Scientific paper summarization has gained rapid growth in attention along with explosive demand for scientific research, especially in fields akin to chemistry, which is more challenging than typical summarization due to specialized terminologies, domain-specific information, and requirements on vital needs of scientific researchers. Inspired by the prominent capabilities shown by graph-structured data in conveying self-defined and a variety of information, we propose to leverage knowledge graphs to enhance scientific summarization in a chemistry scope. In this paper, we first construct a chemistry paper summarization dataset, Hydrogen Evolution Reaction with Carbon (HERC), which includes 2,254 selected chemical research papers under chemistry expert guidance. Each paper is associated with its citation information and 9 categories of expert-labelled knowledge. Based on that, we create citation and knowledge graphs for each paper. Furthermore, we develop a knowledge-aware summarizer that incorporates the information from both the citation and knowledge graphs and showcases the improved performance with the usage of knowledge graphs.