I think there are 3 explanations why you don't see many "cutting edge" research questions on internet communities
If a scientist finds a really good question, he should and will try to solve it and this should only be a matter of time and earn him a publication, it's his job to develop and answer good and interesting current research questions. That's why you see most top rep user on SE mostly answering questions in their field of expertise and asking questions in their hobby fields
Some months ago, theoreticalphysics.se was closed. There were a lot phd's and professionals registered, but they didn't generate any questions. I made a longer analysis here. One important point is that older but also younger scientists mix up asking a question with ignorance. On biology.se I saw also many good questions from the top and professional user related to research (more interesting to me than reading physics.se with many reddit one liners). This is probably due to the fact that biology is a even broader and more diversified topic than physics and its even harder to be a master of all underlying factors and theories like genetics, biochemistry, bioinformatics, biophyiscs etc for exploring biological systems. So it's a more interdiscplinary field where scientist have to rely on knowledge of other experts and it's normal to ask sub and side-questions to progress with your own main question. Many physicists work in biology research groups, but not the other way around. So there are fields like biology or theoretical computer science (programmers love forums ;) ) that have probably a culture of asking each other research or related to research questions. Personally I think the quality of questions on biology.se is pretty good for such a broad topic, raising the quality means you have to make a sub-topic site and community
It was clear to me that chemistry.se (as a not popular science and in between branch of physics or biology) will have a hard time to fulfill the SE beta metrics, the stats don't look very healthy currently. A lot of the theoretical questions in chemistry are already answered on physics.se (as chemistry is basically only a expansion of physics, a lot of overlap in study courses apart from mathematics). The remaining questions on chemistry.se are therefore very localized and relate to practical problems of chemistry students and need very special expertise to be answered. There aren't many practical questions on physics.se, it's hard to drive a community with such questions, maybe with 5000 or more contributing user. But you cannot attract this amount of experts with current AREA51 system. A community can only grow with common interests, if the topic is big, but the questions are very localized, it does not work if there aren't enough experts for such niche subtopics. For example a site like biostar dealing with bioinformatics seems more healthy and useful than bioinformatics-tagged questions on biology.se or stackoverflow. Normally the AREA51 system should show if there is enough interest to start and maintain such a community, but currently it only works for programming and programmer's hobbys. Even if they increase the requirements metrics for non-programmers topics to guarentee a healthier community, after some months they loose their interest or knowledge of commiting to a proposal. A proposal like chemistry or cognitivesciences lasting over 12 months has probably lost most of its commiters when starting beta. They should auto-kill such proposals and the whole process has to be restarted to gain interested user in a maximum period of time (around 6 months). Without initial dynamics sites like chemistry, theoreticalphysics and cognitivesciences are doomed to not fulfill the beta graduating requirements.
The theoreticalphysics.se user wanted to start a Q&A forum on their own with qsqa software after the closing. It seems they abandoned the idea and no one of the many former registered user is complaining. Seems there is not really a demand for a research questions site in physics, setting up a server cannot be the problem for couple of smart guys. I currently think the best you can do here is a expert site with very localized questions relating to research, but no research questions, but this isn't really enough to build a healthy community.
The logical progression would be, take a site like phyiscs.se, wait till it has 1000 experts, than start smaller healthy proposals with those experts. The problem, a site like physics based on reputation and voting is unlikely to attract 1000 experts, this a problem of limited attention, you only participate in a gamification system if you can earn much rep with reasonable effort, so competing with 1000 other experts on the few questions is pointless and a waste of time for most. 100 experienced user are enough to answer most questions on physics.se within an 1 hour!
The conclusio of this whole analysis is that you will never see healthy highly specialized knowledge forums on the internet for experimental sciences. There is no working community growth model for such sites. The competition on reputation doesn't allow more than 100-1000 user on such sites which is below the critical mass for broad experimental natural sciences. 100-1000 seems to be enough for high quality questions in programming and bioinformatics, but not for much broader topics like physics, chemistry and biology.
I'm eager to hear counter-arguments, I really try to find them myself as a advocate of open science. I think the best we can achieve as scientific community is open access journals with commenting function below online published articles where people can exchange information, questions, answers to very localized problems. Trying to get such scientists together via a very broad internet community site like quora or SE seems impossible to me currently. A content filtering problem besides the attention problem of such gamification sites.