J. MATER. CHEM., 1991 1(3), 357-359 Characterization of Epitaxially Grown ZnS :Mn Films on a GaAs(lO0) Substrate prepared by the Hot-wall Epitaxy Technique Takato Nakamura,*a Hitoshi Muramatsu,a Yoji Takeuchi,b Hiroshi Fujiyasub and Yoichiro Nakanishi" a Department of Applied Chemistry and Material Technology, Faculty of Engineering, Shizuoka University, Hamamatsu 432, Japan Department of Electronics, Faculty of Engineering, Shizuoka University, Hamamatsu 432, Japan Research Institute of Electronics, Shizuoka University, Hamamatsu 432, Japan Epitaxially grown ZnS: Mn films on a GaAs(l00) substrate prepared by a hot-wall epitaxy technique have been studied using electron paramagnetic resonance (EPR) spectroscopy. The lowest-energy transition assigned to M,= -512 splits into either a triplet or a quintet depending on the thickness of the film when the magnetic field is applied normal to it.Computer simulation confirmed that the EPR parameters obtained correspond to those for a single crystal of cubic structure, although the apparent lineshape is different. There is no significant difference between the hyperfine coupling constants and splitting parameters. The apparent difference in the EPR lineshape is ascribed to the change in linewidths of the fine structure involved. In particular, linewidths for the transitions Ms= +_5/2-+_3/2 decrease markedly with increasing thickness of the ZnS: Mn film on the GaAs(100) substrate, confirming that uniaxial deformation of the crystal field surrounding manganese@) is derived from the lattice mismatch between them.Keywords: Thin film; Hot-wall epitaxy; Electron paramagnetic resonance spectroscopy Electron paramagnetic resonance (EPR) spectroscopy has been applied to the study of paramagnetic ions in crystals because the resonance observed reflects the surroundings of the ions. Manganese@) ions doped in single crystals and powders of the cubic and hexagonal structures of zinc sulphide have already been examined in order to elucidate the sym- metry and strength of the crystal field surrounding the manga- nese@) in the host.'-4 Recently, Kreissl et ~l.'-~have not only re-investigated the EPR spectra of ZnS :Mn powders with different crystal structures, but also examined those of thin films deposited on glass substrates, prepared by evapor- ation and chemical vapour deposition.This is of interest since thin films of ZnS :Mn (M =metal ion) are electroluminescent materials. According to Mitsui et ~l.,~there has been lattice distortion in ZnS films on GaAs and GaP substrates due to the lattice mismatch between the epilayer and the substrate. In ZnS :Mn films on a GaAs substrate this should cause distortion of the crystal field surrounding the manganese@), so that anomalous EPR resonance should be observed. Therefore, we consider EPR spectra of the ZnS :Mn films with a variety of thicknesses grown on a GaAs(l00) substrate prepared by the hot-wall epitaxy (HWE) technique. Experimental Thin films of ZnS :Mn were grown on a GaAs( 100) substrate under a background pressure of lov6 Torrt by means of the HWE technique.Precise control of the temperatures of the ZnS and Mn-metal sources, the GaAs substrate and the wall was required. In particular, the temperature of the Mn-metal source, which determined the amount of manganese@) doped in the ZnS host, was of importance because too much doping causes a broad signal due to spin exchange, which gives little information about the surroundings. Preliminary experiments showed that the optimum temperature for the manganese metal was 873 K. The ZnS source was evaporated at 973- 993 K, while the substrate was kept at 523 K during the ~ ~~ t 1 Torr ~133.322Pa. deposition. The deposition rate of the film was in the range 1.3-2.7 A s-I.EPR spectra were measured at 10 and 293 K using JEOL JES-RE3X and JES-FE- 1 XG spectrometers, respectively, the former being equipped with a variable-temperature apparatus, ES-LTRSX. Films on the GaAs substrate were attached to a quartz rod with glue, and the rod was set in a cylindrical cavity operating in the TE,,, mode. Manganese doped in MgO and an ECHO NMR field meter were used for the calibration of the magnetic field. Simulation of the EPR spectra was carried out with the aid of an NEC PC-9801 personal computer with Turbo-Pascal 5.0 language. Results and Discussion For manganese@) in single-crystal cubic ZnS the resonance fields of respective hyperfine lines rotated about the [Oll] axis have been worked out theoretically as' H(Ms,MI)=H,-AM1-A2/2H,[35/4 -MI2+MI(2Ms-I)] -(a/64)(35cos46-30cos26+3-5sin46) x (56Ms3-84Ms2-134Ms+81) (1) in which H,=hv/gpB (his Planck's constant, v is the frequency, g is the g-value and pB the Bohr magneton), Ms and MI are the electron and nuclear spin quantum numbers, respectively, A is the hyperfine coupling constant, a is the splitting param- eter and 6 is the angle between the [loo] axis and the applied magnetic field.Fig. 1 shows the EPR spectrum of a film of ZnS :Mn with thickness 0.98 lm on a GaAs(l00) substrate at 6=0, in which the magnetic field is applied normal to the film. Also shown is a calculated spectrum; the EPR parameters are those for a single crystal of cubic ZnS :Mn obtained by Matarrese et al.,' the linewidth is 0.5 mT, and the lineshape is described in terms of a Lorentz function, since the fine-structure lines observed in this study fit a Lorentzian better than a Gaussian.It is immediately noticed that for the observed spectrum the hyperfine lines split into a poorly resolved quintet. Even in I111l lllll Ill I1 I I1 I 1 I) I I1 I,IrI,I,I,I 310 320 330 340 350 360 HImT Fig. 1 Comparison of the EPR spectrum of a ZnS :Mn film, thickness 0.98 lm, on a GaAs(l00) substrate (top) with a calculated spectrum using the EPR parameters of A=-64.1 ~10-~cm-' and a= 1.30 x cm-' for a single crystal with cubic structure (bottom), in which 0 is 0" the EPR spectrum measured at 10K they do not give well resolved fine structure because of the g-strain.It has been reported that there is lattice distortion in the ZnS films grown on GaAs and GaP substrates by metal- organic chemical vapour deposition because of the lattice mismatch between the epilayer and substrate, and that the misfit strain is mostly eliminated by 1 pm.8 If this is true, the surroundings of the manganese@) atoms in the ZnS :Mn films will depend on the thickness of the film. Therefore ZnS layers of 1 pm without manganese were grown prior to the deposition J. MATER. CHEM., 1991, VOL. 1 of the ZnS: Mn, and the EPR spectrum was compared with those of the ZnS:Mn films deposited directly on the GaAs substrate. The EPR spectrum of the ZnS :Mn/ZnS film contains fine structure and a quintet at the resonance assigned to the lowest-energy transition MI= -5/2 was observed, although the resolution corresponding to the transition Ms= -90 -80 70 -60 -so --40 -30 20 --10 n--3 -2 -1 0 1 2 relative HImT Fig.3 Angular dependence of the resonance field of fine-structure lines of the lowest-energy transition MI= -5/2 for a ZnS:Mn film with thickness 1.4 pm on a GaAs(100) substrate (open circle); A, B, C, D and E denote the calculated angular dependence of the resonance field of the transitions M, = +1/2* +3/2, -5/2*-3/2, -1/2* +1/2,+3/2* + 5/2 and -3/2--1/2, respectively, based on eqn. (l),using the EPR parameters of a single cubic crystal 100 mT-300 310 320 330 340 350 360 370 Fig. 2 EPR spectra of a ZnS :Mn film deposited on a ZnS/GaAs( 100) HImT substrate (top) and deposited directly on a GaAs(100) substrate (bottom).In the former the thicknesses of the ZnS:Mn and ZnS Fig. 4 EPR spectral simulation of ZnS :Mn films with thicknesses of layers are 0.6 and 1.0 pm, respectively, and in the latter that of the 0.7 (A, B) and 1.4 pm (C, D) on a GaAs(100) substrate, in which A, ZnS :Mn film is 1.4 pm and C are observed spectra and B and D are calculated J. MATER. CHEM., 1991, VOL. 1 11;, field strength, elucidated theoretically by Watanabe,g it is presumed that the crystal-field strength itself, surrounding manganese@) in the ZnS: Mn film, does not change signifi- cantly with increasing thickness. It is found that some of the linewidths of the fine structure vary with the thickness of the film.In Fig. 5, the linewidths for the transitions Ms= f5/20 & 3/2, f3/20 f1/2 and +1/2--1 /2 are plotted against the thickness of the ZnS :Mn film. A marked decrease in the linewidth as the film thickness increases up to 1.5 pm is observed for the transition between Ms = & 5/2 and & 3/2. Similar phenomena have been observed in the study of the EPR spectra of single-crystal cubic MgO :Mn under external uniaxial stress." The distortion lifts the degeneracy of the spin manifold. This splitting is not resolved, and manifests OO*itself as a linewidth increase in the anisotropic region relative to the isotropic one. The above are consistent with the phenomena for manga- nese@) under uniaxial stress." It is therefore concluded that the anomalous resonance showing three-line fine structure is due to the distortion of the crystal field surrounding manga- Q,5 1,o 1,5 nese@) in the ZnS film near the GaAs substrate. The distortion thickness of film/pm Fig.5 Variation of the linewidths of fine structure with the thickness of the ZnS: Mn film on a GaAs(100) substrate; 0, A and 0 denote the transitions M, = k5/2w f3/2, f3/2w f1/2 and +1/2w-1/2, respectively f512-& 3/2 is poor (Fig. 2). Similar fine structure appears in the EPR spectra of the ZnS: Mn films with a thickness of >1 pm (Fig. 2). Also, Fig. 3 shows that the angular dependence follows eqn. (1). Introducing the linewidth parameters for the transitions M, = f5/2-f3/2, f3/20 f1/2 and +1/20-1/2, calculated spectra were fitted with observed spectra (Fig.4). Obtained values of the hyperfine coupling constant and splitting param- eters are -63.8f0.2+10-4 and 1.30fO.l +lop4 cm-', respectively, independent of the thickness of film. A compari-son of the obtained parameters with those of the single cubic crystal determined by Matarrese et al.' indicates that there is no significant difference between them. Taking account of the parabolic dependence of the splitting parameter on crystal- is almost relaxed at a film thickness of 1.5 pm. We thank Mr. K. Fujii and Mr. K. Mituda, R&D Department, JEOL Ltd., for their help in the EPR measurement at 10 K. References 1 L. M. Matarrese and C. Kikuchi, J. Phys. Chem. Solids, 1956, 1, 117. 2 S. P. Keller, I. L. Gelles and W. V. Smith, Phys. Rev., 1958, 110, 850. 3 H. D. Hershberger and H. N. Leifer, Phys. Rev., 1952, 88, 714. 4 T. Buch, B. Cherjaud, B. Lambert and P. Kovacs, Phys. Rev. B, 1973, 7, 184. 5 J. Kreissl and W. Gehlhoff, Phys. Status Solidi A, 1984, 81, 701. 6 J. Kreissl and D. Backs, Phys. Status Solidi A, 1987, 99,K117. 7 J. Kreissl, Phys. Status Solidi A, 1986, 83, 191. 8 I. Mitsui, H. Mitsuhashi and H. Kukimoto, Jpn. J. Appl. Phys., 1988, 27, L15. 9 H. Watanabe, Prog. Theor. Phys., 1957, 18,405. 10 E. R. Feher, Phys. Rev. A, 1964, 136,145. Paper 0/04860G; Received 29th October, 1990