This copy is for your personal, non-commercial use only.
If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here.
The following resources related to this article are available online at www.sciencemag.org (this information is current as of November 3, 2014 ): Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/content/334/6060/1256.full.html Supporting Online Material can be found at: http://www.sciencemag.org/content/suppl/2011/11/30/334.6060.1256.DC1.html A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/content/334/6060/1256.full.html#related This article cites 38 articles, 3 of which can be accessed free: http://www.sciencemag.org/content/334/6060/1256.full.html#ref-list-1 This article has been cited by 5 articles hosted by HighWire Press; see: http://www.sciencemag.org/content/334/6060/1256.full.html#related-urls This article appears in the following subject collections: Chemistry http://www.sciencemag.org/cgi/collection/chemistry
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2011 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.
Downloaded from www.sciencemag.org on November 3, 2014
Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here.
REPORTS (aL†aR†|vac〉) given an expected mean number of events m0 = d2Nps consistent with zero concurrence, where ps = 1.0 × 10−3 is the probability of generating a Stokes heralding photon and N is the number of experimental runs. Our results (X = 3, N = 1.9 × 1014, and m0 = 9.1 T 0.9) indicate positive concurrence at a 98 T 1% confidence level. Therefore, based on this detection of entanglement between Stokes and anti-Stokes modes, we can infer entanglement between the phonon modes of two macroscopic solids at room temperature. Finally, we examine the quality of entanglement generated between the diamonds by neglecting the vacuum component in Fig. 3, which is only caused by inefficiencies in coupling, detection, and readout of the anti-Stokes mode. To do this, we performed quantum state tomography (25) on the joint polarization state of the Stokes and anti-Stokes modes, postselecting on the detection of both photons. The reconstructed state is shown in Fig. 4, and we have subtracted accidental coincidences calculated from the Stokes and anti-Stokes singles rates. These results provide a more complete estimate of the coherence between the two modes than the interference fringes in Fig. 2. The concurrence of this subspace, 0.85, provides an estimate of the achievable entanglement between the two diamond phonon modes as the readout efficiency, coupling, and detector efficiencies approach unity (i.e., p00 → 0). Further, the fidelity to the nearest Bell state ½jHV 〉 þ
pffiffiffi jVH〉= 2 is 0.91. However, in the presence of real coupling and detection losses, the existence of entanglement can only be inferred from the density matrix in Fig. 3 (22). In our experiment, short-lived quantum correlations were revealed by combining an ultrafast interferometric pump-probe scheme with photon-counting techniques. The large optical bandwidth enabled the resolution of extremely fast dynamics in the solids, and also operation at high data rates, providing sufficient statistics to establish entanglement even in the presence of losses. This approach lays the foundation for future studies of quantum phenomena in manybody, strongly interacting systems coupled to strongly decohering environments and points toward a novel platform for ultrafast quantum information processing at room temperature. References and Notes 1. R. Penrose, in Mathematical Physics, A. Fokas, T. W. B. Kibble, A. Grigoriou, B. Zegarlinski, Eds. (Imperial College Press, London, 2000), pp. 266–282. 2. D. P. DiVincenzo, Fortschr. Phys. 48, 771 (2000). 3. S. Gigan et al., Nature 444, 67 (2006). 4. J. D. Thompson et al., Nature 452, 900 (2008). 5. B. Abbott et al., New J. Phys. 11, 073032 (2009). 6. S. Gerlich et al., Nat Commun 2, 263 (2011). 7. A. D. O’Connell et al., Nature 464, 697 (2010). 8. J. D. Teufel et al., Nature 475, 359 (2011). 9. J. Chan et al., Nature 478, 89 (2011). 10. G. Panitchayangkoon et al., Proc. Natl. Acad. Sci. U.S.A. 107, 12766 (2010).
Enhancing Hydrogen Evolution Activity in Water Splitting by Tailoring Li+-Ni(OH)2-Pt Interfaces Ram Subbaraman,1,2 Dusan Tripkovic,1 Dusan Strmcnik,1 Kee-Chul Chang,1 Masanobu Uchimura,1,3 Arvydas P. Paulikas,1 Vojislav Stamenkovic,1 Nenad M. Markovic1* Improving the sluggish kinetics for the electrochemical reduction of water to molecular hydrogen in alkaline environments is one key to reducing the high overpotentials and associated energy losses in water-alkali and chlor-alkali electrolyzers. We found that a controlled arrangement of nanometer-scale Ni (OH)2 clusters on platinum electrode surfaces manifests a factor of 8 activity increase in catalyzing the hydrogen evolution reaction relative to state-of-the-art metal and metal-oxide catalysts. In a bifunctional effect, the edges of the Ni(OH)2 clusters promoted the dissociation of water and the production of hydrogen intermediates that then adsorbed on the nearby Pt surfaces and recombined into molecular hydrogen. The generation of these hydrogen intermediates could be further enhanced via Li+-induced destabilization of the HO–H bond, resulting in a factor of 10 total increase in activity. lectrocatalysis of the hydrogen evolution reaction (HER) is critical to the operation of water-alkali electrolyzers (1–6), in which hydrogen is the main product, and chlor-alkali electrolyzers (5, 6), in which it is a side product.
E
1 Materials Science Division, Argonne National Laboratory, Lemont, IL 60559, USA. 2Nuclear Engineering Division, Argonne National Laboratory, Lemont, IL 60559, USA. 3Advanced Materials Laboratory, Nissan Research Center, Kanagawa 237-8523, Japan.
*To whom correspondence should be addressed. E-mail: [email protected]
1256
These two technologies are highly energy-intensive and are known to account for ~25 to 30% (87,600 to 92,000 GWh/year) of the total electrical energy consumption by industrial processes in the United States (3, 7). The HER is also an electrochemical reaction of fundamental scientific importance; the basic laws of electrode kinetics, as well as many modern concepts in electrocatalysis, were developed and verified by examining the reaction mechanisms related to the charge transfer–induced conversion of protons (in acid solutions) and water (in alkaline solutions) to molecular hydrogen.
2 DECEMBER 2011
VOL 334
SCIENCE
11. E. Collini et al., Nature 463, 644 (2010). 12. L.-M. Duan, M. D. Lukin, J. I. Cirac, P. Zoller, Nature 414, 413 (2001). 13. T. Chanelière et al., Nature 438, 833 (2005). 14. K. S. Choi, H. Deng, J. Laurat, H. J. Kimble, Nature 452, 67 (2008). 15. D. N. Matsukevich, A. Kuzmich, Science 306, 663 (2004). 16. C. W. Chou et al., Nature 438, 828 (2005). 17. G. Milburn, J. Opt. Soc. Am. B 24, 167 (2007). 18. See supplementary information on Science Online. 19. K. C. Lee et al., Diam. Relat. Mater. 19, 1289 (2010). 20. A. Greentree, B. Fairchild, F. Hossain, S. Prawer, Mater. Today 11, 22 (2008). 21. L. Prechtel et al., Nano Lett. 11, 269 (2011). 22. S. J. van Enk, Phys. Rev. A 75, 052318 (2007). 23. W. Wootters, Phys. Rev. Lett. 80, 2245 (1998). 24. G. J. Feldman, R. D. Cousins, Phys. Rev. D Part. Fields 57, 3873 (1998). 25. D. F. V. James, P. G. Kwiat, W. J. Munro, A. G. White, Phys. Rev. A 64, 052312 (2001). Acknowledgments: We thank V. Vedral, A. Datta, and L. Zhang for valuable insights. This work was supported by the Royal Society, Engineering and Physical Sciences Research Council (grant GR/S82176/01), EU IP Q-ESSENCE (grant 248095), EU ITN FASTQUAST, U.S. European Office of Aerospace Research and Development (grant 093020), Clarendon Fund, St. Edmund Hall, and Natural Sciences and Engineering Research Council of Canada.
Supporting Online Material www.sciencemag.org/cgi/content/full/334/6060/1253/DC1 Materials and Methods SOM Text References (26–36) 29 July 2011; accepted 27 October 2011 10.1126/science.1211914
Although previous studies have helped to rationalize which surface properties govern the variations in reactivity among catalysts (8–12), many key questions concerning the HER remain unanswered. For example, it is not clear why the rate of the HER is ~2 to 3 orders of magnitude lower at pH = 13 than at pH = 1, nor why the reaction is sensitive to the catalyst surface structure in alkaline media but largely insensitive in acids (13–17). A practical implication of the slow kinetics in alkaline solution is the lower energy efficiency for both water-alkali and chlor-alkali electrolyzers. For water-alkali electrolyzers, the high overpotentials for the oxygen evolution reaction (OER) at the anode also contribute significantly overall energy losses (18). This has led to various approaches to identify catalysts for both the OER and HER. However, such design strategies have rarely been based on molecular-level understanding of the reaction pathways. In addition, the influence of noncovalent (van der Waals–type) interactions on the overall kinetics of the HER has been underexplored, particularly in light of recent studies highlighting the impact of noncovalent interactions on the rates of many electrochemical reactions such as the oxygen reduction reaction, together with CO and methanol oxidation reactions (19–22). Currently, various combinations of metals (Pt, Pd, Ir, Ru, Ag, Ni), metal alloys (Ni-Co, Ni-Mn, Ni-Mo), metal oxides (RuO2), and Ni sulfides and phosphides are used to catalyze the conver-
www.sciencemag.org
REPORTS sion of H2O to H2 (2, 10–12, 23–27). Although Pt and Pt-based systems offer the highest activity and stability of all materials used for the HER, the benefits have not, to date, warranted the high cost associated with these materials. As a result, conventional electrolyzers generally use high– surface area Raney Ni and Ni alloys as the HER catalysts (2, 23, 24, 28). Several engineering approaches have been used to improve these nonnoble catalyst materials, such as enhancing the surface area, changing the alloy composition, and using higher catalyst loadings (~25 to 40 times the equivalent for Pt) (2). Although these approaches have offered small performance gains, key problems with the use of such non-noble materials remain, including the decrease in activity arising from both the formation of hydrides and the oxidative dissolution of the catalyst during intermittent operation (2). These materials limitations suggest that superior performance might be achieved if lower-cost Pt-based cathode materials could be developed. Indeed, by substantially increasing the activity of Pt and by also decreasing the Pt loading through the use of Pt-shell nanomaterials with non-noble cores (29, 30), it would be possible to envision the use of highly active, durable, and low-cost Pt-based HER electrodes for alkaline electrolyzers. Limitations in the catalytic activities of Pt and Pt-group metals arise from the fact that although most of these materials are good catalysts for the
adsorption and recombination of the reactive hydrogen intermediates (Had), they are generally inefficient in the prior step of water dissociation. Conversely, metal oxides (and in some cases other compounds such as sulfides), although effective for cleaving the HO–H bond, are poor at converting the resulting Had intermediates to H2 (31–33). Hence, optimal HER catalyst design will depend on combining the catalytic proficiencies of metals and metal oxides by creating new bifunctional metal oxide–metal systems (metal oxides deposited on metal substrates) (34–37). Here, we report the design and performance of composite materials to facilitate different parts of the overall multistep HER process in alkaline environments: an oxide to provide the active sites for dissociation of water, and a metal to facilitate adsorption of the atomic hydrogen produced and its subsequent association to form H2 from these intermediates. This involved growth of conductive ultrathin Ni(OH)2 clusters (height 0.7 nm, width 8 to 10 nm) on both pristine Pt singlecrystal surfaces and Pt surfaces modified by twodimensional (2D) Pt ad-islands [Pt-islands/Pt(111)]. We found that, relative to the corresponding Pt single-crystal surfaces, the most active Ni(OH)2/Ptislands/Pt(111) electrodes in KOH solutions are more active for the HER by a factor of ~8 at an overpotential of –0.1 V. Further enhancement of water dissociation is achieved by the intro-
Fig. 1. (A to C) STM images (60 nm by 60 nm) and CV traces for (A) Pt(111), (B) Pt(111) with 2D Pt islands, and (C) Pt(111) modified with 3D Ni(OH)2 clusters in 0.1 M KOH electrolyte. (D) HER activities for Pt(111), Pt-islands/Pt(111), and Ni(OH)2/Pt(111) surfaces in alkaline solution (a, b, and c, respectively). Corresponding HER activities for Pt(111) and Pt-islands/Pt(111) electrodes in acid solutions (a´ and b´) are shown to emphasize large initial differences between kinetics of the HER in alkaline versus acid solutions. The inset shows XANES spectra for Ni(OH)2 on Pt(111) shown for three different potentials: HER (–0.1 V), Hupd (0.1 V), and near OER (1.2 V). Also shown is the reference spectrum for Ni(OH)2. No shift in the edge energy in XANES spectra between HER and Hupd regions is observed. www.sciencemag.org
SCIENCE
VOL 334
duction of solvated Li+ ions into the compact portion of the double layer, resulting in a factor of 10 total increase in activity. Finally, we demonstrate that the knowledge attained by studying single-crystal surfaces can be used for the design of prospective commercial nanocatalysts for alkaline electrolyzers. As a starting point, to develop more complete structure-function relationships for the HER, we used scanning tunneling microscopy (STM) to compare the atomic structures of Pt(111) and Pt(111) modified by electrochemically deposited Pt islands, referred as Pt-islands/Pt(111), (38). In agreement with prior reports (39, 40), the image of Pt(111) in Fig. 1A displays the presence of a few randomly distributed mono-atomic steps and 2D Pt islands with diameters of 1 to 2 nm and monoatomic height. Considering that the shape of the current-potential curve in both the underpotentially deposited hydrogen (Hupd) region [defined as the state of hydrogen adsorbed at a potential that is positive of the Nernst potential for the hydrogen reaction (33)], between 0.05 to 0.35 V, and the region of reversible adsorption of hydroxyl (OHad) species, above 0.6 V, is consistent with earlier reports for a perfect Pt(111) surface, we conclude that these defects are invisible in cyclic voltammetry (CV) traces. In the STM image of the islands shown in Fig. 1B, however, the Pt adatoms can be clearly resolved as 2D features with diameters of ~1 to 3 nm and a height of 1 atomic layer. The CV trace of such a surface encompasses two sharp Hupd peaks centered at 0.23 V and 0.4 V (Fig. 1B). On the basis of prior studies (39, 40), we associate these peaks with hydrogen adsorption at the (111)-(111) and (111)-(100) terrace-step sites, respectively. Consistent with the higher oxophilicity of low-coordinated Pt sites, the onset of OH adsorption starts at more negative potentials on the Pt island–covered electrode than on pristine Pt(111), whereas the OHad peaks are less reversible on the former surface. After 50 potential cycles between – 0.3 V and +0.3 V, the STM images and the CV traces remain the same, indicating that within this potential range, the morphologies of the Pt(111) and Pt-islands/Pt(111) surfaces are stable. Therefore, such well-defined surfaces offer a unique opportunity to correlate the kinetic rates of the HER with a truly atomically resolved surface structure. The mechanism of the HER in alkaline media is typically treated as a combination of three elementary steps: the Volmer step—water dissociation and formation of a reactive intermediate Had (2H2O + M + 2e– ⇆ 2M-Had + 2OH–)—followed by either the Heyrovsky step (H2O + Had-M + e– ⇆ M + H2 + OH–) or the Tafel recombination step (2M-Had ⇆ 2M + H2). Adsorbed hydrogen species Had formed at potentials negative of the Nernst reversible potential for the HER are also referred to as overpotentially deposited hydrogen (Hopd). To distinguish the different states of adsorbed hydrogen, we use a thermodynamic notation,
2 DECEMBER 2011
1257
REPORTS referring to Hupd as the strongly adsorbed state and Had (i.e., Hopd) as a weakly adsorbed state. Any rigorous kinetic analysis of the HER lies beyond the scope of the present discussion. Rather, we focus mainly on the design of interfaces for efficient electrochemical conversion of H2O to H2. For example, Fig. 1D shows that in alkaline solution, relative to the corresponding pristine Pt(111) surface, the Pt-islands/Pt(111) surface is ~5 to 6 times as active for the HER. Figure 1D also shows that in acid solution, the HER on the Pt-islands/Pt(111) electrode is improved by a factor of only ~1.5. In turn, this strong pH effect indicates that the low-coordinated Pt atoms may have a major effect on the ratedetermining step of the HER in alkaline solutions. Because the major difference between the reaction pathways in alkaline and acid solutions is that the hydrogen in alkaline solutions is discharged from water instead of from hydronium ions (H3O+) (13–15), we propose that the large promoting effect of low-coordinated Pt atoms in alkaline solution is due to more facile dissociative adsorption of water. In turn, this would be consistent with the Volmer reaction being the rate-determining step for the HER in alkaline electrolytes. The role of edge-step sites in accelerating dissociative adsorption of water on metal surfaces is well documented in ultrahigh-vacuum (UHV) environments (33). We therefore conclude that for materials with near-optimal M-Had energetics (such as Pt), surface reactivity for the HER can be further improved by tailoring the active sites for more efficient dissociative adsorption of water molecules. To fulfill this requirement, we modified Pt(111) and Pt-island/Pt(111) surfaces by depositing Ni-(hydr)oxide clusters (38), as the 3d-transition metal oxides might be even more active for water dissociation than Pt defect sites (31). The facile water dissociation properties of Ni(hydroxy) oxides relative to other transition metal oxides have been well established (26–28), motivating the use of Ni(OH)2 for this work. The local symmetry, the oxidation state of Ni atoms, and the number and identities of nearest-neighbor atoms and the distances between them were determined by in situ x-ray absorption spectroscopy (XAS) measurements (41, 42). For example, from the analysis of the x-ray absorption near-edge structure (XANES) and extended x-ray absorption fine structure (EXAFS) of the XAS spectra (inset of Fig. 1D; see also fig. S3), we found Ni-O and Ni-Ni bond distances of 2.05 T 0.01 Å and 3.08 T 0.01 Å, and we also determined that Ni remains mostly in the +2 valence state, even after multiple hours of holding the electrode potential at –0.1 V. Furthermore, from the edge shift (defined as the half-height energy of the normalized XANES edge step), we concluded that between –0.1 Vand +0.8 V, the change in the oxidation state of Ni is less than 0.5. These results suggest that stable Ni(OH)2 clusters are the predominant hydr(oxide) form on the Pt(111) and Pt-islands/Pt(111) surfaces, especially in the HER potential region.
1258
Because the octahedral symmetry of the a and/or b phases of Ni(OH)2 prevents p-d hybridization, the prominent pre-edge from 1s → 3d transitions implies that the Ni(OH)2 species are rich in defects that, according to prior reports (31–33), are particularly active for dissociative adsorption of water molecules. The surface morphologies of Ni(OH)2/Pt(111) and Ni(OH)2/Pt-islands/Pt(111) were probed by STM and in situ surface x-ray crystal truncation rod (CTR) measurements (43, 44) (see fig. S4 for the CTR data). Although atomic resolution could
not be obtained, the STM image in Fig. 1C clearly shows that the Ni(OH)2 clusters are randomly distributed across the (111) terraces. All Ni(OH)2 clusters exhibited hemisphere-like shapes, with characteristic diameters of ~8 to 10 nm and heights of ~0.7 nm, the latter corresponding to two layers of Ni(OH)2. This result indicates that the oxide exhibits Volmer-Weber (VW)–type growth whereby 3D clusters of Ni(OH)2 grow even at the lowest coverages (45). VW growth, in turn, is possible if the heat of adsorption of Ni(OH)2 on Pt is lower than the cohesive energy of Ni(OH)2.
Fig. 2. (A) STM image (60 nm by 60 nm) and CV trace of the Ni(OH)2/Pt-islands/Pt(111) surface. Clusters of Ni(OH)2 in the STM image appear ellipsoidal with particle sizes between 4 and 12 nm. (B) Comparison of HER activities with Pt(111) as the substrate. Incremental improvements in activities for the HER in 0.1 M KOH from the unmodified Pt(111) surface are shown for the hierarchical materials [ad-islands, Ni(OH)2, and their combination] as well as the double layer (addition of Li+ cations). The activity for the unmodified Pt(111) surface in 0.1 M HClO4 is shown for reference. Dashed arrow shows the activity trend.
Fig. 3. Schematic representation of water dissociation, formation of M-Had intermediates, and subsequent recombination of two Had atoms to form H2 (magenta arrow) as well as OH– desorption from the Ni(OH)2 domains (red arrows) followed by adsorption of another water molecule on the same site (blue arrows). Water adsorption requires concerted interaction of O atoms with Ni(OH)2 (broken orange spikes) and H atoms with Pt (broken magenta spikes) at the boundary between Ni(OH)2 and Pt domains. The Ni(OH)2-induced stabilization of hydrated cations (AC+) (broken dark blue spikes) likely occurs through noncovalent (van der Waals–type) interactions. Hydrated AC+ can further interact with water molecules (broken yellow spikes), altering the orientation of water as well as the nature and strength of interaction of the oxide with water.
2 DECEMBER 2011
VOL 334
SCIENCE
www.sciencemag.org
REPORTS The surface coverage of Ni(OH)2 on Pt(111) is estimated from the STM image by measuring the area covered by the particles on the Pt(111) substrate. According to such analysis, the cluster density reached a maximum at a surface coverage of ~35%. Upon comparing STM images of Pt(111) and Ni(OH)2/Pt(111) surfaces (Fig. 1, A and C), however, it was not clear whether isolated defects or flat Pt(111) terraces were the preferred nucleation sites for Ni(OH)2. This information was more accessible by comparing STM images of the Ni(OH)2-free (Fig. 1B) and Ni(OH)2-covered Pt-island/Pt(111) (Fig. 2A) electrodes. The STM image in Fig. 2A was acquired after deposition of Ni(OH)2 on a Pt(111) surface modified by ~0.2 monolayer (ML) of Pt islands. We observed formation of both 3D Ni(OH)2 clusters (having a predominantly ellipsoidal shape) and oxide-free terraces. The clusters had approximately constant heights of ~0.8 nm but diameters ranging from 4 to 12 nm. The same STM image, however, revealed no visible presence of 2D Pt islands, which suggests
that Ni(OH)2 preferentially nucleates on the Pt surface defects and that most of the Pt islands are covered by Ni(OH)2. Having established the nature and morphology of the Ni(OH)2 clusters, we briefly summarize the role of the Ni(OH)2 clusters in the formation of Hupd and OHad adlayers on Pt(111) and Ptislands/Pt(111) electrodes. Addition of Ni(OH)2 on the surface of Pt(111) as well as on the Pt(111) surface covered with Pt islands led to a decrease (~35%) in the coverage of Hupd. This finding suggests that the Ni(OH)2 clusters selectively block the Pt sites corresponding to Hupd. Furthermore, the two sharp Hupd peaks characteristic of hydrogen adsorption/desorption on the Pt(111) electrode modified by the 2D Pt islands are completely suppressed on the surface covered by the 3D Ni(OH)2 clusters (Fig. 2A). This is additional evidence, consistent with the STM results, that defects serve as the nucleation centers for electrodeposition of Ni(OH)2. We therefore conclude that Pt islands are predominantly covered by Ni(OH)2. In contrast to the Hupd potential re-
Fig. 4. (A) STM image (50 nm by 50 nm) and CV trace for Ni(OH)2/Pt-islands/Pt(110) single-crystal surface. STM image shows Ni(OH)2 randomly distributed on inherently rough Pt(110).. (B) Comparison of HER activities with Pt(110) as the substrate. Incremental improvements in activities for the HER in 0.1 M KOH from the unmodified Pt(110) surface are shown for the hierarchical materials [Ni(OH)2 and the combination of ad-islands with Ni(OH)2] as well as the double layer (addition of Li+ cations). The activity for the unmodified Pt(110) surface in 0.1 M HClO4 is shown for reference. Dashed arrow shows the activity trend. (C) Transmission electron micrograph image (50 nm by 50 nm) and corresponding CV trace of Ni(OH)2-free Pt-nano catalysts (TKK) with an average particle size of 5 nm. (D) Comparison of HER activities with commercial nanocatalyst Pt/C (TKK) as the substrate. Incremental improvements in activities for the HER in 0.1 M KOH from unmodified Pt/C are shown for the hierarchical materials [surface covered with Ni(OH)2] as well as the double layer (addition of Li+ cations). The activity for the unmodified Pt/C surface in 0.1 M HClO4 is shown for reference. Dashed arrow shows the activity trend. All the current densities for the TKK catalyst system are normalized by the geometric surface area of the glassy carbon substrate. www.sciencemag.org
SCIENCE
VOL 334
gion, an enhanced adsorption of OHad, which is accompanied by irreversible reduction of OHad on the negative-going sweep, is observed on both electrodes, arising from the higher oxophilicity of the surface elements. Although in the presence of Ni(OH)2 clusters there are 35% fewer Pt sites available for the HER than on the bare Pt(111) substrate, the Ni(OH)2/Pt(111) electrode is ~7 times as active for the HER relative to Pt(111) (Fig. 1D). Moreover, Fig. 2B shows that the activity is further enhanced [by a factor of ~8 relative to bare Pt(111)] on the Ni(OH)2/Pt-island/Pt(111) surface; at 6 mA/cm2, the difference in overpotential between the HER in alkaline and acid solutions is reduced to only 100 mV. Clearly, then, on both surfaces, Ni(OH)2 must play a promoting role in the dissociation of water and thereby enhances the rate of formation of Had intermediates on the metal surface. As schematically depicted in Fig. 3, we propose that water adsorption requires concerted interaction of O atoms with Ni(OH)2 and H atoms with Pt at the boundary between Ni(OH)2 and Pt domains. Water adsorption is then followed by water dissociation and hydrogen adsorption (Had) on the nearby vacant Pt sites. Finally, two Had atoms on the Pt surface recombine to form H2 (H2 desorption step) and OH– desorbs from the Ni(OH)2 domains, followed by adsorption of another water molecule on the same site. From a surface reactivity standpoint, fruitful kinetic synergy (bifunctionality) between Ni(OH)2 and Pt appears to be the key to maximizing the rate of the HER. As shown in Fig. 2B, this bifunctionality in turn brings the activity of the HER in alkaline solutions very close to the activity of Pt in acid solutions. To verify this conclusion, we have also compared the HER on Au(111) and Ni(OH)2/Au(111) in alkaline solution. The relatively weak interaction between Au and Had offsets the benefit of the enhanced water dissociation at the Au/Ni(OH)2 interface (38). As a result, the rate of the HER on the Ni(OH)2/Au(111) surface is much lower than on the Pt(111)/Ni(OH)2 surface. This further emphasizes the importance of choosing the correct metal oxide–metal pairs in optimizing the kinetics of the HER. In what follows, we demonstrate that the nature of interactions in double layer is equally important, and the rate-determining Volmer step can be further enhanced by tuning the double-layer properties. Several recent studies have unambiguously shown that the rate of electrochemical reactions on Pt in alkaline solutions is controlled by the presence of alkali-metal cations (AC+) (20, 21). However, these effects have been entirely restricted to the potential region of a critical OHad coverage (for electrode potential E > 0.6 V), the latter species serving to stabilize hydrated cations in the compact part of the double layer through noncovalent (van der Waals–type) interactions. This stabilization leads to the formation of OHad-AC+(H2O)x complexes that can either decrease the reactivity of Pt by blocking
2 DECEMBER 2011
1259
REPORTS the active sites for adsorption of reactants such as O2, H2, and CH3OH (20, 21) or, as in the case of the CO oxidation reaction, improve the reactivity of Pt via enhanced adsorption of OHad (22). For the present reaction, the key questions are whether hydrated alkali cations on the Ni(OH)2/Pt-islands/Pt(111) surface can be stabilized through interactions with OH species in the Ni(OH)2 cluster and, if so, whether these hydrated cations could affect the alkaline HER kinetics. For these purposes, the effect of hydrated Li cations was probed mainly because, in alkaline environments, Li+ is known to interact with H2O and OHad more strongly than K+. In line with the previous electrochemical report (20), we found that Li+ cations have no effect on the HER on Pt(111) surfaces. However, the results in Fig. 2B revealed that the HER on Ni(OH)2/Pt-islands/Pt(111) is enhanced by almost a factor of 2 in the presence of Li+ cations. This increase in activity has substantially narrowed the gap between the rates of HER on Pt in acid and alkaline solution; more specifically, inspection of Fig. 2B shows that, at 5 mA/cm2, the difference in overpotential between acid and alkaline environments is narrowed to only 35 mV. The fact that the activity of the HER is affected by the nature of alkali metal cations strongly suggests that Ni(OH)2-Li+-OH-H complexes are present in the compact portion of the double layer. The presence of this complex does not completely explain the factor of 2 increase in HER activity. It is likely that the probability of water dissociation is enhanced via possible Li+-induced steric and/or electronic effects on the interfacial water structure and reactivity, as shown schematically in Fig. 3. Thus, Ni(OH)2 plays a dual role: In addition to assisting with water dissociation, it also provides an anchor to hold the beneficial Li+ ions in the compact portion of the double layer. Having established the methodology of tuning interfacial activity for the HER on Pt(111), we demonstrate that the very same guiding principles for accelerating the Volmer reaction step in alkaline solutions are equally applicable to Pt(110). For simplicity, only the CV and STM data for the most active system, Ni(OH)2/Ptislands/Pt(110), are shown in Fig. 4A. The general characteristics (both structural and electrochemical) are similar to what was observed for the corresponding Pt(111) systems. The current denities for the HER on Pt(110) and Pt-nanocatalyst systems are presented in the logarithmic Tafel form (Fig. 4, B and D) to offer the most condensed representation of surface characterization and polarization curves. As expected, the systematic modification of Pt(110), first with Pt islands and then with Ni(OH)2, exhibits an HER activity trend (Fig. 4B) with the same order as the driving force for dissociative adsorption of water molecules, as discussed above for the Pt(111) systems: Pt(110) < Ni(OH)2/Pt(110)
If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here.
The following resources related to this article are available online at www.sciencemag.org (this information is current as of November 3, 2014 ): Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/content/334/6060/1256.full.html Supporting Online Material can be found at: http://www.sciencemag.org/content/suppl/2011/11/30/334.6060.1256.DC1.html A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/content/334/6060/1256.full.html#related This article cites 38 articles, 3 of which can be accessed free: http://www.sciencemag.org/content/334/6060/1256.full.html#ref-list-1 This article has been cited by 5 articles hosted by HighWire Press; see: http://www.sciencemag.org/content/334/6060/1256.full.html#related-urls This article appears in the following subject collections: Chemistry http://www.sciencemag.org/cgi/collection/chemistry
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2011 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.
Downloaded from www.sciencemag.org on November 3, 2014
Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here.
REPORTS (aL†aR†|vac〉) given an expected mean number of events m0 = d2Nps consistent with zero concurrence, where ps = 1.0 × 10−3 is the probability of generating a Stokes heralding photon and N is the number of experimental runs. Our results (X = 3, N = 1.9 × 1014, and m0 = 9.1 T 0.9) indicate positive concurrence at a 98 T 1% confidence level. Therefore, based on this detection of entanglement between Stokes and anti-Stokes modes, we can infer entanglement between the phonon modes of two macroscopic solids at room temperature. Finally, we examine the quality of entanglement generated between the diamonds by neglecting the vacuum component in Fig. 3, which is only caused by inefficiencies in coupling, detection, and readout of the anti-Stokes mode. To do this, we performed quantum state tomography (25) on the joint polarization state of the Stokes and anti-Stokes modes, postselecting on the detection of both photons. The reconstructed state is shown in Fig. 4, and we have subtracted accidental coincidences calculated from the Stokes and anti-Stokes singles rates. These results provide a more complete estimate of the coherence between the two modes than the interference fringes in Fig. 2. The concurrence of this subspace, 0.85, provides an estimate of the achievable entanglement between the two diamond phonon modes as the readout efficiency, coupling, and detector efficiencies approach unity (i.e., p00 → 0). Further, the fidelity to the nearest Bell state ½jHV 〉 þ
pffiffiffi jVH〉= 2 is 0.91. However, in the presence of real coupling and detection losses, the existence of entanglement can only be inferred from the density matrix in Fig. 3 (22). In our experiment, short-lived quantum correlations were revealed by combining an ultrafast interferometric pump-probe scheme with photon-counting techniques. The large optical bandwidth enabled the resolution of extremely fast dynamics in the solids, and also operation at high data rates, providing sufficient statistics to establish entanglement even in the presence of losses. This approach lays the foundation for future studies of quantum phenomena in manybody, strongly interacting systems coupled to strongly decohering environments and points toward a novel platform for ultrafast quantum information processing at room temperature. References and Notes 1. R. Penrose, in Mathematical Physics, A. Fokas, T. W. B. Kibble, A. Grigoriou, B. Zegarlinski, Eds. (Imperial College Press, London, 2000), pp. 266–282. 2. D. P. DiVincenzo, Fortschr. Phys. 48, 771 (2000). 3. S. Gigan et al., Nature 444, 67 (2006). 4. J. D. Thompson et al., Nature 452, 900 (2008). 5. B. Abbott et al., New J. Phys. 11, 073032 (2009). 6. S. Gerlich et al., Nat Commun 2, 263 (2011). 7. A. D. O’Connell et al., Nature 464, 697 (2010). 8. J. D. Teufel et al., Nature 475, 359 (2011). 9. J. Chan et al., Nature 478, 89 (2011). 10. G. Panitchayangkoon et al., Proc. Natl. Acad. Sci. U.S.A. 107, 12766 (2010).
Enhancing Hydrogen Evolution Activity in Water Splitting by Tailoring Li+-Ni(OH)2-Pt Interfaces Ram Subbaraman,1,2 Dusan Tripkovic,1 Dusan Strmcnik,1 Kee-Chul Chang,1 Masanobu Uchimura,1,3 Arvydas P. Paulikas,1 Vojislav Stamenkovic,1 Nenad M. Markovic1* Improving the sluggish kinetics for the electrochemical reduction of water to molecular hydrogen in alkaline environments is one key to reducing the high overpotentials and associated energy losses in water-alkali and chlor-alkali electrolyzers. We found that a controlled arrangement of nanometer-scale Ni (OH)2 clusters on platinum electrode surfaces manifests a factor of 8 activity increase in catalyzing the hydrogen evolution reaction relative to state-of-the-art metal and metal-oxide catalysts. In a bifunctional effect, the edges of the Ni(OH)2 clusters promoted the dissociation of water and the production of hydrogen intermediates that then adsorbed on the nearby Pt surfaces and recombined into molecular hydrogen. The generation of these hydrogen intermediates could be further enhanced via Li+-induced destabilization of the HO–H bond, resulting in a factor of 10 total increase in activity. lectrocatalysis of the hydrogen evolution reaction (HER) is critical to the operation of water-alkali electrolyzers (1–6), in which hydrogen is the main product, and chlor-alkali electrolyzers (5, 6), in which it is a side product.
E
1 Materials Science Division, Argonne National Laboratory, Lemont, IL 60559, USA. 2Nuclear Engineering Division, Argonne National Laboratory, Lemont, IL 60559, USA. 3Advanced Materials Laboratory, Nissan Research Center, Kanagawa 237-8523, Japan.
*To whom correspondence should be addressed. E-mail: [email protected]
1256
These two technologies are highly energy-intensive and are known to account for ~25 to 30% (87,600 to 92,000 GWh/year) of the total electrical energy consumption by industrial processes in the United States (3, 7). The HER is also an electrochemical reaction of fundamental scientific importance; the basic laws of electrode kinetics, as well as many modern concepts in electrocatalysis, were developed and verified by examining the reaction mechanisms related to the charge transfer–induced conversion of protons (in acid solutions) and water (in alkaline solutions) to molecular hydrogen.
2 DECEMBER 2011
VOL 334
SCIENCE
11. E. Collini et al., Nature 463, 644 (2010). 12. L.-M. Duan, M. D. Lukin, J. I. Cirac, P. Zoller, Nature 414, 413 (2001). 13. T. Chanelière et al., Nature 438, 833 (2005). 14. K. S. Choi, H. Deng, J. Laurat, H. J. Kimble, Nature 452, 67 (2008). 15. D. N. Matsukevich, A. Kuzmich, Science 306, 663 (2004). 16. C. W. Chou et al., Nature 438, 828 (2005). 17. G. Milburn, J. Opt. Soc. Am. B 24, 167 (2007). 18. See supplementary information on Science Online. 19. K. C. Lee et al., Diam. Relat. Mater. 19, 1289 (2010). 20. A. Greentree, B. Fairchild, F. Hossain, S. Prawer, Mater. Today 11, 22 (2008). 21. L. Prechtel et al., Nano Lett. 11, 269 (2011). 22. S. J. van Enk, Phys. Rev. A 75, 052318 (2007). 23. W. Wootters, Phys. Rev. Lett. 80, 2245 (1998). 24. G. J. Feldman, R. D. Cousins, Phys. Rev. D Part. Fields 57, 3873 (1998). 25. D. F. V. James, P. G. Kwiat, W. J. Munro, A. G. White, Phys. Rev. A 64, 052312 (2001). Acknowledgments: We thank V. Vedral, A. Datta, and L. Zhang for valuable insights. This work was supported by the Royal Society, Engineering and Physical Sciences Research Council (grant GR/S82176/01), EU IP Q-ESSENCE (grant 248095), EU ITN FASTQUAST, U.S. European Office of Aerospace Research and Development (grant 093020), Clarendon Fund, St. Edmund Hall, and Natural Sciences and Engineering Research Council of Canada.
Supporting Online Material www.sciencemag.org/cgi/content/full/334/6060/1253/DC1 Materials and Methods SOM Text References (26–36) 29 July 2011; accepted 27 October 2011 10.1126/science.1211914
Although previous studies have helped to rationalize which surface properties govern the variations in reactivity among catalysts (8–12), many key questions concerning the HER remain unanswered. For example, it is not clear why the rate of the HER is ~2 to 3 orders of magnitude lower at pH = 13 than at pH = 1, nor why the reaction is sensitive to the catalyst surface structure in alkaline media but largely insensitive in acids (13–17). A practical implication of the slow kinetics in alkaline solution is the lower energy efficiency for both water-alkali and chlor-alkali electrolyzers. For water-alkali electrolyzers, the high overpotentials for the oxygen evolution reaction (OER) at the anode also contribute significantly overall energy losses (18). This has led to various approaches to identify catalysts for both the OER and HER. However, such design strategies have rarely been based on molecular-level understanding of the reaction pathways. In addition, the influence of noncovalent (van der Waals–type) interactions on the overall kinetics of the HER has been underexplored, particularly in light of recent studies highlighting the impact of noncovalent interactions on the rates of many electrochemical reactions such as the oxygen reduction reaction, together with CO and methanol oxidation reactions (19–22). Currently, various combinations of metals (Pt, Pd, Ir, Ru, Ag, Ni), metal alloys (Ni-Co, Ni-Mn, Ni-Mo), metal oxides (RuO2), and Ni sulfides and phosphides are used to catalyze the conver-
www.sciencemag.org
REPORTS sion of H2O to H2 (2, 10–12, 23–27). Although Pt and Pt-based systems offer the highest activity and stability of all materials used for the HER, the benefits have not, to date, warranted the high cost associated with these materials. As a result, conventional electrolyzers generally use high– surface area Raney Ni and Ni alloys as the HER catalysts (2, 23, 24, 28). Several engineering approaches have been used to improve these nonnoble catalyst materials, such as enhancing the surface area, changing the alloy composition, and using higher catalyst loadings (~25 to 40 times the equivalent for Pt) (2). Although these approaches have offered small performance gains, key problems with the use of such non-noble materials remain, including the decrease in activity arising from both the formation of hydrides and the oxidative dissolution of the catalyst during intermittent operation (2). These materials limitations suggest that superior performance might be achieved if lower-cost Pt-based cathode materials could be developed. Indeed, by substantially increasing the activity of Pt and by also decreasing the Pt loading through the use of Pt-shell nanomaterials with non-noble cores (29, 30), it would be possible to envision the use of highly active, durable, and low-cost Pt-based HER electrodes for alkaline electrolyzers. Limitations in the catalytic activities of Pt and Pt-group metals arise from the fact that although most of these materials are good catalysts for the
adsorption and recombination of the reactive hydrogen intermediates (Had), they are generally inefficient in the prior step of water dissociation. Conversely, metal oxides (and in some cases other compounds such as sulfides), although effective for cleaving the HO–H bond, are poor at converting the resulting Had intermediates to H2 (31–33). Hence, optimal HER catalyst design will depend on combining the catalytic proficiencies of metals and metal oxides by creating new bifunctional metal oxide–metal systems (metal oxides deposited on metal substrates) (34–37). Here, we report the design and performance of composite materials to facilitate different parts of the overall multistep HER process in alkaline environments: an oxide to provide the active sites for dissociation of water, and a metal to facilitate adsorption of the atomic hydrogen produced and its subsequent association to form H2 from these intermediates. This involved growth of conductive ultrathin Ni(OH)2 clusters (height 0.7 nm, width 8 to 10 nm) on both pristine Pt singlecrystal surfaces and Pt surfaces modified by twodimensional (2D) Pt ad-islands [Pt-islands/Pt(111)]. We found that, relative to the corresponding Pt single-crystal surfaces, the most active Ni(OH)2/Ptislands/Pt(111) electrodes in KOH solutions are more active for the HER by a factor of ~8 at an overpotential of –0.1 V. Further enhancement of water dissociation is achieved by the intro-
Fig. 1. (A to C) STM images (60 nm by 60 nm) and CV traces for (A) Pt(111), (B) Pt(111) with 2D Pt islands, and (C) Pt(111) modified with 3D Ni(OH)2 clusters in 0.1 M KOH electrolyte. (D) HER activities for Pt(111), Pt-islands/Pt(111), and Ni(OH)2/Pt(111) surfaces in alkaline solution (a, b, and c, respectively). Corresponding HER activities for Pt(111) and Pt-islands/Pt(111) electrodes in acid solutions (a´ and b´) are shown to emphasize large initial differences between kinetics of the HER in alkaline versus acid solutions. The inset shows XANES spectra for Ni(OH)2 on Pt(111) shown for three different potentials: HER (–0.1 V), Hupd (0.1 V), and near OER (1.2 V). Also shown is the reference spectrum for Ni(OH)2. No shift in the edge energy in XANES spectra between HER and Hupd regions is observed. www.sciencemag.org
SCIENCE
VOL 334
duction of solvated Li+ ions into the compact portion of the double layer, resulting in a factor of 10 total increase in activity. Finally, we demonstrate that the knowledge attained by studying single-crystal surfaces can be used for the design of prospective commercial nanocatalysts for alkaline electrolyzers. As a starting point, to develop more complete structure-function relationships for the HER, we used scanning tunneling microscopy (STM) to compare the atomic structures of Pt(111) and Pt(111) modified by electrochemically deposited Pt islands, referred as Pt-islands/Pt(111), (38). In agreement with prior reports (39, 40), the image of Pt(111) in Fig. 1A displays the presence of a few randomly distributed mono-atomic steps and 2D Pt islands with diameters of 1 to 2 nm and monoatomic height. Considering that the shape of the current-potential curve in both the underpotentially deposited hydrogen (Hupd) region [defined as the state of hydrogen adsorbed at a potential that is positive of the Nernst potential for the hydrogen reaction (33)], between 0.05 to 0.35 V, and the region of reversible adsorption of hydroxyl (OHad) species, above 0.6 V, is consistent with earlier reports for a perfect Pt(111) surface, we conclude that these defects are invisible in cyclic voltammetry (CV) traces. In the STM image of the islands shown in Fig. 1B, however, the Pt adatoms can be clearly resolved as 2D features with diameters of ~1 to 3 nm and a height of 1 atomic layer. The CV trace of such a surface encompasses two sharp Hupd peaks centered at 0.23 V and 0.4 V (Fig. 1B). On the basis of prior studies (39, 40), we associate these peaks with hydrogen adsorption at the (111)-(111) and (111)-(100) terrace-step sites, respectively. Consistent with the higher oxophilicity of low-coordinated Pt sites, the onset of OH adsorption starts at more negative potentials on the Pt island–covered electrode than on pristine Pt(111), whereas the OHad peaks are less reversible on the former surface. After 50 potential cycles between – 0.3 V and +0.3 V, the STM images and the CV traces remain the same, indicating that within this potential range, the morphologies of the Pt(111) and Pt-islands/Pt(111) surfaces are stable. Therefore, such well-defined surfaces offer a unique opportunity to correlate the kinetic rates of the HER with a truly atomically resolved surface structure. The mechanism of the HER in alkaline media is typically treated as a combination of three elementary steps: the Volmer step—water dissociation and formation of a reactive intermediate Had (2H2O + M + 2e– ⇆ 2M-Had + 2OH–)—followed by either the Heyrovsky step (H2O + Had-M + e– ⇆ M + H2 + OH–) or the Tafel recombination step (2M-Had ⇆ 2M + H2). Adsorbed hydrogen species Had formed at potentials negative of the Nernst reversible potential for the HER are also referred to as overpotentially deposited hydrogen (Hopd). To distinguish the different states of adsorbed hydrogen, we use a thermodynamic notation,
2 DECEMBER 2011
1257
REPORTS referring to Hupd as the strongly adsorbed state and Had (i.e., Hopd) as a weakly adsorbed state. Any rigorous kinetic analysis of the HER lies beyond the scope of the present discussion. Rather, we focus mainly on the design of interfaces for efficient electrochemical conversion of H2O to H2. For example, Fig. 1D shows that in alkaline solution, relative to the corresponding pristine Pt(111) surface, the Pt-islands/Pt(111) surface is ~5 to 6 times as active for the HER. Figure 1D also shows that in acid solution, the HER on the Pt-islands/Pt(111) electrode is improved by a factor of only ~1.5. In turn, this strong pH effect indicates that the low-coordinated Pt atoms may have a major effect on the ratedetermining step of the HER in alkaline solutions. Because the major difference between the reaction pathways in alkaline and acid solutions is that the hydrogen in alkaline solutions is discharged from water instead of from hydronium ions (H3O+) (13–15), we propose that the large promoting effect of low-coordinated Pt atoms in alkaline solution is due to more facile dissociative adsorption of water. In turn, this would be consistent with the Volmer reaction being the rate-determining step for the HER in alkaline electrolytes. The role of edge-step sites in accelerating dissociative adsorption of water on metal surfaces is well documented in ultrahigh-vacuum (UHV) environments (33). We therefore conclude that for materials with near-optimal M-Had energetics (such as Pt), surface reactivity for the HER can be further improved by tailoring the active sites for more efficient dissociative adsorption of water molecules. To fulfill this requirement, we modified Pt(111) and Pt-island/Pt(111) surfaces by depositing Ni-(hydr)oxide clusters (38), as the 3d-transition metal oxides might be even more active for water dissociation than Pt defect sites (31). The facile water dissociation properties of Ni(hydroxy) oxides relative to other transition metal oxides have been well established (26–28), motivating the use of Ni(OH)2 for this work. The local symmetry, the oxidation state of Ni atoms, and the number and identities of nearest-neighbor atoms and the distances between them were determined by in situ x-ray absorption spectroscopy (XAS) measurements (41, 42). For example, from the analysis of the x-ray absorption near-edge structure (XANES) and extended x-ray absorption fine structure (EXAFS) of the XAS spectra (inset of Fig. 1D; see also fig. S3), we found Ni-O and Ni-Ni bond distances of 2.05 T 0.01 Å and 3.08 T 0.01 Å, and we also determined that Ni remains mostly in the +2 valence state, even after multiple hours of holding the electrode potential at –0.1 V. Furthermore, from the edge shift (defined as the half-height energy of the normalized XANES edge step), we concluded that between –0.1 Vand +0.8 V, the change in the oxidation state of Ni is less than 0.5. These results suggest that stable Ni(OH)2 clusters are the predominant hydr(oxide) form on the Pt(111) and Pt-islands/Pt(111) surfaces, especially in the HER potential region.
1258
Because the octahedral symmetry of the a and/or b phases of Ni(OH)2 prevents p-d hybridization, the prominent pre-edge from 1s → 3d transitions implies that the Ni(OH)2 species are rich in defects that, according to prior reports (31–33), are particularly active for dissociative adsorption of water molecules. The surface morphologies of Ni(OH)2/Pt(111) and Ni(OH)2/Pt-islands/Pt(111) were probed by STM and in situ surface x-ray crystal truncation rod (CTR) measurements (43, 44) (see fig. S4 for the CTR data). Although atomic resolution could
not be obtained, the STM image in Fig. 1C clearly shows that the Ni(OH)2 clusters are randomly distributed across the (111) terraces. All Ni(OH)2 clusters exhibited hemisphere-like shapes, with characteristic diameters of ~8 to 10 nm and heights of ~0.7 nm, the latter corresponding to two layers of Ni(OH)2. This result indicates that the oxide exhibits Volmer-Weber (VW)–type growth whereby 3D clusters of Ni(OH)2 grow even at the lowest coverages (45). VW growth, in turn, is possible if the heat of adsorption of Ni(OH)2 on Pt is lower than the cohesive energy of Ni(OH)2.
Fig. 2. (A) STM image (60 nm by 60 nm) and CV trace of the Ni(OH)2/Pt-islands/Pt(111) surface. Clusters of Ni(OH)2 in the STM image appear ellipsoidal with particle sizes between 4 and 12 nm. (B) Comparison of HER activities with Pt(111) as the substrate. Incremental improvements in activities for the HER in 0.1 M KOH from the unmodified Pt(111) surface are shown for the hierarchical materials [ad-islands, Ni(OH)2, and their combination] as well as the double layer (addition of Li+ cations). The activity for the unmodified Pt(111) surface in 0.1 M HClO4 is shown for reference. Dashed arrow shows the activity trend.
Fig. 3. Schematic representation of water dissociation, formation of M-Had intermediates, and subsequent recombination of two Had atoms to form H2 (magenta arrow) as well as OH– desorption from the Ni(OH)2 domains (red arrows) followed by adsorption of another water molecule on the same site (blue arrows). Water adsorption requires concerted interaction of O atoms with Ni(OH)2 (broken orange spikes) and H atoms with Pt (broken magenta spikes) at the boundary between Ni(OH)2 and Pt domains. The Ni(OH)2-induced stabilization of hydrated cations (AC+) (broken dark blue spikes) likely occurs through noncovalent (van der Waals–type) interactions. Hydrated AC+ can further interact with water molecules (broken yellow spikes), altering the orientation of water as well as the nature and strength of interaction of the oxide with water.
2 DECEMBER 2011
VOL 334
SCIENCE
www.sciencemag.org
REPORTS The surface coverage of Ni(OH)2 on Pt(111) is estimated from the STM image by measuring the area covered by the particles on the Pt(111) substrate. According to such analysis, the cluster density reached a maximum at a surface coverage of ~35%. Upon comparing STM images of Pt(111) and Ni(OH)2/Pt(111) surfaces (Fig. 1, A and C), however, it was not clear whether isolated defects or flat Pt(111) terraces were the preferred nucleation sites for Ni(OH)2. This information was more accessible by comparing STM images of the Ni(OH)2-free (Fig. 1B) and Ni(OH)2-covered Pt-island/Pt(111) (Fig. 2A) electrodes. The STM image in Fig. 2A was acquired after deposition of Ni(OH)2 on a Pt(111) surface modified by ~0.2 monolayer (ML) of Pt islands. We observed formation of both 3D Ni(OH)2 clusters (having a predominantly ellipsoidal shape) and oxide-free terraces. The clusters had approximately constant heights of ~0.8 nm but diameters ranging from 4 to 12 nm. The same STM image, however, revealed no visible presence of 2D Pt islands, which suggests
that Ni(OH)2 preferentially nucleates on the Pt surface defects and that most of the Pt islands are covered by Ni(OH)2. Having established the nature and morphology of the Ni(OH)2 clusters, we briefly summarize the role of the Ni(OH)2 clusters in the formation of Hupd and OHad adlayers on Pt(111) and Ptislands/Pt(111) electrodes. Addition of Ni(OH)2 on the surface of Pt(111) as well as on the Pt(111) surface covered with Pt islands led to a decrease (~35%) in the coverage of Hupd. This finding suggests that the Ni(OH)2 clusters selectively block the Pt sites corresponding to Hupd. Furthermore, the two sharp Hupd peaks characteristic of hydrogen adsorption/desorption on the Pt(111) electrode modified by the 2D Pt islands are completely suppressed on the surface covered by the 3D Ni(OH)2 clusters (Fig. 2A). This is additional evidence, consistent with the STM results, that defects serve as the nucleation centers for electrodeposition of Ni(OH)2. We therefore conclude that Pt islands are predominantly covered by Ni(OH)2. In contrast to the Hupd potential re-
Fig. 4. (A) STM image (50 nm by 50 nm) and CV trace for Ni(OH)2/Pt-islands/Pt(110) single-crystal surface. STM image shows Ni(OH)2 randomly distributed on inherently rough Pt(110).. (B) Comparison of HER activities with Pt(110) as the substrate. Incremental improvements in activities for the HER in 0.1 M KOH from the unmodified Pt(110) surface are shown for the hierarchical materials [Ni(OH)2 and the combination of ad-islands with Ni(OH)2] as well as the double layer (addition of Li+ cations). The activity for the unmodified Pt(110) surface in 0.1 M HClO4 is shown for reference. Dashed arrow shows the activity trend. (C) Transmission electron micrograph image (50 nm by 50 nm) and corresponding CV trace of Ni(OH)2-free Pt-nano catalysts (TKK) with an average particle size of 5 nm. (D) Comparison of HER activities with commercial nanocatalyst Pt/C (TKK) as the substrate. Incremental improvements in activities for the HER in 0.1 M KOH from unmodified Pt/C are shown for the hierarchical materials [surface covered with Ni(OH)2] as well as the double layer (addition of Li+ cations). The activity for the unmodified Pt/C surface in 0.1 M HClO4 is shown for reference. Dashed arrow shows the activity trend. All the current densities for the TKK catalyst system are normalized by the geometric surface area of the glassy carbon substrate. www.sciencemag.org
SCIENCE
VOL 334
gion, an enhanced adsorption of OHad, which is accompanied by irreversible reduction of OHad on the negative-going sweep, is observed on both electrodes, arising from the higher oxophilicity of the surface elements. Although in the presence of Ni(OH)2 clusters there are 35% fewer Pt sites available for the HER than on the bare Pt(111) substrate, the Ni(OH)2/Pt(111) electrode is ~7 times as active for the HER relative to Pt(111) (Fig. 1D). Moreover, Fig. 2B shows that the activity is further enhanced [by a factor of ~8 relative to bare Pt(111)] on the Ni(OH)2/Pt-island/Pt(111) surface; at 6 mA/cm2, the difference in overpotential between the HER in alkaline and acid solutions is reduced to only 100 mV. Clearly, then, on both surfaces, Ni(OH)2 must play a promoting role in the dissociation of water and thereby enhances the rate of formation of Had intermediates on the metal surface. As schematically depicted in Fig. 3, we propose that water adsorption requires concerted interaction of O atoms with Ni(OH)2 and H atoms with Pt at the boundary between Ni(OH)2 and Pt domains. Water adsorption is then followed by water dissociation and hydrogen adsorption (Had) on the nearby vacant Pt sites. Finally, two Had atoms on the Pt surface recombine to form H2 (H2 desorption step) and OH– desorbs from the Ni(OH)2 domains, followed by adsorption of another water molecule on the same site. From a surface reactivity standpoint, fruitful kinetic synergy (bifunctionality) between Ni(OH)2 and Pt appears to be the key to maximizing the rate of the HER. As shown in Fig. 2B, this bifunctionality in turn brings the activity of the HER in alkaline solutions very close to the activity of Pt in acid solutions. To verify this conclusion, we have also compared the HER on Au(111) and Ni(OH)2/Au(111) in alkaline solution. The relatively weak interaction between Au and Had offsets the benefit of the enhanced water dissociation at the Au/Ni(OH)2 interface (38). As a result, the rate of the HER on the Ni(OH)2/Au(111) surface is much lower than on the Pt(111)/Ni(OH)2 surface. This further emphasizes the importance of choosing the correct metal oxide–metal pairs in optimizing the kinetics of the HER. In what follows, we demonstrate that the nature of interactions in double layer is equally important, and the rate-determining Volmer step can be further enhanced by tuning the double-layer properties. Several recent studies have unambiguously shown that the rate of electrochemical reactions on Pt in alkaline solutions is controlled by the presence of alkali-metal cations (AC+) (20, 21). However, these effects have been entirely restricted to the potential region of a critical OHad coverage (for electrode potential E > 0.6 V), the latter species serving to stabilize hydrated cations in the compact part of the double layer through noncovalent (van der Waals–type) interactions. This stabilization leads to the formation of OHad-AC+(H2O)x complexes that can either decrease the reactivity of Pt by blocking
2 DECEMBER 2011
1259
REPORTS the active sites for adsorption of reactants such as O2, H2, and CH3OH (20, 21) or, as in the case of the CO oxidation reaction, improve the reactivity of Pt via enhanced adsorption of OHad (22). For the present reaction, the key questions are whether hydrated alkali cations on the Ni(OH)2/Pt-islands/Pt(111) surface can be stabilized through interactions with OH species in the Ni(OH)2 cluster and, if so, whether these hydrated cations could affect the alkaline HER kinetics. For these purposes, the effect of hydrated Li cations was probed mainly because, in alkaline environments, Li+ is known to interact with H2O and OHad more strongly than K+. In line with the previous electrochemical report (20), we found that Li+ cations have no effect on the HER on Pt(111) surfaces. However, the results in Fig. 2B revealed that the HER on Ni(OH)2/Pt-islands/Pt(111) is enhanced by almost a factor of 2 in the presence of Li+ cations. This increase in activity has substantially narrowed the gap between the rates of HER on Pt in acid and alkaline solution; more specifically, inspection of Fig. 2B shows that, at 5 mA/cm2, the difference in overpotential between acid and alkaline environments is narrowed to only 35 mV. The fact that the activity of the HER is affected by the nature of alkali metal cations strongly suggests that Ni(OH)2-Li+-OH-H complexes are present in the compact portion of the double layer. The presence of this complex does not completely explain the factor of 2 increase in HER activity. It is likely that the probability of water dissociation is enhanced via possible Li+-induced steric and/or electronic effects on the interfacial water structure and reactivity, as shown schematically in Fig. 3. Thus, Ni(OH)2 plays a dual role: In addition to assisting with water dissociation, it also provides an anchor to hold the beneficial Li+ ions in the compact portion of the double layer. Having established the methodology of tuning interfacial activity for the HER on Pt(111), we demonstrate that the very same guiding principles for accelerating the Volmer reaction step in alkaline solutions are equally applicable to Pt(110). For simplicity, only the CV and STM data for the most active system, Ni(OH)2/Ptislands/Pt(110), are shown in Fig. 4A. The general characteristics (both structural and electrochemical) are similar to what was observed for the corresponding Pt(111) systems. The current denities for the HER on Pt(110) and Pt-nanocatalyst systems are presented in the logarithmic Tafel form (Fig. 4, B and D) to offer the most condensed representation of surface characterization and polarization curves. As expected, the systematic modification of Pt(110), first with Pt islands and then with Ni(OH)2, exhibits an HER activity trend (Fig. 4B) with the same order as the driving force for dissociative adsorption of water molecules, as discussed above for the Pt(111) systems: Pt(110) < Ni(OH)2/Pt(110)
