A review of cadmium sorption mechanisms on soil mineral surfaces revealed from synchrotron-based X-ray absorption fine structure spectroscopy:Implications for soil remediation

2021-03-17 09:05:46XinxinMOMatthewSIEBECKERWenxianGOULingLIandWeiLI
Pedosphere 2021年1期

Xinxin MO,Matthew G.SIEBECKER,Wenxian GOU,Ling LI and Wei LI

1Key Laboratory of Surficial Geochemistry,Ministry of Education,School of Earth Sciences and Engineering,Nanjing University,Nanjing 210023(China)

2Department of Plant and Soil Science,Texas Tech University,Lubbock TX 79409(USA)

3State Key Laboratory of Environmental Geochemistry,Institute of Geochemistry,Chinese Academy of Sciences,Guiyang 550081(China)

(Received February 17,2020;revised March 24,2020)

ABSTRACT The sorption of cadmium(Cd)is one of the most important chemical processes in soil,affecting its fate and mobility in both soil and water and ultimately controlling its bioavailability.In order to fundamentally understand the sorption/desorption of Cd in soil systems,X-ray absorption fine structure spectroscopy(XAFS)has been applied in numerous studies to provide molecular-level information that can be used to characterize the surface adsorption and precipitation reactions that Cd can undergo.This information greatly improves our current knowledge of the possible chemical reactions of Cd in soil.This paper critically reviews the mechanisms of Cd sorption/desorption at the mineral-water interface based on XAFS studies performed over the past twenty years.An introduction to the basic concepts of sorption processes is provided,followed by a detailed interpretation of XAFS theory and experimental data collection and processing,ending finally with a discussion of the atomic/molecular-scale Cd sorption mechanisms that occur at the soil mineral-water interface.Particular emphasis is placed on literature that discusses Cd adsorption and speciation when associated with iron,manganese,and aluminum oxides and aluminosilicate minerals.Multiple sorption mechanisms by which Cd is sorbed by these minerals have been found,spanning from outer-sphere to inner-sphere to surface precipitation,depending on mineral type,surface loading,and pH.In addition,the application of complementary techniques(e.g.,113Cd nuclear magnetic resonance(NMR)and molecular dynamics simulation)for probing Cd sorption mechanisms is discussed.This review can help to develop appropriate strategies for the environmental remediation of Cd-contaminated soils.

Key Words:aluminosilicate minerals,aluminum oxide,bioavailability,heavy metal,iron oxide,manganese oxide,mineral-water interface,molecular-scale sorption mechanisms

INTRODUCTION

Cadmium(Cd)is the second element in the IIB group and has been recognized as a toxic,carcinogenic,and teratogenic heavy metal(ATSDR,2012).This element is classed as a human carcinogen(group 1)by the International Agency for Research on Cancer of the World Health Organization(WHO).The predominant pathway of human exposure to Cd is through diet,especially rice consumption(Järup andÅkesson,2009;Sataruget al.,2010;Chaney,2012).The world’s first documented incidence of mass poisoning through Cd-polluted rice occurred in Japan in the 1950s,resulting in approximately 300 people suffering from osteomalacia and marked decalcification associated with Itai-Itai disease(Nogawa and Kido,1993).China,which is a major producer and consumer of rice,suffers from significant amounts of rice Cd contamination.The background level of Cd in the soils of China ranges from approximately 0.01 to 1.80 mg kg−1with an average of 0.163 mg kg−1,which is slightly less than the global concentration of approximately 0.01-2.0 mg kg−1with an average of 0.35 mg kg−1.The soil quality standard,i.e.,the maximum allowable concentration of Cd in soil,is specified at approximately 0.3-0.6 mg kg−1in China,whereas standards worldwide vary between 0.8 and 5.0 mg kg−1(Zhaoet al.,2014).If the Cd content in soil exceeds the standard value,control and treatment could be required.According to Ministry of Environmental Protection and Ministry of Land and Resources of China(2014),Cd is at the top of the list of dangerous inorganic contaminates and exceeds the safety standard by approximately 7.0%.Additionally,several regions have suffered from excessive levels and incidents of pollution with Cd around the world,for example,Hunan Province in China.

Anthropogenic activities,e.g.,mining and smelting,sewage irrigation,and the application of fertilizer,are the major cause of Cd input into agricultural soils(Luoet al.,2009;Hu and Cheng,2013;Wenet al.,2015).The natural weathering of Cd-rich sedimentary rock can also lead to elevated Cd levels in soil.The naturally occurring Cd in arable soils that are influenced by the parent rocks in the black rock series of the Three Gorges region of southwestern China can reach 42 mg kg−1(Liuet al.,2015),which is around 80 times the amount generally observed in surface soil(0.53 mg kg−1)(Kabata-Pendias and Pendias,2001).

The negative impact of Cd in soil is determined by its bioavailability,which characterizes the proportion of the total Cd that can be taken up by biota.Several empirical methods can be used to evaluate soil Cd bioavailability,includingin-situdiffusive gradients in thin film(DGT)andex-situsingle-step extraction methods(Smildeet al.,1992;Krishnamurtiet al.,1995;Wanget al.,2017).However,in principle,Cd bioavailability is controlled by the chemical reactions(Krishnamurtiet al.,1995;Chenet al.,2000)that take place at the soil mineral-water interface,including adsorption/desorption,precipitation/dissolution,and the formation of Cd-ligand complexes(Bolanet al.,2013).

Because of the importance of soil mineral-water interface reactions,a mechanistic understanding of the sorption reactions of Cd is fundamentally important for risk assessment and soil remediation.X-ray absorption fine structure spectroscopy(XAFS)is an element-specific technique that yields molecular-and atomic-level structural information concerning both interatomic distance and the number and type of atomic coordination(Kellyet al.,2008).In the past thirty years,XAFS has been widely applied to elucidate metal sorption mechanisms at the molecular level(Hayeset al.,1987;Chisholm-Brauseet al.,1990;Scheideggeret al.,1996;Gräfeet al.,2007;Liet al.,2011).The mechanisms for outer-sphere complexation,inner-sphere surface complexation,and surface precipitation can be distinguished based on these structural parameters.In this review,we provide an introduction to the basic concepts of the sorption of metals by soil minerals,followed by a detailed description of the XAFS data collection and the methods used to process this data and a thorough review of XAFS studies that have investigated the sorption mechanisms of Cd at the soil-mineral interface.Furthermore,future perspectives on the application of new methods such as113Cd nuclear magnetic resonance(NMR)(Di Leo and O’Brien,1999)for probing Cd interface chemistry are discussed,and molecular dynamics(MD)simulation(Zhanget al.,2016,2018).

GENERAL CONCEPTS OF METAL SORPTION REACTIONS IN SOIL CHEMICAL PROCESSES

The bioavailability of contaminants in soil and sediment is affected by a series of individual physical,chemical,and biological interactions that determine the exposure of plants and animals to chemicals(National Research Council,2003),as shown schematically in Fig.1.Processes affecting bioavailability include:i)contaminant interaction between phases,as a result of which contaminants can reside in a bound state(associated with soil or sediment particles)or a released state(dissolved in a liquid or gas phase);ii)the transport of contaminants(in both bound and released forms)to organisms;and iii)the passage across physiological membranes.Sorption reactions at the mineral-water interface are of great significance in controlling the speciation,mobility,fate,and bioavailability of contaminants in soil environments.

Sorption can be defined as the phenomenon in which aqueous ions,complexes,or molecules accumulate at the interface between a solid substance and an aqueous solution.In general,there are four main mechanisms by which sorption can occur(Fig.1):i)outer-sphere and ii)inner-sphere surface complexation,where soluble species aggregate on the surface of minerals without forming a three-dimensional molecular arrangement in a process that is collectively referred to as adsorption,in which inner-sphere complexes are generally more strongly bound to surfaces than outer-sphere complexes;iii)surface precipitation or polymerization,where products with a structure or composition that is different from that of the host mineral are formedviaprocesses such as“epitaxial overgrowth”in petrology,or small multinuclear inorganic complexes such as dimers or trimers are formed;and iv)coprecipitation or substitution,where solute species are integrated into the mineral structure through the dissolution and re-precipitation of the mineral,orviadiffusion and isomorphic substitution within the mineral lattice(Krauskopf and Bird,1995).The general concepts of sorption,adsorption,and absorption are easily confused and need to be distinguished in order to identify the potential risk of releasing Cd back into the pore water in soil.

Adsorption is a two-dimensional process in which solute molecules or atoms(i.e.,adsorptives)leave the solution to become attached to the surface of a mineral(i.e.,adsorbent).The molecules and atoms on the mineral surface(i.e.,adsorbates)are boundviasurface complexes,which can be divided into two types on the basis of the atomic arrangements and the bonding properties between adsorbate and adsorbent.One type of the surface complex is the outer-sphere complex,where one or more water molecules are present between the mineral surface and the adsorbed species.Outer-sphere complexes are affected by weak electrostatic attractions.The other type of the surface complex is the inner-sphere complex,where strong,short-range ionic or covalent bonds form at a characteristic crystal site without the occurrence of any water molecules between the adsorbed species and the mineral surface(Stumm and Morgan,1981;Sparks,1995).Inner-sphere complexes can be monodentate,bidentate,or tridentate depending on the site on the surface of the adsorbent(Fig.1).This method of classifying bonding is based on a polyhedral approach to surface complexes.

Fig.1 Schematic diagram of metal bioavailability processes for contaminants in soil(from National Research Council(2003),reprinted with permission from Bioavailability of Contaminants in Soils and Sediments,copyright(2003)by the National Academies of Sciences,courtesy of the National Academies Press,Washington,D.C.).A-contaminant interaction between phases;B-transport of bound contaminants to organisms;C-transport of released contaminants to organisms;D-passage across a physiological membrane;E-circulation within an organism,accumulation in a target organ,toxicokinetics,and the toxic effects.Possible sorption mechanisms at the mineral-water interface include outer-sphere surface complexes(I),inner-sphere surface complexes(II),surface precipitation(III),and substitution(or incorporation at the surface)(IV).

Recent evidence for a continuous transition between surface complexation(adsorption)and surface precipitation indicates that as the amounts of metal cations or anions adsorbed onto a surface(surface coverage or surface loading)increase to cover a larger surface area,precipitates can form on the surface(Siebeckeret al.,2014)(Fig.1).The traditional way to view surface precipitation is that at low surface loading,surface complexation tends to dominate.As surface loading increases,surface polymers or multinuclear surface complexes may form,accompanied by further nucleation on the surface.Surface precipitation becomes the predominant mechanism as the surface coverage further increases.However,surface precipitation has also been demonstrated at a surface coverage of less than 5%(Scheideggeret al.,1996).

Sorption is one of the most significant chemical processes that take place in soil systems,determining the quantity of plant nutrients,metals,radionuclides,pesticides,and other organic chemicals that are retained on the surface of the soil.Hence,it is one of the key processes that affect the migration of nutrients and contaminants within soil.Sorption also affects the electrostatic properties of suspended particles and colloids,which influence coagulation and settlement(Stumm,1992;Sparks,1995).

In certain cases,surface incorporation(the substitution of atoms on the surface of a mineral)can occur on a dynamic surface.For example,when Cd2+is sorbed onto birnessite in the presence of dissolved Mn2+,an interface coupled dissolution-precipitation(ICDP)process takes place,leading to a rapid atomic exchange of the aqueous Mn2+with the manganese(Mn)atoms on the surface of the birnessite,and this exchange induces the incorporation of Cd2+into the mineralviaits sorption at the birnessite-solution interface(Sunet al.,2019a).These mechanisms only exhibit structural differences on a molecular scale,meaning that a spectroscopic technique is required in order to distinguish them.One of the most widely applied spectroscopic techniques is XAFS.

BRIEF INTRODUCTION TO XAFS

As an advanced structural analysis technique,XAFS is based on a synchrotron radiation X-ray source.The general theoretical concepts,facilities,experimental methods,and data analysis of XAFS are briefly introduced in this section.Other resources describing the use of synchrotron X-rays in soil science and geosciences are available in Kellyet al.(2008)and Calvin(2013).

XAFS spectra

In XAFS,X-ray-excited photoelectrons are scattered by surrounding coordinating atoms,causing the X-ray absorption intensity to oscillate with energy.By studying these oscillating signals,it is possible to obtain the electronic and geometric local structures of the studied system.One important distinction of XAFS-based methods is that they are element specific,meaning that a single element(e.g.,Cd)can be investigated in a heterogeneous medium such as soil.The XAFS spectra are usually divided into two regions:the extended X-ray absorption fine structure(EXAFS)and the near edge X-ray absorption fine structure(NEXAFS)or X-ray absorption near edge structure(XANES).

The XANES portion of the XAFS spectrum,which is characterized by a continuous sharp and strong increase,includes three detailed components:i)the pre-edge region that arises from the transition of the inner-core electrons to the unoccupied bound states;ii)the elemental absorption edge at the ionization energy(E0),which is generally the most prominent feature in a spectrum and whose energy position increases with atomic number;and iii)the region of 50-100 eV above the elemental absorption edge,which arises from the multiple scattering of the excited photoelectrons onto the outlying neighboring atoms(Hawthorne,1988).The XANES spectrum features mainly provide electronic structural information about the absorbing atom due to the sensitivity of this method to the valence and neighbor coordination within a structure(Schulze and Bertsch,1995).The acquisition of an XANES spectrum is typically fast(ranging from seconds to tens of minutes,depending on the experimental setup and sample),meaning that XANES analysis is suitable for time-resolved study.The spectral features seen in the XANES region are often referred to as the“fingerprint”.The simple“fingerprint”characteristics of an XANES spectrum can then be applied for the rapid identification of chemical species(such as valence states).

Unlike that of XANES,the EXAFS region is characterized by a continuous gentle and weak oscillation.The range of energies in the EXAFS region is up to about 800 eV above the absorption edge.The underlying principle of the EXAFS spectrum is that the incident X-ray photons are energetic enough to eject electrons completely from the absorber atom,and the excited photoelectrons are then coherently scattered by neighboring atoms.The vibrational frequency over this energy range(800 eV)is negatively correlated with the distance between the absorber atom and the neighboring atoms,and the amplitude is highly related to the type of ligand present and the coordination number(Schulze and Bertsch,1995).Therefore,EXAFS analysis can provide molecular information about the coordination number and interatomic distances in the geometric local environment of an absorber atom.The shape of the EXAFS spectrum is related to the short-range ordering(SRO)of several coordination shell atoms that represent the nearest atoms neighboring the element of interest(within 0.6 nm)(Brown and Parks,1989),and thus the spectra will change if the crystal structure changes.These scattering processes render EXAFS a useful complementary technique to X-ray diffraction(XRD),and it is suitable for probing the local structure of material with poorly crystalline structure and sorbed species at solid-water interface.The critical aspects of these two environments are that they dictate elemental reactivity(as in the adsorption/desorption reactions)and impact the environmental fate of contaminants.Additionally,both environments lack long-range ordering(LRO),which is typically probed using XRD.

Synchrotron facilities and beamline setup of XAFS

Synchrotron radiation facilities generally produce high energies and large fluxes of electromagnetic radiation(light).The combination of high energy and large flux means that these appliances are extremely bright sources of light.The light is produced when charged particles(electrons)move at velocities close to the speed of light in an ultra-high vacuum environment(as a storage ring).Using a series of magnets,the electrons are forced to change direction while moving around the ring,and radiation is emitted in the form of light as a result of these changes.A synchrotron facility comprises three main components:a booster ring,a storage ring,and beamlines(Fig.2).Synchrotron radiation-based X-rays are produced in the electronic storage ring for use during operation.These X-rays possess a number of unique and excellent properties,such as a wide energy spectrum(between several hundred eV and>100 keV),high brightness(at least 1×1013photons per second),high stability,and highly polarized and time-structured emission.There are more than 70 synchrotron light sources around the world(either operational or under construction)(https://lightsources.org/),of which three synchrotron radiation sources are currently operating in the mainland of China,including Beijing Synchrotron Radiation Facility(BSRF),Hefei Light Source(HLS),and Shanghai Synchrotron Radiation Facility(SSRF).There are more than 10 light sources(under construction and/or operation)in the Americas.

Fig.2 Photograph(a)and schematic of Shanghai Synchrotron Radiation Facility(SSRF)(b and c)(http://ssrf.sinap.cas.cn/).In c,numbers 1-20 indicate experimental stations,beamlines(BLs)in red(BL08U1-A,BL08U1-B,BL13W1,BL14W1,BL14B1,BL15U1,BL16B1,and BL17U1)are those constructed at the first stage,BLs in dark yellow(BL01B1,BL09U,BL17B,BL18U,BL19U1,BL19U2,BL02B,and BL03U)are for users,and BLs in light yellow(BL02U1,BL02U2,BL03SS-ID,BL03SB,BL05U,BL06B1,BL07U,BL09B,BL10U1,BL10U2,BL11B,BL12SW,BL13U,BL13SS-W,BL16U1,BL16U2,BL18B,BL20U1,and BL20U2)are under construction(as of 2016).

The general experimental setup(detectors,optics,and sample positioners)of the XAFS beamlines is shown in Fig.3.Five major components are included in the setup:i)X-ray source including the bending magnet or insertion device,both of which are located in the electron storage ring,ii)optics including the slits,mirrors,and monochromator,iii)sample and sample holder,iv)detectors,and v)the amplifiers and computers.Different groups of mirrors can be used for different purposes.For example,focusing mirrors and slits can ensure that the beam is approximately 1 mm×5 mm in size for experiments where a bulk representation of the sample is desired.However,mirrors can also be used to focus the X-ray beam to the micrometer scale for investigating soils and environmental geochemical systems.The purpose of analysis at this scale is to probe the chemical heterogeneity that is commonly observed in soil.Using a microfocused beamline,less common but more reactive species can be observed,which may not be observed in bulk sample analysis.In terms of the sample setup,two main measurement geometries can be used for bulk sample measurement.In the transmission mode,the transmission of X-rays is registeredviathe incident flux(intensity of incident X-rays,I0)and transmitted flux(intensity of transmitted X-rays,I)monitors(also termed ionization chambers).The quantity of X-rays absorbed by the sample can be measured and plotted as the natural logarithm of the ratio ofI0toI.In the fluorescence mode,the intensities of the fluorescence photon or secondary electron emissions are detected,while the sample is placed at an angle of 45°to the incident X-ray beam.

Sample preparation

The XAFS measurements can be performed on solids,liquids,gasses,wet pastes,suspensions,and even biological samples.In-situXAFS measurements of these samples at room temperature and atmospheric pressure can also be carried out,with a measurable concentration close to that of the natural environment(0.1%,weight/weight)(Brownet al.,2005).This is particularly important for soil and environmental samples because trace metals,which are often environmentally hazardous,are often present at low concentrations.

The quality of an XAFS spectrum often depends on the quality of the sample preparation,meaning that the preparation and preservation of samples are of utmost importance.For bulk XAFS measurements,the speciation of the absorber atoms should be well preserved and uniformly distributed within the sample during preparation.The size of the particles within a sample should also ensure that the element of interest is bound to particles of less than one absorption length at the energy of interest.Additionally,the concentration of an absorber element should be above that of the detection limit for the beamline used.Different modes of data collection are generally required for different samples.The transmission mode is valid for samples containing the element of interest at concentrations greater than 5%(weight/weight),while the fluorescence mode is advantageous for thin samples of high concentration or thick samples in which the element of interest is diluted to tens or hundreds of ppm.The methods of sample preparation and containment depend on the properties of the sample.It is critical to understand the concentration and thickness of a sample in order to avoid the production of artifacts in the raw data.Improperly prepared samples will result in unusable data.

Fig.3 Schematic diagram of a general X-ray absorption fine structure spectroscopy(XAFS)experimental beamline setup,where the ring on the left represents the electron storage ring(booster ring not indicated),the X-rays produced by the electrons moving around the ring are focused and filtered via the slits,monochromator,incident flux(intensity of incident X-rays,I0)monitor,fluorescence intensity of X-rays emitted from the sample(If)detector,and transmitted flux(intensity of transmitted X-rays,I)monitor located in a sealed(but accessible)experimental room called the hutch,and the sample holder is often motorized and can be controlled from the computer outside the hutch(a)and photographs of beamlines and experimental stations in Shanghai Synchrotron Radiation Facility(SSRF),China(b)and the hutch inside the sampling chamber of an XAFS beamline,BL14W1,in SSRF(c).

Solid samples can be finely ground and uniformly smeared over the sticky side of a high-purity,heat-resistant Kapton tape,after which the thickness of the sample can be adjusted by folding the tape into layers(Fig.4c)(Denecke,2006;Kellyet al.,2008).Another practical way of preparing a powder sample is to disperse the sample(if reducing the concentration is necessary)in a diluting agent(such as boron nitride powder or poly ethylene glycol(PEG))and press the resulting material to form a pellet.The agent PEG is often convenient because it helps the pellet stick together.Using the pellet pressing method means that the mass and diameter of the pellet are known,which can be useful for calculating the X-ray absorption coefficient.Pellets can be used in both transmission and fluorescent modes,depending on the composition of a sample.Liquid samples can be placed into specialized sealed containers that include windows for the entry of incidental and transmitted X-rays,while the length of the sample path can be adjusted by varying the volume.An example of a vessel that can be used with liquid samples is shown in Fig.4a;the receptacle selected depends on the composition of the sample.Moist paste samples are usually uniformly spread onto the windows of specialized sample holders(Fig.4a,b),which are constructed using various thicknesses of polymethyl methacrylate(PMMA)or Teflon.Samples obtained from mineral slurries can be filtered and mounted onto sample filters for analysis(Laffertyet al.,2010).

Fig.4 Common sample preparation methods used for collecting X-ray absorption fine structure spectroscopy(XAFS)data:plastic sample holders of various thickness onto which wet or dry samples can be mounted(a and b),folded Kapton tape with a dry powder sample spread evenly across the sticky surface of the tape,which is then folded several times to avoid pin-holes in the sample(c),and three metal pieces that can be assembled to form a metal die,which is then used to form a pellet from a dry sample(d).The metal die in d has a diameter of 7 mm and allows a user to create a precisely weighted sample of known diameter for analysis at the beamline.PMMA=polymethyl methacrylate.

Data acquisition and processing

The energy dependence of the X-ray absorption coeffi-cientµ(E)in transmission and fluorescence modes can be measured using Eqs.1 and 2,respectively:

whereIfis the detected intensity of the fluorescence of X-rays emitted from the sample,which is proportional to the absorption.Data are generally calibrated using a known reference material such as metal foil.

For both XANES and EXAFS data analyses,the initial step in the data reduction process is the pre-edge background subtraction and normalization of the edge step value from the raw data.The data are normalized to a scale between 0 and 1,which permits the comparison of spectra produced by samples at different concentrations.The XANES or EXAFS spectra can then be obtained from the raw XAFS data(Fig.5).The XANES portion is obtained by simply adjusting the energy to within an appropriate range(Fig.5b).The XANES region is usually−20 to 50 eV from the absorption edge,but can vary slightly depending on the specific needs of the user.However,processing the data in the EXAFS portion is considerably more complicated.The technique used to process EXAFS data can be divided into several steps.i)The post-edge data is subtracted from the background in order to removeµ0(E),which is the absorption coefficient in the absence of neighboring atoms(Fig.5c).ii)The energy axis is converted to the photoelectron wave vector,k,using the equation:

Fig.5 Basic steps for X-ray absorption fine structure spectroscopy(XAFS)data reduction:measured experimental raw spectra of an aqueous Cd solution and solid Cd(OH)2 reference compounds,collected in transmission modes(a),normalized,background-subtracted absorption spectra on a scale of 0-1(b),k3-weighted extended X-ray absorption fine structure(EXAFS)spectra extracted from the raw data,where k is the photoelectron wave vector calculated using Eq.3(c),magnitude of Fourier transform of EXAFS data(from k space into radial structure function(R)space)(d),and normalized,background-subtracted near edge x-ray absorption fine structure(XANES)spectra,ranging between−50 and 100 eV from the edge(e).

wheremeis the mass of an electron,ħis Plank’s constant divided by 2π,Eis the energy of the incident X-rays,andE0is the ionization energy.iii)The higherkportion of the spectra is weighted,where the amplitude of the EXAFS oscillations decays rapidly with increasingk.The purpose of weighting the spectra is to amplify the oscillations,which contain structural information.The data are generally multiplied byk2ork3for weighting(Fig.5d).iv)A Fourier transform(FT)is performed in order to filter out the frequency and amplitude of the different oscillations from which the EXAFS data are composed.The FT changes the data from wave vector(k)space into radial structure function(R)space,allowing the direct observation of interatomic scattering distances in nm.The scattering contribution to the EXAFS oscillations by each shell is decomposed and can be viewed as independent peaks(Fig.5e).InRspace(Fig.5e),the positions of the peaks are correlated with the atomic distances,but do not account for phase shifts,while the intensities of the peaks are correlated with the coordination numbers.The intensity of the peak is also affected by constructive and destructive interference of the scattering photoelectric waves.Thus,constructive interference increases the amplitude of the peak in the FT,while destructive interference decreases it.

After the data reduction steps(Fig.5),two types of data analyses are generally used for EXAFS data,the choice of which will depend on the specific experimental objectives and setup.One such method is linear combination fitting(LCF),which is a form of statistical quantitative analysis,and the other is a structural refinement method that is generally known as shell-by-shell fitting.In LCF,the spectra of an unknown sample are compared to multiple spectra of well characterized samples.This method can be used for both XANES and EXAFS analyses.In shell-by-shell fitting,the peaks seen in the FT(Fig.5d)are calculated using theoretical scattering paths that are based on a structural model.If the structural model used to fit the data is accurate,then the statistical results for each of the fitted EXAFS spectrum variable will be physically reasonable,thus indicating a good fit.The shell-by-shell fitting method can be used to obtain information on interatomic distances and coordination numbers.

CADIMUM SORPTION MECHANISMS AT SOIL MINERAL SURFACES-XAFS STUDIES

The speciation of Cd(II)in aqueous solutions is dominantly in the form of the hexahydrate cation(Cd(H2O)2+6)under conditions of low Cd concentrations(e.g.,<0.1 mol L−1)and acidic and weakly alkaline pH(pH 2-8).Cadmium is mononuclear(Cd(OH)(2−n)+n(n=1-4))at higher pH because of the formation of hydrolysis products(e.g.,pH≥8).Multinuclear hydrolysis products(Cd2OH3+and Cd4(OH)4+4)also start to form in small amounts as a result of hydroxide precipitation in concentrated Cd(II)solutions(Baes and Mesmer,1986;Bochatayet al.,2000).In solution,Cd(II)is generally coordinated by six oxygen atoms,viaoctahedral coordination(Greenwood and Earnshaw,1984).Also,Cd(II)can exhibit higher coordination numbers that range from 8 to 12 owing to its longer average ionic radius of 0.095 nm,when coordinated with six oxygen atoms in solid form,as compared with 0.070 nm forVINi2+,0.073 nm forVICu2+,0.0734 nm forVICo2+,and 0.075 nm forVIZn2+(Shannon,1976).

Several macroscopic batch and spectroscopic studies of Cd(II)sorption have been conducted using several sorbents,including iron hydroxides(Spadiniet al.,1994;Randallet al.,1999;Manceauet al.,2000),aluminum oxides(Papelis,1995;Papeliset al.,1995;Sunet al.,2019b),manganese(hydr)oxides(Randallet al.,1998;Bochatayet al.,2000;Sunet al.,2018,2019a),and alumino silicates(Farquharet al.,1997;Gräfeet al.,2007;Vasconceloset al.,2008).In all cases,Cd(II)was found to be coordinated by six oxygen atoms.There are four possible mechanisms for the removal of Cd(II)from the solution phaseviainteraction with the mineral surfaces:i)outer-sphere surface complexation,ii)innersphere surface complexation,iii)precipitation,and iv)ion exchange.Of these mechanisms,inner-sphere complexation has been observed in all these systems,particularly the iron/manganese(hydr)oxide systems.The precipitation of Cd-rich solids has mainly been observed in aluminum oxide systems,whereas outer-sphere complexation has mainly been observed in alumino silicate systems.The ion exchange mechanism has occasionally been observed in the systems of some minerals with interlayer or tunnel structure,such as biotite and cryptomelane.

Iron oxides

The sorption of Cd2+by ferric hydroxides,including hydrous ferric oxide(HFO),goethite(α-FeOOH),and lepidocrocite(γ-FeOOH),has been studied with XAFS(Spadiniet al.,1994;Randallet al.,1999;Manceauet al.,2000).The EXAFS study by Spadiniet al.(1994)showed that Cd atoms sorb onto goethite by both sharing edges and corners with surface iron(Fe)octahedra at surface coverage from 11% to 100% and mainly forming double corner-sharing complexes(referred to as the bidentate binuclear surface complex)as the surface coverage increases,i.e.,under conditions of moderate to high Cd surface loadings.The investigation by Randallet al.(1999)also provided evidence for cornersharing surface complexes on goethite over a wide range of pH and at higher surface loadings,which was consistent with the results of Spadiniet al.(1994).The EXAFS data from Randallet al.(1999)suggested that the goethite morphology inhibited Cd sorption at edge-sharing sites because of the lower proportion of edge-sharing sites(approximately 2%)than corner-sharing sites on the surface of goethite although quantum mechanical modeling indicates that edge sharing adsorption is more energetically favorable than corner sharing adsorption.Randallet al.(1999)also indicated that the sorption of Cd onto lepidocrocite is mainly a result of sharing edges with the surface Fe octahedra,which is distinctly different from the mechanism of sorption onto goethite;this was interpreted as being due to the far greater proportion of available edge-sharing sites.However,this explanation is refuted by Manceauet al.(2000)as a result of a calculation of the surface site density,which indicates that there are not enough edge-sharing surface sites available for the sorbed Cd on lepidocrocite crystallites.This difference,found by Manceauet al.(2000),was further interpreted by the difference in bulk structures and the stacking mode of anion layers(O2−and OH−)which is hexagonal in goethite and cubic in lepidocrocite.

In addition,Cd(II)was also found to prefer higher affinity sites(edge-sharing sites)on HFO rather than those onα-FeOOH by Spadiniet al.(1994).Considering all these bonding environments,the dominant mechanism of Cd(II)adsorption by iron oxides is inner-sphere surface complexation,irrespective of edge sharing or corner sharing.According to theab initiocalculation results and the geometries of the molecular clusters(Fig.6)of Cd on goethite in Randallet al.(1999),edge-sharing bindings were characterized by significantly shorter Cd-Fe distances(0.287-0.334 nm)than corner-sharing bindings(0.352-0.376 nm).The Cd-Fe distance in the single corner-sharing structure(0.352 nm)is shorter than that of the double corner-sharing structure(0.376 nm).The geometries of Cd sorption onto iron oxides that were derived from EXAFS data in previous studies(Table I)were essentially consistent with the prediction model:the first shell Cd-O distances of the inner-sphere complexes are similar and range from 0.224 to 0.230 nm,whereas the second shell Cd-Fe distances of edge-sharing complexes(0.326-0.332 nm)are considerably shorter than those of the corner-sharing complexes(0.348-0.380 nm).

Manganese(hydr)oxides

Fig.6 Crystal structure of goethite(a and b)and optimized geometries of density functional theory(DFT)-optimized molecular clusters that represent Cd(II)adsorption complexes on goethite(c-f)(reproduced from Randall et al.(1999)with copyright permission from Elsevier).DCd-O,DCd-O(AV),DCd-Fe,DCd-Fe(1),and DCd-Fe(2)are the Cd-O,average Cd-O,Cd-Fe,short Cd-Fe,and long Cd-Fe distances in nm,respectively.

TABLE ISummary of Cd sorption mechanisms on soil minerals

The sorption of Cd(II)by manganese(hydr)oxides,including cryptomelane(KMn4O8),manganite(γ-MnOOH),andδ-MnO2,has also been studied(Randallet al.,1998;Bochatayet al.,2000;Sunet al.,2018).The study of Cd2+sorption onto cryptomelane by Randallet al.(1998)suggested that Cd2+is preferentially sorbed within the tunnel sitesviaa charge-balancing exchange with H+.Bochatayet al.(2000)suggested the formation of six-coordinated Cd(II)edge-sharing inner-sphere complexes at the watermanganite interface.The most common manganese oxide in the environment is birnessite;most studies have focused too much on the sorption behavior of Cd(II)onto the surface of birnessite while neglecting the molecular binding mechanisms and the impact of coexisting reductive ions such as Fe2+and Mn2+.One study involving the combination of powder XRD and X-ray photoelectron spectroscopy(XPS)revealed that Cd(II)was mainly adsorbed at the vacancy sites in the interlayer of hexagonal birnessite according to the relationship between Cd(II)sorption behavior and the average oxidation state of Mn(Wanget al.,2012).A differential pair distribution function(d-PDF)analysis carried out by van Genuchten and Peña(2016)demonstrated that Cd(II)was mainly adsorbed onto the cation vacancy(internal)sites,with a small portion binding with the edge(external)sites of birnessite particles.A more recent study by Sunet al.(2019a)combined XPS,XRD,and EXAFS and investigated the effect of Mn(II)on Cd(II)sorption by studyingδ-MnO2and Mn(III)-richδ-MnO2.This study revealed that in the absence of Mn(II),Cd mainly forms triple corner-sharing(TCS)inner-sphere complexes at Mn(IV)vacancy sites with a minor amount of double corner-sharing(DCS)complexes at the Mn(IV)edge sites on the surface ofδ-MnO2,while forming DCS and double edge-sharing(DES)complexes at Mn(III)edge sites on Mn(III)-richδ-MnO2,and Mn(II)competes with Cd(II)for the sorption sites,inducing partial Cd(II)migration from vacancy sites to less stable edge sites at low Mn(II)concentrations under both acidic or alkaline conditions(Fig.7).Furthermore,Mn(II)reacts with birnessite at higher pH and Mn(II)concentrations,leading to the formation of an amorphous Cd(II)-Mn(III)coprecipitate.The Cd-Mn distance of an edge-sharing complex derived from second-shell analysis varies from 0.320 to 0.337 nm in manganese oxide systems,similar to iron oxide systems,whereas the Cd-Mn distance that indicates the presence of a corner-sharing complex varies from 0.366 to 0.376 nm(Table I).

Aluminum(Al)-rich minerals

The sorption of Cd(II)onto Al-bearing minerals,including Al(hydro)oxides(e.g.,alumina and gibbsite)and aluminosilicates(e.g.,kaolinite,feldspar,and mica),has been studied using XAFS(Papelis,1995;Papeliset al.,1995;Farquharet al.,1997;Gräfeet al.,2007;Vasconceloset al.,2008;Sunet al.,2019b).Outer-sphere complexed Cd(II)species were found in aluminosilicate systems.According to studies by Gräfeet al.(2007)and Vasconceloset al.(2008),outer-sphere surface complexes form at low pH(e.g.,pH<7)in kaolinite systems,whereas inner-sphere surface complexes form at higher pH(e.g.,pH 9)(Fig.8).In the XAFS study of Cd(II)sorption by perthite,biotite,feldspar,and muscovite(Farquharet al.,1997),it was suggested that Cd(II)forms only outer-sphere complexes on perthite,feldspar,and muscovite,whereas Cd(II)was thought to form stronger surface complexes on biotite,possibly as inner-sphere complexes coupled with some ion exchange.

Fig.7 Mechanisms of Cd(II)immobilization by birnessite inferred from the results of extended X-ray absorption fine structure(EXAFS)spectroscopy(from Sun et al.(2019a),with copyright permission from Elsevier).

Fig.8 Simplified graphical presentations of inner-sphere Cd complexes on gibbsite-(a)and outer-sphere complexes at the kaolinite-water interface(b)(from Gräfe et al.(2007),with copyright permission from Elsevier).

In addition to surface complexes,Cd(II)-containing precipitates were also reported in Al(hydro)oxide systems.According to Papeliset al.(1995),Cd(II)was mainly sorbed onto Al oxides as mononuclear surface complexes and formed a disordered Cd hydroxide or Cd hydroxocarbonate precipitate at higher Cd(II)concentrations(10−3mol L−1).The first shell Cd-O distance was 0.235 nm,and the second shell Cd-Cd distance was 0.384 nm.Owing to the much larger radius of Cd2+(0.097 nm)than Al3+(0.054 nm),Cd had not been shown to form layered double hydroxide(LDH)viasurface sorption by Al-bearing minerals(Papelis,1995;Papeliset al.,1995;Gräfeet al.,2007;Vasconceloset al.,2008),differing from Ni2+,Zn2+,and Co2+as discussed in previous studies(De la Caillerieet al.,1995;Scheideggeret al.,1996,1998;Towleet al.,1997;Scheinostet al.,1999;Thompsonet al.,1999;Scheinost and Sparks,2000;Ford and Sparks,2000;Nachtegaal and Sparks,2004;Liet al.,2012;Gouet al.,2018).However,a recent study by Sunet al.(2019b)confirmed the thermodynamic feasibility of Cd-Al LDH formation by laboratory synthesis of Cd-Al LDH and density functional theory(DFT)calculations.They identified the formation of Cd-Al LDH onγ-Al2O3by LCF analysis of EXAFS and XRD data.

CADMIUM SORPTION MECHANISMS AT SOIL INTERFACES-NEW PROBING METHODS

Solid-state NMR spectroscopy

In addition to XAFS,solid-state NMR is also an effective technique that can be used to obtain information on the bonding and chemical environment of an adsorbed species at environmental mineral-water interfaces.The113Cd NMR method is sensitive and effective because the113Cd isotope possesses:i)a filleddorbital,which ensures that its coordination geometry is not impacted by the ligand field effect;ii)a spin of 1/2,which means that it makes no quadrupolar contribution to NMR relaxation and the NMR signal maintains a narrow width;iii)a relatively high natural abundance of 12.26%;and iv)an observed chemical dispersion of>900 ppm(Ellis,1983).

Several studies utilizing solid-state NMR spectroscopy have investigated the interaction between Cd and layered silicates in detail,especially the Cd-smectite interaction.Banket al.(1989)applied113Cd magic-angle spinning(MAS)NMR to study Cd-exchanged montmorillonite at a pH of 7 and observed one peak with a chemical shift of approximately 15.6 ppm.The simulated spectrum included two components of different peak widths,which was interpreted as being due to the heterogeneity of the Cd adsorption sites.Edge sites which were more specific and likely to involve a permanent charge produced a narrower signal with a peak width of 1 kHz,and interlayer sites(in positions that can be near Fe in the octahedral sheets)where nonspecific adsorption could occur produced a broader signal with a peak width of 6-9 kHz.In contrast to the conclusions of Banket al.(1989),Tinetet al.(1991)interpreted113Cd NMR peaks with two components as the result of anisotropy in the signal which was due to interaction between the interlayer Cd and hydroxyl groups in the hexagonal cavities,instead of resulting from the effect of multisite or different chemical environments.

Junet al.(1996)investigated the interaction between montmorillonite and aqueous CdCl2solution(4.11×10−2mol L−1CdCl2)and recorded a single peak at a chemical shift of−8 ppm of the sedimented montmorillonite using113Cd MAS NMR,indicating that only free Cd2+is selectively adsorbed onto the negatively charged solid surface.Consistent with the results of Junet al.(1996),Di Leo and O’Brien(1999)also observed one resonance peak at a chemical shift of approximately 10 ppm in the MAS NMR spectra,indicating that only one site is available for Cd2+sorption on montmorillonite at a low concentration of Cd(0.1 mol L−1),while at a high Cd concentration(1 mol L−1),three components were detected and interpreted as being a combination of Cd2+in the interlayer(at a chemical shift of approximately 10 ppm),Cd2+on the external surface sites(at a chemical shift of approximately−40 ppm),and CdCl+in the interlayer(at a chemical shift of approximately 116 ppm)(Fig.9).It should be noted that Cd would not usually be found at such high concentrations in a soil solution.Sullivanet al.(1998,2000)performed27Al-113Cd spin-echo doubleresonance(SEDOR)and 2D-exchange NMR experiments in order to study Cd-smectite binding and the mobility of the cations.They demonstrated that Cd was rigidly bound inside the pseudo-hexagonal cavities of the tetrahedral sheet in the dehydrated Cd-smectite binding.This was determined because27Al-113Cd coupling would not have been observed otherwise with the rather long distance between the interlayer Cd and the A1 atoms in the octahedral sheet(>0.5 nm).More recently,Di Leo and Cuadros(2003)combined113Cd and1H MAS NMR spectroscopy with Fourier transform infrared(FTIR)spectroscopy to investigate the chemical environment of Cd species on the surfaces of the dioctahedral and trioctahedral smectites,montmorillonite,and hectorite,respectively.Their study confirmed that when the Cd concentration is sufficiently low and close to those of surface water and groundwater,Cd2+is mainly adsorbed onto smectite in the interlayer and surrounded by water molecules.Their study also indicated that Cd occupies a single site in thecenter of the interlayer along thecaxis in smectites with no tetrahedral substitution as a result of symmetrical attraction from the surrounding sheets,whereas Cd atoms interact more strongly with the basal oxygen atoms in the tetrahedral sheet in smectites with tetrahedral substitution.This is because substitution in the sheet provides excess negative charges,creating secondary sites in addition to those at the center of the interlayer.

Fig.9 113Cd magic-angle spinning solid-state nuclear magnetic resonance(MAS NMR)spectra of 1 mol L−1 CdCl2-treated montmorillonite for two sets of samples with particle-size modes of 12.85(a)and 8.45(b)µm(reproduced from Di Leo and O’Brien(1999)with copyright permission from Clays and Clay Minerals).Peak A is at an average chemical shift of 116 ppm,Peak B is at an average chemical shift of−11 ppm,and Peak C is at an average chemical shift of−41 ppm.

Molecular dynamics simulation

The MD simulation involving first-principle(i.e.,ab initio)methods that are based on DFT is another practical alternative that can provide critical insight into understanding of the fundamental mechanisms controlling the thermodynamic and kinetic processes occurring at the mineral-water interface.This technique is usually used to assist and complement the interpretation of spectroscopies,such as EXAFS data.

Randallet al.(1999)usedab initioMD(AIMD)to optimize the geometries and relative stabilities of four small molecular clusters,i.e.,the four types of Cd surface complexes(double corner-,single corner-,double edge-,and single-edge sharing)on goethite.Their results verified the predicted Cd-Fe distances:double edge,0.287-0.334 nm;single edge,0.317 nm;single corner,0.352 nm;and double corner,0.376 nm.They also calculated the relative total free energies of formation:double corner,0 kJ mol−1;single edge,−115 kJ mol−1;and double edge,−309 kJ mol−1,which suggests the order of increasing cluster stability and supports the widely held theory that edge-sharing adsorption is more energetically favorable than corner-sharing adsorption.

Zhanget al.(2016)conducted first-principle MD(FPMD)simulations to model the stable complexes of Cd(II)adsorbed onto the(010)and(110)interfaces of a neutral 2:1 phyllosilicate framework(i.e.,pyrophyllite)and obtained the microscopic structures and complex free energies for the three possible binding sites(≡SiO,≡Al(OH)2/≡AlOH≡AlSiO,and vacancy sites,corresponding to monodentate,bidentate,and tetradentate complexes,respectively).Their results demonstrated the Cd-O(or Cd-apical O)distances and the desorption free energies on the(010)interface to be 0.213 nm and 5.4±0.7 kJ mol−1at the≡SiO site,0.221-0.222 nm and 8.7±1.0 kJ mol−1at the≡Al(OH)2site,and 0.221-0.222 nm(or 0.291-0.296 nm)and 16.8±2.0 kJ mol−1at the vacancy sites,respectively(Fig.10),with similar results for the(110)interface,suggesting that the vacancy sites are the most favorable and stable bonding sites.Zhanget al.(2018)further carried out FPMD simulations at elevated temperature(423 K),under which the conditions for the surface complexation of divalent and trivalent cations were significantly enhanced.Their results showed the average Cd-O(or Cd-apical O)distances and the desorption free energies on the(010)interface to be 0.217 nm and 6.3±0.3 kJ mol−1at the≡SiO site,0.225 nm and 9.6±0.5 kJ mol−1at the≡Al(OH)2site,and 0.224 nm(or 0.291 nm)and 16.3±0.9 kJ mol−1at the vacancy site.The main conclusions deduced from this study in terms of surface complex structures,free energies,and pKavalues were that they were all similar to those found by Zhanget al.(2016).Additional simulations indicated that multinuclear surface complexes of Cd(II)are stable at 423 K and that these multinuclear phases may act as precursors for the epitaxial growth of new mineral phases.

Fig.10 Equilibrium structures of Cd(II)complexes adsorbed onto three binding sites at the(010)type interface of neutral framework 2:1 phyllosilicates:≡SiO site(a),≡Al(OH)2 site(b),and vacancy site(c)(reproduced from Zhang et al.(2016)with copyright permission from Clays and Clay Minerals).The vacancy site is on the edge of the phyllosilicate mineral;however,it represents a site where Cd can bind to 2 Al atoms.For clarity,the other water molecules have been removed.Solid parts are shown as sticks;Cd(II)and the coordinated water molecules are shown as ball-and-stick models:O,red and black;H,white;Si,yellow and gray;Al,purple and dark gray;Cd,gold and light gray.Numbers in the figure indicate the Cd-O distance in nm.

IMPLICATIONS FOR SOIL REMEDIATION

Over the past forty years,immobilization or stabilization with mineral-based material to reduce the leachability and/or bioavailability of heavy metals in contaminated soils is gaining acceptance as a widely appliedin-situremediation technology.Naturally occurring rock-forming minerals that possess a high specific surface area and provide countless bounding sites with a strong affinity for metal ions have been shown to play an important role in effectively stabilizing heavy metal contaminants in soil(Berti and Cunningham,1997;Lothenbachet al.,1997;García-Sánchezet al.,1999;Chenet al.,2000).According to our literature review,Albearing minerals,especially Al oxides,are expected to act more effectively than iron and manganese oxides for reducing Cd mobility.This is because surface precipitation,which occurs commonly in the systems of Al-bearing minerals,but rarely appears in the systems of iron and manganese oxides,is a relatively valid sequestration mechanism as a result of the stable three-dimensional local structure of the precipitates.It can therefore be reasonably assumed that Al-bearing minerals may be more applicable for amending Cd-contaminated soils,in terms of both the types of sorption products and long-term stability.Furthermore,Al-containing clay minerals show great potential for the practical remediation of contaminated land in a cost-effective and environmentally friendly manner.However,there is still a lack of knowledge about the mechanisms by which Cd sorption occurs in clay minerals,meaning that more comprehensive and systematic in-depth studies are needed for better decisions on soil remediation.

Because of the high diversity and heterogeneity of contaminated soils,there is no one-size-fits-all methodology in remediation.Another long-standing challenge of soil remediation is the unknown or unexpected long-term risks that remain following remediation.This requires more precise control of the long-term fate and transport of trace metal contaminants in the complicated subsurface environment.In this regard,fundamental research on the mechanisms of metal sorption reactions at the soil mineral-water interface is not only important,but also indispensable in soil,environmental,and geochemical sciences.Without such knowledge,it is challenging to understand metal speciation in soil,difficult to make feasible and cost-effective strategies that can be applied to field-scale remediation,and impossible to establish complete risk assessment models.The speciation of trace metals in natural soil systems is complex,especially in samples with inherently low background concentrations,meaning that characterization of these species is difficult.The complex soil system is a microcosm and can be considered an aggregation of innumerable single minerals,and the systematic exploration of molecular scale sequestering of trace metals by soil minerals such as iron oxides,manganese oxides,aluminum-containing minerals(i.e.,aluminum oxides and phyllosilicates)may therefore provide necessary reference and a basis for further understanding of these complicated systems.Therefore,this review takes Cd(II),one of the most typical heavy metal contaminants in soil,as an example to summarize the Cd sorption mechanisms that occur at the mineral-water interface,aiming to provide molecular theoretical foundation for predicting the mobility of trace metals and ultimately for the remediation of polluted soils.

CONCLUSIONS AND PERSPECTIVES

The sorption of Cd at the soil mineral-water interface is a complex process,which depends on pH,concentration,and especially the type of mineral present.However,as an important and toxic metal contaminant,the environmental soil chemistry of Cd is still not well understood,and this topic needs more research to better predict Cd mobility in the environment.In this review,we underscore several areas that are worthy of further exploration.First and foremost,most of the reported molecular-level studies were based on laboratory research(Wattset al.,2019);thus,the results of these studies,while important to understanding of fundamental adsorption and surface precipitation mechanisms,may not be directly applicable in the field where Cd concentrations are much lower(ppm level).Thus,future studies should focus more on natural samples with lower Cd concentrations.As discussed above,the current synchrotron light sources only allow the adequate characterization of samples with concentrations of down to a few hundred ppm,which are still higher than those found in many soil environments.An update of the current beamline facilities,to provide higher fluxes and better sensitivity,may therefore be necessary.Besides,for Cd as a heavy element,the wave vector(k)space of its EXAFS spectra(for any of the above discussed sample)is usually dominated by the oscillation of the first Cd-O shell,which significantly suppresses the signals contributed by further shells(e.g.,Randallet al.,1999).This intrinsic limitation may raise great challenges for characterizing samples with a mixture of phases.To solve this problem,it may be necessary to develop more complementary techniques,such as a combination of DFT-based MD simulation with the calculation of Cd L-edge XANES spectra with FEFF9(Wattset al.,2019).Over the past decade,the application of metal stable isotopes in environmental science has become a promising direction for research.The stable Cd isotope is an excellent tool that can be used to track the sources of Cd pollution and provide new insights into the fate of Cd in plant-soil systems(Wiggenhauseret al.,2016).Currently,most of the reportedδ110Cd isotopic compositions have focused on natural rock samples in order to decipher geological records,while measurements of the chemical processes occurring in soil are sparse(Liet al.,2019).This means that the establishment of a stable isotope tracing protocol for investigating the metal sorption process is feasible(Schauble,2004;Gouet al.,2018),which is expected to offer new opportunities for understanding the interface chemistry of Cd in soil within the natural environment(Horneret al.,2011;Wasylenkiet al.,2014).

ACKNOWLEDGEMENTS

This research was funded by the National Natural Science Foundation of China(Nos.41722303,41977267,and 41473084)and the National Key R&D Program of China(No.2017YFD0800303).Special thanks for the financial support of the Opening Fund of State Key Laboratory of Environmental Geochemistry(No.SKLEG2019712)and the National 1 000 Youth Talent Program of China.


登录APP查看全文