AVS 45th International Symposium
    Plasma Science and Technology Division Thursday Sessions
       Session PS-ThM

Paper PS-ThM8
Simulations for Process Optimization Issues in Ionized Metal PVD

Thursday, November 5, 1998, 10:40 am, Room 318/319/320

Session: Plasma Applications in Copper Metallization
Presenter: P.L.G. Ventzek, Motorola Inc.
Authors: P.L.G. Ventzek, Motorola Inc.
M. Hartig, Motorola Inc.
V. Arunachalam, Motorola Inc.
D.G. Coronell, Motorola Inc.
D. Denning, Motorola Inc.
Correspondent: Click to Email
Magnetron plasma sources for ionized metal plasma physical vapor deposition (IMP-PVD) are being exploited for metal deposition in semiconductor device manufacturing applications because of their ability to lay down films with the required step coverage at high rates and reasonable uniformity. Challenges exist to extend the technology to ever more aggressive features with tighter tolerances on process parameters (e.g. uniformity). Despite their importance, multidimensional numerical models that treat the plasma dynamics in IMP chambers are relatively rare. This is possibly due to the difficulty in considering the complex magnet arrangements that characterize these systems. In this paper, we will present results obtained using the Hybrid Plasma Equipment Model (HPEM)@footnote 1@ and MAXWELL3D@footnote 2@ to simulate the behavior of a generic IMP source for metal deposition. This generic source employs internal coils to ionize metal sputtered from a target in a dc magnetron. As is typical, the dc magnetron structure is azimuthally asymmetric and will consist of concentric spirals of oppositely oriented permanent magnets. A special feature of the model is that energy and angle dependencies in the sputter yield are considered and that we have developed a methodology for looking at 3- dimensional effects.@footnote 3@ First we will present design-of-experiment studies of various process parameters and magnet configurations. In general, reasonable agreement has been found between ion fluxes to the target and experimentally determined erosion profiles. Second we will illustrate how various process parameters affect across wafer thickness uniformity, compositional uniformity and step coverage. @FootnoteText@ @Footnote 1@M. Grapperhaus et al., J. Appl. Phys., 83, 39 (1998) @Footnote 2@Ansoft Corp. @Footnote 3@see AVS paper by J. Kress et al.