This report is based on information obtained from the daily reactor operating sheets (HAN-45954), from various monthly reports of the Technical and Production Divisions, from a history of the design and procurement of Hanford (INC-06263) and from a draft copy of a history begun by John Marshall (HW-7-2493).
I worked as a physicist at Hanford's N-reactor in the early 70's and continued work in various areas of nuclear weapons material production after N-reactor was decommissioned. I spent 3 years at the International Atomic Energy Agency in Vienna, Austria and worked for various US organizations in the area of nuclear nonproliferation. I retired in 2000. I originally wrote this short report in 1997 and put it on a website now long gone. I came across the files while cleaning out a file cabinet and decided to put it on a blog. I hope you enjoy it.
Selection of the Hanford Site
The process leading to the desire to construct a plutonium production plant began with research conducted at Columbia University in 1939. In may 1940, Princeton University suggested a process for producing plutonium based on the Columbia research. This was followed by a University of Chicago (UofC) recommendation that there be a more in-depth investigation into the possible military uses of plutonium. In July 1941, a National Academy of Sciences (NAS) committee took this recommendation under advisement. In November 1941, NAS proposed to the Office of Scientific Research and Development (OSRD) that a "vigorous" investigation be undertaken to study the military uses of plutonium. In December 1941, OSRD initiated this investigation at the University of Chicago Metallurgical Laboratory under the guidance of Dr. A.H. Compton.
In June 1942 the Army Corps of Engineers (ACE) was directed to assume responsibility for the direction of the plutonium program even though there was no experimental proof that such a process would work. While the research was continuing at UofC, the ACE arranged for Stone and Webster to design and construct experimental facilities at UofC, a pilot plant at Argonne, and a full scale production plant at Clinton Tennessee.
It was originally proposed that the Argonne site be used to experimentally confirm that the process would work. However, labor difficulties at Argonne delayed the construction work there. As a result, the experimental work was conducted under the west stands of the UofC athletic field. There Fermi's CP-1 reactor achieved a critical reaction on December 2, 1942.
In September 1942, Stone and Webster became more concerned about the almost insurmountable difficulties associated with design and construction of a full scale plant to produce and recover plutonium. In response to their request to obtain additional assistance, the ACE made initial contacts with E.I. du Pont Company. Du Pont's initial response was to decline due to the lack of experience in such an endeavor and the unproven nature of the processes. They later acceded to the government's request, and agreed to take on the contract "...for securing the development of design, procurement of equipment, and construction of a plant for manufacturing Product X-10."
The contract with Du Pont called for the large scale production plant to be built near Knoxville, Tennessee. However, due to the unpredictable nature of the processes and the potential hazards, Du Pont was unwilling to proceed with the location near Knoxville. It was at this point that the following site criteria were established.
- A manufacturing area of approximately 12x16 miles.
- No public highways or railroad tracks within 10 miles.
- No town with a population greater than 1000 within 20 miles.
- The availability of at least 25,000 gpm of pure water.
- The availability of at least 100,000 kw of electricity.
Reactor Design
Design of the production scale plant was initiated by the University of Chicago Metallurgical Laboratory. They turned over their initial design report to Du Pont which outlined a production reactor design with the following features.
- The reactor core was to be a 20' diameter by 20' tall graphite cylinder containing uranium in a lattice configuration.
- The core was to be encased in a steel container with a funnel shaped lower end and a hemispherical top. The top was to be removed for refueling.
- The reactor was to be surrounded by a 38' diameter water filled concrete tank that extended 8' above the top of the reactor.
- Vertical, 5" diameter ducts were to be used for the fuel and helium coolant.
- Fuel assemblies were to be made up of flat uranium plates set side by side in a graphite cassette 4 15/16" in diameter by 11" long with provisions for coolant flow on all surfaces of the fuel.
- Fuel discharge was accomplished through a perforated drum valve at the outlet of each fuel duct and was operated by a reach rod from outside the biological shield.
Design of this intermediate unit was halted in December when a new directive stated that the main production plant was to consist of four units, each 250,000 kw. These four plants, when operating at a 60% capacity factor, would provide approximately 600 grams of plutonium daily. This directive initiated a new round of design meetings between Du Pont and UofC. The new design from these meetings differed only slightly from the initial design. It was still to be a helium cooled unit.
Design of the helium cooling system was proving to be one of the most perplexing problems in the development of the new, higher power unit. This issue ceased to be a problem on January 9, 1943 when UofC issued a report (CE-407) describing a water cooled unit capable of operating at 500,000 kw. This unit was to have long aluminum clad uranium rods in thin walled aluminum tubes with water flowing through the annulus. Both horizontal and vertical orientations were considered. On February 16, design work on the helium cooled unit was discontinued.
During the next six months work continued on various facets of the design. Lattice spacing for the fuel channels ranged from 11 inches, to 8 inches and finally 8 3/8 inches. In August, there was enough information to proceed with the final design. On August 27, 1943 official construction of the first large scale unit began.
The original vision for Hanford was for the construction of four, 250,000 kw reactors; however, this was later reduced to three. Because of the hazards, each pile was to be located in a separate one-square-mile operating area (designated "100 areas") situated at six mile intervals along the Columbia River. This wide spacing allowed for additional reactors to be sited and still maintain three mile separation. The three initial reactors were designated B, D and F.
Startup and Operation of the Reactors
A large contingent of physicists provided support during the startup and operation of the reactors. All were members of the 100 Area Technical Group under the direction of J.B. Miles. Most had previous experience at the Metallurgical Laboratory or at the Clinton Laboratories. Among them was one woman, Leona Marshall, whose husband, John, was also one of the physicists.
B-Reactor
Physics support at B-Reactor (also known as Olive) was lead by G.L. Weil. There were two physicists plus five technical assistants on each shift. The physicists were W.R. Kanne, Leona Marshall, H. Newson, D. Hughes, John Marshall and W.E. Jordan. Enrico Fermi (also known as Mr. Farmer) was present during the startup and personally analyzed much of the startup data. Fermi and J.A. Wheeler were responsible for the rapid analysis of the Xe-135 poison problem encountered shortly after power operation began.
Initial loading of B-Reactor began at 5:44 pm on September 13, 1944. The fuel was loaded in layers, 22 columns wide starting at the midplane. Successive rows were added above and below the midplane. During the loading process, indium foil measurements were made in the central shim rod hole and proportional counter measurements were made in an empty central fuel channel to provide updated predictions of the critical loading.
No water was present during the initial loading (all the Hanford reactors, with the exception of N-Reactor, were more reactive without the cooling water). The reactor achieved dry critical with 400 tubes charged at 2:30 pm on September 15, 1944. Additional tubes were charged to give it sufficient excess reactivity to perform experiments. A 162 second period was achieved with 404 tubes loaded and 418 tubes gave a 35 second period.
After completion of the experiments, water was admitted to the pile and refueling resumed. The pile achieved a 10 minute period with 838 tubes charged in a roughly cylindrical pattern and a 31.5 second period achieved with 903 tubes. Two tubes were discharged due to flow problems and power operation began at 10:48 pm on September 26, 1944. Power was raised to 9 Mw at 1:40 pm on September 27. Approximately 6 hours later, the reactor began to shut itself down due to a continuing loss of reactivity.
During the following days, calculations were performed to assess all possible reactivity effects and to identify the cause of the shutdown. Mainly through the work of Fermi and Wheeler, the culprit was identified as the fission product Xe-135. It was clear that additional reactivity would be required to achieve power operation.
Armed with a new understanding of fission product poisoning, calculations were made to predict the maximum power possible at various loadings. The reactor was loaded in stages and operated to test the accuracy of the predictions. The following table summarizes the results.
-
- *B-Reactor could only achieve 220 Mw at the beginning of life due to limited reactivity even at full loading. The other reactors had higher purity graphite and uranium, and could achieve design power with 2004 tubes loaded. In B-Reactor's case, design power was only possible after a buildup of Pu-239 provided the required amount of additional reactivity.
On November 30, operation was resumed with 1595 tubes charged. Fifteen hundred tubes contained 32 fuel slugs and 95 tubes contained 35 slugs. The reactor was operated at 125 Mw until December 21, at which time the pile was shutdown to discharge approximately 6 tons of 45 MWD/Ton fuel. During this shutdown, the remaining tubes were charged to give a full reactor load of 2004 tubes.
When operation resumed on December 29, power was raised until 215 Mw was achieved on January 15, 1945. On January 18 the reactor was shutdown to discharge 6.6 tons of 70 MWD/Ton metal. When operation was resumed on January 19, power was further raised until the design power 250 Mw was achieved on February 4, 1945.
B-Reactor operated at a nominal power of 250 Mw until August 25, 1945, when the nominal power was reduced to 225 Mw. On March 14, 1946, the reactor began shutting down to be put in standby. The decision to put B-Reactor in standby was caused by the concern for damage to the reactor due to expansion of the graphite moderator. The reactor was restarted in July 1948.
As expected, the initial operation of B-Reactor had its problems; however, by March of 1945, just five months after startup, reactor operation had become relatively routine. The reactor power level had reached its design of 250 MW and fuel was being discharged with a nominal exposure of 200 MWD/Ton in batches of approximately 10 Tons. The length of time required for a discharge was approximately 20 hours early in 1945, but was reduced to approximately 16 hours by May 1945. There were, of course, variations due to unforeseen problems or additional maintenance.
D-Reactor
Physics support for D-Reactor (also known as Pine) startup was lead by John Marshall with assistance from Leona Marshall. Weil, Hughes and Newson provided rotating shift coverage with the aid of 3 assistants. Enrico Fermi was also present during the startup.
Dry loading of D-Reactor began on December 5, 1944 in almost the same manner as B-Reactor. Improvements in the proportional counters had been made, allowing more reliance to be placed on them than on the cumbersome indium foils.
D-Reactor was fully loaded with 2004 fuel channels (35 slugs per channel) before water was admitted to the pile. This allowed for testing of the control rods ability to provided shutdown capability under "adverse conditions". The results of this test showed the vertical safety rods (VSR), by themselves, were adequate to hold the dry pile subcritical. Twenty three of the 29 VSRs were found to be adequate to hold the pile subcritical.
Water was admitted to the pile and further rod tests revealed that the horizontal rods alone were not adequate to hold the wet pile subcritical. The insertion of one central VSR was required. Since it was not possible to operate the reactor at power with an uncooled VSR in place, an equivalent amount of poisoning was achieved by loading four fuel channels with lead-cadmium slugs. Power operation began on December 17, 1944.
The escalation of power at D-Reactor proceeded rapidly. By the end of December, power was at 185 Mw and design power of 250 MW was achieved on February 11, 1945. On October 23, 1947 nominal power was raised to 275 Mw.
Fuel charge/discharge was virtually routine from the start with the exception of a single tube discharge of 34 MWD/Ton metal on January 20. The next discharge did not occur until March 16 when a 10 ton batch of 107 MWD/Ton metal was discharged. The exposure of each subsequent batch was increased until a maximum exposure batch of 250 MWD/Ton fuel was discharged on August 6. Batch exposures were reduced successively, down to the nominal 200 MWD/Ton in September 1945. There were special tests that varied from 200 MWD/Ton, but, for the most part, all batch exposures were 200 MWD/Ton.
F-Reactor
The startup of F-Reactor (also known as Maple) was practically a repeat of D-Reactor. Physics support was reduced at F-Reactor, and at no time after initial startup was there continuous shift coverage. Weil had the lead with the assistance of Paul Gast. During startup, shift coverage was supplied by D. Hall, G.D. Monk, and T. Jacot.
Dry loading began on February 14, 1945 with dry critical achieved on the next evening. The pile was fully charged with 2004, 35 slug columns on February 19. The same tests were performed as at D to determine the ability of the rod system to hold the dry reactor subcritical.
As expected, the F-Reactor pile was found to be more reactive than D pile due to improved graphite and uranium purity. Twenty five VSRs were required to hold the dry pile subcritical as opposed to 23 VSRs in D pile. The wet pile was observed to have approximately 75 ih (1 inhour = 0.0024 %dk/k) more reactivity than the wet D-Pile. This excess reactivity required that a total of 10 poison columns had to be charged for startup. Shortly after startup, the reactor was shutdown to discharge three of the poison columns and reload them with fuel.
F-Reactor reached its design power of 250 MW on March 8, 1945 just 12 days after initial startup. On November 7, 1947 nominal power was raised to 275 Mw.
F-Reactor operation was routine from the start. The first fuel discharge did not occur until July 20 when 1.2 tons of 204 MWD/Ton metal was discharge. As in the other two reactors, fuel was discharged at a nominal exposure of 200 MWD/Ton in approximately 10 ton batches. The high and low exposure batches were usually special tests.
Thanks ... a fascinating episode in nuclear history ... great to see this being documented from the original sources.
ReplyDeleteAlso liked your blog on the J1709 monitor, which I'm aiming to replicate ... any further experiences on this you could share on the blog?