Civilian reactor plutonium cannot be made weapons grade with chemical separation because of contamination with Pu-240 (which occurs when Pu-239 absorbs a neutron as is likely to happen if the fuel spends a long time exposed to neutrons as it is in a civilian reactor). Pu-240 is a problem for several reasons - it has a high rate of spontaneous fission which prematurely detonates the bomb, its decay products damage the bomb, and it produces a lot of heat.
All of these issues can be worked around. The premature detonation can be mitigated by clever implosion designs, the bomb can be cooled and the pit replaced more frequently. However, all of this increases maintenance costs and reduces reliability and readiness. Nobody who has a choice in the matter would use reactor grade plutonium to make weapons.
It should be possible to purify reactor grade plutonium to weapons grade using laser isotope separation. Unfortunately, there's very little publicly accessible about this topic, and what little information there is discusses it in the context of uranium enrichment. The predominant isotope separation technology, centrifugation, isn't used with plutonium because the mass difference between Pu-239 and Pu-240 is too small (in comparison to U-235 and U-238).
Yes. In China's case this was done in Plant 404 at Jiuquan. There's been some activity there and at a nearby industrial park
The reprocessing facilities are being expanded; however, the reactor itself remains idle.
I expect that what China is doing in expanding its nuclear arsenal is sequencing construction such that the most unambiguous (and thus most provocative) actions are done last. For example, expanding plutonium reprocessing facilities is necessary and has dual civilian/military use (so China can bullshit about it), therefore it's done first. However, refurbishing a graphite moderated production reactor very clearly telegraphs one's intent, so China would do this near the end.
