Fundamental Mechanisms in Laser and Electron Beam Processing of Semiconductors
- PDF / 707,634 Bytes
- 14 Pages / 418.68 x 637.2 pts Page_size
- 16 Downloads / 226 Views
FUNDAMENTAL MECHANISMS IN LASER AND ELECTRON BEAM PROCESSING OF SEMICONDUCTORS
WALTER L. BROWN Bell Laboratories,
Murray Hill,
N. J.
07974,
USA
ABSTRACT A few of the new developments in understanding transient laser and electron beam processes in semiconductors and a few of the still unresolved questions are summarized. I.
THE QUESTION OF ENERGY TRANSFER
An enormous number of experiments have been reported in the last 3 years on "laser annealing", or "directed energy processing" of semiconductors. The proceedings of the 1978 and 1979 Materials Research Society annual meetings contain reports or summaries of much of this work. [1] [21 Many of the experiments have involved pulsed beams of light, electrons or even low mass ions which couple their energy into electronic excitation and ionization of the semiconductor. This energy is ultimately converted to random atomic motion: heat. Only in the case of ion beams is even a fraction of the energy put directly into motion of the atoms in a material, and for the proton experiments so far reported [3], this fraction is negligibly small. Beams of heavy ions would shift this balance, and indeed such experiments are underway 141, In addition, experiments are planned using light which couples energy directly into vibrational states of the lattice system [5], but no results are yet available. At present, therefore, the time and spatial scales for the transfer of energy from the electronic to the atomic system are subjects of continuing interest and discussion. Historically, (if 3 years can be regarded as history), thermal melting has been identified as a simple unifying basis [6] by which a large body of experimental results from pulsed excitation could be understood and meaningful predictions made. These have steadily evolved and now include pulsed energy and wavelength dependence of amorphous silicon regrowth as single or polycrystal [7]; the observation of melt-type surface topography [8]; the redistribution of impurities in a '11 p layer near the excited surface [9], depending on pulse energy, pulse length, wavelength, and specific impurity; the dissolution of precipitates [10]; the dissolution and growth of dislocations [11]; the cleaning and ordering of surfaces in vacuum [12]; and the formation of amorphous silicon from single crystal silicon at very high solidification velocities [13]. Optical reflectivity has provided important time resolved information. In particular, the data of Auston et al. [14] using a nanosecond pulsed Nd:glass laser at 1.06 p and frequency doubled at .53 p indicate the existence of a molten layer. The layer develops at the semiconductor surface and persists for a time that is consistent with the melting and refreezing of a region whose thickness depends on the energy, pulse length and wavelength of the exciting pulse. Fig. 1 illustrates a typical result. It has been suggested that the material changes observed are not thermal in origin but result from atomic rearrangements allowed by a persistent high density plasma formed by an exciting
Data Loading...
