Reactive physical vapor deposition of TixAlyN: Integrated plasma-surface modeling characterization

• Published in: Journal of vacuum science & technology. A, Vacuum, surfaces, and films Vacuum, surfaces, and films, 2004-03, Vol.22 (2), p.264-271
• Main Authors: Zhang, Da, Schaeffer, J. K.
• Format: Article
• Language: English
• ISSN: 0734-2101; 1520-8559
• DOI: 10.1116/1.1641048

Reactive physical vapor deposition (RPVD) has been widely applied in the microelectronic industry for producing thin films. Fundamental understanding of RPVD mechanisms is needed for successful process development due to the high sensitivity of film properties on process conditions. An integrated plasma equipment-target nitridation modeling infrastructure for RPVD has therefore been developed to provide mechanistic insights and assist optimal process design. The target nitridation model computes target nitride coverage based on self-consistently derived plasma characteristics from the plasma equipment model; target sputter yields needed in the plasma equipment model are also self-consistently derived taking into account the yield-suppressing effect from nitridation. The integrated modeling infrastructure has been applied to investigating RPVD processing with a Ti0.8Al0.2 compound target and an Ar/N2 gas supply. It has been found that the process produces athermal metal neutrals as the primary deposition precursor. The metal stoichiometry in the deposited film is close to the target composition due to the predominance of athermal species in the flux that reaches the substrate. Correlations between process parameters (N2 flow, target power), plasma characteristics, surface conditions, and deposition kinetics have been studied with the model. The deposition process is characterized by two regimes when the N2 flow rate is varied. When N2 is dilute relative to argon, target nitride coverage increases rapidly with increasing N2 flow. The sputter yield and deposition rate consequently decrease. For less dilute N2 mixtures, the sputter yield and deposition rate are stable due to the saturation of target nitridation. With increasing target power, the electron density increases nearly linearly while the variation of N generation is much smaller. Target nitridation and its suppression of the sputter yield saturate at high N2 flow rendering these parameters insensitive to target power variation. The deposition rate, however, increases with target power as a result of the increased ion energy and flux at the target.