July 17, 1931
Professor Arthur B. Lamb,
Laboratory of Chemistry,
Harvard University,
Cambridge, Mass.
Dear Professor Lamb:
I have found the paper on m-dinitrobenzene and related topics by Hendricks and Hilbert interesting. So far as I can tell, it contains no mistakes justifying refusing publication. Nevertheless, it is my opinion that it should not be published before it has been thoroughly revised. The authors may be of a different opinion, but I advise that my remarks be communicated to them with the request that they consider revision. I do not desire to criticize from behind the cloak of anonymity, and so I do not request that my identity be kept secret.
The Crystal structure of m-dinitrobenzene.
The structure reported for the crystal, with the C and N atoms in a symmetry plane and O above and below the plane, depends entirely on the assumption of holohodry. The authors mention on p. 13 that the face development shows no evidence of homihodry; however, only two forms, s and q, which would bear on the question are reported in Groth, so the evidence is not strong. They also report a negative result from a pyroclectric experiment. This evidence suggests holohodry, but by no means proves it.
I myself believe it highly probable that the point-group symmetry of the crystal is lower than D2h. This would be by no means the first crystal found to have lower symmetry than that assigned crystallographically. My reason for not accepting the assigned structure is that the oxygen atoms are so close together. 1.91 Å might be all right for the distance within the nitro group, especially since it is possible that a bond is formed, between those two atoms, but it is surely much too small for two oxygen atoms in different molecules. The authors cannot expect any new or striking conclusions deduced by them from the assumption of holohodry to be accepted.
Following are some minor criticisms:
1. Space-group symbols should be given relative to the axes used, and preferably the full symbol (not the new abbreviated ones till some time has passed). Thus 16 - Pbnm rather than Pnma.
2. The choice of D16
2h - Pbnm rather than D5
2h - Pbnm, say, depends only on the absence of the two reflections given in the table (201) and (102). This is hardly sufficient evidence when dealing with absences.
3. I believe the error Herbel is supposed to have made should be discussed, unless it has been pointed out elsewhere.
The discussion of molecular structure.
1. Bottom of p. 6. Rupture of l-quantization will not permit five L electrons in nitrogen to have parallel spins. But removing one to an outer shell (whether l-quantization is ruptured or not) will.
2. P. 6. There is no state 2s2sp25S of carbon; 2s sp5
5s is meant.
3. p. 7. Pascal attributes a susceptibility of +18 x 10-7 to oxygen in the compounds mentioned. But spin susceptibility as of the O2 molecule, is around 2 x 10-3 at room temperature. In other words, Pascal’s paramagnetism has nothing in common with that of molecular O2 and can not be due to a permanently unpaired electron. Accordingly the formulas with free bonds seem unjustified.
4. I believe the formulas on pp. 8, 9, etc. should be written electronically. As given, they do not answer questions such as the following: does C or N in IV of p. 9 have the unshared pair of electrons?
5. What is meant on p. 11 by saying that nitrogen is quinauivalent in organic nitro compounds? This statement would be made definite by giving an electronic structure.
Very truly yours,
Linus Pauling
LP:M
July
17, 1931
