© T. Dullweber et al., Published by EDP Sciences, 2025

1 Introduction

PV mass production technology is rapidly shifting from PERC (passivated emitter and rear cell) and bifacial PERC+ solar cells toward TOPCon (tunneling oxide passivated contact) solar cells. TOPCon cells are contacted on both sides with a diffused pn-junction and a charge carrier-selective n-type polysilicon (n-poly-Si) contact, which presently achieve efficiencies of around 25% [1]. The carrier-selective poly-Si contacts have been first published in 2013 as TOPCon [2] and POLO (polysilicon on oxide) [3]. As the next solar cell technology step, IBC (interdigitated back contact) solar cells enable higher short-circuit current densities, as both the metal contacts and the charge carrier-selective poly-Si contacts are located at the rear side and therefore do not cause any optical reflection or absorption losses on the front side. In addition, IBC designs allow to apply passivating contacts for both polarities thereby increasing the open circuit voltage. Thus, IBC cell designs offer 0.5% − 1.0%_abs. higher conversion efficiencies compared to TOPCon cells [1]. Consequently, IBC designs are presently leading the world record efficiencies for silicon solar cells with 27.8% [4] and for commercial modules with 24.4% [5]. Last but not least, IBC modules are attractive for the rooftop, building-integrated PV, and vehicle integrated PV markets due to their aesthetic appearance and high power density. In the future, with lower IBC production costs, IBC cells could be attractive as well for PV power plant applications. Currently, IBC cells account for less than 5% of the globally produced PV cells per year [1]. However, the interest of leading cell manufacturers in IBC cell designs has risen sharply in recent years due to the reasons mentioned before. Hence, in addition to the IBC pioneer SunPower / Maxeon, leading cell manufacturers in China such as LONGi and Aiko are now also producing IBC cells on a GW_p scale. According to the ITRPV forecast [1], the share of IBC cells in the global PV market will increase to 30% within the next 10 years.

To address these technical advantages and the high market interest in IBC solar cells, in the past years ISFH has developed a lean manufacturing process sequence for POLO IBC solar cells applying p-type wafers, an Al-BSF base contact and a local PECVD deposition of the SiO_xN_y/n-poly-Si emitter through a glass shadow mask [6–8]. In this paper, we report a new best POLO IBC cell efficiency of 24.5% processed at ISFH on M2 wafer size using industry-type processing equipment at the ISFH SolarTeC. The POLO IBC process with shadow masks at ISFH is very similar to the typical industrial bifacial PERC+ mass production sequence, allowing to re-use most of the existing PERC+ production tools [7,8]. To convert a PERC+ production line to POLO IBC, only the POCl₃ furnace and laser doping tool have to be replaced by a new PECVD tool for local SiO_xN_y/n-poly-Si deposition [7,8], see also Figure 1. In 2024, Kalyon PV and ISFH signed a technology licensing agreement and started to transfer the POLO IBC process from the ISFH SolarTeC to the Kalyon PV PERC+ manufacturing line using their M10 sized Ga-doped Cz wafers [9] and their PERC+ cell production tools. In this paper, we publish first promising POLO IBC test wafer results processed at Kalyon PV obtaining implied V_oc values up to 727 mV demonstrating a suitable passivation quality of Kalyon PVs wet chemistry and PECVD AlO_x/SiN tools.

However, the POLO IBC efficiency will be limited to below 25.5% by the carrier recombination at the Al-BSF base contact [10]. To overcome this limitation aiming at n-type polysilicon (n-poly-Si) / p-type polysilicon (p-poly-Si) POLO² IBC cell efficiencies beyond 26%, in collaboration with EnPV a carrier selective SiO_x/p-poly-Si layer stack has been developed at ISFH using industrial tools for the wet chemically grown SiO_x and the in-situ doped p-poly-Si deposited by LPCVD. Last year, we have published a best J₀ = 2.3 fA/cm² however with a high spread of J₀= 6 ± 4 fA/cm² [11]. In this paper, we optimize the p-poly-Si recipe yielding new best median value of J₀= 4 ± 1 fA/cm² with a reduced spread. ISFH is applying the SiO_x/p-poly-Si layer stack to develop a novel industrial processing sequence for the POLO² IBC solar cell [12]. We deposit both poly layers in-situ-doped full-area on M2-sized n-type Cz wafers and laser-structure both poly-Si polarities in a novel IBC trench layout targeting very cost-effective processes for etch barrier formation and poly-Si etching [12]. In this paper, we present first promising test wafer results demonstrating an implied V_oc up to 735 mV for POLO² IBC solar cells without metal contacts processed with industrial tools in the ISFH SolarTeC.

Fig. 1 Comparison of the PERC+ mass production process sequence at Kalyon PV to the ISFH process sequence of the 24.1% efficient POLO IBC cell. Green frames and lines connect very similar tools and process recipes. Hence, almost all PERC+ mass production tools from Kalyon PV can be re-used to manufacture POLO IBC cells, requiring only one new tool for the local PECVD SiOxNy/n-poly-Si deposition. Graph adapted from [7,8].

2 24%-efficient POLO IBC cells developed by ISFH and technology transfer to Kalyon PV

So far, the POLO IBC solar cells are processed at the ISFH SolarTeC. We use 0.8 − 1.5 Ωcm p-type Ga-doped M2-sized Cz wafers and apply the process flow detailed in Figure 1, right [7,8]. Processing starts with double-sided alkaline texturing and subsequent acidic polishing at the rear side. We grow a wet chemical interfacial SiO_x in a DI/O₃ solution and afterwards deposit the amorphous n-a-Si layer through a glass shadow mask provided by LPKF using an industrial PECVD tool. A glass shadow mask is used for about 10 consecutive poly-Si depositions. Afterwards, the poly-Si is removed from the glass mask by a KOH etch allowing to continue to use the glass mask for the following 10 PECVD poly depositions. At ISFH, the glass masks are loaded manually in the PECVD boat. In production, the PECVD loading robot will be adjusted to load wafers and masks into the boat. The POLO IBC process flow continues with annealing the n-a-Si layer to n-poly-Si in nitrogen atmosphere at 825°C. Both wafer sides are passivated by a AlO_x/SiN layer stack followed by laser contact opening (LCO) of the AlO_x/SiN at the rear side base region. Finally, the Al fingers are screen-printed on top of the LCOs and the Ag fingers are printed aligned to the n-poly-Si. The process sequence is completed by firing the wafers at around 825°C set temperature where the Al contacts locally alloy with the silicon wafer forming an Al-BSF and where the Ag paste dissolves the AlO_x/SiN layer contacting the n-poly-Si. Compared to an industry-typical PERC+ process e.g. at the Kalyon PV manufacturing line, the POLO IBC process uses the same PERC+ process steps such as texturing, rear polishing, AlO_x/SiN deposition, LCO and screen-printing of Al and Ag pastes. Only the POCl₃ diffusion and laser doping is replaced by the local PECVD SiO_xN_y/n-a-Si deposition as displayed in Figure 1. The resulting POLO IBC solar cell is shown as photographs in Figure 2. Whereas the front side is dark blue without metal contacts, the rear side shows the horizontal interdigitated Ag and Al fingers connected to the vertical Ag and Al busbars. Whereas the POLO IBC solar cell in Figure 2 still uses straight busbars, the new best 24.1% and 24.5%-efficient POLO IBC cells, see Table 1, include pads in the busbars to improve contacting by the IV test chuck. Figure 4 displays a schematic drawing of the POLO IBC solar cell design.

Applying this POLO IBC process sequence, previously the best POLO IBC cell efficiency was 23.9% [9]. In this paper, we report a new best POLO IBC cell efficiency of 24.1% independently confirmed by ISFH CalTeC with the current-voltage (IV) parameters summarized in Table 1. The high open circuit voltage V_oc = 723 mV demonstrates the good surface passivation quality of the AlO_x/SiN and SiO_x/n-poly-Si layers with saturation current densities J₀ around 3 fA/cm² [9]. The efficiency improvement from 23.9% to 24.1% was obtained by applying a 25°C lower n-poly-Si annealing temperature which is now optimized for the wet chemical SiO_x and which increased the V_oc from 720 mV to 723 mV, see Table 1. In addition, integrating contact pads in the busbars and applying a new IV test chuck with contact pins ensured a more precise contacting thereby reducing the series resistance from 0.8 to 0.67 mΩcm² as shown in Table 1. This new IV test chuck was qualified for high-precision IV test at ISFH CalTeC enabling an independent certified efficiency measurement. Subsequently, we optimized the anti-reflection coating applying a triple-layer AlO_x/SiN/SiON stack instead of the AlO_x/SiN layer which reduces the average reflectance from 1.7% to 0.9% thereby increasing the J_sc by 0.4 mA/cm² leading to a new best POLO IBC efficiency of 24.5% displayed in Table 1. The IV parameters of the 24.5% cell have been independently confirmed by ISFH CalTeC. As can be seen in Table 1, the POLO IBC series resistance R_s = 0.67 Ωcm² is still relatively high thereby limiting the FF to 82.0%. The high R_s is caused by a high Ag to n-poly contact resistance of 25 mΩcm² as published in reference [13] which is subject to further optimization aiming at future POLO IBC efficiency improvements. In addition, please note that the Ga-doped Cz wafer resistivity varies between 1.0 and 1.6 Ωcm for the 3 cells in Table 1 which slightly contributes to J_sc vs. FF variations between the cells.

As shown in Figure 1, the POLO IBC process with shadow masks at ISFH is very similar to the industrial bifacial PERC+ mass production process sequence at Kalyon PV, allowing to re-use most of their existing PERC+ production tools. To convert the PERC+ production line at Kalyon PV with efficiencies around 23.5% to POLO IBC with efficiencies above 24%, only the POCl₃ furnace and laser doping tool have to be replaced by a new PECVD tool for local SiO_xN_y/n-poly-Si deposition as shown in Figure 1. Kalyon PV and ISFH signed a POLO IBC technology licensing and technology transfer agreement in 2024. Afterwards, we qualified the Ga-doped Cz wafers produced by Kalyon PV for POLO IBC by increasing the ingot resistivity to be centered around 1.5 Ωcm [9] and started first POLO IBC process evaluations on test wafers as follows.

Kalyon PV has processed the first M10-sized p-type Ga-doped test wafers to apply and optimize the PERC+ production recipes for alkaline texturing, acidic polishing, wet chemical cleaning, and AlO_x/SiN front and rear passivation for POLO IBC. Double-sided textured and AlO_x/SiN-passivated wafers exhibit iV_oc values up to 731 mV. When adding rear polishing after texturing according to the test wafer process sequence in Figure 3 (top), first test wafers achieved an iV_oc value of only 687 mV, see Figure 3 bottom left. By optimizing the existing cleaning process post polishing and by selecting higher resistivity wafers with higher bulk lifetimes, the median iV_oc improves to 717 mV with a best iV_oc value of 727 mV as shown in Figure 3 bottom right. Hence, the iV_oc values obtained with Kalyon PV wet chemistry and AlO_x/SiN production tools are approaching the best POLO IBC iV_oc values of 740 mV obtained at ISFH SolarTeC [9]. As a next future step, Kalyon PV will process first complete POLO IBC solar cells using their M10 sized Cz wafers and PERC+ processing equipment with support by ISFH and assess the manufacturability via a small-scale pilot production.

Fig. 2 Photographs of a POLO IBC solar cell processed at the ISFH SolarTeC. The left image shows front and rear side of the cell, the right image adds the glass shadow mask in the front. This POLO IBC solar cell still uses straight busbars, whereas the 24.1% and 24.5%-efficient cells in Table 1 include pads in the busbars which improves contacting by the IV chuck.

Fig. 3 Process flow and QSSPC measurement results of POLO IBC lifetime precursors processed at Kalyon PV using their Ga-doped M10 Cz wafers and their PERC+ mass production tools for wet chemistry, PECVD AlOx/SiN, and firing. The lifetime samples achieve an implied Voc up to 727 mV demonstrating a suitable surface passivation quality of the industrial process tools.

Fig. 4 Schematic drawing of the 24.5% efficient POLO IBC solar cell (top) using p-type Cz wafers, an Al-BSF base contact and a carrier selective n-poly-Si contact (red layer). We are targeting efficiencies beyond 26% by replacing the Al-BSF contact with a carrier selective p-poly-Si contact (green layer) in the POLO2 IBC solar cell (bottom) and develop a novel industrially manufacturable process sequence applying n-type Cz wafers shown in Figure 6.

3 Future upgrade towards 26%-efficient POLO² IBC cells

The POLO IBC cell efficiency is limited to below 25.5% by carrier recombination at the alloyed Al-BSF base contact as simulated in [10] and as evident from the large gap between iV_oc = 740 mV [9] and V_oc = 723 mV in Table 1. Hence, ISFH is developing an industrial POLO² IBC design with carrier selective n-poly-Si and p-poly-Si contacts sketched in Figure 4 (bottom graph) which replaces the Al-BSF contact with a passivating SiO_x/p-poly-Si contact thereby targeting conversion efficiencies above 26%. Whereas the POLO IBC design is limited to p-type wafers, the POLO² IBC design is in principle compatible with both, p-type and n-type wafers. For industrial POLO² IBC development, we follow the world-wide industry trend towards n-type wafers due to their higher tolerance to impurities like Fe and less sensitivity to light-induced degradation.

Whereas other industrial IBC designs with passivating poly-Si contacts apply ex-situ doping of intrinsic poly-Si through a boron containing doping layer such as boron-silicate glass (BSG) [14–17], ISFH and EnPV have developed an in-situ doped p-poly-Si deposition process [11] with the goal to simplify the IBC manufacturing process. We deposit the carrier selective SiO_x/p-poly-Si layer stack applying industrial tools for the wet chemically grown SiO_x and for the in-situ doped p-poly-Si deposited by LPCVD [11]. As shown in Figure 5 the saturation current densities measured by QSSPC obtained with the wet chemical SiO_x are at least as good or even better compared to the thermally grown SiO_x. We adjust the boron doping profile of the p-poly-Si to reduce diffusion of boron through the interfacial oxide into the silicon wafer, apply an AlO_x/SiN capping layer and firing. As shown in Figure 5, green data points, the improved p-poly recipe obtains a new best median J₀ = 4 ± 1 fA/cm² with significantly reduced spread compared to previously published results [11].

ISFH applies the improved SiO_x/p-poly-Si process of Figure 5 to develop a novel POLO² IBC solar cell processing sequence. Using mostly lab-type tools and p-type float zone silicon wafers, a small-area POLO² IBC solar cell with 25.5% efficiency has been developed at ISFH [16]. In this paper, we report for the first time results obtained with a novel industrial POLO² IBC process sequence displayed in Figure 6 [12]. For the industrial approach, we choose M2-sized n-type Cz wafers due to their higher tolerance to Fe contamination and LeTID (light and elevated temperature induced degradation) and use solely industrial processing tools at the ISFH SolarTeC. We deposit the SiO_x/p-poly-Si layer in-situ-doped full-area (process steps 1 and 2 in Fig. 6) as explained in the previous section. During the high-temperature poly-anneal in step 3 we use O₂ and hence grow a thin oxide layer on the p-poly-Si. We laser ablate the oxide and apply a KOH etch in steps 4 and 5 to locally remove the p-poly-Si and a few micrometer of the silicon substrate. The final HF dip in step 5 fully removes the barrier oxide. Afterwards we wet chemically grow the second interfacial oxide (step 5) and deposit the n-poly-Si in-situ doped by LPCVD (step 6). In the future, we plan to replace these two steps by an in-situ PECVD SiON/n-poly-Si deposition as published in [8]. In the subsequent high-temperature poly-Si anneal in step 7 we again use O₂ and grow a thin oxide layer on the n-poly-Si. We remove the oxide layer from the front by a single sided etching step 8 and laser ablate the oxide on the rear side at the edges of the p-poly-Si in step 9. Hence, the following texture (step 10) removes the two poly-Si layers and interfacial oxide layers and textures the front side as well as the rear side where the oxide has been removed. Thereby we form a textured trench region on the rear side where all poly layers are fully removed hence insulating the two solar cell polarities. In the p-poly-Si emitter region the interfacial oxide and n-poly-Si layer form a tunneling contact with the p-poly-Si layer. The final HF dip post texture (step 10) removes the remaining barrier oxide from the rear side. Afterwards, we deposit AlO_x/SiN layer stacks on both sides (steps 11 + 12) and screen print Ag contacts on the rear side and fire the cells at around 800 °C set temperature (steps 13–15).

We find the POLO² IBC process sequence in Figure 6 to be a rather short and hence potentially cost-effective process sequences for manufacturing IBC solar cells with n-poly-Si and p-poly-Si passivating contacts. In addition, the trench insulation in steps 9 and 10 requires to laser only about 10% of the rear side wafer area whereas other IBC process sequences e.g. in Refs.14, 16, 17 have to laser about 50% of the rear side area which is challenging in terms of laser process time and production throughput. Another benefit is that the novel POLO² IBC process sequence in Figure 6 allows to use UV or cheaper green or IR lasers where the latter may create silicon defects at the wafer surface which are removed by the texture etch. In contrast, IBC process sequences e.g. in Refs.14, 16, 17 laser on top of poly-Si layers without etching these layers and hence they have to strictly avoid creating laser damage in the silicon substrate which may limit the choice of suitable laser sources. Finally, both poly-Si polarities in Figure 6 terminate with an n-poly-Si layer which allows to use the same screen printing paste for both solar cell metal polarities, e.g. an industry-typical TOPCon Ag paste.

To assess the passivation quality and V_oc potential of the novel industrial POLO² IBC process flow, we process test wafers as POLO² IBC solar cells with the process sequence in Figure 6 but without metal contacts using solely industrial processing tools at the ISFH SolarTeC. The test wafers mostly contain the IBC-typical interdigitated finger layout as sketched in Figures 4 and 6. We add two 4 × 4 cm² areas on the wafer where we change the laser patterning processes to obtain a field representing the n-poly-Si base contact (top left) and another field representing the p-poly-Si/n-poly-Si emitter contact (bottom right). Figure 7 shows a photoluminescence (PL) image of the measured carrier lifetime t at an illumination intensity of 0.78 suns of the resulting test wafer with the POLO² IBC finger layout including the vertical busbars as well as the n-poly-Si only (top left) and p-poly-Si/n-poly-Si only (bottom right) fields. In these three different areas, we also measure the injection dependent lifetime by QSSPC thereby determining iV_oc, iFF and the surface J₀.

As shown in Figure 7, the POLO² IBC area yields an iV_oc up to 735 mV which is in-between the n-poly-Si field with iV_oc = 740 mV and the p-poly-Si/n-poly-Si field with 722 mV. The implied fill factor iFF of the POLO² IBC area ranges up to iFF = 86.0%. To estimate the efficiency potential of the novel industrial POLO² IBC process sequence, we calculate the so-called implied efficiency ih as follows

$$\begin{array}{l}i\eta =i{V}_{oc}\times iFF\times J_{sc}/P_{light}\\ =735\text{mV}\times 86.0\%\times 41.7\text{mA}/{\text{cm}}^2/100\text{mW}/{\text{cm}}^2\\ =26.4\%.\end{array}$$

Here, we assume a J_sc = 41.7 mA/cm² which has been simulated for a POLO² IBC solar cell in reference [10]. Due to the nature of the passivating poly-Si contacts, we target to maintain the good iV_oc value as V_oc value in the POLO² IBC cell when including screen-printed metal contacts by stopping the metal contact alloying within the poly-Si keeping the SiO_x passivation intact. However, the fill factor of the POLO² IBC cell including screen-printed metal contacts will be lower than the iFF due to additional resistive losses of the metallization. Nevertheless, we assess the measured iV_oc and iFF values in Figure 7 as promising indication of an POLO² IBC efficiency potential of about 26%.

From symmetrical SiO_x/n-poly-Si lifetime test wafers, we determine a J_0,n-poly = 2 fA/cm²_. Similarly, from symmetrical SiO_x/p-poly-Si lifetime test wafers (without n-poly-Si tunnel contact), we determine a J_0,p-poly = 8 fA/cm². This value is significantly higher than the 4 ± 1 fA/cm² shown in Figure 5, likely because we had to use a different external LPCVD furnace for p-poly-Si deposition in this run due to a temporary LPCVD tool issue at ISFH, where the p-poly-Si recipe of the external LPCVD furnace was not yet optimized. We will soon apply the p-poly-Si LPCVD tool used in Figure 5 to POLO² IBC cell processing which may further increase the iV_oc values. Figure 7 shows multiple vertical yellow lines which exhibit a lower carrier lifetime than all other wafer areas. These yellow lines correspond to the poly-Si areas where in a metallized POLO² IBC cell metal busbars would be printed similar to the POLO IBC layout in Figure 2. The root cause of the slightly lower lifetime in these areas is not yet understood and is subject to further analysis and optimization.

We evaluate three different screen-printing Ag pastes and their optimum firing conditions using test wafers coated with SiO_x/n-poly-Si and AlO_x/SiN capping. The Ag test screen prints four fields on the wafer with different Ag finger pitches. By photoluminescence measurement we determine the average saturation current density J₀ value of each field and by interpolation determine the J₀ value of the Ag metallization J_0,met as e.g. described in reference [18]. By applying the transfer length method (TLM) we determine the specific contact resistivity r_c of the Ag paste to the n-poly-Si layer. We apply different n-poly-Si deposition methods such as LPCVD, PECVD and sputtering (PVD) in order to assess their impact on the electrical contact properties. The LPCVD and PECVD n-poly-Si layers are 200 nm thick, the PVD n-poly-Si layers were deposited with 120 nm thickness. As summarized in Figure 8, the LPCVD n-poly-Si layers obtain high J_0,met > 50 fA/cm² and high r_c > 4 mΩcm² for different Ag pastes and firing conditions revealing no suitable metallization option. In contrast, the PVD n-poly-Si layers obtain low J_0,met < 20 fA/cm² and low r_c < 3 mΩcm² which is the target range to enable POLO² IBC conversion efficiencies > 26%. The PECVD n-poly-Si layers are in-between with a few promising values around J_0,met = 20 fA/cm² and r_c= 3 mΩcm². Hence, for future processing of POLO² IBC cells with Ag contacts we will replace the LPCVD n-poly-Si deposition (step 6 in Fig. 6) with the PECVD n-poly-Si deposition which we have already optimized for POLO IBC cells, see step 3 in Figure 1. To our knowledge, we report for the first time that the n-poly-Si deposition method has a strong impact on the electrical Ag contact properties. The physical root cause is not yet fully understood and is subject to a future detailed analysis.

Fig. 5 Saturation current density J0 values of SiOx/p-poly Si surface passivation after coating with AlOx/SiN and firing. The wet chemical SiOx with improved in-situ-doped LPCVD p-poly recipe yields a new best median J0 = 4 ± 1 fA/cm2 with very low spread. Each data point corresponds to one test wafer.

Fig. 6 Schematic drawing of the novel industrial POLO2 IBC process sequence [12]. We apply an oxide barrier formed in the n-poly-Si anneal, a subsequent laser ablation of the oxide barrier at the p-poly edges followed by a texture etch to remove the poly-Si layers in a narrow trench region in order to insulate the n-poly-Si base contact layer from the p-poly-Si/n-poly-Si emitter tunneling contact.

Fig. 7 Photoluminescence (PL) mapping of the carrier lifetime t of a M2-sized n-type Cz test wafer representing a POLO2 IBC solar cell without metal contacts applying the novel process sequence in Figure 6 revealing a promising iVoc = 735 mV. The n-poly-Si field in the upper left corner represents the n-poly-Si base area of the POLO2 IBC cell with an excellent J0,n-poly = 2 fA/cm2 . The p-poly-Si/n-poly-Si field on the lower right represents the p-poly-Si/n-poly-Si emitter area of the POLO2 IBC cell yielding a J0,p-poly = 8 fA/cm2 which is a bit higher than the best values in Figure 5.

Fig. 8  Measured metallization current density J0,met and contact resistivity rc of screen-printed Ag contacts on n-poly-Si layers capped with PECVD AlOx/SiN stack. Whereas LPCVD n-poly-Si layers result in high J0,met values and high rc values, PECVD n-poly-Si and in particular PVD n-poly-Si obtain low J0,met < 20 fA/cm2 and low rc < 3 mΩcm2 which is required for high POLO2 IBC conversion efficiencies.

4 Discussion and conclusions

By optimizing the poly-Si annealing temperature, introducing contact pads for the IV test, and optimizing the anti-reflective coating, we obtained a new best POLO IBC cell efficiency of 24.5% processed at ISFH on M2-sized p-type wafers. With support by ISFH, Kalyon PV is currently transferring and implementing the POLO IBC process to their PERC+ cell manufacturing line using their in-house M10-sized Ga-doped Cz wafers. Textured, rear side polished and AlO_x/SiN passivated test wafers obtain an iV_oc up to 727 mV demonstrating a suitable passivation quality of Kalyon PVs wet chemistry and PECVD AlO_x/SiN production tools. Kalyon PV targets to process first M10-sized POLO IBC solar cells till end of 2025. As a next IBC technology upgrade aiming at 26% efficiency, ISFH is currently developing a novel industrial processing sequence for POLO² IBC solar cells with n-poly-Si and p-poly-Si carrier selective contacts. Using industrial tools for the wet chemically grown SiO_x and the in-situ doped LPCVD p-poly-Si, EnPV and ISFH optimized the SiO_x/p-poly-Si layer stack yielding a new best median J₀ = 4 ± 1 fA/cm². ISFH applied this SiO_x/p-poly-Si layer stack to the novel industrial processing sequence for POLO² IBC solar cells which includes in-situ doped and full-area deposited poly-Si layers and laser-structuring of both poly-Si polarities in a novel IBC trench layout. Using partly lab-type tools, a small-area POLO² IBC solar cell with 25.5% efficiency has been developed. Using solely industrial processing tools in our ISFH SolarTeC, M2 sized n-type Cz wafers, and our novel IBC trench insulation process sequence, we measured iV_oc = 735 mV and iFF = 86.0% of POLO² IBC cells processed without metal contacts. Since the poly-Si contacts minimize carrier recombination at metal contacts, the iV_oc value demonstrates the V_oc potential and an efficiency potential of up to 26.4% of this promising new POLO² IBC manufacturing process. Ag screen-printed metallization test structures reveal that the contact resistivity and the contact recombination strongly depend on the n-poly-Si deposition method favoring PECVD or PVD over LPCVD. Hence, in the future we will apply a PECVD n-poly-Si process for our fully-processed POLO² IBC solar cells.

Acknowledgments

The authors thank the companies LPKF Laser & Electronics SE, Germany, for manufacturing the glass shadow masks, and TOYO ALUMINIUM K.K., Japan, for providing the aluminum paste. We thank Jan Krügener, Leibniz University Hannover, Germany, for support with LPCVD depositions. We thank our ISFH colleagues Karsten Bothe and Gerrit Lange for support with IV test chuck optimization, Tobias Neubert for laser recipe setup, and Welmoed Veurman for Quokka 3 simulations.

Funding

This research was funded by the European Union in the research project IBC4EU (grant number 101084259) and by the German Federal Ministry of Economics and Energy BMWE in the research project Olivia (grant number 03EE1184C). Part of the research was funded by EnPV GmbH, Karlsruhe. In addition, the State of Lower Saxony in Germany contributed with the Institutional funding of ISFH.

The authors thank all funding agencies for financially supporting our research work presented in this paper.

Conflicts of interest

The authors have nothing to disclose.

Data availability statement

This article has no associated data generated or analyzed.

Author contribution statement

Conceptualization and Methodology, Yevgeniya Larionova, Philip Jäger, Verena Mertens, Udo Römer, Özlem Coşkun, Geoffrey Gregory, Erik Hoffmann, and Thorsten Dullweber; Solar cell processing, measurements and analysis, Sabrina Schimanke, Melanie Ripke, Ulrike Baumann, Alaa Osman, Gamze Çekerek, and Meriç Çalişkan Arslan; Writing, Thorsten Dullweber; Project Administration, Yevgeniya Larionova, Philip Jäger, Geoffrey Gregory, Özlem Coşkun. Supervision and Funding Acquisition, Thorsten Dullweber, Robby Peibst, Rolf Brendel, Özlem Coşkun, and Massimo Centazzo.

References

All Tables

All Figures

Fig. 1 Comparison of the PERC+ mass production process sequence at Kalyon PV to the ISFH process sequence of the 24.1% efficient POLO IBC cell. Green frames and lines connect very similar tools and process recipes. Hence, almost all PERC+ mass production tools from Kalyon PV can be re-used to manufacture POLO IBC cells, requiring only one new tool for the local PECVD SiOxNy/n-poly-Si deposition. Graph adapted from [7,8].

In the text

	Fig. 2 Photographs of a POLO IBC solar cell processed at the ISFH SolarTeC. The left image shows front and rear side of the cell, the right image adds the glass shadow mask in the front. This POLO IBC solar cell still uses straight busbars, whereas the 24.1% and 24.5%-efficient cells in Table 1 include pads in the busbars which improves contacting by the IV chuck.
In the text

	Fig. 3 Process flow and QSSPC measurement results of POLO IBC lifetime precursors processed at Kalyon PV using their Ga-doped M10 Cz wafers and their PERC+ mass production tools for wet chemistry, PECVD AlOx/SiN, and firing. The lifetime samples achieve an implied Voc up to 727 mV demonstrating a suitable surface passivation quality of the industrial process tools.
In the text

Fig. 4 Schematic drawing of the 24.5% efficient POLO IBC solar cell (top) using p-type Cz wafers, an Al-BSF base contact and a carrier selective n-poly-Si contact (red layer). We are targeting efficiencies beyond 26% by replacing the Al-BSF contact with a carrier selective p-poly-Si contact (green layer) in the POLO2 IBC solar cell (bottom) and develop a novel industrially manufacturable process sequence applying n-type Cz wafers shown in Figure 6.

In the text

	Fig. 5 Saturation current density J0 values of SiOx/p-poly Si surface passivation after coating with AlOx/SiN and firing. The wet chemical SiOx with improved in-situ-doped LPCVD p-poly recipe yields a new best median J0 = 4 ± 1 fA/cm2 with very low spread. Each data point corresponds to one test wafer.
In the text

Fig. 6 Schematic drawing of the novel industrial POLO2 IBC process sequence [12]. We apply an oxide barrier formed in the n-poly-Si anneal, a subsequent laser ablation of the oxide barrier at the p-poly edges followed by a texture etch to remove the poly-Si layers in a narrow trench region in order to insulate the n-poly-Si base contact layer from the p-poly-Si/n-poly-Si emitter tunneling contact.

In the text

Fig. 7 Photoluminescence (PL) mapping of the carrier lifetime t of a M2-sized n-type Cz test wafer representing a POLO2 IBC solar cell without metal contacts applying the novel process sequence in Figure 6 revealing a promising iVoc = 735 mV. The n-poly-Si field in the upper left corner represents the n-poly-Si base area of the POLO2 IBC cell with an excellent J0,n-poly = 2 fA/cm2 . The p-poly-Si/n-poly-Si field on the lower right represents the p-poly-Si/n-poly-Si emitter area of the POLO2 IBC cell yielding a J0,p-poly = 8 fA/cm2 which is a bit higher than the best values in Figure 5.

In the text

Fig. 8  Measured metallization current density J0,met and contact resistivity rc of screen-printed Ag contacts on n-poly-Si layers capped with PECVD AlOx/SiN stack. Whereas LPCVD n-poly-Si layers result in high J0,met values and high rc values, PECVD n-poly-Si and in particular PVD n-poly-Si obtain low J0,met < 20 fA/cm2 and low rc < 3 mΩcm2 which is required for high POLO2 IBC conversion efficiencies.

In the text