Abstract
We report industrial fabrication of different kinds of nanostructured multicrystalline silicon solar cells via normal acid texturing, reactive ion etching (RIE), and metal-assisted chemical etching (MACE) processes on diamond wire sawing wafer. The effect of different surface structure on reflectivity, lifetime, and electrical performance was systematically studied in this paper. The difference between industrial acid, RIE, and MACE textured multicrystalline silicon solar cells to our knowledge has not been investigated previously. The resulting efficiency indicates that low reflectivity surface structure with the size of 0.2–0.8 μm via RIE and MACE process do not always lead to low lifetime compared with acid texturing process. Both RIE and MACE process is promising candidate for high efficiency processes for future industrial diamond wire sawing multicrystalline silicon solar cells.
1. Introduction
Recently, diamond wire sawing technique on multicrystalline silicon wafer grew up at a very high speed due to cost-effectiveness compared to traditional slurry sawing process [1, 2]. It was demonstrated that diamond wire sawing (DWS) has several superiorities over multiwire slurry sawing (MWSS), that is, higher slicing speed, less saw damage layer, better environmental friendship, and potential for cutting thinner wafers [3, 4]. However, diamond wire sawing brings a big texturing problem due to the surface with less saw damage layer and less start point for normal acid texturing [5]. For mass production, there are now three main types of industrial techniques for fabricating DWS mc-Si solar cells: acid texturing with additive [6], reactive ion etching (RIE) [7], and metal-catalyzed chemical etching (MACE) [8–10].
In this work, the comparison between different process and the effect of different surface structure on reflectivity, lifetime, and electrical performance was systematically studied. We also demonstrate that DWS mc-Si solar cells can achieve average efficiency as high as 19.05% and 18.86% by RIE and MACE, respectively.
2. Experimental
The acid texturing, RIE, and MACE processes presented in this work are applied in the following solar cell fabrication process using industrial production machine:
Acid texturing for DWS mc silicon wafers by etching in HF/HNO3/H2O/additive solution (49 wt% HF: 65 wt% HNO3: H2O = 1 : 2 : 1.2 by volume, 10°C) and subsequent cleaning in KOH solution and HF/HCl solution via RENA Inline text machine. Acid texturing for MWSS mc silicon wafers as reference by etching in HF/HNO3/H2O solution (49 wt% HF: 65 wt% HNO3: H2O = 1 : 3 : 2 by volume, 10°C) and subsequent cleaning in KOH solution and HF/HCl solution via RENA Inline text machine. RIE for DWS mc silicon wafers by etching in HF/HNO3/H2O solution (49 wt% HF: 65 wt% HNO3: H2O = 1 : 3 : 2 by volume, 10°C) to remove saw damage layer, maskless RIE in SF6/O2/Cl2 plasma with 13.56 MHz radiofrequency using Belight system, and subsequent etching in HF/NH4F/H2O2 solution. MACE for DWS mc silicon wafers by etching in HF/HNO3/H2O/AgNO3 solution (49 wt% HF: 65 wt% HNO3: H2O = 1 : 2 : 2 by volume, 25°C), cleaning in hot HNO3 solution and subsequent etching in HF/HNO3/H2O solution. The etch depth of all texturing process are about 2.9 μm per each wafer side.
Emitter formation using a tube furnace from CT Systems with liquid POCl3 as dopant source at a temperature of 835°C and atmospheric pressure for 60 min in O2 and N2 ambient followed by edge isolation and removal of phosphor-silicate glass (PSG) using Rena tool. Plasma enhanced chemical vapor deposition (PECVD) of 80 nm hydrogenated amorphous silicon nitride (SiNx:H) antireflective coating at 400°C using a CT tool. Screen printing of Al rear contact, Ag rear contact, and Ag front contact with standard paste, which was fired using a Despatch infrared fast-firing furnace, with a peak temperature set point of 925°C and a belt speed of 6000 mm/min.
Samples used for lifetime measurement were deposited by SiN on both side and fired at the same temperature as the cell. The microstructure, reflection, lifetime of minority carrier, and IV curves of the resulted mc-Si wafers or solar cells were measured by SEM (Hitachi, S4800, Japan), reflectometer (Radiation Technology D8, China), IV measurement system (Halm, Germany), and WT2000 (Sinton WCT120, USA), respectively.
3. Results and Discussion
3.1. Microstructure of mc-Si Wafers with Different Texture Process
Figure 1 shows the SEM pictures of different surface structures including acid etching on MWSS wafer and acid etching, RIE, and MACE on DWS wafer. Figures 1(a) and 1(b) compare the surface morphologies of acid etching on MWSS and DWS wafer. Obviously, the size of oval pits is quite similar and about 2–5 μm. A uniform nanostructure fabricated by RIE was successfully integrated into the former etching pit on DWS wafer in Figure 1(c). The diameter of the nanostructure is in the range of 200–400 nm. Figure 1(d) shows that uniform small round etching pits with the size of 500–800 nm can be obtained by MACE process on DWS wafer.
(a) Acid-etched MWSS mc-Si wafer
(b) Acid-etched DWS mc-Si wafer
(c) RIE DWS mc-Si wafer
(d) MACE DWS mc-Si wafer
3.2. Reflectance of mc-Si Wafers and SiN-Deposited Wafer with Different Texture Process
Figure 2 shows the reflectivity R versus the wavelength (350–1050 nm) curves of wafers with different microstructures. The average reflectivity Rave of the acid etching DWS wafer (31.0%) is 5.7% higher than that of the MWSS wafer (25.3%), while the Rave of RIE DWS mc-Si wafer (15.1%) and MACE DWS mc-Si wafer (23.2%) is 15.9% and 7.8% less than that of the acid-etched DWS mc-Si wafer (31.0%). The reflectivity difference between acid-etched DWS and MWSS may be attributed to the deeper etching pit in the MWSS wafer than those of the DWS wafer because a thin amorphous silicon layer only found in the DWS wafer will prevent effective etching of the surface [8, 11]. The size of surface structure should be responsible for the significant reflectivity reduction of RIE and MACE DWS mc-Si wafer compared to acid-etched ones.
Figure 3 shows the reflectivity R versus the wavelength (350–1050 nm) curves of SiN-deposited wafers with different microstructures. The average reflectivity Rave of the acid-etching DWS wafer (7.7%) is 0.8% higher than that of the MWSS wafer (6.9%), while the Rave of RIE DWS mc-Si wafer (2.1%) and MACE DWS mc-Si wafer (5.8%) is 5.6% and 1.9% less than that of the acid-etched DWS mc-Si wafer (7.7%).
3.3. Minority Carrier Lifetime of mc-Si Wafers with Different Texture Process
Figure 4 shows the minority carrier lifetime of wafers with different microstructures. The average lifetime of the acid-etching DWS mc-Si wafer (58.1 us) is 15.5 us higher than that of the MWSS mc-Si wafer (42.6 us), while the Rave of RIE DWS mc-Si wafer (66.2 us) and MACE DWS mc-Si wafer (101.3 us) is 8.1 us and 43.2 us higher than that of the acid-etched DWS mc-Si wafer (58.1 us). The increasement of lifetime of acid-etched DWS mc-Si wafer compared to that of MWSS wafer was due to the thinner saw damage layer of DWS mc-Si wafer. It is important that even the size of surface structure decrease to several hundred nanometers, good surface passivation can still be obtained on DWS mc-Si wafer with proper process. It can be inferred that within the size of several hundred nanometers, the surface area is not the key factor to affect surface passivation. Especially for MACE process, the HF/HNO3/H2O treatment after Ag-removal acts like “rounding” process in texturing for HIT solar cells, providing possibility to deposit high-quality passivation film [12]. Both RIE and MACE processes have the potential to optimize surface recombination velocity and reflectance simultaneously.
3.4. Electrical Parameters of mc-Si Solar Cells with Different Surface Structure
Table 1 shows electrical performance of mc-Si solar cells with different surface structure. Even with some additive to improve the reflectance of acid-etched DWS mc-Si wafer, higher reflectivity leads to lower short-circuit current density (36.25 mA/cm2) and efficiency (18.51%) compared to that of acid-etched MWSS mc-Si wafer (36.60 mA/cm2 and 18.60%). Both RIE and MACE processes show significant current improvement compared to acid-etching process on DWS mc-Si solar cells due to the lower reflectance. Comparing RIE DWS with MACE DWS mc-Si solar cells, the RIE process shows higher Jsc and MACE process shows higher Voc which is in accord with the reflectivity and lifetime results. The average efficiency and open-circuit voltage of RIE (19.05%) and MACE (18.86%) also demonstrate their potential to achieve higher performance and the possibility to be mainstream texturing process for DWS mc-Si solar cells.
4. Conclusion
In summary, 19.05% and 18.86% efficiency DWS mc-Si solar cells with uniform nanotexture have been fabricated through RIE and MACE processes on an industrial production line. The results confirmed that both RIE and MACE process technique can provide an effective texture for DWS mc-Si solar cells. Because of the significant improvement of light-trapping ability on RIE and surface passivation ability on MACE DWS mc-Si solar cells, the Jsc of RIE DWS mc-Si solar cell is 1.23 mA/cm2 higher and the Voc of MACE DWS mc-Si solar cell is 0.7 mV higher than that of an acid-etched one. Those results demonstrate the biggest obstacle for industrial wide spread of DWS mc-Si wafer has been removed by the application of RIE or MACE process and this will benefit for high efficiency and low cost of the PV industry.
Conflicts of Interest
The authors declare that there is no conflict of interest regarding the publication of this paper.
Acknowledgments
The authors express their thanks to the National Natural Science Foundation of China under Grant no. 61575019; the Fundamental Research Funds for the Central Universities with Grant no. 2017RC034, no. 2017RC015, and no. 2017JBZ105; the Doctoral Program of Higher Education no. 20130009130001; and Jiangsu Planned Projects for Postdoctoral Research Funds (1701072B).