Can interface charge enhance selectivity in tunnel layer passivated contacts? Using negatively charged aluminium oxide capped with dopant free PEDOT or boron doped polysilicon

Highlights

• YES, interface charge can enhance selectivity, as shown for low-temperature, dopant-free PEDOT:PSS passivated contacts.

• ‘negative charge’ AlO_x/PEDOT:PSS beats conventional SiO_x/PEDOT:PSS passivated contacts.

• Estimated efficiency potential is 1.1% higher (26% as compared to 24.9%).

• AlO_x/p⁺-poly-Si does not beat conventional SiO_x/p⁺-poly-Si, due to high-temperature p⁺-poly-Si processing.

• Optimum thickness for ultra-thin AlO_x tunnel layer: 1.5 nm.

Abstract

Tunnel layer passivated contacts are widely used for selective extraction of electrons or holes in silicon solar cells. We investigate if a tunnel layer with high negative interface charge can further enhance the hole extraction ability (selectivity) of a passivated contact and thus the efficiency of silicon solar cells. High negative-charged, atomic layer deposited AlO_x dielectric tunnel layers, capped with either conventional boron doped polysilicon (p⁺-poly-Si) or low-temperature, dopant-free poly (3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS), are investigated in order to form hole selective passivated contacts. Ultra-thin AlO_x films in the tunnel regime (0.13–2 nm) are studied, measuring their passivation quality (τ_eff), fixed charge density (Q_tot) and defect density (D_it) prior to capping, and contact recombination density (J_0,contact) and effective contact resistivity (ρ_contact) after capping with either PEDOT:PSS or p⁺-poly-Si. The measured J_0,contact and ρ_contact values are then used to calculate carrier selectivity and efficiency potential if rear-side integrated in a solar cell.

Low temperature deposited PEDOT:PSS capped passivated contacts behave differently compared to high temperature deposited p⁺-poly-Si capped passivated contacts. Yes, negatively charged AlO_x tunnel layers can enhance hole extraction capability/selectivity of a passivated contact. For PEDOT:PSS capped passivated contacts, the efficiency potential deploying optimized negatively charged AlO_x tunnel layers (1.5 nm, annealed at 425 °C) was measured to be 1.1% higher than conventional reference-moderately charged SiO_x tunnel layer (26% as compared to 24.9%). However, this is not observed for p⁺-poly-Si capped passivated contacts due to boron in-diffusion and associated charge compensation mechanism, a consequence of the high-temperature diffusion process.

Introduction

Tunnel layer passivating contacts have proven to enable high energy conversion efficiencies for crystalline silicon (c-Si) solar cells at low processing complexity. Sandwiching an ultrathin dielectric tunnel layer (passivating the c-Si wafer interface) and a highly p-/n-doped or a high/low work function capping layer between the metal contact and the c-Si wafer results in reduced contact recombination and thus in improved solar cell efficiencies [[1], [2], [3], [4], [5], [6], [7]]. The tunnel oxide is the core element of a tunnel layer passivated contact. It must decrease minority carrier recombination near the contact (for instance, electron recombination in case of a hole extracting contact as studied here, Fig. 1.), leading to lower recombination current density J_contact, and simultaneously allow majority carriers (holes in our case) to pass through it, resulting in a low effective contact resistivity ρ_contact (corresponding to a high hole current J_holes in our case studied), see Fig. 1. In order to form ‘conventional’ SiO_x tunnel layers, different oxidation technologies like applying a standard RCA oxide [8,9], wet-chemical HNO₃ [1,[10], [11], [12]], a dry-grown UV/O₃ oxide [13], thermal oxides [10,11], etc., were studied. Using ultra-thin atomic layer deposited (ALD) aluminium oxide (Al₂O₃) tunnel layers, exhibiting a high negative charge density (~10¹² cm⁻²), as opposed to SiO_x tunnel layers, which exhibit only a moderate positive fixed charge density (~10¹⁰ cm⁻²), might enhance the excess hole extraction (selectivity) of the passivated contact formed as shown in Fig. 1 (as the negative tunnel layer charge in Al₂O₃ films leads to an enhanced accumulation of holes nearby the contact). ALD Al₂O₃ layers not only provide the benefit of high field effect passivation, but also improved chemical passivation by decreasing the interface defect density, thereby making it a promising candidate for tunnelling passivation layer for solar cell contacts in MIS structure [14] or when coupled with separate carrier selective material like doped polysilicon [[15], [16], [17]], molybdenum oxide [18], zinc oxide [19] or even high work function materials [20]. But does high negative interface charge of AlO_x actually improve hole selectivity and subsequently solar cell efficiency?

In order to test this, we compared two different capping layers suited for selective hole extraction: dopant-free poly (3,4-ethylenedioxythiophene) polystyrene sulphonate (PEDOT:PSS, an organic, low-cost, high work function material) and the conventionally used capping layer, i.e. boron doped polycrystalline silicon (p⁺-poly-Si). Please note, PEDOT:PSS is processed at low temperature (spin coating at room temperature, followed by an anneal at ~140 °C), whereas p⁺-poly-Si deposition is a high temperature process (we used low pressure chemical vapour deposition (LPCVD) at ~570 °C and subsequent boron diffusion at ~800–850 °C). As low temperature capping layer processing would likely not alter the tunnel layer properties, this will allow us to better understand the effect of tunnel layer variation without facing additional problems of high temperature processing as encountered for polysilicon based contacts. ‘Conventional’ SiO_x/p⁺-poly-Si [21,22] and SiO_x/PEDOT:PSS [23,24] passivated contacts, as well as ‘negatively charged’ AlO_x/p⁺-poly-Si [15,17] and AlO_x/PEDOT:PSS [20,25] passivated contacts are already realized and have resulted in cell efficiencies above 20%. In this work, we systematically investigate whether the presence of a high negative fixed charge density at the c-Si/AlO_x interface helps to selectively extract holes in the above mentioned hole selective passivated contact configurations.

Xin et al. [15] replaced the SiO_x tunnel layer beneath the p⁺ poly-Si by AlO_x tunnel layer to achieve improved hole selectivity for thicknesses < 1 nm [15]. However, the negative charge in AlO_x films is typically located within the first 1 nm of the c-Si/AlO_x interface [26] and thickness of the AlO_x tunnel layers plays a crucial role here. Since, it is important to grow tunnel layers with precise thickness control therefore, we used industrially relevant, high throughput spatial ALD tool to deposit AlO_x films. The passivation properties of ALD AlO_x tunnel layers with varying thicknesses were studied before and after annealing at 425 °C in air ambient (as charge formation in AlO_x layers takes place upon annealing). After a subsequent PEDOT:PSS or p⁺-poly-Si capping, the resulting hole extracting passivated contacts were further characterized, i.e. the contact recombination parameter (J_0,contact) and the effective contact resistivity (ρ_contact) were measured. Finally, the selectivity of the processed passivated contacts and their efficiency potential if deployed rear-side within a solar cell were calculated according to Brendel's model [27]. Overall, two hole-selective passivated contacts, AlO_x/PEDOT:PSS and AlO_x/p⁺-poly-Si deploying negative charge at interface were optimized as a function of ALD AlO_x thickness (0.13–2 nm) and compared to their ‘conventional’ reference system SiO_x/PEDOT:PSS and SiO_x/p⁺-poly-Si.