Kurchatov used American materials for the dual purpose of double-checking the scientific results obtained by members of his team, and for evaluating the probability that the stolen secrets might contain purposely planted disinformation. Inside the Russian nuclear establishment legends were told of how his subordinates — the theoretical physicists — would report to Kurchatov with freshly calculated formulas. According to their accounts, Kurchatov would look carefully at their work, then silently open the safe with the precious stolen American secrets to compare the results. “No, it is not right,” he would say. “You have to work more and come again.”
And as for the theoretical physicists, so also for the chemists who analyzed the purity of F-l graphite. “Data of this sort [i.e., unacceptably high cross sections]… made the physicists depressed,” writes Panasyuk, “but they didn't stop measuring new batches of graphite from various sources of raw materials that were processed under various technological conditions. Finally, measurements that had just been finished of a recent shipment were put on Kurchatov's desk. Victory! For the first time a cross section of 8.6–0.4 ? 10–27 cm2 was obtained!” Kurchatov then personally checked a few samples from the earlier “bad” batches against the pure new graphite and proved that the chemists were wrong. “The method for measuring impurities was improved,” concludes Panasyuk of this discouraging episode, “and it appeared that the ‘bad’ batches were slightly contaminated by boron and rare-earth elements.”
The quality soon stabilized; the new graphite that began arriving was not only purer but also denser. By August there was enough graphite on hand — about five hundred tons — to assemble the reactor. To make sure the material was adequate, Kurchatov had the entire mass stacked together in the pit under the tent into a vast black cube twenty feet on a side to measure its average absorption cross section. “Measurements and calculations,” notes Panasyuk, “determined that cross section to be 4 ? 10–27 cm2.” Loaded with sufficient uranium, the reactor should work. Teams began drilling single blind holes in the graphite blocks for the uranium slugs, a total of thirty thousand holes by the time the work was done. Each uranium-loaded block would be surrounded on four sides by a set of plain graphite blocks to make a grid spacing of about eight inches — room for fission neutrons from one slug to bounce off carbon atoms in the graphite and slow down enough to resist absorption by U238 nuclei until they encountered a fresh U235 nucleus in the next slug along.
In the meantime, the Assembly Workshops went up around the reactor pit, a handsome brick building 130 feet long, fifty feet wide and two stories high. The pit floor and walls were lined with poured concrete, the sandy soil itself serving as the reactor's primary shielding. Construction workers dug a passageway from the sub- basement-level floor of the pit up past a baffle of special walls made of lead blocks and hollow bricks filled with a mixture of boron and paraffin to an underground control room. A laboratory on the first floor, into which the dome of the reactor would protrude, went unshielded. Radioactive gases would be removed through a fan system and vented into the Moscow air.
As a first step, workmen suspended three cadmium control rods above the center of the pit. Steel wires from the control rods ran over reels hung in the attic framing above the main laboratory and down into the underground control room, which was fitted with both electrically and manually operated winches. A bright brass submarine periscope in the control room would allow the F-l operators to observe the notches on the control rod above that would indicate its position within the reactor block. The two emergency rods were unnotched; they would be positioned either entirely out or fully within the reactor and could be scrammed (“emergency dropped,” the Soviet scientists called the operation) with the winches to quench the chain reaction. Workmen also installed red lights and sirens in the pit connected to radiation monitors. Electricians wired in backup electrical supplies to support the winches and the control panel.
Just as Enrico Fermi had done building the first man-made nuclear reactor in Chicago in the fall and early winter of 1942, Kurchatov proceeded toward a full-scale assembly by directing the construction of a series of smaller, subcritical assemblies. These would not contain enough uranium and graphite to achieve a self-sustaining chain reaction, but they would carry measurements and calculations incrementally forward in that direction and enable the scientists to learn what to expect from the novel process as they went along. Panasyuk calls these smaller assemblies “model” assemblies (Fermi's people called theirs “exponential” since an exponent entered into the calculation of their approach to critical mass). It was standard laboratory procedure in experimental physics in those days before computer simulation to build such functioning models to accumulate data that could be extrapolated to predict the operation of a full-scale machine.
“Kurchatov,” writes Golovin, “proposed to attain the critical dimensions by… increasing the diameter of the sphere each time and using all the available uranium prepared up to then.” Supplies and approximations would thus keep pace across the weeks. The reactor team laid down a flooring of graphite blocks a meter deep and began building the first model assembly in the center of the flooring. The core layers got the uranium and graphite of the highest purity, with less pure materials reserved for the periphery. When they dismantled the model assemblies, they stored the graphite blocks and uranium slugs on the floor of the main hall upstairs, moving the dirty, greasy, heavy materials by hand onto and off of a belt lift.
Kurchatov's team completed the first of the four model assemblies on August 1,1946, using 1.4 tons of uranium and 32 tons of graphite. “Layer by layer they would put in graphite and uranium,” team member Boris Dubov-sky recalls, “conducting measurements at the same time and processing the results. The work, as a norm, went on all around the clock.” Because of the backwardness of the Soviet radio industry,[26] the measurements group used electronics salvaged from German aircraft shot down during the war.
The third model assembly “alarmed everyone,” says Golovin. It demonstrated hardly any increase in neutron multiplication over model number two despite its larger loading of uranium; they feared some “essential miscalculation.” Kurchatov ordered additional measurements, which showed “that the third batch of uranium was considerably less pure than the rest.” “This discovery,” writes Panasyuk, “called for urgent reorganization of physical quality control for all batches of uranium produced by our industry.” The reorganization was effective; Golovin says the fourth model assembly, completed in early November, “reassured everyone that success was imminent.”
Kurchatov's team began assembling the full-scale reactor on November 10, 1946. By now the work was routine. The assembly materialized a layer at a time, swelling outward in a roughly spherical configuration that began to crowd the pit, looming overhead as it reached ground level, the black graphite soaking up the light, graphite dust adding to the gloom. They did not feel gloomy. They worked with increasing excitement.
There was barely enough uranium. Before they were through they had used up all the uranium metal available in the USSR at that time — forty-five tons. It was not enough. They had ninety kilograms of uranium oxide and 218 kilograms of powdered metallic uranium on hand that they had used for physical measurements in the years before Vannikov organized Soviet production of uranium metal. They pressed the impure materials into briquettes, loaded the briquettes into graphite blocks and laid the blocks around the periphery of the lattice. Three tons of graphite added near the end, a cockade of sorts, was American Lend-Lease graphite shipped during the war for searchlight electrodes.
Fortunately, the reactor approached criticality well before their rough initial calculations had predicted it would. That consequence, like the oxide loading and much else in the Soviet program, repeated the earlier American experience and defines the boundaries of what remote and fugitive espionage could pirate. “The density of neutrons in the effective center of… the reactor approximately doubled between layer 53 and layer 58,” notes Panasyuk; “because of this change, it was clear already at layer 58 that the… forecast dimensions (76 layers) were highly exaggerated.”
“It became obvious,” Golovin reports, “that there was going to be a chain reaction. The final layers of uranium were stacked behind extra shielding in case of an unforeseen runaway reaction.” They finished layer 61 on December 24,1946, in the evening. On the graph of neutron intensity that Kurcha-tov had been maintaining, it was obvious that layer 62 would cross the threshold of criticality. Kurchatov sent the workers home for a rest. People began trickling back to the pit in the middle of the night. Layer 62 was laid with all three control rods inserted into the reactor by two in the afternoon on Christmas Day. “Kurchatov was in another building at the time,” Panasyuk remembers. “We telephoned him and notified him that the reactor was ready.”
Kurchatov arrived in the underground control room to direct the startup. He was sufficiently concerned about a possible accident that he cleared the building even of guards and had the area cordoned, authorizing only his crew of four immediate assistants, which included Boris Dubovsky and Igor Panasyuk, to remain. Fermi had trusted his calculations more; there had been a crowd on hand when CP-1 had started up in Chicago four years previously, even
