Performance characterization of monolithic thin film resistors

Abstract

Thin film resistors have a large resistance range and stable performance under high temperature operating condition. Thin film resistors trimmed by laser beam are able to achieve very high precision on resistance value. As a result, thin film resistors have been widely used to improve the performance of integrated circuits such as operational amplifier, analog-to-digital (A/D) and digital-to-analog (D/A) converters, etc.. In this dissertation, a new class of thin film resistors, silicon chrome (SiCr) thin film resistors, has been investigated at length. From thin film characterization to aging behavior modelling, we have carried out a series of engineering activities. The characteristics of the SiCr thin film incorporated into three bipolar processes were first determined. After laser trimming, we have measured a couple of physical parameters of the SiCr film in the heat affected zone (HAZ). This is the first time the sheet resistance and the temperature coefficient of resistance (TCR) of thin film in the HAZ have been characterized. Both thermal and d.c. load accelerated aging tests were performed. The test structures were subjected to the aging for 1000 hours. Based on the test data, we not only evaluated the classical thermal aging model for untrimmed thin film resistors, but also established several empirical thermal aging models for trimmed resistors and d.c. load aging models for both trimmed and untrimmed thin film resistors. All the experiments were carried out for both conventional bar resistors and our new Swiss Cheese (SC) resistors. For the first time, the performance of laser trimmed SC resistors, which was experimentally evaluated, shown a clear superiority over that of trimmed bar resistors. Besides these experiments, we have examined different die attach techniques and their effects on thin film resistors. Also, we have developed a number of hardware systems and software tools, such as a temperature controller, d.c. current source, temperature control program, automated resistance measurement program and laser system control program. Particularly, with a low cost, high temperature accuracy, two stage adaptive control system, we were able to implement a test environment with only [plus or minus]0.03°C temperature variation. This temperature stable test environment significantly improved the accuracy of the experimental results.