A Flexible Piezoelectric PVDF/MXene Pressure Sensor for Roughness Discrimination

The accurate assessment of surface roughness is critical for numerous applications ranging from quality control in manufacturing to material characterization. To achieve a functional capability of roughness perception, a flexible pressure sensor based on polyvinylidene fluoride-Ti3C2 (PVDF/MXene) nanocomposite is developed. The sensor consists of electrospun PVDF nanofibers embedded with 2-D MXene nanosheets. The MXene enhances the piezoelectric \beta -phase content of the PVDF up to 97.2% at optimal loading of 2.5 wt%. The PVDF/MXene nanocomposite exhibited high piezoelectric voltage sensitivity up to 0.059 V kPa ^{-{1}} under applied pressures. The wavelet transform analysis of signals obtained by scanning the sensor on sandpapers of varying roughness showed distinct time-frequency patterns corresponding to different surface roughness levels. Unsupervised dimensionality reduction using t-distributed stochastic neighbor embedding (t-SNE) revealed clustering of roughness data into distinct categories. A convolutional neural network (CNN) classifier achieved 98% accuracy in categorizing the surface roughness based on the sensor signal wavelet transforms. The piezoelectric nanocomposite sensor shows promise for surface metrology applications.