Problems in Large Size Amorphous Silicon Plate Manufacturing
- PDF / 779,033 Bytes
- 12 Pages / 420.48 x 639 pts Page_size
- 49 Downloads / 212 Views
PROBLEMS IN LARGE SIZE AMORPHOUS SILICON PLATE MANUFACTURING J.P.M. SCOPM1T SOLEMS S.A. 3, rue l,_n Blum, 91120 Palaiseau, France. ABSTRACT The growing aSi:H based component industry needs large PECVD machines for high throughputcoating offlat substrates up to 1 meter in size. Deposition uniformity can be affected by gas depletion in diffusion and laminargas flow geometries. Deposition rate is shown to be in conflict with large size uniformity. Main uniformity problems are found to originatefrom uneven RF voltage distribution, the effect is shown to be amplified near plasma equilibrium transitions. Contamination, a key factor for production quality, is controlled by the hot pressurized plasma box by both reducing unwanted background gas near UHV level and by allowing systematic fast plasma cleaning without machine corrosion or post contamination. Powderformation and dynamic is discussed based on several novel experiments and powder formation threshold is shown to be pushed away if the plasma is made intermittent.
INTRODUCTION Amorphous silicon and its electronic grade PECVD derivatives such as Silicon Nitride and alloys are used for the synthesis of a wide range of large size components ranging from solar photovoltaic modules to active matrix LCD displays [ 1,2]. All these techniques speculate on the ability of PECVD to deposit on large panels made of relatively low cost material such as glass (thanks to the moderate process temperature). The growing market for such products requires PECVD production machines which shall be able to coat uniformly flat plates up to 1 meter in size at rates larger than 10 plate/hour without allowing yield problems related to film uniformity or contamination. This paper is focussing on some aspects of a large collective effort regrouping SOLEMS and its German sister company PST, two equipment manufacturers, ALCATEL-NEXTRAL and PFEIFFER-BALZERS, both licencees of the plasma box design and several universities such as Ecole Polytechnique in Palaiseau (J. Perrin, P. Roca, A. Lloret...), Universities of Orl6ans (A. Bouchoule...), Toulouse (J.P. Boeuf, J.P. Couderc...), Barcelona (E. Bertran...) and EPF from Lausanne (A. Howling, Ch. Hollenstein,...). All these teams were supported by government funding (AFME, BMFT, DG XWI) for the design of photovoltaic module mass production and by the EEC (BR1TE-EURAM) for the specific aspect of powder formation. Many interesting and difficult problems are to be solved in order to optimize the design of large PECVD machines. The deposition system architecture (single chamber or multichamber) in itself requires a very sophisticated analysis involving the device design and special interface processing [3,4]. Even for a very simple geometry such as the RF plasma capacitor, a full set of scaling rules must be carefully respected in designing a family of PECVD machines from the lab size to the full size in order to allow straightforward process transfer [5]. A minimum understanding of the basic mechanisms for the plasma, the gas phase and the surface chemis
Data Loading...
