Temperature-cycling acceleration factors for aluminum metallization failure in VLSI applications : C. F. Dunn and J. W. McPherson. 28th A. Proc. Reliab. Phys. Symp. (IEEE), 252 (March 1990)

Abstract

Pulse shape distributions resulting from the interaction of γ-rays in planar germanium detectors were studied as a function of field intensity, γ-ray energy, and detector width. It was observed that at 2300 V/cm the initial rate of charge collection approached 10 nsec/mm. Due to a slow component the total charge collection was about 30% longer. At 1000 V/cm the collection time was somewhat longer, and at 250 V/cm it was considerably longer. These observations appeared to be independent of γ-ray energy in the range of 0.6–3 MeV and detector widths of 3–11 mm. Events interacting near the center of the detector result in fast rising pulses and those interacting near one of the electrodes have a risetime which is approximately twice as long, thus giving rise to a distribution of pulse shapes. With low energy γ-rays all shapes occur with equal probability, but at higher energy the pulses tend to have a more uniform shape. Due to the spread in risetimes, signals from monoenergetic γ-rays do not cross a given pulse height level at the same time and therefore produce timing uncertainties. With a 1 cm detector timing uncertainties of more than 30 nsec have been observed at the 50% level, and less than 5 nsec at the 5% level. It therefore appears that for accurate timing of all signals, the timing discriminator should be set at the lowest level consistent with the system noise. The variations in risetime can also affect the energy resolution. If a 1 cm detector is used in conjunction with a pulse shaping network consisting of a 1 μsec RC differentiator and integrator, the variations in risetimes may cause an amplitude spread on the order of 0.03%. At γ-ray energies of 10 MeV this is comparable to the spread due to statistics.