In medical text mining, Named Entity Recognition (NER) and Relation Extraction (RE) play a vital role in extracting information from published medical articles. However, these traditional approaches like dictionary- or rule-based methods and machine learning heavily rely on large-scale dictionaries, target-specific rules, or well-constructed corpora. These methods to NER have been superseded by the deep learning-based approach independent of hand-crafted features. The progress of deep learning in graphs has enabled the use of dependencies in text via the application of a Graph Convolution Network (GCN) on the dependency tree. This proposed research NR-GCN investigates the use of Graph Convolutional Networks (GCN) in Named Entity Recognition (NER) and Relation Extraction (RE). The named entities and relationship (NR-GCN) model, which focuses on the nanomaterial (Hydroxyapatite) domain, is implemented based on available research and is compared to the current leading approach. It examines many features that NR-GCN may perform within the overall architecture of the model to improve the performance of both tasks. This work proposes the Graph Convolution network to extract contextual elements further. The research also conducts a comparison between the NR-GCN, GCN, and non-GCN models and concludes that the NR-GCN model exhibits enhanced performance. The experimental findings demonstrate that the enhanced model surpasses previous comparison network models in terms of evaluation metrics such as Recall, Precision, and F1 score.