RSC Advances

COMMUNICATION

Publ

ishe

d on

04

Sept

embe

r 20

14. D

ownl

oade

d by

Tul

ane

Uni

vers

ity o

n 29

/09/

2014

19:

35:0

6.

View Article OnlineView Journal | View Issue

Diverse redox ch

State Environmental Protection Engineerin

Control in Textile Industry, College of E

Donghua University, Shanghai, 201620, Ch

Fax: +86-21-6779-2522

† Electronic supplementary informationquantum yield at 366 nm. See DOI: 10.10

Cite this: RSC Adv., 2014, 4, 44654

Received 15th July 2014Accepted 4th September 2014

DOI: 10.1039/c4ra07153k

www.rsc.org/advances

44654 | RSC Adv., 2014, 4, 44654–446

emistry of photo/ferrioxalatesystem†

Zhaohui Wang,* Dongxue Xiao and Jianshe Liu

The diverse redox processes of the photo/ferrioxalate system

(PFS) were investigated by varying the concentrations of Fe(III),

oxalate and oxygen. Photoreactivity of PFS is determined by the

prevalence of the most photolabile Fe(III) and abundance of Fe(III)

and oxalate, which is critical for the operation optimization of PFS

in wastewater treatment.

Introduction

Ferrioxalate complex, mostly known as a chemical actinometer,recently nds its photochemical application in wastewatertreatment.1–4 Various reactive intermediates, such as Fe(II) andreactive oxygen species (ROS) including O2c

�, cOH and H2O2,are generated because of the photolyzation of Fe(III)–oxalatecomplexes (photo/ferrioxalate system, PFS) (eqn (1)–(7)).5–9 Thephotoproduced cOH is signicant for the oxidative decontami-nation of organic substances in natural and engineeredsystems.1,6,9 The high treatment efficiency can be sustained withUV irradiation in the presence of an excess of oxalate.�

FeðOxÞ3�3� ���!hv �

FeðOxÞ2�� þ 2CO2

:� (1)

�FeðOxÞ3

�3� ���!hv �FeðOxÞ2

�2� þ C2O4:� (2)

CO2c� + [Fe(Ox)3]

3� / [Fe(Ox)2]2� + CO2 + C2O4

2� (3)

CO2c� + O2 / CO2 + O2c

� (4)

O2c� + H+ 4 HO2c (5)

2HO2c / H2O2 + O2 (6)

Fe(II) + H2O2 / Fe(III) + cOH + OH� (7)

g Center for Pollution Treatment and

nvironmental Science and Engineering,

ina. E-mail: zhaohuiwang@dhu.edu.cn;

(ESI) available: TOC result; estimated39/c4ra07153k

58

Although the application of PFS in water purication is well-known, the mechanism of photolysis and subsequent reactionsin PFS still remains controversial.7–9 There are two different basicideas about the mechanism of the primary photolysis reaction offerrioxalate. Excitation of ferrioxalate is followed by photo-dissolution without electron transfer between iron and oxalateligand (eqn (1)),7 or by an intramolecular ligand-to-metal chargetransfer (LMCT) process (eqn (2)).5,6,8 Despite a debatable argu-ment on the two pathways, both mechanisms will lead to theformation of Fe(II) and at least one radical anion (C2O4c

�/CO2c�).

CO2c� is a strong reducing agent (E0 ¼ �1.8 V vs. NHE) and can

react with another ferrioxalate molecule or instead reduce O2 tosuperoxide anion (O2c

�) at a near-diffusion-controlled rate (k ¼2.4 � 109 M�1 s�1).10 Oxalate can form three kinds of complexwith Fe(III), i.e. Fe(Ox)+, Fe(Ox)2

� and Fe(Ox)33�, in different

proportions depending on the pH and the Fe(III)-to-oxalate ratio.Vincze and Papp11 measured the quantum yields at 254 nm of 0,1.18 and 1.60 for Fe(Ox)+, Fe(Ox)2

� and Fe(Ox)33�, respectively.

With regards to their markedly different photoactivity, a diversephotoredox chemistry of PFS is expected with the changes in Fespeciation caused by the rapid consumption of oxalate. However,most of the related work was performed in the presence of excessof oxalate.1–7 Few investigations focus on the photochemistry ofPFS at low Fe(III)/oxalate ratio to resemble the scenario of oxalatedepletion, despite its importance in understanding the funda-mental reaction mechanism and practical operation.

In this study, the photoredox processes of PFS at differentFe(III)/oxalate ratios was investigated by determining thephotoproduction of Fe(II) and H2O2. AO7, a nonbiodegradableorganic pollutant, was selected to probe the cOH photopro-duction. A preliminary mechanism was proposed based on theresults of Fe speciation and electron spin resonance (ESR).

Experimental sectionChemicals

Iron(III) perchlorate hydrate and Acid Orange 7 (AO7: C16-H11N2O4SNa) were purchased from Aldrich. 5,5-Dimethyl-1-

This journal is © The Royal Society of Chemistry 2014

Fig. 1 (a) H2O2 and (b) Fe(II) photoproduction as a function of irradi-ation time at pH ¼ 3.0.

Communication RSC Advances

Publ

ishe

d on

04

Sept

embe

r 20

14. D

ownl

oade

d by

Tul

ane

Uni

vers

ity o

n 29

/09/

2014

19:

35:0

6.

View Article Online

pyrroline-N-oxide (DMPO) and N,N-diethyl-p-phenylenedi-amine (DPD) were obtained from Sigma Chemical Co.Horseradish peroxidase (POD) was purchased from HuameiBiological Engineering, China. Oxalic acid (Ox), sodiumhydroxide, perchloric acid, potassium phosphate monobasic(KH2PO4), sodium phosphate monobasic (NaH2PO4) and1,10-phenanthroline, were of reagent grade and used assupplied. Barnstead UltraPure water (18.3 MU cm) was usedfor all experiments.

Experimental procedures

All photochemical experiments were conducted in a 70 mLcylindrical Pyrex vessel (3.8 cm in diameter, containing 50 mLsolution) with continuous magnetic stirring. The ultravioletirradiation source was a 100WHg lamp (Toshiba SHL-100UVQ-2)as specied elsewhere.12 The UV light, which had wavelengthsbelow 290 nm, was ltered with the Pyrex glass. The reactionmixture for photolysis was freshly prepared by the dilution ofstock solutions of 0.01 M oxalic acid, 5 mM Fe(III) at pH of 1.5(HClO4). AO7 (50 mM) was used as a probe compound forhydroxyl radicals. The initial pH was adjusted with dilute HClO4

or NaOH. The pH variations during all experiments were lessthan �0.2 pH units. Deaeration, when desired, was accom-plished by argon bubbling for at least 20 min before irradiationand throughout the experiments. During each kinetic experi-ment, 1 mL (for determination of Fe(II) and H2O2) or 2 mL (fordye measurement) aliquots were sampled with a new syringeeach time for the consequent analysis, which was immediatelydisposed off.

Methods and analysis

The concentration of Fe(II) was spectrophotometrically moni-tored by a modied phenanthroline method.12,13 H2O2 deter-mination was performed using a POD/DPD method.14 The dyedecoloration was monitored by measuring the absorbance at485 nm on a Hitachi U-2900 spectrophotometer.15 The photo-degradation rates were well described by pseudo-rst-ordermodel. TOC was determined on a Tekmar Dohrmann Apollo9000 TOC analyzer. ESR spectra of DMPO spin-trapping radicalswere recorded at room temperature on a Bruker EPR ELEXSYS500 spectrometer, equipped with an in situ irradiation source (aQuanta-Ray ND : YAG laser system l ¼ 355 nm). The simula-tions of ESR spectra were obtained with using WinSim EPRsimulation soware. The spin adduct assignment was based onits distinctive hyperne splitting parameters (similarity >90%).The detailed parameters were previously reported.12 Fe specia-tion was calculated using the MEDUSA soware, a chemicalequilibrium program with a built-in thermodynamic database.16

Results and discussionFormation of H2O2 and Fe(II)

H2O2 and Fe(II) are the major long-lived photolyzed products ofFe(III)–oxalato complexes.5 Their amounts at different concen-trations of Fe(III) and oxalate (Fe(III), 100 and 200 mM; oxalate,100–1000 mM) were monitored (Fig. 1). Fig. 1a shows that the

This journal is © The Royal Society of Chemistry 2014

amount of H2O2 can only be measured at low Fe(III) concen-trations during 30 min of irradiation. It should be noted that allthe concentrations of H2O2 measured were accumulatedconcentrations, that is, these just represented the net outcomeof H2O2 formed, and decomposed by a subsequent rapid darkreaction with residual Fe(II).5 Therefore, no evidence of H2O2

accumulation at 200 mMof Fe(III) did not mean that there was noformation of H2O2, but probably indicated a rapid consumptionof H2O2 aer formation.

Fig. 1b shows that a large amount of Fe(II) was generated at200 mM of Fe(III), with the consumption of oxalate (Fig. S1†),indicating that a typical photolysis proceeded even at high Fe(III)concentration. Because Fe(II)–oxalate complex (k ¼ 3.1 � 104

M�1 s�1) reacts with H2O2 three orders of magnitude faster thanthe Fenton reaction (k ¼ 53–76 M�1 s�1),17 the newly generatedH2O2 was expected to be rapidly consumed by Fe(II). Therefore,the extremely low levels of H2O2 at high Fe(III) dosage should notbe ascribed to the poor photoreactivity of Fe(III)–oxalatecomplex, but should be partially ascribed to the rapid decom-position of H2O2.

AO7 degradation kinetics

The most important feature of PFS is the efficient generation ofhighly reactive cOH radicals.1–3 The photoproduction of cOHwasdetermined with the addition of AO7 as a cOH scavenger(kcOH+AO7 ¼ (1.10 � 0.04) � 1010 M�1 s�1).18 The reaction of AO7with CO2c

� is not considered under aerated conditions because

RSC Adv., 2014, 4, 44654–44658 | 44655

RSC Advances Communication

Publ

ishe

d on

04

Sept

embe

r 20

14. D

ownl

oade

d by

Tul

ane

Uni

vers

ity o

n 29

/09/

2014

19:

35:0

6.

View Article Online

of the rapid quenching reaction of CO2c� by molecular oxygen.10

Fig. 2 demonstrates the variation of degradation rates (k, min�1)at the different Fe(III)/oxalate concentrations. Interestingly, eachk prole shows a “V” shaped curve, with a reection point(minimum degradation rate) at 50 mM of oxalate for 100 mM ofFe(III), 100 mM of oxalate for 200 mM of Fe(III) and 20 mM ofoxalate for 400 mM of Fe(III). There were no obvious changes of kat Fe(III) concentrations below 200 mM. The degradation rate ofAO7 decreased at 400 mM of Fe(III) and high oxalate loading(>200 mM). This clearly indicates that the addition of excessreactants (i.e. Fe(III) and oxalate) would not always lead to anenhanced degradation rate. To maintain the high efficiency ofPFS, it is recommended to always keep oxalate concentrationabove the reection points in the solution for the specic Fe(III)loading.

Fe(III) speciation

The photoproduction of cOH, as reected by the photo-degradation efficiency of AO7, should be closely associated withthe photoreactivity of Fe complexes and Fe speciation. Fig. 3shows the speciation of Fe(III) as a function of total oxalateconcentration ([Ox2�]T) at pH ¼ 3.0. With the increasingconcentration of oxalate over the range of 0–1000 mM, the Fespeciation gradually changes with increasing amount of Fe(III)–oxalato complexes accompanied by decreasing amount of hex-aaquo and hydroxylated Fe(III) complexes (i.e. Fe3+, FeOH2+,FeOH2

+). Mono(oxalato) complex (Fe(Ox)+) is the predominantspecies of Fe(III)–oxalato complexes at 1 : 1 Fe(III)-to-oxalateratio, whereas Fe(Ox)2

� and Fe(Ox)33� are the major species at

Fe(III) : oxalate <1 : 2. Because Fe(Ox)2� and Fe(Ox)3

3� areconsiderably more efficiently photolyzed than Fe(Ox)+ andFeOH2+,11 the Fe(II) quantum yield, a collective value contrib-uted by each photolabile Fe complexes, should signicantly varyat different Fe(III)/oxalate concentrations.

Due to the difficulty to obtain the quantum yield of eachphotolabile species within the range of 300–400 nm,19 herequantum yields at the main emission wavelength (366 nm) ofthe Hg lamp were used to approximately evaluate the reactivity

Fig. 2 The degradation rate of AO7 as a function of oxalate concen-tration at pH ¼ 3.0. Inset: variation of k with initial Fe(III) concentrationin 50 mM AO7.

Fig. 3 Molar fraction distribution of the Fe(III)–Ox complexes atdifferent Fe(III) concentrations: (a) 100 mM; (b) 200 mM; (c) 400 mM atpH ¼ 3.0.

44656 | RSC Adv., 2014, 4, 44654–44658

under polychromatic irradiation as used in the study. Becauseof their similar spectral absorption, FeOH2

+ was assumed tohave photoreduction reactivity roughly close to that of FeOH2+

(F366 nm z 0.017),20,21 and therefore was regarded as equivalentto FeOH2+ in this study. According to the recently reportedvalues by Weller et al.,19 the Fe(II) quantum yield of Fe(Ox)2

� andFe(Ox)3

3� is 1.17 and 0.91, respectively. The total Fe(II) quantumyield (Ftot) at 366 nm can be estimated by the followingequation:

Ftot ¼Xi

Fi3i fi

,Xi

3i fi (8)

This journal is © The Royal Society of Chemistry 2014

Fig. 4 ESR spectra of DMPO radical-adducts formed during thephotolysis of Fe(III)/Ox solution under argon-purged conditions in 0.04Mof DMPO at pH¼ 3.0. (a) Ox only, 1000 mM. (b) Fe(III) only, 200 mM. (c)Simulated data of (b). (d) Fe(III), 200 mM; Ox, 100 mM. (e) Simulated dataof (d). (f) Fe(III), 200 mM; Ox, 400 mM. (g) Fe(III), 200 mM; Ox, 1000 mM. (h)Fe(III), 100 mM; Ox, 1000 mM. (i) Simulated data of (h).

Communication RSC Advances

Publ

ishe

d on

04

Sept

embe

r 20

14. D

ownl

oade

d by

Tul

ane

Uni

vers

ity o

n 29

/09/

2014

19:

35:0

6.

View Article Online

where fi is the molar fraction of a specic species i; 3i is themolar extinction coefficient of species i; Fi is the individualquantum yield of species i for Fe(II) formation; F(Fe(Ox)+) and3(Fe(Ox)+) were estimated as 0 and 753 M�1 cm�1,11,19 respec-tively. Other parameters were from ref. 19 and 20.

Fig. S2† shows the estimated Ftot prole at different oxalateconcentrations. At [Ox2�] <50 mM (for 100 mM Fe(III)), [Ox2�]<100 mM (for 200 mM Fe(III)) and [Ox2�] <200 mM (for 400 mMFe(III)), Ftot was extremely low and even lower than the controlsystem (without oxalate ligand). With a further increase in[Ox2�], the corresponding Ftot gradually increased to themaximum value (z1.1). This result well aligns with thedependence of the AO7 degradation rate on the total oxalateconcentrations, as demonstrated in Fig. 2. It implies that reac-tivity of PFS is strongly related to the Fe speciation and thephotoreactivity of the prevalent Fe species. It can also beexpected that the photodegradation efficiency of PFS wouldgreatly diminish as oxalate ligand is quickly depleted.

Free radicals involved

The photogeneration of Fe(II) may occur through two differentways: the primary photoreduction of Fe(III) complexes (eqn (1)and (2)) and the secondary reduction by the intermediateradical, such as CO2c

� (eqn (3)). The Ftot for Fe(II) generationwould be signicantly affected by the competitive reaction ofCO2c

� with O2. In contrast, the excess of Fe(III) (relative to O2)would inhibit the formation of H2O2 by quenching CO2c

�. Here,ESR with DMPO as the spin-trapping reagent was employed todetermine reactive radicals involved in the reaction systems. Toavoid the interference of dioxygen with carbon-centered radicalmeasurement, the ESR experiments were conducted underanaerobic conditions. As shown in Fig. 4, two kinds of radicalswere identied. The six-line ESR signal was assigned to DMPO–CO2c

� adduct22 (aH ¼ 18.9 G, aN ¼ 15.8 G) (Fig. 4d and h),whereas the four-line signal was known to be of DMPO–cOH(aN ¼ aH ¼ 14.8 G)23 (Fig. 4b). In the absence of oxalate,photolysis of hydroxylated Fe(III) complexes generates cOHradicals (Fig. 4b). With the addition of 100 mM oxalate, thesignal of DMPO–CO2c

� adduct appeared. According to Fig. 3b,about 40% of Fe(III) was present as photo-inert Fe(Ox)+,hydroxylated Fe(III) complexes being half of the total Fe(III).Thus, CO2c

� should not have originated from the photolysis ofFe(Ox)+, but from the attack of oxalate by photogenerated cOH(eqn (9)).24 Interestingly, a further increase of oxalate concen-tration (i.e. 400 mM and 1000 mM) led to the disappearance ofthe CO2c

� signal (Fig. 4f and g). However, CO2c� was measur-

able in the presence of 100 mM Fe(III) and 1000 mM oxalate(Fig. 4h). This signicant comparison indicates that the yield ofCO2c

� was closely related to the abundance of Fe(III). DMPOmight efficiently trap CO2c

� at a lower Fe(III) content, but it isnot the case at a higher level of Fe(III), where photoproducedCO2c

�might be promptly captured by its adjacent Fe(III)–oxalatocomplexes.

C2O42� + cOH / CO2c

� + CO2 + OH� (9)

This journal is © The Royal Society of Chemistry 2014

Conclusion

In this study, the photochemistry of Fe(III)–oxalato complexeswas investigated at varied levels of Fe(III) and oxalate. Therewas no measurable H2O2 accumulated at 200 mM of Fe(III),partly because of the decrease of H2O2 formation, whichresulted from CO2c

� capture by the excess of Fe(III) and rapidconsumption of H2O2 by photogenerated Fe(II). With theincrease of oxalate concentration, photo-inert Fe(Ox)+, mostphotoactive Fe(Ox)2

� and Fe(Ox)33� dominated the Fe(III)

speciation, and thus controlled the overall photochemicalreactivity of Fe complexes. At Fe(III) : oxalate ratio of <1 : 2,cOH photoproduction efficiency, as probed by AO7 decolor-ation rate, was considerably low compared to those with theexcess of oxalate ligand. This result is very important foroptimizing the operation of a photo/ferrioxalate system in itsapplication of water treatment.

Acknowledgements

This work was nancially supported by National NaturalScience Foundation of China (NSFC) (Grant no. 41273108 and21377023) and Fundamental Research funds for CentralUniversities Central (2232013A3-08). Z. H. W thanks Prof. JincaiZhao at ICCAS for his generous support for ESR measurements.

RSC Adv., 2014, 4, 44654–44658 | 44657

RSC Advances Communication

Publ

ishe

d on

04

Sept

embe

r 20

14. D

ownl

oade

d by

Tul

ane

Uni

vers

ity o

n 29

/09/

2014

19:

35:0

6.

View Article Online

Notes and references

1 M. E. Balmer and B. Sulzberger, Environ. Sci. Technol., 1992,26, 1014.

2 J. Jeong and J. Yoon, Water Res., 2005, 39, 2893.3 J. Jeong and J. Yoon, Water Res., 2004, 38, 3531.4 G. Liu, S. Zheng, X. Xing, Y. Li, D. Yin, Y. Ding and W. Pang,Chemosphere, 2010, 78, 402.

5 B. C. Faust and R. G. Zepp, Environ. Sci. Technol., 1993, 27,2517.

6 Y. Zuo and J. Hoigne, Environ. Sci. Technol., 1992, 26, 1014.7 J. Chen, H. Zhang, I. V. Tomov and P. M. Rentzepis, Inorg.Chem., 2008, 47, 2024.

8 I. P. Pozdnyakov, O. V. Kel, V. F. Plyusnin, V. P. Grivin andN. M. Bazhin, J. Phys. Chem. A, 2008, 112, 8316.

9 Z. Wang, C. Chen, W. Ma and J. Zhao, J. Phys. Chem. Lett.,2012, 3, 2044.

10 P. S. Surdhar, S. T. Mezyk and D. A. Armstrong, J. Phys.Chem., 1989, 93, 3360.

11 L. Vincze and S. Papp, J. Photochem., 1987, 36, 289.12 Z. Wang, X. Chen, W. Ma, C. Chen and J. Zhao, Environ. Sci.

Technol., 2010, 44, 263.

44658 | RSC Adv., 2014, 4, 44654–44658

13 H. Tamura, K. Goto, T. Yotsuyanagi and M. Nagayama,Talanta, 1974, 21, 314.

14 H. Bander, V. Sturzenegger and J. Hoigne, Water Res., 1988,22, 1109.

15 Z. Wang, R. Yuan, Y. Guo, L. Xu and J. Liu, J. Hazard. Mater.,2011, 190, 1083.

16 I. Puigdomenech, MEDUSA-Make Equilibrium DiagramsUsing Sophisticated Algorithms, 2010, http://www.kemi.kth.se/medusa/.

17 D. L. Sedlak and J. Hoigne, Atmos. Environ., 1993, 27A, 2173.18 S. Hammami, N. Bellakhal, N. Oturan, M. A. Oturan and

M. Dachraoui, Chemosphere, 2008, 73, 678.19 C. Weller, S. Horn and H. Herrmann, J. Photochem.

Photobiol., A, 2013, 255, 41.20 B. C. Faust and J. Hoigne, Atmos. Environ., 1990, 24, 79.21 H.-J. Benkelberg and P. Warneck, J. Phys. Chem., 1995, 99,

5214.22 F. A. Villamena, E. J. Locigno, A. Rockenbauer, C. M. Hadad

and J. L. Zweier, J. Phys. Chem. A, 2006, 110, 13253.23 S. J. Klebanoff, A. M. Waltersdorph, B. R. Michel and

H. Rose, J. Biol. Chem., 1989, 264, 19765.24 G. V. Buxton, C. L. Greenstock, W. P. Helman and A. B. Ross,

J. Phys. Chem. Ref. Data, 1988, 17, 513.

This journal is © The Royal Society of Chemistry 2014