I bond a crosslinked SU-8 substrate to an unexposed Su-8 layer to make sealed ufluidic channels. It's easy once you get the hang of it. This technique was borrowed from Rebecca Jackman from MIT. She has a couple of good papers out describing the process. My substrate is 10um thick and my sealing layer is 5um thick. Jackman uses much thicker resists, but my temperature are the same.