Development of Industrial n-PERT Rear Junction Solar Cells with Screen-Printed Aluminum Point Contacts

Abstract

This work is mainly focused on the development of industrial n-PERT-RJ solar cells with screen-printed Al point contact. In Chapter 1, the global market of photovoltaics and the cell technology of the crystalline silicon photovoltaics were introduced. Silicon transportation during the alloying process of the screen-printed Al point contact formation was investigated in Chapter 2 revealing that is severely affected by the amount of Si-add in the Al paste. The uniform distribution of the Si-add in the Al paste effectively suppressed the outdiffusion of bulk-Si during heating up of the firing process, resulting in a higher passivated area ratio (from 94.1 to 98.7%) of the solar cell’s rear side. The longer recrystallization period during the cooling down procedure of firing process, increased the thickness of the Al-doped p+ region (from 0 to a maximum of 4 μm) underneath the point contact, and supplied a better electrical shielding and a lower J0, Met (Al) (decreasing from 2800 to 635 fA/cm2). The proposed Si transportation model can be supported by the analysis of contact microstructures and VOC of solar cell devices (increasing from 560 to 682 mV). The SAM results were shown in Chapter 3 to reveal that no voids were formed under the Al point contacts and that the quality of the Al point contacts was highly improved when having Si-add in the Al paste. The feasibility of screen-printed Al point contact was proven. The discussions of the adverse impacts or sacrifices on JSC and FF of the n-PERT-RJ solar cells caused by the Si-add in the Al paste were followed up. The rear side reflectance has a direct impact on JSC, but it was not affected by any amounts of Si-add in the Al paste, merely directly related to the rear-side reflector covered area ratio, so there was no JSC loss caused by the Si-add. The RS increases when the amount of Si-add in the Al paste is increased, but the major increment didn’t come from the decreasing conductivity of the Al metal layer. Rather, most of the increase is caused by the formation of smaller Al point contacts. With the premise of identical contact characteristics, there was a 0.4% FF loss for the n-PERT-RJ solar cells processed with high wt% Si-add Al

7.1 Summary Page 92 paste, compared to those that were metallized with low wt% Si-add Al paste. Because of better contact shielding, a better pFF value was acquired, compensating the higher FF loss. There were two strategies for the [Si] gradient management: (1) increasing the Si-add in Al paste or (2) decreasing the Al matrix above each LCO (decreasing the LCO pitch). The second part of Chapter 3 showed that the contact width and J0, Met (Al) became gradually independent of the LCO pitch when the amount of the Si-add in the Al paste was modified from none to high wt%. In other words, Si-add dominated the effect of suppressing the outdiffusion of bulk-Si. Hence, a “point-line” contact concept was proposed for the rear side, consisting of a dot-shaped LCO and Al finger grid design metallized with high wt% Si-add Al paste. The concept not only maintained the VOC (693 mV) but had 60% bifaciality, providing an additional bifacial gain which further reduces the levelized cost of electricity (LCOE). The n-PERT-RJ solar cell concept plays a bridging role not only on the material switching from p-type to n-type, but also assists migrating from the traditional mc-BSF to the future advanced solar cell concepts (e.g., IBC or tandem). Many optimizations were shown in Chapter 4 for improving solar cell efficiency: (1) a shallower FSF doping profile, which made a better blue response, yielding a 0.3 mA/cm2 JSC gain and a J0, FSF decrease from 30 to 25 fA/cm2; (2) changing to fine-line printing (finger width reducing from 60 to 40 μm) and applying the dash contact concept reduced the effective metal fraction from 3% to 1% resulting in the area-weighted J0, Met (Ag) decreasing from 37.2 to 12 fA/cm2, which brought a 6 mV VOC gain; (3) screen-printed Al point contacts applied on the rear side with 635 mA/cm2 J0, Met (Al) and 1% effective Al contact area fraction made area-weighted J0, Met (Al) reduce more than 80% to 8.3 fA/cm2. The J0, Total was 77.8 fA/cm2 which translated to a VOC of 693 mV and 22.2% cell efficiency. In addition to the device performance improvements, the detailed loss analysis and the potential performance for the n-PERT-RJ solar cells were performed by Quokka3 simulation in the second part of Chapter 4. The efficiency losses from different mechanisms are shown in the free energy loss analysis (FELA); the recombination loss in the FSF and resistive loss in

Thesis summary and outlook Page 93 the bulk were the largest two of all loss mechanisms which should be optimized as a higher priority. The roadmap for reaching 700 mV VOC and beyond 23% cell efficiency with diffused surfaces and all screen-printed metallization was also shown. In Chapter 5, a more detailed discussion on the device level was proved. First, it was explained why the selective doping FSF structure had less benefit on our solar cells, only increasing the complexity and cost of the process. Then, the impact of different n-bulk parameters on n-PERT-RJ solar cell performance was revealed. The Rbulk between 1-13 Ω·cm can keep a stable solar cell efficiency. τbulk should be higher than 2 ms preventing severe performance decreases. Reducing the bulk thickness to 145 μm can compensate for the higher cost of the n-type wafers. The n-PERT-RJ solar cells in this thesis perform as well as the advanced passivating contact solar cells today. Only standard mass-production tools were used and the COO is comparable to the p-PERC solar cells with the selective emitter (SE). There was nearly no LeTID degradation, which makes these n-PERT-RJ solar cells have high performance, low cost, simple process, and long-term stability. The result of detailed comparisons of the microstructures, the contact electrical properties and the device performance between PVD-Al and SP-Al rear point contact n-PERT-RJ solar cells were shown in Chapter 6. The SP-Al point contact brought better performance than the PVD-Al point contact, not merely in regards to the electrical properties (J0, Met (Al) of 635 fA/cm2 comparing to 1450 fA/cm2) of the contact, but also to the solar cell efficiency (0.3 % higher). Integrating low (e.g. PVD-Al) and high (e.g. SP-Ag) thermal budget metallization is a challenge. In contrast, the rear SP-Al contacts can co-fire with the front SP-Ag, which is easier to realize using existing solar cell production lines. When discussing the further “bifacial” application on the n-PERT-RJ solar cells and the interconnections on the solar modules, the PVDAl metallization will face more challenges.