boils at 77°K, then cycled it through compression and expansion with a piston and cylinder arrangement. It was because of Kapitza's pioneering cryogenic work that Teller and others had worried, after Joe 1, that the Soviet Union might steal a march on the thermonuclear.
In 1946, an MIT physicist named Samuel Collins had developed a helium liquefier superior to Kapitza's, the Collins Helium Cryostat. Collins, writes the physicist who directed the work on deuterium production at the National Bureau of Standards laboratory in Boulder, “using the same principle as Kapitza, redesigned, modified and then superbly engineered his liquefier to provide a complete, relatively inexpensive, reliable and simple-to-operate [liquid helium] facility. Indeed, Collins's contribution to low-temperature research has been aptly compared with that of Ford to the automobile.” Collins commercialized his helium liquefier through Arthur D. Little, Inc., of Cambridge, Massachusetts, and it was to ADL that Wechsler applied late in 1951 to organize the production of big transport dewars by ADL's Cambridge Corporation that could store hundreds of gallons of liquid deuterium and other liquefied gases at Eniwetok. “ADL talked to Sam Collins,” Wechsler remembers. “They asked him, Hey, can we use your helium cryostat to liquefy hydrogen? You wouldn't do that for commercial production, but their idea was, if you could build a big dewar for storage, you could have a little Collins cryostat attached to the big dewar and by keeping on cooling the hydrogen, recirculating it, you'd never lose any. All you had to do was to keep the cryostat running. That's a really unique idea.”
Coordinating the work at the Cambridge Corporation, Wechsler learned to live on the road. “We had to have dewars by late spring at Los Alamos, starting from January 1952. Built from scratch. I got to where I knew all the Pullman conductors, because air travel was not that good.” He would take the train to Boulder to check out the NBS laboratory, fly a redeye to Washington for a meeting there, sleep on the overnight train from Washington to Boston, work with the dewar development team, take the New England States from Boston to Chicago and then change to the Santa Fe, which gave him a compartment where he could sleep on the way back to Los Alamos as well as write up his notes. “That was a standard route for me, one little aspect of what went on in that period of months.” Other managers were circulating around the country on similar routes. American Car and Foundry eventually opened a plant in Albuquerque to make bomb casings, but before that, someone cycled back and forth to Buffalo. Carrier Corporation supplied compressors. Distillation Products supplied pumps. “We had stuff being shipped from all over the country,” Wechsler reminisces. “Not just to Los Alamos. Some of the things never came here. Some of them had to go straight overseas.”
The big storage dewars that Wechsler developed were considerably more sophisticated than James Dewar's original double-walled flask. Wechsler's dewars incorporated nitrogen-cooled shields, Styrofoam and aluminum-foil insulation and other tricks of the cryogenic trade. That was the first time Wechsler saw complex shapes cut from Styrofoam with a hot wire. Each stainless-steel dewar had a capacity of two thousand liters of liquid, about 530 gallons; they were mounted on diesel truck flatbeds with a motor-generator built onto the flatbed ahead of the dewar housing to supply electrical power. Wechsler ran one cross-country from Boston to Boulder fitted with recording accelerometers to measure how well the shock-absorber system worked. All the tubes for transferring the liquids in and out of the dewars were themselves dewars: vacuum-jacketed stainless steel.
Deuterium production started early, Wechsler emphasizes:
Where do you get enough deuterium to do this job? You start with heavy water. Now you've got to electrolyze the heavy water to break it down to oxygen and deuterium. Do you know any easy way to do that? It's tough. We had a little set of electrolytic cells up there at Boulder. They used 55-gallon drums of heavy water, dumping them in, electrolyzing, collecting the deuterium, pumping it at low pressure into big standard gas holders like you use for natural gas. The liquefiers weren't up and running yet and we didn't have the storage dewars yet but we had to get started because making the deuterium was a long, slow deal. So we had to store the deuterium in gaseous form, and that was the way it went out to Eniwetok. They ordered a bunch of tube banks, big 2,000- pound tube trailer banks pulled with semi's that could be loaded into ships. You could use the tubes for nitrogen, oxygen, you could get them for hydrogen. How many tube trailers did you need? How many places in the country make new, clean, safe tube trailers? They had to be specially modified because we didn't want any standard safety valves on them that might vent the deuterium because we didn't want to lose any — those tubes filled with deuterium were worth more than gold.[47]
A plant at Eniwetok operated by Ohio State University would liquefy the gases after their delivery by ship. The Cambridge Corporation began shipping storage dewars for Los Alamos in early April 1952.
The cryogenic system for the Sausage was similar to the system in Wech-sler's storage dewars but simpler. Liquid hydrogen boils at 20°K, liquid deuterium at 23.5°K, which means liquid hydrogen stores a few degrees colder than liquid deuterium. Taking advantage of that difference, Panda cryogenicists designed the double-walled stainless-steel dewar in the Sausage to connect through pipes in its upper end to a reflux condenser set in a big tank of liquid hydrogen. Vaporizing deuterium would flow by convection through a pipe to the reflux condenser, where it would cool to a liquid again and flow through another pipe back into the Sausage dewar. With this system of continual circulation and cooling dumping any heat the deuterium picked up, the cryogenicists were able to dispense with Styrofoam and aluminum foil. The Sausage dewar, suspended within the big Mike casing — a cylindrical double-walled stainless steel tank with a rounded top and bottom and with the sparkplug assembly mounted inside on a central column that ran the length of the tank — would contain several hundred liters of liquid deuterium. A second evacuated assembly would surround it. Between the second, outer assembly and the dewar, the cryogenicists ingeniously interposed a single floating thermal-radiation shield — another thin-walled tank, probably made of copper, a good reflector of radiant heat. The radiation shield “floated” in the sense that it touched the dewar it contained and the outer assembly that surrounded it in as few places as possible, because any contact would allow heat to flow into the dewar by convection. The contact necessary to hold the components in position was probably made with laminated stacks of thin metal disks, which conduct heat as much as two hundred times less efficiently than would a solid metal bolt.
A thermal-radiation shield floating in a vacuum can significantly reduce radiant-heat transport from a warm exterior to a cold interior. Without it, Wechsler observes, “you're talking a cold surface and a warm surface, and the temperature difference is a couple hundred degrees Kelvin. I don't care if you've got a vacuum between them, the heat leak into the cold surface is serious. But there's a neat little trick. If you can put in a surface with an
The outer assembly of the Sausage secondary would be warm as the Eniwetok air and the Panda cryogenicists wanted more heat-loss reduction than a single floating shield could accomplish. Rather than clutter the secondary with multiple floating shields, they borrowed another trick, one that Kapitza had used in his helium liquefier. “If you can make a shield float at a temperature lower than it would normally,” Wechsler says, “then you can pick up the reflectivity of fifty thermal-radiation shields.” To lower the temperature of the copper shield, they welded a pan onto its bottom that they kept filled with liquid nitrogen. That cooled the shield. So the liquid deuterium dewar at around 20°K saw a copper shield cooled to liquid-nitrogen temperatures, around 76°K, and the copper shield in turn saw the ambient-temperature outer assembly of the secondary.
The outer assembly was a marvel, the piece de resistance of the system. To appreciate its design and function requires going back to the point where the Panda designers had concluded that the soft X rays from the primary would not themselves exert enough pressure on the secondary to deliver the high compression necessary to prepare deuterium for thermonuclear burning. Instead, they decided, they needed to line the Mike casing with a material that the X rays could ionize into a hot plasma that would expand rapidly and deliver the necessary shock.
Carson Mark foresaw another complication as well, Wechsler remembers. “He was really, really concerned about higher-Z materials [i.e., materials of higher atomic number] being exposed. In a radiation environment, with high-energy radiation coming down from the primary, anything like steel — because it's so dense — will cause a pressure pulse when it vaporizes. Carson was trying to sustain a deuterium burn, and he was afraid that if things blew off at higher Z that might chop up the fuel, its temperature wouldn't stay high enough.” The solution was to shield Mike's welded-steel outer casing:
