Efficient wide-bandgap perovskite photovoltaics with homogeneous halogen-phase distribution

Wide-bandgap (WBG) mixed-halide perovskites are essential as top absorbers in tandem architectures with narrow-bandgap (NBG) bottom cells (e.g., Si, organic, or Sn–Pb perovskites) to maximize total voltage. However, WBG PSCs suffer larger open-circuit voltage (VOC) loss than NBG devices per the Shockley–Queisser limit. Increasing Br content to widen the bandgap intensifies VOC loss. A major cause is halide-phase separation during crystallization, especially above ~20 mol% Br, due to lower solubility and faster nucleation of Br-rich phases that preferentially crystallize at the film surface while I-rich phases deposit at the bottom, creating an inhomogeneous energy landscape that impairs carrier extraction and increases nonradiative recombination. While NiOx is a stable, transparent HTL for WBG PSCs, interface defects and energy-level mismatch at NiOx/perovskite reduce VOC. Self-assembled monolayers (SAMs) help but residual VOC losses persist. This work targets halogen-phase homogeneity by regulating crystallization kinetics via a double-layer SAM (D-2P) that templates bottom-up uniform phase formation, improves interfacial energetics, suppresses SRH recombination, and boosts VOC and efficiency.

Prior studies link VOC loss in mixed-halide WBG perovskites to phase segregation and interfacial nonradiative recombination. Br-rich phases, due to lower solubility and lower formation energy, crystallize earlier at the top surface, leaving I-rich phases at the bottom, causing energy landscape inhomogeneity and carrier extraction barriers. SAMs on NiOx have reduced buried-interface defects and improved VOC in inverted PSCs. In situ GIWAXS has clarified upward vs downward crystallization pathways, and compositional texture engineering/homogeneous energy landscapes have been shown to mitigate phase segregation and improve stability in wide-bandgap systems. Despite such advances, achieving both homogeneous halogen-phase distribution throughout the thickness and optimized energy alignment in WBG (>1.74 eV) devices remains challenging, with residual VOC deficits. This study builds on these insights by designing a π–π stacked double-layer phosphonic-acid SAM to anchor both [PbI6]4− and [PbBr6]4−, equalize formation energies, and drive uniform bottom-up crystallization.

• Design and formation of D-2P structure: A double-layer structure of 2-(9H-carbazol-9-yl)ethyl phosphonic acid (2P) is formed on ITO/NiOx via spin-coating ethanol solution of 2P (3000 rpm, 30 s) followed by 100 °C anneal (10 min), ethanol rinsing to obtain S-2P, re-coating 2P, and rinsing again to yield D-2P. Water contact angles quantify surface chemistry evolution (35.1° → 61.9° → 34.5° → 22.2°). C 1s XPS confirms π–π interaction (291.6 eV feature). XPS evidences Ni–O bonding between phosphate groups and NiOx, forming a stable bridge.
• DFT simulations: Calculated binding energies for 2P pair stacking (parallel, intersecting, antiparallel). Intersecting stacking is most stable (−10.847 eV vs −0.014 and −0.018 eV). DFT also evaluates formation energies of FAPbI3 and FAPbBr3 with/without 2P, showing reduced and closer formation energies upon 2P interaction, indicating nucleation facilitation for both halides.
• Film formation and in situ crystallization studies: Mixed-halide WBG perovskite FA0.8Cs0.15MA0.05Pb(I0.7Br0.3)3 (1.75 eV) prepared from DMF:DMSO (4:1) and anti-solvent chlorobenzene. In situ GIWAXS at grazing incidence angles θ = 0.3° (top) and 1° (bottom) monitors crystallization dynamics. For controls (S-2P), q shifts indicate top-first crystallization and depth-dependent phase; for D-2P, constant q at both depths indicates uniform phase and bottom-up nucleation.
• Depth-resolved composition/chemistry: Depth-profiling XPS (I 3d, Br 3d, Pb 4f) assesses halogen distribution and interfacial bonding (shifts to higher binding energies at the bottom indicate strong interactions with D-2P). FIB-prepared top/bottom flakes analyzed by TEM-EDS quantify Br/(I+Br) at different depths.
• Morphology/electronic landscape and energetics: KPFM on peeled films measures contact potential difference (CPD) at top and bottom surfaces. UPS (with minimal UV exposure by peeling film) determines Fermi level and VBM; UV–vis supports bandgap estimates and combined with UPS infers CBM alignment.
• Optical/electrical characterization: Steady-state PL and TRPL/TCFM mapping compare carrier extraction and lifetimes. TAS quantifies trap density of states (tDOS) across trap depths. Devices operated as LEDs to measure EL and EQE_EL; TPV/TPC for recombination/extraction times; dark J–V for reverse saturation; impedance spectroscopy for recombination (Rrec) and transport (Rtr) resistances; C–V for built-in potential (Vbi); J–V and EQE for photovoltaic metrics and hysteresis.
• Device fabrication: Inverted (p–i–n) architecture ITO/NiOx/D-2P (or S-2P)/Perovskite/PEAI/C60/BCP/Ag. Process specifics: perovskite spin 5000 rpm 32 s in N2, CB anti-solvent at 22 s, anneal 100 °C 60 min; PEAI 3 mg mL−1 in IPA at 4000 rpm 30 s; thermal evaporation of C60 (35 nm), BCP (6 nm), Ag (100 nm). Measurement area 0.089 cm2; J–V scan rate 100 mV s−1.
• 4-terminal all-perovskite tandem: Semi-transparent WBG top cell with ALD SnO2 (40 nm, 80 °C) and sputtered IZO (100 nm) as transparent top contact; NBG FAPb0.5Sn0.5I3 bottom cell on PEDOT:PSS with C60/BCP/Ag. Tandem performance from filtered spectra.
• Stability and reliability tests: GIXRD (2θ–sin2θ) for residual strain; temperature-dependent conductivity to extract halide ion migration activation energy (Nernst–Einstein); time-dependent PL under continuous 1-sun; SEM–EDS after prolonged illumination to assess halide redistribution; ISOS-L-1 protocol for operational stability of unencapsulated devices under 1-sun in N2.

• Halogen-phase homogeneity: D-2P induces bottom-up templated crystallization with nearly identical GIWAXS (001) q values at top and bottom (≈1.02 nm−1). Depth-profiling XPS shows similar Br 3d and I 3d intensities across depth in D-2P films, contrasting with Br-rich top and I-rich bottom in S-2P. TEM-EDS Br/(I+Br): S-2P top 38.63%, bottom 22.82%; D-2P top 31.42%, bottom 29.16% (close to nominal 30%).
• Interfacial interactions and energetics: XPS binding energy shifts (Pb 4f, Br 3d, I 3d) at the bottom indicate strong bonding of [PbX6]4− to D-2P via −P–OH–X and P=O–Pb2+. DFT shows intersecting π–π stacking is most stable (−10.847 eV) and 2P reduces and balances formation energies for FAPbI3 and FAPbBr3. UPS: D-2P bottom Fermi level 5.0 eV (shallower than S-2P 5.4 eV), improved alignment with 2P HOMO and reduced electron extraction barrier (larger inferred CBM).
• Carrier landscape and extraction: KPFM CPD difference top vs bottom shrinks from 0.482/0.188 V (S-2P) to 0.251/0.223 V (D-2P), indicating a more uniform energy landscape. PL is more strongly quenched and FWHM reduced in D-2P films, consistent with efficient hole extraction and homogeneous phase. TRPL/TCFM shows much shorter lifetimes for D-2P (order-of-magnitude decrease), indicating faster extraction; TAS shows lower trap DOS across the spectrum for D-2P.
• Device performance (1.75 eV WBG): D-2P PSC achieves PCE 20.80% (certified 20.70%) with VOC 1.32 V and JSC 18.81 mA cm−2, negligible hysteresis and stable MPP. S-2P control PCE 18.33% with VOC 1.25 V.
• Recombination and transport: Ideality factor n reduced from 1.88 (S-2P) to 1.21 (D-2P), indicating suppressed SRH recombination; lower dark reverse saturation. Impedance: higher Rrec = 22.52 kΩ and lower Rtr = 125.61 Ω for D-2P. TPV/TPC: τrec = 53.33 μs, τtra = 0.38 μs (improved extraction and slower recombination). C–V: Vbi increased from 1.01 to 1.16 V.
• Radiative efficiency and VOC loss: EQE_EL increases from 0.034% (S-2P) to 0.437% (D-2P), corresponding to a VOC loss reduction of ~0.15 V, consistent with measured VOC = 1.32 V.
• 4-terminal tandem: Semi-transparent WBG top cell PCE 19.35% (VOC 1.30 V); filtered NBG bottom PCE 8.73% (JSC drop from 31.52 to 12.87 mA cm−2). Combined 4T all-perovskite tandem PCE 28.08%.
• Stability and ion migration: Residual lattice strain reduced from 59.8 MPa (S-2P) to 10.0 MPa (D-2P). Halide ion migration activation energy increases from 0.17 eV to 0.58 eV. EL under bias shows no phase separation in D-2P up to 6 V, while S-2P shows phase separation above 3 V. Under continuous 1-sun illumination in air, D-2P PL remains at 708 nm with constant intensity after 400 h; S-2P shows rapid PL splitting and blue shift. SEM–EDS after two months illumination: D-2P maintains homogeneous halide distribution; S-2P becomes inhomogeneous. Operational stability (ISOS-L-1): D-2P retains >90% initial PCE after 2500 h (unencapsulated, N2), while S-2P falls below 50% by 1000 h.

The study addresses VOC loss in WBG PSCs by eliminating depth-dependent halogen-phase segregation that creates energetic barriers and promotes SRH recombination. The π–π stacked D-2P interface presents phosphate groups that simultaneously coordinate with [PbI6]4− and [PbBr6]4−, equalizing formation energies and reversing crystallization to bottom-up nucleation. This yields uniform halogen-phase distribution through the thickness, a more homogeneous electrostatic landscape (confirmed by KPFM), and improved band alignment with NiOx via the 2P layer (UPS). Consequently, carrier extraction is enhanced (short τtra, reduced PL lifetime with strong quenching), recombination is suppressed (lower ideality factor, higher EQE_EL, higher Rrec), and VOC increases to 1.32 V. Mechanical and ionic stability improve via reduced residual strain and higher ion migration barriers, leading to excellent operational stability. The approach directly targets the root cause of VOC deficits in high-Br WBG perovskites and demonstrates its significance for high-efficiency top cells in tandems, as verified by a 28.08% 4T all-perovskite tandem.

A double-layer 2P self-assembled structure (D-2P) at the NiOx/perovskite interface templates bottom-up, homogeneous crystallization of mixed-halide WBG perovskites by anchoring both I- and Br-based octahedra. This homogenizes the energy landscape, optimizes energy-level alignment, enhances carrier extraction, and suppresses nonradiative recombination, enabling a record PCE of 20.80% (certified 20.70%) with VOC = 1.32 V for a 1.75 eV WBG PSC, and a 4T all-perovskite tandem PCE of 28.08%. D-2P also reduces residual lattice strain and elevates ion migration barriers, delivering outstanding electrical and light stability (>90% PCE retained after 2500 h, unencapsulated, 1-sun, N2). This interfacial templating strategy offers a practical route to minimize VOC loss in WBG PSCs and advance high-efficiency perovskite tandems. Future work may explore generalized double-layer SAM chemistries, scalability, and integration with alternative transport layers and perovskite compositions.