Solution-Processed, Antimony-Doped Tin Oxide Colloid Films Enable

Mar 28, 2013 - (Eg) 3.0 eV for rutile,7,8 3.2 eV for anatase9) has been widely investigated for ... shorter (1−100 μs)9,10 than the transport time ...
7 downloads 0 Views 3MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Solution-Processed, Antimony-Doped Tin Oxide Colloid Films Enable High-Performance TiO2 Photoanodes for Water Splitting Qing Peng,*,† Berç Kalanyan,‡ Paul G. Hoertz,§ Andrew Miller,§ Do Han Kim,‡ Kenneth Hanson,∥ Leila Alibabaei,∥ Jie Liu,⊥ Thomas J. Meyer,∥ Gregory N. Parsons,‡,§ and Jeffrey T. Glass*,† †

Electrical and Computer Engineering Department, Duke University, Durham, North Carolina 27708, United States Department of Chemical and Biomolecular Engineering, North Carolina State University, 911 Partners Way, Raleigh, North Carolina 27695, United States § RTI International, Research Triangle Park, North Carolina 27709, United States ∥ Chemistry Department, UNC-Chapel Hill, North Carolina 27599, United States ⊥ Chemistry Department, Duke University, Durham, North Carolina 27708, United States ‡

S Supporting Information *

ABSTRACT: Photoelectrochemical (PEC) water splitting and solar fuels hold great promise for harvesting solar energy. TiO2-based photoelectrodes for water splitting have been intensively investigated since 1972. However, solar-to-fuel conversion efficiencies of TiO2 photoelectrodes are still far lower than theoretical values. This is partially due to the dilemma of a short minority carrier diffusion length, and long optical penetration depth, as well as inefficient electron collection. We report here the synthesis of TiO2 PEC electrodes by coating solution-processed antimony-doped tin oxide nanoparticle films (nanoATO) on FTO glass with TiO2 through atomic layer deposition. The conductive, porous nanoATO film-supported TiO2 electrodes, yielded a highest photocurrent density of 0.58 mA/cm2 under AM 1.5G simulated sunlight of 100 mW/cm2. This is approximately 3× the maximum photocurrent density of planar TiO2 PEC electrodes on FTO glass. The enhancement is ascribed to the conductive interconnected porous nanoATO film, which decouples the dimensions for light absorption and charge carrier diffusion while maintaining efficient electron collection. Transient photocurrent measurements showed that nanoATO films reduce charge recombination by accelerating transport of photoelectrons through the less defined conductive porous nanoATO network. Owing to the large band gap, scalable solution processed porous nanoATO films are promising as a framework to replace other conductive scaffolds for PEC electrodes. KEYWORDS: Photoelectrochemical, water splitting, antimony-doped tin oxide, TiO2, core−shell

S

can potentially decrease charge recombination in TiO2. One strategy is to load nanocrystalline TiO2 onto conductive scaffolds. High-surface-area, porous, and conductive scaffolds could also solve problems arising from short diffusion lengths of the minority carrier, while maintaining sufficient light absorption.18−21 An ideal nanostructured conductive scaffold should (1) consist of earth abundant materials and be fabricated by scalable low-cost methods, (2) allow a significant fraction of the incoming photons in the desired wavelength range to be absorbed by photoactive materials, (3) have the correct energetics for electrons flowing from TiO2 to the conductive scaffold, and (4) have sufficient porosity for electrolyte diffusion to the majority of the nanostructured scaffold after loading photoactive materials.

table photoelectrochemical (PEC) water-splitting devices with high solar-to-fuel efficiencies have not yet been realized but hold great promise for harvesting solar energy and storing the energy in chemical bonds.1−6 Owing to its low production cost, environmental compatibility, and remarkable chemical and photoelectrochemical stability, TiO2 (band gap (Eg) 3.0 eV for rutile,7,8 3.2 eV for anatase9) has been widely investigated for PEC water oxidation electrodes since 1972.7−11 However, the solar-to-fuel conversion efficiency of polycrystalline TiO2 PEC electrodes is still limited by dilemmas posed by short minority carrier diffusion lengths (L = 10−100 nm),11−13 large optical penetration depths,14 and inefficient electron collection.15 TiO2 nanoparticle films decouple light absorption length and minority carrier diffusion lengths, but photocurrent densities of these films are still 12).10 Fast electron transport © 2013 American Chemical Society

Received: December 9, 2012 Revised: March 11, 2013 Published: March 28, 2013 1481

dx.doi.org/10.1021/nl3045525 | Nano Lett. 2013, 13, 1481−1488

Nano Letters

Letter

On the basis of the hole diffusion length (L)11−13 and optical absorption coefficient of polycrystalline TiO2 (α, ∼1.2 μm−1 for photon at 370 nm),14 the conductive scaffold should have a large roughness factor R (i.e., total surface area/substrate surface normal projected area) to load enough TiO2 to absorb the majority of the incident photons in the targeted wavelength range. Without considering optical properties of the supporting scaffold, the calculated roughness factor to achieve 99% light absorption and sufficient minority carrier collection is R ≈ 2/ (L·α), which is 20−200 for a nanostructured TiO2 PEC. Several high-surface-area conductive scaffolds for TiO2 PEC electrodes have been reported including single crystal Si nanowires,20,22,23 TiSi2 nanonets,21 carbon fibers18 and carbon tubes.19 All of them showed enhancement of photocurrent densities resulting from the conductive scaffold. According to the above-mentioned requirements, transparent conducting oxides are better as a supporting scaffold for PEC application, however, limited research has been reported.24 In this paper, we present the application of an antimony doped tin oxide nanoparticle (nanoATO) film on fluorine doped tin oxide glass (FTO-nanoATO) as the transparent conductive scaffold for nanostructured TiO2 PEC electrodes. The nanoATO films were fabricated by using a scalable colloidal dispersion processing technique, which can be easily implemented on large-scale. Moreover, ATO is a transparent conductor (Eg, 3.6−4.4 eV).25,26 It competes much less with TiO2 for light absorption than previous scaffold materials, for example, Si (Eg ≈ 1.1 eV),20,22,23 TiSi2 (Eg ≈ 1.5 eV),21 and carbon.27 We demonstrate that photoelectrodes of FTO-nanoATO with TiO2 shell (FTO-nanoATO/TiO2) synthesized by atomic layer deposition (ALD) produce significant improvements in photocurrent density relative to planar electrodes of TiO2 ALD film-coated FTO glass (FTO/TiO2). Although the porous nanoATO films have a less defined conduction pathways and presumably larger defect densities than forests of single crystal Si nanowires20,23 and TiSi2 nanonets,21 photocurrent densities are comparable.20,21,23 The thickness of the ALD TiO2 photoabsorbing film was found to be an important factor in determining PEC performance. This is primarily due to the reduction of porosity in the FTO-nanoATO/TiO2 films. Results. The preparation of FTO-nanoATO/TiO2 electrodes is illustrated in Figure 1. The porous FTO-nanoATO film, as shown in Figure 1a, was prepared by spin coating colloidal ATO nanoparticles (average diameter of 22−44 nm) on FTO glass followed by sintering at 500 °C in air. The resulting FTOnanoATO film, ∼2 μm in thickness, is a bicontinuous porous network of pores and nanoparticles (Figure 2a) and similar to previously reported porous films of nanoITO28 and nanoTiO2.1 A conformal layer of TiO2 was coated onto the FTO-nanoATO substrate by ALD (Figure 1b). ALD is a stepwise, substrate sitelimited growth method that can deposit conformal coatings with tunable thicknesses (Supporting Information Figure S1), even onto complex nanostructures (Figure 1c).29 Figure 2a shows the morphology of pure nanoATO film on FTO glass obtained by SEM. The SEM measurement shows that the majority of particles have diameters 10% of the initial signal extending out to more than 100 μs. The transient absorption signal from FTO-nanoATO was too small for meaningful interpretation owing to the large band gap of ATO. Discussion. Photocurrent densities for electrodes of FTOnanoATO/TiO2(200) are much higher than for FTO-nanoTiO2 film electrodes shown Supporting Information Figure S3 and other reports.9,10,16,32 Photocurrent density depends mainly on efficiencies of three processes: photon absorption, charge separation, and interfacial charge transfer.33 With the same electrolyte and electrochemical conditions, the interfacial charge transfer for water oxidation at the TiO2/electrolyte interface is favorable with larger surface area of the electro-

wavelengths are presented in Figure 3c. These measurements show that more than 95% of the photocurrent results from irradiation at 400 nm is consistent with the rutile phase of the TiO2 coating in FTO-nanoATO/TiO2 electrodes (Supporting Information Figure S5). The photocurrent produced from irradiation at >550 nm is essentially zero. The results from Figure 3c corroborate that the photocurrent from FTOnanoATO/TiO2(200) electrode was mainly produced from UV excited TiO2. The photochemical stability of the FTO-nanoATO/ TiO2(200) film was measured by monitoring the photocurrent density under illumination over the course of more than 4 days and the results are presented in Figure 3d. There was a nominal decrease in the current response from FTO-nanoATO/TiO2 photoelectrode over the course of the measurement suggesting that the electrode is stable under operating conditions in 1 M KOH aqueous solution. In comparison, the ATO nanoparticles in the untreated FTO-nanoATO film desorbed from the glass substrate within 5 h in the testing electrolyte. On the basis of these comparisons, the stable photocurrent densities in Figure 3d also show that the thin TiO2 ALD layer is conformal and pinhole free and protects the underlying nanoATO film. Otherwise fluctuation of current density would have been observed due to dissolution of ATO nanoparticles and delamination of the TiO2 shells over the more than 4 day testing period. The normalized transient photocurrent traces (at 0 V versus Ag/AgCl) for FTO-nanoATO, FTO-nanoATO/TiO2(200), and FTO-nanoTiO2 film electrodes under 355 nm pulsed 1484

dx.doi.org/10.1021/nl3045525 | Nano Lett. 2013, 13, 1481−1488

Nano Letters

Letter

possible by the conductive nature of the underlying nanoATO network in the core−shell structure owing to the following reasons. First of all, bandgap excitation occurs predominantly within the thin TiO2 shell, which significantly attenuates the likelihood of “bulk” recombination. For TiO2 nanoparticles, “bulk” recombination prior to or following surface trapping of electrons and holes reduces the overall quantum yield of water oxidation. With the core−shell system, electron−hole recombination prior to surface-trapping of holes only occurs within the thickness of the TiO2 shell. As a consequence, the core−shell structure promotes the transfer of electrons from the shell to the ATO core. Furthermore, the core−shell structure significantly reduces the electron transfer distance in the thin polycrystalline TiO2 film (Supporting Information Figure S5) to the underlying ATO nanoparticle collector electrode. Moreover, the shell serves as a physical barrier and controls the distance between electrons in ATO and holes trapped in TiO2 surface states thereby controlling the kinetics of recombination. The higher electron mobility of the ATO and the core−shell structure favors transport and collection of electrons further impeding recombination with surface-trapped holes. In addition, according to the junction model the interfacial properties between TiO2 and ATO are favorable for preventing hole and electron recombination. The Fermi level of antimony doped tin oxide is located in the conduction band of tin oxide37 and its band gap is ∼4.0 eV.37 TiO2 films have Fermi levels at 0.2 to 0.3 eV below the conduction band.20 In nanoscale materials, band bending is small,38 therefore carrier transport occurs from TiO 2 to nanoATO through a tunneling mechanism.39,40 The small offset between conduction bands of TiO2 and ATO1 make the tunneling of electrons much easier,39 while the relatively large offset in their valence band edges1 dramatically decreases the tunneling probability of photogenerated holes from TiO2 to ATO. As in a typical dye-sensitized solar cell, the photocurrent density of FTO-nanoATO/TiO2 is greatly enhanced by the large surface area of the porous, nanostructured electrodes. As shown in Figure 3b, the planar FTO/TiO2 electrode has a maximal photocurrent density at a thickness of ∼50 nm corresponding to 1000 ALD cycles. Given the optical absorption coefficient α of ∼1.2 μm−1 for photons at 370 nm, a TiO2 layer ∼2 μm thick is needed to absorb >90% of 370 nm photons. Therefore, the reduced photocurrent density with increased TiO2 coating thickness is due to the increased degree of charge recombination resulting from longer average electron and hole transfer distance to the collector electrode and semiconductor-electrolyte interface, respectively. In this configuration, the fast charge recombination offsets the increased population of photoexcited carriers within the thick TiO2 layer (>50 nm). The optimized thickness agrees well with the literature-reported hole diffusion length (L = 10−100 nm) in polycrystalline TiO2.11−13 In contrast to the planar FTO/TiO2 electrodes, the large surface area of FTO-nanoATO film supports (roughness factors >120) dramatically increases the loading of TiO2 at a given TiO2 ALD cycles. As a result, the FTO-nanoATO/TiO2 electrodes absorb much more light than the planar FTO/TiO2 PEC electrodes having the similar thickness of TiO2 ALD layer. Simultaneously, the thin coating of TiO2, for example, ∼9 nm from 200 ALD cycles, on the interconnected porous structure of the conductive FTO-nanoATO films, provide an

Figure 4. (a) Normalized current−time traces following laser flash excitation and (b) normalized absorption−time traces of pure FTOnanoATO (blue), FTO-nanoATO/TiO2(200) (red), FTO-nanoTiO2 and background (black). Both experiments were performed in aqueous 1.0 M KOH with 0 V applied bias versus Ag/AgCl (platinum counter electrode; λex = 355 nm, 8.0 mJ/pulse).

des.20,32 FTO-nanoATO/TiO2(200) electrode has less interfacial area on 1 cm2 irradiation area than FTO-nanoTiO2 film (roughness factor of around 1000).34 The nanocrystalline rutile phase of TiO2 in FTO-nanoATO/ TiO2 electrodes (Supporting Information Figure S5) enables additional photon absorption (3.0−3.2 eV) relative to nanoTiO2 (anatase) particle film.32 However, Figure 3b shows that charge recombination loss in 50 nm thick nanocrystalline TiO2 ALD coating limit its PEC performance. This is consistent with early studies, which identified that high charge recombination loss from low electron mobility is the main reason for the low PEC efficiency of nanocrystalline TiO2 materials owing to the trap states.15,32,35,36 Therefore, the difference in photon adsorption is not the main reason for the dramatic enhancement of photocurrent density in FTO-nanoATO/TiO2(200). It has been previously concluded that only the photoexcited surface trapped holes with long lifetimes (hundredths of a second) participate in water oxidation.9,10 due to slow water oxidation kinetics in competition with rapid charge recombination between surface-trapped holes and electrons.10 The normalized transient photocurrent measurements in Figure 4a show that FTO-nanoATO/TiO2(200) has a significantly longer-lived transient photocurrent than the FTO-nanoTiO2. The longer-lived photocurrent is indicative of significantly slowed charge recombination in FTO-nanoATO/TiO2(200). This is a direct consequence of the enhanced electron transport kinetics and electron collection efficiency, which is made 1485

dx.doi.org/10.1021/nl3045525 | Nano Lett. 2013, 13, 1481−1488

Nano Letters

Letter

densities result from a relatively larger population of longerlived charge carriers, which are enabled by the core−shell structure that favors transport of electrons from semiconductor shell to conductive core while inhibiting electron−hole charge recombination. The scaffold enables efficient charge collection and addresses the dilemma of large optical penetration depth and short diffusion length of minority carriers. The photocurrent densities found for FTO-nanoATO/TiO2 depend strongly on TiO2 coating thickness apparently due, at least in part, to a pore-sealing effect of the TiO2 coating. Further improvements in efficiency should be possible by optimizing the porous structure of the nanoATO film. Owing to its transparency over a wide range of wavelengths, the scalable, solution-processed, conductive nanoATO scaffold has great potential to boost solar-to-fuel conversion efficiencies of small bandgap PEC materials, for example, Fe2O3, whose efficiency has been largely limited by the mismatch of hole diffusion length and optical penetration depth, the combination of which leads to inefficient electron collection.9,42 Experimental Section. Antimony-doped tin oxide nanoparticles of average particle size of 22−44 nm, (NanoTek, purity of 99.5%) was purchased from Alfa Aesar and used without further purification. The weight ratio of Sb2O5/SnO2 is 10:90 and the specific surface area of the particles are 20−40 m2/g. Dispersions of the nanoparticles were prepared by using a previous literature method28 and used to prepare the porous nanoATO films (∼2 μm) on FTO glass. The FTO-nanoATO samples were then annealed in air at 500 °C for 1 h with a ramp rate of 5 °C/min to form a good contact among the particles, as well as between the nanoATO film and the FTO substrate. The resistivity of the nanoATO films is around 40 Ω·cm as analyzed by a four-point probe (Jandel Ltd.) on quartz substrates. TiO2 ALD on the substrates was carried out in a custom hotwall tube reactor as described previously.43 In brief, TiO2 ALD was performed at 300 °C with N2 flow rate of 200 sccm. The process pressure was ∼1 Torr. The timing sequence for dose and purge of TiCl4 and H2O was 1.5/10 s and 1.5/10 s, respectively. Before commencing ALD deposition, samples were placed in the reactor for 30 min under N2 flow (200 sccm) to establish thermal equilibrium. The growth rate of TiO2 ALD was monitored by the witness Si samples (Supporting Information Figure S1). The TiCl4 (99%) was purchased from Gelest Inc. and used as received. Deionized water from an onsite water purification system was used as the water source. Ultrahigh purity N2 (99.999%) from National Welders was used as both carrier and purge gas and further purified by a filter (Gatekeeper). Morphologies of nanoATO films before and after ALD TiO2 were analyzed by scanning electron microscope (SEM) on FEI XL30 with electron beam energy of 10 kV. Roughness factor for the nanoATO films before and after ALD TiO2 coating was analyzed by dye (N719 dye, ditetrabutylammonium cisbis(isothiocyanato)- bis(2,2′-bipyridyl-4,4′-dicarboxylato)ruthenium(II)) adsorption−desorption technique according to the literature procedure.21 The roughness factor of the samples was then calculated by assuming each dye molecule has a cross-section area of 1.65 nm2. Although N719 dye molecule is much larger than −OH group, the similar surface chemistry and −OH ligand density on the tested ALD coatings make it a reasonable method to evaluate the change of surface areas of TiO2 coated FTO-nanoATO substrates. To assemble the substrates into PEC electrodes, ohmic electric contacts to TiO2 coated FTO-nanoATO and FTO glass

average diffusion length for photoproduced holes (photoholes) similar to the TiO2 coating thickness. This geometric feature greatly facilitates efficient diffusion of photoholes to the interface between TiO2 and the external electrolyte. It will be discussed in detail in the following section. In summary, the FTO-nanoATO underlying porous films greatly enhance photocurrent densities in FTO-nanoATO/TiO2 photoelectrodes by decoupling the dimensions for optical absorption and hole diffusion. The effect of TiO2 coating thickness on photocurrent density of FTO-nano/TiO2 under PEC conditions can be understood in the following way. When the TiO2 ALD coating is relatively thin, the interconnected pores throughout the nanoATO film is maintained so that electrolyte is free to diffusion through the porous network as shown in Figure 1b and Figure 2b. In this case, the average transport length of photoholes to the TiO2/ electrolyte interface is comparable to the TiO2 coating thickness because the photoholes react with the electrolyte within the nearby pores. This allows the photocurrent density of the FTO-nanoATO/TiO2 PEC electrodes to increase as TiO2 shell thickness increases up to ∼9 nm due to (i) increased light absorption and (ii) a better crystallinity of the film compared to lower thicknesses.41 As the TiO2 thickness increases beyond 9 nm, for example, ∼12 nm (Supporting Information Figure S2) and ∼30 nm (Figure 2c), the TiO2 coating seals more and more pores, thereby reducing interfacial area and diffusion of electrolyte into the film as illustrated in Figure 1c. Under these conditions, a significant fraction of photoholes must travel to the top surface of the FTOnanoATO/TiO2 film to access the electrolyte for water oxidation. This increases the average transport length of photoholes from the TiO2 coating from the nm to μm scale. This is much larger than the effective hole diffusion length in the ALD TiO2 coating (∼50 nm) as shown in Figure 3b. Thus, the probability of charge recombination in the FTO-nanoATO/TiO2 PEC electrodes increases, which reduces the photocurrent density. This mechanism is supported by the fact that the photocurrent density decreased further with increasing TiO2 shell thickness (i.e., ∼30 nm, 600 TiO2 cycles), as increasing numbers of pores sealed or blocked from electrolyte by the thick coating. These results also suggest that the bicontinuous porous structure of FTO-nanoATO film is important for optimizing PEC performance. By designing the pore size and porous structures of nanoATO films, further improvement of the photocurrent density is possible. Since the maximum photocurrent density of planar FTO/TiO2 PEC was found for films with ∼50 nm TiO2, presumably a porous ATO structure with similar thickness would result in higher photocurrent densities obtainable in nanoATO particle films with bigger pore sizes, or in macro-porous ATO structures or open ATO nanowire structures. Comprehensive optimization will be necessary for better performance, including, for example, balancing light scattering from the structure and the thickness of the TiO2 coating, and improvement of interfacial charge transfer efficiency. Conclusion. FTO-nanoATO/TiO2 electrodes, ∼9 nm thick TiO2 shell deposited by ALD, reach photocurrent densities of 0.58 mA/cm2 (at 0 V versus Ag/AgCl) that is comparable to other conductive scaffold-supported TiO2 PEC electrodes.20,21 This photocurrent density is much larger than the maximum photocurrent densities from planar FTO/TiO2 PECs, ∼0.2 mA/cm2 for ∼50 nm TiO2. The enhanced photocurrent 1486

dx.doi.org/10.1021/nl3045525 | Nano Lett. 2013, 13, 1481−1488

Nano Letters

Letter

substrates and FTO-nanoTiO2 film were made by rubbing Ga− In eutectic on part of the FTO surface of the substrates. Electrical connection was made to the sample by contacting this eutectic with a copper wire, which was then sealed with nonconductive epoxy (Hysol 9460, Loctitte) at the end of a glass tubing through which the Cu wire had been directed such that the surface-normal of the substrate was perpendicular to the glass tubing. Epoxy (Hysol 9460, Loctitte) was then used to define the active area of electrodes (∼1 cm2), and cured under a heating gun at ∼80 °C in ambient environment. The active area of the photoelectrodes was measured with ImageJ software by using the photo images of the electrodes. All electrochemical measurements were performed in a custom fabricated PTFE cell having quartz windows on both sides using a three-electrode configuration and all electrodes were inside the electrochemical cell. A potentiostat/galvanostat system (model 1287A by Solartron Analytical) was used for cyclic voltammetry (CV) measurement, in which the potential was swept from −1.2 to 1.0 V versus Ag/AgCl reference electrode at a scan rate of 20 mV/s. The reference electrode was Ag/AgCl in 3.5 M KCl solution (Hanna instruments Inc.). Pt gauze (100 mesh, 99.95% purity Alfa Aesar Co.) served as the counter electrode. All electrodes were tested in 1 M KOH solution (pH = 13.6), which was continuously bubbled with ultrahigh purity Ar (National Welders) to remove oxygen and H2 from the solution. Prior to measuring the photocurrent, CV scans between −1.2 and 1.0 V versus Ag/AgCl with a Pt wire electrode were performed 10 times to eliminate impurities. For photoelectrochemical analysis, the light source used was a Newport Oriel Xe lamp with AM1.5G filter. The light intensity was calibrated so that the photoelectrode produce an equivalent photocurrent density to that obtained under 100 mW/cm2 of AM1.5G illuminations. Optical filters of CGA320, CGA400, and CGA550 (Newport Inc.) were used to filter the light through with the cut-on wavelength of 320, 400, and 500 nm respectively. The tolerance of cut-on wavelength is ±5 nm. The optical transmittance of the filters is ≥90%. In the long term stability testing, the potential of the testing electrode stay at −1.0 and 0.5 V versus Ag/AgCl for 100 s alternatively for more than 4 days under irradiation. Transient photocurrent measurements were acquired by using a CH Instruments model 600D Series Electrochemical Workstation. A custom two-compartment photoelectrochemical cell with a glass frit spacer was employed for both photocurrent and transient absorption measurements. The working electrode, FTO-nanoATO, FTO-nanoATO/ TiO2(200), and FTO-nanoTiO2 film were placed at a 45° angle into the first compartment, a 10 mm path length square cuvette, with a platinum wire counter electrode. The Ag/AgCl reference electrode (BASi, MF-2079) was placed in the second compartment. Both compartments were filled with aqueous 1.0 M KOH and the entire system was kept under a N2 environment. Current−time traces were acquired for 10 s (0.1 ms per data point) with 0 V applied bias versus Ag/AgCl. Laser excitation (λex = 355 nm, 8.0 mJ/pulse) at a repetition rate of 1 Hz was incident perpendicular to the cuvette and at a 45° angle to the working electrode. Transient absorption (TA) measurements were performed by using nanosecond laser pulses produced by a SpectraPhysics Quanta-Ray Lab-170 Nd:YAG laser (355 nm, 5−7 ns, operated at 1 Hz, beam diameter 0.5 cm, ∼8 mJ/pulse) integrated into a commercially available Edinburgh LP920 laser flash photolysis spectrometer system. White light probe pulses

generated by a pulsed 450 W Xe lamp were passed through the sample, perpendicular to the laser beam, focused into the spectrometer (2 nm bandwidth), then detected by a photomultiplier tube (Hamamatsu R928). A 395 nm long pass filter was placed between the sample and the white light probe to prevent direct band gap excitation by the probe light. A 395 nm long pass filter was also placed before the detector to reject unwanted scattered light. Detector outputs were processed using a Tektronix TD.032C Digital Phosphor Oscilloscope interfaced to a PC running Edinburgh’s L900 (version 7.0) software package. Single wavelength kinetic data (averaging 200 laser shots) were acquired using the electrochemical cell described above with 0 V applied bias versus Ag/AgCl.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information and figures. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: (Q.P.) [email protected]; (J.T.G.) jfglass@duke. edu. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors would like to acknowledge RTI International and Research Triangle Solar Fuel Institute (RTSFI) for financial support. The authors would like to thank Jim Trainham for his support on experiments and helpful advice. Q.P. thanks Kyoungmi Lee for sharing the ALD system and Mark Losego and Yingzhen Lu for valuable discussions. Q.P. thanks Shane Di Dona for the figure of content. This work was also funded in part by the UNC Energy Frontier Research Center (EFRC) “Center for Solar Fuels”, an EFRC funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under award DESC0001011, which supported K.H. (T.J.M). Support for L.A., P.G.H., A.M. and Q.P. was provided by RTI International through the Research Triangle Solar Fuels Institute. We acknowledge support for the purchase of instrumentation from the UNC EFRC (Center for Solar Fuels), funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under award number DESC0001011) and UNC SERC (“Solar Energy Research Center Instrumentation Facility” funded by the U.S. Department of Energy Office of Energy Efficiency & Renewable Energy under award number DE-EE0003188).

■

REFERENCES

(1) Gratzel, M. Nature 2001, 414, 338. (2) Bard, A. J.; Fox, M. A. Acc. Chem. Res. 1995, 28, 141. (3) Wrighton, M. S. Acc. Chem. Res. 1979, 12, 303. (4) Meyer, T. J. Acc. Chem. Res. 1989, 22, 163. (5) Lewis, N. S.; Walter, M. G.; Warren, E. L.; McKone, J. R.; Boettcher, S. W.; Mi, Q. X.; Santori, E. A. Chem. Rev. 2010, 110, 6446. (6) Chen, X. B.; Shen, S. H.; Guo, L. J.; Mao, S. S. Chem. Rev. 2010, 110, 6503. (7) Fujishima, A.; Honda, K. Nature 1972, 238, 37. (8) Hwang, Y. J.; Hahn, C.; Liu, B.; Yang, P. D. ACS Nano 2012, 6, 5060. 1487

dx.doi.org/10.1021/nl3045525 | Nano Lett. 2013, 13, 1481−1488

Nano Letters

Letter

(9) Cowan, A. J.; Barnett, C. J.; Pendlebury, S. R.; Barroso, M.; Sivula, K.; Gratzel, M.; Durrant, J. R.; Klug, D. R. J. Am. Chem. Soc. 2011, 133, 10134. (10) Cowan, A. J.; Tang, J. W.; Leng, W. H.; Durrant, J. R.; Klug, D. R. J. Phys. Chem. C 2010, 114, 4208. (11) Hoang, S.; Guo, S. W.; Hahn, N. T.; Bard, A. J.; Mullins, C. B. Nano Lett. 2012, 12, 26. (12) Salvador, P. J. Appl. Phys. 1984, 55, 2977. (13) Linsebigler, A. L.; Lu, G. Q.; Yates, J. T. Chem. Rev. 1995, 95, 735. (14) Aarik, J.; Aidla, A.; Kiisler, A. A.; Uustare, T.; Sammelselg, V. Thin Solid Films 1997, 305, 270. (15) Hendry, E.; Koeberg, M.; O’Regan, B.; Bonn, M. Nano Lett. 2006, 6, 755. (16) Park, J. H.; Kim, S.; Bard, A. J. Nano Lett. 2006, 6, 24. (17) Peter, L. M. J. Phys. Chem. C 2007, 111, 6601. (18) Pint, C. L.; Takei, K.; Kapadia, R.; Zheng, M.; Ford, A. C.; Zhang, J. J.; Jamshidi, A.; Bardhan, R.; Urban, J. J.; Wu, M.; Ager, J. W.; Oye, M. M.; Javey, A. Adv. Energy Mater. 2011, 1, 1040. (19) Kongkanand, A.; Dominguez, R. M.; Kamat, P. V. Nano Lett. 2007, 7, 676. (20) Hwang, Y. J.; Boukai, A.; Yang, P. D. Nano Lett. 2009, 9, 410. (21) Lin, Y. J.; Zhou, S.; Liu, X. H.; Sheehan, S.; Wang, D. W. J. Am. Chem. Soc. 2009, 131, 2772. (22) Shi, J.; Wang, X. D. Energy Environ. Sci. 2012, 5, 7918. (23) Shi, J.; Hara, Y.; Sun, C. L.; Anderson, M. A.; Wang, X. D. Nano Lett. 2011, 11, 3413. (24) Stefik, M.; Cornuz, M.; Mathews, N.; Hisatomi, T.; Mhaisalkar, S.; Gratzel, M. Nano Lett. 2012, 12, 5431. (25) Terrier, C.; Chatelon, J. P.; Roger, J. A. Thin Solid Films 1997, 295, 95. (26) Shanthi, E.; Dutta, V.; Banerjee, A.; Chopra, K. L. J. Appl. Phys. 1980, 51, 6243. (27) Wu, Z. C.; Chen, Z. H.; Du, X.; Logan, J. M.; Sippel, J.; Nikolou, M.; Kamaras, K.; Reynolds, J. R.; Tanner, D. B.; Hebard, A. F.; Rinzler, A. G. Science 2004, 305, 1273. (28) Hoertz, P. G.; Chen, Z. F.; Kent, C. A.; Meyer, T. J. Inorg. Chem. 2010, 49, 8179. (29) George, S. M. Chem Rev 2010, 110, 111. (30) Kay, A.; Gratzel, M. Chem. Mater. 2002, 14, 2930. (31) Hou, K.; Puzzo, D.; Helander, M. G.; Lo, S. S.; Bonifacio, L. D.; Wang, W. D.; Lu, Z. H.; Scholes, G. D.; Ozin, G. A. Adv. Mater. 2009, 21, 2492. (32) Cho, I. S.; Chen, Z. B.; Forman, A. J.; Kim, D. R.; Rao, P. M.; Jaramillo, T. F.; Zheng, X. L. Nano Lett. 2011, 11, 4978. (33) Chen, Z. B.; Jaramillo, T. F.; Deutsch, T. G.; KleimanShwarsctein, A.; Forman, A. J.; Gaillard, N.; Garland, R.; Takanabe, K.; Heske, C.; Sunkara, M.; McFarland, E. W.; Domen, K.; Miller, E. L.; Turner, J. A.; Dinh, H. N. J. Mater. Res. 2010, 25, 3. (34) Kim, D. H.; Koo, H. J.; Jur, J. S.; Woodroof, M.; Kalanyan, B.; Lee, K.; Devine, C. K.; Parsons, G. N. Nanoscale 2012, 4, 4731. (35) Richter, C.; Schmuttenmaer, C. A. Nat. Nanotechnol. 2010, 5, 769. (36) Yagi, E.; Hasiguti, R. R.; Aono, M. Phys. Rev. B 1996, 54, 7945. (37) Shanthi, E.; Dutta, V.; Banerjee, A.; Chopra, K. L. J. Appl. Phys. 1980, 51, 6243. (38) Hagfeldt, A.; Gratzel, M. Chem. Rev. 1995, 95, 49. (39) Ruhle, S.; Cahen, D. J. Phys. Chem. B 2004, 108, 17946. (40) Lee, S.; Noh, J. H.; Han, H. S.; Yim, D. K.; Kim, D. H.; Lee, J. K.; Kim, J. Y.; Jung, H. S.; Hong, K. S. J. Phys. Chem. C 2009, 113, 6878. (41) Law, M.; Greene, L. E.; Radenovic, A.; Kuykendall, T.; Liphardt, J.; Yang, P. D. J. Phys. Chem. B 2006, 110, 22652. (42) Sivula, K.; Le Formal, F.; Gratzel, M. ChemSusChem 2011, 4, 432. (43) Na, J. S.; Peng, Q.; Scarel, G.; Parsons, G. N. Chem. Mater. 2009, 21, 5585.

1488

dx.doi.org/10.1021/nl3045525 | Nano Lett. 2013, 13, 1481−1488