Journal of Electroanalytical Chemistry 704 (2013) 57–66

Contents lists available at SciVerse ScienceDirect

Journal of Electroanalytical Chemistry

journal homepage: www.elsevier .com/locate / je lechem

Effect of heat treatment on the electrocatalytic propertiesof nano-structured Ru cores with Pt shells

1572-6657/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.jelechem.2013.06.007

⇑ Corresponding author at: Department of Materials Science and Engineering,Norwegian University of Science and Technology (NTNU), NO-7491 Trondheim,Norway. Tel.: +47 735 94051; fax: +47 73 59 11 05.

E-mail address: Svein.Sunde@material.ntnu.no (S. Sunde).1 Current address: Grupo de Energía y Química Sostenibles, Instituto de Catálisis y

Petroleoquímica, CSIC c/Marie Curie 2, E-28049 Madrid, Spain.2 ISE member.

Mikhail Tsypkin a,d, José Luis Goméz de la Fuente a,d,1, Sergio García Rodríguez b,d, Yingda Yu a,d,Piotr Ochal a,d, Frode Seland a,d, Olga Safonova c,d, Navaneethan Muthuswamy a,d,1, Magnus Rønning a,d,1,De Chen a,d,1, Svein Sunde a,d,⇑,2

a Department of Materials Science and Engineering, Norwegian University of Science and Technology (NTNU), NO-7491 Trondheim, Norwayb Grupo de Energa y Qumica Sostenibles, Instituto de Catálisis y Petroleoquímica, CSIC c/Marie Curie 2, E-28049 Madrid, Spainc Paul Scherrer Institut, 5232 Villigen PSI, Switzerlandd Department of Chemical Engineering, Norwegian University of Science and Technology (NTNU), NO-7491 Trondheim, Norway

a r t i c l e i n f o a b s t r a c t

Article history:Received 15 December 2012Received in revised form 31 May 2013Accepted 4 June 2013Available online 25 June 2013

Keywords:ElectrocatalystCO electrooxidationMethanol electrooxidationBifunctionalLigand

Ru@Pt core–shell particles are relevant for application as electrocatalysts in fuel cells. The Ru core isexpected to influence the activity of the Pt in the shell through a compression of bond lengths andelectronic interaction with the core. In this work Ru@Pt core–shell (Ru core and Pt shell) and pure Runanoparticles of diameter below less than 5 nm were synthesized and supported on carbon black (VulcanXC-72). The supported catalysts were heat-treated at temperatures up to 500 �C. Analysis of the catalystsby TEM, EXAFS, XRD, and CO-stripping indicates a strongly segregated architecture with Ru in the core ofthe particles. Upon heat-treatment we observed moderate particle growth, increased extent of alloying,and a decrease of the Pt–Pt bond lengths. The Pt–Pt bond lengths decreased uniformly with heat-treat-ment temperature in the entire range. The extent of alloying and particle growth were significant (i.e.beyond measurement uncertainties) only at a heat-treatment temperature of 500 �C. The electrocatalyticactivity for oxidation of adsorbed CO (CO-stripping) increased in the entire temperature interval. Theactivity for methanol oxidation only increased when catalysts were heated to 500 �C. The results indicatethat the surface concentration of ruthenium in the pristine Ru@Pt catalysts is small.

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

Efficient and stable catalysts for oxidation of CO and smallorganic molecules such as methanol are important for low- tomedium temperature fuel cells fueled by CO-containing hydrogenor organic fuels directly. Bimetallic Pt–Ru catalysts often have asuperior activity for these reactions as compared to Pt and Rumono-metallic catalysts. This is frequently explained by a concep-tual splitting of the effects of alloying into a bifunctional mecha-nism [1,2] and a ligand effect [3,4]. The bifunctional mechanismassumes that both metals are available at the surface, formingadsorption sites. The ligand effect only depends on one of themetals being close enough to the surface to influence the otherelectronically.

Often bi-metallic PtRu electrocatalysts containing equalamounts of Pt and Ru are prepared in the form of alloys. The opti-mum fraction of Ru surface sites has, however, been asserted tobe well below 50% [5–7] and is temperature-dependent [5]. Particlesize [8,9], agglomeration and surface structure [10–13], the type ofcarbon support [14–18], and the oxidation state of Ru [19–37] areadditional factors that all appear to influence the activity as well.

Catalyst architectures other than alloys, such as core–shellstructures [38–46], can change the catalytic activity and stabilityof electrocatalysts significantly. Recent results indicate thatencapsulating Ru inside a shell of the more stable Pt may retaina significant fraction of the electrocatalytic activity for oxidationof pre-adsorbed CO (CO-stripping) whilst inhibiting dissolution ofRu [43]. However, the performance of a direct-methanol fuel cellemploying the Ru@Pt catalyst was little different from one employ-ing Pt at the anode [47]. Additionally, the peak potential observedin voltammograms in methanol-containing solutions at the Ru@Ptcatalyst is little different from Pt [47].

The structure of Pt–Ru systems depend on the preparation condi-tions. Thus, the surface of PtRu catalysts has been reported both to beenriched in either Ru [48–51,38,52,53] or Pt [23,37,34]. The pristine

58 M. Tsypkin et al. / Journal of Electroanalytical Chemistry 704 (2013) 57–66

catalysts are typically partially oxidized [48,52,53], but the oxida-tion state can change during operation [48,51]. Heat-treatment ofthe catalysts in hydrogen or vacuum at low or moderate tempera-tures promote alloying [52,53,51]. In some cases indications of Ptsegregation to the surface have been reported [38,54,34]. Heatingin air or CO at similar temperatures leads to de-alloying.

For these reasons it becomes important to investigate the influ-ence of heat treatment on the Ru@Pt core–shell catalysts. Ru@Pt cat-alysts were demonstrated in Ref. [43] to be substantially more stableunder potential cycling than alloys; Heat treatment may serve as analternative way for investigating the stability of the core–shell struc-ture, particularly for their possible application in intermediate-temperature methanol fuel cells [47]. Associated with possiblestructural changes in the catalyst upon heat-treatment one may alsoexpect concomitant changes in electrochemical performance.

Below we report the results of such heat-treatment of Ru@Ptcore–shell electrocatalysts. We first report the structural and elec-trochemical characteristics of the ruthenium cores, and then thestructural and electrochemical characteristics of the Pt-coveredRu-cores, all supported on Vulcan XC-72. We provide further indi-cations of the completeness of the Pt-shells and their ability to re-tain the Ru inside the catalyst. The structural changes appearingduring the heat treatment are then correlated with the electrocat-alytic properties of the catalyst.

2. Experimental

2.1. Materials and chemicals

PtCl2 and Ru(acac)3 (acac = acetylacetonate) were provided byJohnson Matthey. Ethylene glycol (EG) and polyvinylpyrrolidone(PVP, MW = 55000) came from Sigma Aldrich and Vulcan XC-72from Cabot.

2.2. Synthesis of nanoparticles

The Ru@Pt 1:1 core–shell nanoparticles were synthesized byusing a sequential polyol process [41]. Ru(acac)3 was initially re-duced in refluxing ethylene glycol (EG) in the presence of polyvi-nylpyrrolidone (PVP) stabilizer. The resulting Ru cores werethoroughly washed in order to remove traces of chlorine and stabi-lizer (PVP). The Ru particles were then deposited on the carbonsupport by stirring at room temperature, followed by washing withacetone and centrifuging. For the core–shell particles an aliquotRu/PVP (solution) was added to a PtCl2/EG mixture and heated to200 �C. The resulting Ru@Pt particles were deposited on carbonby stirring at room temperature, then washed with acetone andcentrifuging. The uncoated Ru cores (also carbon supported) wereused as a metallic Ru reference.

All nanoparticles were deposited with 9–16% nominal loadingson the carbon support (Vulcan XC-72) and dried at 100 �C. Cata-lysts receiving no further treatment are referred to below as as-prepared, pristine, or labelled as 100 �C.

2.3. Catalyst characterization

A small amount of the catalyst powder was dispersed in etha-nol, pippeted on a copper grid, and transmission electron micros-copy (TEM) images of these were obtained at a Jeol JEM-2100.The loading was determined with a thermogravimetric analyzer(TGA Q500, TA instruments). The sample was placed in an aluminacrucible and heated to 800 �C at a rate of 20 �C/min under a flow ofsynthetic air (80 ml min1 of nitrogen and 20 ml min1 of oxygen).The data were corrected for formation of RuO2. The catalyst weightwas found as the difference in sample weight before and after car-

bon decomposition, from which the nominal loading was calcu-lated. We note that the calculated loading may be affected by thepresence of any remaining PVP in the catalyst. However, the actualloading does not affect the conclusions to be drawn below and thisis therefore of no consequence here. The ratio between platinumand ruthenium was ascertained by EDS/SEM using an Oxfordinstruments Aztec EDS system attached to a Hitachi S-3400N SEM.

High-resolution X-ray data were collected at the BM01B beam-line (ESRF, France) combined XRD and XAS setup. A small amountof the catalyst (10–20 mg) was immobilized between quartz woolin a 2 mm glass capillary. A flow of argon was maintained throughthe compacted catalyst for oxygen removal. The catalyst surfacewas first exposed to argon at room temperature, then rapidlybrought to temperature, and kept there for 30 min. The sampleswere then quenched to room temperature before the XAS spec-trum and XRD diffractogram (wavelength 0.500798 ± 0.000006 Å)were recorded. After a short rest period the sample was broughtquickly to the next temperature and so on. The XRD-data wererecorded in a range of 2h from 7� through 45� with step size0.003. The XRD results were analyzed with help of Topas software(Bruker). A Pawley-type full-pattern fitting approach was used[55].

Extended X-ray absorption fine structure (EXAFS) data were col-lected at the Ru K edge (22.117 keV) and the Pt L3 (11.564 keV) inthe transmission mode (monochromatic radiation, Si(311) doublecrystal) also at the Swiss-Norwegian beamline (SNBL) at the Euro-pean Synchrotron Radiation Facility (ESRF Grenoble, France). AllXAS data were processed using Athena. The extracted EXAFSspectra were fitted to the theoretical IFEFFIT [56] standards usingthe Artemis program. These spectra were multiplied by k3-weight-ing factors and fitted simultaneously in R-space. Standard devia-tions calculated by the software are reported below. (Notehowever that the standard deviations calculated by the softwareare only estimates of precision (reflecting statistical errors in the fit-ting). This may overestimate the accuracy somewhat, particularlyin the case of a high degree of correlation between parameters. Esti-mated standard deviations for the first-shell distances are 0.01–0.02 Å and 10% for the coordination numbers [57–59,53].) Particlediameters were calculated from the EXAFS data according to proce-dures in papers by Frenkel et al. [60] and Karim et al. [61].

2.4. Electrochemical characterization

The electrochemical measurements were conducted in a stan-dard electrochemical glass cell using an Autolab potentiostat, typePGSTAT302N. The working electrode was prepared by applying thesame (small) amount of the catalyst to a glassy carbon disk andcoating with a thin film of Nafion� according to procedures elabo-rated by Schmidt et al. [62]. For all results presented here a Pt plateof 1.5 cm2 was used as a counter electrode. Experiments were per-formed at room temperature in 0.5 mol d m�3 HClO4. Potentialswere determined using a reversible hydrogen electrode (RHE).

CO gas was bubbled through the electrolyte for 140 s at 50 mVto ensure full coverage of the electrode surface by adsorbed CO.The solution was purged then with argon for 30 min again to re-move dissolved CO. All currents for the electrodes tested for meth-anol oxidation are normalized by the active area of the electrode.The active area was determined from the charge of the CO oxida-tion peak in the first CO stripping voltammogram.

3. Results and discussion

3.1. Characterization of uncoated Ru particles on carbon

Fig. 1 shows TEM images of the untreated Ru/C and the sampleheated at 500 �C under argon flow. Most of the particles appear to

Fig. 1. TEM images of uncoated Ru cores (deposited Vulcan XC-72, not heat treated) as made (left) and heat treated at 500 �C (right). Histograms are shown in the insets.

Table 1Particle diameter d and lattice parameters a and c as obtained from an analysis of theresults of the high-resolution diffraction experiments on uncoated Ru/C heat-treatedat various temperatures.

Heat-treatment temperature (�C) d (nm) a (Å) c (Å)

100 1.2 2.713 4.240300 1.2 2.717 4.231400 1.3 2.706 4.221500 3.8 2.703 4.285

M. Tsypkin et al. / Journal of Electroanalytical Chemistry 704 (2013) 57–66 59

be narrowly distributed around 2 nm, although some particles aslarge as 4 nm are already apparent in the as-synthesized sample.The size distribution for the heat-treated sample (500 �C) is cen-tered at slightly larger values and broader. In the heat-treated sam-ples, however, some quite large particles (beyond the size-rangeincluded in the size histograms) are apparent, as large as 20 nm [63].

Fig. 2 show high-resolution X-ray diffraction (HE-XRD) data forthe Ru cores heat-treated at different temperatures. The diffractionpeaks become sharper and more distinct the higher the heat-treat-ment temperature of the sample.

Table 1 lists the a- and c-axis lengths as found from analysis ofhigh-resolution X-ray diffraction data of the uncoated rutheniumcores. An expansion of the c-axis is apparent, whereas the a-axisappears to contract slightly with increasing annealing tempera-ture. The axis lengths are comparable to those of bulk Ru,a = 2.70389 Å and c = 4.28168 Å [64].

Fig. 3 shows an example of a Fourier-transformed k3-weightedEXAFS dataset for the ruthenium cores. The XANES edge (notshown) was rather broad and contained no sharp features associ-ated with electron confinement. This indicates that oxides arenot present [41,32,48,50,53]. This is also apparent directly fromFig. 3 since the Fourier-transformed EXAFS-spectra exhibit no peak(1.6 Å, not phase-corrected) corresponding to the Ru–O bondlength [65]. The peak corresponding to Ru–Ru coordination(2.3 Å) is clearly dominating the spectra, on the other hand. We

Fig. 2. High-resolution X-ray diffractograms measured for Ru cores heat-treated atthe temperatures indicated in the legends (bottom: as prepared, middle: 300 �C,top: 500 �C).

therefore conclude that a metallic state dominates in the Ru coresunder the conditions of the shell synthesis.

Table 2 reports the average Ru–Ru coordination number and theestimated crystallite diameter as found from the EXAFS analysis. Asignificant increase in both particle diameter (from 1.6 nm to3.7 nm) and Ru–Ru coordination number (from 8.5 to 10.1) withtemperature is apparent, the largest change occurring at 500 �C.Bond lengths were substantially closer to bulk values than onestandard deviation and showed no variation with heat-treatmenttemperature. These are therefore not reported.

The XRD and EXAFS analysis appear to confirm the slight in-crease in particle size of the TEM images (Fig. 1), at least qualita-tively. The difference between the crystallite size from X-rayanalysis and particle size from TEM indicates that the particles insamples heat-treated at 500 �C (see Table 2) are not single crystalsbut agglomerates of such. The Ru–Ru coordination number appearsto approach the bulk value as the heat-treatment temperaturereaches 500 �C. This is in agreement with the increasing degreeof crystallization with heat-treatment temperature, indicated bythe XRD peaks becoming narrower and sharper the higher theheat-treatment temperature in Fig. 2, and also with the increasingcrystallite size the higher the heat-treatment temperature inTable 1. The structural changes are not even across the tempera-ture interval, however, and appear to be significantly affected onlyat the highest heat-treatment temperature of 500 �C.

Fig. 4 shows CO-stripping voltammograms for Ru/C heat-trea-ted at various temperatures. The peak potential for CO oxidationappears to decrease more or less continuously as the heat-treat-ment temperature increase. This negative shift indicates that thespecific catalytic activity increases with heat-treatment tempera-ture (see below) [63]. Since disordered Ru(0001) surfaces havebeen found to be more active than ordered surfaces [66] it is some-what difficult to relate the improvement in catalytic activityobserved here directly to the increased crystallinity upon heat-treatment. Oxidation of CO is a very structure-sensitive reaction[67], and the changes in peak potential found here may be relatedto faceting or other structural features not directly probed by theanalysis techniques we employed.

Fig. 3. Fourier-transformed k3-weighted EXAFS data (k3v(k)) measured at atemperatures of top: 100 �C (pristine), middle: 300 �C, and bottom: 500 �C for theRu K-edge of the Ru core catalysts. Fits are included as dashed lines as indicated inthe figure.

Table 2Coordination numbers NRu–Ru and particle diameter d as obtained from an analysis ofthe results of EXAFS experiments on uncoated Ru/C heat-treated at varioustemperatures. Standard deviations as calculated by the software are included inparentheses.

Heat-treatment temperature (�C) NRu–Ru d (nm)

100 8.5 (0.2) 1.6300 8.3 (0.2) 1.5400 8.3 (0.2) 1.5500 10.1 (0.2) 3.7

Fig. 4. CO-stripping voltammogram for uncoated Ru/C and Ru/C heat-treated atvarious temperatures. (The current is normalized with respect to the peak current.)Sweep rate 10 mV/s.

60 M. Tsypkin et al. / Journal of Electroanalytical Chemistry 704 (2013) 57–66

3.2. Characterization of heat-treated Ru@Pt

TEM images of untreated samples and samples heat-treated at300 �C and 500 �C are shown in Fig. 5. Some agglomeration isapparent in the samples heat-treated at 500 �C as reflected in along tail in the size-distribution towards larger values. The agglom-eration is perhaps a little less than that observed with the uncoatedRu-cores, Fig. 1, however.

Fig. 6 displays high-resolution X-ray diffractograms as a func-tion of temperature for the Ru@Pt core–shell catalysts. Only peakscorresponding to Ru, Pt or PtRu alloys are visible for all heat-treat-ment temperatures. The diffraction patterns displayed in Fig. 6 arecomparable to those simulated and measured for similar Ru@Ptcore–shell catalysts in Ref. [42]. The diffractograms in Fig. 6indicates a better stability upon heat-treatment than those of theRu cores, Fig. 1.

Analysis of high-resolution X-ray diffraction data resulted in thediameters and lattice parameters given in Table 3. For the Pt phasewe include the diameter suggested by the XRD analysis and the lat-tice parameters corresponding to a face-centered cubic (fcc) sys-tem. For the ruthenium the lattice parameters for a hexagonalphase (hcp) are included along with the crystallite diameter. Theincrease in the coordination number for Ru–Pt and correspondingdecrease in that for Pt–Pt from the EXAFS analysis (reported in Ta-ble 4 below) indicate that for a heat-treatment temperature of500 �C the coated system is, at least partially, forming an alloy.Therefore, for this heat-treatment temperature we also includedin the analysis an alloy phase. The diameter and the lattice param-eter were calculated assuming the alloy phase to be cubic, i.e. thatRu dissolved in Pt forms a cubic system [68]. A significant increasein the Pt-shell thickness or the formation of separate Pt crystallitesis also apparent from Table 3 for the highest heat-treatment tem-perature (500 �C). The Ru core appears largely unaffected, and onlya slightly larger diameter for the sample heat-treated at 500 �C isapparent. The average value for the Pt–Pt bond distance in Table 3is a=

ffiffiffi2p¼ 2:771 Å. This is slightly less than that found for bulk Pt

foil by EXAFS in Ref. [41] (2.774 Å, bulk value [64]) and slightly lar-ger than that of Ref. [69] ða=

ffiffiffi2p¼ 2:766 ÅÞ.

Fig. 7 shows examples of the Fourier-transformed k3-weightedEXAFS Ru K-edge data for the Ru@Pt core–shell structures. Againthe XANES data (not shown) and the EXAFS data indicated that lit-tle ruthenium oxide is present in the samples. We thereforeconclude that the as-synthesized and heat-treated Ru@Pt core–shell structures contain Ru in a low-valent state [41].

Analysis of the EXAFS data resulted in the coordinationnumbers reported in Table 4. In the as-prepared samples, the

Fig. 5. TEM images and size histograms (insets) of an untreated Ru@Pt/C sample (top left), a Ru@Pt sample heat-treated at 300 �C under Ar flow (top right), and a Ru@Ptsample heat-treated at 500 �C under Ar flow (bottom). Particle size distributions are included in the histograms.

Fig. 6. High-resolution X-ray diffractograms measured for Ru@Pt samples heat-treated at the temperatures indicated in the legends (bottom: as prepared, middle:300 �C, top: 500 �C).

Table 4Coordination numbers Ni�j as obtained from an analysis of the results of EXAFSexperiments on uncoated Ru/C heat-treated at various temperatures. Standarddeviations as calculated by the software are included in parentheses.

Temperature �C NRu–Ru NRu–Pt NPt–Pt NPt–Ru

100 7 (1) 0.4 (1) 8 (1) 0.4 (0.5)300 7.8 (0.7) 0.5 (0.8) 9 (1) 0.6 (0.5)500 7.7 (0.5) 1.2 (0.9) 8 (1) 1.2 (0.5)

M. Tsypkin et al. / Journal of Electroanalytical Chemistry 704 (2013) 57–66 61

coordination numbers are similar or even very close to previouslypublished results for Ru@Pt core–shell structures [41,70]. FromTable 4 it is apparent that the Ru–Ru coordination numbers change

Table 3Crystallite diameters d, i. e. the size of the crystalline domain as obtained from the refinemethe results of the high-resolution diffraction experiments on coated Ru@Pt/C heat-treated

Temperature �C d (alloy), nm a (alloy), Å d (Pt), nm

100 – – 2.8300 – – 4.5500 3.03 3.900 24.0

much less than those of the uncoated cores, Table 2. The Ru–Rucoordination numbers are therefore to be regarded as independentof heat-treatment temperature (within the measurement andanalysis uncertainty). The coordination numbers for Pt–Ru andRu–Pt appear to make a small jump to larger values for the highestheat-treatment temperature of 500 �C. This suggests that some Ptis transported into the Ru core and vice versa.

The Pt–Pt bond-lengths obtained from the EXAFS analysis were�0.020 ± 0.005 Å, �0.025 ± 0.006 Å, and �0.030 ± 0.006 Å for sam-ples heat treated at 100 �C (as prepared), 200 �C, and 500 �C,respectively, relative to a reference bond-length of 2.766 Å [69].For the as-prepared sample the Pt–Pt bond-length shorteningrelative to the bulk is thus in the same order of magnitude as foundpreviously [41]. That the Pt–Pt bond is being continuously

nt of the XRD patterns, and lattice parameters a and c as obtained from an analysis ofat various temperatures.

a (Pt), Å d (Ru), nm a (Ru), Å c (Ru), Å

3.920 1.0 2.734 4.2723.917 1.0 2.733 4.2823.920 1.1 2.704 4.282

Fig. 7. Fourier-transformed EXAFS data measured for the as-prepared sample(100 �C) and samples heat-treated at 300 �C and 500 �C for Ru K-edge of the Ru@Ptcore–shell catalyst. Experimental data are shown as solid lines and the theoretical,fitted Fourier-transform of k3v(k) as dashed lines as indicated.

62 M. Tsypkin et al. / Journal of Electroanalytical Chemistry 704 (2013) 57–66

compressed (more and more negative DR) is consistent with theinclusion of some Pt in the lattice of the smaller Ru. (The decreasein DR for Pt may appear to be inconsistent with the large growth inthe Pt-phase diameter in Table 3. However, EXAFS measures theaverage Pt–Pt distance including both Pt in the alloy and Ptphases.).

An important difference between the Ru@Pt particles and theuncoated Ru-cores is that the increase in the diameter andRu–Ru coordination number of the latter upon heat-treatment(Table 2) is almost absent in the former (Table 4). It thereforeappears that the Pt shell protects the Ru-particles from growing.

Fig. 8 shows cyclic voltammograms for pristine Ru@Pt/C andRu@Pt/C heat-treated at 500 �C in perchloric acid. Voltammograms

for ruthenium (not heat-treated), platinum, and the platinum-ruthenium alloy are also included for comparison. Compared tothe ruthenium catalyst, the presence of weakly and strongly ad-sorbed hydrogen-desorption peaks in the Ru@Pt catalyst, absentin the Ru voltammogram, is the most distinguished feature. Com-pared to the voltammogram for the Pt catalyst, the desorption peakfor weakly adsorbed hydrogen is displaced to lower potential val-ues. Peaks appear at approximately 0.13 V and 0.20 V for theweakly and strongly adsorbed hydrogen on Pt, respectively. Forthe pristine Ru@Pt catalyst the corresponding values are 0.07 Vand 0.20 V. For the heat-treated Ru@Pt catalyst the values are0.05 V and 0.15 V. In addition, the ratios between the peak currentsare distorted compared to the Pt catalyst; In the Ru@Pt the peakcurrent for the strongly adsorbed hydrogen is much lower in thecore–shell catalyst relative to the peak current for the weakly ad-sorbed hydrogen. The voltammogram for the PtRu alloy displayshydrogen desorption peaks more or less at the same potentials asfor platinum and of approximately the same ratio. For the alloythe peaks are less pronounced than they are for Pt, however.

The voltammogram for the Ru@Pt catalyst in Fig. 8 displays ex-actly the same features that are apparent in the voltammogramsfor Ru@Pt recorded by Bernechea et al. [44], and are in agreementwith findings that pseudomorphic monolayers of Pt on Ru(0001)surfaces interact weakly with adsorbed hydrogen [71]. A smallchange is apparent between the pristine and heat-treated Ru@Ptsamples in Fig. 8. However, in view of the variations we observedfrom voltammogram to voltammogram the difference is not signif-icant. The hydrogen desorption peaks are thus distinctly differentfor the Ru@Pt catalysts as compared to PtRu alloys and the purePt catalyst. The potentials and the relative magnitude of the peakcurrents for the adsorption–desorption peaks in the voltammo-grams are not very sensitive to the heat-treatment, however.

The activity for oxidation of an adsorbed monolayer of CO is dis-played in Fig. 9. Taking the peak potential as an expression for thecatalytic activity [72–75], it is apparent that the activity increasessignificantly with increasing heat-treatment temperature.

The activity for the methanol oxidation is shown as a chrono-amperogram in Fig. 10. In this case there is an increase in catalyticactivity with heat treatment temperature to 500 �C, but only forthis heat-treatment temperature. For a heat-treatment tempera-ture of 300 �C there appears to be no significant change in theactivity from that of the pristine catalyst. In other words, for themethanol oxidation reaction the sample heat-treated at 500 �Coutperforms all the others.

The apparent increase in catalytic activity in the CO-strippingvoltammograms may be interpreted in terms of changes in eitherbifunctionality or ligand effects due to the heat-treatment. Forthe ligand contribution the catalytic activity is expected to corre-late with d-band energy [76–79] and thus the Pt–Pt bond-length[78]. The bifunctional effect is expected to depend on the surfacecomposition. These possible contributions to the changes in elec-trocatalytic activity with heat treatment are illustrated in Fig. 11,which shows the Pt–Pt bond distance, the exponent of the peak po-tential Ep for CO stripping exp (�FEp/RT), the methanol-oxidationcurrent iMOR at 1000 s, and the extent of alloying [49] for Pt (JPt),all as a function of heat-treatment temperature.

The extent of alloying JPt for Pt was calculated as in Ref. [49],

JPt ¼NPt—Ru=ðNPt—Ru þ NPt—PtÞ

Prandomð1Þ

where Prandom = 0.5. For the pristine sample these numbers are sim-ilar to those calculated previously for Ru@Pt particles [70]. A signif-icant increase in JPt associated with the changes in the coordinationnumbers at 500 �C is apparent.

We have chosen to represent the electrocatalytic activity foroxidation of adsorbed CO by the exponent of the negative of the

Fig. 8. Cyclic voltammograms for pristine Ru@Pt/C and Ru@Pt/C heat-treated at 500 �C as indicated in the legends. Voltammograms for ruthenium (not heat-treated),platinum, and the platinum-ruthenium alloy are included as indicated by the legends. Sweep rate 10 mV/s. The experiments were performed at room temperature in0.5 mol d m�3 HClO4.

M. Tsypkin et al. / Journal of Electroanalytical Chemistry 704 (2013) 57–66 63

peak potential. This follows from the equation for CO-stripping vol-tammetry proposed by Herrero et al. [73], which we write as:

j ¼ QFvRT

ax expð�xÞ½a expð�xÞ þ 1�2

ð2Þ

In Eq. (2) Q is the charge associated with CO oxidation, v is thesweep rate, kapp is the apparent rate constant, and R, T, and F thegas constant, the temperature, and Faraday’s number. In Eq. (2) xis exponential in the potential E,

x ¼ kappRTFv exp

FERT

� �ð3Þ

and a is approximately given by

a � hCO;ini

1� hCO;inið4Þ

where hCO,ini is the initial coverage of CO. From these equations wefind that the apparent rate constant is related to the peak potentialEp as,

kapp / xp exp � FEp

RT

� �ð5Þ

where xp is the value of x for which the function x exp (�x)/[a exp(�x) + 1]2 has its maximum. The value of xp will depend on a andthus the initial coverage. However, for typical values of hCO,ini thevariations in xp turned out to be small as compared to those ob-served in exp (�FEp/RT), shown in Fig. 9. (For h = 0.1, xp = 1.1 andfor h = 0.9, xp = 2.9. In other words, the variations in xp are verymuch smaller than the two-orders-of-magnitude change in the fac-tor exp (�FEp/RT) with heat treatment temperature.) We thereforetake the latter expression as an acceptable approximation to therate constant. (For a pure Pt catalyst of otherwise similar character-istics to the Ru@Pt catalysts the peak potential for CO-stripping wasapproximately 0.8 V corresponding to an activity on the scale im-plied by Eq. (5) of 3 � 10�14.)

For n monolayers (ML) the extent of alloying would be JPt–Pt = 6/[(n � 1)12 + 9]) � 100. This gives an extent of alloying equal to 67%,29%, 18%, and 13% for one, two, three, and four ML, respectively.

Fig. 9. CO-stripping voltammogram for Ru@Pt/C heat-treated at various temper-atures as indicated in the legend. (Currents are normalized with respect to peakcurrent.) Sweep rate 10 mV/s.

Fig. 10. Chronoamperometry in 0.5 mol l�1 H2SO4 and 2 mol l�1 CH3OH stepped to0.5 V for Ru@Pt/C heat-treated at various temperatures as indicated in the legend.Prior to the step the electrode was kept in solution at 50 mV for 300 s.

Fig. 11. Top: Pt–Pt bond distance (represented as DR) and the exponent exp (�FEp/RT) of the peak potential for CO stripping as a function of heat-treatmenttemperature. Bottom: Methanol-oxidation current iMOR at 1000 s and extent ofalloying for Pt (JPt) as a function of heat-treatment temperature.

64 M. Tsypkin et al. / Journal of Electroanalytical Chemistry 704 (2013) 57–66

These numbers are reasonably consistent with the findings of Chenet al. [80] for similar, but differently produced, Ru@Pt catalysts. Re-sults of Schlapka et al. [81] indicate that electronic effects ar oper-ative at shell thicknesses much larger than this, though. A similarcorrelation to that of Fig. 11 was observed by Shao et al. [77,82]for bond lengths in Pd–Fe catalysts.

The results in Figs. 9 and 11 show that the Ru@Pt catalysts dis-play a significantly improved activity for oxidation of adsorbed COrelative to that of Pt and only slightly less than that of PtRu alloys.The activity for methanol oxidation on the other hand is inferior tothose of PtRu alloy catalysts and similar to that of pure Pt. The oxi-dation of adsorbed CO (stripping) and of (dissolved) methanol arethus not uniquely correlated at the Ru@Pt catalysts as has been re-cently pointed out also for PtRu nanocatalysts [29,75], in spite ofthe central role of CO in the methanol oxidation reaction [72].

The low extent of alloying and the distinct electrochemicalbehaviour presented above are consistent with the pristine catalystbeing of core–shell type with Pt in the shell, but these data are notconsistent with the possible alternative structures that might beconsidered. The data are not consistent with a PtRu alloy structuresince the extent of alloying is far too low (Fig. 11), the voltammo-grams are too different from those of alloys (Fig. 8), and the poten-tial of the CO-stripping peak is not the same. The data are notcompatible with the catalyst consisting of separate Ru and Pt par-ticles, since the extent of alloying is too high, the voltammograms

are too different from those of Pt, and the CO-stripping peak is nei-ther consistent with that of Pt/C nor with that of Ru/C. The dataalso indicate the Ru is not forming the shell (and Pt the core), sincewe would not expect any hydrogen desorption peaks at all in thevoltammograms (Fig. 8) and and the CO-stripping peak appearsat a potential slightly too high (Fig. 4). All XRD results are also con-sistent with these conclusions.

The results indicate that the surface concentration of Ru in theRu@Pt particles increases significantly for the highest heat-treat-ment temperature. This is most obvious from the change in theextent of alloying, Fig. 11. It is, however, also supported by thechanges in bond-length (Fig. 11), since if Ru enters the shell thiswould be expected to lead a compression of Pt–Pt bond lengths.The XRD analysis, Table 3, also indicate an alloy phase at thehighest heat-treatment temperature.

The electrochemical data indicate that the surface concentra-tion of Ru in the pristine catalyst is very low. This is because thesurface concentration of Ru at which the rate of the methanol-oxi-dation reaction reaches its maximum is expected to be below 10%at room temperature [7,5]. Since no maximum is seen in the rate inFig. 11, the pristine catalyst must have a surface concentration wellbelow that corresponding to the maximum, i. e. 10%. (This assumesthat there is no maximum between 300 �C and 500 �C. Thecorresponding maximum for CO is expected to be around 50%[6,7], therefore imposing a less strict criterion for the surfaceconcentration of Ru.)

Our results offer no direct explanation for the low activity forthe methanol oxidation reaction in the pristine catalyst. However,the results indicate that the presence of Ru oxides or hydrousoxides [19,20,23–29] or ruthenium metal in the surface and notonly beneath it is critical for the activity for methanol oxidation.

M. Tsypkin et al. / Journal of Electroanalytical Chemistry 704 (2013) 57–66 65

We have neglected any dependence of coordination numbersand bond lengths on the applied potential in the discussion above.For Pt and decorated Pt catalysts there appears to be little changein the bond length with applied potential [83,84], however. ThePt-O coordination number becomes finite at potential at which Ptoxides are expected to form at the expense of the Pt–Pt coordina-tion number [83,85]. In this work the electrochemical processesinvestigated occur at rather low potentials. We therefore expectthe structural characterization presented above to produce therelevant parameters to be correlated with the electrochemicalperformance.

4. Conclusions

Heat-treatment of Ru@Pt core–shell particles at high tempera-tures (500 �C) led to a significant increase in the extent of alloyingand to some shortening of Pt–Pt bond lengths. This indicates theinterdiffusion of Ru and Pt and a partial alloying of these elementsin the particles. Changes at 300 �C are much more moderate, andcould only be observed through a small change in bond-lengths.Growth of the Ru cores is suppressed in the Ru@Pt core–shell par-ticles as compared to the uncoated Ru particles, on the other hand.Thus, coating the Ru core with a Pt-shell appears to infer substan-tial stability to the catalyst’s cores as compared to the uncoatedcase.

The catalytic activity for oxidation of adsorbed CO, as inferredfrom the peak potential for the process, increases in the entireheat-treatment interval. For methanol the activity for catalystsheat-treated at 500 �C outperforms that of the pristine core–shellcatalyst and those heat-treated at lower temperatures completely.For lower heat-treatment temperatures there is little change in thecatalytic activity for the methanol oxidation reaction, however.Comparison with results for Ru–Pt catalysts with well-defined sur-face concentrations of ruthenium leads to the conclusion that thesurface concentration of ruthenium in the pristine Ru@Pt catalystmust be small, in line with previous results. The Ru@Pt catalystsdisplay a significantly improved activity for oxidation of adsorbedCO, but the activity for methanol oxidation is inferior to those ofPtRu alloy catalysts even for heat-treated catalysts.

Acknowledgments

Work at NTNU was supported by The Norwegian ResearchCouncil through the NANOMAT program, contract no 182044‘‘Oxidation of small organic molecules’’. John Walmsley fromSINTEF, Materials and Chemistry is acknowledged for assistancewith TEM data collection. We acknowledge the EuropeanSynchrotron Radiation Facility for provision of synchrotronradiation facilities and we would like to thank Herman Emerichfor assistance in using beamline BM01B.

References

[1] M. Watanabe, S. Motoo, J. Electroanal. Chem. 60 (1975) 267–273.[2] C. Roth, A.J. Papworth, I. Hussain, R.J. Nichols, D.J. Schiffrin, J. Electroanal.

Chem. 581 (2005) 79–85.[3] A. Honji, L.U. Gron, J.-R. Chang, B.C. Gates, Langmuir 6 (1992) 2716–2719.[4] J.A. Rodriguez, Surf. Sci. Rep. 24 (1996) 225–287.[5] H.A. Gasteiger, N. Markovic, P.N. Ross, E.J. Cairns, J. Phys. Chem. 97 (1993)

12020–12029.[6] H.A. Gasteiger, N. Markovic, P.N. Ross, E.J. Cairns, J. Phys. Chem. 98 (1994) 617–

625.[7] P.N. Ross, in: J. Lipkowski, P.N. Ross (Eds.), Electrocatalysis, Wiley-VCH, New

York, 1998, pp. 43–74.[8] F. Maillard, S. Pronkin, E.R. Savinova, in: M.T.M. Koper (Ed.), Fuel Cell Catalysis

A Surface Science Approach, John Wiley & Sons, Hoboken and New Jersey,2009, pp. 507–566.

[9] K.J.J. Mayrhofer, B.B. Blizanac, M. Arenz, V.R. Stamenkovic, P.N. Ross, N.M.Markovic, J. Phys. Chem. B 109 (2005) 14433–14440.

[10] J.M. Feliú, E. Herrero, V. Climent, in: E. Santos, W. Schmickler (Eds.), Catalysisin Electrochemistry: From Fundamentals to Strategies for Fuel CellDevelopment, John Wiley & Sons, Hoboken and New Jersey, 2011, pp. 127–163.

[11] A. López-Cudero, J. Solla-Gullón, E. Herrero, A. Aldaz, J.M. Feliu, J. Electroanal.Chem. 644 (2010) 117–126.

[12] M.T.M. Koper, Nanoscale 3 (2011) 2054–2073.[13] F. Maillard, S. Schreier, M. Hanzlik, E.R. Savinova, S. Weinkauf, U. Stimming,

Phys. Chem. Chem. Phys. 7 (2005) 385–393.[14] G. Wu, Y.S. Chen, B.Q. Xu, Electrochem. Commun. 7 (2005) 1237–1243.[15] G. Wu, B.-Q. Xu, J. Power Sources 148 (2007) 158–174.[16] E.S. Steigerwalt, G.A. Deluga, C.M. Lukehart, J. Phys. Chem. B 106 (2002) 760–

766.[17] M. Carmo, V.A. Paganin, J.M. Rosolen, E.R. Gonzalez, J. Power Sources 142

(2005) 169–176.[18] J. Nakamura, in: T. Okado, M. Kaneko (Eds.), Molecular Catalysts for Energy

Conversion, Springer Series in Materialc Science, vol. 111, Springer-Verlag,Berlin Heidelberg, 2009, pp. 185–197.

[19] D.R. Rolison, P.L. Hagans, K.E. Swider, J.W. Long, Langmuir 15 (1999) 774–779.[20] J.W. Long, R.M. Stroud, K.E. Swider-Lyons, D.R. Rolison, J. Phys. Chem. B 104

(2000) 9772–9776.[21] K.E. Swider, C.I. Merzbacher, P.L. Hagans, D.R. Rolison, Chem. Mater. 9 (1997)

1248–1255.[22] K.E. Swider, C.I. Merzbacher, P.L. Hagans, D.R. Rolison, J. Non-crystalline Sol.

225 (1998) 348–352.[23] S.y. Huang, S.m. Chang, C.l. Lin, C.h. Chen, C.t. Yeh, J. Phys. Chem. B 110 (2006)

23300–23305.[24] S.y. Huang, S.m. Chang, K.w. Wang, C.t. Yeh, ChemPhysChem 8 (2007) 1774–

1777.[25] Q. Lu, B. Yang, L. Zhuang, J. Lu, J. Phys. Chem. B 109 (2005) 1715–1722.[26] Q. Lu, B. Yang, L. Zhuang, J. Lu, J. Phys. Chem. B 109 (2005) 8873–8879.[27] K. Lasch, L. Jörissen, K.A. Friedrich, J. Garche, J. Solid State Electrochem. 7

(2003) 619–625.[28] H.M. Villullas, F.I. Mattos-Costa, L.O.S.B. oes, J. Phys. Chem. B 108 (2004)

12898–12903.[29] D.R.M. Godoi, J. Perez, H.M. Villullas, J. Phys. Chem. C 113 (2009) 8518–8525.[30] H. Kim, I.R. de Moraes, G. Tremiliosi-Filho, R. Haasch, A. Wieckowski, Surf. Sci.

Lett. 474 (2001) L203–L212.[31] W.E. O’Grady, P.L. Hagans, K.I. Pandya, D.L. Maricle, Langmuir 17 (2001) 3047–

3050.[32] R. Viswanathan, G. Hou, R. Liu, S.R. Bare, F. Modica, G. Mickelson, C.U. Segre, N.

Leyarovska, E.S. Smotkin, J. Phys. Chem. B 106 (2002) 3458–3465.[33] A.H.C. Sirk, J.M. Hill, S.K.Y. Kung, V.I. Birss, J. Phys. Chem. B 108 (2004) 689–

695.[34] P.K. Babu, H.S. Kim, S.T. Kuk, J.H. Chung, E. Oldfield, A. Wieckowski, E.S.

Smotkin, J. Phys. Chem. B 109 (2005) 17912–17916.[35] C. Bock, M.-A. Blakely, B. MacDougall, Electrochim. Acta 50 (2005) 2401–2414.[36] Y.-C. Wei, C.-W. Liu, K.-W. Wang, ChemPhysChem 10 (2009) 1230–1237.[37] Y.-C. Wei, C.-W. Liu, W.-J. Chang, K.-W. Wang, J. Alloys Comp. 509 (2011) 535–

541.[38] M.S. Nashner, A.I. Frenkel, D. Somerville, C.W. Hills, J.R. Shapley, R.G. Nuzzo, J.

Am. Chem. Soc. 120 (1998) 8093–8101.[39] J. Luo, L. Wang, D. Mott, P.N. Njoki, Y. Lin, T. He, Z. Xu, B.N. Wanjana, I.-I.S. Lim,

C.-J. Zhong, Adv. Mater. 20 (2008) 4342–4347.[40] K.J.J. Mayrhofer, V. Juhart, K. Hartl, M. Hanzlik, M. Arenz, Angew. Chem. Int. Ed.

48 (2009) 3529–3537.[41] S. Alayoglu, A.U. Nilekar, M. Mavrikakis, B. Eichhorn, Nat. Mater. 7 (2008) 333–

338.[42] S. Alayoglu, P. Zavalij, B. Eichhorn, Q. Wang, A.I. Frenkel, P. Chupas, ACS Nano 3

(10) (2009) 3127–3137.[43] P. Ochal, J.L.G. de la Fuente, M. Tsypkin, F. Seland, S. Sunde, N. Muthuswamy,

M. Rønning, D. Chen, S. Garcia, S. Alayoglu, B. Eichhorn, J. Electroanal. Chem.655 (2011) 140–146.

[44] M. Bernechea, S. García-Rodríguez, P. Terreros, E. de Jesús, J.L.G. Fierro, S. Rojas,J. Phys. Chem. 115 (2011) 1287–1294.

[45] P. Mathew, J.P. Meyers, R. Srivastava, P. Strasser, J. Electrochem. Soc. 159(2012) B554–B563.

[46] P. Lara, M.-J. Casanove, P. Lecante, P.-F. Fazzini, K. Philippot, B. Chaudret, J.Mater. Chem. 22 (2012) 3578–3584.

[47] D. Bokach, J.G. de la Fuente, M. Tsypkin, P. Ochal, I.C. Endsjø, R. Tunold, S.Sunde, F. Seland, Fuel Cells 11 (2011) 735–744.

[48] D.-G. Liu, J.-F. Lee, M.-T. Tang, J. Mol. Catal. A 240 (2005) 197–206.[49] B.J. Hwang, L.S. Sarma, J.-M. Chen, C.-H. Chen, S.-C. Shih, G.-R. Wang, D.-G. Liu,

J.-F. Lee, M.-T. Tang, J. Am. Chem. Soc. 127 (2005) 11140–11145.[50] B.J. Hwang, C.-H. Chen, L.S. Sarma, J.-M. Chen, G.-R. Wang, M.-T. Tang, D.-G. Liu,

J.-F. Lee, J. Phys. Chem. B 110 (2006) 6475–6482.[51] G.A. Camara, M.J. Giz, V.A. Paganin, E.A. Ticianelli, J. Electroanal. Chem. 537

(2002) 21–29.[52] M. Mazurek, N. Benker, C. Roth, T. Buhrmester, H. Fuess, Fuel Cells 1 (2006) 16–

20.[53] C. Roth, N. Benker, M. Mazurek, F. Scheiba, H. Fuess, Appl. Catal. A 319 (2007)

81–90.[54] M.S. Nashner, A.I. Frenkel, D.L. Adler, J.R. Shapley, R.G. Nuzzo, J. Am. Chem. Soc.

119 (1997) 7760–7771.[55] G.S. Pawley, J. Appl. Crystal. 14 (1981) 357–361.[56] http://cars9.uchicago.edu/ifeffit/.

66 M. Tsypkin et al. / Journal of Electroanalytical Chemistry 704 (2013) 57–66

[57] F. Lytle, D. Sayers, E. Stern, Physica B 158 (1989) 701–722.[58] S. Hasnain (Ed.), Report of the International Workshop on Standards

and Criteria in X-ray Absorption Spectroscopy, Ellis Horwood, Chicester,1991.

[59] M. Vaarkamp, Catal. Today 39 (1998) 271–279.[60] A.I. Frenkel, C.W. Hills, R.G. Nuzzo, J. Phys. Chem. B 105 (2001).[61] A.M. Karim, V. Prasad, G. Mpourmpakis, W.W. Lonergan, A.I. Frenkel, J.G. Chen,

D.G. Vlachos, J. Am. Chem. Soc. 131 (2009) 12230–12239.[62] T.J. Schmidt, H.A. Gasteiger, G.D. Stab, P.M. Urban, D.M. Kolb, R.J. Behm, J.

Electrochem. Soc. 145 (1998) 2354–2358.[63] P. Ochal, J.L.G. de la Fuente, M. Tsypkin, S. Garcia-Rodriguez, F. Seland, S.

Sunde, ECS Trans. 28 (2010) 9–17.[64] R.W.G. Wyckoff, Crystal Structures, 2 ed., Interscience Publishers, New York,

1963.[65] Y.-C. Hsieh, L.-C. Chang, P.-W. Wu, Y.-M. Chang, J.-F. Lee, Appl. Catal. B:

Environ. 103 (2011) 116–127.[66] J. Lee, W.B. Wang, M.S. Zei, G. Ertl, Phys. Chem. Chem. Phys 4 (2002) 1393–

1397.[67] S. Brankovic, N. Marinkovic, J. Wang, R. Adzic, J. Electroanal. Chem. 532 (2002)

57–66.[68] S. S�en, F. S�en, G. Gökag�ç, Phys. Chem. Chem. Phys. 13 (2011) 6784–6792.[69] W.P. Davey, Phys. Rev. 25 (1925) 753–761.[70] C.-H. Chen, L.S. Sarma, D.-Y. Wang, F.-J. Lai, C.-C.A. Andra, S.-H. Chang,

D.-G. Liu, C.-C. Chen, J.-F. Lee, B.-J. Hwang, ChemCatChem 2 (2010) 159–166.

[71] H.E. Hoster, R.J. Behm, in: M.T.M. Koper (Ed.), Fuel Cell Catalysis. A SurfaceScience Approach, John Wiley & Sons, Hoboken and New Jersey, 2009, pp. 465–505.

[72] M.T.M. Koper, S.C.S. Lai, E. Herrero, in: M.T.M. Koper (Ed.), Fuel Cell Catalysis. ASurface Science Approach, John Wiley & Sons, Hoboken and New Jersey, 2009,pp. 159–207.

[73] E. Herrero, B. Álvarez, J.M. Feliu, S. Blais, Z. Radovic-Hrapovic, G. Jerkiewicz, J.Electroanal. Chem 567 (2004) 139–149.

[74] A. Velázquez-Palenzuela, E. Brillas, C. Arias, F. Centellas, J.A. Garrido, R.M.Rodriguez, P.L. Cabot, J. Phys. Chem. C 116 (2012) 18469–18478.

[75] D.R.M. Godoi, H.M. Villullas, Langmuir 28 (2012) 1064–1067.[76] B. Hammer, Top. Catal. 37 (2006) 3–16.[77] Y. Xu, M. Shao, M. Mavrikakis, R.R. Adzic, in: M.T.M. Koper (Ed.), Fuel Cell

Catalysis. A Surface Science Approach, John Wiley & Sons, Hoboken and NewJersey, 2009, pp. 271–315.

[78] A. Ruban, B. Hammer, P. Stoltze, H.L. Skriver, J.K. Nørskov, J. Mol. Catal. A:Chem. 115 (1997) 421–429.

[79] J.C. Davies, J. Bonde, A. Logadóttir, J.K. Nørskov, I. Chorkendorff, Fuel Cells 5(2005) 429–435.

[80] T.-Y. Chen, T.-L. Lin, T.-J.M. Luo, Y. Choi, J.-F. Lee, ChemPhysChem 11 (2010)2383–2392.

[81] A. Schlapka, M. Lischka, A. Groß, U. Käsberger, P. Jakob, Phys. Rev. Lett. 91(2003) 016101-1–016101-4.

[82] M.H. Shao, K. Sasaki, R.R. Adzic, J. Am. Chem. Soc. 128 (2006) 3526–3527.[83] V. Croze, F. Ettingshausen, J. Melke, M. Soehn, D. Stuermer, C. Roth, J. Appl.

Electrochem. 40 (2010) 877–883.[84] A. Rose, E.M. Crabb, Y. Qian, M.K. Ravikumar, P.P. Wells, R.J.K. Wiltshire, J. Yao,

R. Bilsborrow, F. Mosselmans, A.E. Russell, Electrochim. Acta 52 (2007) 5556–5564.

[85] B.J. Hwang, Y.W. Tsai, J.F. Lee, P. Borthen, H.H. Strehblow, J. Synchrotron Rad. 8(2001) 484–486.