It makes for a bizarre scene: two of the world’s great physicists taking
time off to bend their minds towards the kitchen, rather than the Universe;
but they took the project seriously.

Einstein, of course, had spent nine years in a patent office. He had
been a clerk in Berne at the turn of the century, whose job it was to pass
judgment on other people’speculiar ideas. He would then take the tram home
and have his own strange ideas.

Szilard, a peripatetic Hungarian, had reached Berlin, where he had met
Einstein at a seminar on quantum mechanics. Szilard was not well-known then
in the physics community, but he had plenty of original ideas, and he knew
everybody.

When they got together to think about refrigerators is not known, and
tends to be dismissed by most biographers, but one or other of them would
have seen the newspaper article, and perhaps mentioned it to the other over
coffee. In those days, a great deal of physics was done over coffee.

Which of them had the first idea is not known. Einstein always claimed
that he very rarely had any new ideas. He was, in his phrase, ‘more a mole
than a butterfly’. So it was probably Szilard who came up with the initiating
spark. He was renowned as an ideasman. The only way he wasable to control
the flow of hisimagination was to put all histhoughts down on postcards
and sendthem off to himself at whichever hotel hewas staying at (he loved
hotels). Then theywould arrive fresh and dated in order andhe could sort
out the good from the bad. Szilard was unquestionably a butterfly.

Their final design was unique. It consisted of conventional fridge piping,
through which, instead of a volatile liquid such as the ammonia compound
that had killed the Berlin family, a liquid metal, or suspension of fine
metal particles, was to be pumped by what their patent called ‘ponderomotive
forces’. Basically, the metal was made to flow smoothly by magnetism. It
was one of the earliest magnetic pumps.

They had spent some months working on the project as a relaxation, as
a kind of game. Einstein’s wife, Elsa, thought that they might make a fortune
with it. Szilard also seems to have been convinced that they were onto a
winner. The German electrical company AEG, attracted by the name of Einstein,
asked for permission to build one, and Szilard was hired as a consultant.

Unfortunately, the prototype proved so noisy, even compared with the
far from silent iceboxes of the day, that the idea never left the factory.

Undeterred, they tried again and, by 1931, a total of 29 patents had
been filed, mostly in connection with refrigeration and magnetic pumping.
But by this time, their association had fallen through. It must have been
disheartening when none of their patents proved of the slightest interest
to anyone.

Throughout the 1930s and into the Second World War, much political energy
was used up, throwing Einstein across the Atlantic to Princeton, along with
many scientists, and, after much delay, Szilard too. Szilard, when he was
not thinking about physics, was thinking about politics, and he had spent
time in England helping to find jobs for refugees from Hungary, living off
his dwindling savings, until he crossed the Atlantic in 1938.

The centre of gravity of physics changed too, and pointed now in the
direction of nuclear energy. In this period, Szilard’s interests had switched
to the problem of the atom, and he had filed with the British Admiralty
a patent that he hoped would remain secret: a description of the release
of energy in a ‘chain reaction’. In the ugly European environment of the
1930s, and with his political awareness, Szilard had realised the potential
of his idea. Unfortunately, he did not have the clout to make anyone listen.
So he contacted Einstein.

The story of how Szilard and Eugene Wigner tracked down Einstein to
his summer house at Cutchogue, Long Island, is well-known. Einstein agreed
to sign the letter that Szilard had drafted, describing the potential for
the new discovery of fission to be misused to construct a bomb, and the
letter was sent off to President Roosevelt. Given the consequences, whether
the letter failed or not seems to be a moot point, but Szilard, when the
war ended, decided that the only honorable thing to do was to vacate physics,
and he turned his back on the subject to study biology; physics, for him,
had failed.

Meanwhile, the refrigerator that he and Einstein had designed surfaced
once again, this time in a nuclear reactor.

The peaceful uses of nuclear energy were being exploited at the University
of Chicago at almost exactly the same time as the weapons laboratory at
Los Alamos was founded. The world’s first ‘pile’ was put together by Enrico
Fermi and his team on top of the squash court at Chicago (some sources say
underneath the tennis court) and it was working in its own sweet and primitive
way by 2 December 1942. Szilard worked there a while, until he sentenced
himself to biology. He did not stay long, but he did make one or two suggestions
to do with cooling the reactor.

Inside the core of a reactor, great amounts of heat are created, and
this is of no use whatsoever unless it can be pumped out and made to heat
water in a conventional steam-turbine generator. Szilard worked on this
problem – and suddenly remembered the noisy fridge at AEG. It was the ideal
heat pump. With modifications, and more advanced motors, it worked quietly
too.

Today, in many reactors, a magnetic pump, having the Einstein-Szilard
design as an ancestor, propels liquid sodium through the heart of the reactor,
in part because of the playfulness of two great men, and in part because
of the tragic deaths of a family as a result of a leaky new fridge in their
apartment in Berlin. Neither Einstein nor Szilard made a penny from it,
however: they decided it was not worth the trouble to take out a patent.

Stephen Moss is a science teacher in Oxford.

IN THIS ambitious book, the author examines essentially all the materials
used in the microelectronics industry, with particular emphasis on materials
for integrated circuits, light-emitting diodes, and photovoltaic devices
for solar cells. His approach includes not only semiconductors but also
metals and insulators, as well as packaging materials. The emphasis throughout
is on the need to take account of the properties and interactions of all
the materials to prepare successful devices. To my knowledge, there is no
comparable book available.

Grosvenor’s book is intended mainly for postgraduates and senior undergraduates
in materials science but will also be of great value to many other scientists
and engineers with interests in microelectronic devices. The author assumes
that readers have some basic understanding of the electrical properties
of semiconductor devices, crystal-structure terminology and phase diagrams.

Two substantial introductory chapters on semiconducting materials, and
contacts, devices and integrated circuits make an excellent beginning. These
chapters are both informative and instructive, giving a valuable perspective
to most of the major areas dealt with at greater length in subsequent chapters.
These topics include the growth of single crystals in bulk and the thin-film
forms of semiconducting materials, polycrystalline films, matallisation
schemes, insulating films, and packaging. You can obtain some idea of the
range of the book by noting that the author writes about multiple quantum-well
structures in ultra-thin layers produced by molecular beam epitaxy, as well
as the more mundane mechanical properties of solders used to attach silicon
chips to supporting substrates, gives you some idea about the range of the
book.

The last chapter of the book is especially valuable. This deals with
failure analysis and the investigation of the structure and composition
of microelectronic materials. Grosvenor uses the results from such studies
throughout the book, but here he explains how the studies were made and
gives useful comparative accounts of various forms of microscopy and spectroscopy.

Grosvenor treats each topic in some depth and with great clarity, achieved
through his mastery of the material and straightforward writing style. I
was especially pleased to see that he regularly used the first person singular
in his explanations so they are free from the awkward structures often created
by misguided efforts to avoid first-person accounts in technical writing.
The clarity of the writing is enhanced throughout by many excellent illustrations.
References appear only infrequently in the text, but each chapter contains
more than 30 references to significant books and research papers (many as
recent as 1987), so that interested readers may obtain further information
easily on any topic.

The book is well produced and the proof reading was clearly excellent;
I spotted only a handful of misprints. However, one unfortunate error remains
in the first chapter, where polyethylene is mentioned (along with polyacetylene)
as potentially a conducting polymer, surely a slip for polyphenylene.

My quibbles about the book are minor. It is a remarkable achievement
for a single author in the range of material and the depth of the treatment.
I believe that it will be the standard work for many years to come.

Stephen Moss is in the department of molecular sciences, Aston University,
Birmingham.