Proceedings of MATSUS Spring 2025 Conference (MATSUSSpring25)
DOI: https://doi.org/10.29363/nanoge.matsusspring.2025.400
Publication date: 16th December 2024
Material and technological research in neuromorphic computing has garnered significant attention in recent years, with numerous examples of memristors serving as simple electrical synapses. In addition to electrical input and output, light is increasingly used as both an input and output signal in optical memristors, where light transmission is modulated.
In my talk, I will introduce a new concept for an optical memory device based on photoluminescence, termed the "memlumor"—a luminophore with memory.[1] The photoluminescence quantum yield of a memlumor "remembers" the history of previous excitations through parameters governing its photophysics and photochemistry. Metal halide perovskites have proven to be highly promising materials for memlumors. We demonstrated the synergetic coexistence of both volatile and non-volatile memory effects in perovskites over a broad timescale, ranging from nanoseconds to days.[1] We identified the origin of this complex response as the phenomena of photodoping and photochemistry triggered by light input, both of which are closely tied to defect states and their photoinduced dynamics.[2] Remarkably, defects and their temporal dynamics—typically viewed as detrimental to device performance—are essential for memlumor operation.
I will showcase the memlumor properties of CsPbBr₃, which exhibits memory effects over timescales from nanoseconds to minutes, with switching energy as low as 3.5 fJ. Additionally, I will discuss a novel method of multi-pulse time-resolved photoluminescence,[3] specifically suited to study trapped charge carriers in luminescent semiconductors and memory effects induced by them.
Memlumors, as novel optical dynamic computing elements, have a potential to provide a new dimension to existing optical technologies, paving the way for their application in photonic neuromorphic computing.