The usage of different forms of ferrate (VI) ion for amoxicillin and ciprofloxacin removal: Density functional theory based modelling of redox decomposition

Research ArticleReceived: 17 November 2014 Revised: 16 December 2014 Accepted article published: 26 December 2014 Published online in Wiley Online Library:

(wileyonlinelibrary.com) DOI 10.1002/jctb.4625

The usage of different forms of ferrate (VI) ionfor amoxicillin and ciprofloxacin removal:density functional theory based modelling ofredox decompositionSibel Bar𝒊sç𝒊,a* Feride Ulu,a Mika Sillanpääb and Anatholy Dimogloa

Abstract

BACKGROUND: Decomposition of amoxicillin (AMX) and ciprofloxacin (CIP) in aqueous suspensions by two forms of ferrate(VI) were investigated. The effect of the initial concentration of antibiotics, pH, and ferrate (VI) dosage were examined. Modelcalculations were made by the Density Functional Theory (DFT) method (RB3LYP) taking into account the environmentalparameters. LanL2DZ and 6-311G++(d, p) were taken as basic functions for the calculations. This was followed by analysis oftwo redox decomposition mechanisms of the ferrate ion, with the O2 molecule formation and electron density distribution, andthe reaction mechanism of superoxide particle formation, which participates in the AMX and CIP oxidation process.

RESULTS: Ferrate (VI) degraded CIP more effciently than AMX in both forms. Electrogenerated ferrate (VI) was more efficientthan direct use of its solid form. The removal efficiencies of CIP and AMX by electrogenerated ferrate (VI) were 80.9% and 63.7%,respectively.

CONCLUSION: This study demonstrates that ferrate (VI), with its high oxidizing capacity and coagulation effect, could beapplied to the removal of antibiotics in wastewater treatment. The results of the AMX and CIP electron structure calculationsdemonstrate that electron transfer to the molecules leads to the formation of meta-stable states and causes the molecules tofragment.© 2014 Society of Chemical Industry

Keywords: amoxicillin; ciprofloxacin; ferrate (VI); electrosynthesis; DFT method; zeta potential

INTRODUCTIONPharmaceuticals such as antibiotics, anti-inflammatories and𝛽-blockers are widely used in daily life. Recently, awareness ofthe occurrence and toxicity of pharmaceuticals in natural surfacewater and treatment plants has increased considerably.1 – 3 Thepharmaceuticals which have received particular attention areantibiotics, owing to their effects on several organisms even atlow concentrations.4,5 Most antibiotics are non-biodegradable,so they escape from conventional wastewater treatment plants.6

New alternatives to prevent water pollution therefore need tobe developed. This study employed a non-biological alternativemethod to treat selected antibiotics.

Pharmaceuticals are excreted after application either in their nat-ural form or as metabolites and reach wastewater plants, water-ways and ground water.7 Humans can therefore be considered asource of both antibiotics and antibiotic resistance genes (ARGs),which may be released into the environment through sewage sys-tems. Sewer systems collect wastewater not only from domesticbut also from industrial and hospital sources. Hospital effluents, inparticular, constitute a special category of waste that are highlyhazardous because of their infectious and toxic characteristicsand they also represent an important source of multi-resistantbacteria and antibiotics.8 Medicines are often adapted to resist

biodegradation and can therefore remain in the environment fora long time. Some pharmaceuticals have been found in drinkingwater, which is a warning sign that the current handling of pharma-ceuticals may lead to future health and environmental problems.9

The occurrence of antibiotics at environmentally relevant con-centration levels has been connected to chronic toxicity and theprevalence of resistance to antibiotics in bacterial species.10

Amoxicillin (AMX) is a semi-synthetic 𝛽-lactam antibiotic usedin human medicine to treat several diseases and also in veteri-nary practice as a growth promoter. Most of the AMX is excretedunchanged in the urine and the possibility that AMX could befound in environmental samples is fairly high. AMX has beendetected at the μg L−1 level in environmental samples such as

∗ Correspondence to: Sibel Bar𝚤sç𝚤, Gebze Institute of Technology, Environ-mental Engineering Department, 41400 Gebze, Kocaeli, Turkey. E-mail:[email protected]

a Gebze Institute of Technology, Environmental Engineering Department, 41400Gebze, Kocaeli, Turkey

b Lappeenranta University of Technology, LUT Savo Sustainable Technologies,Laboratory of Green Chemistry, Sammonkatu 12, FI-50130 Mikkeli, Finland

J Chem Technol Biotechnol (2015) www.soci.org © 2014 Society of Chemical Industry

www.soci.org S Bar𝚤sç𝚤 et al.

secondary treated effluents and surface water,11,12 while in manu-facturers’ effluents it may reach the mg L−1 level.13 For this reasonsome studies have been conducted with high concentrations14 – 16

and the mg L−1 level was also chosen here.Quinolone antibiotics are an important type of commonly used

and non-biodegradable antibiotic.17 Ciprofloxacin (CIP) is a sec-ond generation fluoroquinolone antibiotic which is also frequentlyused. It was identified among the top 10 high priority pharma-ceuticals topical for the water cycle.18 CIP has a high aqueoussolubility under various pH conditions and a high stability insoil and wastewaters.19,20 The chemical structures and propertiesof both AMX and CIP can be seen in Table S1, Supplementarymaterial.

Pharmaceutical treatment technologies including ozonation,photo-catalytic degradation and other advanced oxidation pro-cesses (AOPs) have been investigated by many researchers.21 – 34

Recently ferrate (VI) has been studied comprehensively, as it couldbe an alternative to the other AOPs for the removal of pharmaceu-tical residuals.23,35 – 38 Ferrate (VI) has many advantages due to itsdual function as an oxidant and a coagulant. Ferrate species havea reduction potential of 2.20 V under acidic conditions, which isgreater than ozone. Therefore ferrate (VI) is one of the most power-ful and multipurpose chemicals. In addition, ferrate (VI) has greenchemical properties, as it is reduced to a non-toxic by-product,Fe (III), which is a well-known coagulant in water and wastewatertreatment. Embryo toxicity tests have been applied to raw and fer-rate (VI)-treated pharmaceutical-containing wastewater effluentsand the study clearly showed that raw wastewater effluents pos-sessed toxicity to zebra fish but ferrate treated effluents had noadverse effects.36

The reactivity of ferrate (VI) with a number of organic com-pounds (X) showed second-order behaviour, i.e. first-order in totalconcentration ([Fe (VI)]tot) and first-order in total concentration ofcompound ([X]tot).

39

−d [Fe (VI)] ∕dt = kapp [Fe (VI)]tot [X]tot (1)

where kapp is the apparent second-order rate constant. The val-ues of kapp are typically determined as a function of pH withkinetic runs carried out under pseudo-order conditions, [Fe(VI)]tot = [H3FeO4

+]+ [H2FeO4]+ [HFeO4−]+ [FeO4

2−] and [X]tot isfirst order in total concentration of compound. The followingprobable reactions between Fe (VI) and target compound can bementioned:40

H3FeO+4 + X → Fe (OH)3 + product (s) (2)

H2FeO4 + X → Fe (OH)3 + product (s) (3)

HFeO−4 + X → Fe (OH)3 + product (s) (4)

FeO2−4 + X → Fe (OH)3 + product (s) (5)

Ferrate (VI) production methods are fairly important becauseof its unique properties. These methods can be classified intothree main categories: (i) wet chemical synthesis; (ii) dry chemicalsynthesis; and (iii) electrochemical synthesis. The electrochem-ical method has several advantages, such as simplicity andsafety, as there is no need for explosive chemicals. The elec-trochemical method is also the most cost effective of thesemethods. Ferrate (VI) can be synthesized electrochemicallyusing a dissolving iron anode in highly alkaline media. The

reactions involved in the process are shown in the followingequations:

Anode reaction∶ Fe + 8OH− → FeO−24 + 4H2O + 6e− (6)

Cathode reaction∶ 6H2O → 3H2 + 6OH− − 6e− (7)

Overall reactions (in NaOH media) ∶

Fe + 2OH− + 2H2O → FeO−24 + 3H2 (8)

FeO−24 + 2Na+ → Na2FeO4 (9)

Solid potassium ferrate (VI) is a black–purple powder thatremains stable in the air for days, if moisture is excluded. Ferratecan be applied as a solid, which then dissolves in solution to formits ionic components. Ferrate is considered to be unstable in ionicsolution. The ferrate dianion, FeO4

2−, has a tetrahedral structureslightly distorted in the crystal state. In aqueous solution, the ionremains monomeric, and its four oxygen atoms become equiv-alent and exchange slowly with the solvent. In acidic or neutralsolution, the ferrate ions are quickly reduced by water. pH, alka-line media concentration and temperature are the key stabilityfactors.41

The objective of this research is to assess the removal efficienciesof AMX and CIP by ferrate (VI), and to compare the efficiency ofelectrosynthesized and solid form ferrate (VI). This paper presentsnew findings on the contemporary usage of different ferrate (VI)forms for the removal of antibiotics from water.

MATERIALS AND METHODSChemicalsAmoxicillin, ciprofloxacin (≥98%, HPLC grade), potassium ferrate(K2FeO4, purity of 97%) and sodium hydroxide (pellets, anhydrous,purity of≥ 98%) were purchased from Sigma-Aldrich. The buffersused in this study for pH adjustment included: C8H5KO4-HCl solu-tion (pH 4); C8H5KO4-NaOH solution (pH 5); KH2PO4-NaOH solution(pH 6, 7 and 8); Na2B4O7.10H2O-NaOH solution (pH 9). Stock solu-tions for the treatment study (50 mg L−1 AMX and CIP) were pre-pared weekly in high quality pure water using the Millipore WaterPurification System and stored at 4 ∘C.

Electrochemical ferrate (VI) synthesisFerrate (VI) was generated by the electrochemical method ina rectangular Plexiglas reactor and used simultaneously for thetreatment procedure. NaOH was used for alkaline media with aconcentration of 20 mol L−1. Electrodes were made of high purityiron. The active total surface area of the electrodes was 81.25 cm2

and current density was 1.47 mA cm−2. Constant direct current wassupplied by a GW Instek PSP-405 Programmable DC power source.The electrolyte was stirred with a magnetic stirrer. Temperatureand pH were controlled by a WTW Inolab Multi 9310 IDS pH meter.Electrolysis duration was 1.5 h. The optimum conditions for ferrate(VI) generation were determined in our previous study, which alsocontains more information about its stability.41

Jar test procedureA series of jar test experiments were carried out in a six-unit stirrer(Stuart flocculator SW6) to evaluate the ferrate performance forAMX and CIP degradation. The experiments were conducted usingthree different ferrate (VI) doses. The electrosynthesized ferrate (VI)and water volume ratios (v/v) were taken as 10/1, 5/1 and 3/1, while

wileyonlinelibrary.com/jctb © 2014 Society of Chemical Industry J Chem Technol Biotechnol (2015)

The usage of different forms of ferrate (VI) ion www.soci.org

Figure 1. Schematic diagram of the experimental setup.

the doses for the direct use of solid ferrate (VI) were 20, 40 and67 mg L−1. These concentrations were chosen to allow comparisonwith electrosynthesized ferrate (VI), taking into account that theconcentration of the FeO4

2− produced was 190± 10 mg L−1. Ferrate(VI) was dosed to the jar tester via a peristaltic pump and then thepH was adjusted. The jar test procedure was as follows: fast mixingfor 2 min at 250 rpm; slow mixing for 20 min at 25 rpm; then sed-imentation for 120 min. Synthetic pharmaceutical solutions wereprepared with initial concentrations of 10 and 50 mg L−1 to assessthe effect of initial AMX and CIP concentrations. These concentra-tions were greater than those frequently found in surface waters,but were chosen to assess the process efficiency more accuratelyrelative to residual antibiotic and total organic carbon (TOC) deter-mination with the analytical techniques used. Besides, the removalprocess for the selected pharmaceuticals can be better evaluatedwith concentrations that correspond to those found in manufac-ture effluents in real-scale applications. A schematic diagram of theexperimental setup can be seen in Fig. 1.

Experiments were performed under different pH values (4, 5,6, 7, 8 and 9), to determine the pH dependence. The pH wasfirst adjusted to 2 and then modified to final pH after dosingferrate (VI) with the buffers mentioned above. All experimentswere conducted at room temperature.

Analytical procedureAfter sedimentation, 45 mL of the sample were taken and immedi-ately filtered through 0.45 μm cellulose acetate membrane syringefilters (VWR) to measure UV–VIS, TOC, and chemical oxygendemand (COD) parameters. The supernatant was used without fil-tering for the measurement of zeta potentials (ZP).

The equilibrium concentrations were determined via a UV–VISspectrophotometer (Perkin Elmer Lambda 45) at wavelengthsof 228 and 272 nm for AMX and CIP, respectively. The cali-bration curve was established by nine standards in the range1–50 mg L−1. The coefficients of regression (R2) were 1 for AMX,0.9972 for CIP, and the mean linear regression equations werey= 0.0237x+ 0.0088 for AMX and y= 0.0941x+ 0.1186 for CIP.Recovery studies were conducted to investigate the accuracy andprecision of the established method over a wide range. The meanrecoveries of AMX and CIP were 98.2–104.8% and 95.4–107.2%,respectively (Table S2), i.e. the established method showed goodrecoveries of both AMX and CIP. TOC measurements were per-formed by TOC analyser (TOC-VCPH, Shimadzu, Japan) calibratedwith standard potassium hydrogen phthalate solutions.

COD values were determined by the colorimetric method at440 nm with a Hach spectrophotometer (Hach Lange DR-2800).The zeta potential measurements of the particles in suspensionswere determined from their electrophoretic mobility based on theSmoluchowki model using a Malvern Zetasizer Nano-ZS (Malvern,USA). The cell was washed with deionized water and ethanolbefore taking the measurements. All measurements were carriedout at 25 ∘C. Each result was an average of three readings. Theinitial measurements of AMX and CIP can be seen in Table S3.Oxidation-reduction potential (ORP) values of the solutions weredetermined by pH-meter with an ORP probe.

RESULTS AND DISCUSSIONThe performance of electrosynthesized ferrate (VI)for oxidation of AMX and CIP

Effect of ferrate (VI) dose and initial antibiotic concentrationFor AMX samples with an initial concentration of 10 mg L−1, a max-imum of 63.7% removal efficiency was achieved with a ferrate(VI) dose of 3/1 at [AMXaq]/[FeO4

2−] volume ratio and pH 7. Elec-trosynthesized ferrate (VI) concentration in alkaline media was190± 10 mg L−1 as FeO4

2−. From Table S4 in the supplementarymaterial, the removal efficiency decreased with decreasing fer-rate (VI) dose to 52.74% and 43.98% with ferrate (VI) doses of 5/1and 10/1, respectively. When AMX concentration was increasedto 50 mg L−1, the same behaviour was observed. Yet the removalefficiency was lower when the antibiotic concentration increased.For 10 mg L−1 AMX, TOC removals of 55.68%, 48.55%, 40.22% wereachieved with ferrate (VI) doses of 3/1, 5/1 and 10/1, respectively,and 50 mg L−1 AMX achieved TOC removals of 49.81%, 45.57%,31.24% with ferrate (VI) doses of 3/1, 5/1 and 10/1, respectively.COD removal efficiencies mirrored the results obtained for TOCabatement rates. Since TOC results reflect the original oxidationstate of the chemical pollutants, it is not surprising that TOCremoval is lower than COD removal. The mean ZP values werelower before treatment. Removal of AMX from the test solution byelectrosynthesized ferrate (VI) treatment significantly decreasedthe ZP of both initial concentrations of 10 mg L−1 and 50 mg L−1

with ferrate (VI) doses of 10/1, 5/1 and 3/1 (v/v), while ZP var-ied from −2.07 to 0.0166 mV. This indicates that the solutionunderwent rapid coagulation or flocculation according to ZP testmethods (D4187-82 ASTM Zeta Potential of Colloids in Water andWastewater).

J Chem Technol Biotechnol (2015) © 2014 Society of Chemical Industry wileyonlinelibrary.com/jctb

www.soci.org S Bar𝚤sç𝚤 et al.

(a)

(c)

(b)

(d)

Figure 2. The changes in removal efficiencies according to pH with different ferrate (VI) doses for the case of electrosynthesized ferrate (VI) (a) AMX removal(b) CIP removal and zeta potential (ZP) trends according to pH for the case of electrosynthesized ferrate (VI) (c) AMX (d) CIP.

For CIP samples with 10 mg L−1 initial concentration, 80.89%removal efficiency was achieved with a ferrate dose of 3/1 (v/v).Under the same conditions, TOC removal was 69.09%, whereasCOD removal was 70.12%. Ferrate (VI) dosing of 10/1 (v/v) per-formed poorly, with only 41.27% removal. All removal efficien-cies decreased with the increase of initial CIP concentration to50 mg L−1. The maximum removal (60.01%) was obtained with aferrate dose of 3/1 (v/v). ZP values varied after degradation from−0.366 to 0.0743 mV for an initial CIP concentration of 10 mg L−1

and 0.206–1.576 mV for 50 mg L−1.CIP was more prone than AMX to degradation by electrosynthe-

sized ferrate (VI). This can be related to the chemical structures ofthe antibiotics. CIP belongs to the electron-rich organic moieties,which can be easily transformed during ferrate (VI) oxidation.23,42

In addition, AMX has a phenolic group, and ferrate (VI) may simul-taneously react with this group, together with the amines andcarboxylic acid, which results in lower removal of ferrate (VI).Besides, when Fe(OH)3 particles occur as a ferrate (VI) decompo-sition by-product, this has to be considered as the sorption prop-erty of both antibiotics. Ciprofloxacin exhibits higher sorption thanamoxicillin, and AMX has relatively low pKa values compared withCIP. So AMX dissociates more readily than CIP, causes less attractionto the sorbent and less sorption.43 These results, along with the

ZP values of the treated solutions, indicate that degradation maynot be the only mechanism for the removal of these compoundsfrom wastewater, and hence their sorption properties should alsobe considered.

pH dependenceAn investigation into the effect of pH on AMX and CIP removalwas carried out for the pH range 4–9. The experimental conditionswere AMX and CIP concentrations of 10 mg L−1 and ferrate (VI)doses of 10/1, 5/1 and 3/1 (v/v). At pH 7, for all three doses offerrate (VI), AMX removal was much higher than under other pHconditions with over 60% AMX removal for the dose of 3/1 (v/v), asseen in Fig. 2(a).

The removal efficiency increased for pH 4 to 7 then decreasedat pH 8 and 9. It should be stated, however, that the removalefficiency at pH 8 was close to that achieved at pH 7. As seen inFig. 2(b), CIP removal efficiencies with final pH in the range 4–8were between 70 and 80% and higher than at final pH 9 (60%) witha ferrate (VI) dose of 3/1 (v/v). TOC and COD removal efficienciescan be seen in Fig. S1 and the changes in ZP values also can be seenin Fig. 2(c) and (d) for AMX and CIP, respectively. Figure 2(c) and(d) clearly indicate that ZP values become less negative or closeto zero with more oxidizing conditions. In Fig. 2(d) the ZP profile

wileyonlinelibrary.com/jctb © 2014 Society of Chemical Industry J Chem Technol Biotechnol (2015)

The usage of different forms of ferrate (VI) ion www.soci.org

showed a different trend for ferrate (VI) dose of 10/1. This couldbe attributed to the species and lack of ferrate (VI) in the media.For pH 5, the dominant species were HFeO4

− and CIP+. HFeO4− with

strong oxidizing capacity degraded CIP+ but due to the lower dose,HFeO4

− was consumed and more CIP+ remained in the media. Thiscaused the ZP to move to a positive value. At pH 8 and 9, FeO4

2− wasthe dominant species and CIP was found as CIPo (∼80% and 60%at pH 8 and 9, respectively) and CIP− (∼20% and 40% at pH 8 and9, respectively). Thus FeO4

2− oxidized CIPo easily and consumed itquickly but CIP− still remained in the solution. This caused negativeZP values.

Zeta potential values showed that removal due to coagulationwas more dominant than oxidation at optimum conditions. How-ever pH values should be considered because different pH valuesshowed different results. For instance, acidic pH values showednegative ZP but the removals were not too low compared with theothers. It can be said that degradation due to oxidation was moredominant at acidic pH values. In contrast to this, at neutral pH, ZPvalues were approximately zero and removal reached the highestvalue for both antibiotics. This means that the degradation mech-anism was mainly due to coagulation.

The effect of pH on degradation efficiency is multifaceted andrelated to the ionization states of molecules and active speciesinvolved in the reaction. The effect of pH phenomenon on AMXand CIP removal might be understood by considering pKa valuesof the compounds. As for CIP, it dissociates according to Equation(5), and its pKa values are 6.09 and 8.2, respectively.23

CIP ↔ CIP − O− + H+ (10)

When the pH of the solution was>8, the major species of CIPwould be dissociated CIP-O−, which was oxidized and coagulatedwith relative difficulty.35,37 This leads to the overall decrease in CIPremoval efficiency. In contrast, when the pH was< 8 (pH< pKa), themajor species would be CIP, which was readily oxidized and coag-ulated by electrosynthesized ferrate (VI) and therefore resulted inrelatively high removal performance for all ferrate (VI) doses.

In the case of AMX, its pKa values were 2.69, 7.49 and 9.63,respectively.27 It was reported that diprotonated species of AMX(H2AMX) reacted more slowly with ferrate (VI) than monoproto-nated AMX (HAMX−)38 did, and therefore the best removal effi-ciency was gained at pH 7. According to the behaviour of thepH dependence for the AMX oxidation, the degradation ratedecreases with an increase in pH above 7.5. This trend appears tobe similar to the general pH trend for the oxidation of primary andsecondary amines.44,45 This similarity indicates that the oxidationof 𝛽-lactams by ferrate (VI) may imply amine moieties.

For better understanding of the pH effect on AMX and CIPremoval, ferrate (VI) speciation in solution with variable pH wasalso considered. There are four ferrate (VI) species in aqueous solu-tions that depend on pH: H3FeO4

+, H2FeO4, HFeO4− and FeO4

2−.Their pKa values are 1.6, 3.5 and 7.23, respectively.36 FeO4

2− is thedominant species under alkaline conditions, and HFeO4

− predom-inates under mildly acidic conditions. Ferrate (VI) has higher oxi-dation potential at low pH (2.2 V) than under alkaline conditions(0.72 V). Also, it has been found that the protonated form of fer-rate (VI) (HFeO4

−) reacts faster than the un-protonated species(FeO4

2−).46,47 As mentioned before, the coagulant properties of fer-rate (VI) should also be considered in order to understand theremoval mechanism at the point where ZP values are close to zero,which indicates rapid coagulation. It can be said that the coagulantand oxidant properties of ferrate (VI) make it efficient for AMX and

CIP removal. In the light of all these observations, the most efficientpH for AMX and CIP removal by electrosynthesized ferrate (VI) waspH 7.

The performance of the solid form of ferrate (VI) for AMXand CIP oxidationOxidizing effect of ferrate (VI) dose and initial concentrationof antibioticsThe performance of direct use of a solid form of potassium fer-rate was analysed here and compared with that of electrosynthe-sized ferrate (VI). When the initial concentration was 10 mg L−1 forAMX samples, 53.74% removal efficiency was achieved as maxi-mum value with a 67 mg L−1 ferrate (VI) dose at pH 7. As indicatedin Table S5, the removal efficiency follows a decreasing trend withdecreasing ferrate (VI) dosage. This trend is similar to that for treat-ment with electrosynthesized ferrate (VI). When the AMX concen-tration was 50 mg L−1, the removal efficiency was low in all cases.The TOC removal achieved was up to 25% for 10 mg L−1 AMX. Aswith the electrosynthesized form, COD removal efficiencies werehigher than TOC removal efficiencies. Also ZP values approachedzero as removal occurred, but the values were not as close to zeroas in the case of the electrosynthesized form and varied in therange −1.2 to 5.5 mV.

In the case of CIP, a maximum of 78.26% removal efficiencywas reached for 10 mg L−1 initial concentration with a 67 mg L−1

ferrate (VI) dose. Yet almost the same efficiency was gained with40 mg L−1 ferrate (VI). At applied solid potassium ferrate dosesranging from 20–67 mg L−1, TOC abatement rates varied between42.41 and 58.63% and COD abatement rates varied between 46.78and 69.88%, respectively. As expected, the increase in appliedferrate dose has an adjuvant effect on TOC and COD removal.Yet no such difference was observed between the applied ferratedoses of 40 and 67 mg L−1. From these findings it can be concludedthat a 40 mg L−1 ferrate dose has adequate efficiency for theelimination of TOC and COD from CIP solutions. When the initialCIP concentration was 50 mg L−1, however, removal efficiencieswere different from those for applied ferrate (VI) doses. TOCremoval efficiencies reached 39%. Zeta potentials varied between−4.696 and 1.473 for two different initial concentrations and,as indicated previously, better removal efficiencies provided ZPvalues that were closer to zero.

pH dependenceWhen using all doses of solid potassium ferrate for AMX andCIP removal, the expected behaviour was observed. Removalefficiency showed an increasing tendency up to pH 7 and then theefficiency decreased at pH 8 and 9 for both compounds, as seen inFig. 3.

According to Figure S2, maximum TOC and COD removalsfor AMX were 25.15% and 43.47%, respectively, at pH 7 and a67 mg L−1 dose of ferrate (VI). For CIP, 58.63% and 69.88% efficien-cies were gained under the same conditions for TOC and CODabatement, respectively. ZP values supported this tendency asthey approached zero while removal continued to occur, as seenin Fig. 3(c) and (d).

According to the results, electrosynthesized ferrate (VI) per-formed better for AMX and CIP removal than directly used solidpotassium ferrate. The results are summarized in Table 1.

This study demonstrates that the electrosynthesized form offerrate (VI) is more efficient than the solid form for the removal ofpharmaceuticals from wastewater. The difference in performance

J Chem Technol Biotechnol (2015) © 2014 Society of Chemical Industry wileyonlinelibrary.com/jctb

www.soci.org S Bar𝚤sç𝚤 et al.

(a)

(c)

(b)

(d)

Figure 3. The changes in removal efficiencies according to pH with different ferrate (VI) doses for the case of solid potassium ferrate (a) AMX removal (b)CIP removal and zeta potential (ZP) trends according to pH for the case of solid potassium ferrate (c) AMX (d) CIP.

might be related to the stability of the former, because the stabilityincreases with more alkaline media. Besides, another factor maybe that solid potassium ferrate needs to be hydrolysed to reactwith the pollutants and it needs more reaction time, but elec-trosynthesized ferrate (VI) may react with the pollutants directly,and this makes the electrosynthesized ferrate (VI) more active.

The process of ferrate-ion synthesis is accompanied by changesin the physical–chemical properties of the solution: water passesinto the meta-stable state, possessing a high redox potential. Thus,the ordinary water redox potential was about 200 mV, but for theelectro-chemically processed water it was approximately 1100 mV.Besides, highly active oxidizing agents such as O2

•, HO2•, and

OH• are formed at the anode, while at the cathode highly activesubstances-reducers (HO2

−, OH− , O2−, H2, etc.) are formed. All this

results in a solution with high redox properties, making AMX andCIP destruction more effective in comparison with solid potassiumferrate (Table 1).

Modelling the redox decomposition of ferrate (VI) ionThe process of ferrate decomposition modelling helps us to under-stand the peculiarities of the oxidizing behaviour of the ferrateion under diverse experimental conditions. For different pH, themechanisms of ferrate decomposition differ, and this fact can havean influence on the reaction velocity and intermediate products.

That is why the reaction modelling is seen as the importantmoment in the assessment of energetically preferable processesunder varying conditions (pH, temperature, etc.).

Little research has been devoted to oxidation processes with fer-rate (VI) anion. The most recent attempts sought to study the redoxreaction mechanism used quantum-chemical calculations.48 – 50

Quantum-chemical calculations for some model systems for fer-rate ion participation have been done. It is known51,52 that in a solu-tion the ferrate ion can be represented by either FeO4

2− or bi-ferrateion [H4Fe2O7]2+.51,52 Bi-ferrate ion formation in acidic solution canfollow the scheme below:

2[H3FeO4

]+→

[H4Fe2O7

]2+ + H2O (11)

A number of publications demonstrated that an intermolecularoxo-coupling may occur in the bi-ferrate ion itself, which preservesthe ion’s tetrahedral structure:53,54

(12)

We investigated two mechanisms of ferrate ion redox decom-position with O2 molecule formation. The first is related to the

wileyonlinelibrary.com/jctb © 2014 Society of Chemical Industry J Chem Technol Biotechnol (2015)

The usage of different forms of ferrate (VI) ion www.soci.org

Table 1. Performance comparison of two forms of ferrate (VI) (exper-imental conditions: ferrate (VI) dose= 3:1 (v:v) and pH= 7 for elec-trosynthesized form; ferrate (VI) dose= 67 mg L−1 and pH= 7 for solidform)

Pharmaceuticalcompound Parameter

Ferrate(VI)electrosynthesized

Ferrate(VI)solid form

AMX (Co: 10 mg L−1) UV228removal, % 63.7 53.7TOCremoval, % 55.7 25.2CODremoval, % 59.6 43.5

CIP (Co: 10 mg L−1) UV272removal, % 80.9 78.3TOCremoval, % 69.4 58.6CODremoval, % 70.6 69.9

AMX (Co: 50 mg L−1) UV228removal, % 54.7 25.7TOCremoval, % 49.8 21.8CODremoval, % 53.5 24.6

CIP (Co: 50 mg L−1) UV272removal, % 60 47.5TOCremoval, % 49 39.1CODremoval, % 58.8 45.3

interaction of the [FeO4]2− ion with H2O molecules in a neutralsolution:

FeO2−4 + 2H2O → Fe (OH)3 + O2 + OH− (13)

The second alternative mechanism of reaction between thesame ion and water molecule in acidic solution follows the schemebelow: [

H4Fe2O7

]2+ + H2O → 2Fe (OH)3 + O2 (14)

Fe(OH)3 and oxygen molecule are the final products of thesereactions.

Model calculations using the Density Function Theory (DFT)method (RB3LYP) considered the salvation parameters in an aque-ous environment. LanL2DZ and 6-311G++(d, p) were taken as basicfunctions for the calculations.55 The vibrational frequency calcula-tions in the geometry optimization showed all stationary points asminima or transition states (TS). Intrinsic reaction coordinate (IRC)calculations were also performed for the TS.

Let us consider some steps of the process of FeO42− interac-

tion with H2O as the initial stage of the decomposition. Fron-tier orbitals (highest occupied and lowest unoccupied molecularorbitals, HOMO/LUMO) play an important role in the Redox reac-tion. They are significant in the elementary acts of reactions, andare responsible for electron redistribution in the reacting environ-ment. The energy levels for the frontier orbitals of initial state (IS),transition state (TS), and final state (FS) of the mechanism pre-sented in the reaction (13) are shown in Fig. 4.

From Fig. 5, the HOMO/LUMO energy levels for IS and TS exceedzero, which tells us about the meta-stable state of the system. Wavefunctions of the orbitals consist of d𝜋 orbitals of iron atom and porbitals of oxygen atom, and these are anti-bonding by nature. Thecharacter of the electron density distribution on the atoms causesweakening of Fe–O bonds and redistribution of the former in thesystem. Something different is observed for the frontier orbitals ofthe FS. The energy levels and electron density distribution valuesindicate the stability of the Fe(OH)3 system. The energy profile ofthe reaction under discussion is given for different states of thesystem in Fig. 5.

Analysis of the reaction barriers for different electron states ofFe atom (i.e. for different TS) tells us that the most profitable

FeO42-+2H2O

IS

TS

FS-1.0

Erel, eV

-1.19

0.98

0

1.0

(a) (b)

(c)

Figure 4. HOMO-LUMO energy level of FeO42− +2H2O system and energy

profile of this reaction’s (a) IS state, (b) TS state, (c) product of reaction -Fe(OH)3 +O2 +OH− (FS).

mechanisms of electron density transfer on the Fe atom in theredox process are the synchronous two-electron (FeVI + 2e→ FeIV)and three-electron (FeVI+ 3e→ FeIII) mechanisms.

Under two-electron transfer, the reaction barrier is 3.77 eV, whileunder three-electron transfer we have a barrierless reaction mech-anism realized with the −1.19 eV difference in energy relative to IS(Fig. 5).

It was mentioned before that in acidic solution the bi-ferrate ionstate is protonated [H4Fe2O7]2+. This gives stability to the system(see the placement of the HOMO/LUMO energy levels for IS, TS andFS in Fig. 5).

Decomposition of [H4Fe2O7]2+ occurs in the presence of a H2Omolecule, which is attached to the bi-ferrate ion, leading to theformation of hydrogen bonds. As the result, the system becomesunstable, and the further decomposition of the [H4Fe2O7 +H2O]2+

complex follows the barrierless mechanism (Fig. 5). The energyreward to TS and FS formation is −3.85 and −4.96 eV, corre-spondingly. This discussion of the ferrate ion’s redox reactionshas shown that decomposition may occur in either of the twomodels. As calculations show, the deprotonated form of the fer-rate ion, FeO4

2−(pH≈ 7), is in a meta-stable state, and this pre-defines its rapid decomposition and transition to a stable state(see Equation (3)). In contrast to this, protonated forms (pH< 7)are more stable due to ferrate ion formation. This may presentadditional obstacles to FeO4

2− decomposition and active O2 for-mation, which participates in further reactions with amoxicillin

J Chem Technol Biotechnol (2015) © 2014 Society of Chemical Industry wileyonlinelibrary.com/jctb

www.soci.org S Bar𝚤sç𝚤 et al.

(a)

(b)

Figure 5. (a) Energy profile of reactions H4Fe2O72+ +

H2O→ 2Fe(OH)3 +O2; (b) HOMO-LUMO levels of energy.

and ciprofloxacin. This conclusion agrees with the experimentalresults.

Modelling the redox decomposition of AMX and CIPmoleculesThe next phase of our research into the mechanism of AMX and CIPoxidative decomposition was to calculate their electron structureat different states (basic state and during the transfer of electronsto the AMX and CIP molecules). These states are formed underthe transfer of electrons from oxygen-containing particles as aresult of the ferrate ion decomposition. The AMX and CIP electronstructure analysis showed that redistribution of electron densityon the atoms and bonds of the molecules investigated happens asthe result of the oxidative process under the electrons belongingto the reaction environment acceptation. For both systems, themost substantial changes in the atomic charges are observed onthe oxygen and nitrogen containing molecular fragments (−C=O,−NH2). Slight changes in negative charges are found on separatecarbon atoms in aromatic rings of AMX and CIP (see Fig. S3 in theSupplementary material).

Polarization processes in the systems under investigationdecrease the covalent character of separate bonds and increasetheir ionic character. This is evidenced by the calculation ofWiberg’s indices, which give the distribution of charges on bondsfor different states of AMX and CIP (see Fig. S4 in the Supplemen-tary material). The indices analysis shows that the appearance ofan extra electron in the system leads to the weakening of mostbonds, causing the molecule’s meta-stable state and its furtherdecomposition into molecular fragments.

To evaluate the influence of electronic effects on the energy ofthe system states, we analysed the AMX and CIP energy levels. Thisis limited here to considering frontier orbitals of the molecules (i.e.a few last occupied and first free orbitals – HOMO/LUMO), as arepresented in Fig. 6.

In the neutral state (1) the energy levels of HOMO/LUMO arenegative and are evaluated as −7.05 eV for AMX and −6.09 eV forCIP (HOMO). Negative energy values are characteristic of LUMO

Figure 6. HOMO-LUMO energy levels for different states of AMX and CIP (1- basic state; 2 - state with one electron added; 3 - state with two electronsadded).

as well. The energy gap is 4.78 eV for AMX and 4.13 eV for CIP.Only one extra electron added to the systems (2) changes thepositions of the energy levels. Energetically, they are higher andthe LUMO energy value is positive. Two electrons transferred toAMX and CIP (3) make the HOMO/LUMO orbitals anti-bondingones. This fact is approved by the decrease of the Wiberg’s indexvalues on bonds under the electron acceptance on AMX and CIP.The energy gap between HOMO/LUMO orbitals also diminishesconsiderably. It becomes equal to 0.57 eV for AMX and 1.57 eVfor CIP.

The character of the electron density distribution on the frontierorbitals is shown in Fig. 7. Frontier orbitals play an important rolein the elementary acts of reactions, and they are responsible forelectron redistribution in the reacting environment.

As seen in Fig. 7, electron density on the HOMO/LUMO orbitalsof the AMX and CIP systems in the basic state (1) is distributedthrough different parts of the molecules and has a binding naturethat causes molecule stabilization. The addition of electrons tothe AMX and CIP molecules (systems 2 and 3, see Fig. 7 andS5 in the Supplementary material) results in the fact that thedensity possesses an anti-bonding character and is distributedthroughout the whole molecule. In sum, all these factors result indisproportionation of the systems investigated.

wileyonlinelibrary.com/jctb © 2014 Society of Chemical Industry J Chem Technol Biotechnol (2015)

The usage of different forms of ferrate (VI) ion www.soci.org

Figure 7. HOMO/LUMO orbitals for AMX and CIP (1 - basic state; 3 - state with two electrons added).

CONCLUSIONSAMX and CIP removal was investigated using ferrate (VI) in twoforms – electrosynthesized and solid. Electrosynthesized ferrate(VI) demonstrated better treatment performance for both antibi-otics. CIP removal up to 81% and 64% AMX removal were achievedwith a relatively low ferrate dose, although the initial concentrationof antibiotics was 10 mg L−1. The results increase our understand-ing of the use of ferrate (VI) forms for pharmaceutical removal.They indicate that the use of the electrosynthesized form of fer-rate (VI) is more efficient for AMX and CIP degradation than directuse of its solid form. The pH dependence trend suggests that bothforms of ferrate (VI) were more effective with neutral pH. Further-more, the zeta potential values of the treated solutions indicatethat the coagulant properties of ferrate (VI) are as important asits oxidation effect. According to the theoretical model of ferrate(VI) redox decomposition, the deprotonated form of the ferrateion, FeO4

2−(pH≈ 7), is in a meta-stable state, which predefines itsrapid decomposition and transition to the stable state. This studydemonstrated that ferrate (VI), with its high oxidizing capacity andcoagulation effect, could be applied effectively to oxidize antibi-otics in wastewater.

Quantum-chemistry calculations for some model systems pro-vide theoretical insight into the mechanism of redox reactionwith ferrate ion participation. The AMX and CIP electron structurecalculations indicate that in redox reactions, electron transfer tothe molecules leads to the appearance of meta-stable states andenables fragmentation of the molecules.

ACKNOWLEDGEMENTThe authors gratefully acknowledge the financial support of TheScientific and Technological Research Council of Turkey (TUBITAK).

REFERENCES1 Ternes TA, Joss A and Siegrist H, Peer reviewed: scrutinizing pharma-

ceuticals and personal care products in wastewater treatment. Env-iron Sci Technol 38:392A–399A (2004).

2 Mompelat S, Le Bot B and Thomas O, Occurrence and fate of pharma-ceutical products and by-products, from resource to drinking water.Environ Int 35:803–814 (2009).

3 Liebig M, Moltmann J and Knacker T, Evaluation of measured andpredicted environmental concentrations of selected human phar-maceuticals and personal care products. Environ Sci Pollut Res13:110–119 (2006).

4 Andreozzi R, Caprio V, Ciniglia C, de Champdoré M, Lo Giudice R,Marotta R et al., Antibiotics in the environment: occurrence in ItalianSTPs, fate, and preliminary assessment on algal toxicity of amoxi-cillin. Environ Sci Technol 38:6832–6838 (2004).

5 Foti M, Giacopello C, Bottari T, Fisichella V, Rinaldo D and MamminaC, Antibiotic resistance of gram negatives isolates from loggerheadsea turtles (Caretta-caretta) in the central Mediterranean sea. MarinePollut Bull 58:1363–1366 (2009).

6 Kümmerer K, The presence of pharmaceuticals in the environment dueto human use – present knowledge and future challenges. J EnvironManage 90:2354–2366 (2009).

7 Holm JV, Ruegge K, Bjerg PL and Christensen TH, Occurrence and distri-bution of pharmaceutical organic compounds in the groundwaterdowngradient of a landfill (Grindsted, Denmark). Environ Sci Technol29:1415–1420 (1995).

8 Verlicchi P, Galletti A, Petrovic M and Barceló D, Hospital effluents as asource of emerging pollutants: an overview of micropollutants andsustainable treatment options. J Hydrol 389:416–428 (2010).

9 Michael I, Rizzo L, McArdell C, Manaia C, Merlin C, Schwartz T et al.,Urban wastewater treatment plants as hotspots for the release ofantibiotics in the environment: a review. Water Res 47:957–995(2013).

10 Biswal BK, Mazza A, Masson L, Gehr R and Frigon D, Impact of wastewa-ter treatment processes on antimicrobial resistance genes and theirco-occurrence with virulence genes in Escherichia coli. Water Res50:245–253 (2014).

11 Kasprzyk-Hordern B, Dinsdale RM and Guwy AJ, The occurrenceof pharmaceuticals, personal care products, endocrine disruptorsand illicit drugs in surface water in South Wales, UK. Water Res42:3498–3518 (2008).

12 Schreiber F and Szewzyk U, Environmentally relevant concentra-tions of pharmaceuticals influence the initial adhesion of bacteria.Aquatic Toxicol 87:227–233 (2008).

13 Alaton IA, Dogruel S, Baykal E and Gerone G, Combined chemicaland biological oxidation of penicillin formulation effluent. J EnvironManage 73:155–163 (2004).

14 Dimitrakopoulou D, Rethemiotaki I, Frontistis Z, Xekoukoulotakis NP,Venieri D and Mantzavinos D, Degradation, mineralization and

J Chem Technol Biotechnol (2015) © 2014 Society of Chemical Industry wileyonlinelibrary.com/jctb

www.soci.org S Bar𝚤sç𝚤 et al.

antibiotic inactivation of amoxicillin by UV-A/TiO2 photocatalysis.J Environ Manage 98:168–174 (2012).

15 Hou L, Zhang H, Wang L, Chen L, Xiong Y and Xue X, Removal ofsulfamethoxazole from aqueous solution by sono-ozonation in thepresence of a magnetic catalyst. Sep Purif Technol 117:46–52 (2013).

16 Arslan-Alaton I and Caglayan AE, Ozonation of procaine penicillinG formulation effluent Part I: process optimization and kinetics.Chemosphere 59:31–39 (2005).

17 Kagle J, Porter AW, Murdoch RW, Rivera-Cancel G and Hay AG,Biodegradation of pharmaceutical and personal care products. AdvAppl Microbiol 67:65–108 (2009).

18 Golet EM, Xifra I, Siegrist H, Alder AC and Giger W, Environmentalexposure assessment of fluoroquinolone antibacterial agents fromsewage to soil. Environ Sci Technol 37:3243–3249 (2003).

19 Yang X, Flowers RC, Weinberg HS and Singer PC, Occurrenceand removal of pharmaceuticals and personal care products(PPCPs) in an advanced wastewater reclamation plant. Water Res45:5218–5228 (2011).

20 Miege C, Choubert J, Ribeiro L, Eusèbe M and Coquery M, Fate of phar-maceuticals and personal care products in wastewater treatmentplants - conception of a database and first results. Environ Pollut157:1721–1726 (2009).

21 Kwon M, Kim S, Yoon Y, Jung Y, Hwang T-M and Kang J-W, Predictionof the removal efficiency of pharmaceuticals by a rapid spectropho-tometric method using Rhodamine B in the UV/H2O2 process. ChemEng J 236:438–447 (2014).

22 Elmolla ES and Chaudhuri M, Photocatalytic degradation of amoxi-cillin, ampicillin and cloxacillin antibiotics in aqueous solution usingUV/TiO2 and UV/H2O2/TiO2 photocatalysis. Desalination 252:46–52(2010).

23 Jiang J-Q, Zhou Z and Pahl O, Preliminary study of ciprofloxacin (cip)removal by potassium ferrate (VI). Sep Purif Technol 88:95–98 (2012).

24 Vasconcelos TG, Henriques DM, König A, Martins AF and KümmererK, Photo-degradation of the antimicrobial ciprofloxacin at highpH: identification and biodegradability assessment of the primaryby-products. Chemosphere 76:487–493 (2009).

25 Mendez-Arriaga F, Torres-Palma R, Pétrier C, Esplugas S, GimenezJ and Pulgarin C, Mineralization enhancement of a recalcitrantpharmaceutical pollutant in water by advanced oxidation hybridprocesses. Water Res 43:3984–3991 (2009).

26 De Witte B, Van Langenhove H, Demeestere K, Saerens K, De Wis-pelaere P and Dewulf J, Ciprofloxacin ozonation in hospital wastew-ater treatment plant effluent: effect of pH and H2O2. Chemosphere78:1142–1147 (2010).

27 Trovó AG, Pupo Nogueira RF, Agüera A, Fernandez-Alba AR and MalatoS, Degradation of the antibiotic amoxicillin by photo-Fentonprocess - chemical and toxicological assessment. Water Res45:1394–1402 (2011).

28 Benitez FJ, Acero JL, Real FJ, Roldan G and Casas F, Comparison ofdifferent chemical oxidation treatments for the removal of selectedpharmaceuticals in water matrices. Chem Eng J 168:1149–1156(2011).

29 Homem V, Alves A and Santos L, Microwave-assisted Fenton’s oxida-tion of amoxicillin. Chem Eng J 220:35–44 (2013).

30 González T, Domínguez JR, Palo P, Sánchez-Martín J andCuerda-Correa EM, Development and optimization of theBDD-electrochemical oxidation of the antibiotic trimethoprimin aqueous solution. Desalination 280:197–202 (2011).

31 Snyder SA, Adham S, Redding AM, Cannon FS, DeCarolis J, Oppen-heimer J et al., Role of membranes and activated carbon in theremoval of endocrine disruptors and pharmaceuticals. Desalination202:156–181 (2007).

32 Chidambara Raj C and Li Quen H, Advanced oxidation processes forwastewater treatment: optimization of UV/H2O2 process through astatistical technique. Chem Eng Sci 60:5305–5311 (2005).

33 Li W, Guo C, Su B and Xu J, Photodegradation of four fluoroquinolonecompounds by titanium dioxide under simulated solar light irradia-tion. J Chem Technol Biotechnol 87:643–650 (2012).

34 Wang G, Wu T, Li Y, Sun D, Wang Y, Huang X et al., Removal of ampicillinsodium in solution using activated carbon adsorption integratedwith H2O2 oxidation. J Chem Technol Biotechnol 87:623–628 (2012).

35 Jiang J and Zhou Z, Removal of pharmaceutical residues by Ferrate (VI).PloS One 8:e55729 (2013).

36 Jiang J-Q, Zhou Z, Patibandla S and Shu X, Pharmaceutical removalfrom wastewater by ferrate (VI) and preliminary effluent toxi-city assessments by the zebrafish embryo model. Microchem J110:239–245 (2013).

37 Sharma VK, Li X, Graham N and Doong R-a, Ferrate (VI) oxidation ofendocrine disruptors and antimicrobials in water. J Water Supply: ResTechnol - AQUA 57: (2008).

38 Sharma VK, Liu F, Tolan S, Sohn M, Kim H and Oturan MA, Oxidationof 𝛽-lactam antibiotics by ferrate (VI). Chem Eng J 221:446–451(2013).

39 Sharma VK, Ferrate (VI) and ferrate (V) oxidation of organic com-pounds: kinetics and mechanism. Coordin Chem Rev 257:495–510(2013).

40 Casbeer EM, Sharma VK, Zajickova Z and Dionysiou DD, Kinetics andmechanism of oxidation of tryptophan by ferrate (VI). Environ SciTechnol 47:4572–4580 (2013).

41 Bar𝚤sç𝚤 S, Ulu F, Sarkka H, Dimoglo A and Sillanpaa M, Electrosynthesisof Ferrate (VI) ion using high purity iron electrodes: optimizationof influencing parameters on the process and investigating itsstability. Int J Electrochem Sci 9:3099–3107 (2014).

42 Rakshit S, Sarkar D, Elzinga EJ, Punamiya P and Datta R, Mechanismsof ciprofloxacin removal by nano-sized magnetite. J Hazard Mater246:221–226 (2013).

43 Githinji LJ, Musey MK and Ankumah RO, Evaluation of the fate ofciprofloxacin and amoxicillin in domestic wastewater. Water Air SoilPollut 219:191–201 (2011).

44 Lee Y, Zimmermann SG, Kieu AT and von Gunten U, Ferrate (Fe (VI))application for municipal wastewater treatment: a novel processfor simultaneous micropollutant oxidation and phosphate removal.Environ Sci Technol 43:3831–3838 (2009).

45 Noorhasan N, Patel B and Sharma VK, Ferrate (VI) oxidation of glycineand glycylglycine: kinetics and products. Water Res 44:927–935(2010).

46 Sharma VK, Luther GW 3rd and Millero FJ, Mechanisms of oxida-tion of organosulfur compounds by ferrate(VI). Chemosphere82:1083–1089 (2011).

47 Sharma VK, Ferrate(VI) and ferrate(V) oxidation of organic compounds:kinetics and mechanism. Coordination Chem Rev 257:495–510(2013).

48 Atanasov M, Theoretical studies on the higher oxidation states of iron.Inorgan Chem 38:4942–4948 (1999).

49 Yang X and Baik M-H, cis, cis-[(bpy) 2RuVO] 2O4+ catalyzes water

oxidation formally via in situ generation of radicaloid RuIV-O. J AmChem Soc 128:7476–7485 (2006).

50 Gokhale AA, Kandoi S, Greeley JP, Mavrikakis M and Dumesic JA,Molecular-level descriptions of surface chemistry in kinetic mod-els using density functional theory. Chem Eng Sci 59:4679–4691(2004).

51 Sono M, Roach MP, Coulter ED and Dawson JH, Heme-containingoxygenases. Chem Rev 96:2841–2888 (1996).

52 Meunier B, De Visser SP and Shaik S, Mechanism of oxidation reactionscatalyzed by cytochrome P450 enzymes. Chem Rev 104:3947–3980(2004).

53 Sarma R, Angeles-Boza AM, Brinkley DW and Roth JP, Studies of theDi-iron (VI) intermediate in ferrate-dependent oxygen evolutionfrom water. J Am Chem Soc 134: 15371–15386 (2012).

54 Liu X and Wang F, Transition metal complexes that catalyze oxygenformation from water: 1979–2010. Coord Chem Rev 256:1115–1136(2012).

55 Frisch M, Trucks G, Schlegel H, Scuseria G, Robb M, Cheeseman J et al.,Gaussian 09. Gaussian Inc., Wallingford CT (2009).

wileyonlinelibrary.com/jctb © 2014 Society of Chemical Industry J Chem Technol Biotechnol (2015)