Shell Development Sigma Xi Club Lecture, Berkeley, Feb. 18, 1948

"Bonds between Metal Atoms in Metallic and Non-metallic Substances."

My chairman, ladies and gentlemen. I am not talking about antibodies today, not because I have tired of them, but because I can't repeat my talk here of a year ago. But metals are interesting - and we don't understand them very well yet.

A year ago I mentioned Crellin and his interest in the days in September. Last year he had two teeth extracted - .

To get back to metals - every once in a while a theoretical physicist says something that turns my stomach. This happened a few years ago about metals. Mott and Jones said that in nickel there is a 0.61 bonding electron per atom, and in iron 0.22! This is silly - when we compare tensile strength, etc. of Ni and Fe with K, say, which has 1; or with silk fibers with one covalent bond per string to be broken - there are 12 contacts, and hence there must be ~ 6, say, bonds per iron atom!

Why did Mott and Jones say this? They think of 3d and 4s orbitals as much different. μ = 0.61 for Ni, 0.22 for Fe. But we must hybridize 3d, 4s, and 4p. This would permit 9 resonating bonds for Co. Actually there are ~6 from Cr to Cu. Explanation.

We might ask - since this metal-metal bonds are good bonds, would they not occur in non-metallic compounds? This is an interesting question - there are many interesting questions such as the explanation of the astounding specificity of serological reactions. You remember that here the cause is structural complementariness of antigen and antibody. Is there any analogy among simpler chemical substances? There is a striking one - I don't believe I mentioned it a year and a half ago - I hadn't yet thought of it. It is crystallization.... And there are occasional isolated instances of surprising specificity in inorganic and organic chemistry. For example, tungsten as W ⁺⁺⁺ occurs only as W₂Cl₉ ^--- (also W₂Cl₂₉ ^--). Why not WCl₆ ^---? Why not W₂F₉ ^---? Answer - configuration permits a W-W bond, in addition to the W-Cl bonds.

Are there any other examples? I asked a young man that - UC Ph.D. - and he couldn't think of any. I then said "Have you ever heard of calomel?"

Cl-Hg-Hg-Cl Also Fe₂(CO)₉ diferricdihydroxytetraphenanthroline. Ag₂F, Pb₉4^-, Ta₆Br₁₂ ^--, sulvanite, Mo₃Cl₄ ^--, MoS₂ - molybdenite, like graphite.

What about hydrolysis of Fe(H₂O)₆ ⁺⁺⁺? Can yellow color be due to polymers? Limonite yellow. Lepidocrocite and Goethite red, also FeOCl. Hematite red-black. Also Cu(OH)₂ blue -> CuO black.

This brings me to a matter that has long interested me - effect on color of having an element in two valence states. Fe(OH)₂ white -> green black -> Fe(OH)₃ red brown.

CuCl and SbCl₃ colorless -> brown black -> CuCl₂ and SbCl₅ blue.

Cs₂AuAuCl₆, Biotite, Mica -> Russian blue; magnetite, tourmaline, Fe⁺⁺, Fe⁺⁺⁺, iluvaite, all black.

But why go on? Voltaire said "The secret of being a bore is to tell everything." It is clear that there is much here that we do not understand - perhaps that is its attraction. Thomas Wright in 1601 said "Nothing is so curious and thirsty after knowledge of dark and obscure matters as the nature of man." So let us continue to think about these questions - to look for uniformities among facts that at first seem disconnected - and in this way do our bit in obtaining an understanding of the great and complex and terrifying world about us.