Survey papers are useful and will give you citations if well written.
But for a PhD, you need to add to the existing body of knowledge, not merely reorganize it. Unless it is a massive contribution, such as a reformulation or neater repackaging of existing knowledge (for extreme examples, consider parts of Euclid, Cartan, Wielandt, the latter of which merely found a shorter proof of a known theorem, but it was much shorter), which itself is a scientific contribution, it's a very clear no.
In fact, if you are able to create a simplified map through a difficult and ill-understood landscape, that's definitely a PhD.
Others talked about new tools as worthy of a PhD, in a way, this is a similar situation. Identifying what makes a field hang together is a new tool. Done well, it can change a field.
However, this is most likely to be far harder than to pick a concrete problem and study it. However, it is also far more than just a literature review.