Activation-Free Liquid Metal Bioelectronics Enabled by a Reconfigurable Island-Bridge Nanoarchitecture for Machine-Learning-Assisted Activity Recognition

Abstract

Patternable liquid metal particle (LMP) conductors with stable performance under deformation are highly desirable for wearable electronics but remain limited by activation-dependent conductivity and structural instability. Here, we report an activation-free conductive ink based on nickel-coated carbon nanotubes (NiCNTs) and nanoscale LMPs. A solvent-mediated stabilization strategy suppresses droplet coalescence, yielding uniformly dispersed and stable LMPs. Strong interfacial coupling enabled by hydrogen-bond-assisted adsorption and Ni–Ga alloying promotes the formation of a dynamically reconfigurable “island-bridge” network, where conductive LMP domains are interconnected by compliant NiCNT bridges. This architecture achieves high electrical conductivity (5.9 × 105 S m–1) without external activation and maintains excellent electromechanical stability (4.6 × 105 S m–1 at 100% strain) and printability on various soft substrates. Printed epidermal electrodes enable low-impedance electrocardiogram monitoring over 7 days and support machine-learning-assisted activity recognition with accuracies above 99%, providing a general strategy for intrinsically conductive and mechanically resilient soft electronics.