Analysis of the double-layer α-Si:H emitter with different doping concentrations for α-Si:H/c-Si heterojunction solar cells

Abstract

Double-layer emitters with different doping concentrations (DLE) have been designed and prepared for amorphous silicon/crystalline silicon (α-Si:H/c-Si) heterojunction solar cells. Compared with the traditional single layer emitter, both the experiment and the simulation (AFORS-HET, http://www.paper.edu.cn/html/releasepaper/2014/04/282/) prove that the double-layer emitter increases the short circuit current of the cells significantly. Based on the quantum efficiency (QE) results and the current-voltage-temperature analysis, the mechanism for the experimental results above has been investigated. The possible reasons for the increased current include the enhancement of the QE in the short wavelength range, the increase of the tunneling probability of the current transport and the decrease of the activation energy of the emitter layers.

Graphical abstract

Keywords

double-layer emitter / α-Si:H/c-Si heterojunction solar cell / short circuit current / quantum efficiency / current-voltage-temperature

1 Introduction

Since the a-Si:H/c-Si heterojunction structure (HJT) was used for photovoltaic devices, the structure has been improved for many times. The pioneering work, which is called heterojunction with intrinsic layers (HIT) [1], was done by Sanyo. The essential feature for this structure is the bifacial heterojunction structure with a back surface field [2]. Solar cells with efficiencies of up to 24.7% have been achieved for this structure [3]. After that a-Si:H/c-Si heterojunction solar cells with other structures emerged, including the inverted a-Si:H/c-Si heterojunction structure [4], the interdigitated back-contact HIT-structure [5,6] and silicon thin-film heterojunction solar cells [7].

The double layer emitter with different doping concentrations is also called high-low junction emitter [8] or front surface field [9], which has been used on the p-n junction silicon solar cells. Zhong, et al. [10] proved that a (heavy doped a-Si:H)/(heavy doped c-Si) emitter structure could decrease the series resistance of a-Si:H/c-Si heterojunction solar cells compared with the (heavy doped a-Si:H)/(intrinsic a-Si:H) emitter structure. In this paper, a structure of high-low doped a-Si:H layers (Fig. 1), called the double layer emitter a-Si:H/c-Si heterojunction solar cell (DLE-HJT solar cell) has been proposed for increasing the efficiency of the HJT solar cell. The device has been simulated by the AFORS-HET software [11]. It is shown that for this DLE structure an increase of 0.5% absolute efficiency can be achieved compared with the conventional single layer emitter a-Si:H/c-Si heterojunction solar cell (SLE-HJT solar cell), which could be caused by the higher drift velocity at the front side of the device and the lower contact resistance at the ITO/a-Si:H interface. Besides, the DLE-HJT solar cell has been fabricated. The solar cells has been analyzed and compared with the conventional SLE-HJT solar cell.

2 Experimental

Figure 1 is a comparison of the cross sections of a double layer emitter HJT solar cell and a conventional HJT solar cell.

40 mm× 40 mm n-type Czochralski monocry stalline Si wafers with (100) orientation were used to fabricate solar cells with a thickness of 180 mm and a resistivity of 2 to 5 W·cm. First, the wafers were cleaned by acetone and alcohol. Then, they were etched by ~25 mm on each side using the 20% KOH (wt) solution at 85°C. After that, they were cleaned by the standard SC1 solution (80 mL H₂O₂ + 80 mL NH₄OH+ 400 mL H₂O, 80°C) for 10 min and SC2 solution (80 mL H₂O₂ + 80 mL HCl+ 400 mL H₂O, 80°C) for 10 min. Finally, they were dipped into HF solution (2% (vol)) for one minute to remove the native oxide layers. The wafers are dried by nitrogen gas before they are put into the chamber of a thermal-evaporation setup.

To fabricate the solar cells, an Al film was first deposited on the back side of the wafer by evaporation. Then, the wafer was treated by rapid-thermal-annealing process to form a back-surface-field. After that, the oxide layer on the front surface of the wafer was removed by using HF solution (5% (vol)). The wafer was then transferred to a plasma-enhanced chemical vapor deposition (PECVD) equipment to deposit the intrinsic layer and the doped amorphous silicon (a-Si:H) layers successively. Finally, ITO layers and Ag grids were prepared by sputtering and evaporation, respectively. For the preparation of ITO and Ag grids, different stainless steel masks were used. A solar cell fabricated is shown in Fig. 2. The size of each sub-cell is 10 mm× 10 mm. Except the preparation parameters for the doped a-Si:H layers by PECVD, other parameters for the solar cell fabrication were kept the same.

Two series of the samples were made. One series includes solar cells with single layer emitters with a thickness of 15 nm but with different doping concentrations. The other series includes solar cells with double layer emitters (the thicknesses of the shallow doped layer and the heavy doped layer are 12 nm and 3 nm, respectively). As listed in Table 1, SiH₄, H₂ and PH₃ were used as the source gases for the emitter deposition. The doping concentration of the emitter was changed by the flow rate of PH₃ during the deposition process.

A sun-simulator (AM1.5G, 3A Grade) and an Agilent B2901Ameter were used to measure the light current-voltage (I-V) curves of the samples. Enlitech QE-R3018 was used to test the external quantum efficiency (QE) spectra of the solar cells. The dark current-voltage-temperature (IVT) method was performed to analyse the electrical properties of the solar cells.

3 Results and discussion

As illustrated in Fig. 3, compared with SLE, the biggest improvement of the I-V properties of the samples with DLE are the short circuit current (I_sc). There exists no significant difference in open circuit voltage (V_oc) between the samples. These experimental results are in coincide with the simulation results [11].

Figure 4 demonstrates the QE results of the best cells with DLE and SLE by simulation and by experiments. According to the simulation analysis by AFORS-HET, the increase in I_sc comes from the increase in the quantum efficiency in the short wavelength range. The simulation by AFORS-HET shows that obvious significant increase in QE can be achieved in the short wavelength range by replacing the SLE by DLE. However, the QE measurement of the real samples reveals that the improvement in QE at short wavelength range is not as much as that in simulation.

From the dark I-V test (100–300 K), the ideal factors of the samples were calculated and depicted in Fig. 5. According to Ref. [12], the ideal factor of a semiconductor junction reflects the main mechanism of carriers transport. As shown in Fig. 5, the main mechanism of the carriers transport changes to “tunneling” when the doping concentration of the emitter layers increases. The ideal factor of the solar cells with high doping concentration emitter layers is even higher than 2.0 at the temperature of 100 K to 300 K. This agrees with the previous analysis that the contact resistance at the ITO/a-Si:H interface decreases with increasing the doping concentration of the emitter [11]. Therefore, the difference in the I-V curve and the EQE should be more pronounced.

With the fixed voltage analysis of the IVT result of the samples, the activation energy (E_a) of the devices at different applied voltage was obtained. The result of one sample is displayed in Fig. 6(a). It is found that one of the samples has higher values (E_a1), mostly larger than 0.2 eV at applied voltage= 0 V, while the other sample has a lower value (E_a2), lower than 0.1eV at applied voltage= 0 V. Figure 6(b) and (c) exhibits the E_a1 and E_a2 of all the samples, respectively. No treatment such as smooth was made for the results depicted in the figure. From Fig. 6(b), it can be seen that the peak value of the E_a1-V curves has a rough trend with the total doping concentration to the emitter layer (or layers) of a sample. The peak value decreases as the doping concentration to the emitter layer (or layers) increase. From Fig. 6(c), it can be observed that the peak value of the E_a2-V curves is not related to the doping concentration of the emitter layer (or layers). For E_a1 values could be affected by the doping concentration of the emitter layers, it is believed that E_a1 is the activation energy of the a-Si:H emitter layers. Low activation energy would be beneficial to the carriers transport in the emitter layer, which would also contribute to the improvement of I_sc and transfer efficiency of a solar cell. E_a1 should be the activation energy of the c-Si wafer.

4 Conclusions

A structure of high-low doped a-Si:H layers was used as the emitter of a-Si:H/c-Si heterojunction solar cells. Two series of HJT solar cells with DLE and SLE were prepared. According to the experimental result, the DLE structure could improve the I _sc of HJT cells compared with the SLE structure. The mechanisms for this improvement include the improvement of the QE in the short wavelength range, the increase of the tunneling probability of the current transport and the decrease of E _a1 in solar cells.

References

[1] Tanaka M, Taguchi M, Matsuyama T, Sawada T, Tsuda S, Nakano S, Hanafusa H, Kuwano Y. Development of new α-Si/c-Si heterojunction solar cells: ACJ-HIT (Artificially Constructed Junction-Heterojunction with Intrinsic Thin-Layer). Japanese Journal of Applied Physics, 1992, 31(11): 3518–3522

[2] Sawada T, Terada N, Tsuge S, Baba T, Takahama T, Wakisaka K, Tsuda S, Nakano S. High-efficiency α-Si/c-Si heterojunction solar cell. Conference Record of IEEE Phot. Spec. Conf., 1994, 2(2): 1219–1226

[3] Gifford J. Panasonic hits 24.7% cell efficiency. 2013–04

[4] Wünsch F, Citarella G, Abdallah O, Kunst M J. An inverted a-Si:H/c-Si hetero-junction for solar energy conversion. Journal of Non-Crystalline Solids, 2006, 352(9-20): 1962–1966

[5] Johnson J N, Manivannan V. Photovoltaic device which includes all-back-contact configuration and related fabrication process. US Patent: 2008/0000522 A1, 2008–01–03

[6] Terakawa A, Asaumi T. Photovoltaic cell and method of fabricating the same. US Patent: 20050062041A1, 2005–03–24

[7] Zhang L, Shen H L, Yue Z H, Jiang F, Wu T R, Pan Y Y. Effect of emitter layer doping concentration on the performance of silicon thin film heterojunction solar cell. Chinese Physics B, 2013, 22(1): 016803

[8] Fossum J G, Lindholm F A, Sah C T. Physics underlying improved efficiency of high-low-junction emittersilicon solar cells. International Electronic Devices Meeting, 1977, 23: 222–225

[9] Zouari A, Arab A B. Effect of the front surface field on crystalline silicon solar cell efficiency. Renewable Energy, 2011, 36(6): 1663–1670

[10] Zhong S, Hua X, Shen W. Simulation of high-efficiency crystalline silicon solar cells with homo–hetero junctions. IEEE Transactions on Electron Devices Year, 2013, 60(7): 2104–2110

[11] Huang H B, Gao J, Fahrner W R, Zhou L. Theoretical analysis of the double-layer emitter with different doping concentrations for α-Si:H/c-Si heterojunction solar cells. 2014

[12] Suzue K, Mohammad S N, Fan Z F, Kim W, Aktas O, Botchkarev A E, Morkoç H. Electrical conduction in platinum-gallium nitride Schottky diodes. Journal of Applied Physics, 1996, 80(8): 4467–4478

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg