Academic Search Engines (ASEs) are crucial for navigating the vast landscape of scholarly literature. Traditionally, these engines rely on keyword-based search, supplemented by predefined facets encompassing metadata such as research field, publication year, type, authors, and language. However, ASEs are limited in their ability to generate dynamic facets in real-time based on article contents. This limitation impedes the efficient exploration and navigation of large article collections. We propose an approach that addresses this limitation by dynamically generating facets using article abstracts. We introduce three distinct methods for dynamic facet generation: (1) KB2 (based on Knowledge Bases) utilizes two knowledge bases (KB) to extract facet values and their associated facets; (2) KBLLM (based on a Knowledge Base and a Large Language Model) utilizes a KB for extracting facet values and a large language model (LLM) to categorize these values by predicting facets; finally, (3) KBLLMKA (based on a Knowledge Base and a Large Language Model with Knowledge Augmentation) combines KB-spotting with facet-value pair extraction and adds this information as auxiliary data to enhance LLM’s facet prediction capabilities. We evaluated the effectiveness of these methods with a user study, performance evaluation, and comparative analyses, which showed the effectiveness of the approach.