DPP-DTT Nanowire Phototransistors for Optoelectronic Synapses in EMG and ECG Signal Classification.

Abstract

A neuromorphic phototransistor based on nanowire-patterned diketopyrrolo-pyrrole-dithienylthieno[3,2-b]thiophene (DPP-DTT) is reported. The nanowires, well-aligned with a width of 460 nm, spacing of 8-11 µm, and height of ≈80 nm, are fabricated using the stamping method of soft lithography and exhibit optically stimulated synaptic behavior. Under blue illumination (455 nm, 0.55 mW cm-2), a photogating effect arises at the DPP-DTT/SiO2 interface, leading to threshold voltage shifts up to 6.4 V as a result of electron trapping at the interface. Negative gate pulses (-7 V) facilitate recombination of the trapped electrons, inducing detrapping and consequently leading to a decrease in the threshold voltage. These two behaviors effectively emulate the processes of potentiation and depression. Efficient trap-detrapping dynamics are facilitated by the unique geometry of the nanowire. Synaptic plasticity is modulated by adjusting stimulus intensity (light pulse: 0.26-1.42 mW cm-2, gate pulse: -6--9 V), duration (0.3-2.1 s), frequency (0.47-3.33 Hz), and repetition (1-40 cycles), supporting transitions from short- to long-term behavior. The device is evaluated through artificial intelligence classification tasks, including image recognition and time-dependent physiological analysis. It achieves the classification accuracies of 97.4% for MNIST, 93.4% for electromyography (7 classes), 89.0% for electrocardiography (5 classes), and 83.8% for CIFAR-10.