The IEEE Spectrum article examines the increasing diversity of materials used in Memristor devices, extending far beyond the traditional metal-oxide films employed in early designs. Historically, memristors consisted of a thin layer of titanium dioxide or similar dielectric, sandwiched between two metal electrodes. When voltage is applied, oxygen-vacancy filaments form or break, switching conductivity and encoding memory.
Recent experiments have revealed surprising new material platforms. Researchers at The Ohio State University showed that organic substances, for example, shiitake mushrooms, can exhibit memristive behavior. That raises the possibility of flexible, biodegradable, even radiation-resistant memristors. Other efforts go in the opposite direction: ultrathin, atom-level materials such as layered 2D crystals (e.g., transition metal dichalcogenides) received attention as “atomristors.” These devices can switch resistance by shifting lattice defects rather than oxygen ions, promising lower switching energy, faster operation, and easier integration with flexible electronics.
Beyond materials, the variations in structure matter. Most memristors follow a metal–insulator–metal (MIM) sandwich design. The choice of electrode and insulator materials influences critical performance metrics: switching thresholds, endurance, on/off ratio, retention time, and energy per switch. Some electrodes are inert (e.g., platinum, gold), acting solely as conductors; others, such as silver or copper, actively participate in forming conductive filaments in cation-migration memristors.
Diversifying memristor materials could unlock advantages for emerging applications. Organic and flexible memristors may lead to bio-integrated or disposable electronics. Two-dimensional material memristors promise ultra-low-power, high-density memory or neuromorphic hardware. Meanwhile, composite and ferroelectric variants may offer improved stability, endurance, or multi-state behavior. Collectively, this material innovation expands the design space for memory and computing hardware, moving memristors beyond niche lab experiments toward future memory, AI, and neuromorphic systems.