In this space ads may actually be able to make money since the audience is very targeted. You even know what field everyone is in by looking at what papers they click on and such. It seems easy to have a free+ads plan and an $10/mo no ads plan.
if you could convince labs themselves to pay for it, it'd be a money maker, although this may require more direct marketing at first. many academic institutions hold subscriptions/licenses in the $10k range.
I wouldn't bet on it. Making money in the academic institutional licensing world is tough, especially if you are marketing something that doesn't impact bottom-line figures for people who are concerned. Arguably this use case is as far removed from bottom line value adds as possible. In an optimal world you have a killer app that reduces University lead COA by 30% or something, but this would have to grow organically (so a lot of direct sales approaches are no good).
A friend of mine has been doing a lot of direct sales (quite successfully I might add) to institutions, but the reason he's making money is that his target audience is already doing a lot of number crunching, they have big budgets, and he's directly saving them cash. Something like this is probably a "nice to have" feature that would be near impossible to convince the school to pay for, to speak nothing of a worthy price tag (greater than $2k/yr for instance).
I sounds like a fun project, but I would not expect it to pay much, if anything. And an ad model's monthly revenue would be a joke.