Monte Carlo computer simulation of the deposition of CdTe thin films by close-spaced sublimation

Citation
N. Romeo et al., Monte Carlo computer simulation of the deposition of CdTe thin films by close-spaced sublimation, MATER CH PH, 66(2-3), 2000, pp. 259-265
Citations number
15
Categorie Soggetti
Apllied Physucs/Condensed Matter/Materiales Science","Material Science & Engineering
Journal title
MATERIALS CHEMISTRY AND PHYSICS
ISSN journal
02540584 → ACNP
Volume
66
Issue
2-3
Year of publication
2000
Pages
259 - 265
Database
ISI
SICI code
0254-0584(20001016)66:2-3<259:MCCSOT>2.0.ZU;2-3
Abstract
Thin film solar cells based on the CdS/CdTe heterojunction, where the CdTe polycrystalline layer is deposited making use of the close-spaced sublimati on (CSS) technique have attained a very good level of conversion efficiency in recent times. Despite this apparently consolidated situation the debate on the formation of the CdS/CdTe heterojunction is still open indicating t hat a certain margin of improvement is still possible. In particular, the c onclusion that a CdSxTe1-x alloy is formed at the CdS/CdTe interface due to interdiffusion that takes place during fabrication. The extent of the inte rdiffusion and the value of x depend on many factors such as deposition tem perature, gas pressure and geometrical parameters. The situation is even mo re complicated because while the deposition is in progress. the ratio betwe en the number of Te and the Cd atoms which are reaching the target is invar iably greater than one. To shed some light on these issues it is mandatory to know at least the exa ct Te/Cd ratio as a function of the deposition parameters. To this end we c arried out a Monte Carlo computer simulation of the CdTe deposition in our CSS equipment. The computer code was implemented starting from very simple hypotheses simulating the deposition in perfect gas at constant temperature . Subsequently, we added more physical reality including temperature gradie nt between source and target. The code was then finally optimized introduci ng the usual null-event method which confirmed the consistency of our resul ts. (C) 2000 Elsevier Science S.A. All rights reserved.