Old Thesis Club: Secondary Emission of Electrons from Molybdenum, by H.A. Smith, 1928

As noted in a previous post on Monte Carlo simulation in 1960, we recently came into possession of a large box of old Master's theses. The bulk of these are from the 50's and 60's, but there are some going back much farther. As I pass these every day I'm in the office, I thought it might be amusing to take a look at these for the blog, now and again. I don't plan to do a detailed examination of the quality of the science (at least, not necessarily), but to use this to look at how things have changed over the decades.

The first of these, pictured above, is one of the oldest: Secondary Emission of Electrons from Molybdenum by Hardick Smith, who submitted it in partial fulfillment of the requirements for the MS degree in 1928. There are older theses in the box, I think, but this might be the earliest with a definite date attached.

The subject matter of this is not, in itself, all that interesting (to me). "Secondary emission of electrons" means that they're bombarding a molybdenum surface with electrons, and looking at the extra electrons knocked out in the process. This involves a series of experiments with vacuum tubes, which tend to break down due to overheating ("heated to incandescence" is a recurring phrase) or get contaminated with gas adsorbed onto the surface. It's a finicky measurement to make.

What's fascinating about this, though, is the introduction. 1928 is right at the biggest and most important inflection point in the history of 20th century physics. Louis de Broglie's wave theory of the electron had been proposed in 1923, and Werner Heisenberg's "matrix mechanics" formulation of quantum physics and Erwin Schrödinger's wave equation were produced right around this time. Dirac's relativistic theory of the electron was only two years off. The mass of experiments showing weird stuff happening had gotten to a critical point-- Arthur Holly Compton had just won the 1927 Nobel for the scattering effect that bears his name (discovered in 1923), and the Davisson-Germer experiment (in the US) and George Paget Thomson's electron diffraction experiment in Aberdeen had just shown that the electron behaved like a wave.

The introductory material in this thesis is fascinating, then, both for what it contains and what it doesn't. There's a lengthy discussion of the various modes of electron emission from metals: thermionic emission, where heating a metal drives electrons out; the photoelectric effect, where shining light on a metal drives electrons out; and secondary emission due to electron bombardment, where hitting a surface with electrons knocks out additional electrons. The unsettled nature of physics at this time leads to a general air of puzzlement about the whole thing.

It's clear from context that both Einstein's photon theory of the photoelectric effect and the Bohr model of electrons in special allowed orbits are well accepted at this time, but the overall structure of matter is clearly still in flux, making all three types of emission a puzzle. Thermionic emission is described as the best understood of the lot, with a simple theoretical model given, provided one is willing to accept a major unjustified assumption:

In these experiments we have assumed the presence of free electrons in a metal without accounting for their origination. We are certain that they are free to move about the metal under the influence of an electric field, but whether they are completely unrelated to the atom or not is not understood. It is possible that they are electrons in the outermost orbits of the atom, and are exchanged by Collision from one atom to the other.

So as long as you don't worry about where the electrons are coming from, you can understand how heat drives them out.

The photoelectric effect is also generally accepted, but poses some problems. Einstein's Nobel Prize citation from 1921 specifically cites his model of the photoelectric effect, and Robert Millikan's 1923 Nobel was awarded in part for exacting measurements that confirmed Einstein's predictions in 1916. So the idea that electrons are knocked out by absorption of a photon was well understood, but exactly how that worked remained mysterious:

[T]he quntum of energy hν is transmitted to the receiving electron without loss and is independent of the parentage of the electron receiving it. This is rather mystifying as it is hard to understand how any electron independent of the orbit in which it is can always receive the energy quantum hν and still have the same final energy. On the other hand, if we assume that the free electrons of the metal receive the light quantum, we encounter, as has been pointed out by Milliken [sic], another difficulty, namely, it would be impossible without a violation of the law of conservation of momentum for a quantum of energy whose momentum is given by hν/c... to collide with an electron whose mass is many times greater and give it suffcient energy to escape.

The third method, secondary emission, is described as "still more baffling," with three possibilities suggested for a microscopic mechanism to knock electrons out of the metal: 1) collisions between incoming electrons and those mysterious free electrons, which can work in this case because the masses are equal, 2) collisions with electrons bound in orbits around atoms, and most ingeniously, 3) "soft X-ray radiation produced by bombarding the metal with the electrons may cause a photoelectric emission of the secondary electrons." That is, the incoming electrons are brought to a halt inside the metal, spitting out x-rays via bremsstrahlung. some of these x-rays hit other electrons in the metal, and knock them out via the photoelectric effect.

The specific experiment here is designed to attack the problem of just what the hell is going on with the electrons by looking at the parameters affecting the secondary emission. If it's collisions with free electrons, they would expect a temperature dependence of the secondary emission. If it's collisions with bound electrons, they would expect steps in the number of secondary electrons emitted as they increase the energy of the incoming electrons-- as they cross the binding energy of a new state, they should suddenly be able to knock out more electrons. And if it's photoelectric, the number should be more or less constant, as the bremsstrahlung effect will always be there, independent of energy or temperature. The experiment is hampered by technical problems, but tentatively comes down in favor of the photoelectric explanation.

All of these problems are eventually fixed by the modern band theory of electrons in solids. The "free" electrons are in fact bound into delocalized states that form continuous energy bands inside the metal, and this explains how they are free to move around in response to fields, and why the free electron theory works so well to explain thermionic emission. the momentum issue is resolved because the electrons in these bands are not, in fact, sitting at rest inside the material, so the small photon momentum is not the only thing you have to work with. And the energy of an ejected electron is always the same because they're coming out of a broad band filled with enormous numbers of electrons of very similar energy, not discrete atomic states.

But all of that stuff requires the full wave theory of matter, which was too new to make it into this thesis. Once Heisenberg and Schrödinger and Dirac had done their thing, theorists could work out the consequences of placing electrons into a periodic potential, and build up the band structure model, but that was still mostly in the future in 1928. Thus, bafflement on the part of our long-ago thesis student.

So, this was an interesting glimpse of a really confused time in the history of physics. Stuff that these days seems is casually explained to sophomores taking "modern physics" was still largely unsettled in disturbing ways. It's hard not to wonder whether some blogger-analogue ninety years from now will be looking back on modern research into dark matter and dark energy in the same way.

------

Some non-scientific notes about this old thesis:

-- Weirdly, there's no mention at all of the faculty member who supervised this work. Nor are there any acknowledgements of others.

-- The thesis is typed on really thin paper, basically translucent. One of the senior faculty in the department recognized this as associated with an old way of making copies, but it's well before my time.

-- I'm not entirely sure this is the final draft, as there are a number of typos marked in red, and other errors corrected by hand with a fountain pen. Also, non-standard characters (Greek letters, etc.) are hand-drawn.

-- The data graphs are on different paper, drawn by hand, as seen here:

This might still be less annoying than formatting Excel graphs properly.

There are no figures of the apparatus.

-- The author, Hardick A. Smith, was apparently a 1925 graduate of Cornell, because Googling his name turns up some ancient alumni magazines mentioning him, and that he worked as a physicist for an electrical products company in New York. That's as much snooping as I'm willing to do about him.