Au-TiO 2 Nanocomposites and Efficient Photocatalytic Hydrogen Production under UV-Visible and Visible Light Illuminations : A Comparison of Different Crystalline Forms of TiO 2

Au(∼1wt%)/TiO2(anatase or rutile or P25) nanocomposites were prepared by the solvated metal atom dispersion (SMAD) method, and the as-prepared samples were characterized by diffuse reflectance UV-visible spectroscopy, powder XRD, BET surface analysis measurements, and transmission electron microscopy bright field imaging. The particle size of the embedded Au nanoparticles ranged from 1 to 10 nm. These Au/TiO2 nanocomposites were used for photocatalytic hydrogen production in the presence of a sacrificial electron donor like ethanol or methanol under UV-visible and visible light illumination. These nanocomposites showed very good photocatalytic activity toward hydrogen production under UV-visible conditions, whereas under visible light illumination, there was considerably less hydrogen produced. Au/P25 gave a hydrogen evolution rate of 1600 μmol/h in the presence of ethanol (5 volume %) under UV-visible illumination. In the case of Au/TiO2 prepared by the SMADmethod, the presence of Au nanoparticles serves two purposes: as an electron sink gathering electrons from the conduction band (CB) of TiO2 and as a reactive site for water/ethanol reduction to generate hydrogen gas. We also observed hydrogen production by water splitting in the absence of a sacrificial electron donor using Au/TiO2 nanocomposites under UV-visible illumination.


Introduction
Ever since the first report of water splitting using a TiO 2 photoanode by Fujishima and Honda, researchers all over the world have been trying to develop efficient solar energy harvesting semiconductor materials for water splitting and thereby producing clean hydrogen energy [1].Several strategies like band gap engineering of semiconductor materials, cocatalyst loading, two-step photoexcitation (Z scheme), and so forth have been developed to achieve efficient overall water splitting to produce hydrogen and oxygen [2][3][4][5].Recently, people have been trying to exploit the UV and visible absorption of metal nanoparticles like Au, Ag, and Cu by embedding them on a semiconductor material [6,7].Among the various semiconductor catalysts, TiO 2 is one of the most promising due to its high chemical stability, nontoxicity, and low cost [8].The only limitation of TiO 2 as a photocatalyst is its poor efficiency in the visible region of the solar spectrum due to its wide band gap (3.2 eV for anatase).However, TiO 2 can show some visible photocatalytic activity due to the presence of some rutile form of TiO 2 and/or by doping it with metal or nonmetallic elements [9][10][11] or by deposition of noble metal nanoparticles on TiO 2 [12][13][14][15].Metal nanoparticle loaded TiO 2 showed enhanced photocatalytic activity under UV or visible light illumination.The improved UV activity of these photocatalysts is attributed to the better charge separation by electron transfer from the conduction band (CB) of TiO 2 to Au nanoparticles [16][17][18], whereas the observed visible activity is explained as photoexcitation of Au nanoparticles (due to surface plasmon resonance effect) and charge separation International Journal of Photoenergy by the transfer of photoexcited electrons from Au to the CB of TiO 2 [16,[19][20][21][22][23][24][25].There are some other reports explaining the enhanced visible activity of Au/TiO 2 due to local electric field enhancement near the TiO 2 surface by the Au surface plasmon resonance [26][27][28][29].
There are several methods reported in the literature for the synthesis of TiO 2 supported gold nanoparticles such as adsorption of preformed Au colloids [30], photodeposition [31,32], deposition precipitation [33], impregnation [33], and chemical reduction [34].We used a different approach for the synthesis of these Au loaded TiO 2 nanocomposites: the solvated metal atom dispersion (SMAD) method.The SMAD technique has been used extensively in our laboratories for the gram scale synthesis of nanoparticles of metals like Au, Ag, Cu, metal chalcogenides, [35][36][37][38].The advantages of this method over other routes reported in the literature are that no byproducts are formed, reproducibility and scalability.We used different types of TiO 2 : anatase (UV active), rutile (visible active), and P25 (a mixture of 75% anatase and 25% rutile) for Au loading.These Au (∼1 wt%) loaded TiO 2 nanocomposites, prepared by the SMAD method were used as photocatalysts toward photocatalytic hydrogen production, in the presence or absence of a sacrificial electron donor ethanol (5 volume%) under UV-visible and visible conditions.This paper will discuss our findings regarding the role of Au nanoparticles and the role of different crystalline forms of TiO 2 .

Materials.
The different phases of TiO 2 used were anatase (Alfa Aesar, 99.9% and anatase prepared by aerogel method in our laboratories), rutile (Sigma Aldrich, 99.99%), and P25 (Degussa).Sodium sulfite (98.3%) was purchased from Fischer Scientific.Butanone (Sigma Aldrich, 99.7%) was distilled and freeze pump thaw processed for five cycles prior to use in SMAD experiments.Ethanol (200 Proof) from Decon Labs was used for photochemical reactions.

Instrumentation.
Diffuse reflectance UV-visible spectra (DRS UV) were recorded using a Cary 500 scan UV-Vis-NIR spectrophotometer operating in air at room temperature over the range from 200 to 800 nm.Brunauer-Emmet-Teller (BET) measurements of surface area and pore size distribution of TiO 2 and Au/TiO 2 nanocomposites were determined using a Quantachrome NOVA 1200 N 2 gas adsorption/desorption analyzer at liquid nitrogen temperature.Powder XRD analysis of these samples was carried out using a Scintag-XDS-2000 spectrometer with Cu K radiation with applied voltage of 40 kV and current of 40 mA.TEM bright filed images were taken using a Phillips CM100 electron microscope operating at 100 kV.Bulk elemental analysis was carried out using Perkin Elmer Optima 4300DV spectrometer (ICP-OES) at Galbraith laboratories Inc.

Preparation of
Au/TiO 2 Nanocomposites.Au loaded TiO 2 nanocomposites were prepared by the SMAD method.The details of the SMAD technique are given elsewhere [39].
In a typical experiment, the crucible was loaded with Au shot (∼60 mg) and TiO 2 (P25, anatase commercial, anatase aerogel, or rutile) (∼2960 mg) was placed in the bottom of the reactor.Liquid nitrogen cooling and vacuum were applied to the reactor and the crucible was heated resistively under vacuum in such a manner that there was cocondensation of Au atoms with solvent molecules (butanone) on the walls of the reactor.The reactor was brought up to room temperature under Ar atmosphere once the metal evaporation was complete.Upon matrix (Au-butanone) melt down, the Aubutanone colloid comes into contact with TiO 2 at the bottom of the reactor.The Au-butanone/TiO 2 mixture was stirred vigorously under Ar atmosphere for 2 h.The color of the slurry changed from dark blue to purple during the stirring process.Butanone was removed from the mixture by applying vacuum, and the dried powder was used for further studies.In addition to the vacuum drying, Au/TiO 2 samples for control experiments were calcined in air at 200 ∘ C for 2 h (heating rate, 5 ∘ C/min) to make sure that there was no butanone remaining.The catalysts prepared by the SMAD method are shown in Figure S1 (see Supplementary Material available online at doi: http://dx.doi.org/10.1155/2013/685614), and their color varied from lavender to purple.

Photocatalytic Water Splitting Experiments Using a 450 W Mercury Lamp.
The photocatalytic water splitting experiments were carried out in a glass enclosed reaction chamber with a quartz inner radiation reaction vessel.The glass chamber was connected to a gas circulation evacuation and water cooling system.In a typical experiment, 255 mg catalyst (as prepared Au/TiO 2 or Au/TiO 2 calcined at 200 ∘ C for 2 h), 322 mL distilled water, and 18 mL ethanol (or 18 mL methanol or 20 mM sodium sulfite) as sacrificial electron donor was taken in the glass reactor with a magnetic stir bar.The reaction mixture was evacuated and filled with Ar five times to remove all the dissolved gases.This was followed by irradiation using a 450 W high pressure Hg lamp via a quartz tube.Water at 20 ∘ C was circulated continuously through the outer walls of the reactor and the quartz vessel to make sure that the temperature of the reaction mixture did not exceed 35 ∘ C [40].The activity of these catalysts for hydrogen production was investigated during the first 5 h irradiation period, using a fresh catalyst each time. 2 M NaNO 2 solution was used as a filter to cut off UV radiation in our visible studies (UV-visible absorption spectrum of 2 M NaNO 2 is given in the supporting information).H 2 production was monitored using an online GC system (GOW-MAC 580 model) employing an All Tech molecular 80/100 sieve 5 A column with Ar as the carrier gas and a thermal conductivity detector.powder XRD.The diffuse reflectance spectra of Au/TiO 2 nanocomposites showed 2 sets of peaks (Figure 1): band gap transition band of TiO 2 with a maximum in the UV region and surface plasmon resonance band in the visible region (Au/anatase commercial with a maxima at 526 nm, Au/anatase aerogel with a maxima at 535 nm, Au/rutile with a maxima at 525 nm, and Au/P25 with a maxima at 560 nm).The position and shape of the surface plasmon resonance band depends on particle size, shape, and dielectric constant of the medium [41].The red shift of the surface plasmon band maxima in these nanocomposites can be attributed to the refractive index of the TiO 2 matrix.The band gap values of the respective phases of TiO 2 remained unchanged even after Au loading as per the DRS UV data (anatase = 3.2 eV, rutile = 3.0 eV, and P25 = 3.1 eV).The presence of loaded Au nanoparticles on the TiO 2 matrix was confirmed by TEM bright field imaging.The TEM bright field images of different phases of TiO 2 before and after Au loading are given in Figures 2 and 3. Commercial anatase has micron sized particles (0.1-0.2 m, Figure 2(a)) with a surface area 13 m 2 /g, whereas anatase prepared by the aerogel method consists of small nanometer sized particles (10-15 nm, Figure 2(b)) with a high surface area of 102 m 2 /g (Figure S1 and Table S1).TEM bright field images of these anatase particles showed embedded Au nanoparticles of size ranging from 1 to 9 nm (Figures 2(c) and 2(d)) with an average particle diameter of 2.4 nm and 4.3 nm (Figure 4) for aerogel and commercial anatase, respectively.The surface areas of these Au/anatase nanocomposites were less than the respective TiO 2 matrix used (Table S1).The rutile phase of TiO 2 used in our studies showed micron sized particles (0.2-0.6 m, Figure 3(c)).The surface area of rutile was found to be very low (Table S1).The TEM bright field image (Figure 3(d)) showed embedded Au nanoparticles on the semiconductor with a size ranging from 3 to 8 nm with an average particle diameter of 4.7 nm (Figure 4).In the case of P25, the particles size ranged from 15 to 20 nm (Figure 3(a)) with a surface area of about 48 m 2 /g (Table S1).TEM bright field images exhibited Au nanoparticles of size ranging from 3 to 10 nm (Figure 3(b)) with an average particle diameter of 6.7 nm (Figure 4).The powder XRD pattern (Figure S3) of these Au/TiO 2 nanocomposites did not show any Au reflections due to the low concentration (Table S2) of Au in these samples.

Results and Discussions
A conclusion from these characterization studies is that the final gold particle size depends a great deal on the surface area of the TiO 2 support.Therefore, as the Au-butanone colloid encountered the cold TiO 2 , gold nanoparticle growth is limited/controlled by the TiO 2 surface available.

Photocatalytic Activity of Au/TiO 2 Nanocomposites under
UV-Visible Conditions.All Au/TiO 2 photocatalysts prepared by our SMAD method showed UV-visible activity to generate hydrogen in the presence of a sacrificial electron donor, irrespective of the TiO 2 phase present in the system (Table 1).Among the different catalysts used, Au/P25 showed the highest activity with a hydrogen production rate of 1600 mol/h in the presence of ethanol.In addition to ethanol as a sacrificial agent, we used methanol and sodium sulfite as an electron scavenger in Au/P25 system.The rate of hydrogen evolution was lower compared to ethanol when methanol was used as the sacrificial agent with a photocatalytic hydrogen production of 816 mol/h.The photocatalytic hydrogen production was very low when 20 mM sodium sulfite solution was used as a sacrificial agent (20 mol/h).These results clearly indicate that during sacrificial electron donor mediated hydrogen production, there is a considerable contribution toward hydrogen generation from the sacrificial agent in addition to the hydrogen generation.
Au/TiO 2 photocatalysts were found to be much more active compared to their samples without gold.The higher activity could be attributed to the better charge separation  [17,18,[42][43][44][45][46].This kind of electron transfer from semiconductor to metal continues until Fermi level equilibration takes place.Choi et al. showed that under UV irradiation, Au@TiO 2 core-shell nanoparticles exhibit a blue shift in the surface plasmon resonance band maxima due to the transfer of electrons from TiO 2 to Au core and the Fermi level shift to negative potentials [47].Au cannot undergo this kind of the Fermi level shift without taking electrons from a semiconductor [47].visible light illumination photocatalytic experiments.From the UV-visible spectrum of NaNO 2 (Figure S4) it is clear that this chemical filter cuts off radiation only up to 400 nm.Since rutile does absorb in the 400-410 nm range, some charge transfer from TiO 2 to Au could take place.As expected, the rutile phase of TiO 2 is visible light active, whereas the anatase phase is visible light inactive due to its slightly large band gap.Au loading improved the visible light activity of TiO 2 nanoparticles containing the rutile phase.Thus, the enhanced visible activity of these rutile-containing photocatalysts could be due to better charge separation in the presence of Au nanoparticles by transferring electrons from the CB of TiO 2 to the Au Fermi level [16][17][18].

Photocatalytic Activity of TiO
In addition to the charge transfer from TiO 2 to Au, there are several reports where Au imparts visible activity to wide band gap visible light inactive semiconductor nanoparticles like TiO 2 , SrTiO 3 , and CeO 2 [19][20][21][22][23][24][25][48][49][50][51].García and coworkers prepared visible active Au loaded P25 or CeO 2 nanocomposites for hydrogen or oxygen production by water splitting in the presence of a sacrificial electron donor (methanol or EDTA) or acceptor [16,48].According to these reports, Au nanoparticles absorb visible light by the surface plasmon resonance effect; charge separation at the Au nanoparticle takes place by the transfer of excited electrons from Au to the CB of TiO 2 ; subsequent holes are quenched by the sacrificial electron donor.Kimura et al. showed that the rutile form of TiO 2 more favors this kind of interfacial electron transfer from Au to TiO 2 [19].Essentially, in these reported systems, Au nanoparticle acts as a sensitizer as in the case of dye-sensitized solar cells [49].
However, this kind of electron transfer from Au to the CB of TiO 2 is not energetically favored [26].On the other hand, Tian and Tatsuma showed that this kind of electron transfer is possible due to the formation of a Schottky barrier at Au/TiO 2 junctions [20,50].In the case of Pt (05 wt%) /SrTiO 3 , the Pt metal decreases the band gap energy of SrTiO 3 by the formation of a Schottky barrier at the Pt/SrTiO 3 interface [51].Thus, there is still controversy regarding electron flow to or from metal particles with TiO 2 .
To get a better overall understanding about how Au nanoparticles impart visible activity to TiO 2 semiconductor, we examined the visible activity of Au/anatase nanocomposites.Two different types of anatase were used: a commercial one with micron sized particles and anatase prepared by an International Journal of Photoenergy aerogel method with nanometer sized particles.Regardless of the anatase used for Au loading, we did not see any visible light activity for the first five hours in the presence of water ethanol mixture.However, there was a very small amount of hydrogen evolution sixth-hour onwards.The mechanism of formation of small amount of hydrogen after five hours of visible light illumination is not clear to us, which could be either due to the reduction of water or due to the thermal decomposition of ethanol, aided by small nanoparticles undergoing nonradiative decay forming localized hot spots [52,53].It has also been reported that Au nanoparticles can also impart visible activity to a plasmonic-metal/semiconductor nanocomposite by surface plasmon mediated local electric field enhancement [26][27][28][29].Interaction of a semiconductor nanoparticle with this kind of localized electric field could allow the formation of electron/hole pairs in the near surface region which can migrate to the surface without undergoing electron/hole pair recombination.For example, electromagnetic simulations showed that the increased photocatalytic water splitting under visible illumination of anodic TiO 2 with Au nanoparticles is due to the local electric field enhancement near the TiO 2 surface instead of electron transfer from Au to the CB of TiO 2 [26].Awazu et al. called this kind of surface plasmon-induced localized electric field enhancement and improved photocatalytic activity as "plasmonic photocatalysis" [54].Christopher et al. explained the enhanced visible activity of Ag/TiO 2 as due to the higher concentration of charge carriers in the semiconductor by plasmon mediated radiative transfer of energy from Ag nanoparticles to TiO 2 , no electron flow from Ag to TiO 2 was observed [55].Recently, Seh et al. showed enhanced visible light activity for the Janus shaped anatase Au-TiO 2 nanocomposites as due to the strong localization of plasmonic near fields close to the Au-TiO 2 interface and the coupling of plasmonic near fields to optical transitions involving localized electronic states in amorphous TiO 2 [29].In our case, this kind of localized electric field created by Au nanoparticles could not generate electron/hole pairs in the near surface region of the semiconductor as the surface plasmon absorbance of Au

UV-visible light
Step 1 Step We strengthened this hypothesis by doing photocatalytic water splitting under UV-visible conditions using Au/P25 nanocomposites in the absence of a sacrificial electron donor.There was an evolution of hydrogen (Figure 5) which again supports the proposed mechanism of charge separation achieved by Au nanoparticles (acts as a sink for photogenerated electrons and the Fermi level equilibration) and water molecules (quench the holes in the valence band of TiO 2 ).In order to prove that the hydrogen evolution is not caused by some organic impurities like butanone incorporated into the Au/P25 photocatalyst during sample preparation, we used calcined Au/P25 (at 200 ∘ C for 2 h).This sample showed photocatalytic activity for several cycles (Figure 5).We also did control experiments, wherein we used Degussa P25 and tried photocatalytic water splitting in the absence of ethanol; there was no hydrogen evolution.These results imply a mechanism shown in Scheme 1.The rate of hydrogen evolution was low for Au/P25 in the absence of ethanol.P25 produced no hydrogen in the absence of ethanol; however, in the presence of ethanol, there International Journal of Photoenergy was hydrogen evolution.This implies two things: (1) charge separation in pure TiO 2 takes place only in the presence of a sacrificial electron donor; (2) charge separation takes place in TiO 2 nanoparticles which are in good interfacial contact with Au nanoparticles even in the absence of a sacrificial electron donor.In the absence of ethanol, initially, charge separation takes place only on those TiO 2 particles which are in good interfacial contact with Au nanoparticles by transferring photoelectrons from the CB of TiO 2 to the Au Fermi level; holes are being quenched by water molecules to produce H 2 O 2 or O 2 .Thus, charge separation takes place in the case of photoexcited TiO 2 nanoparticles which are in good interfacial contact with TiO 2 , and the Au/TiO 2 Fermi level equilibration takes place.Subsequent water reduction and hydrogen evolution are observed.The amount of hydrogen evolved in the absence of ethanol is low compared to the hydrogen evolved in the presence of ethanol.This could be due to the low concentration (1 wt%) of Au (which causes the charge separation) in these catalysts.We could not see any oxygen evolution during these experiments.This could be due to several reasons: (1) the reaction mixture was deaerated before photoillumination, and evolved small amount of oxygen could dissolve in water due to the high solubility of oxygen in water compared to hydrogen [57], (2) evolved oxygen can accept photoexcited electron from TiO 2 to form O 2 −• [58,59].The rate of the hydrogen evolution decreases with time as the evolved oxygen acts as a better electron acceptor to form superoxide.

Conclusions
These Au/TiO 2 photocatalysts were found to be resilent photocatalysts for the generation of hydrogen under UV-visible conditions.The photocatalytic hydrogen evolution rate was found to be 1600 mol/h which is higher than Au/TiO 2 nanocomposites prepared by the photodeposition method (in our hands).Comparisons of different TiO 2 crystalline forms with different surface areas suggest that, in our case, the best explanation of any visible light activity is that the rutile form can absorb some energetic visible light, and then electrons are scavenged by the Au nanoparticles.This is the same mechanism as with more energetic UV light, where the process is much more efficient.
When ethanol scavenger is present, these photocatalysts are long lived.In the absence of ethanol, catalyst degradation takes place slowly (over days); this process of degradation is not understood, and hydrogen generation is much slower.
Further work comparing different metal nanoparticles is underway.

Figure 5 :
Figure 5: Hydrogen evolution by photocatalytic water splitting in the absence of sacrificial electron donor using Au/P25 (calcined sample) under UV-visible conditions.
[56]me 1: Elementary steps showing photocatalytic water splitting under UV-visible illumination using Au/TiO 2 nanocomposites: step 1 is electron transfer from the CB of photoexcited TiO 2 to the Au Fermi level, and step 2 is the Au Fermi level shift to negative potential by gathering photoexcited electrons from TiO 2 and water reduction.nanoparticledoesnotoverlapwiththe TiO 2 absorption spectrum.Linic and Ingram observed enhanced visible activity with Ag loaded N-TiO 2 , whereas Au loaded N-TiO 2 did not show any significant enhancement in the photocurrent due to the red shift in the position of surface plasmon absorption band of Au nanoparticles compared to that of Ag[56].3.4.Au Nanoparticles in Au/TiO 2 Nanocomposites Act as an Electron Sink for Better Charge Separation and TotalWater Splitting.Our results indicate that there is a transfer of electrons from TiO 2 to Au and Au nanoparticle acts as a sink for photogenerated electrons.If this is true, Au nanoparticles could aid charge separation without using a sacrificial electron donor.