#16 wire gauge crimp joints of good quality were made tinned wires in the conventional way. Some of them then subjected to twisting in gap between the two wings. All of them were heat-soaked at 135/spl deg/C for 300 hours. They were then into a thermal-shock furnace for thermal cycling from -40/spl deg/C to 125/spl deg/C, each cycle lasting one hour for a total of 570hours. The four-leads method was used to measure the contact resistance for every cycle under the dry circuit condition. The contact resistance of good crimps generally increased at a slower pace over time than the bad ones. The resistance of some good and bad crimps would tend to more or less off after 100-200 cycles while others would continue to increase their resistance without leveling off and some of them might show a sudden jump. Nevertheless, the measured resistance history showed that a crimp with a tinned wire exhibited accepted performance whether it was good or not so good a crimp. The results also showed that the temperature history would effect the contact resistance measured at a later time. For example, the resistance of a crimp at 125/spl deg/C would be somehow different depending on weather the previous temperature started from -40/spl deg/C or from -17/spl deg/C. Data analysis showed that even with severe thermal shocks for 570 hours and possible consequence of repeated shearing, the contact in the crimp was still essentially a metallic one.

Experiments were performed to investigate the contact resistance behavior and breakdown characteristics of the oxidized copper surface. In order to eliminate the influence of the mechanical pressure on the contact resistance, a gallium liquid metal drop was used as the probe. Current was varied from 0.1 mA to 1 A and then ramped back down to 0.1 mA. The voltage across the contact between a gallium drop and the oxidized copper was recorded. The contact resistance was then computed and plotted against the voltage. It was discovered that the A-fritting voltage of the contact was about 2.0 V when the film thickness was about 55 nm, which was consistent with the results of Holm (1967). It was also observed that the film could break down and then recover partially for several cycles before the film was finally broken down completely to settle at a lower voltage. When the film thickness was 8 nm or less, the measured contact voltage at the fritting stage was always less than 0.4 V. In such cases, it was not possible to observe the typical 'avalanche' A-fritting (breakdown), even though the film had not been ruptured by mechanical load.

The occurrence of intermittences in electrical circuits can result in serious reliability problems particularly where low signal voltages and currents are used. It is well known that differential thermal expansion or vibration in electrical connectors can result in micromotion resulting in fretting corrosion. This can lead to an increase in contact resistance and eventual loss of electrical contact. The occurrence of electrical intermittences has been considered a precursor to contact failure associated with fretting corrosion and can cause disruption in digital circuit signals. We report the development of instrumentation that simultaneously measures the occurrence of electrical intermittences along with contact resistance during the fretting of electrical contacts. In addition, the measurement system records contact friction and normal force dynamically. Intermittences can be counted and timed with durations from 20 ns to milliseconds as a function of fretting cycles and correlated with the increase in contact resistance. All systems are integrated under LABVIEW computer control software. Measurements were made on Cu-Cu and Sn-Sn rider/flat combinations. Results will be interpreted in terms of the influence of wear debris on the electrical properties of the contacts.

We report a study of the change in lattice constant when single-crystal silicon is substitutionally doped with boron. The measurements were made using 10-μm-thick epilayers with boron concentrations (NB) in the range 1.7×1019–1.2×1020 cm−3. The influence of elastic strain in the epilayers and their substrates was eliminated by including Bragg reflections from planes that were inclined to the (100) surface. We obtained a value for the lattice contraction coefficient β=(5.19±0.09)×10−24 cm3, where the range is ± one standard error. Specimens with NB≊2×1019 cm−3 were strained to give coherent interfaces with their substrates. Evidence for relaxation (presumably by misfit dislocations at the interface) was observed with NB≊5×1019 cm−3 (corresponding to a lattice mismatch of approximately 2.5×10−4), but this occurred in a patchy fashion and remained incomplete while NB was increased to 1.2×1020 cm−3 (corresponding to a mismatch of 6.5×10−4). This partial relaxation occurred at somewhat smaller mismatches than would be expected from reported studies of silicon-germanium alloy epilayers on silicon substrates. Our technique for analyzing a set of Bragg reflections to separate the effects of elastic strain and lattice mismatch appears to be novel and has general applicability to the study of epilayer/substrate combinations.

Measurements of the residual stress in polysilicon films made by Low Pressure Chemical Vapor Deposition (LPCVD) at different deposition pressures and temperatures are reported. The stress behavior of phosphorus (P)-ion implanted/annealed polysilicon films is also reported. Within the temperature range of deposition, 580 °C to 650 °C, the stress vs deposition temperature plot exhibits a transition region in which the stress of the film changes from highly compressive to highly tensile and back to highly compressive as the deposition temperature increases. This behavior was observed in films that were made by the LPCVD process at reduced pressures of 210 and 320 mTORR. At deposition temperatures below 590 °C the deposit is predominantly amorphous, and the film is highly compressive; at temperatures above 610 °C (110) oriented polycrystalline silicon is formed exhibiting high compressive residual stress.

The microstructure, of unannealed and annealed polysilicon films was studied using TEM and XRD. The LPCVD films were grown at 600°C and 620°C with 320 mTorr of silane, and at 580°C with 220 mTorr of silane. The substrates were [001] Si with a thin oxide film. The stress state of the films changed from compressive at 580°C and 620°C, to tensile in a narrow region around 600°C. The same materials were annealed at 1100°C. The unannealed films vary from partially amorphous at 580°C, where the slowest growth rate was observed, to randomly oriented and equiaxed at 600°C, to columnar and highly oriented at 620°C. The grains in the 620°C material have a high stacking fault and polytype density, and it was proposed that these occurred on growth, and not from dislocation motion. The grain size increased from 40–250run to 100–300 nm in the 600°C samples when annealed at 1100°C, and the density of twins and stacking faults was reduced. The hexagonal phase was observed only in unannealed materials in SAED and as broad “wings” at the base of the 111d.c (diamond cubic) peak in XRD. Within the limits of SAED, no change in lattice parameter of the hexagonal phase was observed.

A review article is presented of the author’s work on light emission from electron tunneling junctions. The light emission arises from the radiative decay of junction surface plasmon modes excited by tunneling electrons. The coupling out of the light is assisted by roughness and resonant particle scattering schemes. As a consequence of the quantum nature of the light generation process, there is a fundamental relation between the applied voltage and the observed emission spectrum. Because of the limited tunneling electron‐surface plasmon coupling, these thin‐film light sources have photometric brightnesses of only a few foot‐Lamberts. Other avenues of research for brighter thin‐film light sources are: development of metal electrodes having a lower surface plasmon frequency excitation of electronic transitions.

The discovery of a new thin-film electroluminescence system based on impurity-doped anodic aluminum oxide is reported. Average brightnesses of 30 f L have been observed for devices at room temperature operating at 1 kHz biased at 60 V rms, resulting in 7 mA electronic current and a 10% duty cycle. A new stabilizing layer of manganese oxide permits long-term operation without catastrophic breakdown. While electrical stability is achieved, loss in electroluminescence efficiency is apparent, possibly due to electromigration of the impurity ions.

We calculate the time-dependent decay of the surface-plasmon-coupled molecular fluorescence from a solution of diffusing dye molecules near a metal. Measurements of this signal for an ethanol solution of Tb acetylacetonate next to a Ag film agree well with the calculation when effects due to the non-local response of Ag are included.

Light emission via inelastic tunneling (LEIT) has been observed from the metal-insulator-semiconductor tunnel-junction Al-Al2O3- Sn-doped indium oxide (ITO). The spectra exhibit the bias-voltage-dependent upper-frequency quantum cutoff relation. These devices were stable at room temperature for months at a time while biased to 3 V. The external quantum efficiency of this metal-insulator-semiconductor system is about 0.1 that of the metal-insulator-metal system Al-Al2O3-Ag. Enhancement of the light emission using silver particle resonators is also reported.

Recently we reported the discovery of a new way of generating light from a solid-state thin-film structure

It has been shown that inelastic tunneling excitation of surface plasmon modes can result in light emission when electrodes are properly prepared. The light emission establishes a fundamental quantum relationship between the maximum optic frequency and the applied voltage. Such a process can be used as a basis for spectroscopy when modulation-derivative techniques are used. The derivative method is very similar to that used in IETS except it is a second derivative of a photo detection current that is measured. We term this LEIT spectroscopy (light emission by inelastic tunneling). We will discuss the physical processes underlying the LEIT effect as well as possible applications. Some of the key factors of this light source are: 1) the precision of the relationship between voltage and maximum optical frequency; 2) mechanics for coupling out light for excited surface plasmon modes; 3) the optical range over which such sources can be operated.