Scalable two-terminal all-perovskite tandem solar modules with a 19.1% efficiency

Two-terminal all-perovskite tandem solar cells combine a wide-bandgap perovskite top cell with a narrow-bandgap perovskite bottom cell to surpass single-junction efficiency limits. While lab-scale devices have reached >23% (certified up to 26.4%), progress toward scalable modules is hindered by reliance on non-scalable spin-coating processes and challenges in monolithic interconnection and sequential deposition of two perovskite absorbers with similar solubilities. This study aims to develop and demonstrate a scalable, laser-scribed, two-terminal all-perovskite tandem module architecture that maintains high geometric fill factor and device performance, addressing interconnection in inert atmosphere and solvent-induced degradation during sequential deposition.

Recent advances in perovskite-based tandems include perovskite/silicon and perovskite/CIGS architectures, with two-terminal all-perovskite tandems achieving certified efficiencies up to 26.4%. Prior 2T devices typically relied on multiple spin-coating steps, limiting areas to ~12 cm², though surface-anchoring zwitterionic molecules have enabled 12 cm² active areas with 21.4% PCE. Scalable deposition methods (thermal evaporation, blade, spray, slot-die) have shown promise for single-junction modules (blade: ~17.8%, slot-die: ~20.8%). Key gaps include implementing laser-scribed P1–P3 interconnections compatible with tandem stacks in inert atmospheres and developing device architectures with robust interlayers to prevent dissolution of the WBG perovskite when depositing the NBG layer. Recombination layers present trade-offs: percolated Au improves scribe yield due to low lateral conductivity but adds parasitic absorption; thin ITO (~15 nm) can induce shunt losses. Improved tunnel/recombination junctions are identified as a future need.

- Device architectures: Reference spin-coated 2T all-perovskite tandem cells in p–i–n architecture with stack: IO:H front electrode | 2PACz HTL | WBG perovskite (FA0.83Cs0.17Pb(I0.83Br0.17)3, Eg=1.78 eV) | LiF | C60/SnO2 ETL | recombination layer (ITO ~15 nm for cells; percolated Au ~1–2 nm for modules) | PEDOT:PSS HTLbottom | NBG perovskite (Cs0.15FA0.85MA0.1Sn0.5Pb0.5I3, Eg=1.26 eV) | PCBM/C60 | BCP | Cu back electrode. - Laser-scribed module interconnection: P1–P3 lines implemented in an inert (N2) atmosphere using a ns 532 nm Nd:YVO4 laser with specific parameters (P1: 10 kHz, 50 mm s−1, 2.00 J cm−2; P2: 10 kHz, 33 mm s−1, 0.45 J cm−2; P3: 10 kHz, 100 mm s−1, 0.40 J cm−2). Scribe widths: P1 60 µm, P2 60 µm, P3 40 µm; total scribing width 240 µm. Module GFF derived from dead area ~210–220 µm measured by SEM/ToF-SIMS. - Scalable fabrication sequence (modules, 12.25 cm² aperture): Combination of blade coating (2PACz, WBG, PEDOT:PSS, NBG, PCBM) and vacuum deposition (LiF, C60, SnO2 by ALD, Au, BCP, MgF2). Blade coating with controlled speeds and temperatures; vacuum quenching for WBG to lower roughness (RMS ~20 nm). ALD SnO2 (35 nm) at 90 °C (TDMASn/H2O) used as barrier/protection. - Vacuum-assisted growth control (VAGC) with nitrogen flow for NBG: After blade coating NBG precursor (DMF:DMSO), immediate transfer to vacuum chamber with simultaneous moderate nitrogen flow (up to 260 cm³ min−1) to rapidly extract solvents, minimizing resting time to <5 s and preventing corrosion of underlying WBG. Conditions compared: PV0 (vacuum only), PV1 (lower N2), PV2 (optimized N2 ~260 cm³ min−1), with pressure-time profiles and crystallization timing. - Characterization: J–V under AM 1.5G (AAA simulator), MPP tracking (including elevated 85 °C tests). EQE of subcells with bias light/filters. LBIC mapping with 530 nm (top subcell) and 850 nm (bottom subcell) to assess photocurrent homogeneity. Electroluminescence imaging with 760 nm longpass and 775 nm shortpass filters to spatially resolve subcell EL. ToF-SIMS 3D depth profiling to verify selective layer removal at P1–P3. - Certification: Fraunhofer ISE CalLab PV Cells certification for 2.56 cm² spin-coated module after encapsulation and masking.

- Spin-coated reference 2T cells (0.1 cm²): champion PCE 23.5%; stable PCE (SPCE) 23.4% via 5 min MPP tracking at room temperature in N2. - Spin-coated laser-scribed modules (aperture 2.56 cm²; GFF 94.7%): in-house champion PCE 22.2% (active area 23.7%); Voc per stripe ~2.01 V; total Voc ~8.0 V (loss ~10 mV per stripe, ~0.5% relative); FF ~75%. MPP tracking: SPCE 21.4% for 15 h (stable power 54.7 mW); stable at 85 °C in N2 for 3.5 h. Certified encapsulated/masked module (2.43 cm²): 17.99% ± 0.63% steady-state MPP (Fraunhofer ISE), in-house I–V up to 19.8%. - ToF-SIMS and SEM confirm selective laser ablation at P1–P3 with dead area ~210–220 µm, yielding GFF 94.7% (±0.1%). - LBIC mapping shows homogeneous photocurrent generation and collection across both subcells; EL imaging reveals localized defects but overall good homogeneity, consistent with low Voc and FF losses in modules. - Scalable modules (aperture 12.25 cm²; 7 cell strips; GFF 94.7%): module PCE 19.1%; MPP 18.3% (~224 mW) for 20 h with <7% drop. Individual 1.75 cm² strip averages PCE ~19.4% with Voc ~1.93 V, FF ~70–71%, Isc ~24.5–24.9 mA; module FF ~70–71%. - Nitrogen-assisted VAGC reduces NBG solution resting time to <5 s, preventing WBG corrosion. Optimized PV2 condition produces microscopic, defect-free NBG morphology; without sufficient N2 flow (PV0/PV1), defects and degradation occur. Small-area 2T cells processed with VAGC+N2 outperform VAGC-only counterparts. - Loss analysis: Scalable modules show ~4% FF and ~100 mV Voc loss vs spin-coated modules, attributed to morphology/inhomogeneity from blade coating. Overall PCE drop from cell-to-module aligns with inactive area from GFF.

The study directly addresses the challenge of translating high-efficiency all-perovskite tandem cells into scalable, monolithically interconnected modules. By implementing inert-atmosphere laser scribing with well-controlled P1–P3 lines and a robust recombination junction, the work achieves high GFF (94.7%) and minimal Voc/FF penalties from interconnection. The introduction of nitrogen-flow-assisted VAGC for the NBG layer solves the core materials challenge of solvent-induced degradation of the WBG top cell during sequential deposition, enabling uniform, defect-minimized films and homogeneous current collection across large areas. The results demonstrate that performance losses upon upscaling are primarily tied to geometric fill factor and manageable morphology variations from blade coating. These findings validate a practical pathway for industrially relevant, all-thin-film, all-perovskite tandem modules using scalable processes and provide diagnostic insights (LBIC/EL/ToF-SIMS) to guide further optimization. Remaining bottlenecks include recombination junction parasitics and stability of Sn/Pb NBG absorbers; addressing these will further narrow the gap to record small-cell performance.

This work demonstrates scalable, two-terminal all-perovskite tandem modules fabricated exclusively by blade coating and vacuum deposition, combined with inert-atmosphere laser scribing. Spin-coated devices reach 23.5% PCE (0.1 cm²) and spin-coated modules achieve 22.2% PCE (2.56 cm²) with GFF 94.7% and stable operation. Scalable mini-modules (12.25 cm² aperture) achieve 19.1% PCE with homogeneous current collection confirmed by LBIC and EL imaging, and selective interconnections verified by ToF-SIMS. Key to non-destructive sequential deposition is nitrogen-assisted vacuum growth control, which minimizes solvent exposure and preserves the WBG top cell during NBG deposition. Future work should focus on: improving recombination/tunnel junctions to reduce parasitic absorption and shunt susceptibility; enhancing NBG absorber stability via compositional engineering and defect passivation; implementing inert-atmosphere encapsulation; and integrating inline gas/vacuum quenching with roll-to-roll or slot-die processes to enable larger-area manufacturing.

- Stability: Mixed Sn/Pb NBG perovskites are sensitive to oxygen/moisture; encapsulation in ambient during certification led to observable degradation and lower certified PCE versus in-house measurements. - Recombination layer trade-offs: Percolated Au reduces shunt formation but adds parasitic optical losses; thin ITO can cause shunting due to lateral conductivity. Neither solution is optimal. - Morphology control: Blade-coated films exhibit higher roughness (~20 nm RMS) than spin-coated (<10 nm), contributing to minor Voc/FF losses and spatial inhomogeneities detected by EL. - Process sensitivity: Sequential deposition requires very fast solvent extraction; insufficient nitrogen flow or delays cause WBG corrosion. - Diagnostic limits: LBIC spatial resolution is limited by spot size (200–500 µm), potentially missing sub-spot-scale defects revealed by EL.