In-situ phosphorus-doped polysilicon prepared using rapid-thermal anneal (RTA) and its application for polysilicon passivated-contact solar cells

https://doi.org/10.1016/j.solmat.2020.110518Get rights and content

Highlights

  • N-type polysilicon passivated contact structure annealed by Rapid-Thermal Anneal (RTA) is studied.

  • The effects of the annealing temperature, annealing time, cooling time, polysilicon thickness on surface passivation are investigated.

  • The whole crystallization period of RTA is reduced to ~15 min with the best iVoc of 727 mV and J0,s of 4.7 fA/cm2 after hydrogenation.

  • Limiting polysilicon thickness to less than 40 nm helps to avoid blistering.

  • The polysilicon passivated-contact solar cell prepared using RTA shows a champion conversion efficiency of 23.04%.

Abstract

A rapid thermal anneal (RTA) is used to crystallize the plasma-enhanced chemical vapor deposition (PECVD) deposited hydrogenated amorphous silicon (a-Si:H) thin film to form the phosphorus-doped polysilicon passivated contact in tunnel oxide passivated contact (TOPCon) solar cells. The effects of annealing temperature, annealing time, cooling time, and the polysilicon thickness on the surface passivation are investigated. The primary advantage of the RTA is reducing the whole crystallization period to ~15 min, shorter than the conventional tube-furnace annealing period of >60 min. We find that the RTA is a robust method to prepare high-quality polysilicon passivated contact without introducing blistering when the thickness of the a-Si:H is less than 40 nm. The optimized RTA process leads to an implied open-circuit voltage (iVoc) of 712 mV and a single-sided dark saturation current density (J0,s) of 12.5 fA/cm2 in the as-annealed state, which is inferior to the surface passivation of the controlled one prepared by a tube furnace annealing. Fortunately, a subsequent Al2O3 capping hydrogenation improves the iVoc and J0,s to 727 mV and 4.7 fA/cm2, respectively. The champion conversion efficiency of 23.04% (Voc = 679.0 mV, Jsc = 41.97 mA/cm2 and FF = 80.86%) is achieved, which demonstrates the effectiveness of RTA for preparing a high-efficiency polysilicon passivated-contact solar cell.

Introduction

Polysilicon (poly-Si) passivated contact consisting of an ultrathin SiOx layer and a heavily-doped poly-Si layer, also named as TOPCon [1,2], POLO [3], monoPoly [4] or i-TOPCon [5], has demonstrated an excellent surface passivation for crystalline silicon (c-Si) solar cells [1]. The single-sided dark saturation current density (J0,s) of the n-type TOPCon structure has been improved down to 1 fA/cm2 [6]. A recent work from Trina shows that the poly-Si passivated contact technology promotes the average solar cell efficiency to >23% in 20000 pieces/day pilot line, which is better than that of ~22.25% of the passivated emitter and rear-contact (PERC) solar cells [5].

For the preparation of poly-Si, one approach is to crystallize a hydrogenated amorphous silicon (a-Si:H) deposited via PECVD with a post-annealing process and a proper hydrogenation treatment, among which tube-furnace annealing is commonly used and has been proven an effective and reliable method for fabricating the poly-Si passivated contact with an excellent surface passivation. Typically, a tube-furnace annealing requires more than 60 min to finish the whole crystallization procedure, including a loading process (~10 min), a temperature ramping up process (~20 min), a annealing process (~30 min), and a unloading process (~10 min), which is one of the potential bottlenecks for the mass production.

Developing a new annealing method to shorten the crystallization period while keeping a good surface passivation is desired for the TOPCon solar cell industrial manufacturing. The belt furnace firing was reported as a short-time annealing process, by which the passivated structure was renamed as “fired passivating contact” [7]. Rapid thermal anneal (RTA), featured with a fast heating rate supported by direct illumination of halogen lamps, is originally applied to recover the lattice damage after ion implantation in semiconductor industry [8,9], which has also been applied to solar cell fabrication, such as shallow diffusion [10], surface passivation [11], emitter formation [12], selective emitter (SE) formation [13], and amorphous silicon crystallization [14]. Therefore, the RTA would be one of the potential methods to fabricate the poly-Si passivated contact within a short processing time. Moreover, the way that the wafers are placed on the plate tray in a RTA system is similar to the placement of wafers on the tray of parallel plate PECVD, which makes the integration of the RTA and parallel plate PECVD into one in-line manufacture line feasible.

We employed the RTA to prepare the poly-Si passivated contact by transforming the PECVD deposited in-situ phosphorus-doped a-Si:H into poly-Si. The effects of annealing temperature, annealing time, cooling time, and subsequent hydrogenation were investigated. Besides, the in-diffusion phosphorus profile, the crystallinity, and the surface morphology of the poly-Si were also investigated. An excellent surface passivation of the RTA-processed passivated contact with the champion implied open-circuit voltage (iVoc) of 727 mV and single-sided dark saturation current density (J0,s) of 4.7 fA/cm2 is achieved after an Al2O3 capping hydrogenation [15]. Furthermore, we found that limiting the poly-Si thickness to less than 40 nm is helping to avoid blistering during a RTA process. To demonstrate its effectiveness, the poly-Si passivated-contact solar cell prepared using the RTA achieved a conversion efficiency of 23.04%. Our work suggests that the RTA is a potential method to fabricate the poly-Si passivated contact with a reduced process time in comparison with the conventional tube-furnace annealing.

Section snippets

Experimental

The 180 μm, 1–3 Ω cm, <100>-oriented n-type Czochralski silicon wafers were used as the substrate for the lifetime samples. Both sides of the wafer were subjected to acidic polishing to remove saw damages, followed by the standard RCA clean and the native SiOx was removed by dipping in the diluted HF. An ultrathin silicon-oxide (SiOx) layer with the thickness of ~1.4 ± 0.1 nm was grown in an 110 °C nitric acid (68 wt%) bath. Afterward, a 40–120 nm thick phosphorus-doped a-Si:H layer was

Results and discussion

Both the iVoc and J0,s of the RTA-processed lifetime samples as a function of the annealing temperature ranging from 780 °C to 900 °C without and with the Al2O3 capping hydrogenation are given in Fig. 1(a and b). The surface passivation is improved with the increase of annealing temperature and reaches the peak value at 820 °C, but it starts to decrease significantly once over 880 °C. The champion iVoc and J0,s of the as-annealed lifetime sample reach 712 mV and 12.5 fA/cm2 at 820 °C, and a

Summary

In this work, the effectiveness of the RTA process for preparing high-quality poly-Si passivated contact is investigated. The effects of annealing temperature, annealing time, cooling time, and the poly-Si thickness on the surface passivation are all taken into account. The advantage of the RTA process is reducing the whole crystallization period to ~15 min maintaining the high-quality passivation of poly-Si passivated contacts. The surface passivation of the RTA as-annealed passivated contact

CRediT authorship contribution statement

Qing Yang: Investigation, Data curation, Formal analysis, Writing - original draft, Writing - review & editing. Mingdun Liao: Investigation. Zhixue Wang: Investigation. Jingming Zheng: Investigation. Yiran lin: Investigation. Xueqi Guo: Investigation. Zhe Rui: Investigation. Dandan Huang: Investigation. Linna Lu: Investigation. Mengmeng Feng: Investigation. Peihong Cheng: Resources. Chunhui Shou: Funding acquisition. Yuheng Zeng: Conceptualization, Formal analysis, Writing - original draft,

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was supported by the National Key R&D Program of China (Grant No. 2018YFB1500403), Zhejiang Energy Group (Project No. znkj-2018-118), National Natural Science Foundation of China (61974178, 61974149, 61574145, 51601210, 61704176, 61874177), Zhejiang Provincial Natural Science Foundation (LY19F040002, LR19E020001), Key Research and Development Program of Zhejiang Province (2019C01080), Ningbo “Innovation 2025” Major Project (2018B10050), Key Project of Zhejiang Province (No. 2018C01034

References (27)

Cited by (0)

1

Co-first author: Qing Yang, Mingdun Liao.

View full text