Nanostructured oxidic semiconductors such as TiO₂, ZrO₂, etc., were widely used as photoelectrochemical electrodes and colloidal photocatalysts for decomposition of toxic pollutants [1]. The importance of the hole-electron recombination ki­netics in determining the photocatalytic activity of colloidal semiconductors has been discussed [2,3]. To improve photochemical properties of the photocatalysts, doping of their surface with transition metal ions was usually applied [1-3]. An EPR study of TiO₂ aqueous colloids doped with Fe, V or Mo has been reported [4]. The authors observed inhibition of hole-electron recombination by Fe³⁺ and V⁴⁺ dopants.
One of the most important characteristics of the doped semiconductor particles for better understanding of their properties and action is knowledge about the structure and spatial distribution of the dopant centers with measurement of the distances < r > between them or their local concentration < C >. The EPR technique allows to obtain such data in the case of paramagnetic centers (PC) in any diamagnetic matrix [5]. In our previous work [6] we have presented some quantitative data for polycrystalline TiO₂ electrodes doped with V⁴⁺ and Cr³⁺ ions. The method for < C > and < r > determination from magnitude of the dipole-dipole interaction in cases of random, cubic or pair spatial distribution is well known and was proved experimentally in numerous works. The problem of determination the parameters analogous to < C > and < r > for systems in which PC were distributed on the surfaces of solid materials or in the interface is not solved yet. In this paper we perform our recent results obtained for the nanosized TiO₂ (Degussa P 25, Hombicat-100 and nano-TiO₂ prepared ourselves) and ZrO₂ particles contained different amount of V(4+) ions on the surface. For the correct calculation of the corresponding parameters < C_surf > and < r_surf > a special theoretical and computer analysis has been done.

Fig.1. EPR spectra of the P-25 particles after: 1 – 10 min, 2 – 6 days,3 –75 days of incubation in 0.15 cm3 of 0.65 M ascorbic acid in C2H5OH-H2O = 3:2 solution. [V4+] = 1.8 x1020 cm–3; T = 77 K.

Fig.1 presents typical EPR spectra of TiO₂ (Degussa P25) particles at different content of V⁴⁺ ions. It is evidently seen, that, as we have observed for the doped matrix of polycrystalline TiO₂ electrodes [6], there are two types of V⁴⁺ dopants: a) 'isolated' centers with well resolved anisotropic EPR spectrum and rather low concentration < C_surf > of them and b) 'aggregates' with a broad singlet line in the EPR spectrum usual for PCs with strong spin-exchange interaction between them and much shorter distances < r_surf > comparing with 'isolated' centers. Similar results were obtained for all TiO₂ and ZrO₂ systems, but appearance of the singlet line depended strongly on the vanadium content.
Figures 2 and 3 present changes in the low-field parts of the V⁴⁺ EPR spectra for the samples prepared using Degussa P 25 and Hombicat-100 powders at various V⁴⁺ contents (Fig. 2) or time after sample’s preparation (Fig. 3). The main regularities of the structural features of V⁴⁺ centers in the TiO₂ and ZrO₂ interface will be discussed in the paper.

Fig.2. Low-field lines of the EPR spectra of the Hombicat-100 particles with [V4+] content: 1 – 1.6 ×1019 cm–3, 2 – 4.0 ×1019 cm–3, 3 – 1.2 ×1020 cm-3 after 15 days of incubation in 0.15 cm3 of 0.75 M ascorbic acid; 4 –0.01M VO2+ in C2H5OH-H2O = 3:2 solution; T = 77 K.	Fig.3. Low-field lines of the EPR spectra of the Hombicat-100 ( 1-3 ) and P-25 ( 4,5 ) particles after: 1 – 10 min, 2, 4 – 25 h, 3, 5 –15 days incubation in 0.15 cm3 of 0.75 M ascorbic acid. [V4+] = 4.0 ×1019 cm–3 ( 1 – 3 ) and 2.0 ×1019 cm–3 ( 4 , 5 ); 6 –0.01M VO2+ in C2H5OH-H2O = 3:2 solution; T = 77 K.

One can see from these figures that there exist up to three types of surface V⁴⁺ centers with slightly different spectra and spin-Hamiltonian parameters. Analysis of these changes showed that in time V⁴⁺ ions from “aggregates” are transformed mainly to the “c” type of the “isolated” centers (Fig. 2, 3), though these “c” ions are still anchored to the surface, and not dissolved in a liquid phase. An attempt to attribute species “a-c” has been done, as well as to characterize quantitatively the peculiarities of spatial distribution of the V⁴⁺ ions on the oxidic surfaces.

Fig. 4. Fourier transformed dipole frequency spectra for 1-, 2- and 3-dimentional distributions of PCs.

For the correct calculation of corresponding parameters < C _surf> and < r_surf> for the studied systems, a special computer analysis has been carried out. It was shown (Fig. 4) that there existing the essential difference in the EPR-line shape for three cases of 3-, 2- and 1- dimensional systems. The obtained experimental results will be discussed in the paper in detail and will be compared with structural data known for the TiO₂ and ZrO₂ crystalline materials doped with V⁴⁺ inside the bulk matrix.
It was theoretically shown that there existing the essential difference in determination of the structural parameters for three-, two- and especially one-dimensional systems [7, 8]. The following equation performs the relation between T _{2 D} and n_D :

T_{2 D}^-1= 4.64 n_D^{3 / D}g²ħ ,

where T_2D is the transversal relaxation time of electron spin, g is an electron magnetogyric ratio, n_D is D -dimensional density of paramagnetic centers and D is the dimension of the PC’s spatial localization [8].

Thus, comparing the structural properties of metal-doped oxidic semiconductors one can conclude that:
1) a tendency to form the metal ion aggregates with high local concentration < C_surf> of the doping ions is typical for V⁴⁺ species on the surface of TiO₂ and ZrO₂ nanoparticles, similar to what has been observed for a matrix of V⁴⁺-doped polycrystalline TiO₂ systems studied in [6].
2) Relative concentrations of different V⁴⁺ species of the 'isolated' and 'aggregated' V⁴⁺ ions depend on vanadium content in the sample and the nature of the carrier.

Various types of V⁴⁺-doped TiO₂ nanoparticles, as well as of ZrO₂ ones, presented different peculiarities of the structure and spatial distribution of metal centres on their surface, that should be taken in consideration for explanation of photocatalytic and photoelectrochemical behaviour of such nanocrystalline systems.

Acknowledgements:
The authors are grateful to the RFBR (grant No.00-03-81168) and INTAS (Grant No. 00-0863) for financial support. We are thankful to Prof. A.Kh. Vorob’ev for providing us his PC EPR-programs and to Prof. M.Ya. Melnikov for valuable discussions.

References

[1]	Photoelectrochemistry, Photocatalysis and Photoreactors. M. Schiavello, Ed. Dordrecht: Reidel Publ. Co., 1985; Photochemical Conversion and Storage of Solar Energy. E. Pelizzetti, M. Schiavello, Eds. Dordrecht: Kluwer, 1991.
[2]	Kalyanasundaram K., Gratzel M., Pellhzetti E., Coord. Chem. Rev., 69, 57 (1985).
[3]	Martin S.T., Morrison C.L., Hoffmann M.R., J. Phys. Chem., 98, 13695 (1994).
[4]	Grätzel M., Howe R.F., J. Phys. Chem., 94, 2566 1990).
[5]	Abragam A., The principles of nuclear magnetism. Clarendon Press, Oxford, 1961.
[6]	Kokorin A.I., Arakelian V.M., Arutyunian V.M., Proc. of 13-th QUANTSOL, Kirchberg, Austria, 2001, p. 13-15. The same , Russian Chem. Bull., 2002, in press.
[7]	Dzeparov F.S., Lundin A.A., Khazanovich T.N., Russion JETP, 92, 342 (1987).
[8]	Atsarkin V.A., Vasneva G.A., Demidov V.V., et al., JETP Letters, 72, 530 (2000).