…we are happy to note that great progress has been made at Los Alamos in setting up quantitative methods for coping with the difficult problems of the hydrogen bomb; we urge the Commission to support these efforts by making computational machinery available to the laboratory.
Our second activity in that interim period was calculating expected results of newly scheduled tests of thermonuclear principles in the Greenhouse test series scheduled for Enewetak[58] in May 1951. I will discuss these in the latter part of this chapter.
Regarding Greenhouse, the GAC, in the report of its September 1950 meeting, had this to say:{6}
We wish to make it clear… that the test, whether successful or not, is neither a proof firing of a possible thermonuclear weapon nor a test of the feasibility of such a weapon. This test is not addressed to resolving the paramount uncertainties which are decisive in evaluating the feasibility of the Super.
Thank heaven all atomic nuclei are positively charged. That means, as discussed in Chapter 6, that they all repel one another and keep their distance—at least at temperatures less than some tens of millions of degrees.[59] If any nucleus were neutral or negatively charged, it could sneak up on one of its positively charged brethren and react with it or unite with it, producing a mini-explosion and releasing a great deal of energy—at least by atomic standards. If the nuclei belonged to relatively light elements, below iron in the periodic table, the result of their union would be an energy release from a hundred thousand to a million times more than is typical for a pair of atoms in a chemical reaction. If the universe were made of nuclei that could so easily get together, it would contain no cold cinders like Earth or Mars. There would be only glowing suns and glowing planets.
Higher temperature means higher speed for the atoms and nuclei. At high enough temperature, such as exists at the center of the Sun or in an exploding H bomb, some of the caroming nuclei, as I discussed in Chapter 6, can surmount the barrier of electric repulsion between them and, so to speak, reach out and touch one another. Think of yourself accelerating a car on a level roadway, then shifting into neutral and seeing how far up a slope you can roll. If the slope is Pike’s Peak, it doesn’t matter whether you approach at 10 miles per hour, or 50, or 200. You still won’t reach the top. That is like nuclei at ordinary, or even somewhat elevated temperature. Now imagine that you could increase the car’s energy a thousand-fold. Then you might have a shot at rolling to the top of the mountain. That is like nuclei at millions of degrees.
What happens then when the nuclei do “touch”? Consider a pair of deuterons. There are two possibilities and they occur with nearly equal probability. Written like chemical formulas, those reactions are:
D + D → 3He + n (energy release 3.3 MeV)
D + D → T + p (energy release 4.0 MeV)
The two reacting deuterons contain a total of two protons and two neutrons. It is easy to see that the products of their reaction, either a helium-3 nucleus and a neutron, or a hydrogen-3 nucleus (denoted by T for triton) and a proton, also contain two protons and two neutrons.
Those energy releases amount to about 1 MeV per nucleon (n or p), about the same as the energy release per nucleon in the fission of a uranium nucleus. But there is more. The triton created in half of the D-D reactions reacts with a deuteron to create an alpha particle, or helium-4 nucleus. The reaction is:
D + T → 4He + n (energy release 17.6 MeV)
The neutron that emerges from this reaction takes 80 percent of the energy that is released, about 14 MeV. Since 14 MeV is considerably greater than the energy of any other particle that is involved, this particular neutron is relatively easy to detect and provides a “signature” of fusion (or, if you prefer, a smoking gun). It is this DT reaction that inspired the graphic used in this book’s section dividers.
With this added reaction, the energy release per unit mass is more than twice that of fission. And the cost of the input material, deuterium, is much less than the cost of uranium-235 or plutonium-239. On top of that, there is no limit to how much deuterium can be stored in one place. Just as for wood or coal or oil, there is no critical mass. And the neutron furnished by the DT reaction has enough energy (14 MeV) to stimulate fission in uranium-238, not just uranium-235. So it is easy to see why, right from that 1942 conference in Berkeley, thermonuclear weapons were seductive for the physicists involved in weapons design.
It turns out that the DT reaction not only releases much more energy than the DD reaction, it also proceeds more vigorously—that is, with higher probability and at somewhat lower (although still enormous) temperature. That is why Ulam and Fermi investigated the possibility of starting with some tritium in the fuel, not waiting for it to be formed in the DD reaction. It would not be surprising if they also considered a scheme like the one first imagined by von Neumann and Fuchs—a Super containing some tritium in its ignition region and only deuterium in its propagation region (the thermonuclear equivalent of a fire started with kindling and kept going with large logs).
Theorists like deuterium because it is a fairly simple substance and this facilitates computation. Engineers and Air Force Generals don’t like deuterium. It is a gas at ordinary temperature, and can be made dense only by liquefying it at a temperature of 23 kelvins above absolute zero, or 250 Celsius degrees below zero. The theorists’ point of view prevailed for the first large-scale thermonuclear test (Mike, November 1, 1952). The Mike device weighed 62 tons{7} and was a very long way from being portable on an airplane or missile (or even a truck). From soon after Mike, and up to the present day, so far as I know, the thermonuclear fuel of choice has been lithium-6 deuteride, or 6Li2H. This substance has two advantages over liquid deuterium. First, it is a solid at ordinary temperature. You might say that the lithium has provided a matrix that enables the deuterium atoms to cozy up to one another. Second, the lithium adds another useful nuclear reaction to the three I described above. It is
36Li + n → T + 24He (energy release 4.9 MeV)
The subscripts show the atomic number (the number of protons in the nucleus, which is the element’s place in the periodic table), the superscripts show the total number of protons and neutrons in the nucleus. As before, T stands for triton, the nucleus of tritium. The neutron on the left side of the reaction equation results from the DD reaction (the first branch in my list above). Only computational complexity kept us from considering lithium-6 deuteride for the first thermonuclear tests.
Nowadays, to the extent that there are any new nuclear weapons in the United States arsenal,[60] they must be designed, built, and deployed without testing (a task not quite as formidable as it might at first sound thanks to the vast store of knowledge from past testing and the power of modern computers).
In the early days of nuclear weapons, on the other hand—during and soon after World War II—testing was considered essential, starting with the Trinity test in Alamogordo in July 1945. (To be sure, the Hiroshima bomb was used without testing, a choice dictated by its conservative design and, significantly, by the short supply of uranium-235 at that time.) Immediately after the end of the war, scientists and managers at Los Alamos began planning for tests as new designs took shape.
59
Kelvin degrees, not meaningfully different from Celsius degrees when you are talking millions. (See page 67.)
60
The US stockpile of nuclear weapons declined from a high of more than 31,000 in 1967 to an estimated 7,700 in 2013, with a projected 3,600 in 2022.{8} It is reasonable to suppose that even as the numbers decline, newly designed weapons are replacing those of older design in the stockpile.