Estimation of the Upper Limit of the Minority-Carrier Diffusion Length in Multicrystalline Silicon: Limitation of the Action of Gettering and Passivation on Dislocations
Estimation of the Upper Limit of the Minority-Carrier Diffusion Length in Multicrystalline Silicon: Limitation of the Action of Gettering and Passivation on Dislocations
复制标题
DOI：
10.4028/www.scientific.net/ssp.95-96.197
复制
发表时间：
2003-09
期刊：
影响因子：
--
通讯作者：
M. Kittler;W. Seifert
中科院分区：
文献类型：
--
作者：
M. Kittler;W. Seifert
We discuss the effect of gettering and hydrogen passivation on the recombina tion activity of contaminated dislocations. It is demonstrated that a residual amount of dislocation states in the order of 10 cm remains active even under optimum processing conditions, which sets an upper limit to the diffusion length in dislocated solar-grade silicon. Bas ed on an analysis of DLTS investigations on misfit dislocations, we show that impurities may either be accommodated in the dislocation core or in a cloud around the dislocation. We suggest that the f orm r are at the origin of the limited efficiency of gettering and passivation. Introduction Dislocations are the defect species in multi-crystalline si licon that has the main influence on the solar cell performance. The minority-carrier diffusion length, L, is determined by the dislocation density (ranging between 10 4 and 10 cm) and their contamination with impurities/metals. Gettering (e.g. phosphorus diffusion gettering during the emitter formation or Al gettering from the back side) and hydrogen passivation techniques are used in solar cell technology to reduce the electrical activity of crystal defects/dislocations to enhance the cel l fficiency. EBIC investigations revealed that the very small intrinsic ac tivity of clean dislocations in Si cannot be restored by neither gettering nor passivation [1,2]. At le ast about 10 -10 impurities/metal atoms per cm dislocation length remain active. The cause of t his limitation is not understood so far. In this paper we deduce the cause of this limit from observations made by DLTS on model defects/misfit dislocations which are localized in a well-defi ned depth position [3]. Taking into account the limits of gettering/passivation we estimate the upper limit of the diffusion length as a function of the dislocation density by using a model that describes t h recombination of contaminated dislocations [4]. It will be shown that the calculat ion is in a good agreement with experimental diffusion length data reported in the literature [5]. The weak intrinsic activity of clean dislocations in Si unrecoverable by neither gettering nor passivation Fig. 1 shows the experimentally observed temperature dependence of the EBIC contrast of clean misfit dislocations which appeared at the interface of a few μm thick SiGe(2%)/Si stacks grown by CVD. The clean dislocations exhibit only very weak intrinsic recombination activity, with a maximum at about 50 K and untraceable activity at room temperature. Contamination by metals leads to a substantial increase of contrast/activity, see e.g . Fig. 2. The c(T) behaviour shown in the upper part of Fig. 2 corresponds to contamination with a density of met al impurities/deep levels of about 3 x 10 per cm dislocation length. The lower part of Fig. 2 represents the c(T) behaviour fter Solid State Phenomena Online: 2003-09-30 ISSN: 1662-9779, Vols. 95-96, pp 197-204 doi:10.4028/www.scientific.net/SSP.95-96.197 © 2004 Trans Tech Publications Ltd, Switzerland All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of Trans Tech Publications Ltd, www.scientific.net. (Semanticscholar.org-18/03/20,13:30:52) phosphorus diffusion gettering, corresponding to a contamination of about 3 x 10 per cm dislocation length. That means that the density of impurities at t he dislocations was reduced by a factor of 100, but gettering was not able to restore the very small intrinsic acti vity. As a general rule for dislocations in different materials (CZ and FZ Si, mc-Si, misfit dislocations in Si/SiGe stacks) we observed that at least about 10-10 impurities/deep levels per cm dislocation length remain electrically active after extrinsi c gettering and/or hydrogen passivation of contaminated dislocations, e.g. [1,2,6]. Consequently, the effectiveness of ge tterin and passivation is limited and does not allow to restore the very weak intrinsic activity of clean dislocations. The cause of this ‘magic limit’ was not understood so far. However, our r ecent observations made by DLTS on misfit dislocations [3] enable us to propose an explanation for this limitation now . 0 20 40 60 80 100 120 140 160 0.2 0.4 0.6 0.8 1.0
