Challenges for ionized physical vapor deposition

jeffrey hopwood,dali mao
DOI: https://doi.org/10.1109/PLASMA.2004.1339998
2004-01-01
Abstract:Summary form only given. Vapor deposition using ionic species has several advantages over neutral vapor deposition methods. These include the ability to use the electric field within the plasma sheath to modify the energy and direction of the depositing particle. Energetic deposition allows one to produce thin film materials with tailored properties. The control of particle direction, on the other hand, allows for the deposition of materials into high aspect ratio microstructures, such as those encountered in modern integrated circuits. For these reasons, ionized physical vapor deposition (IPVD) is an important technology in the manufacturing of microelectronics as well as hard and decorative coatings. Ionized physical vapor deposition encompasses a broad array of plasma sources including laser heated plasma, cathodic and anodic arcs, high-power unbalanced sputtering magnetrons, inductively coupled plasma, and electron cyclotron resonance plasma. A brief review of the operation and performance of these sources is given. From this review, the basic plasma source requirements are developed. In particular, the minimum electron density, temperature, and neutral mean free path are all critical parameters that determine the ionization of the vapor stream. A particularly challenging aspect of IPVD is the deposition of compounds in high aspect ratio microstructures. The reactive deposition of materials such as TiN is aided by the intense plasma due to significant dissociation of the nitrogen molecule. The relative ionization fraction of metals and reactive gases, such as Ti and N/sub 2/, is substantially different due to the large gap in ionization potentials, however. As a result, the metal vapor is usually highly ionized if n/sub e//spl sim/10/sup 12/ cm/sup -3/, but the reactive gas remains only weakly ionized. Because of an anisotropic velocity distribution, metal ions are deposited directly into the bottom of microstructures, but isotropic neutral nitrogen atoms must - ndure multiple reflections within the structure to reach the bottom. Therefore, the stoichiometry of compound thin films is often metal-rich within deep microstructures. Experimental mass spectrometry, thin film characterization, and deposition modeling is used to illustrate this problem, and a few practical solutions are discussed.
What problem does this paper attempt to address?