"Freedom is paramount, nothing else matters" is the only valid starting point. Liberating software is against that normal trend to protect, or profitize, one's work. Your scientists have put a lot of work into their product, and the proprietary 'aspect' must be a very confirmative one, unless exactly they have unlimited or plentiful ressources, including time and engagement, at their hands. Freedom is craving, so if not "Freedom is paramount, nothing else matters", then your discussion will be difficult.
I am not well versed in project management, so my suggestions would be speculative - why not just release the software as free, perhaps under a different name, and continue updating the proprietary project?