7 May 1954

Professor Raymond K. Sheline

Department of Chemistry

Florida State University

Tallahassee, Florida

Dear Professor Sheline:

I hope that my telegram has not caused you great inconvenience. I decided to send it to you when I had reached the conclusion that the structure described in the telegram with Vd symmetry, would give rise to one stretching frequency of the bridging carbonyl groups, and three stretching frequencies, active in the infrared, of the non-bridging carbonyl groups. I now find that I had reached this conclusion because I read a number out of the wrong column in a table in Herzberg's book. This caused me to decide that there were three different infrared active frequencies for the non-bridging carbonyls, whereas (unless I have made another mistake now) there are instead one non-bridging infrared active frequency and another doubly degenerate infrared active frequency. Accordingly the agreement with your infrared spectrum that I had thought to exist with the new structure does not in fact exist, and I suppose that your infrared observations eliminate the new structure.

You might think it worth while, in any case, to mention this structure in your paper, provided that you have not yet read proof on it.

I must say that I like the structure with four bridging carbonyl groups better than the other structures, in that the coordination is more closely similar to that for di-iron enneacarbonyl than for the other structures. It is true that the six bonds from cobalt to carbon are directed toward the corners of a trigonal prism, rather than an octahedron. Nevertheless, the trigonal prism may be just as satisfactory arrangement for the bonds as the octahedron. You remember that in molybdenite the trigonal prism has axial ratio very close to unity. This would make the rectangular face square, and no deformation would be needed in order to build up the new structure, with point-group symmetry Vd. I do not have any argument to indicate that this structure would be more stable or less stable than your structure III.

With either the new structure or your structure III the cobalt-cobalt distance is calculated to be 2.6 Å, assuming cobalt-carbon = 2.0 Å and bond angles as in molybdenite. This is not an unreasonable value.

[page 2]

the iron-iron distance in di-iron enneacarbonyl is stout 2.5 Å. The same distance can, of course, be achieved with your favored structure if the bond angles are suitably chosen.

Sincerely yours,

Linus Pauling:W