Toxicity reduction of ZnO cauliflower-like structure through trivalent neodymium ion substitution and investigation via computer vision and AI image analysis

Abstract

The synthesis of ZnO and ZNDO NPs was carried out via simple chemical precipitation. The synthesized ZnO and ZNDO NPs exhibit hexagonal wurtzite structures from the XRD patterns. In the FESEM analysis, the ZnO and ZNDO NPs formed spherical (stone-like) and nano (cauliflower) structures. The chemical composition was identified by EDX analysis. The PL spectrum identified the surface defects, such as zinc and oxygen vacancies. The ZnO and ZNDO acquire radical rummaging and antioxidant behaviors as estimated by DPPH free radicals, H2O2 radicals, decreasing in power, and hydroxyl scavenging techniques. Our observations imply that ZnO and ZNDO were excellent platforms to scavenge the ROS, and there was a impressive prospective for the chemically created ZnO and ZNDO NPs as a source of antioxidants. From the hemolytic studies, Nd3+ ion ZnO NPs changed the material matrix to slow down the release of Zn2+ ion in the ZNDO NPs. It was indicated that the Nd3+ ion decreases the cytotoxicity of ZNDO compared to the ZnO NPs. Cytotoxicity analyses were carried out for ZnO and ZNDO NPs using health fibroblast (L929) cells; ZNDO NPs exhibit minimum toxicity compared to the ZnO NPs. Furthermore, the pixel perspective of the NPs through computer vision tools gave us the image-based surface morphology that correlates to the relevant toxicology studies. The customized pipeline of AI and computer vision curated algorithms such as autoencoder based data augmentation for sample size enhancement, roughness-induced porosity estimation, roughness metrics Ra and Rq, Edge density estimation using Sobel gradients were put forward to investigate the component's non-toxic behavior from a pixel perspective.