Metal halide perovskite solar cells have achieved power conversion efficiencies (PCEs) above 27% for single junctions and approximately 35% for perovskite–silicon tandems. Industrial deployment, however, requires translating these results into large-area modules that combine high efficiency, spatial uniformity, and long-term reliability. This talk presents molecular engineering of the perovskite absorber within a fully industry-compatible device stack as a route from lab to fab, spanning scalable deposition and interconnection, encapsulation, accelerated testing, and multi-year field validation.

We frame the cell-to-module efficiency gap as a property-distribution problem: a sub-centimetre cell averages local variations in film quality, whereas a series-interconnected module is constrained by its least-performing sub-cell. Controlling spatial optoelectronic heterogeneity, not merely average film quality, is therefore essential for upscaling. To this end, we introduce a multifunctional crystallization-directing molecular additive into the perovskite precursor ink. All devices employ an identical industrial stack — linear-sputtered ITO front electrode, sputtered NiO_x hole transport layer, and evaporated passivation and electron transport layers — with only the absorber deposition adapted to substrate area: blade coating up to 100 cm² and slot-die coating at ~800 cm². This design isolates absorber chemistry as the scaling lever. The additive produces a large-grained, phase-pure absorber and improves optoelectronic uniformity, raising active-area PCE from 22.0% to 23.0% (0.125 cm² cells), 19.0% to 21.3% (3.9 cm² devices), 17.8% to 21.2% (15.5 cm² mini-modules), and 13.7% to 19.0% (100 cm² monolithic modules). The gain widens monotonically with area, from +1.0 to +5.3 percentage points, reducing the absolute upscaling loss by a factor of 2.1. Multi-scale characterization corroborates this improvement: large-area photoluminescence imaging reveals reduced property dispersion, correlative electron microscopy–cathodoluminescence shows recrystallized large grains and suppressed secondary PbI₂, and two-dimensional ¹H–¹H spin-diffusion NMR confirms that the additive–lattice interaction persists across the relevant length scales. Encapsulated mini-modules retain ~90% of initial efficiency after 1000 h damp heat (85 °C, 85% RH), 92% after 500 thermal cycles (−40 °C to 85 °C), and 85.2% after 159 days outdoors in Nicosia, Cyprus.

At fab scale, the identical stack transfers to ~800 cm² substrates with the additive-engineered absorber deposited by slot-die coating. Modules with metal top electrodes reach 18.3% PCE, while a soft-sputtered ITO top electrode yields semi-transparent modules with superior long-term stability: a 781 cm² module achieved 15.2% aperture-area efficiency (16.3% active area), and encapsulated modules maintained T₈₈ after one year of outdoor tracking in Crete, Greece. Ongoing optimization targets 18% PCE for the semi-transparent format, which also enables four-terminal perovskite/silicon tandems.

Finally, because accelerated ageing protocols cannot fully capture coupled environmental stressors or predict the service lifetimes required for bankability, we complement indoor testing with extensive outdoor campaigns. The reproducible process and operationally stable formulation enable meaningful comparison of field performance across samples. We report four years of outdoor data across diverse climates, including arid desert and polar deployments, revealing degradation pathways and failure modes not observed under conventional test conditions and providing benchmarks for material and device design.

These results establish molecular-level control of crystallization within an unchanged industrial stack as a chemically tractable, transferable scaling strategy, carrying a single additive concept from 0.125 cm² cells to ~800 cm² modules across two coating methods. The talk will conclude with remaining bottlenecks and the roadmap toward durable, high-performance perovskite modules and tandems for commercial production.