Would it have to be liquid nitrogen? Could liquid CO2 be a substitute?
http://www.science.uwaterloo.ca/~cchieh/cact/c123/phasesdgm.html
A diagram would be nice but will do what I can with words. Use dry ice (commonly avb) by a feed chamber, say 30-50mm diam and 200mm long cylinder form. A cap would seal from the top to create pressure, once at 8 atm pressure it will turn liquid activating a float valve to drop into a pressurized resevoir. Once drained, the feed chamber would be isolated from high pressure, allowing a bleed valve to release pressure so it can be refilled with another cylinder of dry ice. The stator winding laminates would be hollowed out to allow the liquid to settle into the lower laminates and a pressure relief valve would be installed in the upper winding (outrunner design), allowing the evaporating gasses to be released to atmosphere.
As for gluing the magnets, get rid of the glue. Make the magnets form fitting to the housing in a semi circular "keystone" shape.