Dehalogenation

[hogue %-% at %-% canada.den.mmc.com (Pat Hogue 1-2183)]

 Opinions are solicited on the possible mechanism for dehalogenation described
 below.
      Perfluoropolyethers (PFPEs) have been shown to react even at cryogenic
 temperatures with atomically clean aluminum and iron surfaces.  PFPEs also react
 with (most probably) Lewis acid surface sites at elevated temperatures.  In
 both cases AlF3 or FeF3 and carbonyl fluoride (COF2) is observed.  There is also
 some evidence for ALF6-3 species.
      Is fluorine anion abstraction to Lewis acid site the initiating step?  Or
 is atomic fluorine transferred instead?
      My AM1 model of the small polyether represented by
  F3COCF2OCF2CF2OCF2*CF2OCF2CF3 (where the asterisk represents the location of
  the Fluorine anion abstracted, it has the largest HOMO coefficient)
  indicates that these coefficients changed from a state of delocalization
 over the central portion of the molecule to a state
  where the coefficients are concentrated on the three terminal fluorines and
 oxygen at one end of the molecule.
      I propose that three fluorine atoms leave the terminal -O-CF3 group to form
 FeF3.  I think the resulting carbon triradical can undergo rearrangement with
 oxygen to produce carbon monoxide (also reported by some workers).
  Any opinions?  How about a second three atom transfer to form the hexavalent
 fluoride?