Chemical Engineering Journal 220 (2013) 125–132

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier .com/locate /cej

Soil remediation using soil washing followed by Fenton oxidation

1385-8947/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.cej.2012.11.137

⇑ Corresponding author. Tel./fax: +34 913944106.E-mail address: jmrosas@quim.ucm.es (J.M. Rosas).

J.M. Rosas ⇑, F. Vicente, A. Santos, A. RomeroDpto Ingenieria Quimica, Facultad de Ciencias Químicas, Universidad Complutense Madrid, Ciudad Universitaria S/N, 28040 Madrid, Spain

h i g h l i g h t s

" Soil washing with a nonionic surfactant Tween 80 was analyzed." A model considering surfactant adsorption and p-Cresol extraction was proposed." Soil washing was used in combination with Fenton oxidation." Reaction is selective to the contaminant degradation." Reaction allows the recovery and reuse of the surfactant.

a r t i c l e i n f o

Article history:Received 16 May 2012Received in revised form 10 November 2012Accepted 17 November 2012Available online 26 January 2013

Keywords:Soil washingFenton reagentNonionic surfactantDesorption kinetic

a b s t r a c t

Soil washing was applied to a contaminated soil with p-Cresol by using a nonionic surfactant (Tween 80).A mathematical model has also been proposed to describe both the pollutant desorption and the surfac-tant adsorption, taking place simultaneously. The effect of temperature (20–40 �C) and surfactantconcentration (0.1–10 g L�1) have been analyzed on both kinetic rates. The kinetic desorption rate ofp-Cresol increases as the initial solubilizer concentration. Desorption of p-Cresol was slightly greater withincreasing temperature. The obtained kinetic model represents quite well the experimental results.

Soil washing wastewater (20 mg L�1 of p-Cresol and 0.86 g L�1 of Tween 80) has been treated with Fen-ton Reagent to remove the pollutant extracted (p-Cresol) and to recover the surfactant solution. The pH ofthe soil washing wastewater was about 6.5 and did not change significantly during the Fenton Reagenttreatment. Total conversions of p-Cresol were observed, at very short times, at the conditions testedfor the Fenton reaction (100 mg L�1 of H2O2 and 10 mg L�1 of Fe2+). The hydrogen peroxide was nottotally exhausted, showing conversions near 60% at 120 min. Besides, the removal of Tween 80 duringthe Fenton’s reaction was lower than 10%, which suggests that the reaction is mainly selective to p-Cresoldegradation. The toxicity of the liquids, measured by Microtox bioassay, was significantly reduced afterthe oxidation reaction, suggesting the negligible formation of degradation intermediates with highertoxicity than p-Cresol.

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

Contamination of soils by toxic and/or hazardous organic pollu-tants is a widespread environmental problem and the removal ofhydrophobic organic compounds (HOCs) from them has becomea major concern. A potential technology for rapid removal of HOCssorbed to soils is soil washing with a solubilizer solution.

Soil washing is a process that uses physical and/or chemicaltechniques to separate contaminants from soils and sediments.Contaminants are concentrated into a much smaller volume ofcontaminated residue, which is either recycled or disposed. Thistechnique can be used to treat a wide range of inorganic and organ-

ic contaminants, being used independently or in combination withother treatment technologies.

Among the different HOCs, cresol has been selected as targetpollutant. Cresols are isomeric substituted phenols with a methylsubstituent at either ortho, meta or para position relative to the hy-droxyl group. These compounds are contained in crude oil, coal tar,and fly-ash from coal and wood combustion and are used as sol-vents, disinfectants, and in the production of fragrances, antioxi-dants, dyes, pesticides, resins, and as wood preservatives. Amongcresols, p-Cresol is mainly used in the formulation of antioxidantsfor lubricating oil and motor fuel, rubber, and polymers. p-Cresol isconsidered to be toxic and has been classified as hazardous pollu-tants. p-Cresol has been selected as a representative of HOCs, sinceit is difficult to desorb from subsurface media because of slowdesorption kinetics from soils and sediments [1,2].

126 J.M. Rosas et al. / Chemical Engineering Journal 220 (2013) 125–132

In a previous work [3], the use of different extracting agents onthe p-Cresol extraction from a loamy sand soil was compared. Thebest results were obtained by the nonionic surfactant Tween 80(TW80), with extraction efficiencies higher than 70%. This extract-ing agent was very selective to p-Cresol extraction, minimizing themobilization of soil organic matter and maintaining the natural pHof the soil. This solubilizer has been used, in this work, to evaluateboth the kinetic desorption of p-Cresol and the adsorption of thisnonionic surfactant into the soil at different experimentalconditions.

In literature, various studies reported the use of soil washing incombination with other remediation techniques such as photoca-talysis [4–6], photo-Fenton oxidation [7,8], electrochemical degra-dation [9], and bioremediation [10]. These works are mainlyfocused on the removal of the corresponding pollutant and, theextracting agent in the liquid waste is majority or totally degradedin most of them.

Soil washing has also been used as an initial step to decrease thepollutant concentration in soils. The remaining pollutant in soilwas subsequently oxidized by using a Fenton Like process. [11–13]. Another recent work [14] proposes the simultaneous use ofsurfactants and chemical oxidants to improve the abatement ofNonAqueous Liquids Phases in soils.

To our best knowledge, only few works analyze the contami-nant removal, by coupling soil washing with the oxidation tech-nique based on Fenton’s reagent (FR), to treat the washingsolution [15,16]. In one of them, the authors analyzed the remedi-ation of a contaminated soil with TNT by soil washing, and the sub-sequent treatment with FR to treat the liquid residue. The authorsuse water as extracting agent and, as they reported, they consumedlarge volumes of water to meet remediation goals, increasing sig-nificantly the washwater volume to be treated with Fenton’s re-agent [15]. The other work monitored the removal of PAH, byusing Triton X-100 and Igepal CA-720 as extracting agents, fol-lowed by the use of FR to treat the soil washing washwater. Theseauthors used a wash solution containing 1% surfactant, and theyobtained a high degradation of the surfactant during the Fentonreaction [16], avoiding the recovery and reuse of the surfactantsolution. In this sense, the surfactant cost supposes the majorimplementation cost of the techniques.

In this work, a nonionic surfactant (TW80) previously selectedwas used to extract p-Cresol from a loamy sand at neutral pH witha soil organic matter content of 5%. The liquid contaminant residueobtained by soil washing was followed by Fenton oxidation pro-cess. A kinetic model is proposed to describe the kinetic desorptionof p-Cresol in the presence of TW80, as well as the surfactantadsorption kinetic. The effect of temperature and surfactant con-centration is taken into account in both kinetic expressions. Fur-thermore, the viability of using Fenton oxidation to treat the soilwashing washwater, containing p-Cresol and TW80, is also evalu-ated, in order to remove the pollutant and to get the recovery ofthe most surfactant solution, and, therefore, to improve the econ-omy of the soil washing treatment.

2. Materials and methods

2.1. Reagents

p-Cresol was supplied by Aldrich with a reported purity of 99%.The extracting agent, TW80, was purchased from Sigma–Aldrich.The soil used in this study is a loamy sand from Madrid. Prior touse, the soil was air-dried and sieved (<2 mm), and then character-ized by the following: soil pH (H2O), 7.8; organic matter content,4.9%; cation exchange capacity, 12.8 cmol kg�1, external surfacearea, 2 m2 g�1; pore volume, 0.004 cm3 g�1; total Fe, 8600 mg kg�1

and Mn, 1250 mg kg�1 [3].

2.2. Contaminated soil

The contaminated soil was artificially prepared by using uncon-taminated control soil. The soil was spiked with a solution of p-Cre-sol (400 mg L�1) dissolved in water (aqueous solubility of p-Cresolis �19,000 mg L�1 at 25 �C). The sorption experiments were per-formed by using the standard batch equilibrium method, per-formed in vials, on a shaking water bath (50 rpm), at 20 �C. Theliquid to solid ratio (VL/W) was 2.5 mL g�1, with contact time of48 h. After equilibrium conditions, the samples were collectedand centrifuged. The supernatant fraction was used to determinethe concentration remaining in the water phase.

2.3. Desorption kinetics study

A kinetic study to predict, simultaneously, the p-Cresol desorp-tion and the solubilizer (TW80) adsorption at different experimen-tal conditions was carried out by using a series of batchexperiments, varying the batch temperature between 20 and40 �C. The solubilizer concentration was also varied between 0.1and 10 g L�1 at liquid to solid ratio (VL/W) = 2.5 mL g�1. Theseexperimental conditions were selected based on a previous work[3]. Duplicate tests were performed for each experiment and theaverage value was used for figures with an experimental error low-er than 5%. Desorption kinetic of the pollutant and adsorption ofthe surfactant has been evaluated by fitting experimental data. Dif-ferential equations have been solved by using an Euler integrationmethod coupled to a non-linear regression (Marquardt algorithm)to optimize the corresponding kinetic parameters.

2.4. Fenton oxidation

The liquid contaminant residue derived from the soil washingwith 1 g L�1 of TW80 after 48 h was treated by Fenton’s reagent.The soil washing wastewater contains approximately 20 mg L�1

of p-Cresol, 0.86 g L�1 of Tween 80 and 130 mg L�1 of dissolver or-ganic matter. Fenton oxidation experiments were carried out in600 mL glass batch reactors shaken at an equivalent stirring veloc-ity around 200 rpm for 1 h at room temperature. The reaction vol-ume was 500 mL, and the reactants were added simultaneously tothe liquid waste residue at the beginning of each run. The startingH2O2 concentration was approximately 100 mg L�1, which nearlycorresponds to the stoichiometric amount of H2O2 necessary tooxidize p-Cresol to CO2 and H2O. The Fe2+ dose was 10 mg L�1.The initial pH value of the solution after the soil washing was6.5, when TW80 was used as extracting agent. This pH was notnecessary adjusted in these experimental conditions. Duplicatetests were performed for each experiment and the average valuewas used for figures with an experimental error lower than 5%.

2.5. Analytical methods

p-Cresol concentration was determined by HPLC. The surfactantconcentrations were measured by a potentiometer titration meth-od (by using an Ag/AgCl electrode) supplied by Metrohm (Metr-ohm, Switzerland, Application Bulletin No. 230/1e) [17].Hydrogen peroxide concentration in the supernatant was alsomeasured by using a potentiometer titration method (with a Ptelectrode) supplied by Metrohm. Organic acids were analyzed byion chromatography. The Total Carbon (TOC) was measured by aShimadzu TOC-V CSH solid analyzer, as described in the standardprocedure EN 13137. Fe and Mn extracted from the soil were ana-lyzed by acid extraction/atomic absorption spectroscopy (EPA3050B Method). The toxicity of the liquid samples after treatmentwas determined by means of a bioassay following the standardMicrotox test procedure (ISO 11348-3, 1998), by using a Microtox

J.M. Rosas et al. / Chemical Engineering Journal 220 (2013) 125–132 127

M500 analyzer (Azur Environmental). This standard test is basedon the decrease of light emission by Photobacterium phosphoreumresulting from the exposure to a toxicant. The microorganismswere Microtox Acute Reagent (Vibrio Fischeri) supplied by I.O.Analytical [18].

Fig. 1. Adsorption isotherm of TW80.

3. Results and discussion

3.1. Kinetic model

As a preliminary study, the kinetic desorption of p-Cresol withdifferent solubilizers has been evaluated by two conventionalmodels containing first-order and second-order equations. Desorp-tion extraction data were relatively well described by the modelcontaining a pseudo first order equation, while higher predictedresidual sums of squares were obtained for the second-order equa-tion model.

However, it is important to take into account different aspectsto develop a more rigorous desorption kinetic model, with a simpleimplementation: the cleaning process is based on a roll-up mech-anism induced by surfactant adsorption. If high enough concentra-tions are applied, hydrophobic contaminants can be solubilized byincorporation into micelles, i.e., aggregates of surfactant moleculeswhere the hydrophobic moieties form a core which is insulatedfrom the aqueous environment by the outward-oriented hydro-philic moieties. This self-assembly requires a critical micelle con-centration (CMC) to be exceeded [19]. In this sense, Liu et al.(1992) showed that surfactant micelles can assist in lowering theinterfacial tension if the dosage is above the CMC [20]. Further-more, Park and Bielefeldt (2003) reported a significantly increaseof the partitioning from Non-Aqueous Phase Liquids to the aqueousphase when the dose of non-ionic surfactants was four times aboveCMC [21].

However, Liu et al. (1992) also discovered that surfactant is lostinto the soil due to sorption which may retard the solubilization oforganic pollutants [20]. Therefore, when different solubilizers wereused in remediation of HOC contaminated soil, the adsorption ofthem should be considered, because sorption of the extractingagents onto soils results in an important loss and reduced perfor-mance for the solubilization of HOCs. Furthermore, certain extract-ing agent can be relatively expensive, and their adsorption canincrease the dosage required, leading to an excessively costlyremediation.

Based on these features, the kinetic model must consider thesurfactant concentration. Many desorption kinetic models foundin the literature do not consider the changes of the surfactant con-centration with the time caused by its own adsorption into thesoil; they are only limited to the equilibrium conditions [22–26].

In this sense, some experiments were carried out to evaluatethe adsorption of TW80 in the soil used in this work. Fig. 1 showsthe corresponding adsorption isotherm. The data show a typicaladsorption pattern as a plateau above a certain concentration.The sharp rise in sorption in TW80 can be interpreted as being achange from surfactant monomer sorption to surface coverage bysurfactant aggregates that form such hemimicelles and/or admi-celles. Furthermore, this nonlinear isotherm can influence diffusiveprocesses since the nonlinearity will result in concentration-dependent diffusion coefficients [27]. In our case, the solubilizeradsorption uptake was adequately reproduced by a Freundlichmodel, Eq. (3), with a Freundlich exponent < 1, which indicatesthat the effective diffusion coefficient rises with increasing concen-tration and as a result, equilibrium is reached faster at high surfac-tant concentration.

p-Cresol desorption rate can be described, by a pseudo first or-der equation concerning p-Cresol concentration. However, the ki-

netic model proposed also includes a dependency with thesurfactant concentration, where m is the order of the desorptionrate with respect to the surfactant concentration, Eq. (6). The equa-tions used to describe the kinetics of surfactant adsorption, Eqs.(1)–(4), and contaminant desorption, Eqs. (5)–(7), are detailed asfollows:

Kinetic model for TW80 adsorption:

qs ¼ ðCso � CsÞ �VL

Wð1Þ

dqs

dt¼ ksðqes � qsÞ ð2Þ

qes ¼ KFs � Cns ð3Þ

ks ¼ kso � exp�DHs

RT

� �ð4Þ

Kinetic model for p-Cresol desorption:

qC ¼ ðCCo � CCÞ �VL

Wð5Þ

� dqC

dt¼ kC � Cm

s ðqC � qrCÞ ð6Þ

kC ¼ kCo � exp�DHC

RT

� �ð7Þ

where Cso and Cs are the initial and the time-dependent aqueousphase concentration of TW80 (mg L�1), respectively; CCo and CC

are the initial and the time-dependent aqueous phase concentrationof p-Cresol (mg L�1); qs and qes (mg kg�1) are the amount of surfac-tant adsorbed per mass of soil at any time t (h) and at equilibrium,respectively; qc and qrC (mg kg�1) are the amount of contaminantthat remains adsorbed per mass of soil at any time t (h) andthe residual amount at equilibrium, respectively; KF ((L kg�1)(mg L�1)�n) and n are the Freundlich constant and exponent (whichis a measure of the chemical heterogeneity), respectively; m is theorder of the desorption rate with respect to the surfactant concen-tration; ks and kso (h�1) are the adsorption rate constant and its cor-responding preexponential factor; kC and kCo ((mg L�1)�m h�1) arethe desorption rate constant and its corresponding preexponentialfactor; DHs and DHc (kJ mol�1) are the adsorption enthalpy of thesurfactant and the desorption enthalpy of the contaminant, respec-tively; VL (L�1) is the volume of the liquid phase; W (g) correspondsto the soil mass; T (K) is the adsorption–desorption temperature;and R is the ideal gas constant (J mol�1 K�1)

Fig. 2a shows the kinetics of p-Cresol desorption and b the kinet-ics of TW80 adsorption, with different solubilizer concentrations at

Fig. 2. Kinetics of p-Cresol desorption (a) and TW80 adsorption (b), with differentsolubilizer concentrations at 20 �C. Symbols: experimental data. Lines: predictedvalues from parameters in Table 1.

Fig. 3. Kinetics of p-Cresol desorption (a) and TW80 adsorption (b), with 1 g L�1 ofTW80, at different contact temperatures. Symbols: experimental data. Lines:predicted values from parameters in Table 1.

128 J.M. Rosas et al. / Chemical Engineering Journal 220 (2013) 125–132

20 �C. The initial adsorbed amount of p-Cresol, qc, was 0.28 mg g�1.The entire range of solubilizer concentrations studied is above thecritical micelle concentration (CMC), where solubilization mecha-nism takes place predominantly. Two different regions can be ob-served related to the p-Cresol extraction. A fast p-Cresoldesorption takes place at the initial interval times, followed by asecond region, where a decrease in desorption rate is observed.According to Polanyi’s potential theory, the contaminants sorbedon the lower activation energy sites desorb more quickly than thoseat higher energy sites. Thus, the different regions can be ascribed top-Cresol desorption over sites with different energies.

The results indicate that higher extraction efficiencies are ob-served with the increase of TW80 concentration. Desorption rateof p-Cresol also increases as the initial solubilizer concentration.In this sense, the p-Cresol extraction value is faster reached asmuch high is the initial solubilizer concentration.

As can be seen in Fig. 2, the increase of the initial solubilizerconcentration produces higher loss of the solubilizer by its adsorp-tion into the soil. This tendency was expected because the sorptioncapacity of the soil for the surfactant was not exhausted. However,the residual concentrations of TW80 in the liquid phase are highenough to carry out the solubilization of p-Cresol.

Fig. 3a shows the kinetics of p-Cresol desorption and b the kinet-ics of TW80 adsorption, with 1 g L�1 of TW80, at different contacttemperatures. The influence of the temperature on the p-Cresolextraction is not really marked. However, desorption of p-Cresolwas slightly greater with increasing temperature. The simultaneousadsorption of TW80 is also quite favoured with the temperature in-crease. The adsorbed amount of TW80 at 40 �C is twice the amountadsorbed at 20 �C. In this case, the low increase observed on the p-Cresol extraction with the temperature is devalued with the high

loss of surfactant into the soil. In this sense, some authors reportedthat surfactant adsorption is reversible, although it appears to berate-limited [28]. Therefore, an additional washing step must benecessary for the recovery of the relative expensive solubilizer fora real implementation of this technique. It can be seen that the sim-ulated p-Cresol desorbed amounts and solubilizer adsorbedamounts consistently agree with the experimental ones. However,at low contact times, the predicted amounts of p-Cresol extractedand TW80 adsorbed are slightly deviated from the experimentalones. This fact can be mainly related to the fact that the equilibriumconditions have not been reached at the shorter extraction times.

The kinetic parameters are obtained by fitting the experimentaldata shown in Figs. 2 and 3 to the model developed in Eqs. (1)–(7)by using the following mass balances for p-Cresol and TW80:

qs ¼Z t

0ksðqes � qsÞdt ð8Þ

The time-dependent p-Cresol and TW80 concentrations are ob-tained by substituting these parameters in Eqs. (8) and (9). Thesepredicted values are shown as lines in Figs. 2 and 3. The obtainedkinetic parameters and statistical indicators are shown in Table 1.The corresponding enthalpies, for surfactant and pollutant adsorp-tion, present values of the same order of magnitude. As can beseen, the model reproduces fairly well the experimental p-Cresoldesorbed values. Moreover, the kinetic of TW80 adsorption is alsowell described by the model proposed.

The kinetic desorption constant of p-Cresol is twice the kineticadsorption constant of TW80, indicating that p-Cresol desorptionrate is higher than TW80 adsorption rate. These results suggestthat a high amount of p-Cresol can be extracted with a minimumsolubilizer loss with an adequate optimization study of the contact

Table 1Estimated constants derived from fitting experimental data to the model proposed in Eqs. (1)–(7).

kF ((L kg�1)(mg L�1)�n)

n m ks (20 �C)(h�1)

kso

(h�1)DH1 (kJ/mol)

kc (20 �C)((mg L�1)�m h�1)

kco

((mg L�1)�m h�1)DH2 (kJ/mol)

Residual error (sum ofsquares)

1.147 0.56 0.2 0.054 0.579 5.61 0.043 2.511 9.86 8.93 � 10�2

J.M. Rosas et al. / Chemical Engineering Journal 220 (2013) 125–132 129

time. However, it is important to take into account for a real appli-cation that desorption values in an intentionally spiked soil aregenerally lower than the expected for the same contaminant froma field contaminated soil.

3.2. Fenton oxidation experiments

3.2.1. Conversion profilesSoil washing was used in combination with an oxidation tech-

nique based on the Fenton’s reagent. The liquid extracted fromthe soil, with 1 g L�1 of TW80, was treated with 10 mg L�1 of Fe(II) and the corresponding stoichiometric amount of H2O2 to carryout the complete mineralization of p-Cresol desorbed. The theoret-ical stoichiometric amount of H2O2 required for total mineraliza-tion of p-Cresol was obtained from the following equation:

C7H8Oþ 17H2O2 ! 7CO2 þ 21H2O ð9Þ

A further acidification of the liquid, to obtain the theoreticaloptimum pH for Fenton reaction, was not necessary for the p-Cre-sol degradation, presenting an initial value of 6.5. Despite theknown adsorption of TW80 into the soil, there is a residual concen-tration of TW80 in the liquid residue higher than 0.8 g L�1.

It is important to mention that, when H2O2 was used as an oxi-dant in the absence of Fe2+, oxidation of the pollutant was alwaysnegligible. With the sequential addition of Fe(II) and H2O2 to the li-quid extracted with TW80, the color of the liquid changed from al-most limpid to yellow within 10 min of the reaction time, andfinally to light yellow and limpid. Fig. 4 shows the p-Cresol, H2O2

and TW80 conversions of the liquids extracted with TW80, by Fen-ton’s reagent. As can be seen in Fig. 4, a total conversion of p-Cresolis observed at very short times, while less than 10% TW80 degrada-tion, measured by potentiometer titration, was obtained. Mean-while, the hydrogen peroxide conversion gradually increases tovalues higher than 60% at 120 min. The possible formation of inter-mediates associated to the p-Cresol degradation was initially ana-lyzed by HPLC at different reaction times. However, neitheraromatic species, such as benzoquinone, or cathecol compounds,were detected in the chromatograms. The possible presence ofshort-chain organic acids was evaluated by ion chromatography.Kavita and Palanivelu (2005) reported the formation of oxalic

Fig. 4. p-Cresol, TW80 and H2O2 conversions after addition of [Fe2+]o = 10 mg L�1,[H2O2]o = 100 mg L�1 to the washwater: [TW80] = 0.86 g L�1, [p-C] = 20 mg L�1.

and acetic acid as final products after Fenton oxidation for cresolisomers [29]. In this work, only the formation of oxalic acid was ob-served during the oxidation reaction; however, the concentrationobtained (about 100 mg L�1) was higher than the one expectedfrom the p-Cresol oxidation. This excess can be derived from theoxidation of the nonionic surfactant or maybe because of the oxi-dation of different soil components extracted during the soilwashing.

The possible oxidation of a nonionic surfactant was also ana-lyzed by application of the Fenton Reagent ([H2O2] = 100 mg L�1,[Fe2+] = 10–100 mg L�1) to a solution of 1 g L�1 of Tween 80. Theconcentration of Tween 80, determined by potentiometer titration,remained very similar, and the results derived from the ion chro-matography indicated the negligible formation of oxalic acid.

On the other hand, a peak associated to the presence of oxalicacid was observed after the oxidation by FR of dissolved organicmatter obtained by extraction with pure water, suggesting thatthe formation of oxalic acid could come from the oxidation of dis-solved organic matter derived from the soil.

Most of the works using soil washing in combination with otheroxidative techniques (used for the liquid waste treatment), such asphotocatalysis [4–6], photo-Fenton oxidation [7,8], or electro-chemical degradation [9], practically find the total degradation ofthe nonionic surfactant, which could be due to the employmentof stronger oxidation conditions. However, recent works have re-ported promising results for the surfactant reuse, by using milderconditions in electrochemical degradation [30], or by using physi-cal separation techniques, such as solvent extraction [31], vaccumstripping [32] and selective adsorption by activated carbon [33,34].

The results for the treatment of soil washing wastewater, withthe FR, by using mild oxidation conditions, suggest that the reac-tion is very selective to the contaminant abatement, allowingrecovering the surfactant and its reuse in a next treatment. A fur-ther and more expensive physical separation step could be avoidedin this way.

3.2.2. Pollutant oxidation without external Fe (II) additionSome authors have reported the possibility of extracting iron

from soil by using chelating agents, in order to increase the effi-ciency of the in situ soil remediation, reducing the requirementfor soluble catalyst addition [33–35]. Besides, chelating agentsare widely used to maintain an adequate dissolved transition metalconcentration in near-neutral pH conditions, when chelatingagents are added close to the Fenton reagent in a modified Fentonprocess [36]. In a previous work, we have observed that some che-lating agents favored the extraction of the naturally occurring me-tal oxides from soils, maintaining a considerable amount of Fe insolution [37]. In view of these results, the possible extraction ofFe and Mn with TW80 w]as evaluated to the liquid extracted afterthe soil washing. The concentrations of solubilized Fe and Mn, afterthe soil washing process with TW80, were 0.12 ± 0.01 and0.53 ± 0.03 mg L�1, respectively. The results show a partial solubi-lization of transition metals (as Fe and Mn). However, theseamounts are lower than the 10 mg L�1 added to carry out the Fen-ton oxidation. Anyway, a run was performed by using the liquidextracted with TW80, only by adding H2O2, at the same previousconditions. The results showed that the only addition of H2O2 tothe effluent of the soil washing (containing 0.12 mg L�1 of iron

130 J.M. Rosas et al. / Chemical Engineering Journal 220 (2013) 125–132

lixiviated from soil) produces appreciable p-Cresol degradation,obtaining p-Cresol and H2O2 conversions of 22 and 38%, respec-tively, at 120 min. These results are quite interesting, because theuse of soil washing could provide enough amounts of solubilizedFe, in soils with higher Fe content than the one present in this soil.This could avoid the following addition of Fe (II) to cause the Fen-ton’s reaction to take place and thus limiting the solubilizer(TW80) degradation to values of less than 10%.

A new run was carried out in order to study the influence of ironextraction from soil. Soil washing was carried out by adding also1 g L�1 of sodium citrate in order to extract more iron from the soilthan the one observed by using the surfactant. Concentrations ofsolubilized Fe and Mn in the washwater were, in this case,0.56 ± 0.01 and 1.87 ± 0.09 mg L�1, respectively. The concentrationof p-Cresol after soil washing was similar to the one obtained withthe surfactant. Hydrogen peroxide, [H2O2]o = 100 mg L�1, was alsoadded to this liquid residue. p-Cresol conversion was 39% after120 min.

Therefore, the iron extracted from the soil increases by using achelating agent and, in consequence, the p-Cresol conversion alsodoes. The use or not of an extracting agent in the process, insteadof iron external addition, should be evaluated taking into accountthe time required to extract a significant amount of iron (III), thetime needed to oxidate the pollutant with that iron (III), the costof the chelating agent, the possibility of reusing or not the chelat-ing agent, and the risk of extracting not only iron but toxic metalsfrom the soil that would keep in the final wastewater.

3.2.3. pH evolutionFig. 5a presents the pH profiles of the Fenton’s reaction with the

liquids extracted with TW80 by soil washing. The pH of the liquidextracted with TW80 reaches values near the neutral pH and only aslight increase of the pH was observed. Kavitha and Palanivelu(2005) also reported that the optimum pH for Cresols degradationis ranged in a close window between 3.0 and 3.5 [29]. They ob-served a drop in degradation efficiency at pH > 4.0, due to precipi-tation of Fe(OH)3, which lowers the concentration of free solubleiron species available for reacting with peroxide. However, the li-quid extracted from the soil with TW80 presents an initialpH > 6.5, and, as aforementioned, a complete degradation of p-Cre-sol was observed without any pH adjustment. In this sense, wehave measured the concentration of Fe before and after the addi-tion of H2O2, and this value remained the same.

On the other hand, some authors observed that the pH of thesolution decreases as the oxidation progresses due to the forma-tion of degradation products which are acidic in nature, when ana-lyzed the Fenton oxidation of cresols and phenols in aqueousphase. However, the results, here obtained, when applying the

Fig. 5. Results obtained after applying the Fenton’s Reagent ([Fe2+]o = 10 mg L�1, [H2O2]o

(a) pH profiles obtained in a blank run (Tween and p-Cresol dissolved in distilled water) athe toxicity units (TU).

Fenton reagent to the aqueous phase taken out from to the soil,seem to indicate that the liquid extracted from the soil can act asa buffer solution, maintaining a constant pH and allowing the sta-bilization of Fe in solution. To confirm this, a test was carried outwith the same concentrations of p-Cresol/TW80/Fe (II)/H2O2 foundin aqueous phase extracted from the soil but by using distillatedwater. As can be seen in Fig. 5, the initial pH of the solution in thisblank run was 4.6. This value is considerably lower than the pH va-lue obtained with the soil-extracting agent solution. This differencecan be ascribed to a buffering effect of the soil, as previously re-ported [3]. After reacting for 120 min, the p-Cresol conversion inthe run was 100% while TW80 degradation was lower than 10%,and the pH decreased to 3.4, as can be seen in Fig. 5. Therefore, itseems that the liquid residue obtained by soil washing can extractdifferent components of the soil, which act as a buffer solution,limiting the possible pH changes.

The quick oxidation of p-Cresol obtained at neutral pH, at theconditions tested (Fig. 4), can be explained by the high used iron(II)/pollutant ratio (10–20 mg L�1) that corresponds to a 1:1 Fe2+/p-C molar ratio. The formed hydroxyl radicals attack the p-Cresolmolecule according to the following reactions:

Fe2þ þH2O2 ! Fe3þ þ OH� þ OH� ð10Þ

C7H8Oþ OH� ! oxidation products ð11Þ

Even if p-Cresol is not completely mineralized, at least, it wasconverted to less toxic organic compounds, as can be deduced fromthe ecotoxicity measurements, further commented.

On the other hand, reaction (11) takes place by the generationof hydroxyl radicals from hydrogen peroxide. The iron in solutionwas measured after reactions (11), (12), being this concentrationabout 10 mg L�1. Therefore, the formed Fe3+ remains in solutionprobably because it is stabilized by some organic compounds ex-tracted from the soil, which have also buffer properties. The cata-lytic cycle can be continued by means of Eq. (12) that controlsthe process. However the rate of reaction (11) is much faster thanthe rate of reaction (12).

Fe3þ þH2O2 ! Fe2þ þHO�2 þHþ ð12Þ

The influence of pH and the presence or not of substances ableto keep the Fe3+ in solution, in the reactions (11) and (12), are avery interesting topic that deserve further researching, but is outof the scope of this work.

3.2.4. Carbon content evolutionFig. 5b shows the total organic carbon content (TOC) evolution

during the Fenton’s reaction with the liquid extracted with TW80by soil washing. The initial TOC of the liquid extracted, associated

= 100 mg L�1) to the liquid residue containing [TW80] = 0.86 g L�1, [p-C] = 20 mg L�1

nd with the washwater; (b) total carbon content (TC) evolution and (c) evolution of

J.M. Rosas et al. / Chemical Engineering Journal 220 (2013) 125–132 131

to the presence of p-Cresol and TW80, corresponds to 15.55 and351.76 mg L�1, respectively. These values suggest that the initialcarbon content of the liquid extracted with TW80 (478 mg L�1)makes reference to the contribution of other organic compounds,probably associated to the presence of dissolved organic matterfrom the soil.

The carbon content evolution follows an exponential decay, be-cause of most of the TOC decrease was observed in the first interval(5 min). The total organic carbon content is reduced about 15%,after the extraction with TW80, at 120 min. These amounts sup-pose that p-Cresol degradation must be accompanied by the re-moval of other components present in the liquid phase. Thedegradation of the more labile fraction of the dissolved soil organicmatter can be probably the responsible of this TOC reduction, be-cause the degradation of TW80 is very low.

3.2.5. Study of the toxicityFig. 5c shows the evolution of the toxicity units (TU’s) during

the Fenton’s reaction with the liquid extracted with TW80. The ini-tial toxicity of the liquid residue is mainly associated to the pres-ence of p-Cresol in the residue, due to TW80 is not toxic in therange of concentrations observed [3]. The results indicate thatthe oxidation process significantly reduces the toxicity of the con-taminant residue, probably associated to the total conversion of p-Cresol observed, in spite of the additional toxicity of the remainingH2O2 present at the evaluated times. Furthermore, the toxicity val-ues tendency suggests the negligible formation of degradationintermediates with higher toxicity than p-Cresol.

4. Conclusions

A simplified model which takes into account the nonionic sur-factant (TW80) adsorption and p-Cresol extraction was proposedin this work. The model reproduces fairly well the experimentalp-Cresol desorbed values. The kinetic of TW80 adsorption is alsowell described by the model proposed. Higher extraction efficien-cies are observed with the increase of TW80 concentration. The ki-netic desorption rate of p-Cresol also increases as the initialsolubilizer concentration. However, it also produces higher lossof the solubilizer by its adsorption into the soil. The influence ofthe temperature on the p-Cresol extraction is not really marked.Nevertheless, desorption of p-Cresol was slightly greater withincreasing temperature.

Soil washing with extracting agent TW80 was used in combina-tion with an oxidation technique based on the Fenton’s reagent.The initial pH of the liquid extracted was not adjusted, presentingan initial value of 6.5. Total conversions of p-Cresol were observed,at very short times. However, the hydrogen peroxide was not to-tally exhausted, showing conversions at 120 min near 60%. Thequick oxidation of p-C noticed even at neutral pH could be dueto the Fe2+/p-C molar ratio (1:1) used. This ratio allows generatingthe amounts of hydroxyl radicals initially required to oxidate the p-Cresol in solution. Even if p-Cresol is not totally mineralized, thetoxicity of the liquids was significantly reduced after the oxidationreaction, indicating the negligible formation of degradation inter-mediates with higher toxicity than p-Cresol.

Soil washing treatment allows the solubilization of transitionmetals contained in the soil. Therefore, it would be possible theFenton’s reaction to take place without the addition of Fe (II) insoils with high Fe content. The results with this soil (low Fe con-tent) showed that the addition of H2O2 produces an appreciablep-Cresol degradation of 22% in the absence of added Fe(II).

The total carbon content decreases about 15%, after the extrac-tion with Tween 80. The removal of Tween 80 during the Fenton’s

reaction was lower than 10%, which suggests that the reaction ismainly selective to p-Cresol degradation.

Therefore, the combination of soil washing, followed by Fentonoxidation at low concentration of the oxidant and catalyst could bean adequate method to remediate contaminated soils with p-Cre-sol. Furthermore, they can be satisfactory used to treat the corre-sponding liquid residue obtained after the soil washing, yieldinga total pollutant abatement, but also recovering the surfactantsolution for its reuse in a next treatment, reducing the costs ofthe combined process.

Acknowledgments

The authors acknowledge financial support for this researchfrom the Comunidad Autónoma de Madrid provided throughoutproject CARESOIL (S2009AMB-1648) and from Spanish Ministryof Science and Innovation, projects CTM2006-00317 andCSD2006-00044-T5.

References

[1] M. Badanthadka, H.M. Mehendale, Cresols, Encyclopedia of Toxicology, seconded., Elsevier, USA, 2005. pp. 678–680.

[2] R.A. Evangelista, H.L. Allen, R.M. Manuel, Treatment of phenol and cresolcontaminated soil, J. Hazard. Mater. 25 (1990) 343–360.

[3] J.M. Rosas, F. Vicente, A. Santos, A. Romero, Enhancing p-Cresol extraction fromsoil, Chemosphere 84 (2011) 260–264.

[4] D. Fabbri, A. Bianco Prevot, V. Zelano, M. Ginepro, E. Pramauro, Removal anddegradation of aromatic compounds from a highly polluted site by couplingsoil washing with photocatalysis, Chemosphere 71 (2008) 59–65.

[5] D. Fabbri, A. Crime, Davezza, C. Medana, C. Baiocchi, A. Bianco Prevot, E.Pramauro, Surfactant-assisted removal of swep residues from soil andphotocatalytic treatment of the washing wastes, Appl. Catal. B: Environ. 92(2009) 318–325.

[6] X. Li, J. Guang, G. Yu, S. Deng, Photodestruction of BRE-99 in micellar solutionsof nonionic surfactants of Brij 35 and Brij 58, Chemosphere 78 (2010) 752–759.

[7] R.D. Villa, A.G. Trovo, R.F. Pupo Nogueira, Soil remediation using a coupledprocess: soil washing with surfactant followed by photo-Fenton oxidation, J.Hazard. Mater. 174 (2010) 770–775.

[8] E.R. Bandala, Y. Velasco, L.G. Torres, Decontamination of soil washingwastewater using solar driven advanced oxidation processes, J. Hazard.Mater. 160 (2008) 402–407.

[9] M.T. Alcántara, J. Gómez, M. Pazos, M.A. Sanromán, Combined treatment ofPAHs contaminated soils using the sequence extraction with surfactant–electrochemical degradation, Chemosphere 70 (2008) 1438–1444.

[10] R. Piskonen, M. Itavaara, Evaluation of chemical pretreatment of contaminatedsoil for improved PAH bioremediation, Appl. Microbiol. Biotechnol. 65 (2004)627–634.

[11] T.T. Tsai, C.M. Kao, R.Y. Surampalli, S.H. Liang, Treatment of fuel-oilcontaminated soils by biodegradable surfactant washing followed byFenton-like oxidation, J. Environ. Eng. 135 (2009) 1015–1024.

[12] T.T. Tsai, C.M. Kao, R.Y. Surampalli, H.Y. Chien, Enhanced bioremediation offuel-oil contaminated soils: laboratory feasibility study, J. Environ. Eng. 135(2009) 845–853.

[13] B.J. Shiaou, In situ surfactant and chemical oxidant flushing for completeremediation of contaminants and methods of using same, USA Patent,US007364386 B2, 2008.

[14] G.E. Hoag, J. Collins, Soil remediation method and composition, USA Patent,US007976241B2, 2011.

[15] Z.M. Li, M.M. Peterson, S.D. Comfort, G.L. Horst, P.J. Shea, B.T. Oh, RemedingTNT-contaminated soil by soil washing and Fenton oxidation, Sci. TotalEnviron. 204 (1997) 107–115.

[16] J.K. Saxe, H.E. Allen, G.R. Nicol, Fenton oxidation of polycyclic aromatichydrocarbons after surfactant-enhanced soil washing, Environ. Eng. Sci. 17(2000) 233–244.

[17] M. Samardzica, M. Sak-Bosnara, D. Madunic-Cacicb, Simultaneouspotentiometric determination of cationic and ethoxylated nonionicsurfactants in liquid cleaners and disinfectants, Talanta 81 (2011) 789–794.

[18] A. Romero, A. Santos, J. Tojo, A. Rodríguez, Toxicity and biodegradability ofimidazolium ionic liquids, J. Hazard. Mater. 151 (2008) 268–273.

[19] H. Lippold, U. Gottshalch, H. Kupsch, Joint influence of surfactants and humicmatter on PAH solubility. Are mixed micelles formed?, Chemosphere 70(2008) 1979–1986

[20] Z.B. Liu, D.A. Edwards, R.G. Luthy, Sorption of non-ionic surfactants onto soil,Water Res. 26 (1992) 1337–1345.

[21] S.K. Park, A.R. Bielefeldt, Equilibrium partitioning of a non-ionic surfactant andpentachlorophenol between water and a non-aqueous phase liquid, Water Res.37 (2003) 3412–3420.

132 J.M. Rosas et al. / Chemical Engineering Journal 220 (2013) 125–132

[22] W. Chu, K.H. Chan, W.K. Choy, The partitioning and modelling of pesticideparathion in a surfactant-assisted soil-washing system, Chemosphere 64(2006) 711–716.

[23] K. Yang, L. Zhu, B. Xing, Enhanced soil washing of phenanthrene by mixedsolutions of TX100 and SDBS, Environ. Sci. Technol. 40 (2006) 4274–4280.

[24] Q.-R. Zeng, H.-X. Tang, B.-H. Liao, T. Zhong, C. Tang, Solubilization anddesorption of methyl-parathion from porous media: a comparison of HPCDand two nonionic surfactants, Water Res. 40 (2006) 1351–1358.

[25] C.K. Ahn, Y.M. Kim, S.H. Woo, J.M. Park, Soil washing using various nonionicsurfactants and their recovery by selective adsorption with activated carbon, J.Hazard. Mater. 154 (2008) 153–160.

[26] C.K. Ahn, M.W. Lee, D.S. Lee, S.H. Woo, J.M. Park, Mathematical evaluation ofactivated carbon adsorption for surfactant recovery in a soil washing process, J.Hazard. Mater. 160 (2008) 13–19.

[27] W. Schlebaum, G. Schraa, W.H. Van Riemsdijk, Influence of nonlinear sorptionkinetics on the slow-desorbing organic contaminant fraction in soil, Environ.Sci. Technol. 33 (1999) 1413–1417.

[28] S.K. Park, A.R. Bielefeldt, Non-ionic surfactant flushing of pentachlorophenolfrom NAPL-contaminated soil, Water Res. 39 (2005) 1388–1396.

[29] V. Kavitha, K. Palanivelu, Destruction of cresols by Fenton oxidation process,Water Res. 39 (2005) 3062–3072.

[30] J. Gómez, M.T. Alcántara, M. Pazos, M.A. Sanromán, Remediation of pollutedsoil by a two-stage treatment system: desorption of phenanthrene in soil and

electrochemical treatment to recover the extraction agent, J. Hazard. Mater.173 (2010) 794–798.

[31] D.H. Lee, R.D. Cody, D.J. Kim, Surfactant recycling by solvent extraction insurfactant-aided remediation, Sep. Purif. Technol. 27 (2002) 77–82.

[32] S. Kungsanant, B. Kitiyanan, T. Rirksomboon, S. Osuwan, J.F. Scamehorn,Recovery of nonionic surfactant from VOC-contaminated coacervate phasesolutions by co-current vacuum stripping: Effect of surfactant concentration,temperature, and solute type, Sep. Purif. Technol. 66 (2009) 510–516.

[33] A. Rastogi, S.R. Al-Abed, D.D. Dionysiou, Sulfate radical-based ferrous–peroxymonosulfate oxidative system for PCBs degradation in aqueous andsediment systems, Appl. Catal. B: Environ. 85 (2009) 171–179.

[34] Y. Sun, J.J. Pignatello, Chemical treatment of pesticides wastes. Evaluation ofFe(III) chelates for catalytic hydrogen peroxide oxidation on 2,4-D atcircumneutral pH, J. Agric. Food Chem. 40 (1992) 322–327.

[35] X. Xu, R.X. Thomson, An evaluation of the green chelant EDDS to enhance thestability of hydrogen peroxide in the presence of aquifer solids, Chemosphere69 (2007) 755–762.

[36] A. Rastogi, S.R. Al-Abed, D.D. Dionysiou, Effect of inorganic, synthetic andnaturally occurring chelating agents on Fe(II) mediated advanced oxidation ofchlorophenols, Water Res. 43 (2009) 684–694.

[37] F. Vicente, J.M. Rosas, A. Santos, A. Romero, Improvement soil remediation byusing stabilizers and chelating agents in a Fenton-like process, Chem. Eng. J.172 (2011) 689–697.