Process Optimization of BCl3/Cl2 Plasma Etching of Aluminum with Design of Experiment (DOE)

Reactive ion etching of aluminum (Al) had been carried out with BCl3/Cl2, and process optimization was performed using statistical process modeling technique. I-optimal design was employed to set up the etch experiment with operating parameters, namely, gas composition, RF power, and chamber pressure. Analysis has been performed on etch rate, selectivity, and profile, individually, and the process optimization including three responses of interest, simultaneously. Through the experiment we have confirmed that etch rate was affected by Cl2, but the excessive amount of chlorine decreased etch rate. BCl3 is significantly related with selectivity, and increased power gives good selectivity because RF power range was from 130 to 170 watts. Selectivity was affected by how to fast remove Al2O3 by BCl3. Etch profile was strongly related with all the three parameters, and they affect on anisotropic sidewall property. Suggested optimization simultaneously considers three responses of interests, which is crucial in process development and optimization for quickly ramping up high volume manufacturing. Introduction The present semiconductor fabrication equipments used at a university laboratory are superannuated or utilized relatively for a long period due to its high cost. Although the equipments have a good performance at the beginning of purchase, processing performance is gradually degraded as years over. When experimenters are carried out experiment with existing recipe, obtaining desired result of an experiment is sometimes difficult. In such a case the following question can be arisen; what was the method that reduces an error and predicts a range of fluctuation of the result. In this paper, we will focus our attention on aluminum (Al) etch, which is the most relevant material in interconnects in semiconductor fabrication. Several metallic materials such as aluminum and its alloys, tungsten, copper, platinum are frequently utilized in not only the interconnection layer but also the device packaging process of the semiconductor manufacturing industry. Choice of material is important for achieving a desired interconnection layer condition for better transmission of electrical signal. Material is selected based on its essential properties, such as electrical conductivity, mechanical and chemical stability, adhesion characteristics, film deposition considerations, and patternability [1]. Although copper is increasingly used for metallization process in current leading edge semiconductor manufacturing, aluminum has been generally used for an interconnection material in integrated circuits (IC) because of its bondability to Si and SiO2 and easiness for patternability than other metals. The fluorine containing gases used to etch Si and SiO2 are not suitable for etching Al since etch product, AlF3, has very low vapor pressure at temperatures below 100°C. AlCl3 have sufficiently high vapor pressures to allow plasma etching of Al, and thus chlorinated gases have been exploited to develop dry etch process for aluminum thin film [2]. The most common gases are BCl3/Cl2 composition at Al etch. Chlorine is the primary etchant and BCl3 act as a source of heavy ions for bombardment. Therefore, as the first step in our goal, we have studied how process optimization of mixed BCl3/Cl2 plasma etching of Al with design of experiment (DOE) in superannuated Oxford 80 plus. Process is optimized using a statistically established process model. A designed experiment was the simultaneous evaluation of two or more factors (parameters) for their ability to affect the resultant average or variability of particular product or process characteristics. I-optimal design was employed to set up the etching experiment with operating parameters. The second step we have confirmed that a model established through the analysis of DOE process analysis as far as it goes to predict the range of fluctuation of results of etch rate, profile, and selectivity.