The Graphite Reactor's control room in the 1940s.
'A Visit from Saint Nucleus'
Ed. note: In 1976 ORNL Review published a U.S. Bicentennial edition that also served as a history of ORNL at 25 years. Henry Newson began his reactor career with the Metallurgical Laboratory at the University of Chicago and moved to Oak Ridge for the construction and startup of the X-10 Pile, or Graphite Reactor. He subsequently worked at Hanford, Washington, and Los Alamos, New Mexico, before returning to ORNL in 1946 and eventually moving to Duke University. Newson contributed the following reminiscence, titled "A Visit from Saint Nucleus," of the Nov. 4, 1943, startup of the Graphite Reactor to the issue:
‘Twas the night before critical and a dozen or so creatures (in addition to numerous field mice) were stirring about the Graphite Pile. The monitoring procedures necessary to make sure that we did not accidentally overload had started about noon with the loading itself. The safety rods and the detecting instruments, which would scram the pile in case of an unexpected critical, were of course activated, but they were not sensitive enough to give significant readings yet. Unlike the safety rods, the horizontal control rods were inserted until each scheduled batch of uranium had been loaded; then they were withdrawn and indium foils were inserted at the center of the pile, withdrawn after a few minutes, and taken to the counting room. After a few minutes and a short calculation, the count was compared with the amount of uranium metal in the pile. During all these operations a boron counter was also registering in the count room. This counter was not accurate enough to give us quantitative readings, but it would give a warning if critical were approached much sooner than expected. Thus while watching the instruments and keeping the automatic safety systems activated, we loaded a batch of uranium, pulled out the rods, measured with the indium foils, reinserted the rods, added another batch of uranium, etc. Even the most accurate measurements (those with the indium foils) had little meaning until several tons of uranium had been added in such a way that the part of the graphite cube containing the uranium was close to a cylindrical shape.
However, measurements were made for what they were worth, even though our early points, while indicating roughly that things were proceeding as expected, had little significance. I remember showing an early plot to Martin Whitaker (Clinton Laboratories’ director) in the cafeteria at about dinner time and remarking jocularly that the pile was jumping up and down. This turned out to be a very indiscreet remark, since it was overheard by one of our managerial friends who might have taken alarm and assumed that the “jumping” was unexpected and ominous. Whitaker declared hastily that these early data had “no earthly significance.” Shortly after this incident, we did achieve a sufficiently regular shape of the loaded region so that we could start taking our measurements seriously.
Our routine now went smoothly for the next 8-hr shift; the count taken on the indium foils increased gradually as each batch was loaded. At first this increase was largely due to the fact that each batch of uranium contributed additional neutrons to those already bouncing around inside the graphite cube, but as the loading increased, the effect of the chain reaction became dominant. We could see it by observing the boron counter after each batch was loaded and the horizontal rods were withdrawn completely. The counts continued to increase for several minutes before reaching the constant counting rate which showed that the chain reaction was going on but had not reached critical. At this point the indium foils were inserted, since boron counters in those days were not accurate enough for our purpose; after each foil was counted, we could determine whether the loading of the next scheduled batch could be completed without reaching critical.
This procedure was designed to be tantalizing. We actually plotted the reciprocal of the (saturated) activity of the indium foils against a function of the amount of uranium and extrapolated from the last two points to zero reciprocal counting rate, which meant criticality. The function of the amount of uranium was chosen so that, while the plot could show that the next batch would not take us to critical, we could not tell how much more would be needed.
As midnight approached, the shift changed and the curiosity seekers went home. That left only those of us who were scheduled to stay for the graveyard shift. This was not meant to be an honor, since according to calculations originating at the University of Chicago’s Met Lab, it should take all night to load just up to critical. Enrico Fermi was scheduled to arrive from Chicago in the morning and preside over the finale.
Shortly after midnight we saw that, while the next batch would not bring criticality, the one after that might. Since we turned into the home stretch much sooner than expected, the operating group occupied the loading elevator and handled the uranium, which had been segregated into batches in advance. This group was composed entirely of Du Pont senior engineers under the supervision of Kent Wyatt. The counting room was manned by Haydn Jones, George Packer, and several junior Du Ponters. George Weil (the only Chicago representative present) and I stayed in the plotting room. We were authorized to start or stop the loading as necessary to bring the pile just to critical. In the unlikely event that it should happen before morning, we were to notify Arthur Compton, Whitaker, and a considerable number of others by messenger and wait for further instructions until they arrived at the pile. At that time there was no telephone communication between Oak Ridge houses and “Clinton Labs.”
As we proceeded, the operating group became more and more proficient with the long poles they used to push the uranium slugs into the pile. Batch after batch was loaded, as each extrapolation indicated that two more batches would be enough. It began to look as though the predicted critical mass was correct after all, in which case this would go on all night! But finally a point was plotted which predicted that the next batch would be too much! Unfortunately, we had not told the operators to stop loading during the counting and plotting, when we thought they could only add a negligible amount. I hurried out and gave the word to stop. The operators complied grudgingly. They knew that critical was not expected for hours and thought that I had panicked and stopped the loading prematurely.
I was at least spared this embarrassment. When the operators finished loading the hole they had started, plugged it up, and removed the control rods, the boron counter flashed faster and faster. That last hole was one too many! We were over critical by 3 “inhours.”** I had exceeded my instructions, but by a small amount compared with that needed to operate the pile under design conditions. The next day Fermi ordered enough uranium to be added to bring the activity up to more than 1000 inhours.
*Whitaker was the director of the Clinton Laboratories, later known as ORNL.
** Inhours is a unit of reactivity in a nuclear reactor