Influence of Al₂O₃/SiN_x Rear-Side Stacked Passivation on the Performance of Polycrystalline PERC Solar Cells

Abstract

In recent years, polycrystalline passivated emitter and rear cell (PERC) solar cells have developed rapidly, but less research has been conducted on the preparation process of their rear side passivation layers on standard solar cell production lines. In this work, a Al₂O₃/SiN_x rear side stacked passivation layer for polycrystalline PERC solar cells was prepared using the plasma- enhanced chemical vapor deposition (PECVD) method. The effects of different Al₂O₃ layer thicknesses (6.8~25.6 nm), SiN_x layer thicknesses (65~150 nm) and SiN_x refractive indices (2.0~2.2) on the passivation effect and electrical performance were systematically investigated, which were adjusted by TMA flow rate, conveyor belt speed and the flow ratio of SiH₄ and NH₃, respectively. In addition, external quantum efficiency (EQE) and elevated temperature-induced degradation experiments were also carried out to check the cell performance. The results showed that the best passivation effect was achieved at 10.8 nm Al₂O₃ layer, 120 nm SiN_x layer and 2.2 SiN_x layer refractive index. Under the optimal conditions mentioned above, the highest efficiency was 19.20%, corresponding V_oc was 647 mV, I_sc was 9.21 A and FF was 79.18%. Meanwhile, when the refraction index was 2.2, the EQE of the cell in the long-wavelength band (800–1000 nm) was improved. Moreover, the decrease in conversion efficiency after 45 h LeTID was around 0.55% under the different refraction indices. The above results can provide a reference for the industrial production of polycrystalline PERC solar cells.

1. Introduction

Efficiency improvement and cost reduction are two major development directions in the photovoltaic (PV) industry. Decreasing the thickness of silicon wafers is an important means of reducing costs.

Table 1. Quotation for N-Type 210 mm silicon wafers of different thicknesses.

Date	Price of Silicon Wafers with Different Thicknesses ($)		Price Difference ($)
	130 μm	110 μm
9 August 2023	0.538	0.517	0.021
29 July 2023	0.502	0.482	0.020
9 July 2023	0.478	0.459	0.019
1 June 2023	0.686	0.657	0.029
11 May 2023	0.834	0.801	0.033

shows the prices of two wafers with common thicknesses of 110 μm and 130 μm, using N-Type 210 mm silicon wafers from TCL Zhonghuan Renewable Energy Technology Co., Ltd. (Tianjing, China). On the whole, the quotation for the thickness of 110 μm is less than 130 μm.

However, reducing the thickness of silicon wafers may make the minority carrier diffusion length greater than the thickness of the silicon wafer, resulting in more minority carriers diffusing to the back surface to be recombined and at the same time increasing the transmittance of long-wave photons, ultimately leading to a decrease in conversion efficiency. Thus, ensuring or even improving the conversion efficiency of solar cells while reducing costs is the top priority of this current work.

Thanks to the development of surface passivation technology [1,2], the PERC solar cell has become a mainstream product in the PV industry. PREC is based on the production of conventional solar cells by depositing a layer or stack of passivation layers (currently usually Al₂O₃/SiN_x stack) using the PECVD method on the back surface instead of a conventional aluminum backfield [3]. Because of the high negative charge density of 10¹²–10¹³ cm⁻² in the Al₂O₃ passivation layer [4], carriers in the underlying silicon decrease, resulting in the reduction in surface recombination velocity. So, Al₂O₃ is an excellent kind of field-effect passivation material, which has been confirmed by previous research [5,6,7,8]. However, to improve the firing stability, we often deposit the SiN_x layer on Al₂O₃. Although Al₂O₃ and SiN_x have different charge polarities, stacking of SiN_x does not alter the negative charges of Al₂O₃. In addition, the capping layer of SiN_x on rear-side stacked passivation of Al₂O₃/SiN_x could protect the passivation layers from being damaged by metal pastes and increase the reflection of light, resulting in better optical generation [9,10,11,12].

Being able to obtain high-performance solar cells at a low cost has always been a development goal of the photovoltaic industry. PERC can offer an opportunity to achieve that, considering both its low-cost manufacture and high conversion efficiency [13]. Recent research papers have reported that rear-side passivation film parameters have a significant impact on the electrical performance of PERC. However, on one hand, most of them were small-scale experiments in laboratories and not batch experiments on standard solar cell production lines [8]. On the other hand, to improve solar cell performance, many researchers used monocrystalline silicon rather than polycrystalline silicon [14,15], which would lead to higher costs.

In this study, we carried out our experiment on a standard solar cell production line. We adjusted the TMA flow rate, conveyor belt speed and silane ratio to optimize the relevant parameters of Al₂O₃ and SiN_x, exploring their impact on the performance of polycrystalline PERC.

2. Experimental Section

2.1. Preparation of PERC Solar Cell

A batch of boron-doped P-type polycrystalline silicon wafers was selected to be used in this experiment. The silicon wafers were sliced using diamond wire-cutting technology, with a thickness of 180 ± 20 μm, a size of 156 mm × 156 mm and a resistivity of 1~3 Ω·cm.

The front side of the silicon wafers was first processed using a 10~20% NaOH solutions and then textured with a HF/HNO₃ solution. After that, the textured wafers were diffused with POCl₃, followed by a wet chemical etching process to remove phosphosilicate glass (PSG) and P-N junctions at the edges, resulting in a resistance of 110~130 Ω. The junction depth was about 3 μm. Then, the Al₂O₃ layer was deposited on the rear side and the SiN_x layer was deposited on both sides of the silicon wafers by PECVD in 2in 1 MAiA (Meyer Burger Technology AG, Thun, Switzerland). The thickness and refractive index of the front side SiN_x layer were kept at 75 nm and 2.0, respectively. On the rear side, Al₂O₃/SiN_x rear-side stacked passivation was ablated by a laser (525 nm, DR) to form local openings. The dot-to-dot distance was 36 mm, and the line-to-line distance was 900 mm. Finally, standard screen printing was used to screen-print silver and aluminum on the front and the rear surfaces of wafers, which formed good ohmic contact through the firing process. Figure 1 shows the schematic of the process for the fabrication of PERC in this work.

2.2. Experimental Methods

To investigate the influence of the rear side of the Al₂O₃, SiN_x layer thickness and SiN_x refractive index on the electrical performance of polycrystalline PERC cells, we adjusted these parameters by controlling the flow rate of the TMA in the Al₂O₃ process chamber, the belt speed in the SiN_x process chamber and the SiH₄/NH₃ flow ratios in the SiN_x process chamber of MAiA, which is a rear-side passivation equipment. According to the above method, solar cells with different thicknesses of Al₂O₃ layer, SiN_x layer and SiN_x refractive index were fabricated. Then, the effects of Al₂O₃ layer thickness (6.8, 10.8, 16.0, 20.5 and 25.6 nm), SiN_x layer thicknesses (65, 80, 100, 120 and 150 nm) and SiN_x refractive index (2.0, 2.1 and 2.2) were evaluated. All experimental results in this article were set up with five parallel samples and the presented results were the average of those five experiments.

In the LeTID degradation experiment, the solar cells were put into a pre-degradation test chamber, with a light intensity of 800 W/m², degradation temperature of 70 °C and accumulative degradation time of 45 h. Cells’ electrical performance parameters were measured using a German Halm tester under standard testing conditions after each degradation step, which were 2, 7, 17 and 45 h. It should be tested immediately when the temperature of the sample drops to 25 °C.

2.3. Characterization Methods

The passivization layer thickness and refractive index were measured by a TESCAN GES5E laser ellipsometer (TESCAN ORSAY HOLDING, a.s. Czech, Europe). The external quantum efficiency (EQE) of solar cells was investigated using the solar cell measurement system of Bentham PVE300 (Suzhou Vision Intelligence Technology Co. Ltd., Suzhou, China). The effective minority carrier lifetime mapping of passivated samples was characterized by equipment of WCT-120 (Sinton consulting Inc. Boulder, CO, USA). The electrical performance of solar cells was measured by a German Halm tester (h.a.l.m. Elektronik GmbH, Frankfurt, Germany) under standard testing conditions (AM1.5G, 1000 W/m², 25 ± 2 °C).

3. Results and Discussion

3.1. Effect of Al₂O₃ Layer Thickness on the Passivation and Electrical Performance

The Al₂O₃ film is the first film to be prepared in the Al₂O₃/SiN_x rear-side stacked passivation layer of polycrystalline PERC solar cells. The raw materials for preparing Al₂O₃ are TMA, N₂O and Ar. Ar is a kind of inert gas that does not participate in chemical reactions and mainly plays a blowing role. TMA reacts with N₂O in the reaction chamber to generate Al₂O₃ film, which is deposited on the silicon wafer substrate. To examine the influence of the Al₂O₃ layer thickness on minority carrier lifetime and electrical performance, a series of wafer samples with different Al₂O₃ layer thicknesses were fabricated by adjusting the TMA flow rate and maintaining the N₂O flow rate.

Table 2. Correspondence between TMA flow rate and Al₂O₃ layer thickness.

TMA flow rate (mg/min)	300	420	540	650	760
Al2O3 layer thickness (nm)	6.8	10.8	16.0	20.5	25.6

shows the correspondence between the TMA flow rate and Al₂O₃ layer thickness. It can be seen that by increasing the TMA flow rate within a certain range, which was from 300 to 760 mg/min, the Al₂O₃ layer thickness increased continuously and showed a linear relationship with TMA flow rate.

To investigate the effect of Al₂O₃ thickness on the minority carrier lifetime of PERC before and after sintering, we continue to deposit a layer of SiN_x film on the top of the Al₂O₃ film. The SiN_x layer thickness and the refractive index in this section were set to 110 nm and 2.05, respectively. Figure 2 shows the minority carrier lifetime before and after sintering of the polycrystalline PERC solar cell with different Al₂O₃ layer thicknesses. The difference in minority carrier lifetime before and after sintering was significant at the same thickness. The minority carrier lifetime was in the range of 73~106 μs before sintering and was in the range of 156~238 μs after sintering. The sintering process resulted in an increase in minority carrier lifetime and the increment of that was from 83 to 132 μs under different Al₂O₃ layer thicknesses. When the film thickness was 10.8 nm, the increase in minority carrier lifetime was the largest, which was 132 μs.

The reason to explain the occurrence of the above phenomenon is that hydrogen atoms released from the Al₂O₃ layer can easily cross the Al₂O₃ layer and then enter into the silicon substrate to passivate the hanging bonds on the silicon surface. However, as the thickness of the Al₂O₃ layer was larger than 10.8 nm, the passivation effect would be unsatisfactory, as the diffusion of hydrogen atoms to the silicon bulk is blocked. At the same time, high-temperature sintering would promote the occurrence of this process. The above conclusions are consistent with the previous results [16,17]. Therefore, an appropriate thickness of Al₂O₃ film is necessary to achieve an excellent passivation effect. Based on comprehensive analysis, a TMA flow rate of 420 mg/min and Al₂O₃ film thickness of 10.8 nm are the most suitable parameters.

Table 3. Electrical performance parameters of polycrystalline PERC cells with different Al₂O₃ layer thickness.

Al2O3 Layer Thickness (nm)	Voc (V)	Isc (A)	FF	η (%)
6.8	0.6506	9.19	79.84	19.43
10.8	0.6519	9.2053	79.74	19.48
16	0.6495	9.1916	79.76	19.38
20.5	0.6484	9.1849	79.74	19.33
25.6	0.6468	9.1449	79.56	19.15

shows the electrical performance parameters of the polycrystalline PERC solar cells with different Al₂O₃ layer thicknesses. As can be seen from the table, the conversion efficiency showed a trend of increasing firstly and then decreasing with the thickness of Al₂O₃ increasing. The best performance of the solar cells was achieved when the film thickness was 10.8 nm. Meanwhile, the highest efficiency was 19.48%, corresponding V_oc was 0.6519 V, I_sc was 9.2053 A and FF was 79.74. The overall trend is consistent with Figure 2, which indicates that the change in electrical performance of polycrystalline PERC cells with different Al₂O₃ layer thicknesses is mainly caused by the difference in the effect of backside passivation. Therefore, for polycrystalline PERC, the passivation effect and electrical performance were optimal when the thickness of the Al₂O₃ layer was 10.8 nm.

3.2. Effect of SiN_x Layer Thickness on the Passivation and Electrical Performance

In this section, the substrate was first deposited with an Al₂O₃ layer of the optimum thickness of 10.8 nm, which was learned through Section 3.1. Then, the wafer substrate was transported to the process chamber via a conveyor belt on MAiA equipment to complete the SiN_x deposition using the PECVD method. From

Table 4. Correspondence between the conveyor belt speed and SiN_x layer thickness.

Conveyor belt speed (cm/min)	180	220	260	380	460
SiNx layer thickness (nm)	150	120	100	80	65

we can see that different SiN_x layer thicknesses of 65, 80, 100, 120 and 150 nm can be obtained by changing the conveyor belt speed from 180 to 460 cm/min and maintaining constant process parameters such as SiH₄ and NH₃ flow ratio, microwave power, temperature and deposition rate in the SiN_x process chamber. Due to the constant silane ratio, the refractive index remained unchanged and was still set to 2.05. From Table 4, we can see that the faster the conveyor belt speed, the thinner the SiN_x layer.

Although a single layer of Al₂O₃ can already provide a good passivation effect, its thermal stability is poor. Covering a layer of SiN_x can not only protect the Al₂O₃ layer to improve the thermal stability but also increase the thickness of the back surface layer and reduce light transmission. Similar to Al₂O₃, the thickness of SiN_x also affects the efficiency of the polycrystalline PERC solar cells. Figure 3 shows the minority carrier lifetime before and after sintering with different SiN_x layer thicknesses. It can be seen from the figure that there were some differences in minority carrier lifetime before sintering with different SiN_x layer thicknesses. Before sintering, minority carrier lifetimes of the wafers were between 60 μs and 79 μs, while after sintering, the minority carrier lifetime of the wafers improved to 153~219 μs rapidly, which was more than twice as much as the cells without sintering. When the SiN_x film thickness was 120 nm, the passivation effect was optimal.

To a certain extent, the thicker the rear SiN_x, the better the passivation effect. At the same time, more photons would be reflected to the p-n junction if the photons were to reach the P-type silicon substrate and owing to that, more photon-generated carriers would be produced. However, if the SiN_x layer were to be too thick, the probability of the photon being reflected to p-n would decrease because the thicker SiN_x layer would have more internal defects within its body. Furthermore, this would result in a decrease in the efficiency of polycrystalline PERC solar cells. On the other hand, if the SiN_x layer were to be too thin, the Al paste would locally fire through the SiN_x layer during sintering. Then, additional parasitic Al contacts would be formed which would increase contact recombination, resulting in a decrease in the conversion efficiency of solar cells. As a consequence, an appropriate thickness of SiN_x film is also necessary to achieve an excellent passivation effect. Based on comprehensive analysis, a conveyor belt speed of 220 cm/min and SiN_x film thickness of 120 nm are the most suitable parameters.

Table 5. Electrical performance parameters of polycrystalline PERC cells under different SiN_x.

SiNx Layer Thickness (nm)	Voc (V)	Isc (A)	FF	η (%)
65	0.6468	9.2266	78.96	19.18
80	0.649	9.2264	79.09	19.27
100	0.6503	9.2322	78.95	19.29
120	0.6497	9.2547	79.02	19.34
150	0.6489	9.2018	79.21	19.25

shows the electrical performance parameters of polycrystalline PERC solar cells with different SiN_x layer thicknesses. The conversion efficiency showed a trend of increase before decrease, which was consistent with Figure 3. Therefore, for PERC with Al₂O₃/SiN_x rear-side stacked passivation layers, the optimal thickness of SiN_x was 120 nm. At this point, the conversion efficiency was 19.34%, and the corresponding V_oc, I_sc and FF were 0.6497 V, 9.2547 A and 79.02, respectively.

3.3. Effect of Refractive Index on the Passivation and Electrical Performance

In this section, the SiN_x and Al₂O₃ layer thicknesses on the rear side of PERC were kept constant at 120 nm and 10.8 nm, respectively. By adjusting the flow ratio of SiH₄ and NH₃, SiN_x layers with different refractive indices were prepared, which were 2.0, 2.1 and 2.2. From Figure 4, we can see that the minority carrier lifetime increased with the increase in the refractive index. Before sintering, the minority carrier lifetime was between 51 and 63 μs. After sintering, the minority carrier lifetime increased rapidly and was almost three times that, which was between 152 and 185 μs. The maximum value of the minority carrier lifetime after sintering reached 185 μs at a SiN_x refractive index of 2.2.

With the SiN_x layer thickness remaining constant, increasing the flow ratio of SiH₄ and NH₃ would increase the refractive index of SiN_x, which would increase the Si content in the SiN_x layer. Furthermore, the concentration of the Si-H bond was also found to increase. As is known to all, Si-rich SiN_x films have a higher hydrogen content, thus increasing the number of Si-H bonding, resulting in improving the passivation effect [18,19,20]. However, the refractive index cannot be too large because the greater the difference between the refractive index of the SiN_x and air, the higher the reflectivity will be, which was found in previous studies [21].

Table 6. Electrical performance parameters of polycrystalline PERC solar cells with different SiN_x refractive indices.

SiNx Refractive Index	Voc (V)	Isc (A)	FF	η (%)
2	0.6408	9.0769	79.1	18.73
2.1	0.6449	9.1637	79.1	19.03
2.2	0.647	9.21	79.18	19.2

shows the electrical performance parameters of the polycrystalline PERC solar cells with different SiN_x refractive indices. With the increase of the SiN_x refractive index from 2.0 to 2.2, the conversion efficiency was changed from 18.73% to 19.20%. At the same time, V_oc increased from 0.6408 V to 0.6470 V, I_sc increased from 9.0769 A to 9.2100 A and FF increased from 79.10 to 79.18. The electrical performance parameters at the 2.2 refractive index are the most excellent. On the one hand, an increase in the refractive index of the backside SiN_x would improve the passivation effect. On the other hand, a higher refractive index would reduce the light transmission and increase the light reflection on the backside. So, increasing the refractive index of the backside SiN_x passivation layer within a certain range is beneficial to improving the electrical performance of the polycrystalline PERC solar cell.

PERC solar cells with different SiN_x refractive indices (2.0, 2.1 and 2.2), which were taken from the above experiment, were used to explore the effect of the refractive index on EQE. The EQE results shown in Figure 5 indicate that solar cells exhibited an overall general increase in EQE with different SiN_x refractive indices in the spectral measurement range of 300~730 nm, and the increasing trend was consistent. However, EQE reduced in the spectral range from 730 nm to 1100 nm. From Figure 5, we can see that the largest EQE was 96.18% instead of 100% at wavelengths of 650 nm, which could be attributed to reflection, front surface recombination, bulk recombination and so on. Further, optical losses for the longer wavelength (>650 nm) photons result from poor light-trapping architecture at the rear side. The increase in refractive index can improve the quantum efficiency of polycrystalline silicon PERC batteries in the long wavelength range (800~1000 nm), which was consistent with the increase in electrical performance parameters in Table 6.

The cells with different SiN_x refractive indices obtained in 3.3 were used for light and elevated temperature degradation experiments. Figure 6 shows the trend of the conversion efficiency of polycrystalline PERC cells under 45 h LeTID with different SiN_x refractive indices. From the figure, we can see that conversion efficiency of cells with different refractive indices declined quickly within the first 10 h, then the degradation rate gradually slowed down and finally became stable. The conversion efficiency of cells with a refractive index of 2.2 decreased from 19.20% to 18.66% after 45 h degradation, with a decrease of 0.54%. Meanwhile, the conversion efficiency of cells with 2.0 and 2.1 refractive indices decreased by 0.57% and 0.61%, respectively. On the whole, the degradation pattern of solar cell samples with different refractive indices is consistent and the conversion efficiency was still the highest at 2.2 SiN_x refractive index after 45 h of degradation. The results showed that the different rear-side passivation conditions of SiN_x refractive index basically do not affect the photovoltaic degradation of the polycrystalline PERC cells. Under the different refraction indices, the decrease in conversion efficiency was around 0.55%.

Until now, we have optimized the relevant parameters for the Al₂O₃/SiN_x rear-side stacked passivation of PERC through the standard solar cell production line. To compare with other relevant research results, the optimal conversion efficiency of polycrystalline PERC batteries in other studies is shown in

Table 7. Comparison of conversion efficiency of polycrystalline PERC in different research.

Source	η	Experimental Subjects and Scale	Ref.
Sastrawan et al.	18.00	Polycrystalline PERC, industrial pilot line	[22]
Zhao et al.	18.63	Polycrystalline PERC, bench scale experiment	[23]
Wang et al.	18.61	Polycrystalline PERC, bench scale experiment	[24]
Liu et al.	18.42	Polycrystalline PERC, bench scale experiment	[25]
In our research	19.20	Polycrystalline PERC, standard solar cell production line	-

. The optimal conversion efficiency in our research was 19.20%, which can be seen in Table 6, under the optimized coating parameters of a 10.8 nm Al₂O₃ layer, 120 nm SiN_x layer and 2.2 SiN_x layer refractive index. Within the limited statistics, we can see that our polycrystalline PERC solar cell has certain advantages in conversion efficiency. In addition, due to our optimization about the parameters of Al₂O₃/SiN_x, the stacked passivation film on the rear side of the solar cell was completed on a standard solar cell production line, so the optimized parameters in our research can be directly applied to batch production.

4. Conclusions

The parameters of Al₂O₃ and SiN_x in the rear side passivation layer has a large influence on the electrical performance of the polycrystalline PERC cell. The best rear side passivation effect and electrical performance of the polycrystalline PERC cell were achieved when the Al₂O₃ layer thickness was 10.8 nm, the SiN_x layer thickness was 120 nm and the SiN_x layer refraction index was 2.2. Under the optimal conditions mentioned above, the highest efficiency was 19.20%, corresponding V_oc was 647 mV, I_sc was 9.21 A and FF was 79.18. Furthermore, the increase in refractive index can improve the quantum efficiency of polycrystalline silicon PERC batteries in the long wavelength range (800~1000 nm). The decrease in conversion efficiency after 45 h light LeTID was around 0.55% under the different refraction indices and the conversion efficiency was still the highest at 2.2 SiN_x refractive index after 45 h of degradation.