CCL: Dipole Moments and Molecular Fragments

 Sent to CCL by: "TJ O'Donnell" [tjo.^^.acm.org]
 I agree, in particular:
 > I think chemists have attached too much
 > significance to the undeniable separation of charge that exists between
 > bonded atoms of different electronegativities, mainly because there was
 > no way to demonstrate that these charge separations were not necessarily
 > quantitatively or qualitatively different from any others that might be
 > defined for the system.
 There are many molecular properties that are successfully computed using
 fragments, whether they be atom fragments, bond fragments or group fragments -
 clogp, heats of formation, dipole moments, etc.
 The choice of fragments is arbitrary and similar results can be obtained using
 other
 fragments.  For example, clogp is famously computed using group fragments,
 such as those defined by Al Leo, et. al.  But it can be computed just as well
 (IMHO) using atom fragments.  One can argue statistics about the fit to clogp
 for one method compared to the other, but the results of such arguments only
 lead to a decision about which method is more predictive/useful, not which
 method is
 more correct nor which more accurately represents some underlying physics.
 Any molecular property is truly a property unique to that particular molecule
 and (never?) uniquely attributable to any sums of fragments.  Fragment analyses
 are always arbitrary, albeit superbly practical and useful.
 One exception might be molecular weight, which is uniquely attributable to a sum
 of atomic weights.  But this might be considered more of a definition and it
 surely
 DOES reveal something about the underlying physics.  It arises from the
 fact/observation
 that the nuclear mass of one atom is unaffected by that of neighboring atoms.
 Electrons don't behave like that, as we all know.
 Otherwise chemistry would be simply physics.
 TJ O'Donnell
 gNova Scientific Software
 http://www.gnova.com/
 Phil Hultin hultin.:.cc.umanitoba.ca wrote:
The question about dipole moments of charged species is related to another issue, which is somewhat of a hobby-horse of mine.
In organic chemistry and biochemistry particularly, structural and mechanistic rationales are often based on the idea of “fragment dipoles” and their interactions with one another. In small molecules these are usually dipoles said to be associated with polar covalent bonds, while in proteins the so-called helix dipole is another example.
For many years I accepted the idea that “intramolecular dipole-dipole repulsions” were good explanations for all kinds of phenomena, but I am in serious doubt about that idea now. When you start to dissect the overall dipole moment of a molecule, you enter into the same kind of origin-dependence that you see in ions. What makes the arbitrary choice of a bond dipole or a helix dipole more significant than the infinitude of other point-to-point dipoles that could be defined within the molecular charge envelope? I think chemists have attached too much significance to the undeniable separation of charge that exists between bonded atoms of different electronegativities, mainly because there was no way to demonstrate that these charge separations were not necessarily quantitatively or qualitatively different from any others that might be defined for the system.
We have looked at charge distributions to seek evidence for fragment dipoles and their interactions, and we haven’t seen anything convincing. Do others have any opinions on this? 
 
 Dr. Philip G. Hultin
 Associate Professor of Chemistry,
 University of Manitoba
 Winnipeg, MB
 R3T 2N2
 hultin]^[cc.umanitoba.ca <mailto:hultin]^[cc.umanitoba.ca>
 http://umanitoba.ca/chemistry/people/hultin