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.

Raman Scattering from a 7 nsec pulsed dye laser has been used to determine the onset of recrystallization following an 8 nsec dye laser excitation pulse in ion-implanted silicon. We find essentially complete recrystallization 59 nsec after the first excitation pulse and from Stokes-anti-Stokes ratios we find at 59 nsec a crystalline lattice temperature of 600 ± 200° C. Time-resolved transmission measurements at λ = 1.15 µm also demonstrate that no molten phase has occurred even though the usual reflectivity enhancement is observed.

A set of simultaneous equations for lattice temperature, carrier concentration and carrier temperature in Si is numerically solved for typical nanosecond and picosecond laser pulses. The calculated threshold energies required to reach Si melting temperature are consistent with measured thresholds to reach a flat-top reflectivity of ~ 70% for typical nanosecond pulses at 1.06 and .53 µm. In the picosecond regime, almost the entire pulse energy is at first stored in the carrier system, and a carrier temperature exceeding 30,000 K is achieved for a 3 J/cm2, 30 ps pulse at 1.06 µm. In this case, the surface carrier concentration reaches a high enough value to cause an enhanced reflectivity from the plasma lasting for at least 0.1 ns. The lattice temperature reaches 1410°C, while carriers relax their energy to the lattice after the pulse, and the energy stored in the carrier system would be enough to supply the heat of fusion at the silicon surface.

We have made a systematic investigation of the threshold energy density for recrystallization of ion-implanted silicon by Q-switched laser irradiation as function of thickness of the disordered layer, temperature during implantation, type and dose of implanted impurity, laser wavelength, and substrate orientation. Most results have been obtained with a Q-switched ruby laser. A linear dependence of the threshold on layer thickness (in the region of 60–300 nm) was found for arsneic-implanted silicon, but not for silicon-implanted silicon. For an amorphous layer thickness of 200 nm we found very little dependence of the threshold on type of dopant. In the case of the Nd:YAG laser, however, the lowest threshold was observed for column VI elements, the highest for column IV elements and intermediate and equal thresholds for the elements from column III and B. The influence of temperature during implantation was found to be small, but the threshold appeared to be different for (100)- and (111)- oriented substrates.

Amorphous, implanted, Si layers have been melted by pulsed electron irradiation. Implanted As has been used as a marker for determining melt duration. Systematic differences between As diffusion in initially amorphous or crystalline Si are interpreted in terms of different enthalpies of melting between amorphous (1220 J/g) and crystalline (1790 J/g) Si. The amorphous Si layers melt and crystallize at significantly lower electron energies than those required to melt crystalline Si, indicating that amorphous Si melts at 1170K compared to 1685 K for crystalline Si. We have used these thermodynamic parameters to successfully predict some of the phenomena associated with the laser induced melting and crystallization of amorphous Si.

Electron beams can be used for processing materials both through thermal effects and through nonthermal, or chemical, effects. Although most interest for the last 10 years or more has been on nonthermal effects, there has recently been revived interest in electron beam heating. Such beams can be pulsed flood beams of large area (8 cm diameter), large current density (500 A cm−2) and electron energies of 5–30 keV. At the other extreme are focused cw beams of a few μm diameter and currents of about 50 μA which can be used for selective heating. There are also systems which generate ribbon beams which are effective for well controlled annealing of large areas.

Pulsed laser annealing of silicon implanted by Group (III, V) dopants leads to the formation of supersaturated alloys by nonequilibrium processes occurring in the interfacial region during liquid phase epitaxial regrowth. The distribution coefficient from the melt (k') and the maximum dopant substitutional solubility (CSmax) are far greater than equilibrium values and both are functions of growth velocity. Substitutional solubility is limited by lattice strain and by constitutional supercooling at the interface during regrowth. Values for CSmax obtained at different growth velocities are compared with predictions of thermodynamic limits for solute trapping.

The segregation phenomena of In, Ga and Bi in Si have been investigated as a function of the liquid-solid interface velocity following laser irradiation. The crystallization velocity has been changed within the range 0.8–5 m/s by varying either the substrate temperature during irradiation or the laser pulse duration. The measured interfacial segregation coefficients depend critically on the velocity and on the crystal orientation of the solidifying plane.

In this paper we treat surface tension driven convection effects in pulsed laser formed melts. Mass transport is determined from an approximate solution of the Navier Stokes equation. It is shown that for small laser spot diameters the characteristic mixing times are on the order of 100's of ns. The dependence of the convection mechanism on material and laser parameters is discussed and extended to thin metal films on Si. Experimental results substantiating the theoretical considerations are presented.

Two techniques have been used to measure the velocity of the amorphous-crystalline boundary during scanned laser crystallization of amorphous Ge films on fused-silica substrates. Values in the vicinity of 200 cm sec-1 have been measured by both methods. The results obtained by the first technique, an optical transmission method, confirm our theoretical model for the periodic motion of the boundary. The measurements made by the second technique, which is based on an examination of the structural features obtained at laser scanning rates up to about 600 cm sec-1 , show the boundary velocity to be rather insensitive to film thickness and background temperature. Controlled crystallization is expected to require stability of the laser beam power.

Laser processing of amorphous thin films of amorphous Ge often results in an explosive or self-sustaining crystallization reaction. The reaction is sustained by the heat liberated during crystallization. In a theoretical analysis of the process that was presented at this symposium last year, Gilmer and Leamy postulated the existence of a thin layer of liquid at the propagating interface. The liquid layer forms at temperatures above Ta, the melting point of amorphous Ge, and is predicted to be ~ 0.02 – 0.1 of the film thickness in width. We have obtained experimental confirmation of the presence of this liquid layer.

A single picosecond pulse laser annealing of ion-implanted Si is reviewed as ultra-short pulse laser annealing, comparing them with nanosecond pulse and picosecond-pulse train annealing. In order to clarify the physical mechanism of pulsed laser annealing, the dynamic behavior of the amorphous to crystalline transition has been investigated by means of time-dependent optical reflectivity measurement at 0.63 µm (cw) and 1.06 µm (30-ps pulse itself) under the irradiation of the annealing beam of a single 30-ps laser pulse at 1.06 µm. A tentative model is proposed for explaining the results and further problems which remain to be resolved are discussed.

The low energy electron diffraction (LEED) patterns obtained from clean (111) oriented Si, Ge and GaAs single crystals subsequent to their irradiation with the output of a pulsed ruby laser in an ultra-high vacuum (UHV) environment suggest that metastable (1×1) surface structures are produced in the regrowth process. Conventional LEED analyses of the Si and Ge surfaces suggest that they terminate in registry with the bulk but that the two outermost interlayer spacings differ from those of the bulk. For the case of Si these changes are a contraction of 25.5 ± 2.5% and an expansion of 3.2 ± 1.5% between the first and second and second and third layers respectively.

Amorphous silicon has been produced on a single crystalline silicon surface by intense UV laser radiation at 266 nm followed by rapid quenching. In addition, formation of oxide of several tens of nanometers has been observed when irradiation takes place in air or in O2 ambient. Various experimental techniques including Transmission Electron Microscopy (TEM), sputtered Auger Electron Spectroscopy (AES), and differential Fourier-transform IR spectroscopy (FT-IR) have been employed to study adsorption of oxygen during rapid melting and resolidification process. The present results suggest a new processing technique for forming thin oxide film, namely, “Laser-induced oxidation.”

We report the use of time-resolved optical reflectivity to directly monitor the dynamics of cw laser-induced solid phase epitaxy (SPE) of thin films. This in situ measurement technique utilizes optical interference effects between light reflected from the surface of a sample and from an advancing interface to provide continuous temporal and spatial resolution of crystal growth processes. SPE growth rates of ionimplanted films which are five orders of magnitude faster than previously observed can be induced and accurately measured with the laser method. Arsenic enhances the SPE rate, and spatially resolved measurements show that the growth rate for arsenic implanted films varies in accordance with the ionimplantation profile. Results are reported for silicon selfimplanted samples with and without subsequent arsenic ion implantation, and for silicon samples directly implanted with arsenic.

The surface reflectivity of ion implanted silicon is a function of the entire past history of the material, including the nature of the implanted species and the implant dose. It is influenced by interference effects, arising from discontinuities in index of refraction at the surface and at the boundary between damaged and undamaged material. The reflectivity may be either higher or lower than that of the unimplanted silicon. As the crystalline silicon regrows from below, the reflectivity changes because of variable constructive or destructive interference. This paper describes monitoring of the surface reflectivity during continuous laser annealing. The beam from a continuous argon ion laser is scanned in a raster pattern over the silicon surface. The surface reflectivity for a HeNe laser beam is monitored as a function of time during crystalline regrowth. The reflectivity contains an oscillatory component which arises because of changes in interference due to the decreasing thickness of the noncrystalline layer. The oscillatory behavior produces a signature characteristic of the annealing. The reflectivity monitoring technique is useful for investigating the influence of parameters such as the spatial profile of the laser beam and the implantation dose on the annealing characteristics. The results are correlated with measurements of the depth profile of the implanted ions, as revealed by anodic oxidation and stripping.

We present a thermal analysis which treats the problem of melting as it occurs during cw laser heating. Analytical expressions for sample temperature distributions are derived, and calculated results are compared to experimental measurements.

A standard optical disappearing hot wire micropyrometer was modified with a long working distance magnifier to enlarge the laser heated spot so that the maximum temperature of the center of the spot could be observed and measured with some certainty. In situ temperature measurements were made on As ion implanted Si <100> chips using an Argon Ion laser with the beam slightly defocused to a 50–100μm diameter spot on the Si surface. The data curves show a linear function of measured temperature with laser power up until the melting point is reached. By using the Si m.p. as a calibration point, the temperature can be reproducibly measured within a few tens of degrees. Comparisons of actual re-growth rates with those calculated from the well known solid state Si regrowth expression as a function of temperature yields good agreement between measured and calculated temperature.

EBIC has been used to investigate the beam-induced damage in As+ -implanted Si subjected to scanned laser or electronbeam annealing. Comparison is made between these annealing techniques; in general, electron-beam annealing is found to give superior results. In addition, several different experiments combining these techniques are described. Samples were pre-annealed prior to laser annealing, using either thermal or electron-beam annealing. Other samples were given a thermal post-anneal after laser-annealing. The effect of these treatments on defect formation is discussed.

Electronic states of extremely heavily doped n-type Si obtained by high dose ion implantation and laser annealing are investigated by measuring the infrared optical properties. Free carrier effective mass (m*) and carrier relaxation time (τ) are obtained as a function of carrier concentration (1019−5×1021 cm−3). Values of m* and τ increase and decrease, respectively, with the increase of carrier concentration. These results are discussed in relation to the occupation of electrons in a new valley of the conduction band.