George,
You problem is similar to fluoration
metal-reach samples. We use another
path for fluorating refractory minerals.

According to Ustiov (1991, patent from 1980)
we adding to sample of mineral pure and dried
KCl. The refractory minerals with KCL at the
presence CLF3 (KBr for BrF5) gives an
eutectically liquid and extraction of oxygen
are 100 %.

My comments to this method: we take about 10-15 % KCl
(KBr) from a sample mass and grinding  to powder
under CCl4 (no alcohol). The ammount of ClF3 (BrF5)
must be taken accordingly the KCl (KBr) presence
and we take the amount of ClF3 multiplied to 2-3
times, as for reaction needs. The Ni-vessels we heat
up to 700 C. After that all must be OK.

And at the end. We found about 7 years ago that
the Ni-vessel size are important. We use a long
cylindrical fingers with small inner diameter.
We heat only end of fingers, but the upper
part of fingers are upon "cold" zone, created by
cold water flow. The gas convection and mixind
gives us a fast and complete reaction. 100-% yield
we obtain after 30 minits of heating maximum.

Best regards,
Elena