Thermally stimulated luminescence from rare-earth-doped barium copper oxides

We report the first thermally stimulated luminescence (TSL) and emission-spectra measurements of x-ray-induced defects in Ho&5Ba&5Cu20„and GdBa2Cu30 in the temperature interval 80-300 K. The Gd-doped oxide exhibits a single TSL glow peak at 195 K with concomitant single-band emission with maximum at 500 nm. Two TSL glow peaks (160 and 172 K) are observed in Ho&5Ba&5Cu20„, whose spectral emission is characterized by complex, temperature-dependent maxima at 450 and 660 nm. Possible defect mechanisms are discussed.

i.e. , material stoichiometry, sintering temperature, etc. , is very important in establishing favorable superconducting parameters. The role of oxygen although not fully understood, is important in this process. It is suggested that the oxygen deficiency in YBa2Cu307 is partially accommodated by oxygen vacancies on the copper planes between the barium layers of the orthorhombic unit cell. Unlike pure metallic superconductors these new materials are composed of ceramic oxides with a band gap and defect sites (oxygen vacancies), and, consequently, may be highly susceptible to radiation damage. In this Brief Report we report the first results of thermally stimulated luminescence (TSL) studies of x-irradiated oxide superconductors.
Two-phase samples of Ho~5Ba~sCu20 (T, =46 K), Eu~sBa~5Cuqo, (T, =95 K), and single-phase GdBa2Cu30"(T, =95 K) were studied by TSL techniques in the temperature interval 80-300 K. These materials were prepared and characterized by methods given in Refs. 9 and 10. Radiation damage was provided by an x-ray machine operating at 40 kV p. p. and 25 mA producing an exposure rate at the sample site of 1.4 x 10 R min '. The TSL apparatus has been previously described. " Shown in Fig. 1 are the TSL glow curves for Ho& 5Ba& 5Cu20 and GdBa2Cu30 resulting from x-ray exposures at 80 K of 1.4x 10 R and 2.8 x 10 R, respectively. The Ho-doped sample exhibits peaks at 160 and 172 K, whereas the Gd-doped sample is characterized by a single peak at 195 K. Typical background photomultiplier tube (PMT) current is 1 X 10 A. For x-ray exposures as large as 4.2x10 R the Eu-doped sample yielded no TSL emission. Moreover, we found that the intensity of TSL in Hoand Gd-doped samples appeared to be correlated with the amount of time each sample had been (1.4 && 10 R) and GdBa2Cu30"(2. 8 x 10 R).
dramatic decreases in just three days, and after two weeks exhibited no intensity at all. It is possible that the Eudoped sample had undergone complete degradation before we attempted TSL measurements, and, consequently, we cannot conclude unambiguously that it does not luminesce. Nevertheless, such rapid loss of TSL sensitivity is very unusual in any luminescent material. We note, however, that these samples are opaque and that our measurements correspond to surface rather than bulk luminescence. Thus, our present TSL results probably reAect a greater sensitivity to sample properties than the usual bulk measurements. The Hoand Eu-doped samples contain both the green semiconducting and black superconducting phases. On the other hand, the Gd-doped sample is 90-95% pure su- volume. Apparently the TSL is not solely attributable to the semiconducting phase since both the Hoand Gddoped samples luminesce. However, a second sample of 99% pure superconducting GdBa2Cu30"did not exhibit luminescence. These results suggest that TSL is due to surface properties rather than any particular phase, and that these properties exhibit sample-to-sample variations that account for our observations.
Clues to the origin of TSL are provided by the concomitant emission spectra. The Gd-doped sample exhibits narrow emission [a full width at half maximum (FWHM) of 40 nm] that peaks at 500 nm as shown in Fig. 2 pears. Aparently the 660-nm emission peak is associated with only the 160 K TSL glow peak of Fig. 1. Likewise, the 172-K glow peak seems to be characterized by a single emission peak at 495 nm. However, the overlap of the two glow peaks (160 and 172 K) makes this assignment somewhat tentative.
If we make the analogy to simpler oxide systems, e.g. , A1203, MgO, and BeO, we would expect V-type (trapped holes) and F-type (trapped electrons) centers to exist in x-irradiated rare-earth-doped copper oxides. 320 nm due to recombination of thermally released holes with electrons at F centers (two electrons trapped at oxygen vacancy) and F+ centers (one electron trapped at oxygen vacancy), respectively. Therefore, one possibility for explaining the Ho-doped emission is to assign the 660and 450-nm peaks to F and F+ centers, respectively. The 160 and 172-K TSL glow peaks correspond to energetically difterent trapped holes that, upon thermal excitation, are released and recombine with F and/or F+ centers producing the observed emission. Initially there exists both F and F+ centers. As holes become thermally unstable they recombine with one of the F-center electrons producing an F+ center and 660-nm emission. Also, some of the holes may recombine with F+-center electrons yielding 450-nm emission as shown in Fig. 3 for (T) =135 K. Thus F centers are quickly annihilated leaving only F+ defects. As (T) increases only that emission characteristic of F+ centers is observed. The single 500-nrn peak in Gd-doped barium copper oxide may also be associated with F+ centers. Note, however, the absence of F-center emission.
Either F centers do not exist in the sample, or, alternatively, they exist but are converted to F+ centers by radiationless recombination.
A second plausible explanation for the observed TSL emission is that it results from the deexcitation of rareearth ions. X irradiation either removes or adds electrons to the ions, producing, for example, either Ho + or Ho + (or perhaps both) ions. Upon thermal excitation holes (electrons) are released that recombine with Ho + (Ho +) producing Ho + ions in excited states, the deexcitation yielding the observed emission. Difrerences in the widths of the spectral lines of the Hoand Gd-doped samples would be due to the diA'erent crystal-field environrnents to which each is subjected. With the paucity of available data, however, we cannot deduce the origin of the luminescence centers in these samples, and emphasize that the aforementioned possibilities are speculative.
An interesting feature of our data is the loss of luminescence sensitivity with time, especially when the samples are maintained in vacuum. As shown by earlier work, the oxygen-defect perovskites have a propensity to readily lose or gain oxygen. We suggest that this loss of oxygen contributes to the decreased luminescence sensitivity.
Thus TSL measurements may represent a very sensitive way to investigate the problem of oxygen stability in these materials. Present work is under way to examine the effects of various environments on the luminescence sensi-tivity.
It is noteworthy that oxygen loss should lead to an increase in oxygen vacancies which, upon irradiation, would yield a larger concentration of F-type centers. This would produce increased luminescence intensity rather than decreased, as observed. Alternatively, there may be other mechanisms operative which reduce the hole concentration and consequently reduce the luminescence signal. for Ho~5Ba~5Cu20"are E =0.265 eV, s =2. 14X106 sec ', l = 1 (160-K peak); and E =0.290 eV, s =2.05 x 10 sec ', 1 = 1.8 (176-K peak). The approximate second-order kinetics (l =1. 8) describing the 172-K peak indicate considerable retrapping of charged carriers prior to recombination. More elaborate details of the analysis will be presented in a subsequent publication.
In summary, we have reported the first TSL and emission spectra measurements of rare-earth-doped barium copper oxides. Possible explanations for the radiationinduced defects that produce the luminescence have been discussed. It is suggested that surface rather than bulk properties determined the TSL glow curves and that the instability of oxygen is important in this process. TSL may, in fact, be a sensitive technique by which the oxygen diA'usion can be studied. Glow curve parameters describing the TSL glow peaks have been given.