Impact of agricultural fertilization practices on organo

Science of the Total Environment 651 (2019) 591–600

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Impact of agricultural fertilization practices on organo-mineral associations in four long-term field experiments: Implications for soil C sequestration Yongli Wen a, Wenjuan Liu a, Wenbo Deng a, Xinhua He d,e, Guanghui Yu b,c,⁎ a

Institute of Loess Plateau, Shanxi University, Taiyuan 030006, China Institute of Surface-Earth System Science, Tianjin University, Tianjin 300072, China c Jiangsu Provincial Key Lab for Organic Solid Waste Utilization, Nanjing Agricultural University, Nanjing 210095, China d Centre of Excellence for Soil Biology, College of Resources and Environment, Southwest University, Chongqing 400715, China e School of Biological Sciences, University of Western Australia, Perth, WA 6009, Australia b

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Fertilization regimes had contrasting effects on Fe\\C associations in soils. • Combined spectroscopic methods were applied to evaluate the species of Fe and C. • Poorly crystalline Fe minerals were closely correlated to aromatic C accumulations. • Long-term organic fertilization improved the capacity of soil organic carbon storage.

a r t i c l e

i n f o

Article history: Received 1 August 2018 Received in revised form 18 September 2018 Accepted 18 September 2018 Available online 19 September 2018 Editor: Shuzhen Zhang Keywords: Aromatic functional groups Long-term fertilization Organo-mineral associations Poorly crystalline Fe minerals SOC storage

a b s t r a c t Soil organic carbon (SOC) associated with minerals is considered to be one of the most fundamental long-term SOC storage strategies, but little research has integrated the organo-mineral complexes regulated by long-term fertilization. Here, soil samples under three fertilization treatments (Control, no fertilization; NPK, chemical nitrogen, phosphorus and potassium fertilization; NPKM, NPK plus manure) from four 23–34 years long-term field experiment sites across China were examined. Chemical analyses indicated that vigorous iron (Fe) mobilization could be regulated by long-term fertilization regimes. Meanwhile, Fe K-edge X-ray absorption near-edge fine structure (XANES) demonstrated that compared to NPK treated soils, NPKM treated soils contained significantly higher concentration of poorly crystalline ferrihydrite. Results from both the Fourier transform infrared combined with two-dimensional correlation spectroscopy analyses (FTIR-2DCOS) and C 1 s X-ray photoelectron spectroscopy (XPS) revealed that aliphatic carbohydrate might play an important role in binding exogenous Fe (III) in all tested four soils. In addition, greater amounts of aromatic C (the most resistant soil C fraction) were under long-term treated NPKM than NPK soils. Furthermore, multiple regression analyses showed a significantly positive relationship between poorly crystalline Fe minerals and SOC or aromatic C. Such relationships indicated that aromatic functional groups had been attached to the poorly crystalline Fe minerals, which could also be

⁎ Corresponding author at: College of Resources and Environmental Sciences, Nanjing Agricultural University, 1 Wei Gang, Nanjing 210095, PR China. E-mail address: [email protected] (G. Yu).

https://doi.org/10.1016/j.scitotenv.2018.09.233 0048-9697/© 2018 Elsevier B.V. All rights reserved.

592

Y. Wen et al. / Science of the Total Environment 651 (2019) 591–600

protected from being transformed to the crystalline counterpart. In conclusion, results from our integrated spectroscopic analyses have evidenced greater improvement of both poorly crystalline Fe minerals and aromatic C in organically fertilized than in chemically fertilized soils. © 2018 Elsevier B.V. All rights reserved.

1. Introduction Globally about 2344 Pg carbon (C) is stored in soils, which is threefold greater than those in both the vegetation and atmosphere pools (Davidson and Janssens, 2006). Around 75% of this soil C pool is soil organic carbon (SOC), which plays a key role in the global C budget (Batjes, 1996). Generally, SOC is sensitive to environmental and soil conditions that could reflect the feedback of soil C cycle to climate change. Additionally, SOC is an indicator for soil fertility and quality, playing critical roles in enhancing agricultural productivity. It is of great importance to understand the mechanisms responsible for enhancing the stability and long-term storage of SOC strategies (Schmidt et al., 2011). Studies have showed that soil minerals control the SOC storage by regulating organic inputs and by protecting SOC from microbial attacks (Kleber et al., 2005; Basile-Doelsch et al., 2007). Iron (Fe) is the most abundant transitional metal on the Earth's surface and plays a particularly important role in SOC storage (Weber et al., 2006). Fe (hydr)oxides, which contain abundant of hydroxyl groups at the surfaces, can interactively establish stable organo-mineral associations by interacting with carboxyl groups of SOC (Kleber et al., 2005). Thus SOC is more resistant to degradation when adsorbed to Fe (hydr)oxides. Fe (hydr)oxide sorbent types, such as poorly crystalline ferrihydrite, crystalline goethite and lepidocrocite, play key roles in the SOC preservation because of their variable surface area and reactivity (Cismasu et al., 2016). Meanwhile, up to 21.5% of the organic C in soil is bound to the reactive poorly crystalline Fe minerals (Lalonde et al., 2012). More than 37% of the Earth's land surface is consisted of agricultural land, where human activities strongly affect its SOC stock (The World Bank, 2015). Fertilization, as an important agronomic practice, has long been used to enhance plant growth, boost SOC stocks, and improve soil structure. Continuous organic manure inputs to soil have led to a higher SOC accumulation when compared to non-fertilization and chemical fertilization (Maillard and Angers, 2014). Such a higher C storage is usually due to a direct C input from the manure and an indirect C input through an increased net primary production (Maillard and Angers, 2014). Chemical fertilizers may also indirectly enhance the SOC storage with the crop residue input to soil (Halvorson et al., 1999; Galantini and Rosell, 2006; Tong et al., 2014). On the other hand, soil minerals could be affected by fertilization practices, as fertilization dramatically changes the soil's physical, chemical and biological properties. For instance, long-term application of organic manure increased soil macro-aggregate content for N54% (Tong et al., 2014). Greater macro-aggregate content in organically fertilized soils increased the potential for more anoxic environments for Fe redox microbes, thus supporting diverse and abundant anaerobic microbial communities (Reeve et al., 2010; Wen et al., 2018). These changes could greatly influence the Fe bioavailability and its interaction with other biologically important nutrients. For instance, soil cation exchange capacity positively related to the accumulation of SOC over a decade of intensive graze (Machmuller et al., 2015). In addition, root exudates had important effects on the formation of reactive minerals (Keiluweit et al., 2015). Numerous long-term fertilization experiments have been established to address the impact of fertilizer amendments on crop production while providing valuable information to maintain soil fertility by examining changes in soil processes over time. However, information is scarce in depicting the mechanisms that could explain how fertilization regulates the association between soil Fe minerals and SOC accumulation. Besides, the soil water-dispersible colloids act as the linkage between geosphere and hydrosphere, representing the most reactive component in soils.

They typically consist of mixtures of soil minerals and natural organic matter, and are readily affected by fertilization regimes (Schumacher et al., 2005; Van Oost et al., 2007). A better knowledge of the interaction of Fe species with C functional groups in water-dispersible colloids from different fertilizer-treated soils may contribute to our understanding of soil C sequestration. The Fourier transform infrared spectroscopy (FTIR) is a popular facility to address the organic ligands in various environments, which can differentiate both fluorescent and non-fluorescent substances in comparison with fluorescence excitation–emission matrix (EEM) spectroscopy. However, one-dimensional FTIR usually exhibits a variety of overlapped peaks because of the heterogeneous nature of an examined SOC. Combined with a two-dimensional correlation spectroscopy analyses (2DCOS), it is able to solve such peak overlapping problems by distributing the spectral intensity within a data set along a second dimension. This method could provide valuable information of the complexes of organic ligands with metals in soil dissolved organic matter (DOM) (Yu et al., 2012; Wen et al., 2014a). Recent innovations in the synchrotron-based X-ray absorption near-edge fine structure (XANES) spectroscopy, which is an element-specific technique and sensitive to both the oxidation state and the local structure of the absorber element (Prietzel et al., 2007), make it possible not only identify but also quantify, the mineralogy of Fe presented in soils. Since Fe phases in soils are highly complex, this tool is helpful in comparison with the Mössbauer spectroscopy that might mask magnetically weak phases (Huang et al., 2016). In addition, the X-ray photoelectron spectroscopy (XPS) can provide valuable information on the bonding state of C framework in the surface layers of soil particles, examine the inherently stable structures of SOC, and thus the nature of SOC transformations under different fertilization regimes (Mikutta et al., 2009; Xiao et al., 2015). This study employed four long-term field sites locating in different regions (Gongzhuling, Shenyang, Urumqi, and Qiyang) of China with four contrasting soil types (black soil, Luvic Phaeozems; brown soil, Haplic Alisol; desert soil, Haplic Calcisol; and red soil, Ferrialic Cambisol). All experiments have been performed over 23–34 years. Three soils from each site were examined under three fertilization treatments: (1) no fertilization control; (2) chemical nitrogen, phosphorus and potassium, NPK; and (3) NPK plus manure, NPKM. Several novel techniques including FTIR2DCOS, XANES and XPS were combined to gain direct insight into the local coordination and characteristics of Fe minerals and their function in preserving SOC. Our objectives were to determine: (1) if the functional composition and speciation of Fe and C in soil colloids could be influenced by long-term fertilization; (2) which functional groups in the soil colloids C were preferentially binding with Fe oxides and (3) how long-term fertilization could affect the interaction between Fe matrix and C fraction in the soil colloids. These results would increase our knowledge of how agricultural fertilization practices could influence SOC by influencing the interaction between SOC and Fe oxides and provide insights into the turnover and preservation of SOC in agricultural soils. 2. Material and methods This study consisted of four long-term (23–34 years) experimental sites on upland soils throughout the major grain-producing areas of China (Fig. 1). The location, climate, cropping regimes, and basic soil properties from these sites at the start of the field experiment were summarized in Table S1, which also identified the major soil types of these sites, based on the soil classification of the Food and Agriculture Organization (FAO) of the United Nations.


593

Fig. 1. Concentrations of total Fe in soils (left Y axis) and in water-dispersible soil colloids (right Y axis) at four long-term (23–34) fertilization sites. Data were shown as means ± SD (n = 3).

Soils under the following three fertilization treatments were collected from these long-term experimental sites or plots: (1) without fertilization (Control), (2) chemical fertilization as nitrogen, phosphorus and potassium fertilizers (NPK), and (3) a combination of chemical fertilization and organic amendments (NPKM). Fertilization rates and fertilizer types were summarized in Table S2. All plots were arranged in a randomized block design. Each plot was 20 m long and 10 m wide with a 1.0-m deep cement barrier zone between plots. Soils at 0–20 cm depth were collected during May 2013 using a 5-cm internal diameter auger. Each plot was evenly separated into three subplots, and 5 cores (~50 g soil each) from each subplot were randomly sampled. Each soil sample was composited from 15 random cores collected from a single plot. The fresh soil was mixed thoroughly, air-dried, and sieved through 2-mm and 0.25-mm screens for further analyses.

Soil water-dispersible colloids were isolated with deionized water (1:5 w:v) on a horizontal shaker (170 rpm) at 25 ± 1 °C for 8 h and then centrifuged for 6 min at 2500g (Schumacher et al., 2005). For the DOM extraction, the supernatant suspension of soil colloids was passed through a 0.45-μm membrane filter and further diluted until the dissolved organic carbon (DOC) was b10 mg L−1. Both the soil colloids and DOM supernatant suspensions were stored under darkness at 4 °C for further analyses. 2.1. Chemical analyses Soil pH was measured with a pH meter (soil:distilled water = 1:5). The concentrations of SOC and soil total N (TN) were quantified using a CN analyzer (Vario EL, Elementar GmbH). Samples were wet-digested

594


with a concentrated acid mixture (HNO3, HClO4, and HF) to measure total Fe in soils (Zhou et al., 2007). To determine the Fe in the waterdispersible soil colloids (Feco), the soil colloids were mixed with 10% nitric acid at a ratio of 1:1 (v:v), incubated at 150 °C for 2 h, and then filtered through a 0.45-μm membrane. The quantitative mineralogical analyses of Fe (hydr)oxides in soil samples was performed using selective dissolution procedures. The reactive Fe minerals (Fed) were extracted using a 0.1 M DithioniteCitrate-Bicarbonate (DCB) solution (Lalonde et al., 2012). The poorly crystalline Fe-(hydr)oxides and organically complexed Fe in soil (Feo) were extracted with acid ammonium oxalate at pH 3, with a soil solution ratio of 1:50, and shaken for 4 h under dark. The organically complexed Fe (Fep) was measured after the extraction with 0.1 M sodiumpyrophosphate for 16 h at pH 10. The amount of Fe in the extracts was measured using an Inductively Coupled Plasma-Atomic Emission Spectrometry (ICP-AES, Teledyne Leeman Labss, Hudson, NH 03051, USA). The Feo minus Fep is an indication of the poorly crystalline Fe in bulk soils. 2.2. Fe(III) titration, FTIR measurements and 2DCOS analyses To eliminate free metals, the DOM extracts were passed through a H+-saturated cation exchange resin (AG 50 W-X8; Bio-Rad, Hercules, CA, USA) prior to titration. The Fe(III) titration was conducted in duplicates by adding 0.1 M Fe to a series of brown sealed bottles that contained 30 ml DOM fraction. No N30 μl Fe(III) was titrated into a brown bottle, making the titration range in the final solution being 0–100 μM Fe(III) concentrations in 10 μM steps. To avoid precipitation, pH was maintained at 6.0 with 0.05 M NaOH or HCl after the titration. All solutions were shaken for 24 h to ensure complexation equilibrium. After freeze-drying, the samples were analyzed by Nicolet 370 a FTIR spectroscopy (Thermo Nicolet, USA). The freeze-dried DOM-Fe(III) complex were mixed with potassium bromide (IR grade; sample to KBr ratio = 1:100, w:w), and then this mixture was ground and homogenized. A potassium bromide window was formed by compressing a subsample of the mixture in a hydraulic pressure at 20,000 psi between two clean, polished anvils. The FTIR spectra were obtained by collecting 200 scans over the range of 4000–400 cm−1 with a FTIR spectrometer at 4 cm−1 of resolution. The 2DCOS spectra were produced according to the method of Noda and Ozaki (2004), as explained in details in our previous publications (Yu et al., 2012; Wen et al., 2014a). The spectral data set were normalized by the summed intensities and multiplied by a factor of 1000 after removing the CO2 peak and smoothing the FTIR spectral image. The noise components were reduced by the principal component analyses (PCA). Finally, the 2DCOS analyses were conducted using a 2D Shige software (Kwansei-Gakuin University, Japan). In this study, we focused on the FTIR region from 1800 to 900 cm−1, which contains the major excitation bands of the amide, carboxylic acid, ester, and carbohydrate functional groups (Yu et al., 2012). 2.3. C 1s XPS and Fe K-edge XANES analyses We used the freeze-dried samples of the water-dispersible soil colloids to conduct the C 1 s XPS and Fe K-edge XANES experiments. After seizing to a size of b200 μm, the soil colloids were analyzed in triplicates by XPS using a PHI 5000 Versa Probe X-ray photoelectron spectrometer (UlVAC-PHI5000 VersaProbe, Japan) equipped with a monochromated Al Kα X-ray radiation detector (1486.69 eV). The surface composition analyses of soil colloids up to a depth of ~5 nm included a survey element scan and a high-resolution scan at the C 1s edge. The adventitious hydrocarbon C 1s spectrum (C 1s = 284.8 eV) was used to adjust the binding energy scale. The analyzed area corresponded to a 300 μm by 300 μm elliptical spot. A flood gun at 6 eV was used to balance the surface charge that was induced by the photo ejection process. To optimize the signal-to-noise ratio, the spectra

were recorded at a detector resolution corresponding to 0.125 eV per channel. The base pressure in the spectrometer was 6.7 × 10−10 Torr. The spectra were deconvoluted by a non-linear least squares curve fitting program (XPSPEAK Version 4.1, Hong Kong, China) with a Gaussian-Lorentzian mix function and Shirley background correction. The identification of binding energies was done according to Moulder et al. (1992). The chemical composition of soil colloids was examined by deconvoluting the C 1 s peak into sub-peaks by fitting with the Gaussian–Lorentzian functions, which were freely fit for all of the peaks, and then assigned to different C environments (NIST XPS database, see Monteil-Rivera et al., 2000). No fixed full-width-at-halfmaximum values were determined for the spectra of the soil colloids collected under different fertilization treatments in different sites. After grinding into fine powders, the freeze-dried soil colloids from the four long-term experiment sites were brushed onto tapes and stacked together for the Fe K-edge XANES analyses. Two to six replicated spectra were collected using a Si (111) double-crystal monochromator at the XAFS station on the Beamline 14W of the Shanghai Synchrotron Radiation Facility (SSRF, Shanghai, China). The storage ring was operated at 3.5 GeV, with an electron current decreasing from 240 to 160 mA within approximately 8 h. The intensities of the incident and the transmitted X-rays were monitored in ionization chambers that were filled with nitrogen gas. All XANES data were recorded in the transmission detection mode at room temperature. The incident Xray for the XANES spectra varied from 7100 to 7250 eV with 0.1-eV steps and timing for 10 s. The X-ray energy scale was calibrated to the Fe K-edge (7112.0 eV) using a Fe foil prior to the XAFS acquisition. The spectral data processing was performed using an Athena software (version 2.1.1). A linear combination fitting (LCF) analysis was applied to quantify the percentage of Fe mineral phases in different samples using Ferrihydrite, Goethite, Lepidocrocite, Magnetite, Fe2(SO4)3, FeSO4, Ferric oxalate and Ferrous oxalate as the standard samples, respectively. 2.4. Statistical analyses Statistical analyses of data (means ± SD, n = 3) were performed in the R Studio (Version: 0.99.903). A one-way analyses of variance (ANOVA) was employed to test the effects of long-term fertilization on soil pH, SOC, TN, C/N, and Fe concentration in soils and waterdispersible soil colloids. The principal component analysis was conducted with the “stats” package in R. The multiple linear regression selection was performed with regsubsets in the “leaps” package for SOC and for different C fractions from the long-term fertilization study soils. Regsubsets() selects the best candidate model for each possible number of parameters from 2 to 6, including the intercept, with an exhaustive search that includes every combination of parameters at each level. The dependent variables for SOC and C fractions included Fe fractions used in the XANES fitting (i.e., Ferrihydrite, Goethite, Lepidocrocite, Magnetite, Fe2(SO4)3). The Pearson's correlation coefficient (r) was used to evaluate the linear correlation between reactive Fe minerals and C functional groups. Statistical significances were considered at a P b 0.05. 3. Results 3.1. Soil physicochemical properties, Fe fractions in the field samples In general, soil pH, SOC, TN, C/N and Fe fractions varied significantly among different sites and soils under different fertilization treatments (Table 1, Fig. 1). Soil pH was more acidic in the Red and Brown soils than in the Black and Desert soils. Desert soil had the highest pH (7.79–8.02) (Table 1), whereas Red soils had the lowest pH (4.05–6.01). Within each site, soil pH varied over the long-term fertilization time and fertilization treatments. Compared to Control, NPKM significantly increased the soil pH by 0.23 in Brown soil and 0.66 in


595

Table 1 Soil physicochemical characteristics and Fe fractions under three fertilization treatments in four long-term fertilization experiment sites. Site

Soil type

Fertilization regime

pH

SOC/g kg−1

TN/g kg−1

C/N

Fed/g kg−1

Feo/g kg−1

Fep/g kg−1

Feam/g kg−1

Gongzhuling

Black soil

Shenyang

Brown soil

Urumqi

Desert soil

Qiyang

Red soil

Control NPK NPKM Pa F-ratio Control NPK NPKM P F-ratio Control NPK NPKM P F-ratio Control NPK NPKM P F-ratio

7.57 (0.08) 6.41 (0.02) 7.56 (0.05) b0.001 420.9 5.92 (0.07) 5.09 (0.09) 6.15 (0.12) b0.001 105.4 8.02 (0.03) 7.79 (0.01) 8.01 (0.05) b0.001 39.97 5.35 (0.04) 4.05 (0.02) 6.01 (0.03) b0.001 3037

14.29 (0.21) 13.21 (0.27) 24.56 (0.19) b0.001 2267 9.08 (0.17) 9.90 (0.19) 14.05 (0.35) b0.001 331.5 13.59 (0.16) 14.34 (0.22) 30.39 (1.49) b0.001 353.8 7.27 (0.23) 10.43 (0.24) 13.95 (0.48) b0.001 298.2

1.40 (0.06) 1.39 (0.02) 2.50 (0.01) b0.001 847.8 1.00 (0.04) 1.18 (0.02) 1.73 (0.03) b0.001 419.9 1.04 (0.05) 1.10 (0.02) 2.63 (0.13) b0.001 346 0.92 (0.14) 1.34 (0.02) 1.70 (0.02) b0.001 70.13

10.20 (0.34) 9.51 (0.22) 9.84 (0.05) b0.05 6.27 9.07 (0.23) 8.36 (0.28) 8.12 (0.12) b0.01 15.32 13.07 (0.65) 13.05 (0.08) 11.54 (0.09) b0.01 15.82 8.05 (1.42) 7.77 (0.31) 8.23 (0.34) N0.05 0.21

13.35 (0.21) 12.79 (0.09) 11.14 (0.17) b0.001 151.5 8.81 (0.11) 9.49 (0.27) 7.98 (0.19) b0.001 42.65 6.33 (0.18) 6.02 (0.29) 4.98 (0.79) b0.05 6.10 40.01 (1.09) 48.12 (2.01) 41.52 (1.05) b0.01 26.41

1.93 (0.05) 2.23 (0.11) 2.05 (0.03) b0.01 14.17 4.10 (0.03) 4.02 (0.21) 3.65 (0.08) b0.05 10.44 1.07 (0.06) 1.31 (0.02) 1.26 (0.02) b0.001 35.12 3.12 (0.06) 3.88 (0.04) 4.15 (0.07) b0.001 247.4

0.22 (0.00) 0.73 (0.03) 0.43 (0.01) b0.001 530.2 1.52 (0.08) 2.44 (0.40) 1.84 (0.09) b0.01 11.09 0.18 (0.01) 0.33 (0.02) 0.18 (0.01) b0.001 76.55 2.65 (0.05) 3.53 (0.28) 2.72 (0.24) b0.01 16.04

1.71 (0.05) 1.49 (0.08) 1.62 (0.03) b0.01 11.83 2.58 (0.11) 1.58 (0.32) 1.81 (0.17) b0.01 16.87 0.89 (0.07) 0.98 (0.03) 1.08 (0.03) b0.01 11.54 0.48 (0.06) 0.35 (0.29) 1.43 (0.24) b0.01 21.38

a. P (significant level) and F ratio in ANOVA. Data are Means ± SD, n = 3.

Red soil, whereas no significant increase was detected in the Black and Desert soils, both had relatively high pH values. The sole NPK fertilization acidified the soil significantly and their soil pH was decreased by 0.23–1.3 when compared with the Control. The Black and Desert soils contained greater soil C and N compared to the Red and Brown soils. In addition, higher SOC and TN were observed under NPKM under NPK and Control in all the four sites (Table 1). The C/N values were generally higher in the Black and Desert soils than in the Brown and Red soils. Within each site, soils under NPK had a higher C/N in Black, Brown and Desert soils, but a lower C/N in the Red soil (Table 1). The total Fe in bulk soil was greatest in the Red soil (~50 g kg−1), and was similar among other 3 sites (~28 g kg−1) under different fertilizations. However, Fe in the water-dispersible soil colloids varied significantly among sites (Fig. 1), and was greatest in the Red soil, especially under NPKM (~0.45 g kg−1), greater in the Desert soil (0.02–0.1 g kg−1), and least in the Black and Brown soils (ca. 0.003–0.01 g kg−1). In addition, long-term fertilization had a significant influence on the Fe concentration in the water-dispersible soil colloids. For all sites, Fe concentrations in the water-dispersible soil colloids were significantly higher under NPKM. Whereas the Fe concentration in the soil colloids was significantly decreased under NPK fertilization in the Brown, Red and Desert soils, while a significant increase under NPK in the Black soil (Fig. 1).

The Fe fractions in bulk soils were significantly affected by the longterm fertilization regimes (Table 1). The Red soil contained the greatest DCB extracted Fe (Fed), and chemical fertilization enhanced the concentration of Fed, to a larger extent than organic fertilization did. In other three sites, fertilization decreased the concentration of Fed in bulk soils, especially in soils under NPKM. The concentrations of Feo and Fep were highest under NPK than under NPKM fertilized soils across all soil types, whereas NPKM fertilization increased the concentration of poorly crystalline Fe (Feam) significantly. To determine the composition of Fe minerals in the soil colloids, LCF of the Fe K-edge XANES spectra were performed, using ferrihydrite, goethite, lepidocrocite, magnetite, Fe2(SO4)3, ferric oxalate, FeSO4, and ferrous oxalate as reference materials. The LCF results for the soil colloids (Table 2 and Fig. 2) revealed that goethite was dominant in the Black, Brown and Red soils, whereas lepidocrocite was dominant in the Desert soil. For each site, the remaining Fe phases were composed of poorly crystalline ferrihydrite species, which was significantly higher (5.4% to 14%) in soils under NPKM than under NPK. 3.2. Binding capability of functional groups in soil DOM to Fe The amide, carboxylic acid, ester and carbohydrate functional groups are mostly concentrated in the IR region of 1800–900 cm−1. The synchronous maps generated from the 1800–900 cm−1 region of the FTIR

Table 2 Linear combination fitting (LCF) of the XANES models of the adsorbed Fe spectra with the contributions (in %) of various components required to achieve the best fit*. Site

Soil type


Ferrihydrite

Goethite

Lepidocrocite

Magnetite

Fe2(SO4)3

Ferric oxalate

FeSO4

Ferrous oxalate

R-factor (*10−3)

Gongzhuling

Black soil

Shenyang

Brown soil

Xinjiang

Desert soil

Hunan

Red soil

Control NPK NPKM Control NPK NPKM Control NPK NPKM Control NPK NPKM

8.2 19.8 35.4 17.9 9.8 22.0 10.9 23.5 33.9 21.2 14.4 28.4

72.7 48.7 0 66.1 80.1 39.0 0 0 33.4 74.6 81.3 71.6

0 31.5 0 0 0 0 83.0 75.6 32.7 0 0 0

0 0 5.2 0 0 0 6.1 0 0 0 0 0

19.1 0 59.4 16.0 0 39.0 0 0 0 0 0 0

0 0 0 0 10.1 0 0 0.9 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 4.2 4.3 0

0.46 0.41 4.23 0.29 0.37 1.51 0.39 0.53 0.44 0.06 0.08 0.08

*Note: The fits were conducted based on the spectra of ferrihydrite, goethite, magnetite, lepidocrocite, Ferric oxalate, Fe2(SO4)3, FeSO4, and ferrous oxalate, respectively. Individual fractions normalized to a sum of 100% are reported, together with the effective sum of the fitted fractions. Reference and sample spectra and the LCF fits are shown in Fig. 2. This table provides the proportion of the newly formed nano-precipitates for each sample. Abbreviations: Control, no fertilization; NPK, chemical nitrogen, phosphorus and potassium fertilization; NPKM, NPK plus swine manure.

596


Fig. 2. Fits of the Fe K edge XANES spectra of soil samples at four long-term fertilization sites and standards used in the fitting process.

spectra of soil DOM under NPK and NPKM over Fe(III) at the four longterm sites are shown in Fig. 3. In general, auto-peaks appear at diagonal position and represent the overall susceptibility of the corresponding spectral region to changes in spectral intensity as an external perturbation is applied to the system. For the NPK treatment, the spectra displayed 6, 5, 1, and 2 major auto-peaks, in which the change in the band intensity followed the order of 905 N 1270 ≈ 1380 ≈ 1550 ≈ 1690 in the Black soil, 905 N 1380 N 1270 N 1550 in the Brown soil, 1380 in the Desert soil, and 1380 ≈ 1100 in the Red soil. For the NPKM treatment, only 1, 2, 1, and 1 of the major auto-peaks were observed in the Black, Brown, Desert and Red soils, respectively. The spectral band at 1380 cm−1 was consistent in the Black, Brown and Desert soils, whereas the band at 1010 cm−1 was the only observed peak in the Red soil. The various spectral bands were assigned as follows: the band at 1690 cm−1 was assigned to the C\\O stretching of amide I in the protein compounds, the band at 1550 cm−1 to N − H deformation and C\\N stretching of amide II in the protein compounds, the band at 1380 cm−1 to the CH deformations in the aliphatic groups, the band at 1120 cm−1 to the C\\OH stretching of aliphatic O\\H, and the band at 1010 cm−1 to the C\\O stretching of polysaccharides, the Si\\O of silicate impurities, or phosphate groups (Wen et al., 2014a).

3.3. Functional composition and speciation of C in water-dispersible soil colloids by C 1 s XPS techniques and relationship with Fe fraction Spectral shifts in the core level C 1 s binding energy were assigned to different chemical environments of C: (1) aromatic carbon (Ar-C-C(H): 284.2 eV), (2) aliphatic carbon (C-C(H): 284.8 eV), (3) ether or alcohol carbon (C\\O; 286.2 eV), (4) ketonic or aldehyde carbon (C\\O; 287.9 eV), and (5) carboxylic carbon (C(O)O; 289 eV) (Mikutta et al., 2009). The XPS C 1 s peak-fitting results (Table 3 and Fig. 4) demonstrated that the aliphatic carbon (C\\C/H) was dominant under all three fertilizations at all four sites, ranging from 30.92% to 62.49%. These peaks were significantly higher under NPK than under NPKM, except in the Red soil. For the Aromatic C (Ar-C-C(H)), the percentage ranged from 10.18% to 36.18%, with the highest under NPKM (20.77%– 36.18%) in each site, higher under the Control (11.81%–17.65%) and least under NPK (10.18%–11.87%). The ether or alcohol carbon (C-O (N)) ranged from 16.74% to 44.65% in all sites, with no significant changes under the three contrasting fertilization treatments. The multiple linear regression was performed to measure the relationship between Fe minerals and the chemical speciation of C in the soil colloids under different long-term fertilizations from all sites. The


597

Fig. 3. Synchronous maps generated from the 1800–900 cm−1 region of the FTIR spectra over Fe(III) of soil DOM under NPK and NPKM at the four long-term fertilization sites. The greater red or blue color intensity represents a positive or negative correlation. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

candidate model with the highest r2 explained 56% and 52% of the variation in SOC and aromatic C, respectively (Table S4), indicating that SOC and aromatic C in soil colloids were significantly positively related to ferrihydrite. There were no significant relationships between the Fe minerals and other C species. Results from the coefficient estimations agreed with those from the multiple linear regression (Table S5). 4. Discussion Comparing to the long-term NPK treatment, the NPKM treatment slightly decreased the total Fe concentration in bulk soils but significantly increased the Fe concentration in water-dispersible soil colloids in all the four tested sites. These results indicated that long-term fertilization regimes could regulate Fe mobilization in the different soil types. Furthermore, Fe K-edge XANES spectroscopy indicated that, regardless of soil types, the NPKM treatment could promote the formation of poorly crystalline Fe minerals, i.e., more ferrihydrite, compared to NPK. Although

our previous studies have shown that organic fertilization could facilitate the formation of amorphous Fe and Al minerals in the same Red soil tested in this study (Yu et al., 2012; Wen et al., 2014a; Wen et al., 2014b; Huang et al., 2016; Wen et al., 2018), this study expanded soil types, suggesting that the formation of amorphous minerals driven by the application of organic fertilizers to soil might be one of the most important mechanisms in SOC sequestration and even long-term SOC preservation. Instead of investigating the SOC molecular structure, we focused on the C involved in the soil organo-mineral associations, since this part of C also plays an important role in SOC stabilization by minerals in addition to the effects of chemical properties of mineral sorbents. Our results of FTIR-2DCOS and XPS revealed that aliphatic carbon played an important role in binding exogenous Fe in all soils (Table 3, Figs. 3 and 4), which agreed with other studies showing that soil colloids contained variable amounts of aliphatic and aromatic compounds (Kalbitz et al., 2006; Hansson et al., 2010; Klotzbücher et al., 2013). It was reported that the

598


Table 3 X-ray photoelectron spectroscopy-based speciation of carbon binding environments in soil water extractable colloids under long-term fertilization regimes* (in atom%): aromatic carbon (Ar-C-C(H)), aliphatic carbon (C-C(H)), ether or alcohol carbon (C\ \O), ketonic or aldehyde carbon (C_O), carboxylic carbon (C(O)O). Site

Soil type

Gongzhuling

Black soil

Shenyang

Brown soil

Xinjiang

Desert soil

Hunan

Red soil


Control NPK NPKM Control NPK NPKM Control NPK NPKM Control NPK NPKM

Chemical species Ar-C-C(H) 284.2 eV

C\ \C(H) 284.8 eV

C\ \O 286.2 eV

C_O 287.9 eV

C(O)O 289 eV

14.06 10.99 23.06 11.81 11.87 20.77 17.65 11.39 36.18 17.12 10.18 20.78

46.32 49.03 30.92 35.44 62.49 53.37 60.21 40.13 35.52 60.41 40.78 47.66

27.36 39.20 30.05 37.05 16.74 17.33 21.25 31.99 25.85 13.20 44.65 22.42

1.77 N.D 8.00 8.52 7.33 4.97 0.90 6.09 2.44 3.70 3.05 5.89

10.50 0.78 7.97 7.18 1.57 3.55 N.D 10.41 N.D 5.57 1.34 3.27

*Abbreviations: Control, no fertilization; NPK, chemical nitrogen, phosphorus and potassium fertilization; NPKM, NPK plus swine manure.

stubborn organic fraction, enriched in aliphatic compounds, did not substantially interact with kaolinite, smectite or poorly crystalline Fe because this fraction was bound to crystalline Fe oxides (Barbera et al., 2008). In this study, crystalline Fe oxides were found positively correlated with aliphatic C, though not significantly (P N 0.05, Table S5). In addition, the results obtained by XPS showed that greater amounts of aromatic C were

observed under NPKM than NPK (Table 3 and Fig. 4). Aromatic compounds, which originate from either lignin, tannins, or low molecular components (e.g., simple phenols, phenolic acids, and flavonoids), can affect soil processes by interacting with metals (Hättenschwiler and Vitousek, 2000; Kraus et al., 2003). These aromatic structures, being most resistant to biodegradation, were found to be preferentially sorbed

Fig. 4. XPS peak-fitting images recorded from water-dispersible soil colloids extracted under Control, NPK, NPKM treatments at four long-term fertilization sites across China.


onto reactive poorly crystalline Fe oxides, which might be a major mechanism for long-term SOC sequestration in soil (Kramer et al., 2012; Rumpel et al., 2015). Mineralogy can play a role in the preservation of specific functional groups (Baldock et al., 1992; Stevenson et al., 2016). Therefore, the increase of aromatic C under the NPKM treatment observed in all soil types tested in this study (Table 3 and Fig. 4), might be a result of a larger amount of ferrihydrite being produced in organically fertilized soils, although other minerals such as kaolinite, smectite and allophane might play important roles in binding with aromatic C, which were not addressed in the present study. The best candidate model from the multiple regression analyses revealed a significant positive relationship between SOC/aromatic C and ferrihydrite (Table S4), indicating that the aromatic functional groups had been attached to poorly crystalline Fe minerals, i.e., ferrihydrite, which was highest in organically fertilized soils. Our results also confirmed the findings of Kramer et al. (2012), who noted that aromatics might be preferentially stored in short-range minerals, i.e., with poorly crystalline Fe minerals, as long-term C storage by these minerals occurred via the mechanism of chemical retention with dissolved aromatic acids (Kramer et al., 2012). Take results from those and our studies together, we concluded that various long-term fertilization regimes could differentially regulate Fe mineral compositions, resulting in different Fe mineralogy in soils. That is to say, in organically fertilized soils more poorly crystalline Fe minerals could occur to and store larger amounts of SOC, particularly the aromatic C fraction, which in turn would protect the poorly crystalline Fe minerals from transforming to their crystalline counterpart. However, further work is needed to explore the interactive mechanism of soil organic C fractions and ferrihydrite in soils. 5. Conclusions This study demonstrated that vigorous Fe transformation could be regulated by long-term fertilization regimes. Specifically, for all four tested sites, an increased concentration of poorly crystalline Fe minerals was presented in soils treated with organic fertilizers, whereas an increased concentration of crystalline Fe minerals was in soils treated with NPK fertilizer. In addition, aliphatic carbohydrate might play an important role in binding exogenous Fe(III). However, the aromatic C fraction, being most resistant to biodegradation, was significantly greater under NPKM than under NPK in all the tested four soils. Furthermore, both SOC and aromatic C in the soil colloids were positively related to ferrihydrite in those soils, indicating that aromatic functional groups had been attached to poorly crystalline Fe minerals, in turn protecting the poorly crystalline Fe minerals from being transformed to their crystalline counterpart, especially in organically fertilized soils. Collectively, we demonstrated that long-term organic fertilization could stimulate the formation of poorly crystalline Fe minerals and aromatic C fraction involving in the mineral associated C, regardless of the soil types, which would improve the capacity of long-term SOC sequestration. These findings provide insights into the organo-mineral associations regulated by agricultural fertilization and expand possibilities for further investigating the regulation of soil Fe minerals and their roles in efficient soil C storage. Acknowledgments We thank the staff at the BL14W1 beamline of the Shanghai Synchrotron Radiation Facility (SSRF). This work was funded by the National Key Research and Development Program of China (2017YFD0800803) and the National Natural Science Foundation of China (41371248 and 41371299). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.scitotenv.2018.09.233.

599

References Baldock, J.A., Oades, J.M., Waters, A.G., Peng, X., Vassallo, A.M., Wilson, M.A., 1992. Aspects of the chemical structure of soil organic materials as revealed by solid-state 13C NMR spectroscopy. Biogeochemistry 16, 1–42. Barbera, V., Raimondi, S., Egli, M., Plötze, M., 2008. The influence of weathering processes on labile and stable organic matter in Mediterranean volcanic soils. Geoderma 143, 191–205. Basile-Doelsch, I., Amundson, R., Stone, W.E.E., Borschneck, D., Bottero, J.Y., Moustier, S., Masin, F., Colin, F., 2007. Mineral control of carbon pools in a volcanic soil horizon. Geoderma 137, 477–489. Batjes, N.H., 1996. Total carbon and nitrogen in the soils of the world. Eur. J. Soil Sci. 47, 151–163. Cismasu, A.C., Williams, K.H., Nico, P.S., 2016. Iron and carbon dynamics during aging and reductive transformation of biogenic ferrihydrite. Environ. Sci. Technol. 50, 25–35. Davidson, E.A., Janssens, I.A., 2006. Temperature sensitivity of soil carbon decomposition and feedbacks to climate change. Nature 440, 165–173. Galantini, J., Rosell, R., 2006. Long-term fertilization effects on soil organic matter quality and dynamics under different production systems in semiarid Pampean soils. Soil Tillage Res. 87, 72–79. Halvorson, A.D., Reule, C.A., Follett, R.F., 1999. Nitrogen fertilization effects on soil carbon and nitrogen in a dryland cropping system. Soil Sci. Soc. Am. J. 63, 912–917. Hansson, K., Kleja, D.B., Kalbitz, K., Larsson, H., 2010. Amounts of carbon mineralised and leached as DOC during decomposition of Norway spruce needles and fine roots. Soil Biol. Biochem. 42, 178–185. Hättenschwiler, S., Vitousek, P.M., 2000. The role of polyphenols in terrestrial ecosystem nutrient cycling. Trends Ecol. Evol. 15, 238–243. Huang, C., Liu, S., Li, R., Sun, F., Zhou, Y., Yu, G., 2016. Spectroscopic evidence of the improvement of reactive iron mineral content in red soil by long-term application of swine manure. PLoS One 11, e0146364. Kalbitz, K., Kaiser, K., Bargholz, J., Dardenne, P., 2006. Lignin degradation controls the production of dissolved organic matter in decomposing foliar litter. Eur. J. Soil Sci. 57, 504–516. Keiluweit, M., Bougoure, J.J., Nico, P.S., Pett-Ridge, J., Weber, P.K., Kleber, M., 2015. Mineral protection of soil carbon counteracted by root exudates. Nat. Clim. Chang. 5, 588–595. Kleber, M., Mikutta, R., Torn, M., Jahn, R., 2005. Poorly crystalline mineral phases protect organic matter in acid subsoil horizons. Eur. J. Soil Sci. 56, 717–725. Klotzbücher, T., Kaiser, K., Filley, T.R., Kalbitz, K., 2013. Processes controlling the production of aromatic water-soluble organic matter during litter decomposition. Soil Biol. Biochem. 67, 133–139. Kramer, M.G., Sanderman, J., Chadwick, O.A., Chorover, J., Vitousek, P.M., 2012. Long-term carbon storage through retention of dissolved aromatic acids by reactive particles in soil. Glob. Chang. Biol. 18, 2594–2605. Kraus, T.E.C., Dahlgren, R.A., Zasoski, R.J., 2003. Tannins in nutrient dynamics of forest ecosystems - a review. Plant Soil 256, 41–66. Lalonde, K., Mucci, A., Ouellet, A., Gelinas, Y., 2012. Preservation of organic matter in sediments promoted by iron. Nature 483, 198–200. Machmuller, M.B., Kramer, M.G., Cyle, T.K., Hill, N., Hancock, D., Thompson, A., 2015. Emerging land use practices rapidly increase soil organic matter. Nat. Commun. 6, 6995. Maillard, E., Angers, D.A., 2014. Animal manure application and soil organic carbon stocks: a meta-analysis. Glob. Chang. Biol. 20, 666–679. Mikutta, R., Schaumann, G.E., Gildemeister, D., Bonneville, S., Kramer, M.G., Chorover, J., Chadwick, O.A., Guggenberger, G., 2009. Biogeochemistry of mineral–organic associations across a long-term mineralogical soil gradient (0.3–4100 kyr), Hawaiian Islands. Geochim. Cosmochim. Acta 73, 2034–2060. Monteil-Rivera, F., Brouwer, E.B., Masset, S., Deslandes, Y., Dumonceau, J., 2000. Combination of X-ray photoelectron and solid-state 13C nuclear magnetic resonance spectroscopy in the structural characterisation of humic acids. Anal. Chim. Acta 424, 243–255. Moulder, J.F., Stickle, W.F., Sobol, P.E., Bomben, K.D., 1992. Handbook of X-ray Photoelectron Spectroscopy. Perkin-Elmer Corp, Eden Prairie. Noda, I., Ozaki, Y., 2004. Two-Dimensional Correlation Spectroscopy-Applications in Bibrational and Optical Spectrocopy. John Wiley & Sons, England. Prietzel, J., Thieme, J., Eusterhues, K., Eichert, D., 2007. Iron speciation in soils and soil aggregates by synchrotron-based X-ray microspectroscopy (XANES, μ-XANES). Eur. J. Soil Sci. 58, 1027–1041. Reeve, J.R., Schadt, C.W., Carpenter-Boggs, L., Kang, S., Zhou, J., Reganold, J.P., 2010. Effects of soil type and farm management on soil ecological functional genes and microbial activities. ISME J. 4, 1099–1107. Rumpel, C., Baumann, K., Remusat, L., Dignac, M.F., Barré, P., Deldicque, D., Glasser, G., Lieberwirth, I., Chabbi, A., 2015. Nanoscale evidence of contrasted processes for root-derived organic matter stabilization by mineral interactions depending on soil depth. Soil Biol. Biochem. 85, 82–88. Schmidt, M.W.I., Torn, M.S., Abiven, S., Dittmar, T., Guggenberger, G., Janssens, I.A., Kleber, M., Kogel-Knabner, I., Lehmann, J., Manning, D.A.C., Nannipieri, P., Rasse, D.P., Weiner, S., Trumbore, S.E., 2011. Persistence of soil organic matter as an ecosystem property. Nature 478, 49–56. Schumacher, M., Christl, I., Scheinost, A.C., Jacobsen, C., Kretzschmar, R., 2005. Chemical heterogeneity of organic soil colloids investigated by scanning transmission X-ray microscopy and C-1s NEXAFS microspectroscopy. Environ. Sci. Technol. 39, 9094–9100. Stevenson, B.A., Sarmah, A.K., Smernik, R., Hunter, D.W.F., Fraser, S., 2016. Soil carbon characterization and nutrient ratios across land uses on two contrasting soils: their relationships to microbial biomass and function. Soil Biol. Biochem. 97, 50–62. The World Bank, 2015. World Bank Data. https://data.worldbank.org/indicator/AG.LND. AGRI.ZS?view=chart.

600


Tong, X., Xu, M., Wang, X., Bhattacharyya, R., Zhang, W., Cong, R., 2014. Long-term fertilization effects on organic carbon fractions in a red soil of China. Catena 113, 251–259. Van Oost, K., Quine, T.A., Govers, G., De Gryze, S., Six, J., Harden, J.W., et al., 2007. The impact of agricultural soil erosion on the global carbon cycle. Science 318, 626–629. Weber, K.A., Achenbach, L.A., Coates, J.D., 2006. Microorganisms pumping iron: anaerobic microbial iron oxidation and reduction. Nat. Rev. Microbiol. 4, 752–764. Wen, Y.L., Li, H., Xiao, J., Wang, C., Shen, Q.R., Ran, W., He, X.H., Zhou, Q.S., Yu, G.H., 2014a. Insights into complexation of dissolved organic matter and Al(III) and nanominerals formation in soils under contrasting fertilizations using two-dimensional correlation spectroscopy and high resolution-transmission electron microscopy techniques. Chemosphere 111, 441–449. Wen, Y.L., Xiao, J., Li, H., Shen, Q.R., Ran, W., Zhou, Q.S., Yu, G.H., He, X.H., 2014b. Longterm fertilization practices alter aluminum fractions and coordinate state in soil colloids. Soil Sci. Soc. Am. J. 78, 2083–2089.

Wen, Y.L., Xiao, J., Liu, F.F., Goodman, B.A., Li, W., Jia, Z.J., Ran, W., Zhang, R.F., Shen, Q.R., Yu, G.H., 2018. Constrasting effects of inorganic and organic fertilisation on shifts in Fe redox bacterial communities in red soils. Soil Biol. Biochem. 117, 56–67. Xiao, J., Wen, Y.L., Li, H., Hao, J.L., Shen, Q.R., Ran, W., Mei, X.L., He, X.H., Yu, G.H., 2015. In situ visualisation and characterisation of the capacity of highly reactive minerals to preserve soil organic matter (SOM) in colloids at submicron scale. Chemosphere 138, 225–232. Yu, G.H., Wu, M.J., Wei, G.R., Luo, Y.H., Ran, W., Wang, B.R., Zhang, J.C., Shen, Q.R., 2012. Binding of organic ligands with Al(III) in dissolved organic matter from soil: implications for soil organic carbon storage. Environ. Sci. Technol. 46, 6102–6109. Zhou, J.M., Dang, Z., Cai, M.F., Liu, C.Q., 2007. Soil heavy metal pollution around the Dabaoshan Mine, Guangdong Province, China. Pedosphere 17, 588–594.