Machine Learning-Aided All-Organic Air-Permeable Piezoelectric Nanogenerator

Abstract

All-organic piezoelectric mechanical energy harvesters display an excellent electrical output with higher sensitivity due to the superior electrode compatibility between active materials and organic electrodes in comparison to that of metal electrodes. Herein, a stretchable, breathable, and flexible all-organic piezoelectric nanogenerator, made up of PVDF nanofibers and δ-PVDF nanoparticles, fabricated through the electrospinning process in a single step, has been demonstrated for prospective machine learning applications. The δ-phase PVDF nanoparticles serve as efficient active piezoelectric and ferroelectric components with a piezoelectric coefficient of ∼13 pm/V. In terms of electrical response, a peak-to-peak ∼VOC of 4 V, ISC of 1.8 μA, and maximum power density of ∼1600 μW/m2 were obtained. The fabricated device also exhibits excellent stretchability and air permeability, enabling the properties of robust wearable devices with a water vapor transmission rate of ∼250 g m–2 day–1. Here, we have shown that a machine learning algorithm proposed for the different finger motion responses can predict with 94.6% accuracy. Thus, it could recognize different finger gestures efficiently with the highest possible accuracy and predict the possible source point. This feature could be advantageous for prospective health care and security purposes apart from the device and sensor applications.