At least in my field (theoretical computer science), the corresponding author is typicallyoften (but not necessarily) the person who did most of the work and prepared the final version of the paper. It is therefore completely normal to have PhD students as corresponding authors.