Sunday, January 02, 2011

The Year of Chemistry (but some physics and biology too)

I seem to have ended 2010 with a little cluster of articles here and there. In Physics World I have a feature on single-molecule sequencing of DNA using nanopores – an exciting area that I’m now convinced is going to pay off some time soon, and which will demonstrate that advances in understanding of biology still frequently hinge on the technical capability that physics and chemistry supply. Oddly the December issue of Physics World seems still not to be in circulation or live online, but there’s a preview of the piece here. In Nature I have a couple of pieces to mark the Year of Chemistry in 2011 – an In Retrospect perspective on Linus Pauling’s classic text The Nature of the Chemical Bond and, as the main course to that hors d’oeuvre, an article on changing views of the chemical bond. The first of these is the first item below (the long version, with material that was rightly cut for the published version); the second is too long for that, but will appear in this week’s issue of Nature. I have a follow-up on the Peter Debye story below as my Crucible column in the January Chemistry World; that’s the second item below. And finally, I have a piece in New Humanist that trails my next book Unnatural, coming out in February, which picks up on the forthcoming production of Frankenstein at the National Theatre, directed by Danny Boyle. I’m greatly looking forward to that performance, and hope to be reviewing it for Nature. The NH piece is graced by one of Martin Rowson’s fabulous illustrations – worth the cover price for this alone.

And Happy New Year to everyone.

***********************

Linus Pauling’s The Nature of the Chemical Bond has, like Newton’s Principia or Darwin’s Origin of Species, the kind of legendary status that is commonly deemed to obviate any obligation to read it. Every chemist learns of its transformative role in uniting the prevailing view of molecules as assemblies of atoms with the new quantum-mechanical picture of atomic wavefunctions. But the book is long, by chemists’ standards mathematical, and anyway we now know that there are more versatile and useful approaches to the quantum bond than Pauling’s.

Yet Pauling’s book remains a good primer on the basic facts of chemical bonding – impressive for a book almost 70 years old. That’s not to say that the book should be more widely read – there are naturally better and more relevant treatments of the subject now, and The Nature of the Chemical Bond does not benefit from the elegant prose of Darwin’s works – but it is still bracing to do so. The best preparation is to look first at what more or less contemporary textbooks have to say about bonding. To take two random examples: Inorganic Chemistry, (Macmillan, 1922), by the eminent T. Martin Lowry, professor of physical chemistry at Cambridge, barely gets beyond John Dalton’s symbolic ‘ball’ molecules and Berzelius’s Law of Multiple Proportions (elements combine in simple ratios); Outlines of Physical Chemistry (16th edn, Methuen, 1930) by George Senter of Birkbeck College, a student of Wilhelm Ostwald and Nernst, doesn’t even mention the chemical bond but speaks in terms of affinities. They are products of the nineteenth century.

It’s true that this is not entirely representative, for the problem of how to describe the chemical bond was already by then acknowledging atomic physics. The English chemist Edward Frankland introduced the term in 1866, but regarded it not as a physical connection, as implied by the practice then common of drawing lines between elemental symbols, but as a kind of force akin to that which binds the solar system. Berzelius suspected that this force was electrostatic: the attraction of oppositely charged ions. That view seemed favoured by J. J. Thomson’s discovery of the electron in 1897, since ions could result from an exchange of electrons between nuclei.

But Gilbert Lewis, another Nernst protégé at the University of California at Berkeley, argued that bonding results instead from sharing, not exchange, of electrons. More precisely, this gives rise to what Irving Langmuir later called a covalent bond, as opposed to the ionic bond that comes from electron exchange. In 1916 Lewis outlined the view that atoms are stabilized by having a full ‘octet’ of electrons, visualized as the corners of a cube, and that this might come about by sharing vertices or edges of the cubes. Langmuir popularized (in Lewis’s view, appropriated) this model, which seemed vindicated when Niels Bohr explained how the octets arise from quantum theory, as discrete electron shells.

Yet this remained a rudimentary grafting of quantum theory onto the notions that chemists used to rationalize molecular formulae. Pauling, a supremely gifted young man from a poor family in Oregon who won a scholarship to the prestigious California Institute of Technology in 1922, was convinced that chemical bonding needed instead to be understood from quantum first principles. He wasn’t (as sometimes implied) alone in that – in particular, Richard Tolman at Caltech held the same view. Pauling had a golden opportunity to develop the notion, however, when in 1926 a Guggenheim scholarship allowed him to come to Europe to visit the architects of quantum theory: Bohr at Copenhagem, Arnold Sommerfeld at Munich and Erwin Schrödinger at Zurich. He also met Fritz London and his student Walter Heitler, who in 1927 published their quantum-mechanical description of the hydrogen molecule. Here they found an approximate way to write the wavefunction of the molecule which, when inserted into the Schrödinger equation, allowed them to calculate the binding energy, in reasonable agreement with experiment.

Pauling expanded this treatment to the molecular hydrogen ion H2+, and generalized it into a description called the valence-bond model. He considered that if the wavefunction that offers the lowest energy turns out to be one that is a combination of the wavefunctions of two or more structures, the molecule can be considered to ‘resonate’ between the structures. The molecule is ten stabilized by ‘resonance energy’. “It is found that there are many substances whose properties cannot be accounted for by means of a single electronic structure of the valence-bond type, but which can be fitted into the scheme of classical valence theory by the consideration of resonance among two or more such structures.” For example, the H2+ ion can be considered a resonance between HA+ .HB and HA. HB+ The electron resonates between the two nuclei.

Pauling also showed in a paper of 1928 how the bonding in molecules such as those of four-valent carbon can be explained in terms of the concept of ‘hybridization’, in which atomic electron orbitals (here the so-called 2s and three 2p orbitals) are ‘mixed’ into hybrid orbitals with a new geometric distribution in space: for carbon, they give rise to four sp3 orbitals which create a tetrahedral covalent bonding arrangement. Thiese ideas were published in a series of papers in 1931 in the Journal of the American Chemical Society that formed the core of The Nature of the Chemical Bond. The book remained in print, with (three) revised editions, until 1960. The scope of the book is breathtaking: it brings multiple bond, ionic, metallic and hydrogen bonds all within the framework, and explains how the ideas fit with observations of bond lengths and ionic sizes in X-ray crystallography, the technique that Pauling studied from the outset at Caltech and which eventually led to his seminal work in the 1950s on the structure of proteins and nucleic acids.

Pauling acknowledges in his book that it is a bit arbitrary to divide up the bonding into particular, resonating configurations of nuclei and electrons; but he says we do that all the time. “The description of the propane molecule as involving carbon-carbon single bonds and carbon-hydrogen single bonds is arbitrary; the concepts themselves are idealizations.” The wavefunction is all that really matters.

It is one thing to say it, however, and quite another to accept this arbitrariness in the face of an alternative. In the late 1920s, Robert Mulliken at the University of Chicago and Friedrich Hund in Göttingen devised a different quantum description of chemical bonding which approximated the electron wavefunctions in another way, giving rise to ‘molecular orbitals’ in which electrons were considered to be distributed over several nuclei. This model gave a rather simpler picture for explaining molecular electronic spectra: the quantum energy levels of electrons. What is more, it could offer a single description of some molecules for which the valence-bond approach needed to invoke resonance between a great many discrete structures. This was especially true for aromatic molecules such as benzene: the VB model needed something like 48 separate structures for naphthalene, and, in the case of ferrocene described in the 3rd (1960) edition of The Nature of the Chemical Bond, no fewer than 560. Evidently, while neither the MO nor VB models could lay claim to being more fundamental or ‘correct’, the former had significant advantages from a practical point of view. This was suspected even when Pauling’s book first appeared – some reviewers criticised him for not mentioning the rival theory, while one suspected that the VB method might triumph purely because of Pauling’s superior presentational skills. Pauling himself never accepted that MO theory was generally more useful, although it was the consensus among chemists by the 1970s.

The significance of The Nature of The Chemical Bond was not so much that it pioneered the quantum-mechanical view of bonding – London and Heitler had done that – but that it made this a chemical theory, a description that chemists could appreciate rather than an abstract physical account of wavefunctions. It recognized that, for a mathematical model of physical phenomena to be useful, it needs to accommodate itself to the intuitions and heuristics that scientists need in order to talk coherently about the problem. Emerging from the forefront of physics, this was nevertheless fundamentally a book for chemists.

**********************

In Kurt Vonnegut’s 1961 novel Mother Night, an American writer named Howard Campbell is brought to trial for his crimes as a Nazi propagandist during the Second World War. The apolitical Campbell decided to remain in Germany after Hitler came to power in 1933, where he is persuaded to make English radio broadcasts of Nazi propaganda. But he has also been enlisted by an operative of the US War Department to lace his broadcasts with intelligence messages coded in coughs and pauses. This role is never made public, and Campbell is constantly threatened with exposure of his ‘Nazism’ while trying to lead an anonymous life post-war in New York.

It would be unwise to stretch too far any parallels with the life of Peter Debye, the Dutch physical chemist who won the 1936 chemistry Nobel for his work on molecular structure and dipole moments. But Mother Night came to my mind after hearing the latest suggestion that Debye, who has been reviled in the past for alleged collaboration with the pre-war Nazi regime, might have been passing on information about German war technology to a spy for the British secret service in Berlin.

The evidence for that, outlined in a paper by retired chemist Jurrie Reiding after consulting Debye’s archival documents in America, is extremely circumstantial [1]. Debye was a lifelong friend of Paul Rosbaud, an Austrian chemist who hated the Nazis and spied for the Allies during the war under the codename ‘Griffin’. Reiding says that such a friendship would be inconceivable if Debye was a Nazi sympathizer. But there are no more than vague hints about whether Debye was actually one of Rosbaud’s informants.

Debye’s links with Nazism were asserted in a 2006 book Einstein in Nederland by the Dutch journalist Sybe Rispens, and were outlined in an article ‘Nobel Laureate with dirty hands’ published in a Dutch periodical in connection with the book. Here Rispens explained (as already known to historians) that Debye, as president of the Germany Physical Society (DPG), had signed a letter in 1938 expelling Jews from the society. Panicked by the media exposé, the University of Utrecht removed Debye’s name from its institute for nanomaterials science, while the University of Maastricht withdrew from an annual research prize named after Debye.

A follow-up report on the matter commissioned by the Netherlands Institute for War Documentation (NIOD) changed the accusation of collaboration to one of ‘opportunism’, and the decisions of both universities have now been reversed. But Debye’s name remained tainted in the Netherlands, despite protestation from many scientists both in Europe and in the US, where Debye worked at Cornell University after leaving Germany in 1940.

There’s good reason to think that Debye was no friend of the Nazis. He collected his Nobel prize against their expressed wishes, and they thought him far too friendly to the Jews in his role as DPG president. Indeed, he even – with Rosbaud’s assistance – helped the Jewish nuclear physicist Lise Meitner flee Germany.

And yet why did he stay in Germany so long, when others left? Roald Hoffmann at Cornell has argued that this inevitably taints Debye’s reputation. ‘In the period 1933-39’, he says, ‘Debye took on positions of administration and leadership in German science, aware that such positions would involve collaboration with the Nazi regime. The oppressive, undemocratic, and obsessively anti-Semitic nature of that regime was clear. Debye chose to stay and, through his assumption of prominent state positions within a scientific system that was part of the state, supported the substance and the image of the Nazi regime.’

Clearly Debye’s story is not one of heroic self-sacrifice; the issue is rather where mild resistance blends into passive collusion. Cornelis Gorter, a physicist at Leiden University who knew Debye well, said that (like Howard Campbell) ‘he was not at all a Nazi sympathizer but was apolitical.’ Yet it seems that, also like Campbell, his deeds can tell quite different narratives viewed from different perspectives. The accusation of opportunism in the NIOD report came largely because, having occupied positions of power in Nazi Germany, Debye went on to serve the US war effort enthusiastically, for example through his work on synthetic rubber. That could suggest ingratiating collaboration with any ruling power, but it also fits the picture of Debye striving to limit Nazi abuses before finally fleeing to oppose them more openly.

This situation is reminiscent also of the controversy about Werner Heisenberg, memorably explored in Michael Frayn’s play Copenhagen. Did Heisenberg actively drag his heels to thwart the Nazi efforts to make an atomic bomb, or did he simply get the physics wrong? Did he even know his motives himself? And if not, how can we hope to?

A clue to Debye’s position may lie in a letter he wrote to the physicist Arnold Sommerfeld just before he left Germany for good. His aim, he said, was ‘not to despair and always be ready to grab the Good which whisks by, without granting the Bad any more room than is absolutely necessary. That is a principle of which I have already made much use.’

But maybe the real moral is the one that Vonnegut adduced for Mother Night: ‘We are what we pretend to be, so we must be careful about what we pretend to be.’

1. Reiding, J. Ambix 57, 275-300 (2010).

No comments: