A flow system for hydrogen peroxide production at reticulated

ORIGINAL PAPERA flow system for hydrogen peroxide production at reticulated vitreous carbon via electroreduction of oxygenQian Li&Christopher Batchelor-McAuley&Nathan wrence&Robert S.Hartshorne&Charles J.V.Jones&Richard ptonReceived:5July2013/Revised:30August2013/Accepted:2September2013#Springer-Verlag Berlin Heidelberg2013Abstract In this work,a reticulated vitreous carbon electrode (RVCE,96.5%porosity,24cm−1)was modified with2-anthraquinonyl groups to electrocatalytically reduce dissolved oxygen in neutral aqueous solution(0.1M phosphate buffer solution supported with3M potassium chloride,pH of6.7)to hydrogen peroxide(H2O2)at25°C under atmospheric pres-sure.The obtained current density was ca.3mA cm−2.For the first time,the oxygen reduction was investigated on a novelly designed RVCE housed in a gravity-feed flow system.Frac-tional current conversions obtained on the RVC flow cell were compared and contrasted with those on a two-dimensional electrode,viz.a tubular flow electrode.The modified-on catalyst has the benefit in terms of easy separation of the product from the catalyst.The in situ generated low concen-tration of H2O2provides potential applications to water puri-fication processes and disinfection for water and food. Keywords Electrosynthesis of hydrogen peroxide. Reticulated vitreous carbon.Electrochemical oxygen reduction.Hydrodynamic flow cell.Anthraquinonyl surface modification List of symbols and termsSymbol/term Definition UnitA Electrode surface area cm2C Analyte concentration mol dm−3 C d Capacitance Ff Fractional current conversion–F Faraday constant96,485Cmol−1Δh Height difference of upper and lowerreservoirscmiR Solution ohmic drop Vl Length of electrode cmI lim Limiting current AI max I max=FCV f An Number of electrons transferred–Q Charge passed Cr Radius cmV V olume of voids within a cylindricalcarbon foam of length of0.64cm andradius of0.3cmcm3V f V olume flow rate cm3s−1υScan rate V s−1ΓAQ Surface coverage of anthraquinonylgroupsmol cm−2 AQ Anthraquinonyl group–GCE Glassy carbon electrode–LSV Linear sweep voltammetry–PEEK Polyether ether ketone–RVC(E)Reticulated vitreous carbon electrode(within static solution)–RVC flow cell Reticulated vitreous carbon flow cellTGC Tubular glassy carbon–Q.Li:C.Batchelor-McAuley:C.J.V.Jones:pton(*)Department of Chemistry,Physical and Theoretical ChemistryLaboratory,Oxford University,South Parks Road,Oxford OX13QZ,UKe-mail:pton@wrence:R.S.HartshorneSchlumberger Cambridge Research,High Cross,Madingley Road,Cambridge CB30EL,UKJ Solid State ElectrochemDOI10.1007/s10008-013-2250-9Fractional current conversion(f)I limI max–Percentage conversion of reactantCharge passedMaximum generated chargeÂ100%–IntroductionThe industrial production of hydrogen peroxide(H2O2)has received continuous interest due to its increasing demand each year[1].According to a recent report from Global Industry Analyst,the annual capacity of H2O2is approaching almost five million metric tons in2017.Mass-produced H2O2is primarily used in the paper-and pulp-bleaching industries[2]. The most popular method to synthesise H2O2is via a chemical pathway,viz.the anthraquinone auto-oxidation(AO)process [1,3].This process involves the catalytic reduction of oxygen by anthrahydroquinone to H2O2and anthraquinone,which then returns to the catalytic cycle.Other commercial methods to reduce oxygen to H2O2include the oxidation of isopropyl alcohol and secondary alcohols in general by the Shell Chem-ical and Lyondell Chemical,respectively.The mass-produced product is commonly consumed a long way from the source. As such,both the transportation and storage of the H2O2must be considered in evaluating the economics of the H2O2-gener-ating process.Further concerns relate to its possible explosive reactivity with a range of organic materials and inorganic compounds[3].Consequently,the at-point-of-use synthesis of H2O2via an electrochemical or other pathway is much desired. One of the well-developed electrochemical methods is the ‘Dow process’,which is an on-site alkaline peroxide generation system[4,5].However,this procedure has been limited to high pH[3].Other electrochemical attempts include the use of solid polymer electrolyte electrolysis cell[6,7],gas diffusion elec-trodes(GDE)[8,9],micro-fluidic electrochemical reactors [10],power ultrasound-assisted electrosynthesis[11],and hy-drodynamic tubular flow cell[12].However,the use of three-dimensional electrodes,such as reticulated vitreous carbon, merits particular attention.Reticulated vitreous carbon(RVC)is a disordered glassy porous carbon material,with an open-pore foam structure[13, 14].Its high void volume retains infused materials within con-trolled pore sizes.The distinct characteristics of large surface area and high porosity of the RVC material counteract the limitations of the low space velocity obtained in electrochemical processes with two-dimensional electrodes[14].The use of reticulated vitreous carbon electrode(RVCE)to electrochemi-cally reduce oxygen to H2O2has been studied on a trickle bed cell[15],in an ultrasound-assisted RVC flow cell[11],and with both stationary and rotating RVCEs in alkaline and acidic solu-tions[16,17].However,the electrosynthesis of H2O2is challeng-ing,mainly due to the sluggishness of oxygen electroreduction,particularly in neutralised and acidic solutions,and the enhanced rate of further reduction of H2O2[12,16].Therefore,the use of an electrocatalyst becomes imperative.Previous work on such a process involved solution-phase2-ethylanthraquinone[18],the use of the cationic surfactant trioctylmethylammonium chloride [19,20],and oxidised RVCE[21,22].Within the present work,RVCE was modified with anthraquinonyl groups,which can electrocatalytically reduce oxygen to H2O2.The modified-on catalyst has the benefit in terms of easy separation from the product solution.The current work,for the first time,investigates the oxygen reduction on a novelly designed RVCE,housed in a gravity-feed flow system. The dissolved oxygen in neutral aqueous solution(pH of6.7) was continuously reduced under the applied potentials.Fraction-al current conversions obtained on a RVC flow cell were com-pared and contrasted with those on a two-dimensional electrode, viz.a tubular electrode flow cell.The in situ generated low concentration of H2O2provides potential applications to water purification processes and disinfection for water and food. Experimental sectionChemical reagentsAll chemicals were of analytical grade and used without any further purification.Potassium ferrocyanide,2-aminoanthraquione,and nitrosonium tetrafluoroborate were purchased from Sigma-Aldrich(Gillingham,UK),and dichloromethane(DCM)was from Fisher Scientific(Lough-borough,UK).Anthraquinone-2-diazonium tetrafluoroborate was synthesised according to a previously reported work[23]. All aqueous solutions were prepared with deionised water of resistivity not less than18.2MΩcm at298K(Millipore UHQ,Vivendi,UK).The pH6.7phosphate-buffered solution (PBS)was composed of50mM monobasic potassium phos-phate,50mM dibasic potassium phosphate,and3M potassi-um chloride.High concentration of supporting electrolyte was added to minimise solution ohmic drop. Electrochemical measurementsA continuous flow system is shown in Fig.1a.The RVCE was housed within the flow system under gravity feed.The elec-trode cross section is shown in Fig.1b.The flow system has a three-electrode configuration.A RVCE(Goodfellow,Cam-bridge,UK,thickness of0.32cm,porosity of96.5%,and pore sizes of24cm−1)was fabricated to be the working electrode.A chemically inert polyether ether ketone(PEEK)material was used as the casing.Two pieces of RVC foam,radius(r)of 0.3cm,were stacked within the PEEK tube(Fig.1b)to make a circular cylinder with a length(l)of0.64cm.For direct com-parison,cell dimensions of a tubular glassy carbon(TGC)flowJ Solid State Electrochemcell were set to be the same as thaose of RVCE,i.e.r of 0.3cm and l of 0.64cm.Electrical contact was achieved by inserting a thin graphite rod into the PEEK body.Inside the insulating PEEK tube,the graphite rod was covered with an in-house-produced conducting araldite-based graphite powder glue,and a physical contact was made to the RVC foams.The electrode design was accomplished by capping both ends with PEEK so that the solution was able to flow through a hole of radius 0.15cm.Outside,a copper ring was attached to the graphite rod.The maximum electrical contact resistance of this in-house-fabricated RVCE was below 50Ωfor all experimental measurements.Two plastic O-rings sat in a trench at each end of the PEEK tube to provide a leak-free system.A leakless Ag/AgCl (1M KCl aqueous solution)reference electrode (eDAQ,ET072)was fitted downstream,close to the working electrode.Its potential was frequently monitored exsitu against a standard calomel reference electrode by using a high-impedance digital voltmeter (Fluke 845AB).The flow system assembly was completed by mounting a platinum mesh downstream as the counter electrode.The sequence of these three electrodes in the flow system is essential.Due to the large current signals,a significant ohmic drop,iR ,can be developed between the working and reference electrodes.Modern potentiostats are designed to recognise and partially compensate this voltage drop only when the reference elec-trode is positioned in between the working and counter elec-trodes in the potential profile [12].Fresh reaction solution was equilibrated under atmospheric pressure for 1h under room temperature so that the analyte oxygen from air can fully dissolve into the aqueous solution.The reaction solution then flowed vertically upwards through the cell in order to avoid trapped air bubbles.An upper reservoir was fixed in position and constantly supplied fresh solution via Teflon tubing to the cell.The flow rates were regulated by both glass capillaries and variation in height differences (Δh )of the upper and lower reservoirs.The glass capillaries were situated at downstream with different inner bore sizes,through which the waste solution flowed.The volume flow rate (V f )of a solution was obtained by measuring the volume of waste solu-tion collected in the lower reservoir over the recorded experi-mental time.This procedure was repeated for a combination of glass capillaries at a range of Δh for variable values of V f range.For oxygen-free conditions,nitrogen gas was bubbled into the upper reservoir for at least 30min.In order to prevent oxygen dissolution during solution transport process,a small pre-nitrogen bubble was deliberately set into the transporta-tion Teflon tubing.All experiments were conducted immedi-ately after the N 2bubble passed the flow cell.An Autolab PGSTAT20computer-controlled potentiostat (Eco Chemie,Utrecht,The Netherlands)was used to perform electrochem-ical measurements.All electrochemical experiments were conducted at atmospheric pressure under 25±0.5○C.Errors were evaluated from repeated experimental procedures for at least three times.Modification of RVCEThe chemical modification of 2-anthraquinonyl functional groups onto the RVCE surface was achieved via reaction with anthraquinone-2-diazonium tetrafluoroborate salt at open cir-cuit potential.An aqueous solution containing 5mM diazoni-um salt was injected into a sealed compartment of RVCE to fill up the entire cavity.The RVCE surface was exposed to diazonium solution for 36h under room temperature and pressure.The electrode was then rinsed with pure water to wash off any physisorbed material.The modification process has been previously proposed to occur via spontaneous elec-tron transfer from the Fermi level of the carbon substrateto(a)Reference 1 M KCl (b)PEEK tubePEEK tube= 0.64cm l Fig.1a Scheme of the hydrodynamic flow system equipped with a reticulated vitreous carbon electrode (RVCE).b RVCE cross section,where PEEK represents polyether ether ketoneJ Solid State Electrochemdiazonium salt [24].A covalently attached 2-anthraquinonyl layer resulted on the modified surface.Results and discussionA model redox couple,ferrocyanide/ferricyanide (Fe(CN)64−/Fe(CN)63−),was first studied.Cyclic voltammetry was conducted under a static aqueous solution containing 0.55mM potassium ferrocyanide (K 4[Fe(CN)6])supported with 3M KCl salt.After each scan,fresh solution was flushed through the RVCE.Figure 2shows cyclic voltammograms at variable scan rates from 20to 800mV s −1.The oxidation peak potentials range from +0.385to+0.633V accordingly.A plot of percentage conversion of the reactant as a function of scan rate is shown as an inset in Fig.2.The term can be defined as follows:Percentage conversion of reactant ¼Charge passedMaximum generated chargeÂ100%The charge passed was measured from the peak area under the oxidation waves after blank subtraction.The maximum amount of charge generated can be calculated via the follow-ing equation:Maximum generated charge ¼FCVwhere F is the Faraday constant (96,485C mol −1),C is the concentration of analyte (in moles per cubic decimetre),and V is the volume of voids (in cubic centimetre).Under the studied cell dimensions,the calculated maximum generated charge was 9.2×10−3C.It can be seen that at a scan rate of 20mV s −1,almost 30%of ferrocyanide was electro-oxidised to ferricyanide on the RVCE.However,as theexperimental time scale decreases (at high scan rates),the percentage conversion of the reactant drops sharply.Linear sweep voltammetry (LSV)was next conducted for all studies within flowing solution.The same reaction solution was transported through a gravity-feed hydrodynamic system via the RVC flow cell.Figure 3depicts the LSVs of a single electron oxidation of Fe(CN)64−to Fe(CN)63−at increasing volume flow rates (V f )from 0.02to 0.74cm 3s −1at a scan rate of 5mV s −1.The half-wave potential,E 1/2,rises as the flow rates increased from +0.353to +0.565V .The noise at higher V f likely indicates a switch to a turbulent flow regime.A parameter that determines whether flow is laminar or turbulent is the Reynolds number,Re [25–27],which for a tube can be calculated from Re ¼V f πrvwhere v is the kinematic viscosity of reaction solution under studied temperature (in square centimetres per second)and r is the pore radius (in centimetre).The pores within RVCE can be envisaged as entangled tubes of void.Assuming v is 10−2cm 2s −1(the value of pure water at 20°C),the calculated Re value at 0.74cm 3s −1is ca.1,130,which is approaching the turbulent flow limit.Moreover,local turbulence can set in before any fully developed turbulence.Hence,noise on voltammograms likely suggests a tendency towards turbulent flow.Note that in the limit of V f →0,the analyte will be completely consumed before exiting the RVC flow cell.In this limit,the current (I max ),determined by the rate at which the analyte enters the electrode,is given by the following equation:I max ¼FCV f-2.0-1.6-1.2-0.8-0.40.00.40.81.2I / m AE / V vs. Ag/AgCl (1 M KCl)Fig.2Cyclic voltammograms of potassium ferrocyanide redox signals under static aqueous solution supported by 3M KCl salt on RVCE at increasing scan rates from 20,100,200,400,600to 800mV s −1.The inset shows a plot of percentage conversion as a function of scan rate-0.250.000.250.500.751.001.251.500.00.20.40.60.81.0I / m AE / V vs. Ag/AgCl (1 M KCl)Fig.3Linear sweep voltammograms of potassium ferrocyanide oxida-tion under flowing aqueous solution supported by 3M KCl salt on RVC flow cell at increasing volume flow rates from 0.02,0.07,0.28,0.44to 0.74cm 3s −1at 5mV s −1J Solid State ElectrochemConsequently,the fractional current conversion,f,isf¼I lim I maxfor any experimentally measured limiting current(I lim).The fractional current conversions for oxidative current at increasing volume flow rates for the RVC flow cell are recorded in Table1. It can be seen that f RVC decreases dramatically with increasing V f.On the right-hand side column,a direct comparison was made with a TGC flow cell.From our previous work[12],it was shown that the I lim for a tubular electrode under laminar flow can be suitably described by the Levich equation[28]. Moreover,within the experimentally studied V f range and dimensions of electrode,the steady-state currents lie in the Levich predicted regime[12].Embracing exactly the same cell dimensions as RVCE,the f TGC can be unambiguously predict-ed as shown in Table1.It is obvious that f RVC is more than1order of magnitude higher than f TGC.Hence,as expected for same cell dimensions,the use of the porous RVCE results in far larger fractional current conversions than the TGCE.The elec-trochemical behaviour of RVCE is now further discussed, focussing on the reduction of dissolved oxygen(O2).Direct O2reduction was investigated at the RVCE under static conditions with an air-equilibrated PBS supported with 3M KCl salt.Fresh solution was flushed through after each scan.Cyclic voltammograms measured at increasing scan rates from20to800mV s−1are shown in Fig.4.It can be seen that the irreversible peak for O2reduction shifts from ca.+0.3V to ca.–0.4V(vs.Ag/AgCl–1M KCl)as the scan rate increases. The peak potential of an irreversible redox couple is dependent on both the heterogeneous rate constant and the rate of mass transport to the electrode so that the reduction peak potential moves to greater overpotentials as the scan rate increases[29]. According to the literature[12,16,17],the irreversible peak corresponds to the2-electron reduction of O2to H2O2.Hydrodynamic studies of O2direct reduction in flowing reaction solutions were then carried out.The LSVs are shown in Fig.5with increasing volume flow rates.No clear O2direct reduction was observed at any flow rate.This outcome agrees with the work reported by Alvarez-Gallegos and Pletcher on a RVC flow cell in an O2-saturated hydrochloric acid solution [17].In order to enhance the rate of O2reduction,the RVCETable1Comparison of fractional current conversion of potassium fer-rocyanide oxidation on RVC flow cell(f RVC)and TGC flow cell(f TGC)at various volume flow rates(V f)at25°C and5mVs−1.Both RVC and TGC electrodes have the same dimensions(r of0.3cm and l of0.64cm cylinder)V f/cm3s−1f RVC f TGC a0.020.240.021 0.070.090.008 0.280.040.003 0.440.030.002 0.740.020.002a The Ilimused to calculate f TGC was predicted from the Levich equation [12]I/mAE / V vs. Ag/AgCl (1 M KCl) Fig.4Cyclic voltammograms of oxygen redox signals under static air-equilibrated solution of PBS–3M KCl on RVCE at increasing scan rates from20,100to200mV s−1(a)-1.6-1.2-0.8-0.40.00.4-1.6-1.2-0.8-0.40.00.4I/mAE / V vs. Ag/AgCl (1 M KCl)(b)-1.6-1.2-0.8-0.40.00.4-2.0-1.6-1.2-0.8-0.40.00.4I/mAE / V vs. Ag/AgCl (1 M KCl)Fig.5Linear sweep voltammograms of oxygen reduction under flowing solution of air-equilibrated PBS–3M KCl on both unmodified(black) and anthraquinonyl-modified(red)RVC flow cell at increasing volume flow rates from a0.06cm3s−1to b0.3cm3s−1at5mV s−1J Solid State Electrochemsurface will be modified with a catalyst,and the corresponding electrochemical signals are discussed as follows.A modified 2-anthraquinonyl RVCE (AQ-RVCE),proce-dures as described in the ‘Experimental section ’,was studied electrochemically in a static PBS reaction solution.Figure 6shows the cyclic voltammograms of the AQ-RVCE without and with O 2at a scan rate of 25mV s −1.The surface-bound electrochemical responses in Fig.6demonstrate a 2e −−2H +reduction transfer of AQ –to AQH 2–in the absence of O 2[12].The reaction scheme involves the heterogeneous redox steps,as shown in Fig.7.The non-zero background current is mainly related to the significant ohmic drop across the system.The high supporting electrolyte concentration (3M KCl)helps to mitigate this ohmic drop distortion to a certain extent.Once the electrode has been chemically modified,the sur-face coverage of 2-anthraquinonyl groups,ΓAQ (in moles per square centimetre),can be estimated from the peak area under the redox waves (Fig.6)by using the following equation:ΓAQ ¼Q where Q is the charge passed (in coulomb),n is the number of electrons transferred (n is 2),and A is the surface area of the RVCE (in square centimetre).A can be estimated experimen-tally from blank scans in the same reaction solution without any analyte on the RVCE and a known surface area of glassy carbon electrode (GCE).The blank scan shown in Fig.8was obtained in a nitrogen-saturated PBS reaction solution,where v is the scan rate and C d is the capacitance.It is known that C d is proportional to surface area,and the RVCE is a low-volume disordered glassy porous carbon material [14].Therefore,by relating the experimentally measured capacitances of both RVCE and GCE,and the known surface area of GCE,the surface area of RVCE was estimated to be (4.3±0.5)cm 2per volume of a cylindrical block of RVCE with r being 0.3cm and l being 0.64cm.Such a value can be justified to have a close agreement with the work reported by Ponce de Leon and Pletcher,where a value of 3.7cm 2per volume of a cylindrical block of RVCE with the same cell dimensions can be estimat-ed from their work [16].Hence,the ΓAQ was calculated to be (1.2±0.1)×10−10mol cm −2.A theoretical value of 3.2×10−10mol cm −2was calculated for the maximum surface coverage of vertically aligned close-packed monolayer of-1.2-1.0-0.8-0.6-0.4-0.2-0.4-0.20.00.2I /m AE / V vs. Ag/AgCl (1 M KCl )Fig.6Redox signals of anthraquinonyl-modified RVCE under nitrogen-saturated (black line )and air-equilibrated (red line )solution of PBS –3M KCl at a scan rate of 25mV s −1H O 22O 2+Fig.7Reaction scheme of surface-bound oxygen reduction mediated by 2-anthraquinonyl groupsI / m AE / V vs. Ag/AgCl (1 M KCl)Fig.8Cyclic voltammogram of a blank scan in nitrogen-saturated PBS –3M KCl solution on RVCE at a scan rate of 100mV s −1.v is the scan rate and C d is the capacitanceTable 2Comparison of fractional current conversion of oxygen reduc-tion on AQ-RVC (f AQ-RVC )and TGC (f TGC )flow cells at various volume flow rates (V f )under air-saturated PBS –3M KCl at 25°C and 5mVs −1.Experimentally measured limiting current (I lim )on AQ-RVC flow cell is also shown V f /cm 3s −1Measured I lim /A f AQ-RVC f TGC a 0.0060.62×10−40.890.090.06 5.4×10−40.710.020.36.0×10−40.180.01aThe I lim used to calculate f TGC was predicted from the Levich equation [12]J Solid State ElectrochemAQ–groups[23].Hence,the modified AQ–layer suggests a near monolayer formation on the RVCE surface.Following the characterisation,the modified surface was then studied in an air-equilibrated solution.The electrocata-lytic activity of2-anthraquinonyl groups towards O2reduction becomes apparent,as shown in Fig.6.The large irreversible reduction wave demonstrates that the2-anthraquinonyl groups mediate O2reduction process.The operative electro-catalytic mechanism is shown in Fig.7.With the assistance of surface-bound semiquinonyl intermediate and hydroquinonyl species,O2is reduced to H2O2[30,31].The mediated O2reduction under hydrodynamic conditions was next investigated on the AQ-RVC flow cell.The LSVs are shown in Fig.5at increasing volume flow rates from0.06to 0.3cm3s−1at5mV s−1.The steady-state currents are shown,in comparison with direct O2reduction on an unmodified RVCE. In the absence of O2,no limiting currents can be observed.Note that H2O2is not active at the potentials of the steady-state currents.Consequently,the limiting currents correspond to the two-electron reduction to H2O2which is mediated by AQ–.The limiting currents were measured against the extrapolated front baseline and were recorded in Table2.It can be seen that the O2 reduction overpotentials are significantly reduced on the modi-fied surface.Knowing the concentration of dissolved O2in an air-equilibrated aqueous solution supported with3M KCl salt to be0.132mM[32],consequently,the fractional current conver-sions in the RVC flow cell were calculated and shown in Table2. The generally greater f for O2reduction relates to its higher diffusion coefficient as compared to that of K4[Fe(CN)6] (1.77×10−5cm2s−1for O2and6.3×10−6cm2s−1for K4[Fe(CN)6])[12].Similarly,as discussed earlier,the corre-sponding fractional current conversions on TGC flow cell can be theoretically predicted and were tabulated in Table2.It can be seen that at lower V f,an almost full fractional current conversion of O2to H2O2may be achieved on the RVC flow cell.In comparison,TGCE has a surface area of1.2cm2,bearing the same cell dimensions as the RVCE.Though the surface area of the RVCE is just over threefolds compared to that of the TGCE, the fractional current conversions of the former demonstrate at least1order of magnitude higher.Such an outcome is mainly due to the distinct structure of the RVC material.The porous structure enhances the mass transport of analyte towards the electrochemical interface.Moreover,the high electrolyte con-ductivity determines the potential distribution and hence the effectiveness of the three-dimensional electrodes. ConclusionsThis work has explored the electrochemical reduction of dissolved O2in a neutral solution on a novelly designed RVC flow cell.A continuous synthesis of H2O2at the point of use achieved a fractional current conversion of nearly0.9at a volume flow rate of0.006cm3s−1on an anthraquinonyl-modified RVC flow cell. The much higher fractional current conversions on RVC flow cell compared with a two-dimensional electrode,i.e.tubular glassy carbon,are mainly due to its high electrode surface area and distinct porous structure.The gravity-feed hydrodynamic system provides a promising method to produce H2O2in a continuous manner at the point of use.It is of significant importance to have provided a bench-scale prototype for the electrochemical synthesis of H2O2under continuous flow conditions. Acknowledgments We thank Schlumberger Cambridge Research Lim-ited for funding the study.References1.Hage R,Lienke A(2006)Angew Chem Int Ed45:2062.Jose S(2012)Global Industry Analyst,Global Hydrogen PeroxideMarket09March3.Campos-Martin JM,Blanco-Brieva G,Fierro JLG(2006)AngewChem Int Ed45:69624.Kinoshita K(1992)Electrochemical oxygen technology.Wiley,NewYork,pp.369,85.Oloman C(1996)Electrochemical processing for the pulp and paperindustry.The Electrochemical Consultancy,Romsey,p1436.Yamanaka I,Murayama T(2008)Angew Chem Int Ed47:19007.Murayama T,Yamanaka I(2011)J Phys Chem C115:57928.Brillas E,Sirés I,Oturan MA(2009)Chem Rev109:65709.Gyenge EL,Drillet J-F(2011)J Electrochem Soc159:F2310.Scialdone O,Galia A,Sabatino S(2013)Electrochem Commun26:4511.González-García J,Banks CE,ŠljukićB,Compton RG(2007)Ultrason Sonochem14:40512.Li Q,Henstridge MC,Batchelor-McAuley C,Lawrence NS,Hartshorne RS,Compton RG(2013)Phys Chem Chem Phys15:785413.Wang J(1981)Electrochim Acta26:172114.Friedrich JM,Ponce-de-León C,Reade GW,Walsh FC(2004)JElectroanal Chem561:20315.Davison JB,Kacsir JM,Peerce–Landers PJ,Jasinski R(1983)JElectrochem Soc130:149716.Ponce de Leon C,Pletcher D(1995)J Applied Electrochem25:30717.Alvarez-Gallegos A,Pletcher D(1998)Electrochim Acta44:85318.Huissoud A,Tissot P(1998)J Applied Electrochem28:65319.Gyenge EL,Oloman CW(2005)J Electrochem Soc152:D4220.Gyenge EL,Oloman CW(2003)J Appl Electrochem33:65521.Saleh MM,Awad MI,Ohsaka T(2007)ECS Transactions3:6722.Awad M,Saleh M,Ohsaka T(2008)J Solid State Electrochem12:25123.Li Q,Batchelor-McAuley C,Lawrence NS,Hartshorne RS,Comp-ton RG(2011)New J Chem35:246224.Lehr J,Williamson BE,Downard AJ(2011)J Phys Chem C115:6629pton RG,Unwin PR(1986)J Electroanal Chem205:126.Fisher AC,Compton RG(1991)J Phys Chem95:753827.Cooper JA,Compton RG(1998)Electroanalysis10:14128.Levich VG(1962)Physicochemical hydrodynamics.Prentice-Hall,Englewood Cliffspton RG,Banks CE(2011)Understanding voltammetry,2ndedn.Imperial College Press,London30.Li Q,Batchelor-McAuley C,Lawrence NS,Hartshorne RS,Comp-ton RG(2011)ChemPhysChem12:125531.Trevor Wilson JZ,Oloman CC,Wayner DDM(2006)Int JElectrochem Sci1:99lero F,Huang F,Graham T(2003)J Solution Chem32:473J Solid State Electrochem。