No abstract

not Available.

The diffusion of phosphorus from a thick epitaxial layer into a silicon substrate has been investigated using the spreading-resistance technique. By comparing the diffusion profile under a free oxidized surface with the profile under a masked surface, it has been shown that surface oxidation enhances diffusion at low temperatures and retards diffusion at high temperatures. The conditions which favor enhanced diffusion are those under which stacking faults grow; retarded diffusion is associated with stacking-fault shrinkage. Enhanced diffusion is due to the oxide injecting excess interstitials into the substrate; retarded diffusion is caused by vacancy injection. It is concluded that phosphorus diffuses by the interstitialcy mechanism and that the criterion for interstitial or vacancy injection is the relative value of the anion and cation fluxes across the oxide-silicon interface.

The electrical conductivity of high purity (99.8 pct) A1N has been determined using both a-c and d-c techniques. Measurements were made at temperatures between 700 ~ and 1000~C and in a nitrogen pressure range of 1-10 -5 atm. The electrical conductivity is independent of nitrogen pressure. Galvanic cell results indicate that ionic conduction is negligible. Over the experimental temperature range, the electrical conductivity of A1N follows the equation ,7(ohm-l-cm -~) = 6.89 • 103 exp (--E/kT) where the activation energy (E) is 1.82 • 0.06 eV. It is concluded that the conduction process is extrinsic due to carbon impurities (,,,350 ppm) present in the A1N.Aluminum nitride is a potentially useful refractory material, particularly at elevated temperatures. Its oxidation in air at temperatures up to 600~ is low (1, 2), and it is resistant to attack by acids, molten metals, and water vapor (1, 3). Recent studies indicate that A1N is very resistant to chemical attack by lithium at 400~ (4); thus A1N may be a suitable electronic insulator in alkali metal-molten salt batteries. Aluminum nitride also has potential applications in elevated temperature semiconductor devices (5-7). Although these various applications require a detailed knowledge of its conductivity at elevated temperatures, there is considerable variation in the activation energy and magnitude of the electrical conductivity of A1N measured by previous investigators (1,3,6,(8)(9)(10)(11)(12). In this study the electrical conductivity of high purity (>99%) A1N has been measured using a-c and d-c techniques at temperatures between 700 ~ and 1000~ and at nitrogen pressures between 1 and 10 -5 atm. ExperimentalHot-pressed polycrystalline A1N samples were prepared by Cerac/Pure Incorporated. Aluminum nitride powder was synthesized by the electric discharge method (3) and hot-pressed without binder in a graphite die at ~1600~ under a nitrogen atmosphere. The 1 cm diameter A1N rod was machined to remove surface-impregnated carbon and then cut into pellets of different thicknesses. Sample porosity did not exceed 5%. The pellets were annealed in our laboratory at 1400~ for 76 hr in high purity nitrogen. Chemical analyses performed by the supplier both before and after annealing are listed in Table I. It should be emphasized that the purity of the annealed samples (99.8%) is higher than A1N samples used in previous studies.Gold electrical contacts, which were used in the a-c conductivity measurements, were fabricated using the following procedure. The fiat surfaces of the AIN pellets were polished and then rinsed in acetone and petroleum ether. After coating the fiat surfaces with du Pont thermosetting gold paste (No. 5780), the pellets were heated slowly in air to 200~ held for 10 hr, and then annealed for 2 hr at 400~ Coherent gold contacts with less than 0.1 ohm resistance across 9 Electrochemical Society Active Member. Key words: extrinsic conduction in A1N', effect of carbon impurities, electronic conductivity in A1N, chromium nitride electrodes.

A contactless rf technique has been developed for measurement of the photoconductivity induced in silicon wafers by a flash of light from GaAs laser diodes. The carrier lifetime inferred from the photoconductivity decay correlates well with diffusion length measurements made on solar cells fabricated from the same wafers. The technique has been applied successfully to silicon whose resistivity is as low as 0.1 Ω cm and lifetime as short as 0.2 μs.

Hall measurements and four‐point probe resistivity measurements were used to determine the concentration profile of boron in doped semiconductor silicon ingots grown by Czochralski and Bridgman techniques. The concentration profiles were fitted to the normal segregation equation and the effective segregation coefficient, knormaleff , was calculated. The average value of knormaleff for boron was 0.786±0.036 in Czochralski single crystals and 0.803±0.036 in Bridgman polycrystals.

A technique is developed to analyze boron and phosphorus concentrations in silicon by determining net carrier concentrations from Hall‐effect measurements. The goal of the technique is to measure these impurities in purified metallurgical silicon (PMS) which is to be ultimately used for fabricating solar cells. Net carrier concentrations are measured on two wafers cut from a second‐generation Czochralski ingot pulled from PMS. Boron and phosphorus concentrations in the PMS and the ingots are calculated from the net carrier concentrations by applying the segregation equations of the impurities. The accuracy of the boron measurements between 1 and 30 ppma is equivalent to that obtained on similar samples analyzed by multiple‐pass float zoning. Phosphorus levels can be measured down to about 20% of that of the boron concentrations. Excellent precision of the Hall‐effect technique is demonstrated. Sensitivity calculations show that net carrier concentration is the most critical parameter affecting the accuracy of the method. Overall, the technique is considered to be sufficient for making quality‐control decisions with respect to boron and phosphorus levels in PMS and ingots grown from it.

Durch Reduktion von SiO2 mit Kohle im 100 kW Lichtbogenofen dargestelltes Si besitzt eine B‐Verunreinigung von 1.8 ppma.