Electron Transfer Rate vs Recombination Losses in Photocatalytic H2

Electron Transfer Rate vs Recombination Losses in Photocatalytic H2 Generation on PtDecorated CdS Nanorods Thomas Simon, Michael T. Carlson, Jacek K. Stolarczyk,* and Jochen Feldmann Photonics and Optoelectronics Group, Department of Physics and Center for Nanoscience (CeNS), Ludwig-Maximilians-Universität München, Amalienstraße 54, 80799 Munich, Germany Nanosystems Initiative Munich (NIM), Schellingstraße 4, 80799 Munich, Germany S Supporting Information *

ABSTRACT: Cadmium chalcogenide nanocrystals combined with co-catalyst nanoparticles hold promise for efficient solar to hydrogen conversion. Despite the progress, achieving high efficiency is hampered by high charge recombination rates and sample degradation. Here, we vary the decoration of platinum nanoparticles on CdS nanorods to demonstrate the important role of pathways for the photoelectrons to the co-catalyst. Contrary to expectations, the shortening of the path, by increasing the number of co-catalyst particles, increases the transfer rate but decreases the photocatalytic performance. This is because subsequent electron transfer to the acceptor is much slower; therefore, the recombination rate with the nearby holes increases. We show that with tip-decorated nanorods, the quantum yield of H2 production can reach and sustain nearly 90%, provided an efficient mechanism of mediated hole extraction is employed. The approach demonstrates that highly efficient photocatalysts may be prepared with only a minimal amount of co-catalyst and thereby suggests future pathways for solar to H2 conversion with semiconductor nanocrystals.

H

particularly promising class of NCs. They combine facile access to the surface and quantum confinement in the radial direction with capacity for spatial charge separation in the axial direction.9 The decoration of II−VI semiconductor NRs with metal nanoparticles10,11 enables electron transfer to the particles due to favorable alignment of the conduction band with the Fermi energy of the metal. This not only enables longlived charge separation across the semiconductor−metal interface but also confers catalytic properties for water reduction on the heterostructure.9,12,13 Nonetheless, as charge balance must be maintained, achieving high efficiency of H2 generation is contingent on balancing electron and hole transfer rates. Both rates are relevant, even if only the products of the reduction half-reaction are of interest while the other side is replaced with a surrogate reaction involving a sacrificial agent (cf.Figure 1). In this context, previous studies on Cd chalcogenide nanostructures have suggested the limiting role of hole transfer.14−17 Localization of the photogenerated holes away from the reaction site, instead of random (isotropic) hole trapping on the surface, has been proposed to address this

arvesting sunlight could provide a sustainable and fossil-fuel-independent energy source, considering that just 1 h of incident solar radiation exceeds annual global demand.1 To this end, impressive strides have been made in fabrication of efficient photovoltaic devices. Nonetheless, the intermittent and highly variable nature of solar illumination means that the direct solar to electricity conversion cannot meet the demand at all times due to inherent lack of storage capability. Photocatalytic solar fuel generation, wherein the incident photon energy drives uphill chemical reactions, offers a possible solution to this problem. The energy is stored in the chemical bonds of the products and is easily retrieved by combustion. Hydrogen, a product of photocatalytic water reduction,2,3 and methane or methanol, the desired products of CO2 reduction,4 are the most common examples. The viability of this approach for storage is manifested by the energy density of solar fuels, which is up to 3 orders of magnitude higher than that of present day batteries.5 Since its inception more than 40 years ago,6 the field of photocatalytic H2 generation with colloidal semiconductor nanocrystals (NCs) has grown significantly.2,7 It has benefited from advances in colloidal chemistry that enabled synthesis of NCs of desired size, shape, surface chemistry, and optoelectronic properties.8 Colloidal nanorods (NRs), several nanometers in diameter and tens of nanometers in length, are a © XXXX American Chemical Society

Received: September 23, 2016 Accepted: November 6, 2016 Published: November 7, 2016 1137

DOI: 10.1021/acsenergylett.6b00468 ACS Energy Lett. 2016, 1, 1137−1142

Letter

http://pubs.acs.org/journal/aelccp

Letter

ACS Energy Letters

In this Letter, we varied the decoration techniques of Pt nanoparticles on CdS NRs and applied redox mediators to manipulate both carrier pathways in the photocatalyst. The cocatalyst nanoparticles were deposited either randomly on the NR surface or selectively at the tip. Using time-resolved transient absorption measurements, we demonstrate that shortening the distance for electron transfer to the co-catalyst can be counterproductive because the benefit of increased transfer rate is outweighed by the concurrent increase in the charge recombination rate. We show that with tip decoration, which balances the electron transfer rate and recombination losses, and with the use of hole transfer mediators, the external quantum yield of H2 production can reach 90%. This is an unprecedented value for such a simple system composed of only CdS NRs and Pt nanoparticles. The CdS NRs were synthesized according to a wellestablished method,24,25 the details of which are given in the Supporting Information. The dimensions of the NRs were determined by TEM analysis as 3−5 nm in diameter and 50− 60 nm in length (see Figure S1). The diameter was smaller than the exciton Bohr radius of CdS (5.5 nm);26,27 therefore, the rods exhibited effects of quantum confinement in the radial direction. Specifically, the absorption onset observed at 470 nm (see Figure 2a) was accordingly blue-shifted with respect to the bulk value (ca. 510 nm, corresponding to a band gap of 2.4 eV). In addition, the absorption of the 1S exciton was strongly enhanced, leading to the dominant peak at 460 nm. Platinum nanoparticles were attached to the CdS NRs with two different deposition techniques (see Figure 2b), which ensured good contact at the semiconductor−metal interface. In the first approach, a single particle was selectively grown at the

Figure 1. Scheme of charge carrier transport for sacrificial hydrogen production involving hole transfer to a scavenger molecule.

issue.18 In particular, CdSe−CdS dot-in-rod structures can localize the photohole in the CdSe seed, owing to the higher valence band edge of CdSe with respect to that of CdS,19,20 and have accordingly exhibited high photocatalytic efficiency.21,22 Alternatively, the rate of hole removal can be enhanced by either increasing the overpotential for the process14 or using a suitable redox mediator.23 These are intriguing developments because they seemingly contradict the common expectation that the final catalytic reaction, that is, electron transfer from the co-catalyst to the acceptor, is the slowest process.20 Therefore, it remains unclear how the transfer rates in the complete electron pathway, first to the co-catalyst and then to the acceptor molecule, can be optimized to improve the photocatalytic yield.

Figure 2. (a) Absorption spectra of bare (black line) and platinum-decorated CdS NRs in water. (b) Synthetic scheme of the decoration procedures and phase transfer to water. The red background indicates an organic environment, while the blue background indicates an aqueous environment. (c) TEM image of tip-decorated CdS NRs. (d) TEM image of cluster-decorated CdS NRs. 1138

DOI: 10.1021/acsenergylett.6b00468 ACS Energy Lett. 2016, 1, 1137−1142

Letter

ACS Energy Letters

Figure 3. (a) Transient absorption spectra of bare CdS NRs at indicated delay times measured with 400 nm excitation. (b) Relaxation of the bleach of the 1S exciton in transient absorption for all three structures. (c) Transient absorption spectra of randomly decorated CdS NRs at indicated delay times measured with 400 nm excitation. (d) Hydrogen generation by CdS NRs under illumination in neutral aqueous solution in the presence of sulfite ions.

3a). A distinct minimum with ΔOD < 0 at the position of the 1S exciton could be observed, which formed due to thermalization at about 300 fs after excitation. The minimum can be understood on the basis of phase-space filling by the excited carriers that occupy the excitonic states after relaxing from the band continuum. This bleaches the transition for further photons arriving. Small spectral regions with induced absorption (ΔOD > 0) result from a red shift of the absorption features possibly due to band gap renormalization by the excited carriers.31 The transient absorption behavior in CdS is mainly attributed to electrons because of their lower effective mass than that of the holes and due to the degeneracy of the valence band.32 Therefore, the time dynamics of the bleach of the 1S exciton predominantly corresponds to the pathway of the photoelectrons. The spectrum in Figure 3b (black line) shows the temporal relaxation of −ΔOD normalized by the maximum depth of the 1S bleach. After an initial decay to about 90% of the original signal within the first few picoseconds, the bleach remains stable until 1.1 ns, which was the longest delay time acquired in our experiment. This observation is common for many cadmium chalcogenide nanoparticles. It is considered that the photohole trapping shortly after illumination significantly delays the recombination.33,34 Therefore, the minor initial decay, observed for early times in Figure 3a, can be attributed to fast hole trapping and the much longer decay to the recombination of the photoelectron with the trapped hole. At high pump intensity, Auger recombination constitutes an additional relaxation pathway and leads to faster exciton bleach decay (see Figure S3).35 In order to avoid complicating the charge carrier dynamics, we kept the pump fluence in the transient absorption measurements at 14 mW/cm2, so that on average one electron−hole pair would be created per NR within the CdS exciton lifetime.

tip of the NR by thermal decomposition of Pt acetylacetonate precursor in organic solvent (further details in the Supporting Information), following a procedure published previously.28,29 The decorated NRs were subsequently transferred to the aqueous phase by ligand exchange with cysteine. TEM measurements (Figure 2c) confirmed the presence of single particles at one end of each NR, while only a very small fraction of the NRs either had Pt nanoparticles at both tips (so-called dumbbell structures) or were not decorated at all. As shown in Figure 2a, the absorption spectrum markedly changed for wavelengths longer than the 1S exciton. This broad feature has been assigned to a d−sp interband transition within the Pt particles.30 Importantly, neither the spectral position nor the shape of the exciton is affected by decoration, implying that the morphology of the semiconductor NC remained unchanged after deposition. In the second approach, the two preparative steps were performed in the inverse order. The bare NRs were first transferred to water using cysteine and then decorated with Pt by photodeposition.11,25 This procedure is well-known to result in nucleation and growth of multiple small Pt clusters over the whole surface of the NR,25 as shown in the highresolution TEM image in Figure S2. In addition, a small number of larger, 4−5 nm in diameter, randomly located particles can be seen in the TEM images (Figure 2d), in agreement with earlier studies.25 During photodeposition, the growing cluster acts as an electron sink from CdS, resulting in facile electron transfer from CdS to Pt also in the formed heterostructure. The absorption spectrum is very similar to the one obtained with the first approach, with no change in the excitonic peak. Electron transfer to the co-catalyst can be monitored with femtosecond transient absorption spectroscopy. For reference, we first excited the bare CdS NRs in an aqueous dispersion with a 100 fs laser pulse at 400 nm and acquired a differential absorption spectrum with a broad-band probe pulse (see Figure 1139

DOI: 10.1021/acsenergylett.6b00468 ACS Energy Lett. 2016, 1, 1137−1142

Letter

ACS Energy Letters

nanoparticles can in fact be more efficient.38 This observation is potentially of great benefit for photocatalytic systems that require costly noble metal catalysts. It can also be noted in Figure 3d that the H2 formation rate clearly decreases for the randomly decorated sample during illumination. This is because of a degradation of the catalytic material, likely due to photooxidation by the hole. The mechanism could involve either oxidation of the lattice sulfide ions (E° = 0.48 V at pH 0)39 leading to dissolution of the NRs or oxidation of the thiol moieties of the cysteine ligands (E° = 0.073 V at pH 0)40 leading to their desorption. Unprotected NCs tend to aggregate, thus decreasing the photocatalytic efficiency. Highly reducing hole scavengers, such as the sulfite anions in the experiment,41 inhibit both pathways; however, the degradation is clearly not entirely prevented. These results emphasize the need to remove the photogenerated hole from the NC as quickly as possible. We have adopted the redox mediator mechanism previously applied to CdS randomly decorated with nickel particles.23 In this approach, the hole is not removed by the scavenger itself but instead by hydroxyl anions that subsequently form hydroxyl radicals. The latter then oxidize the scavenger away from the semiconductor surface. This means that the redox couple OH•/ OH− relays the hole in a two-step process that is much faster than a direct transfer to the scavenger (see Figure 4a). This mechanism was applied to the tip-decorated NR, shown to be more efficient without mediated hole removal. To this end, we illuminated the sample in highly alkaline conditions (pH 14.7) in the presence of either sodium sulfite or ethanol as hole scavengers. As presented in Figure 4b, the photocatalytic activity in an alkaline environment was markedly enhanced with

Decoration of the NRs with Pt at the tip retains the features of the spectra, indicating that there is no response from the Pt particles (cf. Figure S4). By contrast, the dynamics of ΔOD(t), shown in Figure 3b (orange line), significantly changes. Specifically, the decay rate increases, exhibiting a very fast initial component down to about 30% of the starting value. Afterward, it decays on a slower time scale. The increase can be explained with the thermodynamically feasible transfer of the photoelectron to the Pt particle that constitutes an additional decay channel, confirming good contact between the NR and the Pt particle.21 This was verified by experiments performed in the presence of methyl viologen, MV2+, a well-known electron scavenger.20 In this case, the exciton bleach decays rapidly and monoexponentially to zero (see Figure S5). The rate of decay is independent of the presence of Pt, in agreement with a direct electron transfer pathway from CdS to MV2+. In CdS NRs decorated with multiple Pt clusters, the distance to the co-catalyst is much shorter. Hence, it was expected that the rate of electron transfer would be enhanced. Indeed, as shown in Figure 3b,c, relaxation of the 1S exciton bleach becomes much faster in comparison to that of the NRs with a single Pt particle at the tip. Only 10% of the ΔOD signal remains after 75 ps, which then decays with a rate similar to that of the other samples. The bleach decay is clearly nonexponential in Pt-decorated samples, but a comparison of half-life times (82 vs 4 ps for tip and random decoration, respectively) reflects the substantial increase in the electron transfer rate to the co-catalyst in the latter structure. Moreover, the results also show that the number of electrons still remaining on the semiconductor and not transferred to the cocatalyst is diminished for decoration with multiple Pt clusters. Consequently, in terms of the electron pathway, the structure with multiple co-catalyst particles appears more promising because more electrons are transferred with a higher rate on the picosecond time scale. To study the effects of electron transfer on photocatalytic hydrogen generation, we illuminated all three systemsbare, tipped, and randomly decorated CdS NRsin aqueous solution at pH ≈ 9. Sodium sulfite, Na2SO3, was used as a hole scavenger. As shown in Figure 3d, hydrogen evolution could be observed for both hybrid structures with Pt cocatalyst. As expected, bare CdS did not produce any measurable amount of H2. Interestingly, the randomly decorated structures performed worse than the tip-decorated structures. Only 2.6 μmol of H2 was produced by the former over 6 h of illumination (see details in the Supporting Information), while 7.4 μmol was produced by the latter. These results correspond to external quantum yields of 0.6 and 1.7% (averaged over 6 h), respectively. The maximum values were 0.9 and 2.2%. These findings appear surprising in the context of the electron transfer rates. The differences in the sizes of the Pt nanoparticles cannot account for these results because the optimal size of the particles was found in earlier studies to be less than 2 nm,36,37 with a decrease in performance for larger particles, such as those located at the tips of the NRs. They can, however, be explained considering the multistep character of the process. This suggests that the rate-limiting step for the electron is not transfer to the catalyst but subsequent transfer from the catalyst to the acceptor. In effect, the short distance between the trapped hole and the electron on the Pt particle results in a faster recombination rate, outweighing the benefits of faster electron transfer. Interestingly, the result proves that the increase in the amount of co-catalyst can be counterproductive, and semiconductor NC systems with very few co-catalyst

Figure 4. (a) Schematic of electron and hole transfer from CdS to hydrogen and the hole scavenger (HS), respectively. Both direct and mediated hole transfers are shown. (b) H2 generation of tipdecorated CdS NRs at neutral pH with SO32− as a hole scavenger (orange line) and in a highly alkaline environment with SO32− (blue line) and ethanol (green line) as hole scavengers. 1140

DOI: 10.1021/acsenergylett.6b00468 ACS Energy Lett. 2016, 1, 1137−1142

ACS Energy Letters

■

both scavengers with respect to values at neutral pH. The average formation rate over 5.5 h reached 28 and 14.7 mmol/g· h for ethanol and sulfite as scavengers, respectively. These rates correspond to 83 and 44% external quantum efficiency of the incident photon to hydrogen conversion. The highest rates over 1 h of illumination were measured to be 33.4 and 21.8 mmol/g· h with these two scavengers. The former value corresponds to over 90% quantum efficiency. Interestingly, a similarly highefficiency result has been reported recently for CdSe−CdS dotin-rod structures, which take advantage of localization of the hole in the CdSe dot.42 We show here that a near-unity efficiency can be achieved provided only that the holes are removed quickly and are not in the vicinity of the electrons located on the Pt nanoparticles. Even though proton reduction on the co-catalyst takes more time, extraction of the hole by the OH•/OH− couple means that recombination is strongly suppressed and high quantum yields are possible. Moreover, hole extraction also prevents degradation of the photocatalyst. This is illustrated in Figure S6. At neutral pH, the absorption spectrum in the region of the 1S exciton changes significantly after 5 h of illumination. The total absorption decreases, suggesting partial dissolution of the NCs. In addition, the excitonic peak broadens significantly. This is likely due to oxidation of the ligands and a subsequent loss of colloidal stability and defined morphology. In contrast, the sample at high pH preserves the absorption properties after illumination. The external quantum efficiency denotes the ratio of incident photon flux to twice the rate of hydrogen molecules formed but does not directly reflect the power conversion efficiency of a phototocatalyst. To evaluate the latter, we calculated the difference between proton/water and acetaldehyde/ethanol redox potentials (−0.22 V),40 which implies that the stored energy is 0.22 eV per photon. Taking into account that the incident photon energy is 2.78 eV (447 nm), this translates into 7% (=0.9 × 0.22/2.78) power conversion efficiency at the highest quantum yield. This demonstrates that the process can be viable for solar fuel production, even in the presence of sacrificial agents. The results here imply that the process is not specific to any reducing agent, which holds promise for finding sustainable material with further research. In summary, the results demonstrate the important role of pathways for photoelectrons to the co-catalyst. Contrary to expectations, the shortening of the path, by increasing the number of co-catalyst particles, increases the transfer rate but actually decreases the photocatalytic performance. This is because the subsequent electron transfer to the acceptor is much slower. In effect, the recombination rate with the nearby holes increases, thereby canceling any benefit of the increased electron transfer rate. In this context, the improvement can be engineered by removing the photoholes so that the electrons on the Pt have no hole with which to recombine. The added benefit is that degradation by the photohole is limited. Accordingly, we have shown that with appropriate design of the photocatalyst, which takes the balance between the rates of electron transfer and recombination into account, mediated hole transfer can increase the quantum yield of H2 production up to 90%. The results also demonstrate that the amount of cocatalyst can be minimized without detriment to the efficiency. On the contrary, the efficiency can even be improved with only a single co-catalyst particle at the extremity of the NC, suggesting future pathways for solar to hydrogen conversion with semiconductor NCs.

Letter

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsenergylett.6b00468. Experimental procedures, additional TEM images, and transient and static absorption spectra (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Bavarian State Ministry of Science, Research, and Arts through the grant “Solar Technologies go Hybrid (SolTech)” and the European Commission through the FP7-NMP program (project UNION, Grant No. 310250). The authors thank Dr. Markus Dö b linger for HR-TEM measurements and Christoph Hohmann (Nanosystems Initiative Munich) for his support in graphics design.

■

REFERENCES

(1) Lewis, N. S.; Nocera, D. G. Powering the Planet: Chemical Challenges in Solar Energy Utilization. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 15729−15735. (2) Chen, X.; Shen, S.; Guo, L.; Mao, S. S. Semiconductor-based Photocatalytic Hydrogen Generation. Chem. Rev. 2010, 110, 6503− 6570. (3) Osterloh, F. E. Inorganic Nanostructures for Photoelectrochemical and Photocatalytic Water Splitting. Chem. Soc. Rev. 2013, 42, 2294−2320. (4) Habisreutinger, S. N.; Schmidt-Mende, L.; Stolarczyk, J. K. Photocatalytic Reduction of CO2 on TiO2 and Other Semiconductors. Angew. Chem., Int. Ed. 2013, 52, 7372−7408. (5) Bruce, P. G.; Freunberger, S. A.; Hardwick, L. J.; Tarascon, J.-M. Li-O2 and Li-S Batteries with High Energy Storage. Nat. Mater. 2012, 11, 19−29. (6) Fujishima, A.; Honda, K. Electrochemical Photolysis of Water at a Semiconductor Electrode. Nature 1972, 238, 37−38. (7) Maeda, K.; Domen, K. Photocatalytic Water Splitting: Recent Progress and Future Challenges. J. Phys. Chem. Lett. 2010, 1, 2655− 2661. (8) Vaneski, A.; Susha, A. S.; Rodríguez-Fernández, J.; Berr, M.; Jäckel, F.; Feldmann, J.; Rogach, A. L. Hybrid Colloidal Heterostructures of Anisotropic Semiconductor Nanocrystals Decorated with Noble Metals: Synthesis and Function. Adv. Funct. Mater. 2011, 21, 1547−1556. (9) Wu, K.; Lian, T. Quantum Confined Colloidal Nanorod Heterostructures for Solar-to-Fuel Conversion. Chem. Soc. Rev. 2016, 45, 3781−810. (10) Costi, R.; Saunders, A. E.; Banin, U. Colloidal Hybrid Nanostructures: A New Type of Functional Materials. Angew. Chem., Int. Ed. 2010, 49, 4878−4897. (11) Dukovic, G.; Merkle, M. G.; Nelson, J. H.; Hughes, S. M.; Alivisatos, A. P. Photodeposition of Pt on Colloidal CdS and CdSe/ CdS Semiconductor Nanostructures. Adv. Mater. 2008, 20, 4306− 4311. (12) Wilker, M. B.; Schnitzenbaumer, K. J.; Dukovic, G. Recent Progress in Photocatalysis Mediated by Colloidal II-VI Nanocrystals. Isr. J. Chem. 2012, 52, 1002−1015. (13) Xu, Y.; Huang, Y.; Zhang, B. Rational Design of Semiconductorbased Photocatalysts for Advanced Photocatalytic Hydrogen Produc1141

DOI: 10.1021/acsenergylett.6b00468 ACS Energy Lett. 2016, 1, 1137−1142

Letter

ACS Energy Letters tion: the Case of Cadmium Chalcogenides. Inorg. Chem. Front. 2016, 3, 591−615. (14) Berr, M. J.; Wagner, P.; Fischbach, S.; Vaneski, A.; Schneider, J.; Susha, A. S.; Rogach, A. L.; Jackel, F.; Feldmann, J. Hole Scavenger Redox Potentials Determine Quantum Efficiency and Stability of Ptdecorated CdS Nanorods for Photocatalytic Hydrogen Generation. Appl. Phys. Lett. 2012, 100, 223903−3. (15) Wu, K.; Du, Y.; Tang, H.; Chen, Z.; Lian, T. Efficient Extraction of Trapped Holes from Colloidal CdS Nanorods. J. Am. Chem. Soc. 2015, 137, 10224−10230. (16) Khon, E.; Lambright, K.; Khnayzer, R. S.; Moroz, P.; Perera, D.; Butaeva, E.; Lambright, S.; Castellano, F. N.; Zamkov, M. Improving the Catalytic Activity of Semiconductor Nanocrystals through Selective Domain Etching. Nano Lett. 2013, 13, 2016−2023. (17) Wu, K.; Chen, Z.; Lv, H.; Zhu, H.; Hill, C. L.; Lian, T. Hole Removal Rate Limits Photodriven H2 Generation Efficiency in CdS-Pt and CdSe/CdS-Pt Semiconductor Nanorod−Metal Tip Heterostructures. J. Am. Chem. Soc. 2014, 136, 7708−7716. (18) Wu, K.; Rodríguez-Córdoba, W. E.; Liu, Z.; Zhu, H.; Lian, T. Beyond Band Alignment: Hole Localization Driven Formation of Three Spatially Separated Long-Lived Exciton States in CdSe/CdS Nanorods. ACS Nano 2013, 7, 7173−7185. (19) Amirav, L.; Alivisatos, A. P. Photocatalytic Hydrogen Production with Tunable Nanorod Heterostructures. J. Phys. Chem. Lett. 2010, 1, 1051−1054. (20) Zhu, H.; Song, N.; Lv, H.; Hill, C. L.; Lian, T. Near Unity Quantum Yield of Light-Driven Redox Mediator Reduction and Efficient H2 Generation Using Colloidal Nanorod Heterostructures. J. Am. Chem. Soc. 2012, 134, 11701−11708. (21) Wu, K.; Zhu, H.; Liu, Z.; Rodríguez-Córdoba, W.; Lian, T. Ultrafast Charge Separation and Long-Lived Charge Separated State in Photocatalytic CdS−Pt Nanorod Heterostructures. J. Am. Chem. Soc. 2012, 134, 10337−10340. (22) Wu, K. F.; Li, Q. Y.; Du, Y. L.; Chen, Z. Y.; Lian, T. Q. Ultrafast Exciton Quenching by Energy and Electron Transfer in Colloidal CdSe Nanosheet-Pt Heterostructures. Chem. Sci. 2015, 6, 1049−1054. (23) Simon, T.; Bouchonville, N.; Berr, M. J.; Vaneski, A.; Adrović, A.; Volbers, D.; Wyrwich, R.; Döblinger, M.; Susha, A. S.; Rogach, A. L.; et al. Redox Shuttle Mechanism Enhances Photocatalytic H2 Generation on Ni-decorated CdS Nanorods. Nat. Mater. 2014, 13, 1013−1018. (24) Saunders, A. E.; Ghezelbash, A.; Sood, P.; Korgel, B. A. Synthesis of High Aspect Ratio Quantum-Size CdS Nanorods and Their Surface-Dependent Photoluminescence. Langmuir 2008, 24, 9043−9049. (25) Berr, M.; Vaneski, A.; Susha, A. S.; Rodriguez-Fernandez, J.; Doblinger, M.; Jackel, F.; Rogach, A. L.; Feldmann, J. Colloidal CdS Nanorods Decorated with Subnanometer Sized Pt Clusters for Photocatalytic Hydrogen Generation. Appl. Phys. Lett. 2010, 97, 093108−3. (26) Vossmeyer, T.; Katsikas, L.; Giersig, M.; Popovic, I. G.; Diesner, K.; Chemseddine, A.; Eychmueller, A.; Weller, H. CdS Nanoclusters: Synthesis, Characterization, Size Dependent Oscillator Strength, Temperature Shift of the Excitonic Transition Energy, and Reversible Absorbance Shift. J. Phys. Chem. 1994, 98, 7665−7673. (27) Puthussery, J.; Lan, A.; Kosel, T. H.; Kuno, M. Band-Filling of Solution-Synthesized CdS Nanowires. ACS Nano 2008, 2, 357−367. (28) Mokari, T.; Rothenberg, E.; Popov, I.; Costi, R.; Banin, U. Selective Growth of Metal Tips onto Semiconductor Quantum Rods and Tetrapods. Science 2004, 304, 1787−1790. (29) Habas, S. E.; Yang, P.; Mokari, T. Selective Growth of Metal and Binary Metal Tips on CdS Nanorods. J. Am. Chem. Soc. 2008, 130, 3294−3295. (30) Weaver, J. H. Optical Properties of Rh, Pd, Ir, and Pt. Phys. Rev. B 1975, 11, 1416−1425. (31) Schmitt-Rink, S.; Chemla, D. S.; Miller, D. A. B. Linear and Nonlinear Optical Properties of Semiconductor Quantum Wells. Adv. Phys. 1989, 38, 89−188.

(32) Klimov, V. I. Optical Nonlinearities and Ultrafast Carrier Dynamics in Semiconductor Nanocrystals. J. Phys. Chem. B 2000, 104, 6112−6123. (33) Yang, Y.; Wu, K.; Shabaev, A.; Efros, A. L.; Lian, T.; Beard, M. C. Direct Observation of Photoexcited Hole Localization in CdSe Nanorods. ACS Energy Letters 2016, 1, 76−81. (34) Utterback, J. K.; Grennell, A. N.; Wilker, M. B.; Pearce, O. M.; Eaves, J. D.; Dukovic, G. Observation of Trapped-hole Diffusion on the Surfaces of CdS nanorods. Nat. Chem. 2016, 8, 1061−1066. (35) Klimov, V. I.; Mikhailovsky, A. A.; McBranch, D. W.; Leatherdale, C. A.; Bawendi, M. G. Quantization of Multiparticle Auger Rates in Semiconductor Quantum Dots. Science 2000, 287, 1011. (36) Ben-Shahar, Y.; Scotognella, F.; Kriegel, I.; Moretti, L.; Cerullo, G.; Rabani, E.; Banin, U. Optimal Metal Domain Size for Photocatalysis with Hybrid Semiconductor-Metal Nanorods. Nat. Commun. 2016, 7, 10413. (37) Schweinberger, F. F.; Berr, M. J.; Döblinger, M.; Wolff, C.; Sanwald, K. E.; Crampton, A. S.; Ridge, C. J.; Jäckel, F.; Feldmann, J.; Tschurl, M.; et al. Cluster Size Effects in the Photocatalytic Hydrogen Evolution Reaction. J. Am. Chem. Soc. 2013, 135, 13262−13265. (38) Nakibli, Y.; Kalisman, P.; Amirav, L. Less Is More: The Case of Metal Cocatalysts. J. Phys. Chem. Lett. 2015, 6, 2265−2268. (39) Chen, S.; Wang, L.-W. Thermodynamic Oxidation and Reduction Potentials of Photocatalytic Semiconductors in Aqueous Solution. Chem. Mater. 2012, 24, 3659−3666. (40) Karp, G. Cell and Molecular Biology: Concepts and Experiments, 6th ed.; Wiley, 2009. (41) CRC Handbook of Chemistry and Physics, 91st ed.; Haynes, W. M., Ed.; CRC Press, 2010. (42) Kalisman, P.; Nakibli, Y.; Amirav, L. Perfect Photon-toHydrogen Conversion Efficiency. Nano Lett. 2016, 16, 1776−1781.

1142

DOI: 10.1021/acsenergylett.6b00468 ACS Energy Lett. 2016, 1, 1137−1142