Defect concentration and Δn change in light- and elevated temperature-induced degradation

• Published in: Journal of physics. D, Applied physics Applied physics, 2022-02, Vol.55 (8), p.85106
• Main Authors: Kim, Moonyong, Wright, Matthew, Chen, Daniel, Chan, Catherine, Ciesla, Alison, Abbott, Malcolm, Hallam, Brett
• Format: Article
• Language: English
• Subjects: boron–oxygen defects; crystalline silicon solar cells; defects in silicon; light- and elevated temperature-induced degradation; light-induced degradation
• ISSN: 0022-3727; 1361-6463
• DOI: 10.1088/1361-6463/ac34a8

The wide variety of silicon materials used by various groups to investigate LeTID make it difficult to directly compare the defect concentrations ( N t ) using the typical normalised defect density (NDD) metric. Here, we propose a new formulation for a relative defect concentration ( β ) as a correction for NDD that allows flexibility to perform lifetime analysis at arbitrary injection levels (Δ n ), away from the required ratio between Δ n and the background doping density ( N dop ) for NDD of Δ n/N dop = 0.1. As such, β allows for a meaningful comparison of the maximum degradation extent between different samples in different studies and also gives a more accurate representative value to estimate the defect concentration. It also allows an extraction at the cross-over point in the undesirable presence of iron or flexibility to reduce the impact of modulation in surface passivation. Although the accurate determination of β at a given Δ n requires knowledge of the capture cross-section ratio ( k ), the injection-independent property of the β formulation allows a self-consistent determination of k . Experimental verification is also demonstrated for boron-oxygen related defects and LeTID defects, yielding k -values of 10.6 ± 3.2 and 30.7 ± 4.0, respectively, which are within the ranges reported in the literature. With this, when extracting the defect density at different Δ n ranging between 10 14 cm −3 to 10 15 cm −3 with N dop = 9.1 × 10 15 cm −3 , the error is less than 12% using β , allowing for a greatly improved understanding of the defect concentration in a material.