Vikram M. Bhosle

Value of thermal oxide for ion implant based c-Si solar cells

In this paper we present the improvements in cell performance for ion implanted C-Si solar cells by thermal oxidation. We have investigated the effect of oxidation on cell efficiency (CE) and systematically evaluated the effect of oxide thickness on emitter quality for the ion implanted emitters. It is shown that the CE of ion implanted cells increases by ~0.3% absolute with increase in oxide thickness and is attributed to lower J0e. Additionally, the effect of oxidation on dopant profile of these emitters is also characterized using sheet rho (Rsheet) measurements and SIMS analysis, which demonstrate that an optimal profile can be achieved with a single step anneal/oxidation of the ion implanted emitters. The lower J0e achieved for the ion implanted emitters with an oxide layer (10-30nm) is due to superior passivation quality of the thermal oxide and the optimized dopant profile. It is to be noted here that the thermal oxidation is part of the post implant anneal and does not lead to excessive drive in and allows controlling the dopant profile/junction depth (xj) very precisely. This unique ion implant based process flow also offers a distinct advantage over two step diffusion process(es) (POCl3 → PSG etch → high temperature thermal oxidation) which are required to passivate emitters fabricated using POCl3 with a thermal oxide. Thus, the proposed approach of single step anneal/oxidation of ion implanted emitters offers an attractive solution for passivating surfaces and manufacturing high efficiency solar cells with advanced architectures.

Leveraging patterned ion implantation to develop high efficiency selective emitter solar cells

This paper reports the benefits of using patterned ion implantation to create higher efficiency selective emitter solar cells. This doping approach uses in-situ masking to enable selective doping of the contact and field regions of the emitter in a single step. The cell efficiency benefits of implantation versus current POCl3 diffusion processes are also explained, highlighting the improved junction quality, the ability to perform single side doping and the elimination of a junction isolation step which reduces cell efficiency. The implanted selective emitter solar cell process described reduces cell cost per watt through a combination of higher cell efficiency, improved cell binning and fewer process steps. Alignment between the doped regions of the solar cell and the subsequent downstream metallization process is key to enabling this process for adoption in PV manufacturing. The benefits for process integration of the implanted process will be described including the ability to optically align to the se...

AlO x passivation of ion implanted boron emitters using batch AlO x process

In the present paper we highlight the results on passivation of the ion implanted boron emitters using a batch atomic layer deposition (ALD) - AlOx process. We demonstrate that through optimization of the batch ALD process we can achieve extremely low levels of surface recombination uniformly across both sides of multiple wafers, processed during a single run. The double-sided nature of the batch AlOx process is of particular value when passivating symmetric samples, which constitutes an essential step of development and optimization of ion implanted boron emitters. In this work we explore the effect of various ALD process parameters (deposition temperature, pulse and purge duration) on the coverage and passivation quality of the AlOx films using the stacks of symmetric samples, i.e. ion implanted B emitters and high lifetime n-type wafers. We find that passivation quality shows a strong dependence on the film coverage/thickness and is primarily driven by the trimethyl aluminum (TMA) pulse pressure. Additionally, we have investigated the effect of separation distance between the wafers and showed that we can achieve uniform coverage with separation distance as small as 700um. This parameter (aspect ratio) is critical as it can lead to precursor starvation during the batch ALD process. However, to maximize the throughput of the lab scale equipment, the distance between wafers needs to be minimized. Using the optimized ion implant conditions and batch AlOx process, we achieve an excellent value for implied Voc of ~688mV and emitter saturation current density J0e as low as 23fA/cm2 for a ~100Ω/sq implanted B emitter. Furthermore, we show that excellent inversion layer passivation can be achieved for lightly doped n-type wafers, with Seff<;5 cm/s using this process, which could make batch AlOx process an attractive option for fabricating high efficiency implanted n-type rear contact cells.

Fabricating high efficiency solar cells with high sheet resistance emitters by ion implantation and contact resistance modeling

We present improvements in c-Si solar cell performance for high sheet resistance (R_sheet) emitters fabricated by ion implantation. We have investigated the effect of sheet resistance (60-115 Ω/sq) on cell efficiency (CE) and also evaluated the effect of dopant profile shape on the contact resistance for the ion implanted emitters. High efficiency cells, with average CE>19.3%, can be achieved with ion implanted high R_sheet emitters (60-90 Ω/sq) using commercially available screen printed Ag paste. It is to be noted that the best results were obtained for those cells with emitter R_sheet ~ 70-75 Ω/sq, as the cell performance is limited by the FF, namely front contact resistance (R_c) for emitters with R_sheet > 75 Ω/sq. To better understand the effect of emitter R_sheet and the dopant profile on contact resistance we have used VSEs bottom-up physics-based Technology Computer-Aided Design (TCAD) model to simulate these experimental results. We found that the traditional model of evaluating R_c using the phosphorus surface concentration (N_s) does not accurately predict the increase in R_c and consequently the loss in FF for high R_sheet emitters. We propose an alternative approach to model R_c where the contact depth and its associated dopant concentration (N_d) is employed to calculate R_c. This contact depth is not necessarily zero and may lie below the original Si surface. Our simulated results show that the use of N_d at a depth of the order of 10s of nm below the Si surface leads to better agreement between the experimental and simulated R_series, FF and CE than assuming that the contact is made with Si at the original wafer surface. The implications of these findings with regards to emitter profile engineering via ion implantation and formulation of new pastes to lower R_c of high R_sheet emitters are discussed.