The evolution of inkjet-printed electronics has opened transformative pathways for intelligent sensing platforms that seamlessly integrate photonic detection with in-memory computing. This work presents a holistic approach to scalable optoelectronics, bridging the gap between robust logic-in-memory architectures and optically tunable synaptic plasticity for next-generation biosensors.

Our work began with the development of fully inkjet-printed metal-insulator-metal (MIM) structures using high-k HfO₂ dielectrics. These devices demonstrated low-power, non-volatile switching, strong retention, and uniformity suitable for passive memory and selector applications, establishing a scalable platform for printed memory arrays. Building upon this, we transitioned to 2D materials, particularly hexagonal boron nitride (h-BN), achieving devices with high endurance, reproducibility, and tolerance to stochastic variation. Establishing the computational backend, we first demonstrate fully inkjet-printed hexagonal boron nitride (h-BN) memristors capable of current-controlled stateful MAGIC logic as h-BN used for thee logic-in-memory. By utilizing the volatile and non-volatile resistive switching behaviors of Ag/h-BN/Au structures, we achieved high endurance (up to 250,000 cycles) and improved input stability for NOR gate operations. This provides the necessary logic density for processing sensor data directly in memory without the Von Neumann bottleneck.

This foundation set the stage for the next leap: the integration of halide perovskites as multifunctional active layers. Leveraging the photoelectric tunability of Sn-based and CsPbBr₃ nanocrystal perovskites, we demonstrated a new generation of printed devices capable of modulating light in response to electrical and optical history. In particular, TEA₂SnI₄, PEA₂SnI₄, and FASnI₃-based devices were inkjet-patterned within vertical cavity and multilayer structures to exhibit persistent photoconductivity, state-dependent photoluminescence, and feedback-tunable random lasing—hallmarks of optical memristors and photonic synapses.

Bridging the gap between logic and light, we investigated the memristive properties of CsPbBr₃ perovskite devices within a p-i-n diode architecture (PEDOT:PSS/NiO/CsPbBr₃/SnO₂). Beyond their function as LEDs, these devices exhibit reversible resistive switching driven by field-assisted ion migration. Crucially, we demonstrate that this synaptic plasticity is optically tunable: illumination lowers the activation energy for ionic motion, reducing switching thresholds and modulating the hysteresis loop. This dual-mode response allows the material to function as an artificial synapse where weights are dynamically adjusted by visual input, mimicking biological neuromodulation and allowing optical control of synaptic plasticity.

Culminating in the sensory frontend, we optimized this perovskite system into a mixed-phase CsPbBr₃/Cs₄PbBr₆ nanocrystal architecture integrated with chemical vapor deposition (CVD) grown single-layer graphene (SLG). Through precise annealing control, we engineered a "raisin-bread" structure where photoactive CsPbBr₃ nanocrystals are embedded within a protective Cs₄PbBr₆ matrix. This synergistic design yields exceptional device performance:

By uniting the computational robustness of h-BN logic, the adaptive learning capabilities of optically controlled perovskite synapses, and the superior detectivity of graphene-enhanced quantum dots, this work establishes a scalable, material-efficient route toward monolithic intelligent optical biosensors capable of in-sensor computing.