Stress Effects on Raman Measurements of Pulsed Laser Annealed Silicon
- PDF / 385,944 Bytes
- 6 Pages / 420.48 x 639 pts Page_size
- 73 Downloads / 201 Views
STRESS EFFECTS ON RAMAN MEASUREMENTS OF PULSED LASER ANNEALED SILICON
G. E. JELLISON, JR. AND R. F. WOOD Solid State Division, Oak Ridge National Laboratory Oak Ridge, Tennessee 37831 ABSTRACT It has recently been shown that the front surface region of the silicon lattice is severely strained during pulsed laser irradiation. This uniaxial strain reduces the symmetry of the front surface region, resulting in additional shifts and splittings of the phonon frequency and changes in the Raman scattering tensor. It is shown that, for the case of pulsed laser irradiation, the phonon frequency is increased, and the 3-fold degenerate optical phonon is split into a singlet and a doublet. The changes in the Raman scattering tensor make it non-symmetric, and generally invalidate the technique used by Compaan et al. to determine the cross section experimentally. The complications introduced by the presence of stress during pulsed laser annealing, coupled with the temperature dependence of the optical and Raman tensors, make a simple interpretation of the Stokes to anti-Stokes ratio in terms of lattice temperature extremely unreliable.
INTRODUCTION It has been accepted for some time by most workers that the results of pulsed laser irradiation of silicon in the nsec regime can be explained primarily in classical heat flow terms, and that annealing takes place by a melting and rapid epitaxial regrowth mechanism [1,2]. This model implies that the front surface region becomes molten for a short time (-100 ns) for laser pulses of sufficient power. However, Compaan et al. [see Ref. 3 and references therein] and von der Linde et al. [4] have performed a series of time-resolved Raman measurements utilizing the Stokes to anti-Stokes ratio to measure lattice temperature during and immediately after the laser pulse, where they find that the lattice temperatures obtained from this method (300-4509C) are far below those which would be expected if the front surface region were to become molten. Compaan et al. [3] also measured a 1 0 phonon frequency shift of 13 cm- , which corresponds to -470 C, if only temperature effects on the phonon frequency are included. Upon close examination of the pulsed Raman method for the measurement of lattice temperature, it has been found that this technique has many difficulties. For example, (1) There are pulse-to-pulse fluctuations in the probe and annealing beam energy densities and in the probe pulse "jitter" [5]. (2) There are fluctuations of the transverse energy density of both the annealing and probe beams over x-y distances as small as I [m. (3) There are large temperature variations in parameters such as the optical absorption coefficient and Raman matrix element [6]. These effects can drastically affect the temperature determined from the Stokes to anti-Stokes ratio [5-7], but there are other effects which also can affect both the phonon frequency and the Raman intensity, such as (4) stress effects, and (5) the effect of the excess electron-hole pairs present during and immediately after the pulse
Data Loading...
