Defects in inorganic photorefractive materials and their investigations

18 B. Briat et al.
[99, 100]. EPR/optical investigations on this system [21, 101] showed that a
main part among the occurring photo-induced charge transfers involves the
three defects Rh 3+ ,Rh 4+ and Rh 5+ , fulfilling the ’3-valence model’ (Fig. 1)
[102]. On this basis and relying on experimentally determined values of the
effective trap density N eff , Huot et al. [103], and Corner et al. [104] analyzed
the charge transfer properties of the system quantitatively. However, because
the available experimental information was not sufficient to determine all relevant
parameters, a simplified theoretical basis was employed. Later, using
the EPR/optical method, a complete solution of the problem was possible
[105, 106].
The procedure starts with investigating the wavelength dependence of the
photochromic coloration of a BT:Rh crystal [98], induced by a series of rising
pump light energies. The result is plotted over the field of the pump light,
E pump , and probe light, E probe , energies (Fig. 9c). Here and in the following
only a rather brief sketch of the method is given. For further details see Refs.
[105, 106]. The main features in Fig. 9c are a strong light induced transparency
at E probe =1.9 eV and pronounced absorption increases at 1.6 eV and 3.0 eV.
Simultaneous measurements of the EPR of Rh 4+ , the only EPR-active Rh
charge state in ’as grown’ BT:Rh - observable at T ≤ 20 K, show that the
intensity of this EPR signal has an identical dependence on E pump as that of
the transparency along E probe =1.9 eV. This assigns the band at 1.9 eV to
Rh 4+ . Further EPR studies [107] indicate that the Rh 4+ intensity is decreased
by the transfer of a valence band electron to Rh 4+ . In this way the EPR-silent
charge state Rh 3+ is created. The band at 3.0 eV is attributed to Rh 3+ ,lying
higher than Rh 4+ because of its lower charge. The hole created in the valence
band by the electron transfer to Rh 4+ is expected to be captured by another
Rh 4+ ,causingRh 5+ , which is also EPR-silent. Because of its higher charge,
less energy is needed to excite a valence band electron to Rh 5+ . Therefore the
other strong feature in Fig. 9c, at 1.6 eV, is assigned to Rh 5+ . In a similar way
the further structures in Fig. 9c are attributed to various charge states of Fe;
this element is usually present in BT as an unintended background impurity.
These assignments fulfill systematic topological constraints typical for plots
of the type of Fig. 9c [108].
Summarizing this part: By combined EPR/optical absorption studies,
based on the photochromic behavior of BT, the optical absorption bands indicated
by vertical dashed lines in Fig. 9c, have been identified. Among these
only the defects Rh 4+ ,Fe 5+ and Fe 3+ are EPR-active. The position of the
Fe 3+ band, covered by the fundamental absorption, has been inferred from
its peak energy in KTaO 3 , Fig. 10. The other optical bands, belonging to the
EPR-silent defects Rh 3+ ,Rh 5+ and Fe 4+ , are assigned by consistency arguments
of the type as forwarded for Rh 3+ and Rh 5+ . This identification of the
optical absorption bands allows that they can be used as ’fingerprints’ of the
corresponding defects. They can be employed likewise at room temperature;
EPR-measurements, on the other hand, usually require low temperatures. As
shown, also absorption bands corresponding to EPR-silent defects could be