The First War of Physics (42 page)

Enriched uranium began to arrive at Los Alamos from Oak Ridge in early 1945, largely thanks to a revelation that Oppenheimer had experienced about eight months previously. The scientists had backed each method of isotope separation like horses in a race, without realising that connecting the methods in series – using the output from one method as the feedstock for another – might offer a more efficient route to enriched uranium and weapons-grade U-235.

As early as January 1943, Philip Abelson had proposed to use a liquid thermal diffusion technique to enrich uranium for reactor research at the Met Lab. Abelson was working for the US Navy, and compartmentalisation meant that Oppenheimer did not learn of this work until over a year later. In April 1944 he had realised that the horse-race analogy was ‘a terrible scientific blunder’. He now saw that using even slightly enriched uranium as a feedstock for the calutrons at Oak Pudge would greatly increase their efficiency. Abelson’s thermal diffusion plant could provide a temporary alternative to the gaseous diffusion plant K-25, still delayed by problems with the porous barrier materials. Groves authorised the construction of a thermal diffusion plant at Oak Ridge in June 1944. He gave the contractors a 90-day deadline.

In fact, satisfactory barrier materials for K-25 were delivered to Oak Ridge and the first charge of uranium hexafluoride passed through the plant on 20 January 1945. The thermal diffusion plant, S-50, was enriching uranium by March.

With regular supplies of enriched uranium now arriving on the Hill, Frisch devised an ingenious way to test their theories and, at the same time, determine precisely how much fissile material would be required to make a bomb. He submitted a proposal for a series of experiments to the Co-ordinating Council, which oversaw the various projects at Los Alamos. Much to his surprise, his proposal was approved.

The group had plenty of experience with assemblies formed from stacked blocks of uranium hydride, and had advanced towards a critical
assembly by reducing the hydrogen content as the proportion of U-235 was increased. Such a ‘naked’ assembly, which Frisch called a ‘Lady Godiva’, was quite dangerous. Frisch himself almost suffered what would have been a fatal dose of radiation when he leaned too closely to one such sub-critical assembly. His own body reflected back some of the neutrons that would otherwise have escaped harmlessly, and these reflected neutrons caused the assembly to go critical. He noticed that the small red lamps monitoring the neutron intensity had stopped flickering. Instead, they were glowing continuously, the neutron counters by now overwhelmed. He hastily shut the experiment down.

The challenge was to figure out how to work with critical and supercritical assemblies in relative safety. Frisch’s idea was to assemble blocks of enriched uranium hydride into a near-critical configuration, but with a hole running through the centre. Through this hole would then be dropped another block of enriched uranium hydride – called the core – sufficient to make the whole assembly go critical just for an instant as the core passed through the hole and dropped out the bottom.

Feynman, sitting in judgement on the Co-ordinating Council, found the experiment intuitively appealing. He said it was ‘like tickling the tail of a sleeping dragon’. The experiment was henceforth known as the Dragon experiment. ‘It was as near as we could possibly go towards starting an atomic explosion without actually being blown up.’ There were obvious dangers. If the core were to get stuck on its way through, the assembly would go critical and irradiate the physicists with potentially lethal doses of radiation. Frisch was confident that the experiments could be done safely, however, and stipulated that nobody should ever work on the assembly alone.

Now working as a group leader, Frisch built the first of a series of such assemblies at a small laboratory in Omega Canyon, some distance from the main Los Alamos laboratory facilities, during the winter months of early 1945. He worked around the clock to make the first precise measurements of critical mass in U-235. The experiments were very successful. In the split second during which the core passed through the assembly, a large burst of neutrons was produced and the temperature of the apparatus increased by
several degrees. The largest measured energy production was twenty million watts, for just three thousandths of a second, raising the temperature of the assembly by six degrees. This was the first time a super-critical mass of enriched uranium had been studied in the laboratory. By early April 1945, sufficient U-235 was available to form assemblies from blocks of pure metal.

Following the script

The experimental X-10 reactor at Oak Ridge represented an intermediate scale between the first Chicago pile and the large reactors constructed at Hanford. The Chicago pile had operated at a very modest output – barely one watt. X-10 had achieved one million watts. The three Hanford reactors, designated B, D and F, were built and operated by Du Pont and designed to run at 250 million watts. Each reactor consisted of a cylinder of graphite measuring 28 feet by 36 feet, weighing about 1,200 tons, containing 2,004 equally spaced aluminium tubes drilled along its length. Uranium slugs the size of rolls of quarters were sealed in aluminium cans and inserted into the tubes. Cooling water was pumped through the tubes and around the uranium slugs at the rate of 75,000 gallons per minute. The reactors had one purpose only: to produce plutonium. No attempt was made to capture the heat energy released by the reactor and convert it to electricity.

Fermi oversaw the loading of the uranium fuel slugs into the Hanford B reactor on 15 September 1944 and brought the reactor to ‘dry’ criticality, low-power operation close to the threshold of criticality which did not require cooling water. The physicists at Hanford then added more fuel slugs and carried out more experiments. Everything was in order. The reactor behaved precisely as it should. It was pushed towards full power on 26 September.

John Wheeler had joined Wigner’s group at the Met Lab in early 1942 and moved to the Du Pont offices in Wilmington, Delaware in March 1943. A year later he relocated once again to Hanford. Due some late nights ‘babysitting’ the reactor later in the week, he had decided to go home and get some sleep on the night of the 26th. But when he got to his office the
next morning he found that not all had gone to plan: ‘the reactor was not exactly following the script.’

The reactor had started up as expected and had reached a record output of nine million watts. Then it began to lose reactivity, and as the output declined the operators had tried to maintain it by withdrawing the cadmium control rods. ‘It was as if the engine of your car got sick as you were driving along a level road,’ Wheeler later wrote, ‘and you had to push farther and farther down on the accelerator pedal to maintain speed; eventually the pedal would be all the way to the floor and the car would start to slow down.’ By mid-afternoon on 27 September, the control rods had been pulled virtually all the way out to maintain the reactor output. By early evening the control rods were all the way out, yet still the reactor shut down.

Fermi wondered if water had leaked into the reactor, but Wheeler suspected something else. Only a few weeks after joining the Met Lab he had written a report on the possibility of ‘self-poisoning’ of the reactor by its own reaction products.

When in 1938 Hahn and Strassman found barium among the products resulting from the bombardment of uranium with neutrons, they had discovered the stable end result of a long and complex series of nuclear reactions. When U-235 absorbs a neutron, the unstable U-236 nucleus fissions. One of the possible fission reactions produces zirconium, Zr-98, and tellurium, Te-135, and three neutrons. The zirconium isotope is radioactive, and goes on to produce niobium and then molybdenum. Likewise, the radioactive tellurium isotope decays first into iodine, then xenon, then caesium, and finally barium.

If just one of these decay products has a high affinity for neutrons, Wheeler had reasoned, then it would tend to inhibit the chain reaction, soaking up free neutrons until there were insufficient numbers to keep the chain reaction going. As more and more of the ‘poison’ was produced, it would become more and more difficult to maintain the reactor output. Eventually, the poison would overwhelm the reaction and the reactor would shut down. Wheeler had made some further estimates in April 1942 and concluded that self-poisoning would be a significant problem only if
one of the intermediate reactor products had an appetite for slow, ‘thermal’ neutrons about 150 times that of U-235 itself.

When the B reactor was examined, no water contamination could be found. Self-poisoning became the most obvious conclusion. In the early hours of 28 September the reactor recovered, building again to nine million watts by the afternoon before declining again. This further suggested that the poison was itself radioactive, with a half-life of about eleven hours, roughly the time it took the reactor to come back to life. Wheeler checked a table of measured half-lives and identified the culprit. It was the isotope of xenon, Xe-135, later found to have an appetite for neutrons more than 4,000 times greater than that of U-235.

Fixing the problem was relatively straightforward once it was understood. Of course, nothing could be done about the physics of the nuclear reactions. The reactor would always poison itself with Xe-135. The solution was to play a numbers game, adding more uranium fuel to the reactor to ensure that more neutrons would be produced than could be absorbed by Xe-135 at its steady-state concentration. Fortunately, the reactor design had allowed for such contingencies: extra tubes had been drilled at substantial cost and delay to the reactor construction. Wheeler’s prudence now paid off. The extra uranium fuel that was required could be added without a major redesign and rebuild.

The Hanford D reactor went critical on 17 December 1944, and the repaired B reactor went critical once more eleven days later. The F reactor went critical in February 1945. On 4 February the reactors hit the designed operating output of 250 million watts. Plutonium production was now in full swing, with a theoretical yield of 21 kilos of plutonium per month. A jubilant Groves estimated that they would have enough plutonium for eighteen atomic bombs in the second half of 1945.

Solid core compression

But the plutonium would serve no purpose unless a way could be found to detonate it. Kistiakowsky’s X Division had laboured hard through the winter of 1944–45. The woods surrounding Los Alamos reverberated to
an endless series of explosions that increased in intensity as the scientists scaled up their experiments. The group consumed about a ton of high-performance explosives each day, cast into moulds to produce shaped charges weighing about 50 pounds each and machined with elaborate precision.

To support the push for implosion, G Division scientists had developed a series of diagnostic tests that could be used to tell just how symmetrical the implosive Shockwave was. In addition to the Ra-La experiments that Hall was involved with, the scientists were also using various types of X-ray photography, high-speed photography and magnetic field measurements. Von Neumann had designed an arrangement of explosive lenses consisting of a fast-burning outer layer and a slow-burning inner component which worked together like a magnifying glass, shaping and directing the Shockwave towards the bomb core. Each lens transformed the initial explosive burst from a spherical Shockwave expanding outwards to a spherical shockwave converging to a central point. A second layer of fast-burning explosive accelerated and strengthened the implosive wave.

On 7 February a Ra-La test showed encouraging improvement, although spherical compression of a solid core could not yet be demonstrated. It was progress, but lens development was still behind schedule. At a meeting on 28 February involving, among others, Oppenheimer, Groves, Conant, Bethe and Kistiakowsky, the chemical composition for the explosive lenses was finalised and the overall design approach for a plutonium bomb agreed. On 1 March Oppenheimer established the ‘Cowpuncher’ committee, headed by physicist Samuel Allison – recently freed from his responsibilities at the Met Lab – and including Bacher and Kistiakowsky. The committee’s job was to ‘ride herd’ on the final stages of plutonium bomb development. A few days later Oppenheimer called a halt to any further refinement of the explosive lens design.

Kistiakowsky was wary of Allison, who he suspected had been tasked by Oppenheimer to watch over his shoulder. As pressure mounted, nerves began to fray. Although Oak Ridge was now reliably producing weapons-grade U-235, in the likely timeframe of the war, the production rate was sufficient only for a single bomb. Hanford was now reliably producing
plutonium sufficient for several bombs. But a simple gun-type detonation scheme could not be used successfully with reactor-bred plutonium. Everything now hinged on the development of implosion. Kistiakowsky, a chemist in an elite community of physicists, was now confronted with battle lines drawn up based on scientific discipline. ‘On one occasion I was forced to say to Oppie in this top level council where I was the only chemist, “You’re all ganging up on me because I’m not a physicist.” To which Oppie replied smilingly, “George, you’re an outstanding third rate physicist”.’

Although their confidence had grown as the work had progressed, the many uncertainties of plutonium bomb design were still painfully apparent. No matter how valuable the quantities of plutonium that were now beginning to arrive from Hanford, the Los Alamos scientists could not be sure that the Fat Man design would work without a full-scale test.

Plans for such a test had been laid the previous year. A site had been identified at one corner of the US Air Force’s Alamogordo bombing range in the New Mexico desert. It measured 24 miles long and eighteen miles wide. Inspired by a John Donne sonnet – ‘Batter my heart, three person’d God … Your force, to break, blow, burn, and make me new’ – Oppenheimer codenamed the site Trinity. Oppenheimer appointed Harvard physicist Kenneth Bainbridge to plan and direct the test.

By mid-March, experimental evidence had been obtained for solid core compression in an implosive Shockwave so smoothly symmetrical that the results agreed closely with theoretical predictions. The news must have been accompanied by considerable sighs of relief all round. On 11 April, Oppenheimer wrote to Groves to tell him the good news. The production rate of U-235 at Oak Ridge had suggested that a uranium bomb would be ready by 1 July. Oppenheimer now advised Groves that a plutonium bomb would be available by 1 August.