In my first month or so in Los Alamos (summer 1950—a year before Matterhorn got going), I, along with other members of Wheeler’s small team, focused on studying and calculating a variety of aspects of the classical Super and other thermonuclear ideas such as the alarm clock. (We also found time to start modeling the “shots” scheduled for the next May in Enewetak—see below.) I was, as the saying goes, getting up to speed, and was helping Wheeler and Teller prepare for their major presentation to the AEC’s General Advisory Committee that September. By some time in August, as I described in Chapter 10, we had put together a large report, “Thermonuclear Status Report Part I,” officially authored by Teller and Wheeler, and, as I mentioned earlier, unofficially known as the “telephone book” because of its bulk.{2} With suitable scientific caution, the report said, “As of August 1950 it is still impossible to say whether or not any thermonuclear weapon is feasible or economically sensible.” Despite that judgment, Teller and Wheeler were very far from advocating a slowdown or hiatus in thermonuclear research. Just the opposite. They attributed the uncertainty to an insufficiency of research. They said, in effect: With more scientific talent, especially in theoretical physics, and with more time and more calculating, there is a good chance that a thermonuclear weapon can be successfully designed.
A strength of the report was that it provided a good summary of all that was then known or hypothesized about thermonuclear weapons. (Recall that this was pre-radiation implosion.) From my personal perspective, pitching in to help prepare it was a great way to get rapidly into the subject that was to be the focus of my professional attention for the next two years.
In its report following its September meeting,{3} the GAC cautioned against letting the fusion program at Los Alamos drain too many resources from the fission program, but, as I noted in Chapter 10, also praised the thermonuclear efforts and recommended that the AEC provide more computing resources for those efforts (was it in the hope that more calculations would prove, once and for all, the impossibility of building the weapon—who knows?). Oppenheimer’s reaction was to express his “frustrated gratitude” to Teller and Wheeler and their troops.{4}
By the time of the “telephone book” report, Los Alamos, guided by its Family Committee, had decided on a test of thermonuclear burning the next spring, in what would be the George shot of the Greenhouse test series. The unusual cylindrical design of George’s fission-bomb trigger allowed for quicker flow of radiation from the fission bomb along its axis to the capsule of deuterium-tritium (DT), whose ignition was the goal.{5} The purpose of the design was only to get the energy promptly to the DT where it was needed, not specifically to compress the DT. As it turned out, the test was entirely successful, and even provided evidence that the radiation had caused some implosion of the DT capsule. The result of the George shot was very satisfying to all who had worked so hard to make it happen, yet it was not exactly cause for euphoria either, because by the time it took place, the idea of radiation implosion was in the air, and we were all optimistic that there was now a clear path to a workable weapon, even without the reinforcing evidence of the George shot.
What the George shot did do was provide what was very likely the first example on Earth of fusion triggered by heat. DT reactions, with their resulting 14-MeV neutrons, were commonplace in accelerator laboratories. This was the first time that those telltale neutrons were propelled from a reaction not as the result of an accelerator beam causing deuterons and tritons to fuse, but because of heat so intense that the D’s and the T’s acquired thermal energy sufficient to cause their fusion.
The Item shot, a couple of weeks after George in May 1951, also involved DT burning (and the concomitant release of high-energy neutrons), but, as I described in Chapter 10, in quite a different configuration. The thermonuclear fuel, instead of being off to one side of a fission bomb, rested in a central cavity within a spherical fission bomb. This was the “boosting” principle, in which the purpose of the DT burning is not so much to add fusion energy as to intensify fission energy. It, too, was entirely successful.{6}
In the months following the marathon effort in the summer of 1950 to prepare the report for the GAC, I spent time on calculations specific to George and Item, both of which were largely designed by the fall of 1950, and also pursued burning calculations on cylinders of deuterium and on other configurations such as deuterium bubbles in uranium or plutonium (“Swiss cheese”). By August 1950 Fermi and Ulam, using hand calculations (carried out largely by Miriam Planck and Josephine Elliott), had confirmed the pessimistic outlook of the earlier Ulam-Everett calculations for the classical Super. This dampened but certainly did not extinguish our enthusiasm for what we were about. Wheeler, like Teller, had an abiding faith that what was possible in principle could be—and would be—achieved in practice if we just kept scratching for new ideas and exploring a lot of them. Wheeler’s young colleagues—John Toll, Burt Freeman, and I—took our cues from our elders (Wheeler was all of thirty-nine and Teller was forty-two). We went at our calculations and modeling and theorizing with a sense that we would triumph if we worked hard enough. It was not unlike grappling with a particularly challenging homework problem.
Preparing materials for the human computers was my introduction to programming. Instructing these young women was actually not so different from instructing a mechanical computer. One had to lay out the arithmetic steps in detail, not only commands such as “multiply A times B and subtract C,” but also conditional commands such as “If the result of A times B is greater than D, go to step E; otherwise go to step F.” So my transition to programming for the IBM CPC (card-programmed calculator) was straightforward, and by the midpoint in my year at Los Alamos, I was regularly using CPCs for numerous calculations. But I wasn’t the only one. Others working on thermonuclear reactions, such as John Toll, as well as the team developing new fission weapons all came to rely more and more on the CPCs. Our “elders” counted on us “youngsters” to do the actual programming.
The CPC, first offered for sale by IBM in 1949,{7} was a modified accounting machine, about the size and weight of a commercial refrigerator. Instructions and data were fed into it with a stack of punched cards, each card having eighty columns and each column having space for up to a dozen small rectangular punches. A stack of, say, a hundred cards, would be placed in a hopper, then sucked into the machine at the rate of a little more than one per second. If there was a break of a few seconds in the otherwise regular click-click-click of the card reader, it meant that the machine was pausing to do something complicated like taking a square root. The cards emerged into a lower hopper from where they could be lifted back into the input hopper. At any given moment, there were only one or two cards from the input stack inside the machine. Still in the future—the very near future, in fact—were the so-called stored-program machines in which instructions residing within the machine directed the computer’s operations without further individual intervention.