Organic-inorganic hybrid perovskite solar cells (PSCs) are a promising means of producing renewable energy from the abundant solar resource. They present the potential for cost-effective materials and manufacturing, coupled with avenues for achieving high efficiency and physical flexibility.

PSCs have already seen impressive power conversion efficiencies (PCEs) of 26%. However, when it comes to scaled modules, maintaining this high efficiency has been a challenge: this comes in part from defects that arise during solution processing. Transitioning from laboratory-scale cells (< 0.1 cm²) to larger modules (> 10 cm²) poses a notable challenge. In fact, the highest certified stabilized efficiency for perovskite solar modules exceeding 30 cm2 in area has recently been observed to plateau at approximately 19.5%.

Prior researchers have explored the use of additives, such as Lewis acids and bases, in the precursor solution, to prevent defect formation. These, unfortunately, tend to aggregate within the perovskite film during co-precipitation in the course of solvent evaporation.

This led us to formulate a hypothesis: exploring alternative processing methods beyond the commonly used evaporative precipitation and crystallization could potentially decrease defects and the presence of insulating inclusions.

Our concept was to shift towards the use of additives that maintain a liquid phase during perovskite crystallization, thereby ushering materials processing into a new era. We conjectured that thermotropic liquid crystals might meet these criteria. Due to their characteristic mesomorphic phase transition from a solid to an isotropic liquid phase when heated, they would remain in the liquid phase while the perovskite solidified.

We screened a family of six liquid crystals, looking at their phase transition temperatures, their solubilities, and their binding energies with defect sites on perovskite (Fig. 1). Among these, we noted one candidate - (3,4,5-trifluoro-4'-(trans-4-propylcyclohexyl)biphenyl (TFPCBP) - that showed particular promise.

Figure 1. Characterization of liquid crystals and interactions with perovskite. a. Comparison of solubility and melting points of liquid crystals. b. Binding energy of -C₆H₂F₃ group attached to the iodine vacancy defect.

Another crucial criterion for liquid crystals selection would be uniformity in the distribution of additives throughout films. It is crucial that these molecules do not build up on the surface, among other things. To achieve this, the molecules must possess the property of efficient diffusion within a mixed liquid-solid system during annealing. Employing time-of-flight secondary ion mass spectrometry (ToF-SIMS), we found that TFPCBP molecules were uniformly distributed throughout the film (Fig. 2).

We also studied the passivating power of the molecules and found that TFPCBP enhanced the photoluminescence quantum yield (PLQY) of perovskite films. Kelvin probe force microscopy-based studies improved film homogeneity seen in a lowered root-mean-square deviation of contact potential difference across films.

Figure 2. Characterization of perovskite films. a. SIMS depth profile of C₂₁H₂₃F₃⁺ for the perovskite film with TFPCBP. b. PLQY of the fabricated perovskite films. c. Kelvin probe force microscopy potential distribution of perovskite films over a 5 × 5 µm² region.

We leveraged these insights to fabricate high-performing perovskite absorbers. In an n-i-p device configuration, we improved the cell efficiency to 25.6 % (fast-scan) (Fig. 3). When deploying the new liquid crystal strategy in perovskite solar modules, we achieved PCEs of 22.4% and 21.8% (fast-scan) with aperture areas of 15 cm² and 31 cm², respectively.

We also obtained external certification of the new modules via the National PV Industry Measurement and Testing Center (NPVM). We document as a result maximum-power-point PCE of 21.1% for 31 cm² modules consisting of nine subcells. This is the highest certified stabilized PCE reported among n-i-p perovskite solar modules exceeding 10 cm².

Figure 3. Photovoltaic performance of PSCs and perovskite solar modules. a. J-V characteristics of the best-performing device. b. J-V characteristics of two perovskite minimodules with aperture areas of 15.28 and 30.86 cm², respectively. c. Statistics of certified stabilized PCEs as a function of device area of n-i-p perovskite solar modules.

Our molecular strategy also led to enhanced operating stability of devices operating at the maximum power point under 1-sun illumination (Fig. 4). The reverse bias stability was also improved, as was the damp-heat stability of encapsulated modules under 85% RH at 85°C – this reaching 1200 hours. This represents the longest-lived damp-heat stability among n-i-p perovskite solar modules.

Figure 4. Stability of PSCs and perovskite solar modules. a. Continuous MPP tracking of the unencapsulated small-sized PSCs under 1-sun illumination at 25°C with a N₂ flow. b. The reverse bias stability of unencapsulated perovskite modules (8 subcells) biasing at the -V_mpp in ambient air of ~30% RH in the dark at room temperature. c. Stability of the encapsulated perovskite modules stored in ambient air with ~85% RH and heating at 85°C. d. Statistics of the combination of certified PCEs and damp-heat stability of n-i-p perovskite solar modules reported in the literature.

For more details, please check out our paper “A thermotropic liquid crystal enables efficient and stable perovskite solar modules” in Nature Energy (https://www.nature.com/articles/s41560-023-01444-z).