Oxidation-resistant all-perovskite tandem solar cells in substrate configuration

All-perovskite (perovskite–perovskite) monolithic tandem solar cells combine a wide-bandgap (WBG, ~1.8 eV) front subcell and a narrow-bandgap (NBG, ~1.2 eV) back subcell, and have surpassed single-junction perovskite cells in record efficiency. However, oxygen-related instability, especially the rapid oxidation of Sn2+ to Sn4+ in mixed Pb–Sn NBG perovskites, limits commercialization. Conventional superstrate architectures place the oxygen-sensitive NBG absorber at the top, exposing it to air and processing, while flexible tandems typically require transparent polymer substrates, restricting materials and cost. This study asks whether reversing the build order to a substrate configuration—depositing the NBG back subcell first and burying it beneath the recombination junction and WBG front subcell—can enhance oxygen resistance and expand substrate choices without sacrificing efficiency.

Prior work established high-efficiency all-perovskite tandems using superstrate designs with transparent conductive substrates (e.g., glass/ITO or PEN/ITO), but highlighted oxygen-induced degradation of Pb–Sn perovskites due to Sn2+ oxidation. Mitigation strategies include reducing additives (e.g., SnF2), antioxidant passivation, and thin ALD metal oxide barriers (e.g., SnO2) to slow oxygen ingress, yet long-term protection remains limited. Recombination junctions leveraging Au, solution-processed ITO nanocrystals, or sputtered ITO have enabled efficient charge recombination on smooth WBG surfaces but can suffer on rough NBG layers. Wide-bandgap perovskites suffer from halogen-related deep defects and ion migration; large organic cations and pseudo-halogen additives have been used for defect passivation. Flexible tandems have been demonstrated on transparent substrates; however, opaque, low-cost, high-temperature tolerant substrates are attractive if optical access can be from the top side as in a substrate configuration.

Device architecture and process flow were inverted to a substrate configuration to bury the oxygen-sensitive NBG subcell beneath the recombination junction (TRJ) and WBG front subcell. Steps: (I) Deposit 80 nm Cu by thermal evaporation on glass (or PEN), then sputter ~10–15 nm ITO to prevent Cu–perovskite reactions (Cu-only for Cu-foil substrates gets ~10 nm ITO). (II) Form the back (NBG) subcell with PEDOT:PSS/NBG perovskite/C60 using a two-step spin coat (1000 rpm 10 s, 4000 rpm 40 s, ethyl acetate antisolvent), anneal 100 °C 10 min, and evaporate 20 nm C60. NBG composition: FA0.7MA0.3Pb0.5Sn0.5I3 with additives (SnF2, tin powder, formamidine sulfinate) to suppress Sn4+ and improve film quality. (III) Deposit ALD SnO2 (~30 nm) at 75 °C as part of TRJ and as a solvent barrier. (IV) Evaluate recombination layers on rough NBG/C60/ALD-SnO2: Au (evaporated), spin-coated ITO nanocrystals, or magnetron-sputtered ITO (MS-ITO). Due to roughness-induced nonuniform coverage causing Schottky barriers and S-shaped JV curves with Au and ITO NCs, select MS-ITO (~20 nm) to form an Ohmic recombination contact with improved stability and minimal parasitic absorption. (V) Deposit the front WBG subcell with NiO/SAM (2PACz:MeO-2PACz) hole contact, WBG perovskite, C60, and ALD-SnO2. To balance thermal sensitivity of the underlying NBG with WBG crystallization needs, adopt a two-step anneal (85 °C for 15 min + 100 °C for 5 min). Incorporate 1.5 mol% guanidinium tetrafluoroborate (GuaBF4) into the WBG precursor FA0.8Cs0.2Pb(I0.6Br0.4)3 (~1.77 eV) to passivate halide vacancies via Gua+ hydrogen bonding and Pb–F interactions from BF4−. (VI) Sputter IZO (~60 nm) as the transparent top electrode and apply an anti-reflection film to reduce front reflection. Edge scraping of the NBG stack prior to TRJ/WBG deposition allows self-encapsulation of the NBG at device perimeters. Materials characterization includes SEM/AFM of ALD-SnO2 on NBG vs WBG surfaces, XPS (Pb 4f, Sn 3d), XRD, PL/TRPL, TPV, and contact angle. Device characterization includes JV under AM1.5G, EQE with appropriate bias illumination for subcells, and MPP tracking. Stability tests assess dark shelf in dry air (RH<20%), continuous MPP operation under 1 sun for unencapsulated and encapsulated devices, and damp heat (85 °C/85% RH). Flexible devices are fabricated on Cu-coated PEN and Cu foil using the same inverted stack and assessed for bending durability.

- Substrate-configured tandems with MS-ITO recombination layer initially achieved PCE 22.7% (Voc 1.980 V, Jsc 15.4 mA cm−2, FF 74.1%). Au and ITO nanocrystal recombination layers yielded S-shaped JV and low FF due to poor coverage on rough NBG surfaces. - GuaBF4 additive (1.5 mol% vs Pb2+) in WBG perovskite improves semitransparent WBG single-junction performance: substrate configuration champion PCE 17.3% (Voc 1.265 V, Jsc 16.7 mA cm−2, FF 81.9%) vs control 15.8% (Voc 1.214 V, Jsc 16.5, FF 78.1). Average gains: Voc 1.199→1.244 V; FF 77.9→80.8%; PCE 15.3→16.7%. - In superstrate WBG cells, GuaBF4 also improves performance: champion PCE 19.1% (Voc 1.274 V, Jsc 17.7, FF 84.5%) vs control 17.2%. - Mechanism: XPS shows increased Pb binding energy with GuaBF4; XRD indicates enhanced (100) orientation and reduced PbI2; PL intensifies and TRPL lifetime increases (198 ns vs 32 ns); TPV recombination lifetime increases (15.4 µs vs 2.5 µs); ideality factor improves (1.11 vs 1.44). Contact angle increases (75° vs 48°), indicating more hydrophobic surface. - Champion substrate-configured tandem with GuaBF4: 25.3% PCE (reverse) with Voc 2.041 V, Jsc 15.6 mA cm−2, FF 78.9%; forward 25.1%; stabilized PCE 25.1% over 60 s. EQE-integrated Jsc: WBG 15.8, NBG 16.2 mA cm−2. - Superstrate-configured comparator achieved up to 25.6% (Voc 2.034 V, Jsc 16.1, FF 78.2%). - Stability: Unencapsulated superstrate devices degrade severely within ~40 h in dry air, while substrate-configured devices remain stable >250 h; with self-encapsulated edges, no degradation over 1000 h. EQE degradation in superstrate arises from NBG; substrate devices maintain both subcells’ EQE. XPS after aging shows Sn4+ formation in superstrate after 20 h, while substrate devices show no significant Sn signal after 200 h. - Operational stability under 1-sun MPP: unencapsulated substrate devices stable >200 h; unencapsulated superstrate degrade within ~20 h. Encapsulated substrate devices retain ~100% initial efficiency after 600 h; encapsulated superstrate degrade after ~340 h. Substrate devices also outperform under 85 °C/85% RH and moisture/light soaking. - Flexible tandems (substrate configuration): on PEN/Cu, champion reverse PCE 24.1% (EQE-integrated Jsc: front 15.6, back 15.4 mA cm−2) with >88% PCE retention after 10,000 bending cycles at 15 mm radius. On Cu foil, champion reverse PCE 20.3% (Voc 1.957 V, Jsc 14.1, FF 73.4%); forward 20.8%. Performance on Cu foil limited by film cracks and substrate roughness. - Identified optical reflection at the top surface as a key factor limiting Jsc and tandem PCE in the substrate configuration; anti-reflection films help but further optical optimization is needed.

Reversing the device build to a substrate configuration successfully protects the oxygen-sensitive mixed Pb–Sn NBG perovskite by burying it beneath the recombination junction and WBG front subcell. This self-encapsulation significantly suppresses Sn2+ oxidation during storage and operation, directly addressing oxygen-related degradation prevalent in superstrate tandems. Achieving a uniform, Ohmic recombination junction over the rough NBG surface is critical; sputtered ITO (MS-ITO) over ALD SnO2 forms a robust TRJ where Au or ITO nanocrystals fail, enabling high FF. Incorporating GuaBF4 into the WBG perovskite reduces halogen vacancy defects via strong hydrogen bonding (Gua+) and Pb–F interactions (BF4−), reducing non-radiative recombination and improving Voc and FF in both single-junction and tandem contexts. The approach retains state-of-the-art tandem efficiencies (25.3%) while markedly improving dark shelf and operational stability, both unencapsulated and encapsulated. Moreover, substrate configuration decouples the need for transparent substrates, enabling efficient flexible tandems on opaque, lower-cost, and thermally tolerant substrates (Cu-coated PEN and Cu foil). Remaining performance limits stem largely from optical reflection at the front and mechanical challenges on rough metal foils; targeted optical management and substrate/film engineering can further raise efficiency and durability.

This work demonstrates oxidation-resistant all-perovskite tandems in a substrate configuration that buries the oxygen-sensitive NBG subcell, enabling high efficiency (25.3%) with outstanding air and operational stability. A GuaBF4 additive in the WBG perovskite improves Voc and FF by passivating halogen-related defects, contributing to higher tandem performance. The architecture broadens substrate options, enabling efficient flexible tandems on opaque metal-coated polymers and metal foils, with robust bending durability. Future research should address optical losses (front-surface reflection) through anti-reflection and nanophotonic light-trapping strategies and pursue thinner absorber designs to enhance flexibility without sacrificing photocurrent. With optical optimization, tandem PCEs exceeding 30% are anticipated. Further improvements in recombination junctions on rough surfaces and substrate surface engineering (e.g., smoothing layers for metal foils) could boost mechanical reliability and performance.

- Optical losses from front-surface reflection reduce Jsc in substrate-configured tandems; current anti-reflection foils only partially mitigate this. - Building a low-resistance TRJ on the rough NBG perovskite surface is challenging; metals or solution-processed RLs can form nonuniform contacts, lowering FF. - Thermal sensitivity of MA-containing NBG subcells constrains annealing of the overlying WBG, requiring carefully balanced thermal budgets. - Flexible devices on Cu foil suffer from perovskite film cracking and substrate roughness, limiting efficiency and mechanical stability at small bending radii. - Thick absorbers needed for strong absorption reduce mechanical flexibility; thinner absorbers with light trapping are needed for improved bendability.