An Empirical Analysis of LLMs for Countering Misinformation