Metal halide perovskites have emerged as premier candidates for next-generation solar energy harvesting because they combine exceptional optoelectronic performance with low-cost solution processability. Although solution-based manufacturing inherently introduces high densities of crystalline imperfections, extensive research has established that most point defects within the bulk lattice yield benign, shallow trap states. Instead, carrier recombination and operational degradation have predominantly been traced to deep trap states located at external surfaces, interfaces, and grain boundaries. Recent investigations, however, reveal that extended defects located inside individual crystalline grains, known as intragrain planar defects, also exert a profound and detrimental influence on carrier dynamics. These microscopic boundaries, which include ferroelastic twin walls and stacking faults, introduce local lattice distortions and altered chemical bonding configurations that trap mobile charge carriers or impede their transport across grain interiors. Resolving these performance losses has proved exceptionally challenging because electron-beam irradiation during standard transmission electron microscopy rapidly alters or erases these fragile planar features, obscuring the physical rules that dictate their formation during synthesis and processing.
To overcome this measurement bottleneck, researchers implemented low-dose transmission electron microscopy protocols tailored with defect-specific electron beam damage thresholds to systematically track intragrain planar defect formation in lead halide perovskite films. In conventional bright-field imaging, electron doses exceeding only a few electrons per square angstrom can deform or erase fragile ferroelastic twin boundaries, causing twin-bearing grains to falsely appear defect-free under the microscope. The researchers calibrated beam tolerances in methylammonium lead iodide films, discovering that the underlying perovskite crystal lattice remains structurally intact up to roughly ten electrons per square angstrom, whereas ferroelastic domains begin distorting at approximately 3.5 electrons per square angstrom and vanish almost completely at 5.9 electrons per square angstrom.
The experimental campaign focused on two primary classes of intragrain planar defects that populate distinct crystallographic phases: {112}t-type ferroelastic twins, which emerge in tetragonal or orthorhombic perovskite phases such as methylammonium lead iodide, and {111}c-type twins and stacking faults, which typically form in cubic lattice structures. Ferroelastic {112}t twins consist of periodic lamellar domains that spontaneously develop to relieve severe mechanical shear stresses generated when the material cools from its high-temperature cubic annealing phase down into its lower-symmetry room-temperature tetragonal phase.
By systematically varying thermal annealing durations between ten and thirty minutes under both inert nitrogen atmospheres and solvent vapor environments, the investigators established that {112}t ferroelastic twin architectures are governed strictly by physical grain size rather than direct chemical interactions with the surrounding atmosphere. In films annealed under inert gas for thirty minutes, average grain sizes measured 345 nanometers by scanning electron microscopy, and transmission electron microscopy revealed predominantly sub-200-nanometer crystallites that adopted single-domain, twin-free configurations. Conversely, solvent vapor annealing accelerated precursor and ionic mobility, driving rapid grain boundary migration that yielded larger crystallites averaging 550 nanometers after ten minutes and 950 nanometers after thirty minutes. Within these expanded grains, lamellar domain stripes and complex intersecting cross-lamellae appeared in abundance, confirming that larger grain volumes provide sufficient spatial dimensions and accumulated transformation strain to trigger ferroelastic twin wall nucleation during post-annealing cooling through the 330 Kelvin cubic-to-tetragonal phase transition.
The quantitative relationship between twin domain width and crystal grain dimensions was shown to closely follow a classical thermodynamic equilibrium scaling model for ferroic materials. When adjacent crystallites clamp together in a dense polycrystalline thin film, the balance between twin wall interfacial surface energy and stored elastic strain energy dictates an equilibrium lamella width proportional to the square root of the grain diameter. Statistical measurements gathered from over one hundred individual grains confirmed this square-root scaling law across all annealing protocols. Furthermore, the thermodynamic model predicted a definitive critical grain size below which the elastic strain energy is insufficient to offset the energetic cost of creating a domain boundary, leaving the grain entirely twin-free. In pure methylammonium lead iodide films, this critical grain threshold was experimentally determined to be approximately 220 nanometers, explaining why small grains remain single domains while larger grains inevitably develop dense internal twinning networks.
To suppress the thermodynamic driving force responsible for {112}t twin formation, the investigators modulated the degree of lattice tetragonality, defined as the ratio between the elongated c-axis and the orthogonal a-axis, by partially substituting methylammonium with larger guanidinium cations across an alloying range spanning zero to twenty-five atomic percent. X-ray diffraction patterns verified that incorporating guanidinium systematically compressed the tetragonal distortion toward unity without disrupting the three-dimensional perovskite framework or inducing secondary phases. For inertly annealed films, the tetragonality ratio dropped monotonically from 1.0133 in pristine methylammonium lead iodide down to 1.0025 at twenty-five percent guanidinium content. Four-dimensional scanning transmission electron microscopy nanobeam diffraction demonstrated that local intragrain strain varied continuously by only 0.5 to 1.0 percent, confirming homogeneous macroscopic cation incorporation without severe localized phase segregation.
This compositional engineering dramatically suppressed ferroelastic twinning by lowering internal transformation strain and elevating the critical grain size required for domain nucleation. As tetragonality decreased toward pseudo-cubic symmetry, the experimental critical grain size for twin formation expanded from 220 nanometers at zero percent guanidinium up to approximately 1.7 micrometers at a tetragonality of 1.0059, which corresponds to fifteen percent guanidinium substitution. In films possessing this reduced tetragonality, grains measuring less than two micrometers remained entirely free of ferroelastic twin lamellae even under extended solvent annealing conditions that produced substantial grain growth. Consequently, tuning the crystallographic tetragonality ratio provided an intrinsic thermodynamic knob capable of completely extinguishing {112}t twin boundaries across the grain size distributions typically found in high-efficiency perovskite solar cells.
However, lowering tetragonality too aggressively triggered a secondary, competing defect mechanism: the emergence of deleterious {111}c-type twins and stacking faults. When guanidinium concentrations reached or exceeded twenty atomic percent, lowering tetragonality below 1.005, X-ray diffraction identified a distinct diffraction reflection near 11.2 degrees two-theta corresponding to the hexagonal non-perovskite delta polytype, which shares identical face-sharing octahedral motifs with {111}c planar faults. To count these irregular defects accurately, the researchers utilized low-dose four-dimensional scanning transmission electron microscopy nanobeam diffraction combined with non-negative matrix factorization. Defect-bearing grains were rigorously isolated by detecting a first-order 3.66-angstrom diffraction peak accompanied by a collinear diffraction streak along the hexagonal c-axis. Quantitative image analysis revealed that the areal defect fraction jumped from 3.3 percent at twenty percent guanidinium to 13.9 percent at twenty-five percent guanidinium, while density functional theory calculations corroborated that the formation energy of {111}c boundaries drops steeply as tetragonality approaches unity.
Crucially, statistical analysis incorporating size-sampling corrections demonstrated that {111}c defect formation does not depend on crystallite dimensions, confirming that these stacking errors originate during initial crystal growth rather than through post-growth phase transformations. Instead, their prevalence was found to be highly sensitive to crystallization kinetics. Slower crystallization during inert annealing consistently yielded lower normalized delta-phase diffraction intensities compared to rapid, solvent-assisted growth. Synthesizing these dual observations allowed the research team to define a clear, quantitative processing window that simultaneously suppresses both detrimental defect varieties. For typical sub-500-nanometer photovoltaic grains, maintaining the lattice tetragonality between an upper threshold of approximately 1.010 and a lower boundary of approximately 1.005 eliminates {112}t twins without nucleating {111}c stacking faults, while for larger one-micrometer crystallites, the upper threshold narrows to 1.008.
The practical value of this defect-suppression processing window was conclusively validated through the fabrication and characterization of operational planar perovskite solar cell devices. Devices constructed under mild thirty-minute inert annealing conditions incorporating ten to fifteen percent guanidinium exhibited dramatic optoelectronic improvements that directly mirrored the suppression of intragrain defect boundaries. Time-resolved photoluminescence decay measurements revealed that the slow carrier lifetime component, which reflects non-radiative bulk recombination inside grain interiors, surged from 13.97 nanoseconds in control devices to 96.90 nanoseconds at fifteen percent guanidinium, producing an average carrier lifetime of 93.87 nanoseconds. Furthermore, champion device power conversion efficiency climbed from 16.22 percent in unmodified films to a peak of 18.12 percent at fifteen percent guanidinium, accompanied by an increase in fill factor from 0.68 to 0.76.
Subject of Research: Quantitative control and suppression of intragrain planar defects in halide perovskite photovoltaics
Article Title: Design Rules for Quantitative Control of Intragrain Planar Defects in Halide Perovskites
Article References: Gil, B., Park, S. J., Kim, J. Y., & Kim, M. (2026). Design Rules for Quantitative Control of Intragrain Planar Defects in Halide Perovskites. Advanced Science, 13(63), Article e77365. https://doi.org/10.1002/advs.77365
Image Credits: AI Generated
DOI: 10.1002/advs.77365
Keywords: Halide Perovskites, Planar Defects, Ferroelastic Twins, Stacking Faults, Transmission Electron Microscopy, 4D-STEM, Photovoltaic Efficiency, Carrier Lifetime, Crystallization Kinetics, Tetragonality, Guanidinium Substitution, Solar Cells, Material Science, Defect Engineering, Domain Wall Dynamics, Electron Microscopy, Thin Film Photovoltaics
Cite Scienmag News
Denise Maddox. (November 13, 2026). Processing Rules Suppress Intragrain Planar Defects in Halide Perovskite Solar Cells. Scienmag. https://scienmag.com/processing-rules-suppress-intragrain-planar-defects-in-halide-perovskite-solar-cells/
Denise Maddox. "Processing Rules Suppress Intragrain Planar Defects in Halide Perovskite Solar Cells." Scienmag, 13 November 2026, https://scienmag.com/processing-rules-suppress-intragrain-planar-defects-in-halide-perovskite-solar-cells/. Accessed 18 November 2026.
Denise Maddox. "Processing Rules Suppress Intragrain Planar Defects in Halide Perovskite Solar Cells." Scienmag. November 13, 2026. https://scienmag.com/processing-rules-suppress-intragrain-planar-defects-in-halide-perovskite-solar-cells/








