Mechanical Characterization For Low-k Film Of ULSI Interconnects By LSAWs

Interconnection has become the bottle-neck for developing high-speed, high-reliability, and high-density integrated circuit. The major effort made to solve this problem is to integrate low-k dielectric into interconnection system. To eliminate the problem caused by mechanical incompatibility with current integrate-circuit processing in introducing low-k dielectric into interconnection, the key is to determine the hardness of low-k film accurately. For the merits of quickness, preciseness and non-destructive, LSAWs is the most favored technique to determine the Young's Modulus of fragile, soft, and ultra thin film.This paper develops both the theoretical calculation of dispersion curve of surface acoustic waves propagating on layered films and the methods for processing the experiment data.For theoretical calculation, this research developed an"isotropic film/ Si substrate"calculation model from the wave equation in the propagation media, which is suitable for mimicking the ULSI interconnection system. Based on large quantity of calculation results, this paper investigated the affection of various film-parameters such as thickness, density, and Young's Modulus on the propagation characteristic of surface acoustic waves. Nano-porous SiO2 film is the most promising candidate in future low-k interconnection. However, today's calculation model is not adequate to imitate the characteristic of films with nano-pores distributed periodically. This issue was solved by establishing a new calculation model of"transversal isotropic film/ Si substrate". According to this calculation model and its sufficient calculation results, this paper found out how calculation options such as wave propagation direction and how nano-porous film's parameters affect the surface acoustic wave's dispersion curve.On the side of experimental data processing, this research firstly established an adequate sampling system according to the Nyquist's theorem. To prohibit the disturbances of the experiment noise in the collected signals, a FIR software filter with linear phase characteristic was designed by using loading-window designing method. During designing the filter, the spectrums of various windows are carefully investigated. The parameters'effect of Kaiser Window were also carefully checked in this paper, and at last this paper chose Kaiser Window as the truncation window in the design due to its well-flexibility. This paper used FFT based on Tukey-Cooley algorithm to analysis the spectrum of LSAWs experimental signals. Based on the analysis of various common windows'spectrum characteristic, this paper adopted a suitable window to eliminate the spectrum leakage effect. Palisades effect was decreased by adding zeros to the end of the original signals. To fit the theoretical curves to the experiment dispersion curve, least-squares procedure was applied to fit the discrete data points into a smooth dispersion curve. The resolution of couples of closer dispersion curves was improved by conducting derivate on the curves. At last, this paper developed a program based on dichotomizing searching to automatically find the fittest curves, which gives the Young's Modulus value of the sample under test quickly and within precision.