Superlattices and Microstructures, Vol. 9, No. 4, 1991 517 OBSERVATION OF SURFACE AND BULK PLASMONS IN SEMICONDUCTOR SUPERLATTICES T. Dumclow, A.A. Hamilton, and T.J. Parker Department of Physics, Royal Holloway and Bedford New College, Egham Hill, Egham, Surrey TW20 0EX, UK. D.R. Tillcy, B. Samson, and S.R.P. Smith Departmcnt of Physics, University of Essex, Wivenhoe Park, Colchester, Essex CO4 3SQ, UK. R.B. Beall and J.J. Harris Philips Research Laboratories, Cross Oak Lane, Redhill, Surrey RH1 5HA, UK. (Received 13 August, 1990) We report far infrared measurements of oblique incidence power reflectivity and attenuated total reflection (ATR) on multiply Si 5--doped GaAs samples. The reflectivity spectrum in s-polarisation probes the in-plane component of the superlattice dielectric function, and is sensitive to the overall 3D electron density. The p-polarisation spectrum, however, also probes the out-of-plane component, resulting in an extra mode which is sensitive to the distribution of free electrons between wells and barriers. The p-polarisation ATR spectrum shows surface modes which are also sensitive to this distribution. The results are compared with a bulk slab model of the dielectric function. I. Introduction There have been s number of far infrared studies of the phonon properties of superlatticest-7. It is usual in such studies to take a long wavelength assumption and treat the superlattice as a single uniaxial medium when considering its dielectric function. Experiments have been performed which probe both the In-plaue component exx and the out- of-plane compouent ezz of the resultant tensor. £n the case of samples containing free carriers, and hence having a plasmon response as well as a phonon response, only investigations of the in-plane component have so far been reportede,9, other than in a preliminary report of the work discussed heret0. We here describe the use of far infrared reflectivity and attenuated total reflection (ATR) experiments in probing both components of multiply 8-doped GaAs specimens, and show how these experiments yield important information on the distribution of electrons in the superlattice. Providing a superlattice has layers of sufficient thickness its structure may be regarded as alternating slabs having bulk dielectric functions et and e2 and thicknesses d! and d2 respectively. In the long wavelength limit the principal components of the dielectric function of the resultant uniaxial medium are then given byt1,12: exx ffi (dte1+d2s2)/(dt+d2) (I) Ezz-1 ffi (dtel-t+d2s2-1)/(dt+d2) (2) where the z direction is normal to the superlattice layers. The bulk dielectric function of each layer type at frequency e can be considered to take the geueral form: E:u ffi Son[ (o2-eL.2) - raz, n2 e2-eT.2-tev. (02-iel"a ]. n=l.2 (3) where so. is the background dielectric constant, and eTu, eLn, and ep, are the transverse and longitudinal optical phonon and the plasma frequencies respectively for medium n. y. and r. are the phonon and plasmon damping terms. The plasma frequency is given by epn2 = Nme2/e0~onmB* (4) where N. is the 3D density of carriers of effective Inass mn ~ . 2. Experimental Each sample consisted of 5 pm of a G-doped GaAs superlattice grown on a GaAs substrate by molecular beam epitaxy. The superlattice portion contained I00 spike doped silicon layers, acting as electron donors, at 50 nm intervals. Growth temperatures and silicon doping levels are shown In Table 1. Table I. Superlsttice growth conditions Growth SUicon Doping Sample Temperature Level (cm-2) MV1338 650oc lx1013 MV1340 400oc 2x10 t3 MV1342 400°C ~x1053 MV1347 400oc 8x1013 0749-6036/91/040517 + 04 S02.00/0 © 1991 Academic Press Limited