Solvophobic interaction enabled nano-homogeneous perovskite@polymer films for high-performance photonic synaptic transistors

Semiconducting perovskites and conjugated polymers as two extensively studied optoelectronic materials hold high promise for developing photonic synaptic devices for artificial intelligence. However, their inherent immiscibility hinders charge transfer between them within a hybrid film, thus leading to low device performance. Here, we propose an ultrasound-assisted solvophobic interaction strategy to homogenize perovskite@conjugated-polymer composite films to a nanoscale by exploting the critical role of a minimal amount of poor solvent. We demonstrate that not only can this poor solvent enhance electrostatic and van der Waals interactions between the perovskite quantum dots (QDs) and the polymers, more importantly, it also significantly reduces the surface energy difference between them, thereby improving miscibility and suppressing the self-aggregation of quantum dots. Consequently, the interfaces between the perovskite QDs and polymer crystalline domains are significantly expanded, thus considerably enhancing charge separation and reducing transport barriers, with a charge transfer efficiency up to 71.70 % and an ultralow energy consumption of 4.1 aJ. The device exhibits a broad spectral response ranging from 254 nm to 850 nm. Furthermore, these unique features of the composite films allowed an impressive demonstration of advanced artificial neural pathway and neuromorphic computing simulation. This work provides a novel approach for developing ultralow-energy photonic synaptic transistors and integrating multifunctional applications toward the artificial intelligence.