Radical HOMO/LUMO

[hogue "at@at" denmark.den.mmc.com (Pat Hogue 1-2183)]

 Dear Netters:
      I'm coming up on a MS thesis defense and could appreciate
 any dear souls willing to consider whether I am interpreting
 my results correctly, especially the radical.
      As near as I can tell there are four ways a linear
 perfluoropolyether molecule can decompose on a surface:
 1. Lewis acid complexation with ether oxygens, resulting in
 trans-oxygen fluorine transfer, molecular disproportionation and
 subsequent thermally induced decomposition of fluorofomate to F2CO gas.
 The Lewis acid could be Al2O3.  This is Paul Kasai's mechanism.
 2. Lewis acid abstraction of fluorine through attack on F
 lone pairs to leave a cation which can be stabilized by
 resonance with ether oxygens, this is especially likely
 for an acetal (-O-CF2-O-) group.  Subsequent depolymerization
 can also release the F2CO gas always found experimentally. The Lewis acid FeF3
 fills this bill.  D. Carre, W. Jones, M. Zehe and O. Faut mechanism.
 3. Bronsted acid attack on ether oxygen, presumably to cleave the bond
 and result in some kind of surface bound species and alcohol fragment.
 Gama Fe2O3 works here.  W. Morales (NASA, Lewis).
 4. Nucleophillic (electron donation) on fluorine, resulting in
 abstraction of F- leaving a polymer radical that undergoes radical
 depolymerization, also releasing F2CO.  This mechanism
 probably operates on bare metal (Al, Fe) surfaces. (P. Stair, M. Napier)
      My thesis only looks at mechanisms 2 and 4.
      I recently modelled a small perfluoropolyether with the
 following formula (CF3OCF2OCF2CF2)2O using AM1.  I then removed one
 fluorine (mechanism #2 above) near the center of the molecule
 (one of the F's participating in the HOMO) to simulate
 abstraction by a Lewis acid, this left a polymer cation.
 The HOMO shifted from the center of the
 molecule to three terminal fluorines at one end of the cation and the oxygen
 adjacent to them (it had a big coefficient).  I take this
 result to mean that this trifluoromethoxy group is ready to leave
 with the next available Lewis acid site.  Does the Net agree?
      Next I abstracted a F- to leave a polymer radical
 (mechanism #4 above).  The HOMO was delocalized over the whole
 molecule (except for the oxygen that had the high HOMO coefficient
 in the cation described above).  The LUMO was delocalized as well,
 but not quite to the extent the HOMO was.  The  terminal oxygen
 did participate in this LUMO.
      This radical case is a little harder for me to make sense of.
 The radical had both HOMO and Lumo coefficients on sigma
 orbitals.  Cation case HOMO/LUMO orbitals were associated with p orbitals only.
 Does the involvement of sigma orbitals in the radical case signal
 incipient fragmentation?  Or is the radical content to simply
 delocalize the odd electron and float in the vacuum for ever?  It
 certainly seems like another electron would break a bond.  I didn't
 model this though.
 10e6 thanks to all who are willing to offer (helpful) comments.
 Pat Hogue