Photocatalytic H2 Production Using Pt-TiO2 in the Presence of Oxalic Acid: Influence of the Noble Metal Size and the Carrier Gas Flow Rate

The primary objective of the experiments was to investigate the differences in the photocatalytic performance when commercially available Aeroxide P25 TiO2 photocatalyst was deposited with differently sized Pt nanoparticles with identical platinum content (1 wt%). The noble metal deposition onto the TiO2 surface was achieved by in situ chemical reduction (CRIS) or by mixing chemically reduced Pt nanoparticle containing sols to the aqueous suspensions of the photocatalysts (sol-impregnated samples, CRSIM). Fine and low-scale control of the size of resulting Pt nanoparticles was obtained through variation of the trisodium citrate concentration during the syntheses. The reducing reagent was NaBH4. Photocatalytic activity of the samples and the reaction mechanism were examined during UV irradiation (λmax = 365 nm) in the presence of oxalic acid (50 mM) as a sacrificial hole scavenger component. The H2 evolution rates proved to be strongly dependent on the Pt particle size, as well as the irradiation time. A significant change of H2 formation rate during the oxalic acid transformation was observed which is unusual. It is probably regulated both by the decomposition rate of accumulated oxalic acid and the H+/H2 redox potential on the surface of the catalyst. The later potential is influenced by the concentration of the dissolved H2 gas in the reaction mixture.


Introduction
Hydrogen is one of the cleanest energy sources on Earth. The field of heterogeneous photocatalysis can help convert the energy of sunlight to chemical energy via the production of H2 gas from water [1,2]. TiO2 photocatalysts can be excited within the UV spectral range, although in most cases bare titanias have low efficiency for photocatalytic water splitting. The overvoltage of H2 evolution and the recombination rate of electron-hole pairs can be lowered by depositing noble metal particles onto the surface of TiO2 [3]. In the literature Pd [4,5], Pt [6][7][8], Ru [9], Rh [4], Au [10,11] and Ag [12] are used to enhance the photocatalytic efficiency of H2 generation on TiO2. Noble metal particles can be deposited onto the catalyst's surface by either sonochemical method [13], photoreduction [14][15][16], or chemical reduction (in liquid phase with e.g., hydrazine [17,18] or sodium borohydride [19,20]).
The size and amount of noble metal particles on the TiO2 surface are very important factors that can affect the overall activity of the catalyst. Larger quantities of metal particles on the titania can block the active sites and large number of metal deposits can decrease charge carrier space distance which leads to increased electron-hole recombination rates [5,21]. In most cases, noble metal content of 1 wt% [22][23][24] or below [25,26] is considered to be optimal. The size and dispersion of noble metal nanoparticles can determine the number of these particles per one TiO2 particle. Theoretically, larger Pt particles form fewer noble metal islands on TiO2 than smaller ones do if noble metal content is identical. The electrostatic effect of noble metal islands is limited to a group of nearby TiO2 particles. According to this phenomenon, smaller sized noble metal particles are desirable theoretically. However, if the deposit size is too small, it is hard to establish sufficient electrical contact for efficient charge transfer [27]. The size of noble metal nanoparticles can be gently controlled by adding surfactants to the metal precursor containing solution before the reduction process. The dimensions of the so forming nanoparticles are dependent upon the surfactant's quality and concentration [28,29]. In regard to the important influence of metal nanoparticle size on the catalyst's activity, developing an easily adjustable synthesis method with good reproducibility and using additives that can be readily removed after reduction are substantial steps to make the process economic. Noble metal nanoparticles below 10 nm deserve special attention in these experiments.
During the photocatalytic reactions, the presence of easily oxidizable organic compounds in the reaction mixture (instead of using only pure water) is desirable. These electron donors (so called sacrificial reagents) react irreversibly with the photoinduced holes resulting in suppressed electron-hole recombination rates. In numerous papers, different kinds of alcohols (especially methanol) are used as hole scavengers [30][31][32][33]. Some carboxylic acids also can act like effective hole scavengers, such as formic acid [34], chloroacetic acid [35] or oxalic acid [16]. The enhancement in H2 production also depends on the concentration of the sacrificial reagent. While there are numerous publications about enhancing the hydrogen evolution efficiency by depositing Au on TiO2, Pt loaded TiO2 might be a more efficient photocatalyst for H2 production due to its better ability to act as an electron trap as proved recently [36].

Experimental Section
1 wt% (mPt/mTiO2) Pt-modified TiO2 photocatalysts were prepared containing differently sized Pt nanoparticles. Particle size was gently adjustable by varying only one factor during the syntheses: the concentration of the surfactant. Thus, photocatalytic activities of these catalysts are well comparable according to the Pt particle size. Effect of the noble metal particle size on the photocatalytic H2 production was thoroughly investigated in the range of 2.5-4.5 nm.

Catalyst Preparation
Aeroxide P25 (Evonik Industries AG, Essen, Germany) TiO2 (Dav = 25.4 nm, 90% anatase and 10% rutile) was used as bare catalyst. All the syntheses and photocatalytic tests were carried out in Millipore MilliQ ultrapure water as medium. Pt nanoparticles were deposited onto the TiO2 surface by chemical reduction (CR) or photoreduction (PR) methods.

Chemical Reduction Method
During this synthesis method, different concentrations of trisodium citrate (2.50 × 10 −4 M; 1.88 × 10 −4 M; 1.25 × 10 −4 M and 0.63 × 10 −4 M) were used. Citrate anions, which help stabilize the Pt nanoparticles in the growing step after nucleation, provide good conditions for size focusing and synthesizing monodisperse noble metal nanoparticles. The reaction mixture was thermostated at 20 °C . Trisodium citrate was added to the TiO2 suspension (cTiO2 = 5 g/L), followed by H2PtCl6 (final concentration: 2.5 × 10 −4 M). Finally, freshly prepared, ice-cold NaBH4 solution was added as a reducing agent (final concentration: 3 × 10 −3 M). The suspension immediately turned grey. As the reduction took place in the presence of TiO2, this procedure was designated CRIS (chemical reduction, in situ). After 1 h the suspension was washed by centrifugation, re-suspending the particles in oxalic acid solution (5.0 × 10 −2 M) to improve the sedimentation and to get rid of the chloride and sodium remnants. The final suspension was used freshly for photocatalytic test without any further processing.
Other portion of Pt-modified TiO2 catalysts were prepared by mixing the previously chemically reduced Pt sol with the TiO2 suspension (chemically reduced, sol-impregnated samples, CRSIM). The washing procedure was the same as for the in situ prepared Pt-TiO2.

Photoreduction Procedure
To promote the photoreduction of Pt(IV) ions, oxalic acid (OA) was added to the reaction mixture as hole-scavenger organic compound (PROA samples). UV photons excite the TiO2, the organic compound is oxidized by the holes, and at the same time Pt(IV) is reduced by the excited electrons.
The suspension was next subjected to UV irradiation for 2 h to allow photoreduction of the noble metal ions. Within 1-5 min, there was a characteristic color change, from white to gray, which indicated the formation of Pt nanoparticles. Then, the suspension was washed by centrifugation as described in Section 2.1.1. For the photocatalytic experiment, the redispersed catalyst was used immediately. During this synthesis method, the reduction process is much slower than in the case of chemical reduction; therefore, slightly bigger Pt nanoparticles are formed.

Spectrophotometry
The UV-VIS spectra of Pt sols were measured in 1 cm quartz cells in an Agilent 8453 diode array spectrophotometer (Agilent Technologies, Santa Clara, CA, USA), with Millipore MilliQ ultrapure water as blank reference.

Transmission Electron Microscopy (TEM)
The average size of the Pt nanoparticles deposited on Aeroxide P25 TiO2 was calculated according to TEM images recorded with a 200 kV Fei Tecnai G2 20 Xtwin instrument (FEI, Hillsboro, OR, USA). The catalyst samples were investigated immediately after preparation.

ICP Measurements
Concentration of Pt that may have been detached or dissolved from the surface was determined by ICP-MS spectrometry (Model 7700x, Agilent Technologies, Santa Clara, CA, USA), equipped with a Micromist nebulizer, Peltier-cooled spray chamber and an integrated autosampler (G3160B, Agilent). Prior their use, all labware (flasks, autosampler vials, etc.) were thoroughly cleaned using trace quality HNO3 (Suprapur, Merck KGaA, Darmstadt, Germany) and HCl. External calibration was used based on the signal from the 195 Pt isotope, with a good linear fit (r 2 = 0.9998). Standard solutions were diluted from Pt stock solution (ICP-MS standard, Certipur, Merck) using Millipore MilliQ quality deionized water.

H2 Production Measurements
The suspension of the freshly-prepared, washed catalyst in 50 mM oxalic acid solution was poured into a glass reactor (total volume: 150 mL), surrounded by ten 15 W UV lamps (λmax = 365 nm). The well-stirred suspension (ccatalyst = 1 g/L) was purged with N2 at a flow rate of 50 mL/min (in the majority of experiments) to ensure O2-free conditions. The reactor was connected through a PTFE tube to a Hewlett Packard 5890 gas chromatograph fitted with 5Å molecular sieve column and a thermal conductivity detector. Gas samples were taken with a 2 mL sampling valve, every 10 min in the first hour of the experiment and every 20 min in the second hour. The rate of H2 evolution was calculated with regard to the GC calibration (carried out with certified 5% H2:N2 gas) and the N2 flow rate.
The reactor was characterized by actinometric measurement according to the iron oxalate method (IA = 6.35 × 10 −5 Einstein/sL).

UV Decomposition of Oxalic Acid
These experiments were carried out under the same conditions as H2 production measurements, but liquid samples were taken from the suspensions at predetermined intervals during the reaction, and the residual oxalic acid concentration was measured. After centrifugation and filtration with a Whatman Anotop 25 0.02 μm syringe filter, the HPLC measurements were performed on a Merck Hitachi device fitted with an L-4250 UV-VIS detector and a GROM Resin ZH 8 μm column.

Adsorption of Oxalic Acid on the Catalyst
Oxalic acid adsorption was investigated on bare and platinized TiO2. The applied oxalic acid concentrations were in the range of 0.1-50 mM. The 1 g/L suspensions were thermostated at 25 °C and stirred for 4 h in the dark. Then the samples were filtered using a Whatman Anotop 25 0.02 μm syringe filter and the residual oxalic acid concentration was determined by HPLC detailed in Section 2.4.

Reduction of Pt(IV)
Pt(IV) ions have a light absorption peak at ~258 nm [37]. This peak completely disappeared after the chemical reduction procedure, confirming that Pt(IV) ions were reduced to Pt(0) during the 1 h reaction time ( Figure 1). It should be mentioned that with all procedures, the supernatants of Pt-TiO2 suspensions were colorless, which indicates that the Pt nanoparticles were all well stabilized on the surface of TiO2.

Size of the Pt Particles on TiO2
By measuring ~200 particles of Pt on each sample's TEM images, the size of the noble metal nanoparticles appeared to be dependent on the stabilizing trisodium citrate concentration during the chemical reduction process. Larger average Pt nanoparticle size was obtained with lower stabilizer concentration. The size distribution of the formed Pt nanoparticles is shown in Figure 2. The Pt particles were mainly spherical.

Effect of Surface Pt Nanoparticles on the Oxalic Acid Adsorption
Oxalic acid is adsorbed on the TiO2 surface by dissociative adsorption [38]. The adsorption properties of oxalic acid were investigated on bare and Pt modified catalysts in 1 g/L suspensions at 25 °C. The bare Aeroxide P25 TiO2 adsorbed almost the same amount of oxalic acid on its surface than did the 1 wt% Pt-modified Aeroxide P25 (Figure 3). At the initial oxalic acid concentration applied in the H2 production experiments (50 mM), all of the binding sites on the TiO2 surface were likely to be covered by oxalate ions (adsorbed amount is 3 mmol/g). It can be concluded that modifying the titania surface with 1 wt% Pt nanoparticles does not affect considerably the amount of binding sites on the TiO2 surface within experimental error.

H2 Production from Oxalic Acid Solution
H2 evolution was measured in the presence of 50 mM oxalic acid in N2-purged suspensions. Relatively high oxalic acid concentration was applied in order to keep the substrate concentration decrease negligible: during the measurement oxalic acid can decompose mostly to CO2 and H2 under O2-free conditions [16]. Moreover, the initial concentration of oxalic acid was at least 200 times higher than the citrate concentration used during the syntheses. However, most of the citrate ions were presumably washed out during centrifugation. There was a significant decrease in H2 production in the first 40 min of irradiation using any Pt-TiO2. After this period, the H2 production curves reached saturation level and the H2 evolution rate subsequently remained almost constant for the rest of the experiment with all Pt-modified TiO2 catalysts.
Some experiments were performed to find an explanation for this phenomenon. We investigated the possibility of the change in Pt nanoparticle size during the photocatalytic reaction. Comparing the average platinum particle size in TEM images before the experiment and after 2 h of UV-irradiation in oxalic acid solution, we determined that irradiating these catalysts does not affect considerably the average particle diameter and the size distribution of noble metal nanoparticles (Figure 4). Monitoring the platinum concentration in the solution phase by ICP measurement, we did not observe any significant Pt dissolution from the surface of the catalyst during the experiment (Table 1). The platinum concentration in the solution phase was 1.87 ppb and remained below 1 ppb during irradiation which is less than 0.01% of the total platinum concentration in the suspension. Therefore, the disappearance of metallic platinum from the surface of catalyst is insignificant. However, changing the flow rate of purging gas had an important influence over the hydrogen formation at the first stage of experiments. The initial maximum of H2 evolution prolonged by increasing and disappeared by decreasing the flow rate of N2, as is shown in Figure 5. It is theoretically plausible that at the first stage of hydrogen formation the reaction is regulated electrochemically by the redox potential of H-electrode formed by the hydrogen absorption-desorption rate in platinum [39]. With a higher N2 flow, the hydrogen absorption is limited, but at low flow rates, a steady state redox potential can set much faster. Thus, no initial peak in the hydrogen evolution curve was observed at 25 cm 3 /min N2 flow, while there is a prolate maximum curve when adjusting the carrier gas to higher flow rates.

Effect of Pt Particle Size on Photocatalytic H2 Generation
With each chemical reduction method, the highest steady state H2 evolution rates were achieved with the following catalysts: (i) the 1.25 × 10 −4 M citrate stabilized Pt-TiO2 CRIS sample (DPt = 3.2 nm; rH2,steady = 7.21 μmol/min), (ii) the 1.88 × 10 −4 M citrate stabilized Pt-TiO2 CRSIM sample (DPt = 3.5 nm; rH2,steady = 4.78 μmol/min). The correlation between the platinum particle size and the steady state or the initial maximal H2 evolution rates are demonstrated in Figure 6. The distribution of Pt particles on the TiO2 surface appeared to be more uniform in the samples synthesized by in situ chemical reduction method. As a result, CRIS samples had higher hydrogen producing capabilities than the sol-impregnated ones.
At DPt = 3.2 nm average particle size, there was a peak in photocatalytic activity. Smaller Pt nanoparticles are unfavorable, probably due to the loss in metallic character [40,41]. Using photoreduced Pt-TiO2 resulted in the lowest photocatalytic activity with the largest Pt nanoparticle size (DPt = 4.5 nm) on the surface. The average Pt particle size on each catalyst and the respective steady state and maximal H2 evolution rates are demonstrated in Table 2.  It was concluded that the most important parameter influencing the H2 production efficiency of Pt-TiO2 photocatalysts appeared to be the Pt nanoparticle size. As Pt clusters contain different numbers of atoms, the structure and surface bonding is changed which leads to different catalytic activities [42]. The optimal size of the Pt nanoparticles with the best distribution on the TiO2 surface was achieved with the CRIS method.
To investigate the long term stability of the best performing Pt-TiO2 catalyst, the suspension was irradiated for several hours. The H2 evolution rate was almost constant until 400 min, when the oxalic acid (cinitial = 50 mM) totally decomposed. After full mineralization of oxalic acid on the catalyst in about 7 h, the concentration of oxalic acid was re-adjusted to 50 mM in dark. The UV irradiation was restarted after 30 min of stirring and nitrogen purging, in order to provide enough time to achieve the adsorption equilibrium in dark. In the re-initiated reaction the H2 evolution rate was almost the same as at the beginning of the experiment with decrease of the reaction rate to the same level after 40 min (Figure 7). In another experiment, oxalic acid was also fully mineralized but the concentration was re-adjusted without switching off the lamps and without waiting for the hydrogen to completely purge out from the suspension and Pt as well. The initial H2 evolution rate was about 40% of the original one and it decreased to about the same steady state level that was observed in the first cycle. These results support the plausible hydrogen absorption mechanism and formation of H-electrode described above.
We also compared the results from our parallel study using Au modified titanias with our best performing Pt-TiO2 sample. In one case, the Au nanoparticles were almost the same size as Pt (DAu = 3.5 nm, DPt = 3.2 nm), while in another case, exactly the same chemical reduction method was used (DAu = 5.7 nm). The optimally sized Au-TiO2 catalyst showed an almost constant and (in the steady state) nearly double H2 evolution rate than in the case of the above mentioned best Pt-TiO2 photocatalyst. It is also noticeable that modifying Aeroxide P25 TiO2 with Au or Pt nanoparticles in almost the same size resulted in roughly the same H2 evolution rates in 50 mM oxalic acid solution (Figure 8).
In case of Au modified catalysts, initial maximum in hydrogen formation was not observed, indicating no H + /H2 electrode formation on metallic Au deposited onto TiO2.  Au-TiO2 with similar average noble metal particle size () or synthetized by the same chemical reduction method (▲).

Decomposition of Oxalic Acid under Anaerobic Conditions
We investigated the correlation between the H2 production and the decomposition of oxalic acid under the same conditions. The residual oxalic acid concentration was determined by HPLC ( Figure 9). For this experiment, we used the Pt-modified photocatalyst that performed the best in the H2 production measurements (DPt = 3.2 nm, CRIS sample). Correspondingly to the H2 evolution, the curve for oxalic acid decomposition is in accordance with hydrogen accumulation. The assumed sum reaction is: (COOH)2 → 2 CO2 + H2 (1) Figure 9. Photocatalytic decomposition of oxalic acid in the solution phase () and H2 production () under UV irradiation and anaerobic conditions with Pt-P25 CRIS (DPt = 3.2 nm) photocatalyst.
After turning on the lamps, the decomposition and adsorption reaches an equilibrium state, so the H2 evolution becomes nearly constant. It is important to emphasize that the oxalic acid concentration was measured in the solution phase by HPLC, thus the decomposition of the adsorbed oxalic acid cannot be seen in this representation. It is plausible to think that at the beginning the total decomposition rate of oxalic acid chemisorbed on the surface is significantly higher than the decomposition rate in the solution phase. Therefore, the measured hydrogen evolution rate can be accordingly higher than the oxalic acid decomposition rate in the solution phase. At later stages, the decomposition rate of oxalic acid is nearly the same as the hydrogen evolution rate.

Conclusions
Differently sized Pt nanoparticles were synthetized onto Aeroxide P25 TiO2 by chemical reduction and by photoreduction with constant platinum content (1 wt%). We were able to finely control the size of Pt nanoparticles through the use of different concentrations of the stabilizing agent. Two developed chemical reduction methods (CRIS and CRSIM) were utilized and the efficiency of the catalysts for H2 production and the size distribution of surface Pt particles were also investigated. In UV-irradiated O2-free suspensions, the Pt-modified TiO2 catalysts showed high H2 production activities in the presence of oxalic acid. The photocatalytic activity proved to depend strongly on the average Pt particle diameter: there was an optimum in photocatalytic activity at ~3.2 nm average Pt nanoparticle size on the TiO2 surface. In the majority of publications, platinum modified TiO2 photocatalysts are used for H2 production. However, comparing our best performing Pt-TiO2 catalyst with our previous experiments with Au-TiO2 prepared by similar chemical reduction method, the performance of Au-TiO2 catalysts (with same Au size or made by exactly the same procedure as Pt-TiO2) was as good as or even better than the Pt-TiO2 when oxalic acid was used as model compound.
The long-term usability of our best performing catalyst was also investigated: until the irradiated suspension contains any easily oxidizable organic compound (e.g., oxalic acid), H2 evolution rate remains nearly constant. Considering this result, these catalysts might be utilized effectively for H2 production without the loss of photocatalytic activity.
We also investigated the phenomenon of a significant decrease in H2 evolution rate at the first stage of experiments when oxalic acid was used as a model compound. It turned out that the formation of H-electrode on platinum might be the key reason of the limited H2 generation.