An exact model of the IR reflectance and transmission spectra of inhomogeneously doped silicon has been developed. By comparing theoretically and experimentally generated spectra in the range 1.2 to 25 μm, information about the doping profile may be deduced. The technique has been applied to laser and electron beam annealed silicon (as well as furnace annealed) and comparisons made with theoretical profiles and SIMS analysis in the latter case. No values consistent with other published data have been found for the optical properties of silicon at high doping levels (>1020 cm−3).

A simple method of improving laser beam homogenity has been developed. A cylindrical quartz pipe, acting as a lightguide, has been applied for dispersion of the components of the laser beam (microbeams), thus providing more uniform illumination of the sample's surface. The effects of such a homogenizer have been studied by means of X-ray transmission topography [4–5] and 2-MeV 4 He-ion channeling. The samples were ≤111≥ Si wafers implanted with 100 keV As and Bi ions to a dose of 1016 /cm2 . It has been observed that the shallow residual damage layer which exists in the directly irradiated samples is not present in the samples irradiatedthrough the homogenizer.

Low-energy electron beam pulses of 0.2–0.8 s duration have been used to anneal silicon wafers implanted with 1.4×1015 P+/cm2 at 30 and 100 keV. Sheet resistivity, carrier concentration profiles and phosphorus atom distributions have been measured. It has been found that the impurity is completely activated without a noticeable diffusion at the energy density of 30 J/cm2 and higher. The annealing process appears to take place via a solid phase diffusion and annihilation of radiation defects.

DLTS has been used to investigate the nature of CW laser-induced defects in ion-implanted Si. A dominant hole trap (∼Ev + 0.45 eV), whose concentration depends on laser power, was observed immediately after sample preparation. This defect is not stable at room temperature; instead, it decays as a function of time, transmuting to a shallow level at Ev + 0.10 eV. The recovery of the Ev + 0.45 eV level can be stimulated by low temperature thermal annealing or by minority carrier injection. By comparing these defects in laser-annealed samples with defects produced by furnace annealing followed by rapid cooling, and with other published results, the laser-induced defects have beenidentified as interstitial Fe and Fe-B pairs. Experiments suggest that elevated substrate temperature during laser annealing may inhibit the formation of these deep hole traps.

Annealing of divacancies which were produced by 11B ion implantation was investigated under furnace, pulsed e-beam and pulsed ruby laser exposures. Despite orders of magnitude shorter exposure times for annealing and the concomitant expected high levels of electronic excitation and layer stress, we find that the thermal annealing mechanism observed for furnace annealing is an adequate description for divacancy annealing under e-beam exposure. The observed need for melting to remove divacancies by Q-switched laser annealing is also consistent with predictions based upon extrapolations from furnace annealing.

Deep Level Transient Spectroscopy has been used to investigate cw laser induced defects in virgin silicon. Two main regimes have been found. In the solid phase regime, two well defined deep levels at Ec−0.19 eV and Ec−0.45 eV are observed. This point defect introduction is proposed to be involved in the degradation of ion-implanted cw laser annealed junctions. The mechanism leading to point defects generation is likely to involve trapping of in–diffused vacancies, quenched–in from the high temperature state. In the slip lines or melt regimes, additionnal deep levels are detected, which are ascribed to dislocations.

To determine the nature and effect of defects which occur in the growth of crystalline semiconductor films on amorphous substrates, one must carry out microstructural and electrical studies on the same material regions. In previous publications we have demonstrated that by controlling optical absorption in and around delineated silicon areas, the nucleation and growth processes can be controlled. In this way we have been able to produce single crystal islands >20 μm on a side. Using a simple, novel parylene lift –off technique we can now routinely study the material crystallized on bulk glass using transmission electron microscopy. We report here preliminary results of correlated electrical and structural studies on pre – patterned silicon films crystallized under various conditions. Electrical characterization consists mainly of I – V measurements on diodes and MOSFETS fabricated in the crystallized material by conventional silicon IC processing techniques. For microstructural characterization of the same devices we rely on TEM, EBIC and spatially– resolved Raman scattering.

Picosecond laser pulse irradiation of single crystalline Si can produce severe damage, including amorphisation. Morphology changes are studied by means of optical and high-voltage-electron microscopy and are found to depend on energy fluence, surface orientation, and wavelength. The detailed analysis of the damage distributions and thresholds lead to the conclusion, that energy-confinement precedes the introduction of structural changes, which are formed by rapid quenching from the liquid state.

Directed energy sources have been used to recrystallize implanted amorphous layers in bulk single - crystal silicon and to crystallize silicon thin films on insulating amorphous substrates. All investigated forms of beam annealing leave residual electronic defects in and near the recrystallized layer on bulk silicon, with densities in excess of those obtained by conventional furnace annealing. This paper summarizes the general observations that may be drawn from numerous experimental investigations of electronic defects in CW beam - recrystallized bulk silicon and affirms the timeliness of similar comprehensive studies of residual defects in crystallized - silicon thin films.

EPR, photocarrier cyclotron resonance and DLTS measurements have been performed on laser-annealed Si wafers.

In virgin Si, an isotropic EPR line at g1 = 2.0055± 0.0005 and a carrier lifetime decrease, assessed by cyclotron resonance, are correlated to the creation of vacancy complexes associated to Oxygen.

In implanted Si, only defects due to the tailing effect of implantation process are observed by DLTS. These levels can be divided into two groups, one of them is stronglycorrelated with the implanted parameter. They are ascribed to primary or complex associations of point defects.

The EBIC mode of the SEM has been used to investigate the perfection of cw CO2 laser annealed Si. Even in material that contains no slip lines, non-uniform charge collection is found. A combined X-ray and electron microscopy (TEM, SEM) study identified the residual defects responsible for the EBIC contrast as interstitial submicron dislocation loops. Scanning cw laser annealing independent of the wavelength (10.6μm or 0.514μm) always introduces residual defects which act as recombination centers and reduce minority carrier lifetime.

The application of low temperature ion bombardment and pulsed laser or electron beam irradiation for the production of rapidly quenched superconducting alloys is described and compared with other more conventionally used non-equilibrium techniques. The comparison is based on the identification of the metastable phases produced by their specific electrical and superconducting properties. It is concluded that low temperature heavy ion bombardment is comparable with vapor quenching whereas laser quenching yields results similar to liquid quenching. Finally, the interesting superconducting properties of these rapidly quenched metastable alloys are discussed.

The microstructures found in pulse melted metals are reviewed. Epitaxial resolidification and the defects found in pure metals are discussed. Pulse melted alloys are then examined in order of increasing complexity. First, single phase alloys which solidify into metastable solid solutions are discussed. For higher concentrations and liquid-insoluble species, phase separation can occur, and the requirements for such separation are identified. If precipitation does not reduce the impurity concentration of the liquid, the liquid may be quenched to an amorphous solid. Comparisons of ion-implantation and pulse-melt produced amorphous alloys are given.

We have investigated the generation of point defects and dislocations, and the formation of dislocation loops as a function of pulse energy density in laser annealed Al, Ni, and MgO containing nickel precipitates. In the case of Al where vacancies are mobile above 200 K, mostly vacancy loops were observed at room temperature in laser melted layers. Dislocations are formed below the laser-melted layers as well as in specimens treated with pulses below the melting threshold, due to thermal stresses. In the case of Ni (where vacancies are mobile above 373 K) the microstructure in room temperature laser annealed specimens consists of primarily dislocations and their tangles. In MgO:Ni crystals, enough laser energy was absorbed to melt nickel precipitates. The dislocation structure around the precipitates and the transformation of nickel precipitates from coherent into incoherent, provided information on melting and crystal growth of these precipitates.

Silicon wafers with (001), (011), and (111) orientations were irradiated with pulsed Ba beams that caused shallow surface melting. Upon resolidification the dislocations are trapped with configurations which are characteristically different for each orientation. A novel feature are dislocations with a [100] Burgers vectors. The dislocations accommodate part of the thermal misfit between substrate and solidifying layer.

Lattice defects and precipitates induced in unimplanted and Sb-implanted <110> single crystal Al by single pulse irradiation with a Q-switched ruby laser were studied using ion beam analysis and electron microscopy. The absorbed laser energy during irradiation is directly measured in these studies to allow precise numerical modeling of the melt times and temperature profiles. For unimplanted Al, slip deformation gives rise to increased channeled yields throughout the analyzed depth and occurs for energies well below the melt threshold energy of 3.5 J/cm2. Slip deformation is also observed for irradiation energies above the melt threshold energy, and melting is accompanied by a discontinuous increase in the minimum channeling yield, X min- Implanted Sb (to ∼2 at.% peak concentrations) is found to impede epitaxial regrowth and result in polycrystalline Al formation for laser energies such that the melt front is believed not to penetrate through the Sb-containing region. For deeper melt depths, a metastable alloy is formed with up to 35% of the Sb located in substitutional sites. AlSb precipitate formation in the melt was not observed for room temperature irradiations; however, randomly oriented AlSb precipitates are observed for irradiation at substrate temperatures of 100 and 200 °C These measurements yield an estimated time for nucleation of AlSb precipitates in molten Al of 5 nsec < tnuc < 25 nsec.

Pulsed (50 nsec) electron beams with deposited energies of 1.1 ­ 2.4 J/cm2 have been used to rapidly melt a surface layer of Fe. Calculations show that this range of energies produces melt depths from 0.4–1.2 μm and melt times of 100–500 nsec. Optical microscopy and SEM of pulse treated polycrystalline foils show slip traces, as well as a general smoothing of surface features which shows that melting has occurred. TEM shows that the resolidified material is bcc, and that the material within a grain is epitaxial with the substrate. TEM also shows slip traces of {110} planes, as well as a high density of dislocations, both extended and loop. At the highest energy, subgrain boundaries are observed. Some samples were implanted with 1×1016 Sn/cm2 at 150 keV. After pulse treatment, the Sn depth profile was observed to have broadened, consistent with liquid phase diffusion. The Sn had the unexpected effect of suppressing slip at the sample surface.

Thin films of Cr and Cr films capped with Ni–8V have been magnetron sputter deposited on both single crystal and polycrystalline Cu and Cu alloy substrates. Surface alloying has been accomplished with scanning second harmonic Q-switched Nd-YAG and single shot Q-switched Ruby lasers and with 1 MeV Xe ions. The films and surface alloys have been characterized with RBS and channeling, AES and sputter profiling, optical microscopy, SEM, TEM and glancing angle XRD. Surface alloys containing as much as 50 at. % Cr have been produced. Following laser melting it is possible to produce substitutional solid solutions of Cr in Cu far in excess of equilibrium values.

Thin films of Cu on Zr and of Zr on Fe, Ni, Al, V, and Ti have been laser-pulse alloyed with the substrate metals using a raster-scanned Q-switched Nd:YAG laser. The high quench rates from the liquid state attendant to laser-pulse melting can promote the retention of a variety of metastable alloy phases. Structure and composition of the surface alloy layers of submicron thickness were studied with Rutherford backscattering (RBS)/channeling spectroscopy, optical and transmission electron microscopy and glancing angle x-ray diffraction (XRD). Composition profiles are compared with expected liquid-state diffusivities and structures are related to properties of bulk alloys. For example, in the well-known glass-forming system Cu-Zr, an amorphous phase was formed which crystallized upon annealing above Tc. In the Zr-Ti system, which shows complete solid solubility across the binary phase diagram in both the α and β phases, the hexagonal α-phase alloy is formed. Additional results on these and other systems are described below.

In the growing amount of the experimental data on the diffu sion processes of implanted impurities in a metal sample irradiated by highly energetic and very short pulses, some order may be introduced by the knowledge of the structure of induced thermal transient. The differences in such a structure due to two different energy sources (laser and electron beam) are de scribed with reference to the conditions under which Soret diffusion may be experimentally observed as distinct from nor mal diffusion. Laser induced transient seem to offer this possibility.

Some previously reported diffusion processes in laser irra diated systems are reinterpreted on this basis, showing the importance of the Soret effect.