Marek Laniecki, Faculty of Chemistry, A. Mickiewicz University, Grunwaldzka 6,

60-780 Poznan, Poland

Maciej Zalas, Faculty of Chemistry, A. Mickiewicz University, Grunwaldzka 6,

60-780 Poznan, Poland

Industrial methods of hydrogen generation based on fossil fuels are well known since the end of XIX and beginning of XX century. Recent reports of the big oil companies clearly indicate that estimated reserves of natural gas and oil will last at least for 50 years. The constant and increasing depletion of fossil fuels, especially natural gas, applied in the steam-methane reforming (SMR), requires the research and implementation of new methods of hydrogen production based on renewable energy resources. It seems that one of the very promising routes of hydrogen generation directly from water is photocatalysis. Different “clean” methods of hydrogen generation are currently considered, however, only two of them, namely biophotolysis and photocatalysis seems to be the only simple ways of transformation of solar energy into chemical energy stored in hydrogen. Although these methods are still in the “juvenile” period of research, it is belived that after improvement of the yields of photogenerated hydrogen they can become one of the very important way of solar energy storage.

This paper is an attempt to clarify the doping effect of platinized titania applied as photocatalyst in hydrogen generation from water.

General remarks

The early studies [1,2] of water photocatalysis showed that this method, applying titania as photocatalysts, gives rather low yields of hydrogen and requires illumination mainly in the region of ultraviolet light. Later works disclosed that different semiconductors such as sulfides, perovskites and many other combinations of transition metal oxides can operate as the efficient photocatalysts as well [3,4]. Unfortunately the most efficient photocatalyst — CdS, operating in the region of visible light, due to the environmental reasons and very rapid photocorrosion has been eliminated from further application considerations. Since the early seventies of the XX century different laboratories tried to improve both the hydrogen yield and photostability of the applied photocatalysts. Recently such materials like indium titanate, indium tantalate [5] as well as copper(I) managanate(III) [6] were tested under the visible light in water splitting. However, the yields of photocatalytically generated hydrogen were still far beyond the expectations.

This situation prompted many research groups to improve the particular semiconducting systems by modifying the surface of the photocatalyst. To date four different methods of modification have been studied:

— modification of the semiconductor surface with metal (e. g. Pt, Au, Ag, Ni )

— generation of composite semiconductors (e. g. CdS-TiO2 systems [7])

— surface sensitization (e. g. titania with chemisorbed Ru(bpy)32+complexes)

— transition metal doping.

The benefit of transition metal doping species is better capability of trapping electrons to inhibit electron-hole recombination during illumination. The concentration of the beneficial transition metal oxides dopants is very small whereas large concentrations are detrimental. Literature data survey indicate that transition metal oxides such as iron (III)[8] or copper (II) [8,9] actually inhibit electron-hole recombination. Other authors claim that doping titania with e. g. Cr3+ ions [10] leads to the decrease of photoactivity due to the very fast
recombination in electron-hole process. It is belived that these transition metals create acceptor and donor centers where direct recombination occurs.

The different opinions about the inhibition of electron-hole recombination indicate that this field of research requires further studies which in consequence can lead to the significant improvement of the photocatalytical splitting of water.

The research in our group since few years was concentrated on the doping of titania with yttria and oxides of the lanthanide group [11-13], and the present paper is the contribution to a better understanding of this phenomenon during photocatalytic generation of hydrogen.

Experimental

Two series of platinized titania photocatalysts were prepared by hydrolysis of titanium chloride containing appropriate amounts of lanthanides nitrates. Lanthanide oxides after dissolution in nitric acid were placed in 1 dm3 of distilled water and next titanium chloride was added dropwise at room temperature. The hydrolysis process was completed after adding ammonia till pH = 9. The obtained precipitate after removal of chloride and nitride ions was dried and calcined at 675 K for two hours.

Powdered samples containing 0.1 or 0.5 mol % of lanthanides were impregnated by the incipient wetness method with chloroplatinic acid. The concentration of the impregnating solution was adjusted to obtain 0.3 wt. % of supported platinum. The supported H2PtCl6 was decomposed by calcining of the samples at 675 K. The reduction of platinum oxide to the supported platinum particles was performed in situ in photocatalytic reactor at the initial stage of reaction.

Photocatalytic experiments were performed in Pyrex glass round bottom flask (250 cc) applying four Ultra-Vitalux bulbs (300 watts each — Osram ). The applied source of light indicate very similar spectrum of light as the natural solar irradiation. In order to mimic natural conditions the temperature of the irradiated reactor was close to 355 K. The evolved hydrogen in the argon carrier gas was directed towards gas-sampling loop and further directed onto TCD of Varian 3800 gas chromatograph.

In all cases 0.1 gram of photocatalyst was applied and the mixture of water and methanol (50:1) was used as the reaction medium. Methanol was applied as the sacrificial agent. Occasionally other higher alcohols were applied as the sacrificial reagents.

The applied catalysts were characterized with XRD, BET surface measurements, UV-VIS reflectance spectroscopy, temperature programmed reduction whereas dispersion of platinum was measured with hydrogen chemisorption.

Results

The results presenting the amount of the evolved hydrogen after 100 minutes of the photoreaction are shown on Figure 1. As it was shown in our earlier studies [11] the best performance for the samples containing only 0.1 mole % of lanthanide were found for yttria doped catalysts. Relatively low yields of hydrogen obtained for this series of catalysts (low lanthanide oxide content) indicate that doping of titania with lanthanides oxides, depending on the element applied, is very complicated process and still requires much efforts to understand this phenomenon. No simple relationship between the amount of electrons located on f-orbital of the dopant and catalytic photoactivity in water splitting reaction can be established for the samples containing 0.1 mole % of the lanthanides oxides. Certainly this amount of the dopant probably speeds up the electron-hole recombination, however, other explanations can not be excluded. It is also possible that platinum dispersion, which is strongly related with the rate of chloroplatinic acid decomposition can influence the yield

of hydrogen. Photoreduction of platinum is the multi step process what is visible for all studied catalysts as the hydrogen evolution rate at the initial stage of reaction. Figure 2 presents the rate of hydrogen evolution for three selected catalysts, but similar shapes were always observed for all studied samples. The characteristic steps visible during first hour of reaction clearly shows that independently of the applied lanthanide oxide and its concentration, the stabilization is always attained after first hour of reaction.

Figure 1B (concentration of lanthanide oxide — 0.5 mole %) which shows the amounts of hydrogen evolved after 100 minutes of reaction, indicate in certain cases (Gd, Eu, Sm, Ho) hydrogen yields higher than for pure platinized titania (1.66 mmol/100 min.). It is worth to notice that in the absence of metallic (Pt) photocatode the yields of hydrogen were on the level of detection. The analysis of the hydrogen yields for samples containing 0.5 mol% of such ions like Ce4+, Tb4+, Pr4+ indicate that these ions can protect Ti4+ ions against their photoreduction towards Ti3+and in consequence to very rapid recombination of electrons. In contrast, those samples containing samarium, europium gadolinium or holmium indicate

that ions of these elements incorporated within the anatase (XRD measurements) structure protect the excited electrons from rapid recombination.

However, the temperature programmed reduction (see Figure 3) in the case of cerium indicated that reduction of cerium doped titania occurs within the same region of tempera­tures like pure non-doped titania (850-1100 K). Moreover, the presence of 0.5 mol % of ceria in catalysts containing 0.3 wt.% of platinum significantly lowers the temperature reduction of supported platinum particles. The presence of two maxima at « 480 K and 620 K suggests formation of two different platinum species. This specifically low temperature of platinum particles reduction requires further studies. It is also unclear why certain surface segregation of platinum particles occurs.

The application of sacrificial compounds, other than methanol, indicate that in the case of water splitting other donors of electrons (such organic compounds like ethanol, glycol, glycerine or glucose) can fulfill their role as well. However, the best results were always obtained for methanol or ethanol and the worse for glucose or fructose. This results indicate that in the future experiments also certain wastes from food industry can be applied as the source of electrons for photoreduction of water.

An attempt to clarify the role of lanthanide oxide as a dopant of titania has been made on the basis of electronic spectra of synthesized non-platinized photocatalysts. Spectra presented on Figure 4 show typical shapes, band positions and adsorption intensities for the studied catalysts. The comparison of these spectra within 200-400 nm region indicate that the ratio between the electronic absorption band at 385 nm and the bands at 270 and 250 nm can serve as the indicator of better or worse activity in photocatalytic water

Figure 4.

UV-VIS reflection spectra of selected non-platinized photocatalysts containing 0.5 mol % of lanthanide oxide.

splitting. The activity in hydrogen photoevolution in the case of samples containing 0.5 mol % of lanthanide oxide can be related with standard oxidation potentials versus hydrogen electrode. It was established that those ions which indicate high potential of tetravalent strongly acidic aqua ions eg. Gd (7.9 V) Eu (6.4 V), Yb (7.1 V) or Ho (6.0 V) can be responsible for relatively high hydrogen yield in photocatalytic splitting of water.

The preliminary experiments under natural irradiation and comparison of these results with those performed under the laboratory conditions with the same catalysts indicate that both methods with same catalysts leads to very similar results. Those catalysts which indicated good performance in laboratory were also good photocatalysts under natural irradiation.

References

1. A. Fujishima, K. Honda, Nature, 238(1972)37

2. J. Augustynski, J. Electrochim. Acta, 38(1993)43

3. A. L. Linsebigler, G. Lu, J. T. Yates Jr, Chem. Rev. 95(1995)735

4. A. Mills, S. La Hunte, J. Photochem. Photobiol., A:Chem. 108(1997)1

5. Z. Zhou, J. Ye, K. Sayama, H. Arakawa, Nature, 414(2001)625

6. Y. Baasekouad, M. Trari, J. P. Doumerc, Int. J. Hydr. Energy, 28(2003)43

7. K. R. Gopidas, M. Bohoroquez, P. V. Kamat, J. Phys. Chem.,94(1990)6435

8. E. C. Butler, A. P. Davis, J. Photochem. Photobiol., A: Chem. 70(1993)273

9. M. Fujihara, Y. Satoh, T. Osa, Bull. Chem. Soc. Japan, 55(1982)666

10. J-M. Herrmann, J. Disdier, P. Pichat, Chem. Phys. Letters 108(1984)618

11. M. Laniecki, M. Zalas, S. Manas, C. Richter, Proceed. 14 WhEc Conference, June 9-13, 2002, Montreal, CD-ROM edition

12. M. Laniecki, M. Zalas, IHP Programme Materials, Access Camapaigne 2002, Almeria 2003, Spain, pp.75-80 (2003)

13. M. Zalas, M. Laniecki, Proceed. 15 WHEC, June 27-July 3, 2004, Yokohama, CD — ROM edition