Control of Molecular Contaminants in Porous Low-k Dielectric Films and in UHP Gas Delivery Systems

As compared to silicon oxide, porous low-k dielectric materials are more susceptible to molecular contaminants. As the device feature size decreases, control of molecular contaminants in porous low-k dielectric films and in UHP gas delivery systems becomes increasingly more challenging. Moisture was selected as the principal model contaminant in this research because the moisture impurity retained in the dielectric films not only increases the effective dielectric constant (k) value of the films but also degrades the reliability of the device. Dry-down of moisture contaminated UHP systems takes days to weeks, which significantly decreases the process throughput. In this research, the fundamental interaction mechanisms of moisture with spin-on porous methylsilsesquioxane (p-MSQ) and Black Diamond IIx (BDIIx) dielectric films were investigated through isothermal challenge-purge processes at different exposure environments. Mass spectrometers (APIMS and EIMS), and cavity ring-down spectroscopy were used to detect moisture concentration in the gas phase. The moisture concentration in the thin films was directly analyzed by Fourier transform infrared spectroscopy (FTIR). Transmission Electron Microscope (TEM) micrographs were used to evaluate how patterning processes change the films. Moisture solubility, impact of temperature and gas flow rate on moisture removal, and dynamics of moisture uptake and removal in the films were determined by experimental study.Two process models were developed. The first one is capable of predicting the dynamic aspects of moisture adsorption and desorption in the films, and the second one is used to predict dry-down of moisture-contaminated gas delivery systems. The parameters in the models, such as moisture solubility and diffusivity in the films and rate constants of adsorption and desorption on the surface of the electro-polished stainless steel tube, were extracted through fitting these models to the experimental data. The models can be used to optimize key operating conditions such as purge temperature, purge gas purity, and purge gas flow rate. The models are also valuable tools for developing efficient contamination control strategies and process recipes for contamination removal in porous low-k dielectric films and for minimizing the gas usage in gas delivery systems