Innovative-self-powered e-tattoo sensor: A new frontier in human–machine interaction
João C. L. Folhadela, Miguel A. S. Almeida, João. P. Maganinho, Ana L. Pires, André M. Pereira · APL Electronic Devices · 2025
The emergence of Electronic Tattoos marks a significant shift in non-invasive health monitoring and human–machine interaction. In this work, we present the development of a Self-powered Intelligent e-Tattoo capable of localizing varying thermal stimuli, serving as a versatile sensor for temperature anomalies and touch, using the concept of the thermoelectric phenomenon. By employing a planar design crossing screen printed stripes of p-type and n-type Bi2Te3, the fabricated device was able to monitor up to 21 distinct positions. The design was validated via numerical simulations with COMSOL Multiphysics, leading to the training of a Multi-Label Classifier Neural Network that achieved 99.5% accuracy in predicting heat applied to combinations of nine positions. Herein, the prototypes were microfabricated using screen-printing techniques on both Kapton (KS) and temporary Tattoo Paper (TPS) substrates. These prototypes, which combined Bi2Te3 powder with H3PO4-doped PVA, demonstrated remarkable Power Factors (PF) of 12.83 μW K−2 m−1 (n-type) and 1.45 μW K−2 m−1 (p-type), particularly on KS substrates. The collected device data served as pivotal training sets for Multi-Class Classifier Deep Neural Networks (DNN), yielding exceptional F1-scores of 94% and 85% for the KS and TPS devices, respectively, on test sets. Furthermore, a comprehensive study involving several individuals corroborated the e-Tattoo’s effectiveness as an efficient touch sensor. By utilizing the developed DNN models, this validation achieved an impressive F1-score of up to 78%. These findings underscore the growing promise of e-Tattoos in revolutionizing human–machine interactions, indicating a significant shift toward improved personal health monitoring and digital engagement, offering autonomous operation without the need for external power sources.