This Week in Nuclear History: Vannevar Bush worries about French Patents.
On March 7, 1942, Vannevar Bush — the director of the Office of Scientific Research and Development — wrote an unusual letter to Conway Coe, the US Commissioner of Patents. Bush, who was essentially in charge of the nascent US atomic research program at that point, was sending up a plea for assistance in a matter of some delicacy:
When you find a moment, I would like to talk to you about patents in a special field of some difficulty. There is a particular point in connection with the applications by Joliot and Halbane, Serial Nos. 328160 and 328372. I thought you might care to note what the status of these is before we get together, and I will be available to discuss the subject any time that we can arrange. There is a matter of general policy of some difficulty involved on which I certainly need your guidance.
The "special field of some difficulty" was, of course, nuclear fission. "Joliot and Halbane" were Frédéric Joliot and Hans von Halban (whose name Bush, or his stenographer, misspelled), nuclear physicists working out of the Collège de France in Paris, part of the same team that had broken Leo Szilard's self-censorship ring and published on the ability of uranium to sustain a chain reaction in 1939.
The patent applications in question are for none other than the nuclear reactor. Neither were ever granted in the United States — a point we'll come back to. But they were later granted abroad, so we can see what they said with some ease.
The first application was issued as UK patent 614,156 in 1948, with a priority/filing date of May 1, 1939. For its time, it had a provocative title: "Apparatus for the production of energy by nuclear fission." It describes a basic fission reactor: uranium fission chain reactions being used to generate useful heat. It discusses the use of a moderator (particularly heavy water) to slow the neutrons down, and is acutely concerned with the problem of an explosive or run-away chain reaction. The British version even uses the term "critical mass" — though it doesn't quite know how to calculate it — but that might have been added at a later date; I don't know. It also describes the use of a neutron reflector, and gestures towards how the nuclear energy could be converted into a useful form (e.g. put in contact with water and turned into steam). Furthermore, it also claims that you could use thorium instead of uranium throughout the entire system. Not quite a perfect description, but pretty good for 1939.
The second application was issued as UK patent 614,386 in 1948 as well, with a filing date of May 2, 1939. Its title was the same as the other one. It builds on the first patent, proposing several ways that the rate of the chain reaction could be controlled. One of these is rather straightforward by modern standards: the use of neutron-absorbing control rods. But the others are pretty unusual: varying the use of the neutron reflector in a dynamic way; vibrating the entire device; moving the fuel in and out of configuration; using fuel whose density would change as a factor of temperature.
So you can see why Vannevar Bush might have been a little unhappy about these, now that they had been filed as application in the UK, France, and — now — the United States.
Reasons to be unhappy, enumerated:
- Patents are public documents; atomic energy was a secret. Patent applications are not as public as granted patents, but a granted patent is as public as it gets: in exchange for publicity, the patent applicant gets a temporary monopoly. So if the French physicists' patent applications were granted, then the idea of a working nuclear reactor — at the very least — was out in public. This would probably have lead to some news attention, which means more of a focus on whether the United States was or was not interested in uranium fission. All of this was rather antithetical to the "absolute secrecy" that Roosevelt had mandated for the nuclear effort. It also didn't help that time and time again the patents emphasized that uranium fission could lead to catastrophic explosions — a nice spur for anyone thinking about bombs. (A separate patent application prepared at the same time, which I'm not sure was filed in the United States, was explicitly about making atomic bombs. It would be granted in France in 1951 as 971,324.)
- The French were not Americans. A tautological statement, obviously, but worth pointing out explicitly. Bush didn't want atomic energy in control of foreign hands. The French were wildcards as far as Bush was concerned. Where did Joliot's loyalties lie? Unclear, since Joliot was presently in Nazi-occupied Paris. Halban's motivations in pushing the patent applications (he had made it out of France to the UK) were not benign: he wanted to secure France a position in the post-war nuclear world, as equals to the United States and the UK. (The UK, incidentally, would take Halban up on this offer, but didn't tell the United States. This led to a major diplomatic flap later in the war as it was a violation of the Quebec Agreement. It didn't help that by that time, it was known that Joliot had joined the Communist Party.) Bush wasn't even a fan of the British being considered a post-war nuclear equal to the US, and would later that year push Roosevelt to cut them out of the project. (FDR, as is well known, declined.)
- To dispute the patents would be to talk about atomic energy publicly. If the United States wanted to avoid giving the French control over basically all nuclear reactor usage, it would have to dispute the patent applications, and perhaps even deal with the courts. This, again, is the last thing one wants to do with regards to a secret project.
So what's an atomic administrator to do? This was what Bush was going to the Commissioner of Patents to talk about. The situation seemed insuperable.
Coe had an answer thought. During the first World War, Congress had passed a law which allowed the Commissioner of Patents to declare temporarily secret any patent application that he felt had national security implications. The concern at the time was the "wonder weapon" of their day, the submarine, and the hope was that such a law would prevent future "wonder weapons" from diffusing to enemy powers. The law had been resurrected at the beginning of World War II, given some more legal "teeth," and could now be applied, Coe explained, to these pesky French patent applications.
Once a patent application was declared secret, it went to "sleep." Nobody acted upon it, nobody worried about. Once it was determined that it was not longer secret, then, and only then, could the matters of priority or compensation (in case the government used it in the interim) be discussed. Out of sight, out of mind.
So Bush had them declared secret. From there, he started thinking about hiring a full-time patent censor for the bomb project, while at the same time feeling secure in his desire to obtain all nuclear technology inventions in the hands of the US government.1
But the "French problem," as it became known within Manhattan Project circles, would not go away. In 1944, John Lansdale, General Groves' head of security, opined that, "This French situation is going to plague us for some time to come." He was right.
As late as 1959, the matter of whether the once-secret French reactor patents were valid in the United States was still be hashed out in the courts. By then the patent applications were no longer secret. The valid patent for a nuclear reactor, the US argued, was not this vague French specification, but US patent 2,708,656 for a "Neutronic Reactor" by Enrico Fermi and Leo Szilard, filed in 1944 but kept secret until 1955, when it was finally issued. (This patent had its own "rights-taking" issues associated with it, because Szilard didn't really want to sign it off to the government, and only did so when Groves threatened to kick him out of the Manhattan Project.)
In 1959, the US Court of Customs and Patent Appeals denied Joliot et al.'s bid for the validity of their two reactor patent applications. The problem, the Court ruled, was that the French reactor patent writeup "is insufficient to teach those skilled in the art how to make and use the invention," as required by law. Compare the French patents with the Fermi and Szilard patent, and it's clear which of them in a vague idea about how to run a reactor, and which of them is an actual reactor.
Still, in retrospect, it seems to me a bit unfair to look back on a patent application written up 20 years earlier, for a technology which, by the time of writing, actually existed, and to ask whether the original write-up was sufficiently specific to be a valid patent claim. The French patents can't help but look elementary, but if they were for any other industry, and hadn't been declared secret, it's pretty likely they would have been granted at the time they were applied for.
What would the consequences have been if the French patents had been ruled valid? For one thing, the US would have had to pay the French physicists fair compensation for their use of reactors during the war and afterwards. The US probably would have declared it "in the public interest" under the Atomic Energy Act, which would have prevented the French physicists from dictating the terms under which reactors were used, but it still would have granted them royalties.
In retrospect, that doesn't look like the worst possible outcome, but in 1942, it wasn't clear how nuclear technology was going to be handled in the post-war, or how much control a patent-holder would have over US nuclear ambitions. And that, in the end, is what mandated an unusual letter, and unusual secrecy.
- For patent censor history, see my article "Inside the Atomic Patent Office," and for more on the whole Manhattan Project patent program, see my "Patenting the Bomb." [↩]