Xiaona LIYang SONG*Mingyun JIAFang WANGYongrong BIAN and Xin JIANG

1 CAS Key Laborator y of Soil Environment and Pollution Remediation，Institute of Soil Science，Chinese Academy of Sciences，Nanjing 210008(China)

2 University of Chinese Academy of Sciences，Beijing 100049(China)

3 Institute of Botany，Jiangsu Province and Chinese Academy of Sciences，Nanjing 210014(China)

ABSTRACT Organic matter(OM)is the most critical factor in controlling the sorption-desorption of SMZ in soil，however，few studies have explored the effects of OM removal on these important behaviors among different soils.Batch experiments were conducted to investigate the sorption and desorption characteristics of SMZ in three different soils:f luvo-aquic soil(FS)，paddy soil(PS)，and red soil(RS).The SMZ sorption in the evaluated soils was dominated by physisorption.The SMZ sorption capacities of FS and PS，which had a relatively higher OM content than RS，were higher than that of RS.The SMZ sorption in FS was dominated by linear partitioning.In contrast，the SMZ sorption in PS and RS was mainly nonlinear surface adsorption.After OM removal，the SMZ sorption capacity was signif icantly reduced in FS but increased in PS and RS.Furthermore，OM removal restrained the sorption intensity of SMZ in soils.Relatively higher OM and clay contents inhibited the SMZ desorption in FS and PS.The strong negative desorption hysteresis of SMZ in the three soils indicated that SMZ was able to move into the soil solution，thereby posing a risk to humans.Taken together，the f indings of this study showed that OM indeed plays an important role during SMZ sorption-desorption in soil.

Key Words: antibiotics，isotherm，kinetics，sorption-desorption，soil organic matter

INTRODUCTION

Antibiotics，a type of“pseudopersistent”contaminant，are continuously introduced into ecosystems with the extensive reuse of manure and sewage sludge(Çalışkan and Göktürk，2010).These antibiotics can be transferred into and accumulate in biota，thereby promoting their antibiotic resistance(Jechalkeet al.，2014)and posing potential ecological toxicity and risks(Liuetal.，2017).Almost 162 000 t of antibiotics entered the environment from domestic waste materials in China during 2013，which is approximately 150 times higher than that in the United States(Zhanget al.，2015;Zhaoet al.，2016).The widespread release of antibiotics into the environment has attracted increasing concern worldwide.Of these antibiotics，sulfamethazine(SMZ)，used widely in livestock and aquaculture industries，is considered a predominant bacteriostatic sulfonamide antibiotic contaminant(Zhanget al.，2015).

Sorption is an important process that affects the environmental fate of organic contaminants in soil(Weiet al.，2017).Soil properties，e.g.，pH，cation exchange capacity(CEC)，organic matter(OM)，and minerals，can strongly affect the sorption of antibiotics in soil(Lertpaitoonpanetal.，2009;Chuet al.，2013).Because of the hydrophilic and ionizable properties of SMZ，its sorption to soil is correlated positively with soil OM content but negatively with soil pH value(Lertpaitoonpanetal.，2009).Even the SMZ speciation in solution has been shown to vary with pH values(Jiaet al.，2018).Pavlovićet al.(2014)concluded that the OM content of soil is the most critical factor in SMZ sorption.Functional groups in OM contribute to the interaction between soil and antibiotics.More than 90%of the solid phase of soils is minerals;therefore，their contribution to antibiotic retention is considerable，especially in soils with a low OM content.The interaction between OM and minerals can affect the sorption of SMZ in soils.Soil OM might shield some sorption sites，resulting in competition with adsorbates for available sorption sites on minerals，leading to reduced sorption(Sibley and Pedersen，2008).The contribution of inorganic minerals to the sorption of antibiotics is considerable and varies with mineral types(Gao and Pedersen，2005).Only a few studies have considered the effect of a single OM factor on SMZ sorption-desorption characteristics in soils.In addition，the changes in SMZ sorption onto various soil minerals upon OM removal have rarely been reported.

Consequently，this study aimed to i)compare the sorption-desorption behaviors of SMZ in soils with different soil properties and ii)investigate the effect of a single OM factor on SMZ sorption-desorption in different soils.Therefore，the sorption-desorption of SMZ in three different soils was investigated and interpreted based on kinetic and thermodynamic analysis.Moreover，the soils with OM removal were prepared to conduct SMZ sorption-desorption experiments to illustrate the effect of a single SOM factor on SMZ sorption in soils.

MATERIALS AND METHODS

Chemicals and soil samples

Chemicals，i.e.，calcium chloride(CaCl2)，hydrochloric acid(HCl)，sodium hydroxide(NaOH)，and hydrogen peroxide(H2O2)，were supplied by Sinopharm Chemical Reagent Co.，Ltd.(Nanjing，China).The SMZ standard compound(＞98.5%purity)was purchased from the Dr.Ehrenstorfer GmbH Company(Augsburg，Germany).Acetonitrile and glacial acetic acid were obtained from Tedia Company(Fairf ield，USA).The physicochemical characteristics of SMZ are shown in Table I.Ultra-pure water(Millipore synergy，USA)was used in all experiments.

TABLE I Physicochemical characteristics a)of sulfamethazine(SMZ)used in this study

The upper horizon of three different soils(0—20 cm)was sampled.Fluvo-aquic soil(FS)was sampled from the Fengqiu Agro-ecological Experimental Station(Chinese Academy of Sciences)located in Feng Qiu(Henan，China)，while red soil(RS)and paddy soil(PS)were sampled from the Red Soil Ecological Experiment Station(Chinese Academy of Sciences)located in Yingtan(Jiangxi，China).All the soil samples were air-dried and then sieved through a 2-mm sieve for the experiments.The physicochemical characteristics of these soils are listed in Table II.Soil pH was measured in 0.01 mol L-1CaCl2(soil:liquid=1:2.5)using a pH meter(Orion 4 star，Massachusetts，USA)，total organic carbon(OC)content in soil was quantif ied by the potassium dichromate oxidation method，soil CEC was determined using the method of NH4Ac exchange，dissolved organic carbon(DOC)in soil was determined using a TOCV CPH total carbon analyzer(Shimazu，Japan)equipped with an automatic sampling system and TN M-1 unit，the total contents of nitrogen，phosphorous，and potassium were also measured by the conventional standard methods(Lu，2000)，and the soil texture(the contents of sand，silt，and clay)was determined with a Bouyoucos densitometer.The mineral composition of these soils was determined using an X-ray diffractometer(XRD，Ultima V，Japan)equipped with Debye-Scherrer transmission geometry，and the samples were introduced into a 1 mm-diameter Lindemann glass capillary.The XRD data were compared with mineral standard data(International Center of Diffraction Data，ICDD)(Fig.1).The mineral percentage in each soil sample was calculated by analyzing the ratio of the maximum diffraction peak of soil minerals and standard minerals(Table III).

Fig.1 X-ray diffraction(XRD)of the f luvo-aquic soil(FS)，red soil(RS)，and paddy soil(PS)studied and those from which organic matter had been removed(FS-RO，RS-RO，and PS-RO，respectively).

TABLE II Physicochemical characteristics a)of the studied soils

Sorption kinetics

Batch equilibration experiments were conducted to determine the sorption kinetics of SMZ in soil.Brief ly，500 mg of soil was weighed in a 20-mL brown bottle(to prevent photolysis)with polytetraf luoroethylene-lined lips.Then，10 mL of 1.4 mg L-1SMZ stock solution dissolved in 0.005 mol L-1CaCl2was added to the bottles.Mercuric chloride(100 mg L-1)was added to prevent the microbial degradation of SMZ.The bottles were shaken at a constant temperature(25°C)and agitation speed(120 r min-1)for 0.5，2，5，10，16，24，48，72，96，and 168 h.The supernatant in each bottle was sampled and f iltered through a 0.22-μm membrane before analysisviahigh-performance liquid chromatography(HPLC;LC-20A，Shimadzu，Japan).The pH values of the SMZ solution were 5.97，5.45，and 5.00 for FS，RS，and PS，respectively.A blank control was established with a pure chemical solution，and three replicates were sacrif iced for analysis.

TABLE III Mineral composition of the paddy soil(PS)，red soil(RS)，and f luvo-aquic soil(FS)studied and those from which organic matter had been removed(FS-RO，PS-RO，and RS-RO，respectively)

Sorption kinetic models quantitatively describe the solute uptake rates that control the residence time of the adsorbates at the solid-solution interface.The following sorption kinetic models were used to describe the sorption of SMZ to soil:pseudo-f irst-order，

pseudo-second-order，

intra-particle diffusion，

and elovich，

whereQt(mg kg-1)is the amount of adsorbate adsorbed per unit of soil at time(t)，Qfe(mg kg-1)is the amount of adsorbate adsorbed per unit of soil at equilibrium，obtained from the pseudo-f irst-order model，Qse(mg kg-1)is the amount of solid SMZ at equilibrium，obtained from the pseudo-second-order model，t(h)is the sorption time，kf(h-1)is the sorption rate constant of the pseudo-f irst-order model，ks(kg(mg h)-1)is the sorption rate constant from the pseudo-second-order model，ki(mg(kg h0.5)-1)represents theith sorption rate constant of the intra-particle diffusion model，Ci(mg kg-1)represents the boundary layer thickness，a(mg kg-1)is the initial sorption rate，andb(mg(kg h)-1)represents the desorption constant，indicating the degree of surface coverage and activation energy of chemical sorption.

The Boyd model was further used to verify whether f ilm diffusion and intra-particle diffusion control the total sorption rate(Boydet al.，1947;Dekaet al.，2014)，as follows:

whereF(F=Qt/Qe)indicates the fractional attainment of equilibrium at timet(h)，Btis a mathematical function ofF，and if the value ofBtplotted againsttresulted in a line passing through the origin，it means that the diffusion coefficient does not vary withFover the range of time(Viegaset al.，2014).In other words，intra-particle diffusion，rather than f ilm diffusion，is the sorption rate controlling step.

Sorption/desorption isotherms

The SMZ sorption was determined in the three types of soils.Brief ly，500 mg of soil was added to 10 mL of SMZ solution in 20-mL brown glass f lasks，which were then tightly sealed with polytetraf luoroethylene bottle caps during the experiments.The initial concentrations of SMZ were 0，0.1，0.2，0.5，1.0，2.0，3.0，4.8，7.5，and 9.8 mg L-1，in 0.005 mol L-1CaCl2solution.The f lasks were shaken for one week(168 h)and then centrifuged.Subsequently，the suspensions were sampled and f iltered through a 0.22-μm membrane.The SMZ concentration in solution was analyzed by HPLC.The pHvalues of remaining supernatants were also measured with a pH detector(Thermo Scientif ic，China).Simultaneously，500 mg of soil was added to 10 mL of 0.005 mol L-1CaCl2and shaken for 168 h as a blank control.The initial pH values of the SMZ solution were 5.97，5.45，and 5.00 in FS，RS，and PS，respectively，whereas these values were 7.02，5.02，and 5.31 in the soils from which OM have been removed(RO)，respectively.Each test was carried out in triplicate.

Two typical isotherms were used to f it the sorption and desorption data:Freundlich，

and Dubinin-Radushkevich，

whereQe(mg kg-1)is the equilibrium SMZ sorption/desorption amount andCe(mg L-1)is the equilibrium SMZ sorption/desorption concentration in solution，Kf((mg kg-1)(mg L-1)-n)represents the sorption/desorption capacity，1/nis the linear constant in the Freundlich isotherm，representing the sorption/desorption intensity，Qmax(mg kg-1)is the maximum sorption/desorption ability，β(mol2kJ-2)is the constant for the mean free energy of sorption/desorption，andεp(kJmol-1)is the Polanyi potential，withεp=RTln(1+1/Ce).The mean free energy of sorption/desorption(E)was calculated asE=(2β)-0.5(Zhanget al.，2011).

The thermodynamic parameter of Gibbs free energy(ΔG)was specif ied as

whereRis the gas constant(8.314 J(mol K)-1)，Tis the temperature in Kelvin(K)，andT=298.15 K(Fenget al.，2015).

Desorption experiments were conducted immediately after sorption equilibration.All bottles used in sorption isotherm experiments were centrifuged，and the supernatant was discarded.The desorption experiments were then conducted using the bottles.In brief，10 mL of 0.005 mol L-1CaCl2were added to the bottles，followed by shaking for 168 h.The sampling and measurement methods were the same as those used in the sorption isotherm experiments.

Effect of OM on SMZ sorption to soils

To eliminate OM，a procedure modif ied from the method presented by Baoet al.(2009)was carried out，wherein more than 80%of the OM and less than 1%of the minerals were removed.Brief ly，soil samples(100 g)were weighed and placed in a glass dish with a diameter of 125 mm，and a small amount of ultra-pure water was added to wet the soil.The soil was evenly mixed with 20 mL of 30%H2O2，after which the glass dish was covered.The soil samples were oxidized in a water bath(approximately 60°C)with continuous stirring during oxidation until the soil color was light and no bubbles were emitted.The H2O2supernatant was discarded after cooling.Finally，the soil samples were dried at 90°C and were ground for the experiments.

The sorption and desorption experiments of the soils from which OM was removed were conducted in the same manner as those of the original soils.

Determination of SMZ

The SMZ concentration in solution was analyzed using an HPLC system(LC-20A，Shimadzu，Japan)equipped with a Supelco C18reversed-phase column(200 mm×4.6 mm，5μm，Supelco，USA)and an ultraviolet detector(UV，SPD-20A，Shimadzu，Japan).The column temperature was maintained at 25°C，and a mobile phase，consisting of acetonitrile(18%)，Milli-Q water(81%)，and acetic acid(1%)，was used at a constant f low rate of 0.8 mL min-1.The UV detection wavelength was 265 nm.An external standard method was used for the quantif ication of SMZ，and the calibration curve was linear in the concentration range of 0.01 to 200 mg L-1(P＜0.05，R2=0.9997).

Quality control and data analyses

The average recovery of SMZ was higher than 97%，and the SMZstandard samples were retested every 10 samples.To quantitatively describe the rates of sorption and the limiting steps，f ive sorption kinetic models，i.e.，pseudo-f irst-order，pseudo-second-order，Elovich，intra-particle diffusion，and Boyd models，were used to f it the kinetic data.Moreover，Freundlich and Dubinin-Radushkevich isotherms were used to f it the sorption-desorption isotherm data of SMZ in soil，to further understand the sorption-desorption process.A desorption hysteresis parameter(hysteresis index，HI)was calculated as the ratio between sorption and desorption intensities from the Freundlich isotherm，and was used to evaluate the irreversible sorption process(Coxet al.，1997).

Statistical analyses

The f itting of data to various models was conducted with SigmaPlot 12.0.Signif icant differences atP＜0.05 were assessed using analysis of variance(ANOVA)in SPSS Statistics 19.0.

RESULTS AND DISCUSSION

Sorption kinetics of SMZ in soil

Sorption kinetics models are often used to quantify the rates of sorption and limiting factors.Pseudo-f irst-order and pseudo-second-order models showed that the sorption equilibrium of SMZ in PS and RS occurred within 72 h，whereas that in FS occurred until 168 h(Fig.2a).The rapid initial sorption of SMZ in soil might be due to the high SMZ concentration gradient between the solution and soil，which facilitates SMZ sorption(Fenget al.，2015).The Elovich model f itted the sorption kinetics data better than did the pseudo-f irst-order and pseudo-second-order models，yielding higherR2values(Fig.2b and Table IV).The rate constant(a，mg kg-1)showed that the initial rate of sorption followed the order of PS＞RS＞FS(Table IV)，providing conf irmation of the slowest sorption equilibrium of SMZ in FS.

The intra-particle diffusion model was used to determine whether intra-particle diffusion is a rate-limiting step.Fig.2c shows the two-stage linear form of SMZ sorption in FS and the ternary forms in PS and RS.The rate constant in the f irstsorption step(k1)in this model followed the order PS＞RS＞FS，which is consistent with the results from the Elovich model.With respect to adsorption-diffusion，the sorption process involves three steps:f ilm diffusion，intra-particle diffusion，and surface adsorption(Hameedetal.，2008).None of the plotted lines passed through the origin，indicating that intra-particle diffusion was not the sole rate-limiting step.There may be some external diffusion or chemical action in the process of SMZ sorption in soil(Cáceres-Jensenet al.，2013).The rate-limiting step was further examined by the Boyd model(Fig.2d)(Sarkaret al.，2003).None of the plotted lines passed through the origin，indicating that external f ilm diffusion may be one of the important ratelimiting steps during the SMZ sorption process in the three soils.This f inding also supported the results from the pseudof irst-order and pseudo-second-order models，showing that the initial concentration gradient between soil and solution mainly governed the rates of SMZ sorption in the three soils.

Fig.2 Sorption kinetics of sulfamethazine(SMZ)in the f luvo-aquic soil(FS)，red soil(RS)，and paddy soil(PS)studied:pseudo-f irst-order model(solid lines)and pseudo-second-order model(dashed lines)(a)，Elovich model(b)，intra-particle diffusion model(c)，and Boyd model(d).Qt(mg kg-1)is the sorption amount of SMZ in soil within the contact time t(h);Bt is a mathematical function of F，where F(F=Qt/Q e)indicates the fractional attainment of equilibrium at time t.

TABLE IV Kinetic parameters a)of sulfamethazine(SMZ)sorption in f luvo-aquic soil(FS)，red soil(RS)，and paddy soil(PS)

Sorption isotherms of SMZ in natural soils and after OM removal

The effect of OM on the sorption process was studied by comparing the sorption of SMZ in natural soils with that in the soils with OM removal.The sorption isotherms of SMZ in the three soils were all well f itted by the Freundlich model(Fig.3a and Table V).The Freundlich constantsKfand 1/nrepresent the sorption capacity and intensity of soil，respectively(Freundlich，1906).The SMZ sorption capacity，ranked as FS≈PS＞RS，was correlated with the OM contents in the studied soils(Tables II and V).Both soil OM and minerals contributed to the sorption of antibiotics.A relatively high content of OM could expose many functional groups，contributing to the strong sorption of FS and PS to SMZ(Table V)，whereas a relatively low OM content might compete with SMZ for sorption sites，resulting in a poor sorption of RS to SMZ(Table V).In addition，a previous study showed that the linear-partition coefficient(Kd)for SMZ are approximately 2.6 times lower in sandy soil than in loamy sand soil(Thiele-Bruhn，2003)，indicating that higher clay content was able to retain more SMZ in soil.This behavior is also a good explanation for the greater sorption capacities of SMZ in FS and PS than in RS.The sorption capacity values(2.94—5.75(mg kg-1)(mg L-1)-n)for SMZ in these three soils were relatively low，indicating that SMZ was likely to move from the soil into solution，and caused environmental risks through transportation(Pavlovićet al.，2014).

Fig.3 Sorption isotherms of sulfamethazine(SMZ)in the f luvo-aquic soil(FS)，red soil(RS)，and paddy soil(PS)studied and those from which organic matter had been removed(FS-RO，RS-RO，and PS-RO，respectively):Freundlich model(a)and Dubinin-Radushkevich model(b).Vertical bars are standard errors of means(n=3).Q e=the equilibrium SMZ sorption amount;C e=the equilibrium SMZ sorption concentration in solution;εp=the Polanyi potential.

TABLE V Isothermal parameters a)of sulfamethazine(SMZ)sorption and desorption in the paddy soil(PS)，red soil(RS)and f luvo-aquic soil(FS)studied and those from which organic matter was removed(FS-RO，PS-RO，and RS-RO，respectively)

In natural soils，the value of SMZ sorption intensity for FS approached 1(Table V).In addition，mechanisms of SMZ partitioning between soil organic matter and water phase dominated during this sorption process(Chiouet al.，1977).However，the sorption intensity values for PS and RS were considerably lower than 1(Table V)，implying that nonlinear surface adsorption dominated the SMZ sorption in these two soils.Compared with original soils，OM removal resulted in a signif icant decline in the sorption capacity of SMZ in FS，whereas an approximately 2 and 1.5-fold increase was observed in PS and RS，respectively(Table V).This f inding can be ascribed to the varied sorption mechanisms of SMZ in the three studied soils.Removal of OM resulted in exposure of more available adsorption sites in soil mineral，thereby promoting the surface adsorption of SMZ in PS and RS.In addition，all sorption intensity values deviated from 1 upon OM removal(Table V)，which is indicative of decreased sorption intensity.Soil minerals played an important role in SMZ sorption after the removal of OM(Houet al.，2010);therefore，this result also indicated that SMZ sorption in soil minerals was a nonlinear distribution process.As shown in Table III，the mineral compositions of PS and RS were similar to each other but different from that of FS.Specif ically，certain oxides，such as gibbsite and hematite，were present only in PS and RS(Table III).This may also be a good explanation for the difference in sorption mechanisms of SMZ onto PS and RS in contrast to FS.Moreover，the sorption capacity of SMZ in PS-RO and in RS-RO was greater than that in FS-RO.

The sorption behavior of ionizable antibiotics varies with pH values，owing to their complex speciation in the liquid phase.In this study，the pH values(5.0—7.0)of the three soils were in the range of acid/base dissociation constant(pKa1(2.28)to pKa2(7.42))of SMZ;thus，SMZ0dominated in the solution.According to the speciation distribution of SMZ in solutions with different pH values(Jiaet al.，2018)，there is approximately 23.2%SMZ in the FS solution(pH=6.85).Previous results showed that neutral species can be distributed in OM or can form relatively weak electrostatic interactions，including Van der Waals forces，H-bonding with the surface functional groups(—OH，—COOH，—C=C)of soil，and bridging with water(Linget al.，2016)，whereas ionic species can integrate with adsorbents throughπ-π electron donor-acceptor(EDA)interaction，and anions can be adsorbed by proton exchange with nearby water(Teixidoet al.，2011;Guoet al.，2017).As shown in Fig.4，the pH value of solution declined during the SMZ sorption process in PS and RS，whereas it increased in FS.These f indings may prove that SMZ could undergo tautomerization and be converted to ionic SMZ in acidic solution，which could be strongly adsorbed by soil，accompanied by the release of H+(Teixidoet al.，2011).However，proton exchange between SMZ-and surface water might result in increased pH in FS(Niet al.，2011;Teixidoet al.，2011).Moreover，the sorption ability of ionizable compounds in soil varies with speciation(Kahle and Stamm，2007).Another study found that the sorption of ionizable sulfathiazole(ST)on minerals was in the order ST+＞ST0＞ST-(Kahle and Stamm，2007).In this case，the lowest sorption ability of SMZ in FS-RO was may also be ascribed to SMZ-composition.

Fig.4 pH values of supernatants at the sorption and desorption equilibrium of sulfamethazine(SMZ)in the paddy soil(PS)，red soil(RS)，and f luvo-aquic soil(FS)studied(a)and those from which organic matter was removed(PS-RO，RS-RO，and FS-RO，respectively)(b).Vertical bars are standard errors of means(n=3).

The Dubinin-Radushkevich model was used to assess the interaction forces in the sorption process and distinguish chemical and physical sorption in adsorbent(Navikaiteetal.，2016).The sorption energy(E)，def ined as the free energy consumed per unit mol ion transferred to the surface of the adsorbent，was calculated.A physical interaction process typically dominates whenEranged from 0 to 8 kJmol-1，as opposed to a chemical interaction process，which dominated whenEwas between 8 and 16 kJ mol-1(Çalışkan and Göktürk，2010).Here，regardless of whether OM was removed，theEvalues of SMZ sorption in soils were all lower than 8 kJmol-1(Table V)，indicating a dominant physical sorption process.TheEvalues in soils with SOM removal were lower than those in original soils(Table V)，indicating a weaker interaction force between SMZ and soil minerals.This f inding is consistent with the decreased sorption intensity values obtained from the Freundlich isotherm and the increasedβvalues obtained from the Dubinin-Radushkevich model.TheΔGvalues(-2.45 to-5.79 kJmol-1)obtained from the Freundlich isotherm conf irmed that a spontaneous physical sorption process dominated SMZ sorption in soil，which occurs in the range of-20 to 0 kJmol-1(Vithanageet al.，2014).

Desorption isotherms of SMZ from natural soils and af ter OM removal

Desorption isotherms of SMZ from soils were also well f itted by the Freundlich and Dubinin-Radushkevich models(Fig.5 and Table V).As shown in Table V，the value of the desorption capacity was far higher than that of the sorption capacity;in addition，all the desorption intensity values were higher than 1 and followed the order:RS＞PS≈FS.This f inding indicated that the desorption intensity of SMZ in soils was high，which was an opposite trend to that of the OM and clay contents.In this case，OM plays an important role in the immobilization of SMZ in soils，reducing its environmental risks of transportation.It is reasonable to assume that the stronger the SMZ sorption intensity in soil，the relatively harder is the SMZ desorption from soil.Compared with natural soils，the desorption intensity values approached 1 in PS and FS with OM removal，indicating that OM removal favored the release of SMZ from the two soils.Pignatello and Xing(1996)also found that OM dominated the slow desorption of contaminants.However，the desorption intensity value changed non-signif icantly with OM removal in RS，which may be ascribed to the naturally low OM contents in RS.TheΔGvalues(-6.71 to-15.54 kJ mol-1)obtained from the Freundlich isotherm and theEvalues obtained from the Dubinin-Radushkevich model together indicated a spontaneous physical desorption of SMZ from the studied soil.TheEvalue for RS was lower than that for PS and FS(Table V)，indicating that less energy was consumed during the release of SMZ from RS than in PS and FS.This result was consistent with the desorption intensity value obtained from the Freundlich isotherm，wherein the desorption intensity was the highest in RS(Table V).Therefore，the relatively lower OM content in RS was benef icial to the desorption of SMZ from this soil.

Fig.5 Desorption isotherms of sulfamethazine(SMZ)in the f luvo-aquic soil(FS)，red soil(RS)and paddy soil(PS)studied and those from which organic matter had been removed(FS-RO，RS-RO，and PS-RO，respectively):Freundlich model(a)and Dubinin-Radushkevich model(b).Vertical bars are standard errors of means(n=3).Q e=the equilibrium SMZ sorption amount;C e=the equilibrium SMZ sorption concentration in solution;εp=the Polanyi potential.

Desorption hysteresis of SMZ in natural soils and af ter OM removal

The index HI was used to describe an irreversible sorption process.Theoretically，no hysteresis occurs when HI is in the range of 0.70 to 1.00，if the sorption and desorption rates are similar;positive hysteresis is characterized by HI＜0.7，where the rate constant of desorption is slower than that of sorption，whereas negative hysteresis occurs when HI＞1(Barriusoet al.，1994;Zhu and Selim，2000).Based on the sorption and desorption intensity factors obtained from the Freundlich isotherm，the desorption intensity value was far higher than the sorption intensity value，and thus，the HI values for the studied soils were all higher than 1(Fig.6).This condition is indicative of“negative hysteresis.”This negative hysteresis was attributed to SMZ retention only in the external or inter quasi-crystal water(Zhu and Selim，2000;Pils and Laird，2007).Perlingeret al.(1993)explained this phenomenon as structuring forces that occur when particles approach each other closely.This effect was attributed to the nearby water particles that change the intermolecular bonding and facilitate the desorption process of adsorbates(Low，1979).

As shown in Fig.6a，the hysteresis gradually decreased from RS(HI=4.60)to PS(HI=3.39)and FS(HI=1.52).The strongest negative desorption hysteresis occurred in RS，which contained the lowest level of OM.This f inding can be explained by the fact that the relatively higher levels of OM and clay in FS and PS than in sandy RS resulted in their stronger sorption abilities(Coxet al.，1997;Thiele-Bruhnet al.，2004).The OM removal resulted in decreased HI values for PS and FS，and increased HI values for RS，but all the HI values remained higher than 1(Fig.6b).Clay may also play an important role in controlling the SMZ release from soils.The large clay content facilitated the pore retention of SMZ in PS and FS，resulting in restricted negative hysteresis of SMZ in the two soils.In contrast，RS，with sand particles，contains many large pores，thereby exhibiting a poor f ixation capacity and leading to an enhanced negative desorption hysteresis of SMZ in RS with OM removal.This f inding is consistent with the results from the desorption isotherm，as it can be seen that the desorption intensity and capacity values increased for RS with OM removal，whereas they decreased for PS and FS with OM removal(Table V).The negative hysteresis emphasized the mobility of SMZ in soils(Barriusoet al.，1994)，and the strongest negative hysteresis was observed in RS，which contained small amounts of SOM and clay particles.

Fig.6 Freundlich isotherms of sulfamethazine(SMZ)sorption and desorption in the f luvo-aquic soil(FS)，red soil(RS)，and paddy soil(PS)(a)studied and those from which organic matter had been removed(FS-RO，RS-RO，and PS-RO，respectively)(b).HI=hysteresis index.

CONCLUSIONS

In natural soils，the relatively higher OM contents resulted in stronger SMZ sorption capacity in PS and FS than in RS.Removal of OM resulted in increased sorption capacity of SMZ in PS and RS but decreased the SMZ sorption capacity in FS.In addition，OM restrained the desorption of SMZ from soil，as can be seen in the negative desorption hysteresis，which followed the order RS＞PS＞FS.Therefore，OM played an important role in SMZ sorption-desorption within the soils.When OM sorption to SMZ was dominated by linear partition，higher OM content was benef icial for restraining SMZ in the soils.However，when surface adsorption dominated the soil sorption of SMZ，for example，in PS and RS，OM removal promoted the SMZ sorption.The f indings of this study contribute toward understanding the role of OM in the fate and transportation of SMZ in soil.

ACKNOWLEDGEMENTS

This study was f inancially supported by the National Key Research and Development Program of China(No.2017YFD0800704)，the National Natural Science Foundation of China(Nos.41671236 and 41877032)，and the Key Program of Frontier Sciences，Chinese Academy of Sciences(No.QYZDJ-SSW-DQC035).