Abstract

In this study, the efficiency of the multicrystalline was improved by inserting a two-step growth thermal oxide layer as the surface passivation layer. Two-step thermal oxidation process can reduce carrier recombination at the surface and improve cell efficiency. The first oxidation step had a growth temperature of 780°C, a growth time of 5 min, and with N2/O2 gas flow ratio 12 : 1. The second oxidation had a growth temperature of 750°C, growth time of 20 min, and under pure N2 gas environment. Carrier lifetime was increased to 15.45 μs, and reflectance was reduced 0.52% using the two-step growth method as compared to the conventional one-step growth oxide passivation method. Consequently, internal quantum efficiency of the solar cell increased 4.1%, and conversion efficiency increased 0.37%. These results demonstrate that the two-step thermal oxidation process is an efficient way to increase the efficiency of the multicrystalline silicon solar cells.

1. Introduction

Many methods have been proposed to improve multicrystalline silicon (mc-Si) solar cell efficiency. The mc-Si wafers usually contain significant amounts of defects and impurities which affect minority carrier lifetime and limit multicrystalline solar cell efficiency. Solar cell efficiency can also be affected by surface recombination of the carriers. Macdonald et al. [1] and Krotkus et al. [2] successfully enhanced mc-Si solar cell efficiency using several methods, including dry texturing, gettering, selective emitters, hydrogen surface passivation, and using of silicon nitride (SiNx) passivation layer on the emitter surface.

An enhanced lifetime of the photon-generated carriers has also been realized using SiO2, SiNx, and SiCNx films as surface passivation layers [35]. Jana et al. [6] showed the interface trap density (Dit) (2.52 × 1011 cm−2 eV−1) at the SiNx/Si interface. The growth of a high-quality SiO2 layer with low interface trap density (Dit) on silicon is an effective surface passivation method for the solar cells to further improve cell efficiency [79]. Derbali and Ezzaouia [10] reported that thermal SiO2 film is an effective surface passivation layer for highly phosphorus-doped silicon wafers. Ohtsuka et al. [11] showed the Dit (1.0 × 1011 cm−2 eV−1) at the SiO2/Si interface. Comparing Dit of without thermal oxidation (TO) and thermal oxidation (TO) was from 2.52 × 1011 cm−2 eV−1 to 1.0 × 1011 cm−2 eV−1. Kotipallia et al. [12] showed the Dit reduction at the silicon/dielectric interface by using hydrogen gas annealing to reduce the dangling bonds. Schultz et al. [13] showed the influence of thermal oxidation temperature on the Dit of the interface. High temperature (>1000°C) dry oxidation growth of the SiO2 film results in low Dit (1010 cm−2 eV−1) at the SiO2/Si interface, but the process may deteriorate bulk Si quality and reduce bulk carrier lifetime. Reducing the growth temperature to 800°C with wet oxidation (Si + 2H2O → SiO2 + 2H2) can increase the bulk carrier lifetime. The efficiency of the sample annealed at high temperature (1050°C) was 17.20%, while at lower temperature (800°C) was 17.8%. The open circuit voltage increased about 16 mV at lower temperature, this is because the low temperature (800°C) passivation has higher bulk lifetime than the high temperature (1050°C) passivation. But using this temperature with dry oxidation process has not been studied. Chen et al. [14] and Hiroshige et al. [15] showed that low temperature grown SiO2 layers also have low interface Dit with Si wafers when used for surface passivation. Cai and Rohatgi [16] showed that the most effective annealing temperature after PECVD for multicrystalline solar cell wafers was 650°C to 750°C.

Narasimha and Rohatgi [17] showed that growth of the SiO2 layer by rapid thermal oxidation had a low surface recombination velocity (S = 10 cm/s) on the silicon surface, and Chen et al. [18, 19] showed that a SiO2 layer deposited on Si wafer by PECVD followed by rapid thermal annealing could produce lower S and higher carrier lifetime. In this work, the use of lower temperature (750°C–780°C) two-step grown SiOx film as passivation layers using dry oxidation (Si + O2 → SiO2) process was performed and studied. The goal was to improve bulk Si crystalline quality and the bulk carrier lifetime after lower temperature thermal oxidation process and to compare the crystalline quality and carrier lifetime with high temperature (>1000°C) grown oxide. The efficiency enhancement of the mc-Si solar cells was observed after by applying silicon oxide passivation layer grown by two-step thermal oxidation (TO) method. The TO process influence on mc-Si solar cell electrical properties is also investigated and found that by proper design of the oxidation process, the solar cell electrical properties can be effectively improved.

2. Experimental

Figure 1(a) shows the sample structure of the solar cell in this study, fabricated on a p-type mc-Si wafer, and Figure 1(b) shows the fabrication process, consisting of Al/p-Si/n-Si/SiO2/SiNx/Ag layers. The fabrication process commenced with acid texturing, followed by a shallow diffusion layer formed by POCl3 phosphorus diffusion process, resulting in 82 Ω/sq sheet resistance. The phosphosilicate glass was then removed by the wet chemical method, and each wafer was divided into sample A (without TO process) and sample B (with TO process). A high-quality batch-type quartz tube furnace system was used for the two-step TO process, with a temperature profile as shown in Figure 2 to grow oxide on the wafer. The first oxidation step had a growth temperature of 780°C, with growth time 5 min, and N2/O2 gas flow ratio 12 : 1. The second oxidation step had a growth temperature of 750°C, growth time 20 min, and with pure N2 gas environment. To improve incident light absorption, PECVD SiNx film was deposited on both samples after oxidation.

Screen printing was employed to deposit Ag glue on the front side SiNx layer and AgAl glue on the back side of the wafer, with an intermediate drying process. Finally, the samples were annealed at 750–850°C to complete the solar cell fabrication.

3. Results and Discussion

Texturizing of the solar cell surface was performed with HF/HNO3/H2SO4 and deionized water mixture to form a randomly oriented micron-size texture. Figures 3(a) and 3(b) show plane-view optical microscopic (OM) images of pre- and postetched wafer surfaces at 100x magnification.

Schmidt and Aberle [20] showed that the carrier lifetime can be measured by the microwave photoconductance decay (Semilab system, Model WT-2000). Both samples were measured; Figures 4(a) and 4(b) show the carrier lifetime for samples A and B (with and without oxide layer, resp.), and average carrier lifetimes were 5.55 and 21 μs, respectively. Thus, carrier lifetime was significantly improved by the proposed TO process. Lee and Glunz [21] showed that improve carrier lifetime also improved the solar cell efficiency.

Surface reflectance was measured using a UV-VIS-NIR spectrophotometer (HMT, Model MFS-630) in the spectral range 350–1050 nm; after SiNx film deposition are shown in Figure 5(a) test data and summarized in Table 1. Average reflectances were 3.28% and 2.76% for samples A and B, respectively; sample B (with TO) had a lower reflectance of 0.52%. Lee et al. [22] showed that the solar cell absorption band was in the range 400–1200 nm. Thus, the lower average reflectance of sample B results in higher efficiency of the solar cell. Meziani et al. [23] showed that reflectance were 2.77% and 0.94% at 600–800 nm range for SiNx and SiNx/SiO2 passivation, respectively. The refractive index (n) of SiNx and SiO2 were 2 and 1.45, respectively. Thus, adding SiO2 film of low refractive index (n) can decrease reflectance of the light (Figure 6).

Internal quantum efficiency (IQE) was measured using a Solar Cell Scan 100 quantum efficiency measurement system (Model QEX10) at 350–1050 nm spectral range as shown in Figure 5(b), and the results are summarized in Table 1. Average IQE data were 84.62% and 88.72% for samples A and B, respectively, an increase of 4.1% for sample B.

The IQE of a solar cell with and without TO passivation can provide information on the passivation effect from the surface and bulk recombinations. For wavelengths < 800 nm, IQE provides information related to S, while for wavelengths > 800 nm, IQE provides information related to τ. Figure 5(b) shows that for sample B, IQE increased mostly in the spectral range 600–1050 nm, with its reflectance decrease over that range due to the SiOx/SiNx antireflection coating. Thus, the treatments effectively increased light absorption of the solar cell.

Thus, sample B has enhanced IQE over sample A. Since the major IQE enhancement is in the 600–1050 nm range, it is mainly due to reduced reflectance.

Morales-Acevedo and Pérez-Sánchez [24] showed that improve efficiency of without TO and with TO (SiO2) were 11.37% and 11.53%, respectively, and conversion efficiency increased 0.16%. Mack et al. [25] showed that improve efficiency of without TO and with TO (SiO2) were 17.8% and 18.1%, respectively, and conversion efficiency increased 0.3%. Gatz et al. [26] showed that improve efficiency of back side SiNy/SiNx stacks for the surface passivation was from 17.9% to 18.3%, and conversion efficiency increased 0.4%. The absolute efficiency gain by surface passivation in single crystalline from 2006 to 2012 was only 0.16%~0.4% as shown in Figure 7 [2426]. Thus, the increase of 0.4% is quite significant.

In this study, the I-V characteristics of the samples were measured under illumination (Berger Lichttechnik, Model PSS 10 II), as summarized in Table 1. The efficiencies of sample A and sample B were 17.27% and 17.64%, respectively. The conversion efficiency increase of 0.37% is significant. Sample B efficiency was 2.14% higher than that of sample A, open circuit voltage (Voc) was 1.44% higher, and short circuit current (Isc) was 1.51% higher than that of sample A. Thus, the proposed two-step TO process is an effective way to increase mc-Si solar cell efficiency and related characteristics.

4. Conclusions

This study successfully demonstrates enhanced multicrystalline silicon solar cell efficiency by inserting a two-step growth thermal oxide layer to the SiNx passivation layer. The first oxidation step had a growth temperature of 780°C, growth time of 5 min, and with N2/O2 gas flow ratio 12 : 1; the second step had a growth temperature of 750°C, growth time of 20 min, and under pure N2 gas environment. The wafers were annealing at the temperature of 750°C to 780°C after the dry oxidation process. The goal was to improve bulk Si crystalline quality and bulk carrier lifetime after thermal oxidation. The proposed two-step process retained thin oxide layers, preventing effect formation of SiNx layer optical characteristics, which can reduce efficiency.

The mc-Si solar cell efficiency improved to 17.64% from after inserting two-step growth thermal oxide layer between the p-type Si/n-type Si and SiNx passivation layer. The resultant silicon oxide layer increased the absorption of the light in the spectral range 400–1050 nm by 4.1% and thus improved the total absolute efficiency of the mc-Si solar cell.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This work was sponsored by TSMC, NCTU-UCB I-RiCE program and Ministry of Science and Technology, Taiwan, under no. MOST106-2911-I-009-301.