Effect of Light Source on the Catalytic Degradation of Photocatechuic Acid in a Ferrioxalate Assisted Photo Fenton Process

Applied Catalysis B: Environmental 96 (2010) 486495

Contents lists available at ScienceDirect

Applied Catalysis B: Environmental

journa l homepage: www.e lsev ier .com

Effect o ionferriox

J.M. MonUniversity of C Super1, 13071 Ciuda

a r t i c l

Article history:Received 18 JaReceived in reAccepted 3 MaAvailable onlin

Keywords:ProtocatechuicFerrioxalatePhoto-FentonUV lightCPC

echuicial usistinon st, temcatalreml UV-UV lin in t

formed at 200280nm. OH radicalswere themain oxidative intermediate species in the articial UV-A/Cprocess while superoxide and OH radicals played the most signicant roles in the solar process. ArticialUV-A/C light can be used as an alternative to solar CPC on cloudy days.

2010 Elsevier B.V. All rights reserved.

1. Introdu

Agro-indmaking, olsignicant cthe toxicitytions of thesby conventithere is curadvanced oorganic pol

The homeffective syhydroxyl racation of senergy consaddition offormation otion efcienis a photo-the irradiat

CorresponE-mail add

0926-3373/$ doi:10.1016/j.ction

ustrial wastewaters such as those produced in wine-ive-oil extraction or table-olive production containoncentrations of phenolic compounds, contributing toof these efuents. It is well known that high concentra-e compounds prevent mineralization of these efuentsonal aerobic biological treatment processes [1,2]. Thus,rently considerable interest in developing alternativexidation processes (AOPs) for degrading these types oflutants.ogeneous photo-Fenton reaction is one of the most

stems for oxidation, generating mainly highly reactivedicals (OH), as previously reported [3,4]. The appli-olar irradiation to these systems could diminish theumption required for generating hydroxyl radicals. Theoxalic acid to the photo-Fenton system promotes thef ferrioxalate complexes, which increases the oxida-cy of the solar photo-Fenton process, as ferrioxalate

sensitive complex which expands the useful range ofion spectrum up to 550nm [5,6], making the system a

ding author. Tel.: +34 926295300x3888; fax: +34 926295361.ress: [email protected] (J.M. Monteagudo).

more cost-effective and environmentally benign treatment, as wereported previously [7].

A drawback of a solar system is that the reaction times arelonger on cloudy days. This disadvantage could be solved if anarticially UV-irradiated pilot plant with both UV-C (for photol-ysis of H2O2) and UV-A (for ferrioxalate-complex photochemicalreactions) lamps could be used in parallel on these days.

In thiswork, protocatechuic acid (PA) (Fig. 1a), typically found inagro-industrial wastes, was chosen as a model phenolic compoundto evaluate the viability of their degradation under ferrioxalate-assisted articial UV-A/C and solar photo-Fenton processes.

In recent years, ferrioxalate has been used in the photo-Fentonreaction involving ferric compounds [5,810], but there is very lit-tle information on the ferrioxalate-assisted photo-Fenton systemusinga ferrous-initiatedprocess. Theuseof ferrous sulfate is advan-tageous since it is less corrosive than ferric salts, inexpensive andmore soluble than ferric compounds. Additionally, Fe(II) is quicklyconverted into Fe(III) in the presence of hydrogen peroxide, so thatthe degradation of PA solutions is basically a process catalyzed byFe(III)-oxalate.

The degradation of protocatechuic acid solutions has been stud-ied previously using other advanced oxidation processes such asFentonoxidation [11], photo-Fenton [12], O3/UVorH2O2/UVmeth-ods [13]. However, the effect of ferrioxalate complexes on thedegradation efciency of PAusing solar or articial UV light sources

see front matter 2010 Elsevier B.V. All rights reserved.apcatb.2010.03.009f light source on the catalytic degradatalate-assisted photo-Fenton process

teagudo , A. Durn, I. San Martn, M. Aguirreastilla-La Mancha, Grupo IMAES, Department of Chemical Engineering Escuela Tcnicad Real, Spain

e i n f o

nuary 2010vised form 25 February 2010rch 2010e 9 March 2010

acid

a b s t r a c t

The catalytic degradation of protocatprocess irradiated with solar or articarried out either in a pilot plant cona UV-A/C-lamp reactor. An optimizatiincluding the following variables: pHH2O2, Fe(II) and oxalic acid. The photoelimination of the original PA and theand 96% were achieved under articiaoperating conditions. When articialratewas higher in theUV-C system tha/ locate /apcatb

of protocatechuic acid in a

ior de Ingenieros Industriales, Avda. Camilo Jos Cela,

c acid (PA) solutions in a ferrioxalate-assisted photo-Fentonltraviolet light sources was investigated. The reactions wereg of a compound parabolic collector (CPC)-solar reactor or inudy was performed using a multivariate experimental designperature, solar power, air ow and initial concentrations ofytic degradation efciency was determined by measuring theoval of total organic carbon (TOC). TOC-removal rates of 97%A/C and solar light, respectively, but with different optimumght was used in the presence of oxalic acid, the degradationheUV-A systembecause ferrioxalate complexes are primarily

J.M. Monteagudo et al. / Applied Catalysis B: Environmental 96 (2010) 486495 487

Fig. 1. (a) Proaddition on ab

has not beementalDeswas performously (pH,and oxalic a

The exp(NNs) [15,1constants (ied range asaliency anatrue relevanmineralizatcentrationsconcentratidifferent oxalso quantiversus UV-C

2. Experim

2.1. Materi

Protocatprepared bywithout fugrade) andwastewaterin situ becaPA was alw

In all exppH adjustmoxide (30% (initial concesolutions was needed pradical reacwith 1,4-bewere withdexcess Na2dure was pdegradation

xperimental setup based on a CPC and articial UV-A/C pilot plants: (a)b) schematic.

otochemical reactions

Solar-CPC pilot plantcompound parabolic collector (CPC) pilot plant (manufac-y ECOSYSTEM, S.A.) comprises a solar reactor consisting ofnuously stirred tank (50 L), a centrifugal recirculation pump,olar collector unit with an area of 2m2 (concentration fac-. This unit was housed in an aluminum frame connectingtocatechuic acid (PA) chemical structure; (b) Effects of oxalic acidsorption spectra of PA solutions.

n studied until now. In this work, a Multivariate Experi-ign according to the response-surfacemethodology [14]

ed to study the effect of all the variables simultane-air-ow rate and initial concentrations of H2O2, Fe(II)cid).erimental results were tted using Neural Networks6],which allowed the values of kinetic degradation-rateresponse functions) to be estimated within the stud-s a function of the process variables. Additionally, alysis of each variable in the NNs helped to discern theirce. As PA elimination does not necessarily imply totalionof the solution, thedecrease of bothPAandTOCcon-were measured; residual H2O2, dissolved O2 and Fe(II)ons were also determined. Finally, the contributions ofidative intermediate species to overall reaction wereed. The catalytic behavior of Fe and the effects of UV-Airradiation were also evaluated.

ental

Fig. 2. Ephoto; (

2.2. Ph

2.2.1.The

tured ba contiand a stor =1)als

echuic acid (3,4-dihydroxybenzoic acid) solutionsweredissolving PA (SigmaAldrich, 97%) in distilled water

rther purication. FeSO47H2O (Panreac, analyticalH2C2O42H2O (Panreac, 99.5%) were added to theto form ferrioxalate complexes and used immediately

use of their light sensitivity. The initial concentration ofays 20mgL1.eriments, after the addition of Fe (II) andoxalic acid andent, a measured amount of commercial hydrogen per-w/v),Merck)was added to the reactor to bring theH2O2ntration tobetween0and400mgL1. ThepHof thedyeas adjusted with 0.1M H2SO4 and 6M NaOH solutionsrior to degradation. To quantify the oxidation levels bytions, the scavenging of free radicals was accomplishednzoquinone, NaN3 and KI. Before analysis, all samplesrawn from the reactor and immediately treated withSO3 solution to prevent further oxidation (this proce-erformed in order to prevent an overestimation of the).

the tubingform tiltedhorizontal16 borosilicmeasured bfacilitated t(300400naccumulate(Programm

2.2.2. ArtiThe UV

is also shreactor (22(200280nUV-A lampUVC (for pphotochemor simultanborosilicateaccommodand valves and mounted on a xed south-facing plat-39 (latitude of Ciudad Real, Spain) with respect to theplane (Fig. 2). The total illuminated volume inside theate-glass absorber tubes was 16 L. UV irradiation wasy a radiometer (Ecosystem, model ACADUS-85), whichhe measurement of the received irradiation as UV-Am). Data in terms of incident solar power (Wm2) andd solar energy (Wh) were measured by means of a PLCable Logic Controller) coupled to the radiometer.

cial UV-A/C pilot plantpilot plant (FLUORACADUS-08/2.2, ECOSYSTEM, S.A.)own in Fig. 2 and was composed of a 28-L40mm730mm100mm), with two UV-C lampsm; PHILIPS TUV TL D 55W HO SLV UV-C) and twos (320400nm; PHILIPS CLEO Effect 70W SLV). Thehotolysis of H2O2) and UVA (for ferrioxalate-complexical reactions) lamps were both used, either singlyeously. This reactor was constructed of 7-mm-thickand protected externally by a polypropylene tube. It

ated four quartz tubes, 34mm in diameter and 1.5mm

488 J.M. Monteagudo et al. / Applied Catalysis B: Environmental 96 (2010) 486495

Table 1The 5-factor central composite design matrix. Ferrioxalate-assisted articial UV-A/C and solar photo-Fenton systems.

Coded levels Natural levels

[H O ] (mgL1) [Fe (II)] (mgL1) [H2C2O ] (mgL1) Air ow (m3 h1) pH

(+) 60.00() 0.00(+1) 42.61(1) 17.39(0) 30.00Response fu

Solar procArticial U

lar pr

Selected opt[H2O2]o (m 0[Fe (II)]o (m[H2C2O4]oAir ow (m air inpHkPA 0493WkTOC 1759W% PA elimi 0% (1%TOC rem % (18

[protocatechuTemperature tSolar power ty

thick, houscharged witinuously pand afterwatubes, the s

2.3. Analysi

AnalysisperformancAgilent Tecsampling. Awas usedacidic pHwavelengthphotometrito ISO 633The degreedeterminedation

J.M. Monteagudo et al. / Applied Catalysis B: Environmental 96 (2010) 486495 489

Table 2Equations and parameters of Neural Network ttings for protocatechuic acid solutions degradation under the processes: ferrioxalate-assisted UV-A/C or Solar photo-Fenton.

Neural Network ttingEquationa for PA elimination/Mineralization (articial UV process)kPA-UV/kTOC-UV [min1] = N1 (1/(1 + 1/EXP([H2O2]o W11 + [Fe]o W12 + [pH] W13 + [H2C2O4]o W14 + [Temperature] W15 + [Air ow] W16))) + N2 (1/(1 + 1/exp([H2O2]o W21 + [Fe]o W22 + [pH] W23 + [H2C2O4]o W24 + + [Temperature] W25 + [Air ow] W26Equationa for PA elimination/Mineralization (solar process)kPA-solar/kTOC-solar [W1 h1] = N1 (1/(1 + 1/exp([H2O2]o W11 + [Fe]o W12 + [pH] W13 + [H2C2O4]o W14 + [Temperature] W15 + [Solar Power] W16+ [Air ow] W17))) + N2 (1/(1 + 1/EXP([H2O2]o W21 + [Fe]o W22 + [pH] W23 + [H2C2O4]o W24 + [Temperature] W25 + [Solar Power] W26 + [Airow] W27)))

Weight factors Parameter Valuesofneuronsand factors toobtainkPA Values of neurons and factors to obtain kTOC

UV process Solar process UV process Solar process

N1 Neuron 0.8398 0.5686 0.0604 0.1244W11 [H2O2]o 0.7692 3.3961 1.0635 2.4712W12 [Fe(II)]o 0.1254 2.0457 0.0327 1.5996W13 pH 3.8433 0.4852 4.5689 0.5786W14 [H2C2O4]o 22.4343 0.4562 2.6453 0.7341W15 Temperature 3.5056 0.4861 1.6099 0.5150W16 Solar power 1.8059 1.7259W17 Air ow 3.1087 0.5508 0.6165 0.5400N2 Neuron 1.0855 0.3202 0.0472 0.0646W21 [H2O2]o 0.5716 30.404 1.9088 12.7232W22 [Fe(II)]o 0.0617 13.4512 0.0937 5.3229W23 pH 0.9202 7.6703 6.6234 3.6283W24 [H2C2O4]o 7.4190 9.1905 0.4845 5.1059W25 Temperature 0.9095 20.3580 0.6357 8.1274W26 1W27 8

Neural netw [H2C2

Saliency anakPA-UV (mi 59.02kTOC-UV (m 14.83kPA-solar (W 6.42kTOC-solar (W 9.59

a Parameter

2.5. Neural

In this wnation of innetwork wnential actiback-propaSolver toolrithm [17].Finally, a mSolar Power 24.869Air ow 3.1086 10.507

ork output parameters [H2O2]o [Fe]o pH

lysis of the input variables for the neural network (%)n1) 3.24 0.41 8.76in1) 14.02 0.60 52.821 h1) 31.46 16.86 5.921 h1) 27.83 15.11 7.16

s values in equations must be previously normalized to the (0.1) interval.network strategy

ork, a linear basis function was used (a linear combi-puts, Xj, and weight factors, Wij; Table 2). Each neuralas solved with two neurons and used a simple expo-vation function [15,16]. The strategy was based on agation calculation. Parameters were found using thein Excel using the Marquardt non-linear tting algo-Further details can be found in our previous report [7].easure of the saliency of the input variables was made

based on thysis of the(expressed

3. Results

3.1. Previou

Table 3the degrad

Table 3Protocatechuic acid degradation. A previous comparative study.

System PA elimination (%)

Solar 11.78UV-A/C 20.59Solar/UV-A/C 18.60H2O2 0.00UV-A/C/Fe(II) 13.62UV-A/C/H2O2 88.53UV-A/C/oxalic 48.00H2O2/Fe 7.92H2O2/Fe/oxalic 21.00UV-A/C/H2O2/Fe(II) 85.00 (30min); 100.00 (60min)Solar/H2O2/Fe(II) 88.00 (30min); 100.00 (60min)UV-A/C/H2O2/Fe(II)/oxalic 100.00 (20min)UV-A/H2O2/Fe(II)/oxalic 84.00 (20min); 100.00 (40min)UV-C/H2O2/Fe(II)/oxalic 88.00 (20min); 100.00 (30min)Solar/H2O2/Fe(II)/oxalic 100.00 (15Wh; 18min)a

Experimental conditions: [H2O2] =100ppm; [Fe(II)] = 2ppm; [oxalic] = 60ppm; pH=4.a 15Wh accumulated solar energy; 18min reaction time.b 60Wh accumulated solar energy; 44min reaction time.

36Wm2 Average solar power. 11.05750.8858 4.1391

O4]o Temperature Solar power Air ow

8.26 20.3010.63 7.0911.38 20.46 7.5011.26 21.60 7.44e connection weights of the NNs. This allowed for anal-relevance of each variable with respect to the othersas a percentage).

and discussion

s studies

shows the results of an initial comparative study onation of protocatechuic acid solutions in different sys-

TOC removal (%)

14.3415.2615.530.0012.8648.884.9910.5622.0071.3861.2080.00 (40min)57.00 (40min)70.00 (40min)81.00 (60Wh; 44min)b

490 J.M. Monteagudo et al. / Applied Catalysis B: Environmental 96 (2010) 486495

icial

tems. Bothsolutions viarticial UVconrmed treaction ofarticial UVthe PA (13oxidative ingenerated uor oxalic aclight in the pwere increais well-knophotolysis oergistic efferadicals aresufcient todegradationtion and oxafrom approsystem. Thiplexes whiaromatic rinshowing thof the compdegree of micals wereindicating tto break do(H2O2 + Fe(Idegradationas expectedwith solar o

nd threactxpertaineFig. 3. Absorption spectra of (Fe(II)/oxalic) solutions irradiated by art

target-compound elimination and mineralization of PAa direct photolysis using single or combined solar and-A/C light were very inefcient (

J.M. Monteagudo et al. / Applied Catalysis B: Environmental 96 (2010) 486495 491

Fig. 4. Protoc assistemolar ratio an ions: [[PA] =20mgL

Taking icentral-comthe ferrioxasources: (1)offer an ecothis typeofbe used on

3.2. Artici

All the ethe degradaassisted phsolar light.compoundrst-order krespectivelyables pH, aioxalic acid okTOC-UV) twmental desiare presentoperating cobtained focentages. Sboth systemwere measuincluded in

Experimwere in goatechuic acid elimination and mineralization of PA solutions in the ferrioxalate-d remaining concentrations of peroxide and dissolved oxygen. Experimental condit

1, air-ow rate =0.8Nm3 h1. Average temperature: 30 C. Average solar power: 40Wm

nto account the results from this previous study, twoposite experimental designs were applied to optimizelate-assisted photo-Fenton processes under two lightsolar light, as this solar-activatedcatalytic systemcouldnomical and practical alternative for the destruction ofcontaminant; and (2)articialUV-A/C light,whichcouldcloudy days.

al UV-A/C and solar pilot plants

xperiments presented in this section were based ontion of protocatechuic acid solutions in a ferrioxalate-

oto-Fenton system irradiated under articial UV-A/C orIn all of these, both the disappearance of the targetand total organic carbon removal followed pseudo-inetics with respect to the PA and TOC concentrations,, as indicated above. To study the effect of the vari-

r-ow rate and initial concentrations of H2O2, Fe(II) andn the response functions (kPA-solar, kPA-UV, kTOC-solar ando central-composite designs were utilized; the experi-gnmatrix, coded and natural levels, and variable rangesed in Table 1, which also shows the selected optimalonditions for each process as well as the best resultsr kPA, kTOC and PA elimination and TOC removal per-olar power (in the solar process) and temperature (ins) were not controlled during the experiment, but theyred during the reaction and their average values werethe tting.ental results and NN predictions of these constantsod agreement, with an average error lower than 15%

in all casesfor kPA andrelated to th(UV-light pbution paraof each of tpH, air-owpower (in tUV processto the secon

The resuneural netwwas possibied variablethe studiedtration of othe PA elimprocess, thinuence o

3.2.1. ReactAs an e

degradationFenton systexperimentvariation ofiments. Asparameters%TOC remotion)was sid UV-A/C or solar photo-Fenton systems; evolution of Fe(II)/H2O2H O ] =200mgL1; [Fe(II)] = 20mgL1; [H C O ] =30mgL1; pH=4;2 2 2 2 42.

(data not shown). The equation and tting parameterskTOC are shown in Table 2. N1 and N2 are general factorse rst and the second neuron, respectively. W11 to W16rocess) and W11 to W17 (solar process) are the contri-meters to the rst neuron and represent the inuencehe variables in the process ([H2O2], [Fe(II)], [H2C2O4],rate, average temperature and average incident solar

he case of the solar light system); W21 to W26 (articial) and W21 to W27 (solar process) are the contributionsd neuron corresponding to the same variables.lts of a saliency analysis on the input variables for eachork are also shown in Table 2. From these results, it

le to deduce the effects of each parameter on the stud-s, reported as percentages. It was conrmed that overrange in the articialUV-A/Cprocess, the initial concen-xalic and pH were the most signicant factors affectingination and mineralization, respectively. In the solar

e initial concentration of H2O2 had a more signicantn both PA elimination and solutionmineralization rates.

ion analysisxample, Fig. 4 shows a comparative analysis of theof PA solutions in the two ferrioxalate-assisted photo-

ems using articial UV-A/C and solar light sources. Theal conditions used are shown in Fig. 4, although thethe studied parameters was reproducible in all exper-can be seen, during the reaction the evolution of the(molar ratio of remaining Fe(II)/H2O2, %PA elimination,val, remaining H2O2 and dissolved oxygen concentra-milar in both systems. In therst stage of the reaction, in

492 J.M. Monteagudo et al. / Applied Catalysis B: Environmental 96 (2010) 486495

less than 4min (UV process) or 4Wh of accumulated solar energy(solar process; 4min in this case) elimination of approximately85% of the original PA was achieved. However, the mineralizationdegree was only between 15 and 20% for both systems, indicatingthat protoc30min (30Wtial was remattained inDuring thislowed by aH2O2 could

Fe(II) + H2H2O2 O2H2O2 + OHFe(III) + H2H2O2 +HO2

Additionirradiated fof H2O2, reconsumed b

Fe(III)[(C2O

The oxacarbon diox

C2O4 CThis red

CO2 + Fe(When fe

for every Fe

Fe(II) + H2In this i

tration rapithen increathe less signdecrease ofwith the oxwith intermother reseaconcentratireactions (E(Eq. (6)) anHO2 accordepletion aDependingtors, one ofthe reaction

C2O4 +OR + O2 C2O4 + [FO2 +H+ HO2 + HOFe(III) + HO

In contrahigh hydrogachieved anAt that time

ineraoto-Fe10mgC; Av

ingmdow

cult-udyn, butedrioxa2. Anndeancematichinwaverioxapartected.

fect of pH and oxalic acid addition

5 shows (as an example) the effect of pH and then of oxalic acid on the value of the mineralization ratent, kTOC-solar, as simulated by the NNs (using the equa-own in Table 2; operating conditions: [H2O2] =100mgL1;= 2mgL1; [H2C2O4] =60mgL1; pH 4; [PA] =20mgL1;e temperature: 25 C; average solar power =35Wm2). Therevealed that varying the pH and the addition of oxalic acidFe(II)/H2O2 system irradiated with sunlight could increasection rate or cause inhibition effects depending on the oper-onditions. Thus, if the reaction was conducted at pH valuesapproximately 4, the initial concentration of oxalic acid pos-affected the photocatalytic reaction until an optimal valueed, and this optimal value increased as the pH increased.as due to the faster generation of Fe2+ ion by photolysis ofalate and additional hydroxyl radicals produced at that opti-value near 4. Fe(II)may react by Eqs. (5) and (13) generatingOH radicals. An excess of oxalic acid cannot complex withons in solution and the light penetration through irradiatedater decreases. Above pH 44.5, the inuence of oxalic acidsitive over the studied range (between 0 and 60mgL1).atechuic acid rapidly oxidized into intermediates. Afterh in the solar process) of reaction, 100% of the PA ini-

oved. At that time, 75% and 58% TOC removals werethe solar and articial UV-A/C processes, respectively.rst stage, H2O2 concentration rapidly decreased, fol-second, slower, depletion stage. The consumption ofbe due to any of the following reactions:

O2 Fe(III) + OH + OH (5)+H2O (6) HO2 + H2O (7)

O2 Fe(II) + O2 +2H2O (8) O2 + OH + H2O (9)ally, when ferrioxalate complex [Fe(III)(C2O4)3]3 isrom the ultraviolet to the visible range in the presenceactions (10) to (13) take place, with H2O2 also beingy Eq. (13):

4)3]3 + h Fe(II) + 2C2O42 +C2O4 (10)lyl radical anion, C2O4, quickly decomposes to theide radical anion, CO2:

O2 +CO2 (11)ucing agent can produce Fe(II) via reaction (12):

III)[(C2O4)3]3 Fe(II) + CO2 +3C2O42 (12)rrioxalate is irradiated in the presence of H2O2, one OH(II) is generated:

O2 +C2O42 Fe(III) [(C2O4)3]3 +OH + OH (13)nitial stage of the reaction, the dissolved O2 concen-dly decreased until a minimum value was reached andsed until a plateau value was achieved, which indicatesicant role of O2 during the mineralization stage. Thedissolved O2 could be due to the reaction of oxygenalyl radical anion, C2O4, according to Eq. (14) andediate organoradicals by Eq. (15), in agreement with

rchers [18,19]. After this period, the dissolved oxygenon increased, likely due to ferrioxalate photochemistryqs. (10), (16), (17) and (18)), by H2O2 decompositiond by catalytic reactions of Fe(III) with either H2O2 ording to Eqs. (8) and (19), respectively. Both oxygennd formation reactions occur throughout the process.on the availability of the reagents and some other fac-the reactions becomes dominant at different stages of.

2 O2 +2CO2 (14)RO2 + H2O ROH + HO2 (15)e(C2O4)3]3 Fe(II) + 3C2O42 +2CO2 (16)

HO2 (17)

2 H2O2 +O2 (18)

2 Fe(II) + O2 +H+ (19)

st, the Fe(II)/H2O2 molar ratio increased because of theen peroxide consumption until a maximum value wasd then dropped until a minimum value was reached., the dissolved oxygen concentration began to increase.

Fig. 5. Msolar Ph[Fe(II)] =ture: 25

Regardslowedof difther streactioattempthe ferof H2Orapid uabsorbthe fora quenin thethe fermajornot affhigher

3.3. Ef

Fig.additioconstation sh[Fe(II)]averagresultsto thethe reaating cbelowitivelywas usThis wferrioxmal pHmore

ferric iwastewwas polization of protocatechuic acid solutions in the ferrioxalate-assistednton process; effects of pH and oxalic acid dose. [H2O2] =200mgL1;L1; [PA] =20mgL1; air-ow rate =0.8Nm3 h1. Average tempera-erage solar power: 35Wm2.

ineralization of the solutions, the TOCdegradation raten at the end of the reaction, indicating the formation

to-degrade intermediates. It would be of interest to fur-the character of such by-products formed during thet it was outside the scope of this work, which only

to demonstrate the mineralization of PA solutions inlate-catalyzed articial UV/solar system in thepresenceinteresting point is that the mineralization was more

r solar irradiation. This could be due to the signicantof PA between 280 and 310nm (Fig. 1b) and, therefore,on of its possible reaction intermediates as well. Thus,g effect could be produced by PA and its intermediateslength range of the UV-A/C system (200400nm). Aslate photochemistry makes use of light up to 550nm, aof the solar radiation capable of inducing reaction wasby this quenching effect and the TOC removal rate was

J.M. Monteagudo et al. / Applied Catalysis B: Environmental 96 (2010) 486495 493

Fig. 6. Effect ocentration, ansolutions in tconditions: [Hrate: 0.8Nm3

solar power: 3

When thepas the coagof Fe(II) to d

Fig. 6 shconcentratition of PA sthe absencesion of Fe(IIferrioxalateoxygenplayin the presef the presence of oxalic acid on remaining Fe(II), dissolved O2 con-d H2O2 conversion during the degradation of protocatechuic acidhe ferrioxalate-assisted solar photo-Fenton process. Experimental2O2] = 200mgL1; [Fe(II)] = 10mgL1; [H2C2O4] =60mgL1; air-owh1; pH=4; [PA] =20mgL1; Average temperature: 30 C; Average1Wm2.

Hwas greater than4.5, theprocess efciencydecreased,ulation of Fe(III) complexes reduces the catalyst effectecompose H2O2, as we reported previously [7].

ows the evolution of remaining Fe(II), dissolved oxygenon and H2O2 conversion during the degradation reac-olutions in the solar photo-Fenton process at pH 4 inor presence of 60mgL1 oxalic acid. Here, the conver-

) and H2O2 was higher when oxalic acid was added, thephotochemistry beingmore active. In these conditions,ed themost important role in thedegradation reactionsnce of ferrioxalate, as indicated above.

Fig. 7. Minerasolar Photo-F[H2C2O4] =30temperature:

3.4. Effect o

Fig. 7 shof PA soluticess. At a cowith initialoptimal coindicates thof H2O2 to fgen peroxidIt is well kna higher amperoxide reavailable another radicis much smdecompositis also favoconcentratisition reactthe initialthe initialincrease ofeffectwas pPA molecul

Fe(II) + OFe(II) + HOFe(II) + O2

3.5. Degrad

Several aried out tomineralizatation. In thcan be gen(HO2), singThus, the dethe presenclization of protocatechuic acid solutions in the ferrioxalate-assistedenton process; effects of initial concentrations of Fe(II) and H2O2.mgL1; pH=4; [PA] =20mgL1; air-ow rate =0.8Nm3 h1; Average25 C; Average solar power: 35Wm2.

f the initial concentrations of Fe(II) and H2O2

ows (as an example) the mineralization rate constantsons in the ferrioxalate-assisted solar photo-Fenton pro-nstant initial ferrous concentration, kTOC-solar increasedH2O2 concentration over the studied range until an

ncentration of hydrogen peroxide was used. This factat Fe2+ can signicantly accelerate the decompositionorm more hydroxyl radicals. Above this optimal hydro-e concentration, a decrease of kTOC-solar was observed.own that an increase in H2O2 concentration producesount of OH radicals. However, an excess of hydrogenduces catalytic activity, reducing the amount of radicalsd producing the well-known scavenger effect. Althoughals (e.g., HO2) are produced, their oxidation potentialaller than that of the hydroxyl radicals. Additionally,ion of hydrogen peroxide to form water and oxygenred. This optimal value of H2O2 increased as the initialon of Fe(II) increased. Thehydrogenperoxide decompo-

ion and ferrioxalate photochemistry are favored whenFe(II) concentration is increased, as expected. Whenconcentration of H2O2 was smaller than optimal, anFe(II) negatively affected kTOC because an inhibitionresent, possibly due to the excess Fe(II) competingwithes for various radicals, as shown in Eqs. (20) to (22):

H Fe(III) + OH (20)

2 Fe(III)(HO2)2+ (21)

+H+ Fe(III)(HO2)2+ (22)

ation mechanism

dditional different comparative experiments were car-determine the main active species responsible for theion of PA solutions depending of the nature of the radi-ese reactions, various oxidative intermediate specieserated such as hydroxyl radical (OH), hydroperoxyllet oxygen (1O2) and superoxide radical anion (O2).gradation process was developed in the absence and ine of appropriate quenchers of these species [20] under

494 J.M. Monteagudo et al. / Applied Catalysis B: Environmental 96 (2010) 486495

Fig. 8. Minerassisted UV-Aof 2mM scaconditions: [H[PA] =20mgL

the same coing: 1,4-benanion), sodmay also inaquenchercialUV-A/C%TOC remo85%, indicatdiate specieandKI indicdiate specie97 to 56%. Sabsorb phoan importanremoval duengedbyNaof superoxioxygen andphotochem

CO2 +O2In this c

acid pH, an

Relatifree-ns: [HmgL

ilibrin HO

Fe(

eveon soFig. 9.ence ofconditio[PA] =20

at equbetwee

HO2 +

Howradiatialization of protocatechuic acid solutions in the ferrioxalate-/C (a) and solar (b) photo-Fenton process in the presence

venging agents 1,4-benzoquinone, KI and NaN3. Experimental2O2] = 100mgL1; [Fe(II)] = 2mgL1; [H2C2O4] =60mgL1; pH=4;1. (a) Articial UV-A/C process; (b) solar process.

nditions. The scavenging agents used were the follow-zoquinone (C6H4O2, a quencher of superoxide radical

ium azide (NaN3, a quencher of singlet oxygen whichteract with hydroxyl radicals) and potassium iodide (KI,of hydroxyl radicals). Fig. 8a illustrates that, under arti-light, the additionof 1,4-benzoquinone slightly affectedval, reducing the degree of mineralization from 97 toing that O2 was not an important oxidative interme-s. The values of %TOC removal in the presence of NaN3ated that OHradicalswere themainoxidative interme-s, with KI reducing the degree of mineralization frominglet oxygen can be formed when oxygen molecules

tons and are activated [21]. Singlet oxygen also playedt role; the addition of NaN3 reducing the degree of TOC

e to 1O2 from 56 to 21%. When OH and 1O2 were scav-N3, themineralization of PA solutionswas by the actionde radicals (O2) produced in the reaction betweenthe carbon dioxide radical anion formed by ferrioxalateistry:

O2 +CO2 (23)

ase, the degradation efciency was lower because, atamount of superoxide radicals is transformed into HO2

hydroxyl raalization ofto its highealate photoconcentrati

Fig. 9 illu(in the pressumed. Asto the hydroindependenIron, in thisaccording treaction (Eqation of redas previousfree-radicaland H2O2.

4. Conclus

Protocat(9697% TOA/C or solaoperating cpresence oxsystemthanformedat2in the solartion effectspositively awhen H2O2act ashydrothe regenerof extra hyoxidative spand OH radcess. ArticCPC on clouon between %TOC removal and H2O2 consumed in the pres-radical-scavenging agents, articial UV-A/C system. Experimental2O2] = 100mgL1; [Fe(II)] = 2mgL1; [H2C2O4] =60mgL1; pH=4;1.

um according to Eq. (17), which is favored by reaction2 and ferrioxalate according to Eq. (24):

III)(C2O4)33 Fe(II) + H+ +O2 +3C2O42 (24)

r, the role of these species was different with the solarurce, as shown in Fig. 8b. In this case, superoxide anddicals were the main species contributing to the miner-PA solutions. The most signicant role of O2 was duer generation by Eq. (23), as, with solar light, the ferriox-chemistry is more active because the generated CO2

on is higher.strates the relation between the %TOC removal reachedence of radical quenchers) and the amount of H2O2 con-seen here, the mineralization degree was proportionalgen peroxide consumed during the reaction, and it wast of the type of radicals present in the reactionmedium.system, is cycledbetween the+2and+3oxidation stateso Eqs. (10) to (13). Fe(III) can be generated by the Fenton. (5)). However, the oxidation of PA involves the gener-ucing species capable of transforming Fe(III) into Fe(II),ly reported [11], so Fe(II) is not depleted, and differentproduction is limited only by the availability of light

ions

echuic acid solutions can be efciently photodegraded

C removal) using the ferrioxalate-assisted articial UV-r photo-Fenton systems, but with different optimumonditions. When articial UV light was used in thealic acid, the degradation rate was higher in the UV-Cin theUV-Asystemas ferrioxalate complexeswererst

00280nm.ThepHandoxalic acid-additionparametersprocess could increase the reaction rate or cause inhibi-depending on the operating conditions. H2O2 additionffected the rate up to an optimal concentration, butwas in excess the degradation rate decreased, as it mayxyl radical scavenger. TheoptimalpHwasnear4,whereation of Fe(II) by ferrioxalate photolysis and generationdroxyl radicals was faster. OH radicals were the mainecies in the articial UV-A/C process while superoxideicals played the most signicant roles in the solar pro-ial UV-A/C light can be used as an alternative to solardy days.

J.M. Monteagudo et al. / Applied Catalysis B: Environmental 96 (2010) 486495 495

Acknowledgements

Financial support from the Consejera de Educacin y Cienciaof the Junta de Comunidades de Castilla-La Mancha (PCI08-0047-4810) is gratefully acknowledged.

References

[1] M. Rodrguez, S. Malato, C. Pulgarn, S. Contreras, D. Curc, J. Gimnez, S. Esplu-gas, Solar Energy 79 (2005) 360368.

[2] R.F.P. Nogueira, M.R.A. Silva, A.G. Trov, Solar Energy 79 (2005) 384392.[3] A. Durn, J.M. Monteagudo, E. Amores, Appl. Catal. B: Environ. 80 (2008) 42

50.[4] S. Malato, J. Blanco, J. Cceres, A.R. Fernndez-Alba, A. Agera, A. Rodrguez,

Catal. Today 76 (2002) 209220.[5] H. Yao-Hui, T. Shu-Ting, H. Yi-Fong, C. Chuh-Yung, J. Hazard. Mater. 140 (2007)

382388.[6] C. Yong, W. Feng, L. Yixin, D. Nansheng, B. Nikolai, G. Evgeni, J. Hazard. Mater.

148 (2007) 360365.[7] J.M. Monteagudo, A. Durn, I. San Martin, M. Aguirre, Appl. Catal. B: Environ.

89 (2009) 510518.

[8] Y. Lee, J. Jeong, C. Lee, S. Kim, J. Yoon, Chemosphere 51 (2003) 901912.[9] J. Jeong, J. Yoon, Water Res. 38 (2004) 35313540.

[10] Y. Chen, F. Wu, Y. Lin, N. Deng, N. Bazhin, E. Glebov, J. Hazard. Mater. 148 (2007)360365.

[11] F.J. Rivas, J. Frades, M.A. Alonso, C. Montoya, J.M. Monteagudo, J. Agric. FoodChem. 53 (2005) 1009710104.

[12] W. Gernjak, T. Krutzler, A. Glaser, S. Malato, J. Caceres, R. Bauer, A.R. Fernndez-Alba, Chemosphere 50 (2003) 7178.

[13] F.J. Benitez, J. Beltran-Heredia, J.L. Acero, T. Gonzlez, Water Res. 30 (1996)15971604.

[14] G.E.P. Box, W.G. Hunter, J.S. Hunter, Statistics for Experimenters: An Intro-duction to Design, Data Analysis and Model Building, Ed. Wiley, New York,1978.

[15] D.P. Morgan, C.L. Scoeld, Neural Networks and Speech Processing, KluwverAcademic Publishers, London, 1991.

[16] R. Nath, B. Rajagopalan, R. Ryker, Comput. Oper. Res. 24 (1997) 767773.[17] D.W. Marquardt, J. Soc. Ind. Appl. Math. 11 (1963) 431441.[18] K.A. Hislop, J.R. Bolton, Environ. Sci. Technol. 33 (1999) 31193126.[19] A. Aris, P.N. Sharrat, Environ. Technol. 27 (2006) 11531161.[20] W. Li, S. Zhao, B. Qi, Y. Dua, X. Wang, M. Huo, Appl. Catal. B: Environ. 92 (2009)

333340.[21] S. Malhotra, M. Pandit, J.C. Kapoor, D.K. Tyagi, J. Chem. Technol. Biotechnol. 80

(2005) 1319.

Effect of light source on the catalytic degradation of protocatechuic acid in a ferrioxalate-assisted photo-Fenton processIntroductionExperimentalMaterialsPhotochemical reactionsSolar-CPC pilot plantArtificial UV-A/C pilot plant

AnalysisExperimental designNeural network strategy

Results and discussionPrevious studiesArtificial UV-A/C and solar pilot plantsReaction analysis

Effect of pH and oxalic acid additionEffect of the initial concentrations of Fe(II) and H2O2Degradation mechanism

ConclusionsAcknowledgementsReferences