The radiation from high-power, Q-switched lasers has been used recently in semiconductor research to 1) anneal the lattice damage caused by ion implantation, 2) diffuse surface-deposited dopant films, 3) recrystallize doped amorphous films deposited on substrates, and 4) remove precipitates present after conventional high-temperature, long-time, dopant diffusion. All of these phenomena can be understood in terms of a model based on macroscopic diffusion equations for heat and mass transport, cast in a finite-difference form to allow for the temperature- and spatial-dependence of the thermal conductivity, absorption coefficient of the laser radiation, and other quantities. Results of calculations with the model show that the near-surface region of the sample melts and stays molten for a time of the order of 10-7 secs during which dopant diffusion in the liquid state can explain the major features of the experimental results. Detailed results for arsenic-implanted silicon show the importance of non-equilibrium segregation effects.
A brief comparison of heating with pulsed electron bemas and lasers for the annealing of ion implanted semiconductors based on a computer simulation is given. The advantage of monoenergetic electrons is pointed out, and the use of such beam for distribution of impurities is suggested.
The physical mechanisms responsible for the annealing of ion implanted semiconductors for a wide range of materials and annealing conditions have been identified through the use of in-situ dynamical measurements of optical properties. These measurements combined with post-anneal sample characterization, i.e., channeling, electrical measurements, etc., lead to the identification of two different annealing mechanisms. For the case of Q-switched lasers a liquid phase epitaxial regrowth of the damaged layer is indicated. For c.w. lasers solid phase epitaxial regrowth can occur.
Both pulsed and scanning cw lasers operating at wavelengths in the range 0.5-1.06 μm can be used as a method to rapidly heat the surface and underlying bulk of a semiconductor material to a precisely controlled temperature up to melting. As a result, lasers can be utilized to anneal ion-implanted layers, reduce fixed interface charge of deposited oxide films, grow surface oxides, react metal/silicon layers to form silicides and improve and control the electrical properties of deposited silicon films.
Experimental measurements and theoretical calculations of recrystallization induced by single laser scans over amorphous silicon are reported and quantitatively compared. The details of laser-induced crystal regrowth of high dose ion implanted silicon studied over a wide range of experimental parameters are accurately predicted from three dimensional thermal transport considerations and solid phase epitaxial crystal growth. Best agreement between theory and experiment is obtained with the value of 2.24 eV for the activation energy of recrystallization.
High-power laser pulses are being used to replace conventional high-temperature furnace processing in the development of high-speed, low-cost solar cell fabrication methods. Three different approaches to p-n junction formation have been studied: (1) junction formation from laser-annealed, ion-implanted Si in which laser radiation is used to remove radiation damage and to recover the electrical activity in the implanted layer; (2) junction formation by laser-induced diffusion in which a thin film of dopant is first deposited on the substrate and then incorporated into the near-surface region by laser-induced melting of the near-surface region; and (3) laser-induced epitaxial junction formation in which a heavily doped amorphous silicon layer is deposited on a Si substrate and epitaxially regrown from the substrate layer which has been partially melted by the laser radiation. We have demonstrated that all three methods can provide suitable candidates for junction formation in high-efficiency Si solar cells. These laser techniques are particularly attractive for p-n junction formation in thin-film polycrystalline material. In conventional thermal diffusion, enhanced diffusion along grain boundaries can cause short circuits between the emitter region and the substrate. With the laser techniques, the near-surface region of the film actually melts but for such short times (~ 10-7 sec) that significant dopant diffusion to and along the grain boundaries does not occur, thus giving control of grain boundary diffusion and segregation.
Development of low cost solar cells fabrication technology is being sponsored by NASA JPL as part of the Low Cost Solar Array Project (LSA). In conformance to Project requirements ion implantation and laser annealing were evaluated as junction formation techniques offering low cost-high throughput potential. Properties of cells fabricated utilizing this technology were analyzed by electrical, transmission electron microscopy, Rutherford backscattering and secondary ion mass spectrometry techniques. Tests indicated the laser annealed substrates to be damage free and electrically active. Similar analysis of ion implanted furnace annealed substrates revealed the presence of residual defects in the form of dislocation lines and loops with substantial impurity redistribution evident for some anneal temperature/time regimes. Fabricated laser annealed cells exhibited improved spectral response and conversion efficiency in comparison to furnace annealed cells. An economic projection for LSA indicates a potential for considerable savings from laser annealing technology.
Laser chemical vapor deposition (LCVD) uses a focused laser beam to locally heat the substrate and drive the CVD deposition reaction. Several different deposition reactions and substrates have been examined as a function of intensity and irradiation time using a CO2 laser source: Ni on SiO2, TiO2 on SiO2, TiC on SiO2, and TiC/on stainless steel. LCVD film thicknesses range from <100Å to >20μm. Deposition rates of mm/min have been observed for LCVD Ni and TiC and 20 iim/min for LCVD TiO2. The diameter of the deposited films is dependent on irradiation conditions and can be as small as 0.10 the laser beam diameter. The LCVD films exhibit excellent physical properties such as adherence, conductivity, hardness and smoothness. The advantages of LCVD are the same as other laser processing techniques, i.e., control of area and depth heated, rapid heating and cooling, and cleanliness. Possible applications of this technique include: formation of ohmic contacts and localized protective coatings in semiconductor devices; localized coatings and dopants for waveguide optics; surface hardening and alloying of machine surfaces; welding of ceramic materials; and generation of new materials.
Plasma-sprayed coatings of corrosion and wear resisting alloys were laser surface melted, with the objective of densifying them and improving their adhesion to the substrate materials. Corrosion resisting coatings of titanium and of 316L stainless steel on steel substrates were remelted with good results. After laser processing, these initially porous coatings were fully dense and exhibited good surface quality. Results with wear resisting coatings were disappointing beause the hard wear resisting phases crack under the tensile stresses developed during cooling. Results are discussed in terms of the effects of the initial coating properties, the laser beam parameters, and other processing variables such as processing atmosphere and sample translation rates.
A process has been developed by which ideal ceramic powders are synthesized from CO2 laser heated gas phase reactants. Silicon nitride and silicon powders are formed from mixtures of silane and ammonia or silane gases respectively. The resulting powders are spherical, nearly equal diameter, small particle size, high purity and loosely agglomerated. In addition to achieving these ideal characteristics for powder densification processes, the synethesis process is extremely efficient from the criteria of materials utilization and energy consumption per unit mass of powder. The process has been analyzed and modelled; powders have been characterized extensively.
The chemical vapor deposition of titanium carbide using a 1.4 KW CO2 laser as the heat source is examined. The relationship between the laser power density and traverse speed on deposit thickness and properties is described. The laser chemical vapor deposited carbide coating is characterized by SEM, EDAX, optical microscopy and microhardness tests. Hardness up to 2800 Knoop is observed leading to the possibility of extremely wear resistant surface coatings. Advantages such as controllability of input energy and small heat affected zone resulting in little or minimum distortion open the door to a wide variety of selected area chemical vapor deposition processes.
A scanning laser beam was used to melt and normalize the surface layer of sensitized 304 stainless steel. Subsequent Strauss tests indicated a complete resistance to intergranular corrosion. Mechanical testing at strains less than 15% also showed laser surface melting to indefinitely extend specimen life in a stress corrosion environment. At strains greater than 15%, the laser-scanned protective layer was breached by cracks. A maximum critical laser-scanning velocity compatible with normalization of the surface layer is calculated. Similarly, a minimum critical laser-scanning velocity required to avoid resensitization is determined. The stress distribution in a 304 stainless-steel specimen with a laser-melted and self-quenched surface layer is estimated and shown to be compatible with the observed appearance of martensite in the melted surface layer.
A review of reflectors with temperature-dependent absorptance shows that above a critical irradiance, exponential thermal runaway occurs, while below it, the temperature rise approaches either a steady-state value or a slowly increasing value. Calculated damage thresholds, which are generally reduced substantially when the temperature dependence of the absorptance is included, fail to agree with experimental values in most cases because isolated-spot damage or plasma formation occurs. However, excellent agreement is obtained for special ion-cleaned or diamond-turned samples, which damage uniformly for long (~100 ns) pulses.
Anomalous behavior of the specular reflectivity of polished metal surfaces subjected to intense CO2 laser pulses is reported. Irradiations of several metallic elements and alloys were conducted in hard vacuum and time-resolved measurements of specular reflectivity were performed during the 60-ns pulse. Results indicated in many cases a transient drop in reflectivity to a level in the 10-40% range followed by recovery to near unity by the end of the pulse for a stabilized surface. Metal vapor plasma was not detected in most cases and is not believed to be the dominant mechanism operative. Theoretical estimates of surface temperature transients are presented along with correlations of observed surface damage. Mechanisms which may explain the observed anomaly are discussed.
The reflectance behavior of a target surface during laser irradiation determines the laser energy which is directly absorbed. Experimentally, the reflectance of a metal surface has been observed to undergo a sharp and substantial decrease during an intense laser pulse. Explanations have been offered based on an increase in electron-phonon collision frequency as the temperature of the metal surface rises to the melting point of the metal. It is doubtful, however, that the temperature dependence of a Drude-type free-electron model can explain the substantial reflectance changes reported, particularly for high conductivity metals such as copper. Three other classes of explanations have been proposed: 1) deformation of the metal surface, 2) plasma formation in the front of the target, or 3) a nonlinear process causing enhanced absorption within the metal. Specular and total reflectances of metal surfaces during ruby laser irradiation have been measured at the Polytechnic. The laser energy absorbed has been calculated and the temperature history of the metal surface determined using a one-dimensional heat-conduction approach. Reflectance and temperature histories have been related to permanent changes observed in the metal target surface. Only at temperatures significantly above the melting point of the metal does a substantial decrease in the total reflectance occur.
High power continuous lasers can be used to modify the chemical composition of alloy surfaces to depths ranging from 0.01 to 1 mm. Such coatings exhibit potential advantages over more conventional coating techiliques in terms of the integrity of the coating, the character of the interface between the surface alloy and the substrate, and an increased control over the composition of the coating. The processing conditions used in laser surface alloy ing are selected in order to facilitate the mixing of the alloying material with molten substrate material. The manufacture of chromium steel suiface alloys on low carbon steel substrates is described in terms of this mixing and the extension of these results to other systems is discussed. Auger and electron microprobe analyses of chromium steel surface alloys have been performed and indicate that a high degree of compositional uniformity can be obtained with proper control of the processing. The various types of metallurgical and morphological structures are described.
A laser-shock process is described that can greatly improve the fatigue and crack growth resistance of some aluminum aircraft alloys. Additional developments needed for eventual commercial application of the processes are defined. Some of the evolving functional requirements for the laser system are given. It is concluded that a commercial laser system will probably have to be developed specifically for this application.
Carbon Dioxide and YAG lasers have been used to establish a narrow molten zone in a pre-formed polycrystalline silicon ribbon. To achieve the narrow melt across the width of the ribbon, the laser is focussed and scanned across the ribbon by means of mirror scanners. Large grained "macrocrystalline" silicon ribbon is then drawn from the molten zone at growth rates up to 12.5 cm/min. Photovoltaic cells fabricated on this material have shown conversion efficiencies up to 11.3%. The actual absorbed laser power which is required for ribbon growth of interest, e.g. .15 mm thick, 7.5 cm wide, 7.5 cm/min. growth velocity, is only a few hundred watts. Unfortunately, the reflectivity of liquid silicon at 10.6 μm is very high (>90%) so that multikilowatt CO2 lasers will be needed. Nd:YAG lasers offer about a five-fold improvement in laser coupling to liquid silicon but this advantage is offset by the lower efficiency and higher operating cost of Nd:YAG lasers. The effective laser coupling to the melt can be essentially doubled by using spherical reflectors to re-image the reflected laser beam back onto the molten zone. The emissivity change which occurs upon melting, while disadvantageous for laser power requirements, is a real benefit for melt zone control since this produces an inherent stabilizing effect.
While the application of high power (50-5000 W) lasers to materials working is well known, the use of low power (1-5w) CO, lasers has received little attention. This paper presents methods of utilizing low power CO2 lasers in materials processing, such as cutting, drilling, and welding of small organic (e.g., plastic) parts. Laser hardware is discussed and the waveguide laser is presented as an example of low-power materials working hardware. This paper also reports some of the applications which are ideally-handled by low power CO lasers, and reviews the factors which contribute to the successful use of these lasers.
A mathematical model for the transient heat flow in cylindrical specimens is presented. The model predicts the temperature distribution in the vicinity of a moving ring-shaped laser spot around the periphery of the outer surface of a cylinder, or the inner surface of a hollow cylinder. It can be used to predict the depth of case in laser surface transformation hardening. The validity of the model is tested against experimental results obtained on SAE 4140 steel.