Tunnel oxide passivated contacts as an alternative to partial rear contacts

Highlights

• Passivated rear contacts as an appealing alternative to partial rear contact cells.

• Outline of the on-going development of passivated rear contact for n-type cells.

• Champion cell efficiency: 24.4 %.

• Passivated contact for p-type cells is investigated in detail.

• Performance studied on prototype cells (high V_oc and FF).

Abstract

Recently, n-type Si solar cells featuring a passivated rear contact, called TOPCon (Tunnel Oxide Passivated Contact) were reported. The high conversion efficiency of 24.4% and very high FF>82% demonstrates that the efficiency potential of this full-area passivated rear contact is as good as or even better than that of partial rear contact (PRC) schemes like PERL (passivated emitter and rear locally diffused) and in addition avoids complex structuring steps and features a 1D carrier transport. Likewise, a boron-doped passivated rear contact for p-type solar cells (p-TOPCon) is proposed as an alternative to p-PRC cells. The optimum device design of PRC cells has to account for two opposing effects: a low-loss 3D carrier transport requires a high base doping but Shockley–Read–Hall (SRH) recombination within the base due to the formation of boron–oxygen complexes in standard Cz silicon calls for a low base doping level. This conflict might be overcome by p-TOPCon because its performance is less sensitive to base doping. This will be discussed on the base of experimental results. It is shown that its high implied fill factor (iFF) of 84% combined with the 1D carrier transport in the base translates into a higher FF potential. First investigations on planar solar cells prove the good performance of the p-TOPCon with respect to passivation and carrier transport. A V_oc of 694 mV and a FF of 81% underline the efficiency potential of this rear contact.

Introduction

Today the industry is working towards the implementation of PRC [1] schemes in production. Recombination at the local metal-semiconductor contacts is a major source of device recombination which is significantly reduced by simply decreasing the metallized area fraction. However, V_oc and FF losses – the latter arise in particular from the 3D carrier transport at the rear (Fig. 1, left) – have to be carefully balanced by adjusting the pitch of the point contacts. For industrial p-type Cz silicon device optimization is even more complex because a low series resistance R_S, which requires high base doping for an efficient 3D transport, comes at the expense of significant SRH recombination caused by light-induced degradation (LID) [2]. Thus, the formation of boron–oxygen complexes constitutes a major efficiency limitation for these devices.

An appealing alternative to a point contact structure is a full-area passivated contact which decouples the absorber’s passivation from the local metallization. In 1985 Lindholm et al. showed for the first time that heavily-doped polysilicon contacts can reduce the recombination at the rear side to some extent [3]. The benefit of applying polysilicon contacts to Si solar cells in terms of V_oc, i.e. J_0e was also demonstrated by others [4], [5], [6] and was recently revisited by Borden et al. (p⁺-polysilicon/c-Si(n) junction) [7]. In contrast to a-Si:H based heterojunctions the polysilicon contacts are a viable option for conventional solar cells due to their higher tolerance to high-temperature processes. For instance, in Ref. [3] the polysilicon contacts were first realized on the rear side, then capped by a protective layer, and finally exposed to a diffusion process forming the front emitter.

In this work so-called tunnel oxide passivated contacts (TOPCon) for n-type as well as p-type Si solar cells are discussed. The TOPCon structure resembles the polysilicon contacts with deliberately grown interfacial oxide. In contrast to the polysilicon contacts, the TOPCon structure employs a wide bandgap semiconductor layer which contains amorphous and crystalline Si phases. For more details on the contact’s morphology and structure as well as its implications on the blue response (parasitic absorption) the reader is kindly referred to Ref. [8].

The paper first addresses the passivated electron contact which was recently disclosed [9]. Due to its excellent carrier selectivity it enables high V_oc and FF at the same time which was demonstrated on an n-type Si solar cell featuring a boron-diffused emitter and the phosphorus (P)-doped passivated rear contact. In addition, its simple device design enables a 1D current flow in the base and thereby avoids FF losses originating from a 3D carrier transport (Fig. 1, right). Furthermore, it showed a high thermal stability and therefore could be integrated into a conventional diffused solar cell. Likewise, a boron (B)-doped passivated contact for p-type Si solar cells is proposed as an alternative to p-PRC cells which is investigated in more detail. Due to its one-dimensional device design, it might enable the use of lowly-doped wafers. Therefore, p-type cells featuring this passivated rear contact might be less prone to light-induced degradation. To this end, the interface passivation of p-TOPCon is studied on highly and lowly-doped p-type Si. Especially the implied FF (iFF), which describes the injection-level dependence of the passivation, will be discussed and compared to rear side passivation layers typically applied to p-PRC cells. Thereafter, the dark band bending, φ₀, induced by the passivated contact in the c-Si base is probed by means of the surface photovoltage (SPV) technique. It is a measure for the generated built-in potential V_bi of the p-TOPCon/c-Si(n) junction and can be used as an indicator for sufficient doping of the hole contact. Third, the passivated rear contact is integrated into a solar cell featuring n-TOPCon as emitter to facilitate the demonstration of its high V_oc and FF potential.