Detailed Photoluminescence Studies of Heat-Treated InP
- PDF / 332,246 Bytes
- 6 Pages / 420.48 x 639 pts Page_size
- 7 Downloads / 204 Views
DETAILED PHOTOLUMINESCENCE STUDIES OF HEAT-TREATED InP KLAUS PRESSEL, C. HILLER, G. BOHNERT, F. PRINZ, AND A. DORNEN 4. Physikal. Institut, Universitat Stuttgart, 7000 Stuttgart 80, Germany
ABSTRACT We present highly resolved photoluminescence studies on heat-treated nominally undoped InP, which was either unprotected or protected by Si0 2 or Si 3N4 caps during the annealing procedures. Annealing of InP above 350"C leads to six different sharp emissions in the wavelength range between 8790 and 8900 A, which are not observed at 4 K in Zn-doped or Fe-doped samples. Based on temperature-dependent photoluminescence studies, time-resolved measurements and preliminary magnetic field studies we ascribe these emissions to isoelectronic bound exciton transitions. It is also shown that the two emissions at 8883 A (11254 cm"-) (E) and 8889 A (11246 cm') (F) belong to one center. We observe that the lines not only depend on the heat treatment but also on some unintentionally incorporated or residual impurities of a low concentration level. Possible candidates are discussed.
Introduction and Experimental Details Implantation of samples, e.g. for controlled doping, introduces serious radiation damage. In principle these defects can be annealed but a problem arises, because already at -350 °C phosphorous begins to evaporate. Thus for annealing steps at higher temperatures protecting caps are necessary. The caps also can act as source for the formation of defects. Therefore the influence of the caps on thermally-induced defects is interesting to know and was recently the subject of a photoluminescence study by Kim et al. [1]. The authors studied in detail heat-treated InP and distinguished five sharp emissions at 8802 A, 8815 A, 8872 A, 8883 A, and 8891 A that appear in unprotected material [1]. The broader emissions, which they report, are also observed here, but will not be dicussed. As Kim et al. report by the use of Si 3N4 caps none of the lines are generated, whereas additional lines appear when Si0 2 caps are used. The lines reported in Ref. [1] are lying close to lines found after electron-irradiation and subsequent annealing [5,6]. A correlation of the various lines is difficult, because resolution is not good enough and discrepancies in the spectroscopic positions might be due to differences in the monochromator adjustment. But to distinguish the various defects a careful study is necessary which takes into account the processing of the caps as well as the sample prepreparation. To clarify the conditions for the growth of the lines is the topic of this paper. We think that the lines mentioned in Ref. [1-6] have the same origin. In all our measurements we excited the samples by a Kr-ion laser. The light is analyzed either by conventional monochromator technique or by a Fourier-transform-infrared spectrometer (BOMEM DA3.01). All the spectra have been detected with a Ge-detector. The annealing procedures were performed by rapid thermal annealing (RTA), in evacuated glass ampoule and in an oven with Ar flow. We studied Cz-g
Data Loading...
