Thin Solid Films 427 (2003) 411–416 0040-6090/03/$ - see front matter 2002 Elsevier Science B.V. All rights reserved. doi:10.1016/S0040-6090(02)01190-2 High uniformity deposition with chemical beams in high vacuum G. Benvenuti, E. Halary-Wagner,A. Brioude, P. Hoffmann* Institut d’Imagerie et d’Optique Appliquee (IOA), Ecole Polytechnique Federale de Lausanne, CH-1015 Lausanne, EPFL, Switzerland ´ ´´ Abstract A mathematical model is applied to build a compact reactor for uniform thickness deposition in the molecular beam regime. The injector system uses a gas source and is compatible with perpendicular light illumination of the substrate for patterned deposition (etching). Titanium dioxide deposition is achieved and experimental results are compared to the mathematical model. Thickness uniformity better than 2% is experimentally achieved on a 150-mm diameter substrate and is compared to the 1% calculated. This approach allows a compact reactor design and an easy up grading in deposition area size and illumination optics design for light assisted processes. 2002 Elsevier Science B.V. All rights reserved. Keywords: Light assisted deposition; Molecular beams; Uniform thickness; Large area deposition; Compact reactor 1. Introduction The business and technology of large area vacuum coatings has made significant progress in the last two decades and is still proceeding with great expectation for the future w1,2x. Typical applications that require very controlled and uniform thickness deposition on large areas are for example antireflective coatings for architectural and automotive glasses, or solar cells. Another large field with higher added value are appli- cations in micro-opto-electronics. Inside this field, important aspects of light assisted processes are selective deposition (etching),reducedandlocallylimitedthermal damage, modification of materials properties (i.e. den- sification), or increased growth (etching) rates. Deposition of organic molecules, followed by local photo-polymerization also opens interesting possibilities in future polymer devices. Most of the problems encountered both in CVD and MBE technologies related to molecule transport can however be solved with a very promising hybrid tech- nique consisting in molecular beams of CVD precursors w3x (chemical beam epitaxy, MOMBE, or gas source MBE (GSMBE)). In particular, due to the high molec- ular mean free path, boundary layers are absent allowing *Corresponding author. Tel.: q41-216936018; fax: q41- 216933701. E-mail address: patrik.hoffmann@epfl.ch (P. Hoffmann). pulse repetition rates to be increased without depletion effects due to diffusion-limited reactions as reported in a previous publication for higher pressures w4x. Further advantages of chemical beams are to avoid gas phase reactions and light absorption and the simplicity of exact modeling of molecular distributions on the substrate. Molecular beams were extensively studied in the last century. Examples of parameters optimization by ana- lytical modeling w5,6x, Monte Carlo simulations w7–9x, and experimental work w10–12x result in good agree- ment. Very good agreement is usually achieved between experiments, simulations and analytical modeling if adequate precautions are taken w13–15x. In particular, the gas inlet system, consisting in one or more effusive sources with given relative position compared to the deposition area, can be assumed as the unique parameter influencing molecular impinging rate distribution on the deposition area w12,16x. However, one of the difficulties in the molecular flow regime is to achieve simultane- ously both high thickness uniformity and reduced size of the reactor to have a cost effective reactor w1x. In this work, GSMBE is investigated to achieve high uniformitythicknessdepositionwithfulllightprocessing compatibility. Mathematical calculations have been done to construct a relatively compact reactor (55 l) able to achieve thickness uniformity of 1% on a 150-mm diameter substrate. Taking advantage of gas sources, a system composed of several sources is investigated.