Applied Catalysis: Envonmental 2002017} 493-502 Contents lists available at ScienceDirect Applied Catalysis B: Environmental journal homepage: www. elsevier.com/locatelapeath ELSEVIER Preparation and characterization of supported Ruxltj1.)02 nano-oxides using a modified polyol synthesis assisted by microwave activation for energy storage applications Dens Thomas Audichon*, Benoit Guenot™, Steve Baranton*, Marc Cretin®, Claude Lamy”, Christophe Coutanceau** Liner teers. KOMP, MR CNRS 7245, "es and Nor-cnentinel Mei roup. 4 Rue Mel Brant, TSA 51106 46073 Poters edex ARTICLE INFO ABSTRACT Bucy ‘A simple and sofe synthesis method, namely the polyol route, is proposed to synthesize RialtcOa Received 31 May2016 . ‘mixed-oxide nanopaticles deposited on a high surface afeacatbon support, with controlled size and Received in cevised for 20 ly 2015 size distribution, and composition. The samples are characterized by TGA, XRD, TEM. XPS and Raman ‘Accepted 25 Jly 2016 ‘al analyses. Their behaviors as supercapacitors and catalysts for oxygen evolution reaction (OER) have heen investigated in 05M H,S0, medium. Iewas demonstrated that pure RuO, presenced the highest super- ‘capacitive property, and that addition of iridium leads to decrease ths property. On the other hand, ane 2 july 2036 ieyert pure Rio; presented the lower onset potetal or OER, bt tended to Be passivated asthe potential was Miseldes Increased Addon of ium edt nit orto avo the catalyst passivation and wincrease the activi Cage eveladenceacton ‘thigh potential bu soto shi postively the OER onset potenti. then {© 2016 Elsevier BY. Al ight teserved Supeseapactr water electrolysis 1. Introduction ur technological civilization needs more and mare energy. par~ ticularly in emerging and developing countries. Fossil resources, ssuch as coal, natural gas and hydrocarbons are the main primary sources, but their amount is limited and will be exhausted in a few decades. Furthermore they are the main contribution to carbon dioxide emission leading to greenhouse effect, To overcome these energetic and environmental issues, renewable primary energy sources, such as hydroelectric, wind, solar and tidal powers, have known very intense developments for the last decades. But, most ‘of these energy sources are intermittent and thus energy storage solutions have to be developed in parallel, either under chemical or electrochemical forms. The use of supercapacitors isa good means for electricity storage whereas hydrogen is considered as the most effective candidate for chemical storage of the energy, particularly {og its further use in a polymer electrolyte membrane fue cell. Cur rently, 4% of hydrogen is produced by water electrolysis and 95% by steam reforming of fossil fuels | 1,2], whichis not sustainable. Water Covespending authors. E-mail edresss thomas auichen®un-posess f(T. Audehon ‘ope coutanceauuny pater (C. Couaneea} tp tcdolongt0.1016).apeath2016 07.048 (93263373/0 2016 Eisele Al ight reserved electrolysis using renewable primary energy sources is a sustain able process but it still remains an expensive method for producing, high-purity hydrogen. The reason is the high anode overpotential for the oxygen evolution reaction (OER), which leads to high cell voltage, ie. high energy consumption (3) ‘Among the many transition metal oxides, the most success has been achieved using RuO, for supercapacitors due tits advantages of a wide potential window, of highly reversible redox reactions Jong life cycle and metallic type conductivity [4]. Indeed, RuOz hhas a well-known application as Dimensionally Stable Anode in energy storage electrochemical supercapacitots [5]. RU. which exhibits low resistivity, metallic conductivity, and high thermal and chemical stabilities, is also a basic material asa corrosion-resistant electrode at low overpotentias for chlorine and oxygen evolution reactions (6,7). The overpotential of OFR in acid medium is indeed comparatively lower when Ru is used as catalyst than in the case of other metals or oxides. However. in order to increase its stability and activity towards the OER, ruthenium oxide is often mixed with iridium oxide [5], Indeed, due to their high electrochemical per~ formances, ruthenium oxide (RuO, ) and iridium oxide (Ir0} ) have been extensively investigated as anodes for 0, evolution in acidic ‘medium [8,9 Fundamental investigations and extensive researches an syn- thesis methods for the production of RuOz based catalysts have494 Audion ej Applied atass 8: Enionmental 200 (2017) 485-502 ‘been performed in order to obtain the material structure and com- position fulfilling the compromise between property requirements ‘ofthe considered application and cost. The rutile-type RO» surface properties as well as its electrochemical behavior (electrocatalytic activity, selectivity, stability) are indeed known to depend on the precursor materials, preparation conditions and pyrolysis treat- ment [10]. In a previous work, we reported a simple and rapid ‘method (the microwave assisted “Instant method) for the syn- thesis of RuO> and the way fo tune the structute/property of the rutile-type material, for either water electrolysis reaction of eapac- itor application [11 ‘The present study is devoted to a simple, effective and eco- nomic synthesis route for a large scale synthesis of RuO) and IrOz nanoparticles, which respects the following principles: time, ‘energy and atom savings, non-harmful chemical synthesis, safe and ‘environmental-friendly synthesis process. The sft microwave acti- vated polyol route under oxygen atmosphere adopted in this work for synthesizing RuxIr102 mixed oxides leads to the formation ‘of pure and nanosized compounds. The purpose of this work is to use a simple and clean synthesis procedure with special attention paid to the characterization of the obtained Ruhr; 02 particles in terms of microstructures, electrocatalytic behavior towards OERin acidic solution and capacitive properties. Fr this purpose, in order to decrease the metal loading in the electrode and to be able to characterize the own behavior of Rult; x02 mixed oxides at the ‘nanoparticle evel, the materials were dispersed on an inactive high surface area electron-conducting carbon powder. 2, Experimental 2.1. Mixed oxides synthesis Mixed oxides supported on a carbon powder (Vulean XC 72) were synthesized by a modified polyol method, Appropriate amounts of HpltClg.6420 and/or RuCl, 340. (99.9% purity, Alfa ‘Acsat) were dissolved in SOmL of ethylene glycol (putiss. pa, purity =99.5%, Fluka) with nominal metal concentrations corre sponding to Ru:lr atomic ratios of 1:0, 0.75:0.25, 05:05, 025:0.75 ‘and 0:1 in order to obtain a final oxide concentration of 1gL-1. A [NaOH solution (11M) in ethylene glycol was then added drop-wise in the mixture under Nz bubbling (Nitrogen U quality from “Air Liquide”) to adjust the pH to 11. The reactor was equipped with a single crystal of rutile type displays spectrum with three characteristic bands at 561 em), 728cm"! and 752cm~}, corresponding to Eg. Avg and Bpg modes, respectively. Based on these results, the three conti Dutions observed here are attributed to the Eg, Arg and Byg modes and the red-shift could originate from a particle size effect (15.17) Moreover, after measurements with increased laser powers, the spectrum recorded at 0.04 mW (Fig 3a Inset) reveals three peaks located at wavenumber values closer to the characteristic ones. The effect of high laser power on the crystalline structure is similar toa heat treatment, which allows detecting the characteristic bands of ‘therutile lattice of iridium oxide. Two others broad peaks appearing at 820 and 95cm“! with increasing laser power arenot attributed: they could be linked to species formed after carbon combustion induced by laser irradiation, since their presences are observed on all spectra (Fig. 3b) independently on the particle nature, Fig, 3b shows the obtained results forall Rul ¢(2/C samples with a 4mW laser power. The spectrum of Ru03/C sample displays no defined bands [14.18.19], which is observed for al laser power settings (not shown). Raman analyses do not allow defining the nature and the structure of the RuO,/C material. Conversely, all ‘materials containing iridium oxide present in their Raman spectra the three characteristic Eg, Aig and Byg bands of the rutile struc- ture. These bands undergo a red shift when increasing the ratio of ruthenium oxide in the material which could be attributed to the rutile structure of ruthenium oxide with characteristic bands at 528em-!, 64Gcm-) and 716cm-" for Eg, Ayg and Bg modes 20.21, The gradual red shift of the Ey band and the fact that no splitting occurs for the Aig and By, bands may indicate a lattice ‘modification and the formation of mixed metal oxide particles with substitution of Ir atoms in the rutile structure by Ru atoms, XPS analysis were carried out to determine the nature of the particles and che atomic metal surface composition of the Rodtg.9}02/C materials. The Ru 3d and O 1 as well as the Ir 4f and O'1 core level spectra of RuO,/C and [r03/C materials are displayed in Fig. 4. The spectrum in Fig. 4a shows the important ‘contribution of the carbon support (C1 st 284.4 eV), andthe asym- ‘metric peak of Ru at 281.3eV which is characteristic of the Ru 3d core level [22]. The XPS spectrum for l10z/C (Fig. 4c) reveals characteristic shapes of Ir Af care level [23], For both puire oxide materials, the shapes of the O 1's core level spectra (Fig 4b and d) are similar. Only one broad core level peak is observed at 530.6 and 531.2eV for RuO;/C and IrOp/C, respectively. The contribu tion of the C 1s was determined on pure Vulcan XC-72 powder, which serves as support for oxide materials, and was used for the Ru 3d core level spectrum fitting, Based on the literature [24 the Ru 3d XPS spectrum for RuO2/C was fitted with two primary spin- orbit components at 281.1 and 285.3eV for Ru 3dgp and Ru 3d, respectively. addition to the dominant spin-orbit doublets, two satellite peaks at 2828 and 287.0eV are attributed to final-state screening effects (green curves, Fig. 4a), which was explained by [Naslund et al. as being “derived from the strong Coulomb interac- tion between valence electrons and the care hole produced in the photoionization process” [24 For Ir 4f cote level, the peaks located af GIG and 64.6eV ate attributed to the +IV oxidation number of the iridium and correspond to the Afpj2 and 4fgj2 contributions, respectively. At energies 2.0eV higher than that of the spin-orbit doublets, two others broad peaks are observed and attributed to the satellites peaks induced by the final-state screening effects. For Doth pure oxide materials, the O 1 core level signals show four ‘components. The first ones at lower binding energies, 529.9 and 530.5 eV for Ru0g/C and IrOz/C, respectively, are attributed to oxy- gen bonded to the metal (0-) and are directly cortelated to the dominant spin-orbit doublets of Ru 3d and Ir 4f.The second com- ponents at $31.1 and 531.7 eV for RuOs/C and trO2/C, respectively, ‘can also be attributed to O°, but linked tothe satelite peaks of the metallic species and induced by screening effects. However, theseAudion bj ppd Carls Environmental 2002017) 495-502 Table Intensity /a.u, (b) Intensity /a.u za ho ae ee aah ee Binding Energy /eV 0 a on ao Binding Energy / eV (c) ae ie Intensity /.u ere (a) Intensity /a.u me eke Skee Binding Energy /eV Binding Energy / eV ig. 4 XPS spectra of (2) he Ru 34 ad (b} 01 for Ku, material and () I fan (@) O15 for OIC material fitted with Gaussian-Lorentalan curves on Shirley backgrounds components could also be induced by a contribution of hydroxy! ‘groups present on the oxide particles surface [25]. At as, the third and fourth components ar linked to the oxygen boncled tothe car- bon support and ate attributed to ~C- and ~C=0, respectively. The same iting procedure of XPS results was applied for the mixed metal oxides supported on Vulcan XC-72, allowing determining the atomic ratios of each element present in the oxide materials. The obtained values are summarized in Table |. For high tr com- centration in metal oxides, the ratios O1s/(Ru 34 +r 41) indicate a stoichiometry close to 2 in agreement with that expected for pure dioxide material, while for ruthenium rich materia, the stoichiom= cetty is higher than 2. This could come from a more consequent presence of hydroxyl species in Ru oxide materials (22), which tends to demonstrate thatthe formation of hydroxyl groups occurs preferentially on the ruthenium oxide particles. The Ru 3djlr Af sur- face atomic ratios were determined for the mixed oxide materials The obtained results are in accordance with the nominal values, Dut it seems that a very slight iridium surface enrichment takes place. 43.2, Electrochemical characterizations Cyclic voltammograms were recorded at 20mVs~! to deter: sine the nature of the particles synthesized and to characterize their behaviour in H;S0, supporting electrolyte in the potential range from 0.05 to 1.4V vs, RHE (Fig. 5) All currents were nor- malized to the geometric surface area of the working electrode. The voltammogram shapes of the pure oxides are characteristic of ruthenium and iridium oxides supported on carbon, respectively [25]. The CV of Rud, /C obtained is characteristic of amorphous or hydrated RuO, [27], which is in accordance with TGA and Raman, analyses. The broad reversible peaks at 0.6 V vs. RHE are attributed to the Ru(tID/Ru(IV) redox couple, whereas the small shoulders at 08V and 1.15V vs RHE correspond to the redox transitions Ru{IV)/Ru(VE) and Ru (VD/Ru( Vill), respectively [28]. On the other hnand, the cathodic peak often described in the literature for poten tials lower than 0.3V [29,30] is not present in the case of RuQ2/C ‘material prepared by the modified polyol synthesis method. This rnon-reversible peak was attributed to the insertion or absorption ‘of hydrogen in the oxide lattice and directly linked to the sample498 Audion el Applied atass 8: Envionmental 200 (2017) 485-502 jimAom? rates bet follows Eq, (1) [22 Mo,(OH), + 5H 4+ Be ws tea f — brads sears — eve Wet Sete ame ane 109 = 78 § so. <2 E to S25 50 soot @) “as coca oa os os to a2 aa ao a2 as os os vo 42 04 E vs. RHE/V E vs. RHE/V 5. Cyclic voltammograms of Ruxltiy.¢02/C catalysts based electrodes in ms n, * . Silt Hsde ccna recorded se Ome” sa ambience 180 — Somali — som ets twas ts crystallinity 31. In the present case, the mean Ru oxide particle 10.0 Sze fea. snm determined by TEMincices alow cytalitesize 4g. 75 Sveiing ths insertion reaction to take place Fr Ire the CV § 0. is different; two well-defined peaks are distinguishable. By anal- < 2. fn, the peake at 08 and 1.OV vs, RHE ace reasonably ated E oo tthe Ifye¥) and nV eV) redox transitions, tespectvely Sas. [25 For mixed rotheniamirdiom oxide catalysts, the vole. “50. Inogram shapes ate close to that obtained for {9/c, although 3 ‘overlap ofboth rand Riredox processes occurs. When decreasing snot b) the idm content in catalysts, the current densities atria “ne tthe HV) edo cestode wheres os or Ne oa oa ea oe u(RuV) increase and light potential shits observe Cyclic voltammograms of RusI1.x)02/C were recorded at scan E vs. RHE/V cen and 800 ns" fain 08Vto 1 4V¥s RHE Fes oa snd present as examples the CVs obtained for Raj an 03/C oon ft gy “Ew $ Imateras, respectively When inceasng the san rate, the crrent : 4 : densities increase while the CV shape remains the same. The redox Ecol, © il . transitions descbed and attributed previously are accompanied = ys . by proton transfer which iva faradai process occuring only ong TOOL, Baw ov the material surface. The formation of hydroxide surface species, 3 te Ue It 8 coolm § «0 MO, s0Ho Eft. | Miibeweeaitiendt oa here Mis the meta species Ruorir- Inthe studied potential ange g 1% . . the current densities recorded cortespond to the sum of two con- § wlBs a “ - tributions. The first one isthe formation ofthe surface hydroxides "Se 4 4 and the second one is a capacitive contribution resulting of charge 300. + Sccurulation atthe electrodeelectrolyte interac. Because the obo ato ato” ato so recorded current densities are due to electrode/electrolyte inter- ‘Scan rate /mV s* face processes, the capacitances of Rusyl7.y0z/C materials were determined by calculating the average charge and discharge cur- rent density values, The specific capacitances (C in F g-*) were ‘obtained using Eq, (2) by integration of the CV curves at different scan rates: ‘where vis the sean rate (V 5"), mis the mass in mg of oxide species in the Rusj09/C material deposited on the working electrode, £ and E) are the lower and upper potential limits for the integra- tion curves (0.2 and 1.2 Vs. RHE, respectively) and (E) the current density at electrode potential E Fig. Ge shows the changes of the specific capacitance for Rt} 2/C materials as a function of the scan rate. All capac Fir, Ocicvotammogrameof(a}&v03\Cand(b)e0:/Cbared elecrogesreoréed in0'5M HSDs electroiye st diferent sean rte: €) Change ofthe specie cae tances normalized by the metal axide ms, Meare we eifleent ran ater fr Roh 03/0 (RIO @ Real ss0s A Ris 9503, W Rsslt03 ane itances reach values higher than 300 Fg ope, ~"- which reveals the high capacitive properties of the materials. However, the capaci- tance values decrease drastically for scan rate between 1 mVs~ and 50mVs"!. For higher scan rates, the capacitance values sta- bilize, This behavior is linked to the diffusion of electro-reactive species from the bulk electrolyte to adsorption active sites in the catalyticlayer. At high scan rates, the adsorption of electro-reactive species only occurs at the most accessible surface electroactive sites, whereas atlow scan rates the adsorption takes place on active sites in the bulk of the whole material, Although higher capact-Audion et lj pple Cats Environmental 2002017) 495-502 99 Feta ? wo,ie t log ( v/ mV s") Fi. 7. Graph of ou) at nein fog |v obtained fr Ren 0/ Based tance values are obtained for ruthenium oxides, the iridium oxide ‘modified catalysts display capacitance values ofthe same order. Inset in Fig, Sc shows the capacitance values measured at 4 mV“! asa function ofthe mixed-oxides composition, Anincreas- ing linear relationship with the Ru atomic ratio is obtained. Since ‘the mean particles sizes remain constant for all Rly 402 /Cmate- rials, the capacitive properties seem to be higher for ruthenium oxide than for iridium oxide ‘When increasing the scan ate, slight shifts ofthe apparent peaks ‘on the CV are observed at higher potentials for the anodic sweep and at lower potentials for the cathodic sweep. This phenomenon, could come from a non-reversibility of the reactions occurting at the electrolytejelectrode interface. The accumulation of charges ‘an be limited by the diffusion of electro-reactive species towards the active sites or by their adsorption on active sites. To deter- ‘mine the rate determining step which controls the charge transfer ‘occurring at the electrolyte/material interface og (nae) was plot ted as a function of log (v) (Fig. 7). Because oftheir higher intensity, the redox peak corresponding to Ru(lI)/Ru(IV) is considered for RuO,/C and for Rug7slt92502/C electrocatalysts, whereas that of Ir{iV}/i(V1) is considered for the Ir-rich catalysts, Linear plots are ‘obtained. from which the slopes were determined (Table 2). All the slope values are close to 1 which indicates that the reaction at the electrolytematerial interface is limited by adsorption of the electroreactive species. ‘The voltammetric charges (q")are considered to be proportional to the active surface area orto the numberof active sites in the cat- alytic ayer and can be determined from the CVs [33]. The anodic land cathodic charges are calculated from the integration of the voltammograms between 0.2 and 1.2V vs RHE. The ratios of the anodic charge to the cathodic charge are close to 1 for each scan rate, thus the reactions are reversible. The charge q" is determined by averaging of the cathodic and anodic charge values. But, as dis- cussed previously, the reaction occurs on all active sites at very Tow scan rate (1 mVs~!) because the electro-reactive species can. diffuse inside the porosity ofthe catalytic layer, whereas a higher scan rates the reaction can only take place on the most accessi- ble active sites. Ardizzone et al. [24] established the Eqs. (3) and (@) based on the diffusion of electro-reactive species at the elec- ttodeelecttolyte interface in order to determine the total charges (@'raut) and the most accessible charges (q"o.te:) when the scan rates values tend to 0 and +e, respectively. i) aot « Rvoye . © Ruzlgas O/C Toot 4 ustag0/0 . . YF ugseleO/C s ele ow FIC Soca," o w' a) oo 02 04 06 eo vy mv" 7G)! Sense S" z 20x10" D hecttngoye 8x0" 1 Rha fe taco YY PugaesO/ 1.4410" +o wo,ie ° 5 1 1 2 25 wmv 5? tthe mast accessible charges (nan) ane (b)extcaplalin af the fla charges = Ge @ where C) and Cy are constants, vis the sean rate and gate the average charges calculated for each scan rate v. The q' outer val ies ate obtained by determining the intercepts ofthe linear pat of the curves (or scan rate values higher than 20nwV~" (Fg. 63) whereas the q"ro:a1 Values are obtained for scan rate values lower than 20 mV s~ (Fig. 8b). Fig. 9 plots the qoutes and g"yotai Values as. function of the moxed-oxde composition the valves abtaned are ‘summarized in Table 2. The q"ratat obtained for RuO3/C are higher than those obtained for :0,/C (047.4 atd SIB-1CE ouae™Fespec- tively) and mixed oxides led to intermediate values. A decreasing linear relationship with the frais obtained. Considering the ‘most accessible active sites, q” outer, the changes in "outer With the Ru conten i les pronounced but the valve increases from 299.7 {0 387.7C g-"oxide for IrOz/C and RuO,/C respectively. Theactivstesaccessbiliy. one! x, wasthen determined (rig 5b) The highest ative sites acessiliy was found forthe 1rO,/C material, representing SBX ofthe active sts, wiereas for pure ruthenium oxide the value fas down to 42%. Accessiblity Values ofthe mixed oxides are intermediate, fllowing alinearrela- tionship. This obsevatonallows assuming thatthe well-dspersed mixed metal oxide particles on the carbon support and thei small $esleadto linear combination ofthe properties af, and Ruy wethout synergistic effect, since the mean particle sizes ae idenso Audion ej Applied Catass:Enionmental 200 (2017) 485-502 Table? Elecrochemica data obtained fr Rl 0 C materials sythesid bythe modified polyol synthesis methoe (S=slope of the Lox (ou pets. ut total 00 =, 800 fro > 600 e & 500. “& 400. 200 x) = 5s. 3 0 3 as 8 aot b) . 35 . oo 02 as 08 08” 40 Ru,Ir,,,,0, composition / mol site aces (@ sl uc #8 iretion the Re conte in a a Da ogi j/mAem’) W142 43 t4 48 46 47 18 Evs. RHE/V Fig 10, Polarization curves of Ru. 0s wth x=1 (lak ne), x=0.75 (ue line) x=03 redline) x-0.25(ereenlin) and x=0 orange ine recordeé 0.594 HiS0q at smiVer" and T=20-C Inet cortesponding Tt! plot forthe OER. For Interpretation ofthe eferences to colour inthis fgure legen the reser is referred tothesweb version ofthat, tical in all samples and the active sites accessibility is the lowest for RuO2/C, the capacitance results can clearly be attributed to the inherent capacitive properties ofthe ruthenium oxide. ‘To evaluate the electrocatalytic properties toward the oxygen evolution reaction (OER) of the mixed metal oxide catalysts, polar- ization curves were recorded from 0.8 to 1.8 Vvs. RHE atS mV's"?, in ‘order to avoid, as much as possible, the support oxidation (Fig, 10) ‘The currents were normalized by the geometric surface area of the ‘working electrode in orderto compare the performances for asame mixed metal oxide loading, All catalysts led to abundant bubbling ofoxygen on the electrode surface, but the shape ofthe polarization curves is different depending on the material composition. In the case of the pure RuO2/C material, the current density starts to increase from ca, 1.40V Vs. RHE, reaches a maximum val- tues at 1.525 V vs, RHE and then decreases drastically down to ca, 1.7 mAcmY-2, Similar behaviors were already reported for hydrous ‘or non-crystalline ruthenium oxides [11,25] although the involved reactions and mechanisms are not clearly defined. Nevertheless, the anodic current from 1.4 to 1.525V vs. RHE is attributed to the ‘water oxidation, since bubbling was observed at the electrode sur- face, The passivation ofthe electrode material for higher potentials isnotelearly understood but could come from oxide material insta- bilities induced by the hydration, the size and the low crystallinity of the ruthenium oxide particles. The residual cutrent at high over potentials certainly comes from the carbon support oxidation. This, ‘observation confirms that the contribution ofthe support oxidation in the whole current recorded is negligible. Inthe case of the pure ir0;/C material, the OER starts at a poten- tial higher than that for pure ruthenium oxide (at ca. 1475V vs. RHE, close to LAB, ie. at the thermo-neutral voltage), as it has been often described in the literature [35,37]. From ca, 1A40V vs, RHE to ca. L55V vs. RHE, the Ru, material leads to higher effi- ciency for OER than the [rO; material. For higher potentials, the passivation of the IrO3/C catalyst does not occur, and the current density increases monotonously. The high electrocatalytic efft- ciency of the Ir0; particles for the water splitting reaction may come from the higher crystallinity of the particles and from the absence of chemically adsorbed water in the oxide lattice, con- versely to the RuO; material, as observed through Raman and TGA analyses. Inthe case of mixed oxide materials, when increasing the ruthe- nium molar ratio, the OER onset potential decreases down to avalue close to that for pure ruthenium oxide catalyst. At low overpoten- tials the electrocatalytic efficiency for the OFRis higher than that of 1r0/C, whereas for high overpotentials this trend is reversed. This behavior at high potentials can be related to the decrease of the active site number in the catalytic layer, owing to the passivation of the particles by the hydrous and/or non-crystalline ruthenium oxide phases, Especially considering the Rig ystra2502/C material, the shoulder appearing in the polarization curves at ca. L57V vs. RHE confirms the occurrence of the passivation process. Neverthe- less due to the presence of the IrO, sites, activity of mixed oxide catalysts towards the OER remains at higher overpotentials, Looking at the polarization curves in Fig. 10, it ean be proposed, that nan electrolysis cel, the most efficient catalyst compositions, {depending on the cell working point (considering thatthe hydrogen ‘evolution reaction is not limiting the electrolysis process and that the cathode potential remains close to OV): pure RU nanopar- ticles present the best activity for Ug <15V, mixed Rup slt9502 nanoparticles present the best activity for 1.5V< Usey < L65¥ and pure IrOz nanoparticles present the best activity fOr Usey> 1.65. In order to discard the different metal oxide composition, further ‘works will be performed with catalytic metal oxide nanoparticles deposited on a stable support suitable for electrolysis cell working conditions to determine the cell voltage at 1 Acm"2,‘Audion era Apple Catalysis 8 Entronmentl 20 (2017) 495-502 so ‘The Tafel plots are given in the inset of Fig. 10. At low overpo- tentials, it can be assumed that ohmie drop has no effect on the j-E ‘measurements [38,39], so that the Tafel slopes were measured in the potential range from 1.4 and 1.55V vs. RHE, where j<20mA cem-2, The values were determined by fitting the most linear part ‘of the curves and are summarized in Table 2, Due to the complex. reaction involving a 4-electrons transfer for water oxidation, sev- eral mechanisms were considered with different rate-determining steps [40], starting from the water adsorption on the active site (S) with a charge transfer leading to intermediate hydroxyl species: S + H20 + S—OHg, + HY +e7(RI) the second step is the deprotonation ofthe intermediate species leading toa second intermediate specie S~OH yey» S— Oy, + HY He" (R2) finally, the combination of wo intermediate species leads to the formation of O» = One, > 25 4 02(R3} Ithas been shown elsewhere that the Tafel slope values should bbe 120mV dee, 40 mV dec“) and 15mV dee" for (R1), (R2) and (R3) being the rate determining steps (rds), respectively (| How- fever, other authors [41-43] explained that, due to the catalytic layer composition, the bond strength of intermediate species to the active sites could differ and thus could lead to different Tafel slope values. Actually, they dissociated the (Ri) step in two reactions: S410 + S—OHt a, +H +e°(RY) S=OH" gs > S—OH gs (RV) where, the intermediate species § - OH,g, and S— OH gs have the same chemical structure but different bond strengths according to the active site nature. Both steps can occur alternatively or in parallel, and the Tafel slope value is then 60mV dec Inaddition, depending on the bond strength ofthe intermediate species, a chemical reaction can take place in parallel to the step leading to a Tafel slope value of 30mV dec“. The Tafel slope value for the Ruslf1.02/C catalysts (Table 2) decreases when inereasing the ruthenium atomie content, suggesting improvement of the efficiency and of the kinetic parameters for the water oxi- dation reaction. For RuO2/C and l102/C, the Tafel slope values are 49.1 and 70.4mV dec“, respectively, and are in agreement with ‘those given in other works [28,36], although slightly higher (40, and 60mV dee“! for RuO; and IrO,, respectively}. This small dis «crepancy can be due tothe low weight loading of the metal oxidein, the electrode and to the carbon oxidation occurring simultaneously with the water oxidation reaction. Anyway, taking into account the established mechanism, it can be proposed that (R2) is the rds for catalysts with ruthenium atomic ratios higher than 0.5, whereas combination of (RT? and (R1")is the rds forthe Ir rich catalyst, 6. Conclusion Iwas first shown the possibility to synthesize dispersed nanos~ tructured Ruglr.4)02 metal oxides with controlled sizes (from 1 to 3nm) and compositions, displaying supercapacitive and cat- alytic properties. Moreover. the following general highlights can be pointed out: (1) The properties can be tuned by the control of the mixed oxide composition. (2) Pure Ru, present the highest supercapacitive properties, but the addition of iridium leads to decrease this property. (3) Pure RuO, presents the lower onset potential for OER but tends to passivate at higher potentials. (4) Addition of iridium leads to limit or avoid the catalyst passiva- tion and to increase the activity at high potentials, but also to increase the OER onset potential (5) Pure Ru; nanoparticles present the best activity for Use <1.5V, mixed Rusolt5o02 nanoparticles present the best activity for 15V L.65V. (6) The mechanism of the OER at the mixed oxide catalysts was determined from the analysis of the Tafel slopes, and it was shown that this mechanism was dependent on the Rujtr ratio. Acknowledgements ‘The author greatly acknowledge the CNRS Research Grouping HySPaC (GDR n° 3652) and the Institute of Chemistry ofthe “Centre National de la Recherche Scientifique” (CNRS) for supporting this ‘work. One of us (TA) also acknowledges Roger Milly at Soprano for providing him a postdoctoral stay at the IC2MP. References [11 SX MatloomsN Slain, Inivencing fats of water lectrayeis electra etetency Renewable sustainable Energy Rev 16 2012) 4257263, [2] Su saae Potential mporcace ef iverogen as fate solu ence nd upaeion poles) en [BI € Costancens § Baanto,letrochemia conversion fa Inyérogenpreducion: a short overview, WIRES Energy 2 Dedotorg 0 1002 Wene 193 [al J zheng 29 Cygan TR Jo Hyérous utbeniam oxide as an elecioge ‘utero cerochetalcapsctor HesohemSoe 121565 [5] 1 Lu, WG Pe BE conway. sescharge and potential recovery Dheaomena a thermally and eecvochemical prepared KUO, 2016) he ‘upetapactor electrodes, Electrohi. Act 421987) 9541-2552 U6 Foss eco snaestandng the sce of DAP coche 101 S-Trast. 6. Lod. Oxygen ae clorin evolution at conductive metalic fxd andes i. raat 2) Hedsodes of Conductive Metal Oe, [8] M. Carmo, DL Fitz, J Merge, D. Stolten, A comprehensive revie ‘ater cletroiyss Int | Hydrogen Energy 3812015) #80) ~4004 (an Feference cer. Io] T Audion Mayousse. 5. Morisset. ¢ Moras. ¢ comming ‘ipprs Kola Ecouy of uD med nunca ‘cla pot, ne | tyrogen Envey 39 (2014) T6785 16796, Wo] KRU Pelerno, LC Vi laser caleaton method fe the preparation of DSA® pe oxide electrodes, ‘letocem, contun 42008) 933-102 (1) A Devadas § Branton, TW Napporn,CCovtancea,Talring of Ru03 ‘hospi by tcoave tects meted or ener age 112] J-¥. Ki, KH. Kim, S.-H. Park, K-8. Kim, Microway ‘anoetstlinerutieaium otgenanoparides on carbon nanetubes or 7-5. Wa, D-W Wang W fen] Zs, Zoe FHM, cheng Anchoring Jndrous Ru, on graphene shee for hgh petirmanceelecteeciemiel ‘apscitors, dv. ace Mat 202040) 3995-3602, 04) Hokum. B8Popv.charscrerztonofthyrousrthenium exdecarbon ‘orompenteipecapctes reputed by sclindal ete, Powe 105] ¥-T Shin K=¥ ne, ¥-5 ivan, Chaaceration af rg lode ?eab0n ‘nanotube nanocomposites grown ont graphene fr supereepacte.) Alloys Compa 819(2015) 131-157, (16) Ft, 5 Hoang CK Tong, Charactezation af Ry an 40> fs Aeposited on Substrate | Aloys Crp. 317-318 [2001] 96-102, 17 AV Korateo, V5 Huang DS. Te KK long. Raman eeaterng ‘Gharsterztion of vera signed 10 rd, nanecysale grown on single ‘cyst nie substrates, oid tae Commun 137 (2006) 310-314 IR AK be Andrage K BetazzlThity minotes iyo syntesisof02 Audion ej pple Catass 8: Enionmenta 2002017) 485-502 {R9] tow tans Kulu ¥. Nakao Electracatalyie att ot ‘amorphous Ruy electrode fr oxygen evlian nan aqueous ol Hretrochim Acta 30 2011) 2008-2016, [n9) |-¥ kim X-it Kim Hk Kim, Soil Pa KC Rob, KB, Kin, Templateseee "yhens of ruthenum ide nanataes fr high-performance, cfesechenia capac ACS App Maer neces 72013 [20] V5 Huang. FH Polak. Raman investigation ofaile Ru, Sali state Comm, scion} ois [at] Rs: chea Cc ches ¥S. Huang. Cx. Chia Chen DS Ts KX Tieng A fonnparatve study of irostrctue of RO, tanorads via aman seateing Sn fig emssion caning eecton microscopy, Sok State Cor 1 {Boos} 450353, D Rochefort abe, 0. Cuay, PMA. sherwood. XPS investigations of ‘thermally pepased Hud, electrodes in reductive enaltons Heciocun, ‘sa 45 (2003) 2451252 [aa] RS Chem ¥-5 sang, ¥-M. Lang, D-S.Te YC J Kai Grove conc ae eharceneaion of verily lived tO; nanored, Mater Cherm 13 S003) 2505-2520, [24] Lok Naslund AS ington S Holo} Rosen, Fetmstion of 4001), on ud)-based electrodes for hydrogen production} Pjteal Chem C118 (2014) 15315~19923, [25] Me Steer Schl letronic structure investigation of toi aver {epeston utheniam{enide) hin fils using photoemission spectroscopy ‘App Phys 11820131 oess06-1-065300-7. [26] | Cheng zhang Ha H zhong ¥. Zs. Sud of xrbonsupported 103 nd Ro forse the hydrogen evoltion estion ino poiymer leroy eecrolyzer,Hectroctn, Acta 332010) 1835-1861 (a7) ket ehang €-€ Hus €-¥. chou, extra and capuciive characteristics of Iyatotermaliy derived Rio, 0 nanocrystals: independent conta of ttyl size ae water copter. Chem Mater. 192007) 2112-2119. [20] EU nates-costaP de ime Neto SAS Machado, LA Avaca {Gharacesauon of suraces modiiee by soe erved Runa, coatings fovenygen evolution un acid medium Eetrohin. Act 441938) 1995-1523, (29) ke uodtarn} odtayte,V.Sukene A Gagucevitene A els On the fharge storage mechanism at Ruds/0 SM Hy80q interface | So Sate Hectrochem, 122008) 1398-1408 [30] | ibe. Ms. Moats, AK De Ancrade, Morphological and electrochemical Invesiatis of Ru, Ta,0, aide ns plepaed dy the Pehine Aas tethed Appl Electrochem 38 (2008) 767.775, aa {31] Dalz ¢ tantra. s.Trasatt tuchenium dioxide: anew electrode ‘aerials Bebaviout in acid solutions of ier electrolytes, APB Hetrocher 41874) 57-57, [22] Cc Hu cH Chang. Cet vetammetrc deposition of hydrous ruthenium ‘nae for electrocheiclcapactorseflecsococepesing nau ode etrochin. Aes (2000) 2685-2596, [23] 5 Arstzone, 5 Tease nef properties f oxides with technolgicl inact inelectroehersry. Adv Col Inerface Se 641895) 173-251 [4] ScArizone.c.regonarS Tas Inver ab Our active race oF UO tiectogen Elected. Ata 35 (1890) 263-267 [35] T Audion, Te Nagpal, C Morais, C.Comminges.K3. Kokoh. IrOy ested op Ru efit an table eletrasetive necator leerochereal water spit | Phys chem © 1202016) 2562-2573 [36] Keke s stun, Stabilisation of RO, by tO, fo anode oxygen evolution in stad media Hlecsochun. Ata 31 (1985) 1311-1336, 137] [Cheng Zang & chen ¥ Zhang, Stud ae 40y oxides as anodic ‘ecroetalyt lr slid paler eceolste we letra leesochim ‘hes 542009) 6950-5256, {36 LE Owe, Mt Toypin. KS. Wallwork RG Haverkamp.s Sunde, iiumn?vutheniu singe phase ned exes for oxygen evolu ‘composition dependence of eleeoctalii activi, ler achitn ela 70 {Bo1as 158-168 [19] RHP Santana LA, De Faria. Oxygen and chlorine evation on uo, #10, CaO, *Nb,O; mined one elaiodes ctochm. Ata St Ma] ¥-Mavsreto Ete, Eetreataiytepopestis af trasison Ui] La‘Defanas jc Bosses 5 Trasate Elecrocstnye properties of terary ‘nude mixtures of competion 80 STierCes0, oxygen evolution fam ‘Tl tolution app Hectrochers 261296) Yies-t fee, Ua] E Anal chim ny catlyt ane costal for oxygen Fvoeton Reduction Inaidepalymerctectrayre membrane eleralaers ane fel el ACS (Gata. a(zora) 126-1440 Us] Li baba LC Boots, LA, De Fata, oxygen evel at py = CovOs 9 eeettedes Hor ed Saluion lett. Acta 46 (200%), sss,