Solid State Communications, Vol. 68, No. 2, pp. 211-214, 1988. 0038-1098/88 $3.00 + .00 ~$~4 Printed in Great Britain. Pergamon Press plc RAMAN SCATTERING FROM InGaAs/GaAs STRAINED-LAYER SUPERLATTICES F.Iikawa, F.Cerdeira, C.Vazquez-Lopez* and P.Motisuke Instituto de F~sica - UNICAMP, 13081 - Campinas - SP Brasil M.A.Sacilotti CPqD - Telebras, 13100 - Campinas - SP Brasil A.P.Roth Microstructural Sciences - Division of Physics - NRC Ottawa - Canada and R.A.Masut Ecole Polytechnique - Genie Physique, Montreal - Canada (Received May 24, 1988 by Cylon E.T. Goncalves da Silva) We report the results of room temperature Raman scattering experiments on a series of In0.12Ga0.88As/GaAs strained-layer superlattices. From these measurements the amount of strain present in each type of layer is determined quantitatively. This information is used to discuss the question of critical thicknesses for this type of superlattice. The successful growth of InxGa1_xAs/GaAs strained layer superlattices (SLSL) by MOCVD has been recently reported 1,2,3. In these superlattices the lattice mismatch between layers is accomodated by elastic strain. When grown on GaAs substrates, alloy (B) layers are under biaxial compression so that their in-plane lattice constant is equal to that of bulk GaAs (A type layer and substrate). For single alloy layers this is accomplished as long as the layer thickness is smaller than a given composition dependent critical thickness (dB < d~(x)). Theoretical models differ considerably in their predictions of this critical thickness 4,5. Experimental determination of d~ for InxGal_xAs/GaAs single quantum wells and superlattices also give different values, which depend on the technique6u~ed in measuring this quantity -u. Many of these discrepancies are reconciled when the theoretical model takes into account such factors as growth temperature and the motion and multiplication of dislocations 9. Also, it should be noted that not exceeding the critical thickness of each alloy layer does not guarantee that the superlattice, as a whole, is commensurate with the substrate I0. In fact, for the samples used in our experiments, X-ray and photoluminescence results indicate On leave from Instituto de Ciencias Universidad Autonoma de Puebla Apdo J-48 Puebla 72570 Pue Mexico. 211 that stress relaxation occur in some of the thicker SLSL's I-3. We report here the results of room temperature Raman scattering experiments performed on a series of InxGal_xAs/GaAs strained-layer superlattices. These experiments provide quantitative measurements of the amount of strain present in each type of layer and allow us to bring into focus the question of critical thickness. In our experiments we used samples grown by MOCVD on (001) GaAs substrates. The growth procedure and analysis of these samples with double crystal X-ray diffraction, Auger profiling and photoluminescence spectroscopy are described in Refs. i-3. Our Raman measurements were performed on five types of samples, whose characteristic are listed in table I. In the strained layer superlattices (SLSL), alternate layers of thicknessess dA of GaAs and d B of InxGal_xAS (x=0.12) were grown to obtain a twenty period superlattice. Also, a thick (dB = 2~m) and, therefore, strain relaxed alloy layer was grown in order to compare the results obtained with this sample to those from the strained alloy layers in the SLSL. These measurements were also performed on a piece of substrate material (bulk GaAs). Raman spectra were taken at room temperature, in the backscattering configuration, employing 100mWatts of the 5145A line of an Argon-ion laser as exciting radiation. This radiation was focused on the sample through a cylindrical lens in order to avoid heating. The choice of geometry implies that only Raman lines corresponding to