DOI: 10.1002/cphc.200500646 Hydrogen Electrocatalysis** Ludwig A. Kibler*[a] A selection of recent theoretical and experimental studies on electrolytic hydrogen evolution is presented. It is demonstrated with well-defined model surfaces that this reaction is a very structuresensitive process. Crystallographic orientation, defect density and surface composition are parameters that determine the local geometric and electronic surface structure, and are thus crucial for the electrocatalytic activity as characterised by the exchange current density. The observed trends can be understood within a recent theory by J. K. Nørskov et al., which is based on density functional calculations and which emphasises the impact of hydrogen chemisorption energies on the reaction rate, that is, on the exchange current density. The particular electrocatalytic activities of ultrathin metal films and of nanostructures are addressed. 1. Introduction laboratory related to the hydrogen reaction on various metal surfaces of different structure and composition. The reversible electrochemical reaction in Equation (1): Hþ ðaqÞ þ e ! 1=2 H2ðgÞ E 0 ¼ 0 V versus the standard hydrogen electrode ðSHEÞ ð1Þ needs to be catalysed by an electronic conductor, that is, it takes place at an electrode surface. A huge variety of possible materials exist, including metals, semiconductors and enzymes, the active centres of the latter also contain metal atoms, such as Ni or Fe.[1] It is well-known that the electrocatalytic activity of an electrode, which determines the reaction rate at a certain potential, depends strongly on its chemical composition. This dependence has often been reduced to the question of electrode material. However, it has not been fully understood why various materials show differences in activity and, moreover, how structural and electronic properties of an electrode surface impact the reaction rate. While progress in catalysis and electrocatalysis still often relies on empirical laws and observations, a main current interest lies in a basic understanding of fundamental relations between an electrode’s structure (geometric and electronic) and its activity for a given reaction, in order to set up systematic trends. In order to get information at an atomic level, one of the strategies is to combine theoretical with experimental work for a reaction as simple as possible, taking place at the interface of a clean and structurally well-defined electrode surface in contact with an electrolyte free from contaminants. The hydrogen reaction, Equation (1), certainly belongs to the most widely studied electrochemical reactions, first because of its technological importance, and secondly because it is supposed to be one of the apparently simple electrocatalytic reactions that are suited for model studies, for example, with singlecrystal surfaces. This Minireview shall give a survey on selected literature studies, as well as on recent results obtained in the author’s ChemPhysChem 2006, 7, 985 – 991 2. Reaction Kinetics It is generally accepted that there are three possible reaction steps in the hydrogen evolution reaction (HER) in acidic media:[2] 1) Volmer or discharge reaction in Equation (2): Hþ ðaqÞ þ e ! Had ð2Þ 2) Tafel or combination reaction in Equation (3): 2 Had ! H2,ad ð3Þ 3) Heyrowsky or ion + atom reaction in Equation (4): Hþ ðaqÞ þ Had ! H2,ad ð4Þ The catalytic action of the surface is to adsorb hydrogen atoms Had after reaction of protons from solution with electrons from the electrode, and to allow for the formation of molecular hydrogen gas—or vice versa for the hydrogen oxidation reaction (HOR). Free adsorption sites, which are not included in the Equations (2–4), are needed for binding hydrogen at the electrode surface. For constant activity of protons, constant fugacity of molecular hydrogen and constant temper[a] Dr. L. A. Kibler Abteilung Elektrochemie, Universitt Ulm 89069 Ulm (Germany) Fax: (+ 49) 731-502-5409 E-mail: [email protected] [**] Presented in part at the 3rd Gerischer Symposium, “Electrocatalysis: Theory and Experiment”, held at the Harnack-Haus of the Max-Planck-Gesellschaft, Berlin (Germany), July 2005. 9 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 985 L. A. Kibler ature, the electrocatalytic activity is crucially dependent on composition and surface structure of the electrode, and it is usually characterised by the exchange current density, j0, which is proportional to the reaction rate. The exchange current density refers to the anodic or cathodic current density flowing under conditions of dynamic equilibrium, and is defined within the famous Butler–Volmer equation, Equation (5),[3] which relates the reaction current density j to the overpotential h: aFh ð1 aÞFh j ¼ j0 exp exp RT RT “thermoneutral”, that is, DadG0 is near zero. This is especially true for bare platinum and, interestingly, for the active centres in hydrogen-producing enzymes, as shown in Figure 1 with ð5Þ In the following, we want to illuminate the physical origins of apparent j0 values ranging over several orders of magnitude (see Table 1), rather than analysing the reaction mechanism and the meaning of the symmetry factor a (which is usually around 0.5) in greater detail. Figure 1. Free-energy diagram for the HER at various metal surfaces and enzyme models, under standard and equilibrium conditions. Electrocatalysts with free energies of hydrogen adsorption close to zero belong to the more active ones. Courtesy of J. K. Nørskov et al.[11] 3. Volcano Plots As for many other reactions, kinetic data for the HER were empirically found to correlate with thermodynamic properties of related systems in a simple way. For example, a logarithmic plot of j0 for various electrodes versus the respective metal–hydrogen bond strength reveals a so-called volcano curve.[7] The shape of such a volcano plot has been rationalised independently by Gerischer and Parsons,[8, 9] who deduced that j0 is related to the standard free-energy for hydrogen adsorption DadG0. The two linear branches of the volcano are symmetric to DadG0 = 0. This is an example of a so-called linear free-energy relationship, since the logarithm of the reaction rate given by log(j0) is proportional to the free enthalpy of activation, according to transition-state theory. For experimental volcano plots, however, data for adsorption enthalpies from the gas phase reaction have been used, rather than free adsorption enthalpies in solution. In addition, kinetic data taken under various experimental conditions just allow the demonstration of a qualitative trend. Consistent data for DadG0 from both theory and experiment and carefully measured j0 values still give a volcano curve (Section 6.1 and Figure 6) that provides a more quantitative relationship. A consistent set of hydrogen chemisorption energies DEH for 0.25 hydrogen coverage on various metal surfaces was recently obtained by Nørskov et al. by density functional calculations on periodically repeated metal slabs.[10] The free energies of hydrogen adsorption DadG0 were estimated to be given by Equation (6): Dad G0 ¼ DE H þ DE ZPE T DSH ¼ DE H þ 0:24 eV ð6Þ where DEZPE is the difference in zero point energy between adsorbed and gas phase hydrogen and DSH is the entropy of hydrogen adsorption. With the assumption that the presence of solvent molecules does not affect the given trends, the authors demonstrated that the very active electrocatalysts for the HER bind the intermediate atomic hydrogen not too strongly and not too weakly, and allow the overall reaction to be quasi 986 www.chemphyschem.org models for a hydrogenase and a nitrogenase. Starting from these observations, these authors predicted a specific MoS2 structure to be of high catalytic activity, and this was successfully shown in a subsequent experiment with MoS2 nanoparticles.[11] Moreover, these authors developed a simple kinetic model, which explains the two branches of a volcano curve, which are clearly separated at DadG0 = 0 (Figure 2). The reaction rate Figure 2. Measured (top) and calculated (bottom) j0 for HER on various metals as a function of the calculated hydrogen chemisorption energy DEH (top) or of the free energy of hydrogen adsorption DadG0 (bottom). Courtesy of J. K. Nørskov et al.[10] under equilibrium conditions at pH 0 was taken to be proportional to a rate constant k1 and proportional to (1q), where q is the hydrogen coverage. With Equation (7): K¼ q Dad G0 ¼ exp 1q kT 9 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ð7Þ ChemPhysChem 2006, 7, 985 – 991 Hydrogen Electrocatalysis and the assumption that DadG0 can be also viewed as activation energy for proton reduction, j0 can be written as Equation (8): j0 ¼ eNk0 K ðDad G0 0Þ 1þK ð8Þ where e is the elementary charge, k0 = 200 s1 site1 is the only parameter fitted to experimental values, and N is the density of sites. Equation (8) explains the branch on the right of the volcano for positive values of DadG0 in Figure 2. The branch on the left is obtained for zero activation energy, leading to k1 = k0 and Equation (9): j0 ¼ eNk0 1 ðDad G0 < 0Þ 1þK ð9Þ In essence, the free energy of hydrogen adsorption is wellsuited for a correlation with exchange current densities. flat gold surfaces in acidic solution. A further advantage of studying HER on gold single-crystal surfaces is the rather detailed knowledge of local structural properties, as obtained by in situ scanning tunnelling microscopy (STM).[13] However, available literature values for j0 of HER on the low-index faces of gold show a large scatter.[15] Although different acidic solutions have been used by various groups, the presence of anions like sulphate or perchlorate should not play an important role in the hydrogen evolution region. However, the details of the surface structure were often not characterised or considered. Since electrode preparation is crucial for the quality of the surface and the distribution of various active sites, it is important to carefully control and specify the surface structure, as will be demonstrated in the following. Figure 3 shows current–potential curves at 5 mV s1 for the HER at room temperature on AuACHTUNGRE(100), AuACHTUNGRE(111) and AuACHTUNGRE(110), measured under clean conditions and for thermally recon- 4. Measurements of Exchange Current Density It is still a challenge to measure reliable values of j0 for clean and structurally uniform electrode surfaces comparable to the model surfaces used in calculations. Two main experimental problems make a correlation with theory difficult. First, the presence of defects, for example, monoatomic high steps, leads to certain structural and electronic heterogeneity of the surface. As will be shown below (Section 6.2), differences in the metal–hydrogen bonding for each type of adsorption site might be important for the overall activity. As a consequence, variation of the defect density leads to changes in activity for the HER/HOR. Secondly, mass transport is a critical issue for the very active surfaces, especially for HOR. It has recently been shown by the use of small platinum particles, sub-micrometer in size, that j0 of platinum may often been underestimated by more than one order of magnitude.[12] Catalytic effects observed with nanoparticles are discussed in Section 7. Besides the use of microstructures, rotating disk electrodes or flow cells are used by various research groups to avoid mass-transport limitations and to ensure the determination of kinetic data for the hydrogen reaction. In any case, it is reasonable to compare kinetic data, which are gathered by the same methodology, in order to obtain reliable trends that are at least qualitative. The j0 values presented in the following were extracted from simple voltammetric scans performed at room temperature and at a slow scan rate, by fitting the data obtained at overpotentials negative of 0.15 V to the cathodic term of Equation (5) and correcting for the solution resistance. Effects of diffusion have been neglected, because varying the sweep rate leads to similar results. 5. Single-Crystal Surfaces Gold does certainly not belong to the very active materials for the hydrogen reaction, which is in contrast to platinum, rhodium and iridium. However, just for that reason, there should be no big problems with mass-transport limitation for the HER on ChemPhysChem 2006, 7, 985 – 991 Figure 3. Current–potential curves for HER on various reconstructed Au single-crystal electrode surfaces: AuACHTUNGRE(100), AuACHTUNGRE(111) and AuACHTUNGRE(110), in 0.1 m H2SO4 at room temperature. Scan rate v = 5 mV s1. structed surfaces. The catalytic activity increases in the given order (j0 = 0.05, 0.12 and 0.4 mA cm2, respectively), although all three surfaces have a hexagonally densely packed surface. This proves the HER is a very structure-sensitive reaction. It is important to note that these kinetic data are strongly dependent on the quality of each individual crystal, and are only meaningful for careful surface preparation. Three structural modifications shall be discussed, which give rise to a significant change in activity. 1) After formation of the (1 K 1)-structures and subsequent complete potential-induced reconstruction,[13] the activities were about twice as high as those of the (freshly prepared) thermally reconstructed surfaces. Since the unit cells are identical for each surface, it is obvious that the smaller domains and the presence of domain boundaries enhance the HER activity. 2) The use of stepped AuACHTUNGRE(111) surfaces shows that the density of defects, such as monoatomic high steps, plays an important role as well. A systematic increase in the defect density by taking surfaces with a miscut leads to larger j0 values, which approach the activity of AuACHTUNGRE(110). 3) It is not easy 9 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org 987 L. A. Kibler to investigate the HER activity of the unreconstructed Au surfaces, because of potential-induced reconstruction at negative potentials. However, preliminary potential step experiments with AuACHTUNGRE(100) show that the (1 K 1) structure of AuACHTUNGRE(100) has a HER activity about one order of magnitude higher than the thermally reconstructed surface (compare the different surface structures for the STM images in Figure 4). various authors.[6, 18] As will be shown below (Section 6), it can be helpful to take H UPD into account for explaining HER activities. Marković has determined kinetic parameters for HER over the three low-index faces of platinum by measuring close to the equilibrium potential with a rotating disk electrode configuration.[6] The j0 values at room temperature in 0.05 m H2SO4 were found to be around 350, 450 and 900 mA cm2 for PtACHTUNGRE(111), PtACHTUNGRE(100) and PtACHTUNGRE(110), respectively. Provided that mass-transport limitation does not play an important role (see ref. [12] for a detailed discussion), it remains an open issue how surface defects contribute to the overall activities. Nevertheless, the impact of crystallographic orientation on the HER activity has been clearly demonstrated for Pt electrodes. 6. Palladium Electrodes Figure 4. STM images of a AuACHTUNGRE(100) electrode surface in 0.1 m H2SO4 revealing a) the (hex)-reconstruction at 0.2 V and b) the island-covered unreconstructed surface at 0.35 V. Courtesy of Dakkouri and Kolb.[14] An amazing observation is the fact that the reconstructed AuACHTUNGRE(111) surface is more than double as active as the reconstructed AuACHTUNGRE(100) surface (see j0 values in Table 1), although the arrangement of surface atoms is practically the same; a fact which is also reflected in the similar values for the potential of zero charge.[16] It would be interesting to learn more about the local electronic structures of these surfaces, including the effect of the underlying Au bulk structure and the consequences on hydrogen binding energies. In contrast to gold electrodes, hydrogen adsorbs already at potentials more positive than the equilibrium potential on platinum and palladium surfaces and thus belongs to the socalled underpotential deposition (UPD) phenomena. This Volmer-type reaction, see Equation (2), is known to be extremely structure-sensitive for many noble metal surfaces, and often serves as to characterise the metal/electrolyte interface in terms of surface quality and cleanliness.[17] The relevance of UPD hydrogen on platinum has recently been discussed by While platinum is supposed to be the more prominent catalyst material, one of the advantages of palladium for model studies is the easy fabrication of monolayers,[19] sub-monolayer islands and nanoclusters,[20] the HER activities of which shall be discussed in the following. These various structures are especially interesting because of the absence of hydrogen absorption. This reaction often hampers the use of massive Pd electrodes, especially in the case of single crystals, because their bulk structure may suffer from lattice expansion during hydrogen incorporation. The direct participation of absorbed hydrogen in the HER mechanism shall not be addressed here. 6.1. Palladium Monolayers It is plausible that Pd monolayers deposited on foreign substrates do not absorb hydrogen into their bulk structure. It was shown that this reaction only occurs for ultrathin deposits with a thickness larger than two monolayers.[19a] A systematic study of hydrogen adsorption on Pd monolayers deposited onto several hexagonally close-packed noble metal surfaces has revealed that the adsorption potential is strongly influenced by the chemical nature of the substrate (Figure 5).[21] The fact that the properties of such monolayers are different from the properties of the bulk material is explained within the so-called dband model introduced by Hammer and Nørskov.[22] Accordingly, an up-shift of the centre of the d-band leads to a stronger interaction with adsorbates like hydrogen, or vice versa.[23] Both geometric and electronic Table 1. Values of j0 for HER measured over noble metal single-crystal surfaces at room temperature. effects may contribute to such a d-band shift.[24] For Pd monolayElectrolyte Reference Electrode j0 [mA cm2] ers, which often grow pseudothermally reconstructed 0.12 0.1 m H2SO4 [4] Au(111) morphically, it has been ob[4] potential reconstructed 0.20 0.1 m H2SO4 [4] thermally reconstructed 0.05 0.1 m H2SO4 served that lateral compression Au(100) [4] potential reconstructed 0.10 0.1 m H2SO4 leads to a systematic lowering of [4] AuACHTUNGRE(110) thermally reconstructed 0.4 0.1 m H2SO4 the characteristic hydrogen ad[4] RuACHTUNGRE(0001) 3 0.1 m H2SO4 sorption peak.[21] These adsorp[5] AgACHTUNGRE(111) 10 0.1 m H2SO4 PtACHTUNGRE(111) 350 tion potentials, and accordingly [6] PtACHTUNGRE(100) 450 0.05 m H2SO4 the experimentally obtained free PtACHTUNGRE(110) 900 adsorption energies for UPD hy- 988 www.chemphyschem.org 9 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2006, 7, 985 – 991 Hydrogen Electrocatalysis monolayer on a PtRuACHTUNGRE(111) surface, might be underestimated by experiment due to diffusion limitation. This is because j0 values were extracted from simple voltammetric scans, which is not an ideal method for highly active surfaces. It has also to be pointed out that the presence of surface defects may lead to variations in the HER activity. Still, Figure 6 highlights the fact that j0 can be deduced from a calculated set of free energies for hydrogen adsorption.[25] The systematic variation of these free energies for hydrogen on the various Pd monolayers was also found in the experiment.[21] 6.2. Palladium Multilayers By increasing the overlayer’s thickness, the electronic modification of the Pd surface by the Figure 5. Top) Current–potential curves for hydrogen adsorption on Pd monolayers deposited on various subunderlying substrate should dis1 strates in 0.1 m H2SO4 ; v = 10 mV s . Bottom) Scheme for the d-band model shows a down-shift or an up-shift of appear, that is, the so-called the d-band centre for compressive or tensile strain within the pseudomorphic Pd monolayers, respectively.[21] ligand effect vanishes. If the Pd multilayer is still pseudomorphic, altered properties should arise solely from lateral strain. This drogen, show a linear relation with the calculated shift of the was theoretically demonstrated for hydrogen adsorbed on Aud-band centre of the Pd surface.[21] ACHTUNGRE(111) and on AuACHTUNGRE(100) by Roudgar and Gross,[26] who calculated The effect of tuning the hydrogen binding energy on the HER activity is shown in Figure 6, which very much resembles variations in hydrogen adsorption energy with increasing Pd a volcano, as described above (Section 3).[25] The agreement film thickness. Maximum binding energies were found for an overlayer thickness of two monolayers. Similar changes were between theory and experiment is remarkably good. It is seen that the activity of the apparently most active system, a Pd recently observed for hydrogen adsorption on Pd overlayers on AuACHTUNGRE(100).[27] From the above considerations with Pd monolayers, it was expected that the HER activity for the two monolayer system would be lowered, owing to the larger hydrogen binding energy. This was indeed observed in the experiment.[27] However, j0 for HER on the Pd/AuACHTUNGRE(100) surfaces did not increase for Pd overlayers thicker than two monolayers (Figure 7), although hydrogen was again found to adsorb more weakly. This behaviour can be explained by the abovementioned hydrogen absorption reaction, which occurs for Pd coverages above two monolayers. Apparently, the presence of hydrogen in the Pd bulk lattice lowers the activity for hydrogen evolution. It is shown in Figure 7 that j0 does not change a lot for Pd coverages on AuACHTUNGRE(100) between two and 100 monolayer equivalents. However, for Pd overlayers deposited onto PtACHTUNGRE(100), the HER activity drops down continuously with increasing coverage, and is still much higher for a 100-monolayer film compared with the same amount of Pd on AuACHTUNGRE(100) Figure 6. j0 for hydrogen evolution at pH 0 as a function of the differential free-energy of hydrogen adsorption. The curve shows the theoretical predic(Figure 7). It is not very probable that the substrate influences tion based on a simple theoretical model:[10] ^ correspond to the predictions the surface properties of such a relatively thick overlayer, but it for the various pseudomorphic Pd monolayers PdML/X, where X denotes the is assumed that lattice expansion due to hydrogen absorption respective hexagonally closed-packed substrate; * represent experimentally leads to defects in the Pd multilayer. This could explain the obtained j0 values. Courtesy of J. K. Nørskov and co-workers.[25] ChemPhysChem 2006, 7, 985 – 991 9 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org 989 L. A. Kibler Figure 7. j0 for HER over Pd multilayers on AuACHTUNGRE(100) and on PtACHTUNGRE(100) as a function of coverage. higher HER activities of Pd on PtACHTUNGRE(100), because lattice destruction might partially expose the very active first Pd layer.[6a] Figure 9. j0 for HER over Pd sub-monolayer deposits on AuACHTUNGRE(111) as a function of coverage. Contributions from various active sites in dependence of the Pd coverage are included: 1) uncovered AuACHTUNGRE(111) surface (black), 2) Pd islands (red), 3) Pd/Au sites at the perimeter of the Pd islands (blue) and 4) Pd/Pd step sites (green). The presence of these active sites can only explain the activities of Pd coverages above 0.2 monolayers (violet), while extremely high activities per Pd atom are seen at lower coverage. Experimentally obtained data points (squares) are connected with a dotted line as guide for the eye. 6.3. Palladium Sub-Monolayers The HER activity of Pd monolayers deposited on Au substrates was found to be about two orders of magnitude higher than that of the bare substrates.[27, 28] This is illustrated by the current–potential curves shown in Figure 8. In addition, it is seen 1) the uncovered unreconstructed AuACHTUNGRE(111) surface, the activity of which decreases with increasing Pd coverage qPd, Equation (10): ðiÞ Auð111Þ j0 ¼ j0 Auð111Þ ð1 qPd Þ with j0 1 mA cm2 ð10Þ 2) Pd islands, the activity of which is assumed to be proportional to qPd, Equation (11): ðiiÞ j0 ¼ j0PdML qPd with j0PdML 20 mA cm2 ð11Þ 3) Pd/Au sites at the perimeter of the Pd islands. The number of these sites was calculated on the basis of progressive nucleation and growth as seen by in situ STM[19c] to yield Equation (12): ðiiiÞ 2 j0 ¼ j0PdAu ð1 qPd Þð lnð1 qPd ÞÞ =3 with j0PdAu 100 mA cm2 ð12Þ Figure 8. Current–potential curves for hydrogen evolution in 0.1 m H2SO4 on AuACHTUNGRE(111) and a Pd monolayer deposited onto the same, well-ordered AuACHTUNGRE(111) surface and onto a AuACHTUNGRE(111) surface with a 68 miscut; v = 1 mV s1. that the activity of a full monolayer increases with a higher surface-defect density for Pd on AuACHTUNGRE(111), by using single-crystal surfaces of different miscut. Figure 9 shows that j0 for electrodeposited Pd sub-monolayer amounts on AuACHTUNGRE(111) is practically constant between 0.1 and 1 monolayers and exhibits a maximum below that range of coverage.[28] Based on a simple model, which takes various active sites into consideration, the HER activity can be understood for coverages higher than 0.2 monolayers. These active sites are each related with a characteristic activity and involve: 990 www.chemphyschem.org 4) Pd/Pd step sites over monoatomic Au steps, where a normal distribution was assumed, since these steps are only present near qPd = 1. These four contributions can be predicted from independent measurements. The only two parameters fitted to the experiments are j0PdAu and the activity of Pd/Pd step sites, which was obtained from measurements with Pd monolayers deposited on stepped AuACHTUNGRE(111) electrodes, as shown in Figure 8. The various contributions are plotted into Figure 9 together with their sum. It was assumed that each active site for this bimetallic surface is related to a specific hydrogen binding energy, giving rise to a certain activity for hydrogen evolution, as demonstrated above (Section 3). It is clearly observed that mixed Pd–Au step sites show an enhanced activity compared to terrace sites. 9 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2006, 7, 985 – 991 Hydrogen Electrocatalysis Exchange current densities for HER of several hundreds of mA cm2 were measured for Pd coverages below (!) 0.1 monolayers on Au single-crystal electrodes (not shown here).[27, 28] These extreme electrocatalytic activities were observed after just dipping the Au electrodes into a PdSO4 solution at positive potentials. It is assumed that adsorption of Pd2 + and its subsequent reduction in a Pd2 + -free solution leads to the formation of special Pd/Au sites, the structure of which is still unknown. 7. Palladium Nanostructures Extraordinarily high electrocatalytic activities for HER have also been reported for single Pd nanoparticles supported on a AuACHTUNGRE(111) substrate as measured with an STM tip.[29] On the basis of a kinetic model, it was concluded that molecular hydrogen may spill over from the Pd particle and diffuse to the bare Au surface.[30] An alternative explanation for the high activity might be that hydrogen is bound more weakly at a Pd nanocluster on AuACHTUNGRE(111), as calculated by Roudgar and Groß[31] However, it is not clear if the STM tip as used in the experiment is suited for determining the catalytic activity of a nanoparticle.[29] Meanwhile, it is possible to generate large cluster fields of several thousands of Pd particles on AuACHTUNGRE(111) by STM tip-induced cluster formation.[20b] It might be interesting to use the well-defined geometry of a microelectrode in a scanning electrochemical microscope (SECM) to obtain reliable data on the promising electrocatalytic HER activity of Pd nanostructures.[32] 8. Summary and Outlook It has been demonstrated by the use of well-defined singlecrystal surfaces that the HER is an extremely structure-sensitive process. The free energy of hydrogen adsorption is a key parameter in relating surface structure with electrocatalytic activity. The overall activity of a more complex electrode surface is supposed to be determined by the ability of various active sites to chemisorb hydrogen. Defects, such as monoatomic high steps, are often found to be more active than terrace sites. The use of pseudomorphic palladium monolayers allows for a tuning of HER activity over several orders of magnitude, which can be understood with a simple kinetic model. It still remains important to determine experimentally the coverage of hydrogen participating in the HER. A combination of theory and experiment is believed to be necessary for understanding general trends in electrocatalysis. Of special interest are 1) the electrocatalytic activities of nanostructures from both theoretical and experimental points of view and 2) new experimental methods for studying the kinetics of very active surfaces in order to extract reliable values for j0. In the end, a deep understanding of basic principles in electrocatalysis shall allow for the efficient design of new materials. Further studies will be necessary to demonstrate the relevance of findings with single-crystal surfaces for more realistic nanoparticle electrode materials. ChemPhysChem 2006, 7, 985 – 991 Acknowledgements This work was supported by the Fonds der Chemischen Industrie. The author thanks Professor D. M. Kolb for his invaluable help. Keywords: electrochemistry · hydrogen nanostructures · surface chemistry · interfaces · [1] a) A. Volbeda, M. H. Charon, C. Piras, E. C. Hatchikian, M. Frey, J. C. Fontecilla-Camps, Nature 1995, 373, 580; b) J. W. Peters, W. N. Lanzilotta, B. J. Lemon, L. C. Seefeldt, Science 1998, 282, 1853. [2] M. W. Breiter in Handbook of Fuel Cells: Fundamentals, Technology and Applications, Vol. 2 (Eds.: W. Vielstich, A. Lamm, H. A. Gasteiger), Wiley, Chichester, 2003, pp. 361 – 367. [3] K. J. Vetter, Electrochemical Kinetics, Academic Press, New York, 1967. [4] L. A. Kibler, unpublished results. [5] D. Eberhardt, E. Santos, W. Schmickler, J. Electroanal. Chem. 1999, 461, 76. [6] a) N. M. Marcović in Handbook of Fuel Cells: Fundamentals, Technology and Applications, Vol. 2 (Eds.: W. Vielstich, A. Lamm, H. A. Gasteiger), Wiley, Chichester, 2003, pp. 368 – 393; b) N. M. Marcović, P. N. Ross in Interfacial Electrochemistry: Theory, Experiment, and Applications (Ed.: A. Wieckowski), Marcel Dekker, New York, 1999, p. 821. [7] a) S. Trasatti, J. Electroanal. Chem. 1972, 39, 163; b) S. Trasatti, Electrochim. Acta 1994, 39, 1739. [8] H. Gerischer, Bull. Soc. Chim. Belg. 1958, 67, 506. [9] R. Parsons, Trans. Faraday Soc. 1958, 54, 1053. [10] J. K. Nørskov, T. Bligaard, A. Logadottir, J. R. Kitchin, J. G. Chen, S. Pandelov, U. Stimming, J. Electrochem. Soc. 2005, 152, J23. [11] B. Hinnemann, P. G. Moses, J. Bonde, K. P. Jørgensen, J. H. Nielsen, S. Horch, I. Chorkendorff, J. K. Nørskov, J. Am. Chem. Soc. 2005, 127, 5308. [12] S. Chen, A. Kucernak, J. Phys. Chem. B 2004, 108, 13 984. [13] D. M. Kolb, Prog. Surf. Sci. 1996, 51, 109. [14] A. S. Dakkouri, D. M. Kolb in Interfacial Electrochemistry: Theory, Experiment, and Applications (Ed.: A. Wieckowski), Marcel Dekker, New York, 1999, p. 151. [15] J. Perez, E. R. Gonzalez, H. M. Villullas, J. Phys. Chem. B 1998, 102, 10 931. [16] D. M. Kolb, J. Schneider, Electrochim. Acta 1986, 31, 929. [17] J. Clavilier in Interfacial Electrochemistry: Theory, Experiment, and Applications (Ed.: A. Wieckowski), Marcel Dekker, New York, 1999, p. 231. [18] B. E. Conway, G. Jerkiewicz, Solid State Ionics 2002, 150, 93. [19] a) M. Baldauf, D. M. Kolb, Electrochim. Acta 1993, 38, 2145; b) L. A. Kibler, M. Kleinert, R. Randler, D. M. Kolb, Surf. Sci. 1999, 443, 19; c) J. Tang, M. Petri, L. A. Kibler, D. M. Kolb, Electrochim. Acta 2005, 51, 125; d) J. S. Spendelow, A. Wieckowski, Phys. Chem. Chem. Phys. 2004, 6 , 5094. [20] a) G. E. Engelmann, J. C. Ziegler, D. M. Kolb, J. Electrochem. Soc. 1998, 45, L33; b) F. Simeone, D. M. Kolb, Electrochim. Acta 2005, 50, 2989. [21] a) L. A. Kibler, A. M. El-Aziz, R. Hoyer, D. M. Kolb, Angew. Chem. 2005, 117, 2116; Angew. Chem. Int. Ed. 2005, 44, 2080. [22] B. Hammer, J. K. Nørskov, Surf. Sci. 1995, 343, 211. [23] A. Ruban, B. Hammer, P. Stoltze, H. L. Skriver, J. K. Nørskov, J. Mol. Catal. A 1997, 115, 421. [24] J. R. Kitchin, J. K. Nørskov, M. A. Barteau, J. G. Chen, Phys. Rev. Lett. 2004, 93, 156 801. [25] J. Greeley, J. K. Nørskov, L. A. Kibler, A. M. El-Aziz, D. M. Kolb, ChemPhysChem 2006, 7, 1032. [26] a) A. Roudgar, A. Groß, Phys. Rev. B 2003, 67, 033409; b) A. Roudgar, A. Groß, J. Electroanal. Chem. 2003, 548, 121. [27] A. Volkov, Diplomarbeit, University of Ulm, 2005. [28] A. Pediaditakis, Diplomarbeit, University of Ulm, 2005. [29] J. Meier, K. A. Friedrich, U. Stimming, Faraday Discuss. 2002, 121, 365. [30] M. Eikerling, J. Meier, U. Stimming, Z. Phys. Chem. 2003, 217, 395. [31] A. Roudgar, A. Groß, Surf. Sci. 2004, 559, L180. [32] C. G. Zoski, J. Phys. Chem. B 2003, 107, 6401. Received: November 22, 2005 Published online on April 11, 2006 9 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org 991
© Copyright 2026 Paperzz