van der Waals surface summary



2000 May 1
 Here is the summary of  replies to my question:
 ------------
 2000 April 25
 Hello,
 I once read somewhere that an isosurface that encloses 97% of the total
 electron density of a molecule corresponds best to the experimental van
 der Waals surface.
 (1) Has anyone a reference to this, or to any percentage of the electron
 density?
 (2) Is it correct to say that the usual way to measure the "van der
 Waals size" of a molecule is by X-ray diffraction (the molecules in a
 crystal being considered to be in contact)?
 Thanks.
 E. Lewars
 ======
 I thank all who responded. In addition, I wish to point out that on page
 183 of Bader's book (_Atoms in Molecules_, Oxford, New York, 1990) it
 says: for simple molecules (gases) a 0.001 au density surface, enclosing
 98-99 per cent of the electron density, gives molecular sizes in good
 agreement with second virial coefficient and viscocity measurements. For
 presumably bigger molecules, crystal data molecular size is fitted
 reasonably well by a 0.002 au isosurface (I don't know what fraction of
 the total electron density the 0.002 surface encloses; maybe about 95
 per cent).
 Thanks again to all who replied.
 EL
 ===========
 #1
   Date:
          Tue, 25 Apr 2000 18:52:00 -0400 (EDT)
  thought the closest isosurface to the VdW surface was the 0.002
 isodensity surface?  I don't know what percentage of the e- density
 this contains, but I recall somebody like Warren in the early/min
 (mid) 80's publishing this in a paper.
 Joe
 From:  Joe M Leonard <jle (- at -) world.std.com>
 ===========
 #2
 There is one convention that says that for condensed phases, the Van der
 Waals surface is approximated by a 0.002 e/au3 isosurface, while the gas
 phase VdW surface is represented by the 0.001 e/au3 isosurface.
 Curt Breneman
 RPI Chemistry
         From:
              "Prof. Curt M. Breneman" <brenec (- at -) rpi.edu>
  Organization:
              Rensselaer Polytechnic Institute
  ===========
 #3
 Spartan's default value is 0.002 au, but some chemists like to use 0.001
 (they
 have always looked essentially the same in the few cases where I have
 checked).
 Re #2 -
 I'll be interested to hear what you receive on this topic. The two
 comments that I
 (think I) can recall off the top of my head (not a very good place to
 start) are one
 >from Pauling and a much later one from Allinger. I seem to remember
 Pauling in one of
 his books (perhaps Nature of the Chemical Bond) using the intermolecular
 distance
 between chlorine atoms in a solid state chloroarene as a measure of the
 nonbonding
 distance for this atom. I don't recall if Pauling referred to this as a
 "van der
 Waals" distance. Allinger, in his book (Molecular Mechanics), talks at
 some length
 about the difficulty of using intermolecular distances obtained from
 solid state
 structures. For example, suppose one has a solid in which octadecane
 molecules (or
 something like it) are aligned side by side. Does the intermolecular
 distance between
 hydrogens tell us the "van der Waals" distance? Allinger says NO
 because
 the
 intermolecular distance arises from a sum of many interactions, bonded
 and nonbonded.
 Anyway, that's from the top of my head complete with the omissions,
 distortions and
 fabrications that this source often provides. Good luck,
 -Alan
 --------------
 Alan Shusterman
 Department of Chemistry
 Reed College
 Portland, OR
 =============
 #4
 Dear Dr. Elewars : Kindly send me the replies to receive to
 the above mentioned query. Thanks in anticipation!...Shridhar Gadre
 www.reed.edu/~alan
 ===================
 #5
 Dear Dr. Lewars
 In Prof. R.F.W.Bader's group, we integrate the properties densities of
 an
 atom in a molecule over the atomic basin. The basin is the region of
 space
 enclosed by the intersection of Inter-Atomic Surfaces (IAS) and and
 envelope of the density choosen at some value. I will explain HOW AND
 WHY
 we choose this value, with the hope of answering your question.
 The IASs are not arbitrary. They must satisfy a quantum condition that
 arises from the application of Swinger's principle of stationary action
 to
 a subsystem (e.g. an atom in a molecule). An IAS is the unique surface
 separating 2 atoms in a molecule so that the flux in the gradient
 vectror
 field of the density vanishes.
 If one integrates atomic properties up to the 0.001 a.u. isodensity
 envelope, these atomic values sum-up to the corresponding molecular
 (total
 system) values with reasonable accuracy.
 For example, summing atomic energies of medium-sized molecules
 such as the 20 naturally occuring amino acids recovers the SCF energy to
 within 1 kcal/mol (typically 0.5 kcal/mol). The sum of the atomic
 populations
 (or the corresponding atomic charges) is typically 0.001 e for a neutral
 molecule (or within 0.001 e of the molecular charge). The sum of the
 atomic
 volumes (molecular vol.) of the 20 amino acids integrated up to 0.001
 a.u. envelope corrrelates highly with the experimental partial molar
 volumes (r^2=0.971).
 These results are not surprising given that the 0.001 a.u. envelope
 encloses over 98% of the electronic charge of H atoms in hydrocarbons,
 and
 over 99% of the electronic charge of atoms Carbon to Neon.
 The 0.001 a.u. isodensity envelope has been shown to yield values of
 molecular diameters of several gases in good agreement with the
 equilibrium diameters of these molecules determined experimentally
 (e.g. by viscosoity measurements fitted to a 6-12 potential).
 The 0.002 a.u. may be a more realistic measure of the molecular size in
 crystal, however.
    CONCLUSION
    ==========
 The 00.1-00.2 a.u. isodensity envelopes are good approximation to the
 experimentally determined van der Walls sizes of molecules.
    REFERENCES
    ==========
 [1]     Bader RFW. Atoms in molecules: A quantum theory. Oxford,
         U.K.: Oxford University Press, (1990). pages 182,183,199,202.
         And references therein.
         (Radii of isolated ground-state atoms and ions is given in page
         432, for both the 0.001 and 0.002 envelopes).
 [2]     Matta CF, Bader RFW. Atoms-in-molecules study of the
         genetically-encoded amino acids. I. Effect of conformation on
         geometric,atomic, and bond properties. Proteins: Structure,
         Function and Genetics, (2000) [In Press].
 >.......................................................................
  Cherif F. Matta                        tel. (905) 525-9140 ext. 22502
  Chemistry Department                   fax  (905) 522-2509
  McMaster University
  Hamilton, Ontario, CANADA L8S 4M1.
  Member of the Board of Governors of the University
 >...................................................................
 =============
 #6
 Hi,
 I would just like to make a small comment on the side:
 The van der Waals size applies to a very simple picture of a molecular
 system. It is quite useful in many cases but (and this is my perhaps
 misdirected warning) when it comes to scattering processes, e.g.
 collisional energy transfer, the shape of the repulsive wall is of
 course important.
 How to go from electron density to the correct repulsive potential is
 discussed in e.g.:
 C. Nyeland, J. P. Toennies, Chem. Phys. 188 (1994) 205.
 There is also work in progress in this field.
 yours
 :-)         The spring has finally come here in Goeteborg!
 Harald
 Harald Svedung (Ph.Lic.)                phone:          +46-31-7722816
 Department of Chemistry                 fax:            +46-31-167194
 Physical Chemistry                      home phone:     +46-31-240897,
 +46-709223206
 Goeteborg University                    home e-mail:
 harald.svedung (- at -) svedung.pp.se
 SE-412 96 Goeteborg, Sweden             www.che.chalmers.se/~svedung/
 ==========
 #7
 Dear Dr. Lewars,
 A colleague has just given me a copy of your message on the CCL
 concerning electron densities. Part of your question is answered in my
 paper J. Phys Chem. A 102, 6043-6051 (1998) "Behavior of Electron
 Density Functions in Molecular Interactions", in which I suggest that
 the 0.002 au electron density contour corresponds to the van der Waals
 interaction distance in a number of weakly bound systems. The paper you
 were possibly remembering is Bader, Henneker and Cade, J. Chem. Phys.
 46, 3341-3363 (1967) which discusses many facets of the electron density
 functions of diatomic molecules. I think they mention some percentage of
 electron density being enclosed by the .001 or .002 contours. If not
 there, check out other works of that period by Bader, or the more recent
 Bader et al., JACS 109, 7968-7979 (1987). Regarding the question about
 x-ray diffraction, if you look at (for instance) Bondi's paper (J. Phys.
 Chem. 68, 441-51, 1964) I believe he used x-ray diffraction values
 supplemented by other observations. However, if you consider crystalline
 Cl2, the contact distance depends on the relative orientations of the
 molecules. Molecular models consisting of two spherical Cl atoms with
 Van der Waals radii do not account for this, but models based on the
 electron density do. So the warning is: even for a single atom in a
 molecule, the van der Waals radius is not a single value. How big a
 problem this is depends om your application.
 I hope this is of some value to you.
 Sincerely yours,
 John Bentley
 Assistant Director
 Notre Dame Radiation Laboratory
 University of Notre Dame
 Notre Dame, Indiana  46556-0579
 e-mail: bentley.1 (- at -) nd.edu  telephone: 219-631-6117  fax: 219-631-8068
 ===============