p-Doping limit and donor compensation in CdTe polycrystalline thin film solar cells

Experimental evidence shows that non-shallow acceptor states defect complex VCd ClTe 0= and Cu substitution of Cd CuCd 0= play critical roles in p-doping of CdTe in CdS/CdTe thin film solar cells. In this work, two equations are presented by using graphic method, one to determine the limit of p-doping or hole density for such non-shallow acceptor levels, and another to show the quantitative relationship of n-type donor compensation of p-type acceptors in such a material. & 2010 Published by Elsevier B.V. CdTe based solar cells have emerged in recent years as the most cost effective second generation thin film photovoltaic (PV) product by far. Despite intensive effort of research and development (R&D), however, the ratio of CdS/CdTe solar cell’s lab demonstrated light to electricity conversion efficiency (16.5%) [1] vs its theoretical limit has been low, in comparison with other thin film PV technologies, such as CIGS, which has a lower theoretical limit but a higher lab demonstrated efficiency of 19.9% [2]. One of the reasons is that the hole concentration of p-CdTe keeps in the range of 10 –10 cm , instead of the desired level of 10–10 cm 3 [3], resulting in lower junction band bending and difficulty in making ohmic contact. Both effects contribute to a lower open circuit voltage VOC and therefore lower efficiency. Traditional semiconductors, such as Si or GaAs, have a single shallow donor (for n-type) or acceptor (for p-type) impurity level, with typical activation energy r0.05 eV from the band edge. From the local charge neutrality (LCN) condition we have pþND 1 1þgD exp ðEF ED=kTÞ 1⁄4 nþNA 1 1þgA exp ðEA EF=kTÞ ð1Þ where n, p, EF, ED, EA, ND, NA, gD, and gA are electron density, hole density, Fermi level, donor activation energy, acceptor activation energy, donor density, acceptor density, donor degeneracy, and acceptor degeneracy, respectively. For widely used tetrahedral cubic semiconductors, such as Si, GaAs, and CdTe, gD1⁄42 due to electron spin degeneracy, and gA1⁄44 due to heavy hole and light hole in addition to spin degeneracy. For traditional non-degenerate p-type material used in electron devices and integrated circuits, Eq. (1) gives p1⁄4NV exp EV EF kT NA 1 1þgA exp ðEA EF=kTÞ NA ð2Þ under shallow doping condition EF EAckT 1⁄4 0:0259eV ð3Þ where NV is the effective density of states in valence band, EV the valence band maximum (VBM), and T1⁄4300 K, room temperature. The relationship of (2) and (3) are shown in Fig. 1. As for the CdTe polycrystalline thin film in a CdS/CdTe solar cell, it is still being debated: what exactly happens during its p-doping. Yet, it is generally agreed that the Cd vacancy VCd 0= (or the defect complex, VCd ClTe 0= ) and the impurity Cu substitution of Cd CuCd 0= are the responsible p-doping acceptors. Since the Cd vacancy and the defect complex have almost the same measured activation energy [4], we do not distinguish between them. High density of Cd vacancies is naturally formed—which may hinder the formation of other shallow acceptors—during the 5N CdTe crystal growth and polycrystalline CdTe thin film deposition by closed space sublimation (CSS) or other methods. The double acceptor Cd vacancy may turn to defect complex during CdCl2 treatment of the CdS/CdTe junction at 400 1C [1]. Cu also naturally exists ARTICLE IN PRESS