Researchers at Helmholtz-Zentrum Berlin (HZB) have built an all-perovskite triple-junction solar cell that converts 27.3% of incident sunlight into electricity, while retaining 90% of that performance after 770 hours of continuous 1-sun operational tracking. Both numbers matter for anyone weighing perovskite tandem programs against incumbent silicon: the stability figure, achieved at room temperature on encapsulated devices, sets a record for triple-junction all-perovskite configurations, a class long held back by degradation of the buried hole-transport interface.

The work, led by Steve Albrecht and published in Joule, targets the narrowest part of the problem. In multi-junction all-perovskite stacks, the tin-lead bottom sub-cell harvests near-infrared light, and the field has relied on the conductive polymer PEDOT:PSS to extract positive charges from it. That polymer carries two liabilities: it parasitically absorbs photons the bottom cell needs, and its acidic, hygroscopic chemistry attacks the perovskite itself. The result is a standing trade-off between efficiency and lifetime.

Self-assembled monolayers (SAMs) work well in pure-lead perovskites, but HZB's diagnostics showed why they fail on tin-lead. Using fast-hysteresis and bias-assisted charge extraction measurements, the team found that standard carbazole-based SAMs deposited directly onto tin-lead perovskite cause massive ion accumulation, screening the internal electric field needed to separate photogenerated carriers and producing uneven grain growth at the buried interface.

The fix is a bilayer. The researchers first coat the transparent indium tin oxide electrode with an ultrathin, uniform film of hydrophilic graphene oxide, then deposit the SAM molecule MeO-2PACz. Phosphonic acid head groups on the SAM anchor to the graphene oxide's oxygen-containing groups through hydrogen and covalent bonds. That arrangement deepens the layer's electronic work function, raises its conductivity, and gives the perovskite a smooth, hydrophilic surface that accelerates crystal nucleation — yielding films free of the nanovoids that plague SAM-only devices.

The measured gap is large. In single-junction tin-lead cells, the bilayer reached 22.1% power conversion efficiency against 12.0% for SAM-only controls, and optoelectronic characterization attributed the improvement almost entirely to suppressed electronic and ionic extraction losses rather than to optical gains.

For the full device, the team assembled a monolithic triple-junction stack with sub-cell band gaps of 2.00, 1.60 and 1.25 eV. They replaced the conventional gold plus PEDOT:PSS interconnect with an optimized indium tin oxide and graphene oxide/SAM configuration, cutting parasitic near-infrared absorption and lifting the bottom cell's short-circuit current density to 10.3 mA/cm². That current drove the stack to its 27.3% efficiency.

The remaining losses are identified, not speculative. The authors report that current mismatches and series-resistance transport losses across the interconnecting junctions still bound performance, and they project that improving the wider-band-gap perovskite layers and fine-tuning the band alignment of intermediate contacts should push the architecture past 30%.

For R&D managers, the result reframes the interface-engineering budget line: a two-step, solution-processable coating replaces a noble-metal interconnect and an unstable polymer in one move, delivering both the efficiency gain and the 770-hour durability record in the same device. The 30% projection remains a target, but the measured 27.3% and the stability benchmark give perovskite multi-junction portfolios a concrete baseline to build on.