The first demonstration of entirely roll-to-roll fabricated perovskite solar cell modules under ambient room conditions

Organic–inorganic hybrid perovskite solar cells (PeSCs) have achieved record power conversion efficiencies (PCE) up to 26.1%, rivaling crystalline silicon. Yet, translating small-area, lab-scale cells to large-area devices made by high-volume, low-cost manufacturing remains a key challenge. Existing high-efficiency large-area devices often rely on sheet-to-sheet processing, vacuum evaporation, and laser scribing, which add cost and complexity. Flexible PeSCs offer continuous roll-to-roll (R2R) manufacturing and high specific power for applications like space, vehicle- and building-integrated PV, but impose constraints on time/temperature processing. Replacing high-cost components—especially vacuum-deposited Au back electrodes and commercial transparent conductive electrodes (TCEs)—with cheaper, printable alternatives while maintaining performance is essential. Prior work demonstrated the first fully R2R-fabricated small-area PeSCs using printed carbon electrodes (~10.8% PCE), but with significant performance penalties. This study addresses these gaps by demonstrating entirely R2R-fabricated perovskite modules under ambient conditions, enabled by robust scalable deposition, perovskite-friendly printed carbon electrodes, and a high-throughput R2R experimental platform for rapid optimization.

The paper situates its contribution within prior advances in scalable perovskite processing, including spray, blade, and slot-die coating, and notes recent large-area glass-based modules with high efficiencies that still depend on discrete, vacuum-based, laser-scribed processes. Earlier reports on solution-processed back electrodes often required high-temperature steps incompatible with flexible substrates and R2R. A recent milestone (2023) showed fully R2R small-area PeSCs using printed carbon electrodes with 10.8% PCE, indicating the feasibility but also the performance limitations of printed electrodes. The authors also reference their prior PFSD (printing-friendly sequential deposition) approach and subsequent additive strategies, as well as other R2R techniques, highlighting the need for vacuum-free, scalable processes that retain high efficiency.

- Deposition strategy: The study adopts and advances the Printing-Friendly Sequential Deposition (PFSD) technique. A sub-stoichiometric organic cation loading (<50 mol% of PbI2) forms an amorphous precursor thin film with improved film-forming properties. Subsequent deposition of additional organic cation (e.g., MAI) rapidly converts the intermediate to perovskite without time-consuming steps, aligning with R2R process timescales. - Shallow-angle edge blowing: Introduces a shallow-angle N2 blowing at the edge of a roller to create a well-defined solidification zone, minimizing wet-film deformation, reducing crystalline defects, and enabling uniform, mirror-like films. XRD shows no PbI2 residues; SEM indicates compact, homogeneous grains. This approach increases humidity tolerance and processing reliability under ambient conditions. - R2R high-throughput platform: A programmable R2R slot-die (SD) coater enables unmanned fabrication of thousands of unique cells daily. A custom automated R2R solar tester with throughput >10,000 cells/day performs rapid characterization. Device parameters are computed and stored automatically, allowing analysis of thousands of cells in minutes. - Parameter screening: A high-throughput experiment screened 20 parameter combinations to fabricate 1,600 cells, varying PbI2 thickness (yielding ~600–1000 nm perovskite) and MAI flow rate (30–100 µL min−1). Both stoichiometric and off-stoichiometric FA0.45MA0.55PbI3 compositions were explored to study composition-dependent performance. - Hole-transport layer (HTL) system: Developed an HTAB-P3HT HTL stack. HTAB (n-hexyl trimethyl ammonium bromide) passivates perovskite surface traps and promotes P3HT self-assembly. Ultra-thin HTAB was formed via solvent polarity control (chlorobenzene:isopropanol = 9:1 v/v). Uniform P3HT films were achieved with substrate heating at ~45 °C during SD coating. - Printed back electrode: Developed perovskite-friendly carbon inks comprising ethyl cellulose binder, 1:1 carbon black/graphene nanoplatelets as conductive pigment, and PGMEA solvent. A two-stage preparation (high-viscosity milling via three-roll mill, followed by dilution) produced SD-compatible inks. Reverse gravure (RG) coating was also used for module back electrodes. - Device stacks: Flexible PET/TCE/SnO2 ETL/perovskite (FA0.83MA0.17PbI3 or FA0.45MA0.55PbI3)/HTAB/P3HT/printed carbon. Printed Ag grids (screen printing) added atop carbon to reduce series resistance and to interconnect cells in modules. - Cell fabrication specifics (ambient R2R): SnO2 ETL by RG or SD; PFSD perovskite via SD for PbI2:FAI precursor, shallow-angle N2 blowing, then MAI SD coating and brief thermal treatment; HTAB and P3HT SD coated with controlled temperatures (HTAB anneal ~100 °C; P3HT at 45±5 °C); carbon electrode SD coated at ~70 °C then annealed at ~130 °C. Some devices included screen-printed Ag grid (180 mesh, 0.2 mm lines). - Module fabrication: Five-channel SD coater deposits five ~13 mm stripes in parallel on pre-patterned commercial TCE with 2 mm gaps; flows for five stripes scaled from single-stripe parameters (e.g., PbI2:FAI 100 µL min−1, MAI 300 µL min−1, HTAB 140 µL min−1, P3HT 92 µL min−1, carbon 600 µL min−1). RG-coated carbon layer used in modules; Ag grid screen-printed (R2R system), dried with IR and hot air, forming collection grids and series interconnects. Active area ~49.5–50 cm2 (five cells of ~1.1 cm × 9.0 cm). Geometrical fill factor (GFF) ~75%. - Characterization: J–V under 1-sun AM 1.5G using calibrated solar simulators; forward and reverse scans reported; IPCE measured; XRD, SEM, optical density, and TRPL used for materials analysis. Automated testing employed a small aperture (0.025 cm2) for robust alignment on ~0.2 cm2 cell stripes; many experiments conducted under ambient, including high humidity (~60% RH). - Cost modeling: Updated prior cost model to include demonstrated materials, processes, and device architectures. Compared sequences: A (vacuum Au electrode), B (fully printed with carbon electrode and commercial TCE), and C (model architecture eliminating commercial TCE and Ag grids).

- First entirely R2R-fabricated perovskite solar modules produced under ambient room conditions using only industrial printing/coating (SD, RG, screen printing) without vacuum steps. - Small-area flexible cells (fully R2R, vacuum-free, printed carbon electrode, HTAB-P3HT HTL): best PCE 15.5%, Jsc 19.9 mA cm−2, Voc 1.02 V, FF 76.1%; calculated current from IPCE 19.4 mA cm−2. Average PCE production around ~13% under variable humidity (~60% RH), demonstrating robustness. - High-throughput optimization: 1,600 consecutively fabricated cells across 20 deposition parameter sets on a 9 m substrate; identified composition- and thickness-dependent trends (e.g., near-stoichiometric MAI best for thin films; MAI-deficient thicker films improved FF; MAI excess increased Jsc). Enabled rapid, one-day discovery of optimal PFSD parameters. - Film quality improvement: Shallow-angle, edge N2-blowing yields mirror-like perovskite films with no detectable PbI2 by XRD and more homogeneous grain morphology by SEM; enhances device reliability and humidity tolerance. - Module performance (five series-connected stripes, ~50 cm2 active area, printed carbon + screen-printed Ag grids): reverse-scan active-area PCE up to 11.0% with Isc ~192 mA, FF 62.3%, Voc 4.59 V; forward-scan PCE 9.96%. GFF ~75%. Performance limited by TCE sheet resistance and partial solvent damage during screen printing. - Electrical design: Printed carbon sheet resistance ~8000 Ω/sq necessitated added Ag collection grids; without grids, performance significantly degraded. Optimal grid: 0.2 mm lines via 180 mesh screen, minimizing coverage while ensuring adequate conductivity. - Cost modeling: Fully printed configuration (Seq. B) substantially reduces back-electrode costs vs vacuum Au (Seq. A). Estimated manufacturing cost likely <1 USD W−1 for Seq. B at demonstrated efficiencies (15.5% for cells). A model architecture (Seq. C) eliminating commercial TCE and Ag grids could reach <0.5 USD W−1 at 10% efficiency. These improve on prior estimates (~1.5 USD W−1).

The study directly addresses the central challenge of translating lab-scale perovskite performance to scalable, low-cost manufacturing by demonstrating fully R2R-fabricated modules under ambient conditions. The PFSD process with shallow-angle edge blowing enables high-quality perovskite films compatible with continuous processing and varying humidity, overcoming a key barrier to reliability in ambient R2R fabrication. The development of perovskite-friendly printed carbon electrodes removes costly, vacuum-based Au deposition, while the high-throughput R2R platform accelerates optimization by exploring a large parameter space rapidly and statistically. Performance results show that fully printed flexible cells can reach 15.5% PCE, and modules achieve 11% PCE at ~50 cm2 active area, surpassing prior fully R2R demonstrations and approaching thresholds for portable PV markets. The findings elucidate trade-offs in composition and thickness for PFSD and highlight system-level constraints: TCE resistance and solvent interactions from screen printing impact module FF and stability. Cost modeling confirms that printed-carbon-based devices can lower $/W substantially, marking a credible pathway toward competitive manufacturing costs, particularly for applications valuing lightweight, flexibility, and high specific power. Collectively, the results validate the feasibility of vacuum-free, ambient R2R processing for perovskite modules and outline the critical engineering targets (conductive grids/electrodes, higher GFF, TCE alternatives) required to close the performance and cost gap to mainstream PV.

This work demonstrates the first entirely roll-to-roll fabricated perovskite solar cell modules, produced under ambient room conditions using only industrially relevant, vacuum-free processes. Key advances include a robust PFSD method with shallow-angle edge blowing for uniform films, perovskite-friendly printed carbon electrodes, and an automated high-throughput R2R platform that enabled rapid optimization over large parameter spaces. The best fully printed flexible cells reached 15.5% PCE, and modules with ~50 cm2 active area achieved 11.0% PCE. Cost modeling indicates significant reductions in manufacturing cost per W compared to vacuum-based architectures, with fully printed devices likely below 1 USD W−1 and potential future architectures below 0.5 USD W−1. Future work should focus on increasing module GFF, mitigating TCE resistance (or replacing TCEs), eliminating silver grids via highly conductive, perovskite-friendly carbon conductors, minimizing solvent-induced damage during printing, scaling to larger areas, and improving operational stability and encapsulation to meet commercial reliability standards. These steps could further elevate efficiency and reduce costs, opening pathways to niche markets and, ultimately, broader commercialization.

- Module FF and efficiency are limited by the high sheet resistance of commercial flexible TCEs and the high sheet resistance (~8000 Ω/sq) of the printed carbon electrode, necessitating additional Ag grids. - Screen-printed silver grids can induce partial solvent damage to underlying layers and pose long-term corrosion risks, making them suboptimal for commercial durability. - Geometrical fill factor (GFF) is ~75% due to the stripe-pattern approach, lower than laser-scribed modules (up to ~99%), reducing module PCE. - Many measurements were performed on unencapsulated devices and under ambient conditions; long-term operational stability and environmental durability are not reported. - Although ambient robustness is shown (including ~60% RH), performance variability with humidity exists; best devices occur at lower RH. - The lowest-cost architecture (Seq. C) is not experimentally demonstrated; cost projections rely on modeling and assumed efficiency. - Module area is modest (~50 cm2); scalability to larger formats and manufacturing yield at scale remain to be validated.