Prediction of Silicide Formation Sequence from the Principle of the Largest Free Energy Degradation Rate

  • PDF / 1,211,488 Bytes
  • 6 Pages / 414.72 x 648 pts Page_size
  • 34 Downloads / 176 Views

DOWNLOAD

REPORT


PREDICTION OF SILICIDE FORMATION SEQUENCE FROM THE PRINCIPLE OF THE LARGEST FREE ENERGY DEGRADATION RATE

Lin Zhang and Douglas G. Ivey, Department of Mining, Metallurgical, and Petroleum Engineering, University of Alberta, Edmonton, Alberta, Canada, T6G 2G6 ABSTRACT

Recently, a generalized kinetic model for silicide formation has been developed and a principle of the largest free energy degradation rate (largest FEDR) has been proposed. Free energy change rate as a function of diffusion flux has been derived from the kinetic model so that the principle can be justified not only for first phase formation but also for the subsequent processes. Silicide formation sequence predictions, from the model and the FEDR principle, for 15 metal-Si systems have shown very good agreement with the experimental results reported in the literature. In this paper, a brief introduction to the model and the principle is given, followed by examples of the predictions for several metal-Si systems and comparison with experimental results. ]INTRODUCTION During the last two decades, transition metal silicides have attracted tremendous interest frrom scientists around the world. Many silicon-rich silicides have been widely used in very large scale integrated circuit (VLSI) technology and will probably be used in other electronic devices. It is very useful, for both technical applications and theoretical study, to be able to predict silicide formation sequences during thermal annealing of thin metal film-silicon substrate diffusion couples. The formation sequences can be grouped into two categories depending on the metal film thicknesses. One type is single silicide layer sequential formation, for films thinner than about 100 nm, in which only one silicide layer grows at a time. The second type is multiple layer growth, for thicker metal films, in which after a first silicide layer grows over a critical thickness, a second silicide starts to form and then the two layers grow at :same time, and so on. In fact, the latter category can also include multiple phase growth in a lateral metal-Si diffusion couple and the early stages in a bulk couple. A successful theory should be able to predict the experimental phenomena of both categories. Of a number of models and theories proposed for silicide formation kinetics in the past, the Walser and Bend rule [1] is the only one that is most often used to predict first phase nucleation in metal-Si diffusion couples. This rule states that "the first compound nucleated in planar binary (thin film) reaction couples is the most stable congruently melting compound adjacent to the lowest temperature eutectic on the bulk equilibrium phase diagram". From this rule, in a given metal-Si couple, only one stable phase can be the first phase to form. According to earlier experimental results, this rule appeared to work in many cases, so that it was very influential for more than ten years. With better analytical techniques being used for characterizing metal-Si inter aces, more and more experimental results have sho