binding of organic ligands with al(iii) in dissolved organic matter from soil: implications for soil...

Binding of Organic Ligands with Al(III) in Dissolved Organic Matterfrom Soil: Implications for Soil Organic Carbon StorageGuang-Hui Yu,† Min-Jie Wu,† Guan-Ran Wei,‡ Yi-Hong Luo,† Wei Ran,† Bo-Ren Wang,§

Jian−chao Zhang,† and Qi-Rong Shen*,†

†Jiangsu Key Lab for Organic Solid Waste Utilization, Key Laboratory of Plant Nutrition and Fertilization in Low-Middle Reaches ofthe Yangtze River, College of Resources and Environmental Sciences, Nanjing Agricultural University, Nanjing 210095, People'sRepublic of China‡School of Environmental Sciences and Technology, Shanghai Jiaotong University, Shanghai 200240, People's Republic of China§Key Laboratory of Plant Nutrition and Nutrient Cycling, Ministry of Agriculture of China and Institute of Agricultural Resources andRegional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, People's Republic of China

*S Supporting Information

ABSTRACT: The binding characteristics of organic ligandswith Al(III) in soil dissolved organic matter (DOM) isessential to understand soil organic carbon (SOC) storage. Inthis study, two-dimensional (2D) FTIR correlation spectros-copy was developed as a novel tool to explore the binding oforganic ligands with Al(III) in DOM present in soils as part ofa long-term (21-year) fertilization experiment. The resultsshowed that while it is a popular method for characterizing thebinding of organic ligands and metals, fluorescence excitation−emission matrix−parallel factor analysis can only characterizethe binding characteristics of fluorescent substances (i.e.,protein-, humic-, and fulvic-like substances) with Al(III).However, 2D FTIR correlation spectroscopy can characterizethe binding characteristics of both fluorescent and non-fluorescent (i.e., polysaccharides, lipids, and lignin) substanceswith Al(III). Meanwhile, 2D FTIR correlation spectroscopydemonstrated that the sequencing/ordering of organicsbinding with Al(III) could be modified by the use of long-term fertilization strategies. Furthermore, 2D FTIR correlationspectroscopy revealed that the high SOC content in thechemical plus manure (NPKM) treatment in the long termfertilization experiment can be attributed to the formation ofnoncrystalline microparticles (i.e., allophane and imogolite). In summary, 2D FTIR correlation spectroscopy is a promisingapproach for the characterization of metal−organic complexes.

■ INTRODUCTIONFluorescence excitation−emission matrix (EEM) spectroscopycombined with parallel factor (PARAFAC) analysis is a popularmethod for investigating the complexes between Al(III) andorganic ligands in DOM from soil,1−3 water,4,5 and leachates.6

Dissolved organic matter (DOM) in soil is a small but reactivefraction of soil organic matter.7 However, the stability anddynamics of DOM are essential to soil organic carbon storage.Some investigators have shown that noncrystalline micro-particles such as allophane [Al2O3(SiO2)1−2(H2O)2.5−4] andimogolite [(OH)3Al2O3SiOH] in soil DOM control soilorganic carbon sequestration.7,9 Therefore, it is essential toinvestigate the formation of complexes between soil DOM andAl(III). Although the fluorescence EEM-PARAFAC method

can characterize complexes of protein-, humic-, and fulvic-likesubstances with Al(III), one limitation of this method is that itcannot characterize complexes of nonfluorescent substances(i.e., polysaccharides, lipids, and lignin) with Al(III), eventhough such substances are the predominant components insoil.Fourier transform infrared (FTIR) spectroscopy is a

commonly used technique that can distinguish between theprincipal organics in soil DOM, including both fluorescent and

Received: January 18, 2012Revised: April 19, 2012Accepted: May 2, 2012Published: May 2, 2012

Article

pubs.acs.org/est

© 2012 American Chemical Society 6102 dx.doi.org/10.1021/es3002212 | Environ. Sci. Technol. 2012, 46, 6102−6109

pubs.acs.org/est

nonfluorescent substances.10,11 By determining the FTIRspectra of DOM after the titration of Al(III), it is possible toexplore the complexes of organic ligands in DOM and Al(III).However, the peaks in the conventional FTIR spectra oftenoverlap because of the extreme heterogeneity of DOM.12,13

Recent investigations have demonstrated that two-dimensional(2D) correlation spectroscopy can be applied to solve theoverlapped peaks problem by distributing the spectral intensitytrends within a data set, collected as a function of aperturbation sequence (e.g., time, temperature, or concen-tration change), over a second dimension.12−14 Moreimportantly, 2D correlation spectroscopy can be used toprobe the specific sequencing/ordering (sometimes referred toas Noda’s rule) of spectral intensity changes throughasynchronous analysis.14 To our knowledge, 2D FTIRcorrelation spectroscopy has not previously been applied toinvestigate the complexes of organic ligands in soil DOM andmetals.The objectives of the present study were (1) to provide more

information regarding the binding of organic ligands withAl(III) at the molecular level using 2D FTIR correlationspectroscopy combined with EEM-PARAFAC; (2) to give thesequencing of relative affinities of organic ligands for bindingwith Al(III) using 2D FTIR correlation spectroscopy; and (3)to explore the mechanisms of soil organic matter (SOM)sequestration by biogeochemical processes in the soil resultingfrom different fertilization treatments. For these purposes, theeffects of three treatments in a long-term (21-year) fertilizationexperiment were compared on the formation of complexesbetween organic ligands in soil DOM and Al(III): nofertilization (CK), chemical (NPK) fertilization, and chemicalplus pig manure (NPKM) fertilization.

■ MATERIALS AND METHODSSoil Sample and DOM Extraction. Soils were collected

from three treatments in a long-term fertilization experimentstation: CK-, NPK-, and NPKM-treatments. The long-termfertilization experiment was conducted initially in September1990 in fields that were double cropped with wheat and corn atthe experiment station of the Chinese Academy of AgriculturalSciences, Qiyang (26° 45′ N, 111° 52′ E, 120 m altitude),Hunan Province, southern China. The red soil was classified asFerralic Cambisol. A detailed description of the long-termfertilization experiment site can be found elsewhere.15 Soilsamples at depths of 0−20 cm were collected during September2011 using a 5-cm internal diameter auger. Each sample was acomposite of 10 random cores collected from a single plot. Thefresh soil was mixed thoroughly, air-dried, and sieved through2-mm and 0.25-mm screens for further analysis, respectively.The SOM contents of samples collected from plots receivingthe CK-, NPK-, and NPKM-treatments were 14.88 ± 2.02 g/kg(mean ± standard deviation), 18.36 ± 0.16 g/kg, and 25.13 ±2.02 g/kg, respectively (Table S1 of the SupportingInformation, SI).The DOM fraction of soil samples was extracted with

deionized water (solid to water ratio of 1: 2.5 w/v) by shakingfresh soil samples for 24 h on a horizontal shaker at roomtemperature. The DOM was filtered using 0.45 μmpolytetrafluoroethylene (PTFE) filters and further diluteduntil the dissolved organic carbon (DOC) concentration was<10 mg/L. Detailed descriptions of the procedures for thedetermination of pH, TOC, and SOM, as well as for XRF andXRD spectroscopy, are given in the SI. The XRD analysis

revealed that the clay mineralogy in the three fertilizationtreatments was similar with each other, including quartz (SiO2),muscovite [KAl2(Si3Al)O10(OH)2], and cronstedtite-2H1(Fe2+, Fe3+)3(Si, Fe

3+)2O5(OH)4 (Figure S1 of the SI).Al(III) Titration and Complexation Modeling. Aliquots

of 25 mL of the diluted solution of DOM were titrated in 40-mL brown sealed vials with 0.01 mol/L AlCl3 using anautomatic syringe. The Al(III) concentrations in the finalsolutions ranged from 0 to 100 mmol/L in 10 mmol/L steps.To maintain constant pH before and after titration, the metaltitrants were adjusted to pH 6.0, and no more than 25 μL of themetal titrant was added during titration. All titrated solutionswere shaken for 24 h at 25 °C to ensure complexationequilibrium.6 Then, all titrated solutions were analyzed byfluorescence EEM spectroscopy. After being freeze-dried, theywere also analyzed by FTIR spectroscopy.The modified Stern−Volmer equation was used to estimate

the conditional stability constant between Al(III) and thePARAFAC-derived components or the FTIR-derived bands:16

−=

· ·+

FF F f K C f

1 10

0 M M (1)

where F, F0, and f represent the measured fluorescence intensityscore or IR intensity, the initial fluorescence intensity score(i.e., no metal addition) or IR intensity, and the fraction of theinitial fluorescence intensity score or IR intensity, respectively.KM and CM are the conditional stability constant and the totalmetal ion concentration, respectively. The parameters f and KMwere solved by plotting F0/(F0 − F) against 1/CM.

Fluorescence EEM Determination and PARAFACAnalysis. Fluorescence determination is detailed else-where.17−19 Fluorescence EEMs were measured on a VarianEclipse fluorescence spectrophotometer in scan mode.Scanning emission (Em) spectra from 250 to 600 nm wereobtained in 2 nm increments by varying the excitation (Ex)wavelength from 200 to 500 nm in 10 nm increments. Thespectra were recorded at a scan rate of 1200 nm/min, usingexcitation and emission slit band widths of 5 nm. Each scanproduced an Excel data file composed of 171 Em (row) × 31Ex (column) wavelengths. PARAFAC analysis is described indetail elsewhere,17,18 and given in the SI.

Analysis of FTIR and 2D Correlation Spectroscopy.Samples were prepared as a mixture of 1 mg of freeze-driedDOM or of the DOM-Al(III) complex and 100 mg ofpotassium bromide (KBr, IR grade), and this mixture was thenground and homogenized.12,13 A subsample was then com-pressed between two clean, polished iron anvils twice in ahydraulic press at 20 000 psi to form a KBr window. The FTIRspectra were obtained by collecting 200 scans with a Nicolet370 FTIR spectrometer.The 2D correlation spectra were produced according to the

method of Noda and Ozaki.14 In this study, the Al(III)concentration was applied as an external perturbation, and a setof concentration-dependent FTIR spectra was obtained. Let usconsider an analytical spectrum I(x, t). The variable x is theindex variable representing the FTIR spectra induced by theperturbation variable t. We intentionally use x instead of thegeneral notation used in conventional 2D correlation equationsbased on the spectral index v. The analytical spectrum I(x, t) atm evenly spaced points in t (between Tmin and Tmax) can berepresented as follows:

= =I x I x t j m( ) ( , ), 1, 2, ...,j j (2)

Environmental Science & Technology Article

dx.doi.org/10.1021/es3002212 | Environ. Sci. Technol. 2012, 46, 6102−61096103

A set of dynamic spectra is given by the following:

= − I x t I x t I x( , ) ( , ) ( )j (3)

where I(x) denotes the reference spectrum, which is typicallythe average spectrum and is expressed as I(x) 1/m∑j = 1

m I(x,tj).The synchronous correlation intensity can be directly calculatedfrom the following dynamic spectra:

∑φ =−

=

x xm

I x I x( , )1

1( ) ( )

j

m

j j1 21

1 2(4)

Asynchronous correlation can be obtained by the following:

∑ ∑φ =−

= =

x xm

I x N I x( , )1

1( ) ( )

j

m

jk

m

jk j1 21

11

2(5)

The term Njk corresponds to the jth column and the kth rawelement of the discrete Hilbert-Noda transformation matrix,which is defined as follows:

π=

=

−

⎧⎨⎪

⎩⎪N

ifj k

k j

0

1( )

otherwisejk

(6)

The intensity of a synchronous correlation spectrum (ϕ(x1,x2)) represents simultaneous changes in two spectral intensitiesmeasured at x1 and x2 during the interval between Tmin andTmax. In contrast, an asynchronous correlation spectrum (φ(x1,x2)) includes out-of-phase or sequential changes in spectralintensities measuredat x1 and x2.Prior to 2D analysis, the FTIR spectra were normalized by

summing the absorbance from 4000 to 400 cm−1 andmultiplying by 1000. Subsequently, normalized FTIR spectrawere analyzed using principal component analysis (PCA) toreduce the level of noise.20 Finally, 2D correlation spectroscopywas conducted using 2Dshige software (Kwansei-GakuinUniversity, Japan). In this study, the FTIR region from 1800to 900 cm−1 was focused upon because it contains the majorexcitation bands of the amide, carboxylic acid, ester, andcarbohydrate functional groups.20

■ RESULTSTOC and Metal Concentrations in DOM. DOM in soil

from the NPKM treatment had the highest TOC content (14.2mg/L), followed by the NPK- (5.23 mg/L) and CK- (2.31 mg/L) treatments (Table S2 of the SI). The trend of DOMconcentration was similar to that of SOM in the differentfertilization treatments (Table S1 of the SI). Furthermore,fluorescence EEM spectra demonstrated that two, three, andfour peaks were present in samples from the NPKM-, NPK-,and CK- treatments, respectively (Figure 1). Peaks A, B, C, andD were located at Ex/Em of 230/410, 330/410, 270/340, and220/340, respectively. For the NPKM treatment, soil DOMcontained only humic- and fulvic-like substances (Peaks A andB); however, for the NPK treatment, the DOM was composedof both humic- and fulvic-like substances and a small quantityof protein-like substances (Peak C). In contrast, the soil DOMfrom the CK treatment predominantly contained protein-likesubstances (Peaks C and D) and some humic- and fulvic-likesubstances. In combination with FTIR spectroscopy, it wasnoted that a small quantity of protein (i.e., amide I at 1650cm−1) were also detected (Figure S3 of the SI). Therefore, it isconcluded that the high fluorescence intensity of the humic/

fulvic peaks is masking/hiding the protein peak presented inthe EEM contour of the NPKM DOM sample.It is notable that trivalent metals (Al and Fe) had a

significantly different distribution pattern from monovalent (Naand K) and divalent (Ca, Mg, Mn, and Zn) metals in thedifferent treatments (Table S2 of the SI). Specifically, trivalentmetals were mainly present in samples from the NPKMtreatment, and almost no trivalent metals were detected in theCK- and NPK-treatments; however, monovalent and divalentmetals were predominantly presented in the NPK treatmentwith concentrations higher than those observed for the othertwo treatments. The trend of trivalent metals concentrations isimportant for the sequestration of SOM, which will be

Figure 1. Fluorescence excitation−emission matrix (EEM) spectra ofdissolved organic matter (DOM) for soils from the CK-, NPK-, andNPKM-treatments, respectively, in the long term experiment.Specifically, peaks A, B, C, and D were located at Ex/Em of 230/410, 330/410, 270/340, and 220/340, respectively.



discussed in the following sections. Meanwhile, the backgroundmetals will also cause the quenching of EEM contours at acertain extent (Figure S4 of the SI).Fluorescent Components-Derived from EEM-PARAF-

AC Analysis of Al(III) Binding with DOM. All of thefluorescence EEMs collected from 33 different concentrationsof Al(III)-DOM complexes in the CK-, NPK-, and NPKM-treatments could be successfully decomposed by DOMFluor−PARAFAC analysis into a three−component model based onresidual analysis, split half analysis, and excitation and emissionloadings (Figure S1 of the SI). The excitation and emissionloadings of the three PARAFAC-derived components and thefluorescence intensity scores of the titration of the CK- andNPK- treatment-derived DOM components with Al(III) areplotted in Figure 2. Component 1 [Ex/Em = (240, 330)/410]belonged to humic- and fulvic-like substances, component 2[Ex/Em = (250, 340)/490] was mainly ascribed to humic-likesubstances, and component 3 [Ex/Em = (220, 270)/340] wasattributable to protein-like substances, according to theprotocol of Chen et al.21 The surface plots of the threecomponents are presented in Figure S2 of the SI. Thesecomponents had been previously identified in soil environ-ments.1−3

As for the CK treatment, component 3 had the highestfluorescence intensity scores, followed by components 1 and 2.However, in the case of the NPK treatment, component 1 hadthe highest fluorescence intensity scores, followed bycomponents 2 and 3. In contrast, the NPKM treatment onlycontained components 1 and 2 and lacked component 3. Theabove-mentioned results suggested that for the three treat-ments, humic- and fulvic-like substances were detected in the

soil DOM; however, a large amount of protein-like substanceswas present in the CK treatment, and no protein-likesubstances were found in the NPKM treatment.In the case of all three components, Al(III) titration led to

quenching, enhancement, and no change in the fluorescencesignals of samples from the CK-, NPK-, and NPKM-treatments,respectively, reflecting that the structures of organics in theCK-, NPK-, and NPKM-treatments were different (Figure 2).To measure the binding constants between organics andAl(III), the conditional stability constant (log KM) wascalculated using the modified Stern−Volmer equation. Thecalculated log KM was 4.19 (R2 = 0.80) and 1.78 (R2 = 0.64) forcomponents 1 and 2 in the CK treatment, respectively, and3.52 (R2 = 0.75) for component 1 in the NPK treatment (TableS3 of the SI). Therefore, for the CK treatment, Al(III) had astronger binding capability with humic-like substancescompared with fulvic-like substances. However, in the NPKtreatment, Al(III) had a weak binding capability with humic-likesubstances when compared to the CK treatment.

FTIR Synchronous Maps. It can be observed from FigureS3 of the SI that the peaks in one-dimensional FTIR spectrastrongly overlap, and the spectrum of the CK treatment wassimilar with that of the NPK treatment but distinct from that ofthe NPKM treatment. It can be noted that the trends of theFTIR spectra for the three treatments are different from thoseof the EEM measurements, a difference that may be attributableto the contributions of polysaccharides, lipids, or lignindetected in FTIR spectra (Figure S4 of the SI), which hadno fluorescent and thus could not be detected by the EEMmeasurements. Following the titration of a series of organic

Figure 2. Excitation and emission loadings of three PARAFAC-derived components, and fluorescence intensity scores of the titration of the CK-,NPK-, and NPKM-treatment-derived dissolved organic matter components with Al(III).



fractions with Al(III), the intensities of only a few peaks werechangeable, and no further analysis was possible.The synchronous maps generated from the 11 DOM-Al(III)

FTIR spectra (Figure 3) displayed three major autopeaks for

CK- and NPK-treatments but only one autopeak for NPKMtreatment. Note that autopeaks appear at diagonal position andrepresent the overall susceptibility of the corresponding spectralregion to change in spectral intensity as an externalperturbation is applied to the system. The change in bandintensity followed the order 1120 > 1520 > 1690 cm−1 for theCK treatment and 1120 > 1520 > 1380 cm−1 for the NPKtreatment. However, only one autopeak located at 1010 cm−1

varied for the NPKM treatment. The clear peak positions of1690 cm−1 for the CK treatment and 1380 cm−1 for the NPKtreatment could be seen in Figure S6 of the SI. We assignedthese bands as follows: the band at 1690 cm−1 was assigned tothe CO stretching of amide I in protein compounds, theband at 1520 cm−1 to N−H deformation and CN stretchingof amide II in protein compounds, the band at 1380 cm−1 tothe CH deformations in lignin and aliphatic groups, the band at1120 cm−1 to the C−OH stretching of aliphatic O−H, and theband at 1010 cm−1 to the C−O stretching of polysaccharides,the Si−O of silicate impurities, or phosphate groups.10,11,13,22

Off-diagonal peaks (cross-peaks) in the synchronous mapexhibited correlated signals. The cross-peaks showed a positivecorrelation between the bands at 1690 and 1520 cm−1 andanother positive correlation between the bands at 1380 and1120 cm−1, suggesting that amide I and amide II in proteincompounds were cobinding with Al(III), whereas aliphaticgroups (CH and OH) covaried with Al(III). The cross-peaksalso indicated that the band at 1120 cm−1 was negativelycorrelated with the bands at 1690 and 1520 cm−1. Thiscorrelation reveals that with the addition of Al(III), signals ofproteins did not simultaneously vary with those of lipids (i.e.,aliphatic groups).

FTIR Asynchronous Maps. The asynchronous map canprovide useful information on the sequential order of eventsobserved using the spectroscopic technique along the externalvariable.12,13 The asynchronous map is antisymmetric withrespect to the diagonal line and showed no autopeaks.Specifically, three, three, and eight cross-peaks were observedabove the diagonal line in the CK-, NPKM-, and NPK-treatments, respectively. According to Noda’s rule,14 thesequence of the binding affinities of bands with Al(III) followsthe order 1480 → 1120 → 1650 cm−1 in the CK treatment,1540 → 1120 → 1380 cm−1 (1650 cm−1) in the NPKtreatment, and 920 → 1010 → 1110 cm−1 (1380 cm−1) in theNPKM treatment. The asynchronous maps demonstrated thatthe binding of organic ligands in soil DOM with Al(III) followsthe sequence: amide II in proteins > aliphatic O−H > amide Iin proteins in the CK treatment, amide II in proteins > aliphaticO−H > lignin and aliphatic C−H or (amide I in proteins) inthe NPK treatment, and aromatic C−H in cellulose > the C−Ostretching of polysaccharides > aliphatic O−H (lignin andaliphatic C−H) in the NPKM treatment.Moreover, the asynchronous maps showed that the band at

1690 cm−1 was overlapped by bands at 1700 and 1640 cm−1;the band at 1520 cm−1 was overlapped by bands at 1540 and1480 cm−1; the band at 1120 cm−1 was overlapped by bands at1180, 1160, and 1120 cm−1; and the band at 1020 cm−1 wasoverlapped by bands at 1110, 1010, and 920 cm−1.The calculated log KM values based on the modified Stern−

Volmer equation are listed in Table 1. For the CK treatment,

Figure 3. Synchronous and asynchronous2D correlation mapsgenerated from the 1800−900 cm−1 region of the FTIR spectra ofdissolved organic matter in the CK-, NPK-, and NPKM-treatmentsover Al(III). Red represents positive correlation, and blue representsnegative correlation; a higher color intensity indicates a strongerpositive or negative correlation. Table 1. Conditional Stability Constants (log KM) for IR-

Derived Bands Binding to Al(III) as Calculated Using theModified Stern−Volmer Equation

treatmenta IR absorption bands (cm−1) log KM R2

CK 1120 not modeled1540 9.37 0.841690 9.51 0.78

NPK 1120 not modeled1160 9.09 0.591180 not modeled1380 9.30 0.661480 not modeled1520 9.46 0.631540 9.47 0.63

NPKM 920 8.69 0.661010 8.38 0.921020 8.36 0.911110 8.57 0.94

aNote: CK, no fertilization; NPK, chemical fertilization; NPKM,chemical plus manure fertilization.



the calculated log KM values were 9.37 (R2 = 0.84) and 9.51 (R2

= 0.78) for amide I (1690 cm−1) and amide II (1540 cm−1) inproteins, respectively. For the NPK treatment, log KM valueswere approximately 9.46 (R2 = 0.63) and 9.09−9.30 (R2 > 0.59)for amide II (1540 and 1520 cm−1) and lipids (1380 and 1160cm−1), respectively. However, for the NPKM treatment, log KMvalues were in the range of 8.36−8.69 for polysaccharides orsilicates (1110, 1020, 1010, and 920 cm−1). Therefore, onlyproteins had a strong binding capability with Al(III) in the CKtreatment, but both proteins and lipids demonstrated a strongbinding capability with Al(III) in the NPK treatment. However,neither polysaccharides nor silicates had a significant bindingcapability with Al(III) in the CK treatment.

■ DISCUSSIONThe EEM contours of DOM from the different fertilizationstrategies had been shown distinct and thus could be used toassess the soil management effects.23 Ohno et al.23 and Zhanget al.24 investigated the effect of fertilizer treatments (10 years)on EEM contours of DOM, and found that proteins in DOMfrom no fertilization and chemical fertilization were higher thanthose from organic fertilization. Meanwhile, with the increase offertilization time, humic- and fulvic-like materials increasedwhile proteins decreased. Their observations were similar to theresults presented in this study. The DOM from the CK- andNPK-treatments, rather than NPKM treatment, containedproteins that may be attributable to the priming effect ofmanure which rapidly depleted easily available proteins andtransferred them to more stable humic and fulvic-likematerials.25 However, the robust mechanism needs to beelucidated in future investigations.The fluorescence EEM-PARAFAC method could provide

information about Al(III) binding characteristics with fluo-rescent substances such as protein-, humic-, and fulvic-likesubstances. However, this popular approach has been criticizedas being unable to provide information about the bindingcharacteristics of nonfluorescent substances (e.g., polysacchar-ides, lipids, and lignin) with Al(III). In soil environments, bothfluorescent- and nonfluorescent-substances are abundant.10,11

Our results, based on 2D FTIR correlation spectroscopy,clearly demonstrate that this approach can provide informationabout Al(III) binding with both fluorescent (i.e., 1690 and 1540cm−1) and nonfluorescent (i.e., 1380, 1120, and 1010 cm−1)substances. The synchronous maps reveal that the bindingstrength of compounds with Al(III) follows this sequence:lipids (aliphatic O−H) > amide II > amide I in proteins for theCK treatment, lipids (aliphatic O−H) > amide II in proteins >lignin and aliphatic groups for the NPK treatment, and onlypolysaccharides and silicates bind with Al(III) in the NPKMtreatment. Therefore, nonfluorescent substances (i.e., lipids,polysaccharides, and lignin) played an important role in thebinding of organic ligands with Al(III), which could not havebeen detected by the EEM-PARAFAC method. The asynchro-nous maps further demonstrated that the sequencing of Al(III)binding with organic ligands could be modified by the use ofdifferent long-term fertilization strategies. Specifically, proteinsfollowed by lipids were bound with Al(III) in the CK- andNPK-treatments, whereas homopolysaccharides (i.e., cellulose)followed by heteropolysaccharides and lignin were bound withAl(III) in the NPKM treatment.In summary, the 2D FTIR correlation spectroscopy can be

complementary to EEM-PARAFAC method in improving thebinding of organics in DOM with metals characterization.

Therefore, combination of 2D FTIR correlation spectroscopywith EEM-PARAFAC method will help to construct a morecomprehensive picture of the binding of organic ligands inDOM with metals characterization.

Implications for Soil Organic Carbon Storage.Application of organic fertilizer to soil has been shown as auseful way to increase SOM and reduce environmentalpollution.15 However, the mechanism of sequestration of soilorganic carbon is not completely understood. In this study, wefound that a significant amount of OH bonds (3420 cm−1 and1120 cm−1) from polysaccharides were detected in soil samplesfrom all three treatments, with the highest OH content presentin the NPKM treatment (Figures 3 and S2 of the SI).Meanwhile, the results of 2D FTIR correlation spectroscopyfurther showed that the intrahydrogen bond (>3400 cm−1) insoil polysaccharides played an important role in binding withAl(III) in all the three treatments (Figure S3 of the SI).However, the Si−O stretching of silicate (1010 cm−1) was onlydetected in the NPKM treatment and not in samples from theother two treatments. The TEM observation (Figure S8 of theSI) clearly revealed that noncrystalline minerals were formingin soils applied with organic manures. The high-resolutiontransmission electron micrograph (HR-TEM) image and thecorresponding energy dispersive X-ray (EDX) spectrum(Figure S9 of the SI) further support that the main elementsof noncrystalline minerals were Al, Si, and O. Therefore, it isreasonable to surmise that the noncrystalline minerals such asallophane [Al2O3(SiO2)1−2(H2O)2.5−4] or imogolite[(OH)3Al2O3SiOH] were formed in soil DOM from theNPKM treatment. Due to the lack of Si−O stretching fromsilicate in soil DOM, noncrystalline minerals could not beproduced in soil DOM from the CK- and NPK-treatments. Theprevious investigation explored the structure of allophane andimogolite by FTIR spectroscopy, showing the presence of Al−OH and Si−O−Al bonds in the feature peaks (3500, 3440,1000, 940, and 917 cm−1).26,27 Smalley28 indicated that the Si−O−(Al) absorption bands in the 1100−900 cm−1 region of theFTIR spectra. However, it should be mentioned that the bandat 1010 cm−1 is not certainly attributable to the Si−O of silicateimpurities, since it may also be assigned to the C−O stretchingof polysaccharides or phosphate groups.10,11,13 Therefore, itneeds to further confirm the assignment of the band at 1010cm−1 in the future research using 29Si NMR spectroscopy.The proposed mechanism of sequestration soil organic

matter was given in Figure S10 of the SI. In soil aggregates,DOM was the most reactive fraction.7 DOM would be utilizedby crops and microorganisms and released greenhouse gas (e.g.,CO2, CH4, NH3, N2O). Noncrystalline minerals were formed insoil DOM from the NPKM treatment. The formation ofnoncrystalline minerals (i.e., allophane and imogolite) couldsignificantly decrease the reactivity and bioavailability of organicmatter.8 Torn et al.8 demonstrated that noncrystalline mineralssuch as allophane and imogolite controlled soil organic carbonstorage and turnover. Furthermore, Parfitt29 indicated thatunder favorable conditions, the turnover of SOM in allophonicsoils may persist in tephra beds for at least 250 000 years.However, allophane and imogolite have been shown to bestrong absorbents for CO2, CH4, NH3, and N2O.

30,31 Globally,approximately 190−332 million tons of C can be sequestratedby silicate minerals each year.32

Therefore, it can be concluded that the mechanism ofsequestration of SOM in the NPKM treatment is attributable tothe formation of noncrystalline minerals such as allophane and



imogolite in soil DOM, which could decrease the reactivity andbioavailability of organic matter and increase the storage ofCO2, CH4, NH3, and N2O. Knowledge of the binding oforganic ligands and metals contributes to our understanding ofthe sequestration process of SOM and provides novelinformation for fertilization techniques and scientific research.

■ ASSOCIATED CONTENT*S Supporting InformationDetailed descriptions of the chemical characterization assay andPARAFAC analysis; one table listing physiochemical character-istics of bulk soil in the CK-, NPK-, and NPKM-treatments;one table listing TOC and metals concentration in DOM of theCK-, NPK-, and NPKM-treatments; one table presentingconditional stability constants (log KM) for EEM-PARAFAC-derived components binding to Al(III) as calculated using themodified Stern−Volmer equation; one figure showing fluo-rescence EEM contours of composts; one figure showingsurface plots of three PARAFAC-derived components; onefigure presenting synchronous and asynchronous 2D correla-tion maps generated from the 3600−3200 cm−1 region; andtwo figures showing TEM images of aluminosilicate minerals inthe DOM leaching from soil of the NPKM treatment. Thismaterial is available free of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*Phone: +86-25-8439 5212; fax: +86-21-8439 5212; e-mail:[email protected].

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThe work was funded by the National Basic Research Programof China (2011CB100503), the National Natural ScienceFoundation of China (21007027), the Specialized ResearchFund for the Doctoral Program of Higher Education(20100097120015), the China Postdoctoral Science Founda-tion (20100481156, 201104535, and 1002017B), the Funda-mental Research Funds for the Central Universities(KYZ201143), and the Agricultural Ministry of China (2011-G27 and 201103004).

■ REFERENCES(1) Plaza, C.; Brunetti, G.; Senesi, N.; Polo, A. Molecular andquantitative analysis of metal ion binding to humic acids from sewagesludge and sludge-amended soils by fluorescence spectroscopy.Environ. Sci. Technol. 2006, 40, 917−923.(2) Ohno, T.; Fernandez, I. J.; Hiradate, S.; Sherman, J. F. Effects ofsoil acidification and forest type on water soluble soil organic matterproperties. Geoderma 2007, 140, 176−187.(3) Ohno, T.; Amirbahman, A.; Bro, R. Parallel factor analysis ofexcitation−emission matrix fluorescence spectra of water soluble soilorganic matter as basis for the determination of conditional metalbinding parameters. Environ. Sci. Technol. 2008, 42, 186−192.(4) Yamashita, Y.; Jaffe, R. Characterizing the interactions betweentrace metals and dissolved organic matter using excitation-emissionmatrix and parallel factor analysis. Environ. Sci. Technol. 2008, 42,7374−7379.(5) Aiken, G. R.; Hsu-Kim, H.; Ryan, J. N. Influence of dissolvedorganic matter on the environmental fate of metals, nanoparticles, andcolloids. Environ. Sci. Technol. 2011, 45, 3196−3201.

(6) Wu, J.; Zhang, H.; He, P. J.; Shao, L. M. Insight into the heavymetal binding potential of dissolved organic matter in MSW leachateusing EEM quenching combined with PARAFAC analysis. Water Res.2011, 45, 1711−1719.(7) Troyer, I. D.; Amery, F.; Moorleghem, C. V.; Smolders, E.;Merckx, R. Tracing the source and fate of dissolved organic matter insoil after incorporation of 13C labeled residue: A batch incubationstudy. Soil Biol. Biochem. 2011, 43, 513−519.(8) Torn, M. S.; Trumbore, S. E.; Chadwich, O. A.; Vitousek, P. M.;Hendricks, D. M. Mineral control of soil organic carbon storage andturnover. Nature 1997, 389, 170−173.(9) Mikutta, R.; Kaiser, K. Organic matter bound to mineral surfaces:Resistance to chemical and biological oxidation. Soil Biol. Biochem.2011, 43, 1738−1741.(10) Artz, R. R. E.; Chapman, S. J.; Robertson, A. H. J.; Potts, J. M.;Laggoun-Defarge, F.; Gogo, S.; Comont, L.; Disnar, J. R.; Francez, A. J.FTIR spectroscopy can be used as a screening tool for organic matterquality in regenerating cutover peatlands. Soil Biol. Biochem. 2008, 40,5151−527.(11) D’Orazio, V.; Senesi, N. Spectroscopic properties of humic acidsisolated from the rhizosphere and bulk soil compartments andfractionated by size-exclusion chromatography. Soil Biol. Biochem.2009, 41, 1775−1781.(12) Yu, G. H.; Tang, Z.; Xu, Y. C.; Shen, Q. R. Multiple fluorescencelabeling and two dimensional FTIR−13C NMR heterospectralcorrelation spectroscopy to characterize extracellular polymericsubstances in biofilms produced during composting. Environ. Sci.Technol. 2011, 45, 9224−9231.(13) Wang, L. P.; Shen, Q. R.; Yu, G. H.; Ran, W.; Xu, Y. C. Fate ofbiopolymers during rapeseed meal and wheat bran composting asstudied by two-dimensional correlation spectroscopy in combinationwith multiple fluorescence labeling techniques. Bioresour. Technol.2012, 105, 88−94.(14) Noda, I., Ozaki, Y., Eds. Two-Dimensional Correlation Spectros-copy- Applications in Vibrational and Optical Spectroscopy; John Wiley &Sons: London, 2004.(15) Zhang, H. M.; Wang, B. R.; Xu, M. G.; Fan, T. L. Crop yield andsoil responses to long-term fertilization on a red soil in southernChina. Pedosphere 2009, 19, 199−207.(16) Hur, J.; Lee, B. M. Characterization of binding site heterogeneityfor copper within dissolved organic matter fractions using two-dimensional correlation fluorescence spectroscopy. Chemosphere 2011,83, 1603−1611.(17) Yu, G. H.; He, P. J.; Shao, L. M. Novel insights into sludgedewaterability by fluorescence excitation−emission matrix combinedwith parallel factor analysis. Water Res. 2010, 44, 797−806.(18) Yu, G. H.; Luo, Y. H.; Wu, M. J.; Tang, Z.; Liu, D. Y.; Yang, X.M.; Shen, Q. R. PARAFAC modeling of fluorescence excitation−emission spectra for rapid assessment of compost maturity. Bioresour.Technol. 2010, 101, 8244−8251.(19) Yu, G. H.; Wu, M. J.; Luo, Y. H.; Yang, X. M.; Ran, W.; Shen, Q.R. Fluorescence excitation-emission spectroscopy with regionalintegration analysis for assessment of compost maturity. WasteManage. 2011, 31, 1729−1736.(20) Abdulla, H. A.; Minor, E. C.; Dias, R. F.; Hatcher, P. G. Changesin the compound classes of dissolved organic matter along an estuarinetransect: A study using FTIR and 13C NMR. Geochim. Cosmochim.Acta 2010, 74, 3815−3838.(21) Chen, W.; Westerhoff, P.; Leenheer, J. A.; Booksh, K.Fluorescence excitation−emission matrix regional integration toquantify spectra for dissolved organic matter. Environ. Sci. Technol.2003, 37, 5701−5710.(22) He, Z. Q.; Ohno, T.; Cade-Menun, B. J.; Erich, M. S.;Honeycutt, C. W. Spectral and chemical characterization ofphosphates associated with humic substances. Soil Sci. Soc. Am. J.2006, 70, 1741−1751.(23) Ohno, T.; He, Z. Q.; Tazisong, I. A.; Senwo, Z. N. Influence oftillage, cropping, and nitrogen source on the chemical characteristics of



http://pubs.acs.org

http://pubs.acs.org

mailto:[email protected]

humic acid, fulvic acid, and water-soluble soil organic matter fractionsof a long-term cropping system study. Soil Sci. 2009, 174, 652−660.(24) Zhang, M. C.; He, Z. Q.; Zhao, A. Q.; Zhang, H. L.; Endale, D.M.; Schomberg, H. H. Water-extractable soil organic carbon andnitrogen affected by tillage and manure application. Soil Sci. 2011, 176,307−312.(25) Kuzyakov, Y. Priming effects: Interactions between living anddead organic matter. Soil Biol. Biochem. 2010, 42, 1363−1371.(26) Inoue, K.; Huang, P. M. Influence of citric acid on the naturalformation of imogolite. Nature 1984, 308, 58−60.(27) Lou, G.; Huang, P. M. Hydroxy-aluminosilicate interlayers inmontmorillonite: Implications for acidic environments. Nature 1988,335, 625−627.(28) Smalley, I. A spherical structure for allophone. Nature 1979,281, 339.(29) Parfitt, R. L. Allophane and imogolite: Role in soilbiogeochemical processes. Clay Miner. 2009, 44, 135−155.(30) Theng, B. K. G. Adsorption of ammonium and some primary n-alkylammonium cations by soil allophane. Nature 1972, 238, 150−151.(31) Guimaraes, L.; Enyashin, A. N.; Frenzel, J.; Heine, T.; Duarte,H. A.; Seifert, G. Imogolite nanotubes: Stability, electronic, andmechanical properties. ACS Nano 2007, 1, 362−368.(32) Renforth, P.; Washbourne, C. L.; Taylder, J.; Manning, D. A. C.Silicate production and availability for mineral carbonation. Environ.Sci. Technol. 2011, 45, 2035−2041.



binding of organic ligands with al(iii) in dissolved organic matter from soil: implications for soil...

Documents