937. More recentstudies were performed by, for example, Saidy et al. (2012) whoincubated clay mineral and oxide mixtures with dissolved organicmatter (DOC), and Miltner and Zech (1999) who incubated leaflitter in the presence of different oxides. These studies demon-strated the complex interactions between microbial activity,decomposition of OM and presence of different minerals. To gainfurther understanding of these interactions, we tried to simulatetexture and organic carbon (OC) content of natural arable soils,while focussing on the effect of microbial activity alone underfurther constant environmental conditions, in a so-called ‘artificial

G.J. Pronk et al. / Soil Biology & Biochemistry 67 (2013) 235e244236

soil’ incubation experiment (Pronk et al., 2012). Mixtures ofdifferent composition, but the same texture, could so be used todetermine the effect of the presence of different mineral surfacesand charcoal on the development of the system. The importance ofmicrobial biomass in the formation of SOM has been highlighted inrecent years (e.g. Miltner et al., 2009). The turnover and redistri-bution of OC and N species in the artificial soils, where microbialactivity is the only driving force of OM turnover during incubation,may be used to demonstrate the extent to which microbial com-munities are able to contribute to SOM, and help to elucidate thecharacteristics of microbially produced OM in soil.

During degradation, OM is redistributed over the various par-ticle size fractions present in soil as coarse particles are broken upand microbial products are produced. This generally leads to anaccumulation of OC and N in the smaller particle size fractions, ashas been discussed extensively by e.g. Christensen (1992) andKleber et al. (2007) for natural soils, and was also observed forincubation experiments (e.g. Grosbellet et al., 2011; Leifeld et al.,2001). The determination of the particle size distribution of OCand N can therefore be a useful tool to monitor the turnover anddevelopment of OM during incubation. Solid-state 13C nuclearmagnetic resonance (NMR) spectroscopy is often used to charac-terize OC development and composition, and gives an indication ofthe type of compounds that are preferentially preserved ordecomposed (Kögel-Knabner, 2000; Simpson and Simpson, 2012).Several recent studies have indicated that proteins accumulate insoils over time, and that interaction of organic nitrogen species andmicrobial residues with minerals may be important for the stabi-lization of OM in soils (e.g. Kleber et al., 2007; Miltner et al., 2009).Acid hydrolysis is a classic method for the characterization of ni-trogen in soils (Stevenson, 1996). It separates the organic nitrogeninto protein N (determined as a-NH2), non-hydrolysable N andunidentified hydrolysable N. Paul et al. (2006) and Silveira et al.(2008) found that acid hydrolysis provided useful information toestimate different SOM pools. Furthermore, Leinweber andSchulten (2000) demonstrated that non-hydrolysable N is notonly determined by the chemistry of the organic N species present,but also by strong interaction with mineral surfaces. In an incuba-tion system where the origin of the OM is the same for all treat-ments, differences in the amount of non-hydrolysable N mighttherefore be taken as an indication of interaction of nitrogen spe-cies with specific minerals.

We performed an incubation experiment with so-called ‘artifi-cial soils’ composed of mixtures of clean and well-defined modelmaterials, where the development of OM could be followed in asimplified system with known initial conditions. In previousstudies, it was found that macroaggregates were quickly formed inthe artificial soils and increased until around 12e18 month of in-cubation, when they started to decrease. OM associated withmicroaggregates and minerals continued to increase gradually forthe entire incubation time (Pronk et al., 2012). This succession ofmacroaggregation, microaggregation and turnover of microbialdebris was in agreement with the aggregate hierarchy model and

Table 1Model materials present in the different artificial soil mixtures.

Name Model component Soil composition

MT IL FH MT

MT Montmorillonite X XIL Illite X XQC Clay-sized quartz XFH Ferrihydrite XB BoehmiteCH Charcoal

current understanding of interplay between aggregation and mi-crobial activity in soils (e.g. Oades, 1984; Six et al., 2004, 2006). Themicrobial community that developed in the artificial soils depen-ded on charcoal presence and, to a lesser extent, clay mineral typeafter 3 months of incubation (Ding et al., 2013). After 12 months ofincubation, the microbial community composition shifted andmetal oxides and clay minerals were the main minerals influencingthe microbial community, whereas the effect of charcoal decreased(Babin et al., 2013). However, cumulative CO2 respiration wassimilar for all artificial soil compositions, showing no clear effect ofthese differences in microbial community on respiration (Pronket al., 2012). It is still an open question whether the differences inartificial soil composition, and the microbial communitiesfollowing from these compositions, therefore led to differences inOM development within the artificial soils. The objective of thisstudy was to determine the turnover and development of OM withincubation time and to establish the effect of mineral compositionand charcoal presence on OC and N distribution and properties.

2. Materials and methods

2.1. Artificial soil incubation

The artificial soils were composed of themodel materials quartz,illite, montmorillonite, ferrihydrite, boehmite and charcoal. Air-dried and sterilized manure was used as OM source and the mix-tures were inoculated with the water extractable microbial com-munity from a Eutric Cambisol. Manure was added to an equivalentof 15 � 2 mg C g�1 and 1.4 � 0.2 mg N g�1 in all artificial soils. Themixtures contained 40e42% sand (>63 mm), 52e54% coarse andmedium silt (6.3e63 mm) and 6% fine silt and clay (<6.3 mm). Theamount of sand and clay mineral or clay-sized quartz added foreach composition (Table 1) was varied slightly to account for dif-ferences in the particle size distribution of the model materials.These materials were added in such amounts that the mass of thefraction<6.3 mm always represented 5.6% of the total soil mass. Thetexture of the illite, montmorillonite and clay-sized quartz wasdetermined by X-ray attenuation (SediGraph 5100/Master Tech 051,Micromeritics GmbH, Mönchengladbach, Germany) and contained50%, 67% and 74% mass of particles <2 mm respectively. The ferri-hydrite, boehmite and charcoal were accounted as part of the<6.3 mm fraction because of their large specific surface area. Themodel materials, artificial soil composition and incubation weredescribed in detail in Pronk et al. (2012). In total, artificial soils with8 different compositions (Table 1) were incubated for 3, 6,12 and 18months, respectively. Three replicate batches of each compositionwere sampled at each time point.

The artificial soils were separated into fractions with a sizeequivalent to coarse and medium sand (200e2000 mm), fine sand(63e200 mm), coarse silt (20e63 mm) and medium þ fine silt andclay fraction (<20 mm) by wet sieving. The method for aggregatedisruption was adapted from Schmidt et al. (1999) and Amelungand Zech (1999). For ultrasonic dispersion, 150 ml of water was

þ IL MT þ CH IL þ FH IL þ B IL þ FH þ CH

XX X X

X XX

X X

Table 2The organic carbon (OC) content and C/N ratio of the bulk and <20 mm fraction of the artificial soils after 3 and 18 months of incubation. Composition is indicated by MT(montmorillonite), IL (illite), FH (ferrihydrite), CH (charcoal) and B (boehmite), (partially adapted from Pronk et al. (2012)), � represents the standard deviation between threereplicates.

OC mg g�1 bulk soil C/N bulk soil C/N < 20 mm fraction

3 months 18 months 3 months 18 months 3 months 18 months

MT 15 � 0.7 14 � 0.9 11.7 � 0.1 10.2 � 0.1 6.7 � 0.3 6.5 � 0.3IL 15 � 0.6 12 � 0.2 11.8 � 0.1 9.9 � 0.1 6.3 � 0.2 6.6 � 0.9FH 16 � 3.2 13 � 1.0 11.5 � 0.1 9.7 � 0.1 5.8 � 0.1 5.5 � 0.7MT þ IL 14 � 0.9 13 � 1.0 11.6 � 0.1 10.2 � 0.1 6.7 � 0.4 6.3 � 0.1MT þ CH 30 � 2.0 26 � 2.0 20.8 � 1.1 19.4 � 1.1 17.6 � 4.8 13.7 � 1.1IL þ FH 18 � 3.8 12 � 0.5 11.3 � 0.2 9.5 � 0.2 5.9 � 0.3 6.4 � 0.8IL þ B 15 � 0.4 12 � 1.0 11.6 � 0.1 9.8 � 0.1 6.3 � 0.7 5.7 � 0.4IL þ FH þ CH 31 � 2.7 28 � 2.6 19.4 � 1.3 17.9 � 1.0 13.9 � 1.9 11.5 � 1.6

G.J. Pronk et al. / Soil Biology & Biochemistry 67 (2013) 235e244 237

added to 30 g sample and treated with an energy of 60 J ml�1

(dispersion tip Bandelin electronics, Sonoplus UW 2200, Berlin,Germany) to disrupt macroaggregates (250e2000 mm) but notmicroaggregates. All particle size fractions were freeze-dried beforestorage.

Table 3Cumulative CO2 respiration (partially adapted from Pronk et al. (2012)) of the arti-ficial soils in mg CO2eC g dry soil�1 after 3, 6, 12 and 18 months of incubation. Soilcomposition is indicated by MT (montmorillonite), IL (illite), FH (ferrihydrite), CH(charcoal) and B (boehmite), � represents the standard deviation between threereplicates.

Cumulative mg CO2eC g�1

3 months 6 months 12 months 18 months

MT 0.8 � 0.2 1.1 � 0.3 2.1 � 0.6 2.7 � 0.9IL 0.7 � 0.2 1.1 � 0.3 2.0 � 0.5 2.6 � 0.6FH 0.8 � 0.3 1.4 � 0.4 2.7 � 0.7 3.2 � 0.8MT þ IL 0.7 � 0.1 0.9 � 0.2 2.1 � 0.5 2.7 � 0.6MT þ CH 0.6 � 0.1 1.0 � 0.2 2.5 � 0.5 3.2 � 0.7IL þ FH 0.7 � 0.2 1.1 � 0.3 2.4 � 0.5 3.0 � 0.5IL þ B 0.8 � 0.2 1.1 � 0.3 2.3 � 0.6 3.0 � 0.7IL þ FH þ CH 0.9 � 0.3 1.4 � 0.4 2.7 � 0.7 4.4 � 0.9

2.2. Sample characterization

Carbon and nitrogen contents were determined in duplicate bycombustion and chromatographic separation with an elementalanalyser (Hekatech, Euro EA 3000, Wegberg, Germany). The pH ofall incubated samples was measured in 0.01 M CaCl2. The CO2respiration rate was monitored bimonthly in subsamples of 20 g for3 replicates of each composition. The CO2 respired was adsorbed ina 0.1 M NaOH solution and determined by titration to a pH of 8.3according to the method of Isermeyer (1951). For analysis of nitrateand ammonium contents, 5 g sample was added to 25 ml of 0.01 MKCl and shaken for 2 h. The samplewas then centrifuged and nitrateand ammonium contents of the extract were determined with acontinuous flow analyser (Sanþþ, Skalar Analytical B.V., Breda, TheNetherlands).

The composition of the OM of selected samples was measuredwith solid-state 13C NMR spectroscopy. Samples were analysedwith a DSX 200 spectrometer (Bruker Biospin, Rheinstetten, Ger-many) operating with a cross-polarization magic angle spinning(CP MAS) technique at the resonance frequency of 13C. The acqui-sition parameters included a spinning rate of 5 kHz, a contact timeof 0.2 ms and a pulse delay between 0.2 and 0.6 s. Chemical shiftswere calibrated relative to tetramethylsilane (0 ppm). The totalnumber of scans was varied depending on the carbon content of thesample, but a minimum of 11 000 scans were accumulated for eachsample. A line broadening of 75 Hz was applied. The contributionsto the various C groups were determined by integration of theirsignal intensity in their respective chemical shift regions by meansof the instrument software. Chemical shift regions were used asfollows, alkyl C 0e45 ppm, O/N-alkyl C 45e110 ppm, aryl-C 110e160 ppm and carbonyl/carboxyl/amide C 160e220 ppm (Knickerand Lüdemann, 1995).

Protein, hydrolysable and non-hydrolysable N contents weredetermined by acid hydrolysis, adapted from Stevenson (1996).Approximately 1 g of sample was cooked in 5 ml of 6 N HCl and 1 NHCOOH mixture under N2 atmosphere at 110 �C for 12 h. The hy-drolysate was then cooled down and filtered. The residue on thefilter was dried and kept for C and N analysis. The filtrate was driedcompletely overnight on a sand bath and re-suspended in 5ml H2O.The pH of the samples was then adjusted to pH 8 by adding 5 NNaOH. The samples were again heated in an oven at 110 �C for45 min until almost dry and H2O was added to a total volume of25 ml in a volumetric flask. For determination of protein content,

100 ml sample þ 100 ml (0.4 ml/l) Na-citrate þ 100 ml (10 g/l)ninhydrin were mixed and heated for 20 min in boiling water, and1.25 ml ethanol/water mixture (1:1) was added to the samples.Protein N content was then measured with a colorimeter (Spec-tronic 601, Milton Roy, Ivyland, PA, USA) at 570 nm. The non-hydrolysable N content was calculated from the N contentremaining in the sample residue after hydrolysis. The protein N andnon-hydrolysable N content were subtracted from the total Ncontent to determine the unidentified hydrolysable N content ofeach sample.

3. Results

3.1. Development of OM particle size distribution and composition

A general characterization of bulk sample properties can befound in Pronk et al. (2012) and Table 2. The pH of the artificial soilswas 7e7.5 for all compositions and did not change with incubationtime. The cumulative CO2 respiration rate was similar for all arti-ficial soil compositions (Table 3). The OC losses were 15e20% ofinitial OC content (excluding contribution of charcoal-C) after 18months of incubation. About 75e80% of both OC and N were pre-sent in the >200 mm fraction of the manure before incubation. Thisgradually decreased over time, but the >200 mm fraction stillcontained 52e62% of total OC (excluding charcoal-C) and 38e47%of total N after 18 months of incubation. About half of the OC lostfrom the >200 mm fraction was respired as CO2, and the restgradually accumulated in the <20 mm fraction. The OC and Ncontents of the 63e200 mm and 20e63 mm fractions for all artificialsoils that did not contain charcoal were relatively low and constantduring the entire incubation time (Figs. 1 (aef) and 2 (aef)). Asthese fractions consisted almost entirely of quartz, there probably

Fig. 1. Organic carbon stocks of the particle size fractions of the different artificial soil compositions (aeh), with incubation time in months. Soil composition is indicated by MT(montmorillonite), IL (illite), FH (ferrihydrite), CH (charcoal) and B (boehmite). Error margins are based on the difference between 3 replicates per soil composition.

Fig. 2. Nitrogen stocks of the particle size fractions of the different artificial soil compositions (aeh), with incubation time in months. Soil composition is indicated by MT(montmorillonite), IL (illite), FH (ferrihydrite), CH (charcoal) and B (boehmite). Error margins are based on the difference between 3 replicates per soil composition.

G.J. Pronk et al. / Soil Biology & Biochemistry 67 (2013) 235e244240

was very little interaction between minerals and OM. They weretherefore considered to be of less relevance for the development ofthe artificial soils, and the discussion further focuses on the>200 mm fraction, hereafter referred to as ‘coarse fraction’ and the<20 mm fraction, hereafter referred to as ‘fine fraction’. The OCdistribution of the artificial soils containing charcoal varied widelybetween replicates at all time steps, which was probably due todifferences in charcoal distribution over the particle size fractions.Charcoal was gradually broken up into smaller particles duringincubation, resulting in a large variability in particle size distribu-tion between replicates (Fig. 1g and h). Therewas therefore no clearpattern of charcoal distribution over time.

The redistribution of OC and N did not occur linearly with in-cubation time. The OC and N contents of the coarse fractiondecreased and those of the fine fraction increased in the 3 monthsincubated artificial soils compared to the original manure. Between3 and 6 months of incubation, the OC and N content of the coarsefraction stayed relatively the same, and even increased slightly forartificial soils IL and IL þ FH. From 6 to 12 months of incubation, OCand N content of the coarse fraction clearly decreased in all artificialsoils and accumulated in the fine fraction until the end of incuba-tion. Overall, there were no significant differences in OC and Ndistribution between artificial soil compositions.

Solid-state 13C NMR spectra were acquired for the bulk sampleand <20 mm fraction of the original manure, and a selection of 3and 18 months incubated samples of artificial soils. No clear dif-ferences in spectra between artificial soil compositions could beseen. Representative spectra from artificial soil composition IL areshown in Fig. 3a and b. The spectrum for the bulk manure sampleshows a distinct peak at 72 ppm and shoulders at 84 and 64 ppmthat indicate that polysaccharides and some cellulose was present(Knicker et al., 2008; Kögel-Knabner, 1997), probably originatingfrom some remaining plant materials in the manure. There was aclearly defined peak at 56 ppm, indicative of amino-C or methoxyl-C (Kögel-Knabner, 2002) at all incubation times. A general increase

Fig. 3. Selected solid-state 13C NMR spectra of bulk soil and <20 mm particle size fraction of awere obtained from the manure model material.

in the aromatic region relative to the other peaks in the bulksamples occurred from the starting material to the 3 monthsincubated samples. From 3 to 18 months of incubation, there was agradual decrease in the 72 and 105 ppm peaks, which are indicativeof polysaccharides (Kögel-Knabner, 2002), relative to the rest of thespectrum. The amino-C/methoxyl-C peak at 56 ppmwas evenmorestrongly defined in the fine fraction. It decreased from the startingmaterial to the 3 months incubated samples, but increased againafter 18 months of incubation.

3.2. Organic and inorganic nitrogen

Inorganic nitrogen accumulated in the form of nitrate in allartificial soils with incubation time (Fig. 4a). Nitrate contributed amaximum of 0.5% of total nitrogen present in the artificial soils atany time step. The nitrate content of the ferrihydrite containingartificial soils was slightly higher than that of the other composi-tions. This difference may be explained by the iron nitrite fromwhich the ferrihydrite was synthesized, and of which a small res-idue still remained. Ammonium content of the artificial soils wasclose to the detection limit for all compositions except those con-taining montmorillonite (Fig. 4b). Artificial soil MT containedapproximately 8 mg g�1 of ammonium-N after 3 months of incu-bation and MT þ CH and MT þ IL around 3 mg g�1. However, at itsmaximum, these ammonium concentrations were still two ordersof magnitude lower than the nitrate content of the same samplesand can probably be explained by the presence of small amounts ofammonium on the cation exchange sites of montmorillonite. Theammonium content of the montmorillonite containing artificialsoils decreased with time to values similar to the other artificial soilcompositions after 12e18 months.

Protein, unidentified hydrolysable N and non-hydrolysable Ncontents of the bulk artificial soils were determined in triplicateafter 3 and 18months of incubation (Table 4). Total N content of theartificial soils did not change during incubation, so that only the

rtificial soil composition IL at different incubation times. The spectra at incubation start

Fig. 4. Nitrate (a) and ammonium (b) contents of the artificial soils. Composition isindicated by MT (montmorillonite), IL (illite), FH (ferrihydrite), CH (charcoal) and B(boehmite). Error margins are based on the 3 replicates per composition and time step.

G.J. Pronk et al. / Soil Biology & Biochemistry 67 (2013) 235e244 241

distribution of N over these three fractions changed. The amount ofnon-hydrolysable N in the artificial soils was higher than that of theoriginal manure for both the 3 and 18 month incubated samples.The percentage of protein N decreased slightly from 3 to 18 monthsof incubation. Non-hydrolysable N increased for MT and MT þ ILfrom 3 to 18 months of incubation, but decreased for all othercompositions. The protein, unidentified hydrolysable N and non-hydrolysable N contents of the <20 mm fraction were determinedin duplicate for all 4 sampling times (Table 5). Total N content of the<20 mm fraction was similar for the 3 and 6 months incubatedsamples, then increased after 12 months of incubation and againstayed at a relatively similar level after 18 months. The only

Table 4Percentage of protein, unidentified hydrolysable and non-hydrolysable N for bulk samplestandard deviation between three replicates.

Proteins Unidentifi

3 months 18 months 3 months

MT 30.3 � 7.6 22.7 � 3.6 45.0 � 13IL 30.4 � 5.2 24.0 � 2.5 35.4 � 8.4FH 29.4 � 3.9 23.6 � 1.2 39.2 � 2.1MT þ IL 28.8 � 7.4 27.2 � 1.7 41.6 � 11MT þ CH 29.7 � 4.7 20.7 � 3.8 26.6 � 7.1IL þ FH 26.5 � 8.8 23.9 � 6.7 48.6 � 17IL þ B 28.4 � 5.1 23.2 � 3.2 39.4 � 5.0IL þ FH þ CH 26.7 � 3.5 22.3 � 3.3 35.9 � 8.2Manure 29.9 � 2.8 45.4 � 0.3

exception to this is the IL þ B composition, for which the N contentof the small fraction increased gradually with time. The percent-ages of protein, unidentified hydrolysable and non-hydrolysable Nwere similar for all incubation times, although the protein contentof IL and MT is slightly higher and the unidentified hydrolysable Ncontent slightly lower after 3 months of incubation. The non-hydrolysable N content of FH, the artificial soil without clay min-eral, was lower than that of the other compositions at all time steps.Compared to the <20 mm fraction of the original manure used, theprotein content in all artificial soils was much higher, whereas theunidentified hydrolysable N content decreased relative to themanure.

4. Discussion

4.1. OM degradation and turnover in the artificial soils

The artificial soil mixtures developed relatively quickly intoaggregated soil-like materials with established microbial commu-nities. The observed accumulation of inorganic nitrogen indicatesthat other nutrients (most likely P, or some micronutrients) limitedmicrobial activity during incubation. The constant total nitrogencontents over time further show that no significant amount ofdenitrification took place during incubation. Conditions in theartificial soils were likely to be predominantly oxic due to the highsand content of the mixtures.

The redistribution of OM from the coarse to the fine fraction, andthe gradual accumulation of OM with a low C/N ratio in the finefraction observed during the artificial soil incubation experiment(Table 2, Fig. 1) are consistent with literature as discussed by e.g.Christensen (1992) and Kleber et al. (2007). The manure used as asubstrate source consisted of over 70% coarse (>200 mm) OM par-ticles, whereas microbial debris was likely to be in the form of fineparticles and dissolved organic matter. The OC and especially Naccumulated in the fine fraction of the artificial soils was thereforeeither in the form of degradation residues or of microbial origin.About half of the OC and N < 200 mm was lost after 18 months ofincubation, demonstrating the capability of the microbial com-munity to degrade OM. Rapid turnover of coarse particulate organicmatter (POM) and redistribution to the fine fractionwithin one yearwas also found in an incubation experiment with well-developedsoil material and compost by Grosbellet et al. (2011) and forhumus-free sandy mining soil incubated with compost by Leifeldet al. (2001).

The OM composition of the bulk samples as determined bysolid-state 13C NMR spectroscopy did not differ between compo-sitions. The spectra still reflected the composition of the originalmanure after 18 months of incubation (Fig. 3). The decrease of the72 ppm peak with incubation time points to a relative loss of

s and manure after 3 and 18 months of incubation. Error margins (�) represent the

ed hydrolysable N Non-hydrolysable N

18 months 3 months 18 months

.8 45.0 � 8.8 24.7 � 6.4 32.2 � 5.650.0 � 3.9 34.3 � 7.7 26.0 � 4.852.3 � 7.9 31.5 � 6.0 24.0 � 6.9

.3 36.5 � 6.2 29.5 � 3.9 36.3 � 5.950.9 � 5.9 43.7 � 10.5 28.4 � 4.5

.4 55.6 � 9.5 24.9 � 8.7 20.5 � 4.651.4 � 7.0 32.2 � 6.6 25.5 � 4.847.8 � 6.1 37.4 � 5.3 29.9 � 5.6

24.7 � 3.1

Table 5Percentage of protein, unidentified-hydrolysable and non-hydrolysable N in the <20 mm fraction of the artificial soils. Error margin (�) represents the difference between tworeplicates.

Proteins Unidentified hydrolysable N Non-hydrolysable N

3 months 6 months 12 months 18 months 3 months 6 months 12 months 18 months 3 months 6 months 12 months 18 months

MT 35.0 � 2.5 18.4 � 8.0 18.5 � 3.4 21.9 � 0.1 48.9 � 0.5 64.1 � 13 58.6 � 5.7 58.5 � 0.7 16.1 � 2.0 17.6 � 5.8 22.8 � 2.3 19.6 � 0.7IL 27.6 � 3.5 20.1 � 11 15.1 � 1.9 18.0 � 0.6 53.0 � 0.4 63.0 � 12 60.9 � 4.5 62.3 � 1.3 19.4 � 3.9 16.9 � 1.0 24.0 � 2.5 19.8 � 0.7FH 18.1 � 2.5 11.4 � 0.7 17.1 � 14 15.7 � 3.5 69.0 � 2.2 78.0 � 2.5 68.9 � 17 72.7 � 6.9 12.9 � 0.3 10.6 � 1.9 14.0 � 4.0 11.6 � 3.4MT þ IL 17.7 � 0.6 22.6 � 14 18.0 � 2.4 19.1 � 0.1 59.7 � 4.9 59.0 � 21 57.5 � 8.4 64.5 � 0.5 22.6 � 5.6 18.4 � 6.7 24.5 � 6.1 16.4 � 0.6MT þ CH 18.8 � 3.0 18.8 � 10 15.7 � 0.4 19.1 � 3.5 50.5 � 11 57.9 � 8.4 55.5 � 2.2 59.7 � 11 30.7 � 8.7 23.3 � 1.7 28.8 � 1.8 21.2 � 7.0IL þ FH 18.3 � 3.0 16.1 � 0.7 15.6 � 1.5 20.3 � 0.6 65.2 � 5.1 68.7 � 4.3 62.6 � 1.5 60.6 � 0.6 16.5 � 2.1 15.2 � 3.6 21.8 � 0.1 19.1 � 0.1IL þ B 22.0 � 10 23.8 � 8.1 15.6 � 0.6 18.1 � 1.4 58.6 � 18 57.2 � 10 60.7 � 2.1 63.3 � 1.7 19.3 � 7.7 19.0 � 2.2 23.7 � 2.7 18.6 � 3.1IL þ FH þ CH 13.9 � 2.3 13.7 � 7.7 12.8 � 0.6 15.6 � 4.0 65.7 � 3.5 66.2 � 13 65.0 � 0.7 66.5 � 8.3 20.3 � 1.3 20.0 � 5.2 22.2 � 0.1 17.9 � 4.3

G.J. Pronk et al. / Soil Biology & Biochemistry 67 (2013) 235e244242

polysaccharides, indicating that these substances were preferen-tially degraded by the microbial community. The clearly resolvedpeak at 56 ppm is unusual in natural soils (e.g. Baldock et al., 1992;Kögel-Knabner, 1997) and is probably due to the presence of highamounts of C in amino groups of proteins, peptides and aminosugars. In comparisonwith 13C NMR spectra of SOM from long-termfield experiments on arable soils fertilized with manure (e.g. Kiemet al., 2000; Kölbl and Kögel-Knabner, 2004), the artificial soilsshowed relatively low contents of alkyl- and O-alkyl-C, and a highercontent of N-alkyl/methoxyl-C. The manure used as a model ma-terial contained a lot of nitrogen, which accumulated relative tocarbon in the artificial soils over time because there was no outflowof dissolved N species, and no significant denitrification was takingplace. The accumulation of nitrate in the artificial soils shows thatnitrogen was not a limiting nutrient. It may therefore be that therewas no strong incentive for the microbial community to utilizethese easily degradable N-containing compounds, leading to theiraccumulation over time. In addition, N-rich OM may have beenstabilized by interaction with mineral surfaces (Kleber et al., 2007and references therein). The artificial soils were composed of OM-free minerals, which had a large amount of free surfaces thatwere available for interaction with organic compounds from thestart of incubation. The signal at 56 ppm is even more pronouncedin the <20 mm fraction, pointing to the accumulation of amino-Ccontaining molecules like proteins and peptides, which is furthersupported by the low C/N ratio of the fine fraction (Table 2). For themanure at the start of incubation, this can be explained by thepresence of easily degradable nitrogen-containing compounds likee.g. urea that are commonly present in manure. The signal is lowerafter 3 months of incubation, indicating that these compounds hadbeen used by microorganisms, but it increased again with 18months of incubation. The OM that accumulated in the fine fractionof the artificial soils was predominantly of microbial origin, as therewas no input of roots or litter during the experiment. The spectrumof the fine fraction is similar to that of bacteria and fungi (Baldocket al., 1990; Kögel-Knabner, 2002), although those do not have thehigh amino-C signal at 56 ppm. This points to an accumulation ofamino acids or proteins in the artificial soils in addition to thosepresent in the microbial biomass, either due to the release ofextracellular enzymes by the microbial community (Burns et al.,2013), or because of selective preservation of amino-C during in-cubation. The relative distribution of protein N, non-hydrolysable Nand unidentified-hydrolysable N of the fine fraction of the artificialsoils was similar for the entire incubation period. This indicates thatthe composition of the source of OM accumulating in the finefraction did not change with incubation time, which supports thehypothesis that it consists mostly of microbial debris. The clay-sized fractions in natural soils have a larger contribution of O-alkyl and often also long-chain alkyl C compared to the fractionsobtained in the artificial soils. This implies that the accumulation ofthese compounds in clay-sized fractions either needs longer times

to develop, different sources of OM (e.g. plant and root litter), oradditional processes like leaching of dissolved OC and N to thesubsoil, that were not active in the artificial soil experiment.

4.2. Time development of the artificial soil system

Redistribution of N and OC from the coarse to the fine fractiondid not occur continually with time. For the first 6 months, the Ncontent of the fine fractionwas relatively stable, or even decreased.This may indicate that during this time, a large part of microbialactivity took place on coarser particles. As the major part of the OMwas present as coarse POM, it could be that during the first 6months, habitats for microorganisms were mainly formed oncoarse POM particles which were their main food source, and thusthat nitrogen was bound there by these organisms. Other expla-nations may be that fine, un-stabilized OM particles were prefer-entially degraded by microorganisms, leading to a depletion of OMin the fine fraction. After 6e12 months of incubation, nitrogenstarted to accumulate in the fine fraction. This points to a gradualaccumulation of small-sized microbial debris with time, increasinginteractions of OM with clay mineral surfaces, or a shift in strategyof the microbial community, as easily degradable carbon sourcesbecame less available. A further factor may be the development ofaggregation in the artificial soils. Macroaggregates were formedquickly and increased until around 12e18 month of incubation,when they started to decrease (Pronk et al., 2012). This decreasemay have followed the gradual degradation of OM and the accu-mulation of microbial debris in the fine fraction. On the other hand,OM associated with microaggregates and minerals continued toincrease gradually for the entire incubation time, and may bemainly formed by interactions of microbial debris with minerals inthe fine fraction (Pronk et al., 2012). This succession of macro-aggregation, microaggregation and turnover of microbial debris isin agreement with the aggregate hierarchy model and currentunderstanding of interplay between aggregation and microbialactivity in soils (e.g. Oades, 1984; Six et al., 2004, 2006). Further-more, Ding et al. (2013) and Babin et al. (2013) showed that themicrobial community composition and the importance of differentartificial soil components changed from 3 to 12 months of incu-bation, which may indicate a shift in strategies of the microbialcommunities and their behaviour.

The time development of the artificial soils was thereforedetermined by the complex interactions between a changing mi-crobial community, the dynamics of biodegradation of the manure,and the development and decline of physical structure and themicrobial habitats connected to it. The artificial soils started asmixtures of clean materials, uninhabited by microorganisms. Afterinoculation, growth of microbes and fungi occurred and a microbialcommunity was established. As substrate availability declinedduring incubation due to CO2 respiration and increasing formationof microaggregates and organo-mineral associations, release of

G.J. Pronk et al. / Soil Biology & Biochemistry 67 (2013) 235e244 243

microbial produced substrates gradually decreased. A closer ex-amination of specific biomarkers could give more insight into thecontribution of microbial community to the OM present in theartificial soils and how a living microbial community and physicaland chemical properties interact to shape microbial habitats andstructure in this system.

4.3. Effect of mineral composition and charcoal presence

The effect of mineral composition and charcoal presence on theoverall development of the artificial soils was relatively small. Itseems that stable microbial communities with similar pro-ductivities could establish independent of artificial soil composi-tion. However, there were some effects on the distribution ofprotein and non-hydrolysable N. The <20 mm fraction of theartificial soil that did not contain a clay mineral had less non-hydrolysable N as the other mixtures at all time steps. The non-hydrolysable fraction of nitrogen is composed of both organicchemicals that are resistant to hydrolysis and compounds thatresist hydrolysis because they are bound to mineral surfaces orboth (Leinweber and Schulten, 2000). Solid-state 13C NMR spec-troscopy did not indicate differences in composition between thedifferent mixtures. Furthermore, the OM in the fine fraction of theartificial soils had the same origin (i.e. microbial activity) in allmixtures. Therefore, the higher amount of non-hydrolysable N inthe mixtures containing a clay mineral is probably due to organicN that was bound by the clay minerals either by sorption orincorporation in microaggregates. An association of OM with clayminerals in these artificial soils was also found using nano-scalesecondary ion mass spectrometry (nanoSIMS) (Heister et al.,2012). It has long been known that proteins and peptides can bestrongly bound by clay minerals, reducing their availability formicroorganisms, as reviewed by Filip (1973) and Stotzky (1986).Recent research has suggested that binding of labile organiccompounds like proteins and peptides to metal oxides can beimportant for stabilization of OM in soil (Kleber et al., 2007;Knicker, 2011). Here, we can clearly show that the presence of aclay mineral also leads to more non-hydrolysable N in a systemwhere all other incubation conditions were the same. The inter-action of OM with clay minerals may therefore be an importantfactor for the stabilization of OM in soil. There were only smalldifferences in OM development between illite and montmoril-lonite in the artificial soils. This was unexpected, as previousstudies by e.g. D’Acqui et al. (1998), Denef et al. (2002) andWattel-Koekkoek et al. (2003) did find significant effects of clay miner-alogy on OM stabilization. However, these studies all comparedthe (1:1) clay mineral kaolinite to (2:1) clay minerals, whereasboth the illite and montmorillonite used in this study are (2:1)clay minerals, which have a higher surface charge, cation ex-change capacity and specific surface area than 1:1 clay minerals.The difference between illite and montmorillonite is thereforelikely to be smaller than that between (1:1) and (2:1) clay min-erals, as they are closer in structure surface properties. Further-more, drying/wetting cycles, which may increase the effect of claymineral type on soil structure and OM storage (Denef et al., 2002),were excluded in our experiment. It is possible that clay mineraltype becomes more important over longer incubation times, orwith additional input of OM leading to a less nutrient limitedsystem. The presence of illite and montmorillonite did lead to theestablishment of different microbial communities in the artificialsoils (Ding et al., 2013; Babin et al., 2013). More research is neededto determine if the different microbial communities resultingfrom clay mineral type had an effect on OM composition withhigher resolution methods like for example chemical analysis ofmicrobial degradation products.

There was no indication that OM or N preferentially interactedwith ferrihydrite during incubation. There was less non-hydrolysable N in the mixture containing only ferrihydrite andquartz, and in addition, the amount of non-hydrolysable N in themixture containing illite and the one containing illiteþ ferrihydritewas similar. This shows that no more nitrogen was non-hydrolysable when ferrihydrite was present. The pH of the artifi-cial soils was about 7.5 during the entire incubation time, which isclose to the point of zero charge of the ferrihydrite. The affinity ofproteins for the interaction with ferrihydrite may therefore be notas strong, and interactions between proteins and clay minerals maydominate instead. This indicates that ferrihydrite may be importantfor stabilization of protein-like material in soils at lower pH ranges,but that at pH above 7, clay minerals are equally or more importantfor stabilization of organic nitrogen.

The presence of charcoal had no clear effect on the respirationrate in the artificial soils. The charcoal added was highly aromatic(Pronk et al., 2012), and if use of charcoal as a substrate took place atall, its degradation rate was very low in comparison to that of theorganic material originating from the manure. The slightly higheramount of non-hydrolysable N content in the artificial soil mixturescontaining charcoal after 3 months of incubation could be due tosome sorption of OM to charcoal surfaces and in micropores(Lehmann et al., 2011 and references therein). This correspondswith the findings of Ding et al. (2013), who found differences in themicrobial community established in the charcoal containing arti-ficial soils after 3 months of incubation. However, the effect ofcharcoal on hydrolysable N disappeared with longer incubationtime, which was also observed for its effect on the microbialcommunity (Babin et al., 2013). There was no effect of charcoalpresence on the protein or non-hydrolysable N contents in the finefraction. Overall, charcoal may have been an important factor in thedevelopment of the artificial soils at the start of incubation, but itspresence became of less importance with incubation time. It maybe that the charcoal quickly interacted with, and was covered byother organic materials within the first 3 months of incubation andwas no longer available as a reactive surface after that time. There islittle known about the effect of charcoal on the physical andchemical properties of soil (Sohi et al., 2010). Its presence can causeboth positive and negative effects on microbial biomass, activityand OM turnover depending on environmental conditions and theproperties of the soil and the charcoal in question (Lehmann et al.,2011). The interaction between charcoal and other soil componentsand their development of soil structure could therefore be a focusfor further study.

4.4. Conclusions

The development of the artificial soils resulted in complex in-teractions between structure, microbial activity and degradation ofOM. The redistribution of OM from the coarse to the finer particlesize fractions did not proceed linearly, but was instead affected byshifts in strategy and composition of the microbial community,aggregate formation and interactions between OM and minerals.The artificial soil incubation experiment can therefore give insightinto the development and decline of a soil-like system undersimplified and controlled conditions and help to elucidate thecomplex processes that lead to OM degradation in soil.

All different artificial soil compositions developed to relativelysimilar systemswith similar respiration rates. Microbially producedOM accumulated with incubation time mainly in the <20 mmfraction. This material contained a lot of both organic and inorganicnitrogen species that are usually quickly degraded in natural sys-tems. The accumulation of nitrogen in the fine fractions of theartificial soils may be explained because nitrogen was present in

G.J. Pronk et al. / Soil Biology & Biochemistry 67 (2013) 235e244244

excess relative to OC, but the increasing non-hydrolysable N con-tents with time indicate that it was also at least partly stabilized byinteractionwith the many free mineral surfaces that were availablein the artificial soils. The presence of clay minerals led to the for-mation of more non-hydrolysable N, indicating that organic nitro-gen was associated with the clay minerals in particular. Contrary toexpectations, the presence of ferrihydrite did not increase non-hydrolysable N. This shows that at the neutral pH of the artificialsoils, clay minerals may be the main mineral agent for formation oforgano-mineral associations and stabilization on OM and N in soil.

Acknowledgements

This project was carried out within the framework of the pri-ority program 1315 ‘Biogeochemical Interfaces in Soil’ funded bythe German Research Foundation (DFG) (KO 1035/33-1) with thesupport of the Technische Universität München e Institute forAdvanced Study, funded by the German Excellence Initiative. Wethank Cordula Vogel, Gabriele Albert, Maria Greiner, Ulrike Maul,Zohre Javaheri, Samira Ravash and Zahra Kazemi for laboratoryassistance and the preparation of the artificial soils, Anja Miltnerand Esther Cyrus of the Helmholtz Zentrum, Leipzig for the prep-aration of the dried manure and Holger Kirchman for supplying thesoil from Ultuna used for extraction of the microbial inoculant.

References

Amelung, W., Zech, W., 1999. Minimisation of organic matter disruption duringparticle-size fractionation of grassland epipedons. Geoderma 92, 73e85.

Babin, D., Ding, G.-C., Pronk, G.J., Heister, K., Kögel-Knabner, I., Smalla, K., 2013.Metal oxides, clay minerals and charcoal determine the composition of mi-crobial communities in matured artificial soils and their response to phenan-threne. FEMS Microbiol. Ecol. 86, 3e14.

Baldock, J., Oades, J., Waters, A., Peng, X., Vassallo, A., Wilson, M., 1992. Aspects ofthe chemical structure of soil organic materials as revealed by solid-state 13CNMR spectroscopy. Biogeochemistry 16, 1e42.

Baldock, J.A., Oades, J.M., Vassallo, A.M., Wilson, M.A., 1990. Solid-state CP/MAS 13CNMR analysis of bacterial and fungal cultures isolated from a soil incubatedwith glucose. Soil Res. 28, 213e225.

Burns, R.G., DeForest, J.L., Marxsen, J., Sinsabaugh, R.L., Stromberger, M.E.,Wallenstein, M.D., Weintraub, M.N., Zoppini, A., 2013. Soil enzymes in achanging environment: current knowledge and future directions. Soil Biol.Biochem. 58, 216e234.

Christensen, B.T., 1992. Physical fractionation of soil and organic matter in primaryparticle size and density separates. Adv. Soil Sci. 20, 1e90.

D’Acqui, L.P., Daniele, E., Fornasier, F., Radaelli, L., Ristori, G.G., 1998. Interactionbetween clay microstructure, decomposition of plant residues and humifica-tion. Eur. J. Soil Sci. 49, 579e587.

Denef, K., Six, J., Merckx, R., Paustian, K., 2002. Short-term effects of biological andphysical forces on aggregate formation in soils with different clay mineralogy.Plant Soil 246, 185e200.

Ding, G.-C., Pronk, G.J., Babin, D., Heuer, H., Heister, K., Kögel-Knabner, I., Smalla, K.,2013. Mineral composition and charcoal determine the bacterial communitystructure in artificial soils. FEMS Microbiol. Ecol. 86, 15e25.

Filip, Z., 1973. Clay minerals as a factor influencing the biochemical activity of soilmicroorganisms. Folia Microbiol. 18, 56e74.

Grosbellet, C., Vidal-Beaudet, L., Caubel, V., Charpentier, S., 2011. Improvement ofsoil structure formation by degradation of coarse organic matter. Geoderma162, 27e38.

Heister, K., Höschen, C., Pronk, G., Mueller, C., Kögel-Knabner, I., 2012. NanoSIMS asa tool for characterizing soil model compounds and organomineral associationsin artificial soils. J. Soils Sediments 12, 35e47.

Isermeyer, H., 1951. Eine einfache Methode zur Bestimmung der Bodenatmung undder Karbonate im Boden. Z. Pflanzenernaehr. Bodenkd. 56, 26e38.

Kiem, R., Knicker, H., Körschens, M., Kögel-Knabner, I., 2000. Refractory organiccarbon in C-depleted arable soils, as studied by 13C NMR spectroscopy andcarbohydrate analysis. Org. Geochem. 31, 655e668.

Kleber, M., Sollins, P., Sutton, R., 2007. A conceptual model of organo-mineral in-teractions in soils: self-assembly of organic molecular fragments into zonalstructures on mineral surfaces. Biogeochemistry 85, 9e24.

Knicker, H., 2011. Soil organic N e an under-rated player for C sequestration in soils?Soil Biol. Biochem. 43, 1118e1129.

Knicker, H., Hilscher, A., Gonzalez-Vila, F.J., Almendros, G., 2008. A new conceptualmodel for the structural properties of char produced during vegetation fires.Org. Geochem. 39, 935e939.

Knicker, H., Lüdemann, H.-D., 1995. N-15 and C-13 CPMAS and solution NMR studiesof N-15 enriched plant material during 600 days of microbial degradation. Org.Geochem. 23, 329e341.

Kögel-Knabner, I., 1997. 13C and 15N NMR spectroscopy as a tool in soil organicmatter studies. Geoderma 80, 243e270.

Kögel-Knabner, I., 2000. Analytical approaches for characterizing soil organicmatter. Org. Geochem. 31, 609e625.

Kögel-Knabner, I., 2002. The macromolecular organic composition of plant andmicrobial residues as inputs to soil organic matter. Soil Biol. Biochem. 34,139e162.

Kölbl, A., Kögel-Knabner, I., 2004. Content and composition of free and occludedparticulate organic matter in a differently textured arable Cambisol as revealedby solid-state C-13 NMR spectroscopy. J. Plant Nutr. Soil Sci. 167, 45e53.

Lehmann, J., Rillig, M.C., Thies, J., Masiello, C.A., Hockaday, W.C., Crowley, D., 2011.Biochar effects on soil biota e a review. Soil Biol. Biochem. 43, 1812e1836.

Leifeld, J., Siebert, S., Kögel-Knabner, I., 2001. Stabilization of composted organicmatter after application to a humus-free sandy mining soil. J. Environ. Qual. 30,602e607.

Leinweber, P., Schulten, H.-R., 2000. Nonhydrolyzable forms of soil organic nitro-gen: extractability and composition. J. Plant Nutr. Soil Sci. 163, 433e439.

Madhok, M.R., 1937. Synthetic soil as a medium for the study of certain microbio-logical processes. Soil Sci. 44, 319e322.

Miltner, A., Kindler, R., Knicker, H., Richnow, H.-H., Kästner, M., 2009. Fate of mi-crobial biomass-derived amino acids in soil and their contribution to soilorganic matter. Org. Geochem. 40, 978e985.

Miltner, A., Zech, W., 1999. Microbial degradation and resynthesis of proteins duringincubation of beech leaf litter in the presence of mineral phases. Biol. Fertil.Soils 30, 48e51.

Nannipieri, P., Eldor, P., 2009. The chemical and functional characterization of soil Nand its biotic components. Soil Biol. Biochem. 41, 2357e2369.

Oades, J., 1984. Soil organic matter and structural stability: mechanisms and im-plications for management. Plant Soil 76, 319e337.

Paul, E.A., Morris, S.J., Conant, R.T., Plante, A.F., 2006. Does the acid hydrolysis e

incubation method measure meaningful soil organic carbon pools? Soil Sci. Soc.Am. J. 70, 1023e1035.

Pronk, G.J., Heister, K., Ding, G.-C., Smalla, K., Kögel-Knabner, I., 2012. Developmentof biogeochemical interfaces in an artificial soil incubation experiment; ag-gregation and formation of organo-mineral associations. Geoderma 189-190,585e594.

Saidy, A.R., Smernik, R.J., Baldock, J.A., Kaiser, K., Sanderman, J., Macdonald, L.M.,2012. Effects of clay mineralogy and hydrous iron oxides on labile organiccarbon stabilisation. Geoderma 173e174, 104e110.

Schmidt, M.W.I., Rumpel, C., Kögel-Knabner, I., 1999. Evaluation of an ultrasonicdispersion procedure to isolate primary organomineral complexes from soils.Eur. J. Soil Sci. 50, 87e94.

Silveira, M.L., Comerford, N.B., Reddy, K.R., Cooper, W.T., El-Rifai, H., 2008. Charac-terization of soil organic carbon pools by acid hydrolysis. Geoderma 144, 405e414.

Simpson, M., Simpson, A., 2012. The chemical ecology of soil organic matter mo-lecular constituents. J. Chem. Ecol. 38, 768e784.

Six, J., Bossuyt, H., Degryze, S., Denef, K., 2004. A history of research on the linkbetween (micro)aggregates, soil biota, and soil organic matter dynamics. SoilTillage Res. 79, 7e31.

Six, J., Frey, S.D., Thiet, R.K., Batten, K.M., 2006. Bacterial and fungal contributions tocarbon sequestration in agroecosystems. Soil Sci. Soc. Am. J. 70, 555e569.

Sohi, S.P., Krull, E., Lopez-Capel, E., Bol, R., 2010. A review of biochar and its use andfunction in soil. Adv. Agron. 105, 47e82.

Stevenson, F.J., 1996. Nitrogen-organic forms. In: Methods of Soil Analysis. Part 3.Chemical Methods. Soil Science Society of America and American Society ofAgronomy, Madison, WI, USA, pp. 1185e1200.

Stotzky, G., 1986. Influence of soil mineral colloids on metabolic processes, growth,adhesion, and ecology of microbes and viruses. In: Huang, P.M., Schnitzer, M.(Eds.), Interactions of Soil Minerals with Natural Organics and Microbes. SoilScience Society of America, Madison, Wisconsin, USA, pp. 305e428.

Wattel-Koekkoek, E.J.W., Buurman, P., Van Der Plicht, J., Wattel, E., Van Breemen, N.,2003. Mean residence time of soil organic matter associated with kaolinite andsmectite. Eur. J. Soil Sci. 54, 269e278.