# Structure–Function Coupling in Pyridyl Triazole Copolymers for Neuromorphic Synaptic Transistors

## TL;DR

This paper explores how different chemical structures in copolymers affect the performance of neuromorphic synaptic transistors for image recognition.

## Contribution

The study introduces new pyridyl triazole copolymers and shows how their structure influences synaptic behavior and neural network performance.

## Key findings

- The copolymer with a benzothiadiazole linker achieved nearly 80% image recognition accuracy.
- The fluorine-substituted thiophene linker showed no synaptic behavior.
- Interface trap density and morphology directly impact synaptic device performance.

## Abstract

Organic ferroelectric
transistors are excellent candidates as low-cost
alternatives for synaptic devices. Specifically, interfaces with donor–acceptor
semiconducting polymers and copolymers of poly­(vinylidene fluoride)
(PVDF) are attractive for mimicking synaptic responses. By tailoring
the linking unit between the pyridyl triazole (PyTr) acceptors and
thiophene donors, three copolymers are synthesized incorporating selenium-substituted
thiophene, benzothiadiazole, and fluorine-substituted thiophene linkers.
Using the hexafluoropropylene copolymer of PVDF (PVDF-HFP) as the
dielectric layer, the three PyTr semiconductors show p-type transport
in transistor architectures with carrier mobilities between 0.1 and
0.2 cm2 V–1 s–1. The
synaptic plasticity is investigated by applying long-term pulsed voltages
at the gate electrode to emulate potentiation and depression processes.
To assess their neuromorphic functionality, the synaptic responses
of the devices are tested for image recognition in a multilayer perceptron
neural network. The copolymer with the benzothiadiazole linker achieved
recognition accuracy close to 80%, whereas the one with a fluorine-substituted
thiophene linker shows no synaptic behavior, highlighting the critical
role of the semiconductor–dielectric interface. A detailed
study of the interface trap density and morphology is performed to
identify how interfacial properties directly influence synaptic device
performance.