[Wien] Ghost bands & NiO surface

[L. D. Marks]

I'm failing to calculate a NiO surface. It does not seem to matter what
mode (GGA, LDA+U, afm...) I'm getting what look like ghost-bands. The most
obvious feature that I can see is the Ni 3p which instead of having two
levels split by ~0.2Ryd gets split by > 1Ryd. The charges then fluctuate
wildly, and the calculation diverges. N.B., this is not an RMT issue -
I've tried various options.