Buckybowl‐Based Organic Single‐Crystal Photosynapses: Concave Architecture Inducing High Accuracy in Image Recognition

Why It Matters

By proving that molecular geometry can drive high‑accuracy neuromorphic vision, the breakthrough accelerates organic electronics toward practical wearable and biomedical AI applications.

Key Takeaways

• Buckybowl 3S-2Me enables organic single-crystal photosynapse
• Ambient O2 adsorption creates defect levels for persistent photoconductivity
• Devices retain non‑volatile current over 4000 seconds
• Linear multilevel conductance yields ~95% image‑recognition accuracy
• Molecular architecture links structure to neuromorphic performance

Pulse Analysis

Organic photosynaptic devices aim to emulate the eye’s ability to process visual information in hardware, offering a path toward low‑power neuromorphic vision systems. Traditional approaches have focused on circuit engineering, yet the intrinsic properties of the active material dictate speed, retention, and energy efficiency. The recent introduction of a buckybowl‑derived molecule, 3S‑2Me, marks a shift toward molecular‑level optimization. Its concave geometry not only stabilizes single‑crystal growth but also creates a unique surface that interacts strongly with ambient gases, opening new design avenues for organic neuromorphic components.

The key breakthrough lies in the defect states generated when oxygen molecules adsorb onto the buckybowl surface. These states act as shallow traps that prolong carrier lifetimes, producing a persistent photoconductivity (PPC) effect. As a result, the fabricated organic single‑crystal photosynaptic transistors retain a non‑volatile current for more than 4,000 seconds and support linearly programmable multilevel conductance. Such characteristics enable reliable synaptic plasticity emulation, delivering image‑recognition accuracies of 95.2 % on an 8 × 8 pixel dataset and 94.8 % on a 28 × 28 digit benchmark.

From a commercial perspective, these metrics position buckybowl‑based photosynapses as strong candidates for next‑generation wearable AI, implantable biomedical sensors, and edge‑computing vision chips. The ability to achieve high classification performance without complex stacking or external memory reduces fabrication costs and power budgets. Moreover, the demonstrated structure‑property relationship provides a template for further molecular engineering, potentially extending the approach to other organic semiconductors. As the neuromorphic market expands, such scalable, high‑accuracy organic devices could accelerate adoption across robotics, smart cameras, and personalized health monitoring.