Recovered graphene-hydrogel nanocomposites for multi-modal human motion recognition via optimized triboelectrification and machine learning

Thien Trung Luu, Hai Anh Thi Le, Yoonsang Ra, Teklebrahan Gebrekrstos Weldemhret, Hwiyoung Kim, Kyungwho Choi, Dongwhi Choi, Dukhyun Choi, Yong Tae Park

Research output: Contribution to journalArticlepeer-review

Abstract

Hydrogels have extensive applications in portable, flexible, wearable, and self-powered electronic devices based on triboelectric nanogenerators (TENGs). An important issue with hydrogels is their tendency to dehydrate over time, which leads to a decline in both ionic conductivity and mechanical flexibility. Furthermore, the current techniques used to produce these hydrogels mostly rely on the freeze–thaw process, which has limited ability to modify the polymer conformation. Herein, a novel water-assisted recovered hydrogel is proposed using a simple strategy to prepare high-performance hydrogel-based TENGs by optimizing the cross-linking and crystalline domains. Synthesis of the electrostatic electrode in the TENG involved the incorporation of polyethylene oxide (PEO) into a polyvinyl alcohol (PVA) hydrogel network via cross-linking. Graphene nanoplatelets (GNP) were added to precisely tune the electrical conductivity. GNP constructs the backbone structures in the hydrogel and enhances the charge transport capacity. Electrical conductivity is changed by the GNP concentration and thus, electrical output of the hydrogel can be facilely controlled. The water reabsorption increased density and crystallinity of the cross-linking and allowed the hydrogel to show superior performance compared to the original one. The 7th recovery hydrogel produced around 594 V, 40 μA, and 32 nC. The 7th recovery hydrogel had exceptional endurance, with the capacity to withstand over 16,000 cycles of contact separation. Moreover, it could be stretched up to 541 % of its original length and improved by almost twice as much as that without the recovery process. By combining multi-modal graphene-based TENG sensors with machine learning, a state-of-the-art behavioral monitoring system was created that could reliably detect tapping fingers with an average accuracy rate of 95 %. The findings of this research will pave the way for new approaches to the development of autonomous motion sensors and flexible renewable energy sources.

Original languageEnglish
Article number111997
JournalComposites Part B: Engineering
Volume291
DOIs
StatePublished - 15 Feb 2025

Keywords

  • Graphene
  • Hydrogel
  • Machine learning
  • Recovery
  • Triboelectrification

Fingerprint

Dive into the research topics of 'Recovered graphene-hydrogel nanocomposites for multi-modal human motion recognition via optimized triboelectrification and machine learning'. Together they form a unique fingerprint.

Cite this