Hydrogen-Bonded Cations Push Perovskite Solar Cells Toward 26% Efficiency and Lasting Stability

Perovskite solar cells have dazzled the photovoltaics world with a meteoric rise in efficiency, yet their soft, fragile crystal lattices have remained the technology’s Achilles heel. Now, a research team writing in Advanced Science reports an elegant fix that sidesteps one of the field’s most stubborn problems: the instability of the two-dimensional capping layers that are supposed to protect three-dimensional perovskite absorbers. By building the large spacer cations of a 2D perovskite in situ, directly from the perovskite’s own formamidinium cations and a simple phosphonic acid molecule, the researchers created a 2D/3D heterostructure that delivered a champion power conversion efficiency of 26.32 percent on small cells and survived thousands of hours of punishing aging tests with most of its performance intact.

The context for this advance is the remarkable maturation of p-i-n architecture perovskite solar cells built on self-assembled monolayer contacts. Such devices have now reached a record efficiency of 27 percent, and they promise low-cost, industrially scalable photovoltaics that could complement or even challenge silicon. But the organic-inorganic hybrid perovskite film at the heart of these cells is a soft lattice in which ions drift readily under constant illumination, heat and moisture. Ion migration degrades performance over time, and the defects that accumulate at the interface between the perovskite and its charge transport layers are widely recognized as the critical origin of instability and degradation. Any route to commercialization must confront this interface head-on.

The most popular strategy has been to cap the 3D perovskite with a thin layer of 2D perovskite, whose bulky organic spacers physically restrict ion migration. Researchers have used phenethylammonium to form Ruddlesden-Popper type 2D layers, 3-aminomethylpiperidine to build Dion-Jacobson quasi-2D structures lattice-matched to the underlying crystal, and guanidinium to generate ACI-type phases that passivate surface defects. Yet this approach carries a hidden cost. Introducing large spacer cations to replace the native A-site cations, formamidinium and methylammonium, drives uncontrolled ion exchange that leaves vacancies behind and produces a fluctuating n-value, the number of inorganic layers in the 2D structure on which its properties depend. Worse, previous studies have shown that spacer cations can undergo deprotonation under irradiation, with the decomposition products reacting with the perovskite to release volatile hydrogen iodide and ammonia, corroding the very heterostructure meant to protect it.

The new work takes a fundamentally different route. Instead of importing a foreign ammonium cation, the team used n-hexyl phosphonic acid, or HPA, a molecule whose phosphonic acid head group forms multiple hydrogen bonds with the perovskite’s own formamidinium cations. Spectroscopic evidence for this interaction was unambiguous: proton nuclear magnetic resonance showed peak splitting of the formamidinium CH and NH2 signals, Fourier transform infrared spectra shifted in the same direction, and solid-liquid mass spectrometry detected reaction products corresponding to formamidinium bound to HPA and to HPA dimers. When HPA diffuses into the perovskite lattice, it locks onto formamidinium through these hydrogen bonds, and the resulting HPA-FA cations act as spacer cations that induce 2D perovskite formation in place, without any ion exchange displacing the native A-site population.

The team synthesized this 2D phase, dubbed HF-PVK with the formula (HPA-FA)2PbI4, directly from HPA, formamidinium iodide and lead iodide. X-ray diffraction revealed characteristic reflections attributed to the (001), (121), (002) and (003) planes of the 2D structure, while transmission electron microscopy confirmed lattice spacings of 9.3, 5.7 and 3.8 angstroms matching those planes. Ultraviolet-visible absorption showed the characteristic peaks of a 2D perovskite with multiple n-values. Crucially, the phosphonic acid group does double duty: nuclear magnetic resonance and X-ray diffraction showed that the PO(OH)2 group interacts with lead iodide, forming P-O-Pb bonds that passivate uncoordinated lead ions, a well-known source of trap states. In this design, one molecule simultaneously builds the protective layer and heals the defects beneath it.

To construct the working heterostructure, the researchers deposited HPA onto the surface of a 3D perovskite of composition FA0.85MA0.1Cs0.05PbI3. Grazing-incidence wide-angle X-ray scattering showed new 2D perovskite peaks emerging alongside a weakening of the residual lead iodide signal, indicating that HPA consumed unreacted precursors to form the 2D phase. Cross-sectional scanning transmission electron microscopy directly imaged the 2D capping layer, and in-situ photoluminescence during annealing captured the 2D phase appearing within minutes as characteristic signals at 501 and 592 nanometers grew with time. X-ray photoelectron spectroscopy showed the Pb 4f peaks shifting to lower binding energy as P-O-Pb bonds formed, while metallic lead peaks, a signature of deep defects, vanished entirely after treatment.

The electronic consequences were striking. Femtosecond transient absorption spectroscopy revealed that the treated films carried an additional bleaching signal from the 2D phase, and the 3D bleaching peak shifted from 772 to 786 nanometers, consistent with excited-state energy transfer from the 2D layer into the 3D absorber. Photoluminescence quenching at the C60 contact became faster, with the fast decay lifetime shortening from 388 to 233 picoseconds, evidence of more rapid interfacial carrier transfer. Ultraviolet photoelectron spectroscopy showed the work function dropping from 4.95 to 4.14 electron volts and the gap between the Fermi level and conduction band narrowing from 0.58 to 0.32 electron volts, giving the surface stronger n-type character that aligns favorably with the C60 electron transport layer. Kelvin probe force microscopy confirmed the contrast: a PEAI-treated surface swings p-type, which is poorly suited to this interface, while the HPA-treated surface swung decisively the other way.

The devices reaped the rewards. With 1.0 milligram per milliliter HPA, the champion cell improved from 24.29 percent efficiency to 26.32 percent, with open-circuit voltage of 1.16 volts, current density of 26.51 milliamperes per square centimeter and fill factor of 85.6 percent. Across twenty independently fabricated devices the average efficiency rose to 25.79 percent from 23.91 percent. External quantum efficiency exceeded 93 percent with an integrated current matching the measured photocurrent. Space-charge-limited current measurements showed trap density falling from 7.77 x 10^14 to 2.08 x 10^14 per cubic centimeter, impedance spectroscopy showed recombination resistance nearly doubling, and the ideality factor dropped from 2.25 to 1.34, all pointing to suppressed trap-assisted recombination. The strategy also scaled: cells of 1.012 square centimeters reached 24.91 percent, and a 5 by 5 square centimeter rigid module delivered 19.24 percent.

Stability is where the in-situ approach truly separates itself from conventional spacer cations. Density functional theory calculations found that the formation energy of A-site vacancy defects in the HPA-based 2D perovskite exceeds that of both Ruddlesden-Popper and Dion-Jacobson analogues, meaning the n-value stays fixed because formamidinium ions cannot easily migrate away. Time-of-flight secondary ion mass spectrometry after 8 hours at 180 degrees Celsius showed far less formamidinium and halide accumulation in the C60 layer of HPA-treated devices than in untreated ones. While PEAI-treated films showed spectral signatures of deprotonation after an hour at 180 degrees, the HPA-treated films held their chemistry. In aging tests, unencapsulated devices retained 95.5 percent of their initial efficiency after 3000 hours in nitrogen, 92.5 percent after 3000 hours in ambient air at 45 plus or minus 5 percent relative humidity, and about 90 percent after 800 hours at 85 degrees Celsius under the ISOS-D-2I protocol, while untreated devices lost half their efficiency under the same thermal stress.

The broader significance is hard to overstate. Perovskite photovoltaics have long been caught between two demands: the 2D layers that protect the absorber tend to undermine themselves through ion exchange and photochemical decomposition, while the untreated absorber degrades through defect-driven ion migration. By letting the protective cations assemble themselves from the perovskite’s own ingredients, held together by hydrogen bonds rather than fragile ammonium chemistry, this work dissolves that trade-off with a molecule as simple as a hexyl phosphonic acid. Champion efficiencies above 26 percent, working modules, and stability results that clear the thousand-hour mark in air suggest the strategy could translate directly into manufacturing. If in-situ cation construction proves robust across other compositions and module formats, it may become a standard step in the recipe that finally carries perovskite solar cells from the laboratory bench to the rooftop.

Subject of Research: In-situ construction of hydrogen-bonded spacer cations for stable and efficient 2D/3D perovskite heterostructure solar cells

Article Title: In‐Situ Constructed Cations for 2D/3D Perovskite Heterostructure for Stable and Efficient Photovoltaics

Article References: Liu, M., You, Q., Liu, L., Shi, X., Wang, Z., Wu, B., Yang, Z., Zhou, G., Liu, J.-M., Gao, J., & Jiang, Y. (2026). In‐Situ Constructed Cations for 2D/3D Perovskite Heterostructure for Stable and Efficient Photovoltaics. Advanced Science, 13(55), Article e76508. https://doi.org/10.1002/advs.76508

Image Credits: AI Generated

DOI: 10.1002/advs.76508

Keywords: perovskite solar cells, 2D/3D heterostructure, n-hexyl phosphonic acid, hydrogen bonds, ion migration, defect passivation, power conversion efficiency, self-assembled monolayer, stability, photovoltaics, spacer cations, ISOS protocols