The interior design of the uranium plant provided three kinds of spaces that were interspersed around the first and second floors, namely, laboratories, production areas and storage areas. Most operational personnel had access to the production areas, as did escorted visitors, if they could stand the smell of the ammonia, a principal ingredient in the chemical process that converted the incoming uranium hexafluoride to the outgoing uranium dioxide.
The company controlled access to the laboratory spaces in order to protect the sensitive laboratory equipment from routine worker traffic and from the environment of the production area. The laboratory equipment supported manufacturing processes, measured production effectiveness and controlled the quality of the final uranium products. NUMEC allowed only limited access to locked vaults where finished reactor fuels were stored before shipment to clients. Incoming and outgoing shipments occurred at the plant’s single loading dock.
The Apollo plant included chemistry hoods, glove boxes, exhaust fans and filters for dusty or gaseous stages of the processing line. These devices protected workers from inhalation and ingestion of uranium. The company used sealed containers for pellet and powder forms of uranium dioxide to protect workers from inhaling dust.
Uranium is hazardous to human health only if it is inhaled or ingested. Any that enters the body has the potential for both chemical and radiological toxicity. The most sensitive organs are kidneys and lungs. External exposure to uranium is not harmful because the alpha particles it emits travel only a few centimeters in air; obstacles as thin as skin or a sheet of paper are adequate to absorb alpha particles. The isotope U-235 emits an inconsequential amount of gamma radiation in addition to alpha particles. Bare hands are safe for handling well-packaged uranium. The radioactivity of HEU provides no impediment to its theft.
Highly enriched uranium is susceptible to an accidental criticality. That is, the collection of a relatively small volume of HEU in the right geometry (a critical mass) can lead to a spontaneous chain reaction, a so-called criticality accident. Criticality accidents produce lethal quantities of neutrons and other fission products that can injure or kill people in the immediate vicinity. They have occurred in the early stages of nearly every nuclear program in the world, including the United States. They typically generate enough heat to quickly disassemble the critical geometry. To preclude formation of a critical mass in its production processes for HEU, NUMEC collected and stored the material in prescribed arrays of small metallic containers. Design of the arrays of such containers prevented criticality accidents while processing, storing or transporting the HEU. Because such containers were small and sealed, it would have been convenient to pick them up and carry them off, one or several at a time. Operating experience in government facilities handling large quantities of enriched uranium and plutonium has shown that holdup of waste products in ductwork or other confined spaces can create an inadvertent criticality concern. Although NUMEC said holdup could explain some of the material that would eventually go missing at Apollo, no one ever said that such holdup was of sufficient size to create a criticality concern.
In December 1957, AEC modified the license for the plant at Apollo to allow NUMEC to possess and house up to 20 kilograms of U-235 contained in uranium of any enrichment. The AEC allowed only specific processes to be conducted with this uranium, namely, research and development related to fuels for nuclear reactors, recovering U-235 from scrap materials and manufacturing nuclear reactor fuels.
After it obtained this license, NUMEC began producing uranium dioxide at laboratory scale in small batches for experimental purposes and to develop its process design. In addition, soon after startup, in order to establish its business reputation and its cash flow, the company undertook difficult chemical operations to recover U-235 from scrap materials produced by government and other commercial facilities.138
The plant in Apollo began commercial operations in 1959, after Shapiro and Forscher succeeded in developing novel equipment for manufacturing nuclear reactor fuel. Their process converted uranium hexafluoride of any enrichment to uranium dioxide of that same enrichment. Their conversion process and equipment were “sensitive nuclear technology” because they were novel and because they allowed their owner to process uranium into fuel for burning in a nuclear reactor to produce plutonium. Foreign dissemination of such information required formal approval by the AEC and the U.S. Department of Commerce. Shapiro and Forscher also developed proprietary technology for commercial production of hafnium and zirconium, essential ingredients for nuclear reactor control rods and fuel cladding, respectively.
NUMEC’s uranium processing business grew rapidly as word spread about the quality of its product. As a result, the plant soon filled with piping, tanks, glove boxes and various complex machines for processing gaseous, liquid and solid uranium compounds. The Apollo plant could not produce metallic uranium, the form needed for an atom bomb, but the plant at Parks Township could. NUMEC’s income grew more than tenfold in its first seven years of operation, from less than $700,000 in its fiscal year 1957 to about $7 million dollars in its fiscal year 1965. In that same time, its long-term debt grew to nearly $5 million.139
***
In the early years of NUMEC’s operations, only the AEC could own uranium. Thus, when NUMEC processed uranium for a commercial client, the client had to first lease the uranium from the AEC. NUMEC had to keep track of the uranium it received and shipped, and it had to pay the client for any that it lost during processing. These payments were called fines, but they were really reimbursements for the cost of the uranium charged by the AEC.
Uranium was expensive, so there was a strong economic incentive to provide careful accounting of uranium inventories because commercial firms such as NUMEC had to account to their shareholders for profits and losses. A 1964 amendment of the Atomic Energy Act allowed SNM to be owned by private companies, including NUMEC and its commercial customers. The 1964 Act did not change the economic incentive for protection of SNM against theft.140
Enriched uranium hexafluoride came to Apollo in large steel cylinders from AEC’s enrichment plants at Oak Ridge, Tennessee and Portsmouth, Ohio. Natural uranium hexafluoride came to Apollo from the plants that processed ore before it was sent to the enrichment plants. The manufacturing processes at Apollo were essentially the same for uranium of any enrichment.
Some of the uranium dioxide was packaged and shipped to customers in powder form. For other clients, NUMEC applied heat and pressure to the powdered uranium dioxide to produce small, cylindrical, ceramic pellets in a process known as sintering. The processing and form of HEU fuels that NUMEC produced for naval reactors were classified.
Workers at the end of the production line for commercial fuel manually stacked ceramic pellets of low enriched uranium into metal tubes made of stainless steel or a zirconium-based metal alloy known as Zircaloy. Then they welded the tubes shut to produce finished fuel rods for commercial nuclear power plants.
In its first seven years of commercial operations, NUMEC undertook contracts for the production of reactor fuel of low and highly enriched uranium dioxide and contracts to manufacture HEU fuel for naval reactors. One of NUMEC’s larger contracts began in 1964 and involved the manufacture of fuel wafers of HEU (93 percent U-235) for the Shippingport Nuclear Power Plant.141 Bettis Laboratory and AEC’s Office of Naval Reactors built the Shippingport plant for Duquesne Power and Light Company in Pennsylvania. Shapiro developed the fuel design for Shippingport. NUMEC’s early contracts also included uranium fuel of various enrichments for research reactors in the U.S. and abroad.
The contract for the Shippingport fuel was unexpected because Shapiro and Forscher angered Admiral Hyman Rickover, the head of AEC’s Office of Naval Reactors, by their leaving Bettis in 1957 to form NUMEC. When they left, Rickover banned contracting with NUMEC in spite of Shapiro’s explanation that NUMEC could improve the manufacturing and the quality of naval reactor fuel. Rickover relented when Shapiro found a new way to produce
high quality hafnium for control rods in naval reactors. Soon, NUMEC was a major supplier of naval reactor fuel and the navy’s primary source of hafnium.
Despite their mutual dependence, the AEC’s Naval Reactors Office and NUMEC disagreed over safety, quality and cost issues for years. Furthermore, Naval Reactors demanded higher performance of NUMEC than did other NUMEC customers and the AEC regulatory staff. As journalist John Fialka reported, for example, “the losses at Apollo and other submarine fuel facilities [were] so embarrassing that Admiral Rickover [looked] for ways to prevent the release of MUF [material unaccounted for] numbers.”142, 10 Also, in September 1965, AEC Assistant General Manager Howard C. Brown, Jr. recorded Shapiro’s telephonic description of a spat with Rickover, “[Shapiro] said the situation vis-à-vis NUMEC and Admiral Rickover had worsened and he felt he was essentially being boycotted insofar as further Navy nuclear fuel work was concerned.”143
In 2006, the Department of Energy (DOE) declassified a report that listed by year the amount of HEU the AEC and its successors produced at the Portsmouth Gaseous Diffusion Plant. That report includes the 97.7 percent enriched uranium used in naval reactor fuel.144 According to that reference and others, Portsmouth was the only source of this level of uranium enrichment, and that uranium was used exclusively for U.S. naval reactor fuel.
The Portsmouth Gaseous Diffusion Plant was established in 1951, the last of three such plants in the U.S. Its initial mission was to produce HEU for use in nuclear weapons (typically, 93 percent enrichment). In the 1960s, it began serving the nuclear power industry with lower enrichments of uranium (typically, 3 percent enrichment). In 1962, Portsmouth began to produce the first HEU anywhere in the world with an enrichment of 97.7 percent. In 1964, it ceased producing HEU for nuclear weapons but continued to produce HEU of 97.7 percent enrichment. All of this very highly enriched uranium went to the naval nuclear propulsion program.
In addition to the HEU processed at Apollo for naval reactor fuel, NUMEC also processed HEU for other government and private clients, both foreign and domestic. The enrichment of the HEU for such contracts ranged from 20 percent to about 93 percent.
The inventory records that survive show that from 1959 through 1965, the Apollo plant processed a total of 14,693 kilograms of U-235 in the form of high and low enriched uranium. The records show that NUMEC processed about 5,500 kilograms of HEU at Apollo from 1959 through 1965, including fuel for the nuclear navy.
***
During these early years, AEC inspectors observed that NUMEC employed sloppy accounting methods that depended upon borrowing material from follow-on contracts to make up for material lost in preceding contracts. In September 1964 Charles Keller of AEC’s Oak Ridge Operations Office (OROO) wrote to Shapiro to say that OROO had attempted a routine physical inventory at Apollo and failed. Keller said, “Cross-over between different jobs has occurred ... definitely contrary to the provisions of Part 700 of the AEC Manual . . . we are required to inform you that your records are not in condition for audit.”145
Although contrary to good practice and the AEC Manual, crossover between contracts worked well for NUMEC so long as production efficiency (yield) on individual contracts was high and new contracts kept coming. However, by taking advantage of crossover among contracts, NUMEC was coming up short on a cumulative basis. The company reported losses of uranium on individual contracts and then paid the client for discrepancies between receipts and shipments. Such payments came from the profits of the company or from bank loans. Loans to a financially marginal startup enterprise became a source of concern in later investigations of NUMEC’s operations.
***
By the mid 1960s, the uranium plant in Apollo was running on a routine basis. However, the measures in place to control and account for the uranium did not fully satisfy AEC inspectors. NUMEC employed both design features and administrative controls to contain uranium in process and in storage to protect workers and the public from inhaling or ingesting it. Criticality safety measures were in place; contracts were growing; and relations with AEC’s Office of Naval Reactors were stable but acrimonious. The AEC had approved NUMEC’s security measures but those measures were insubstantial. AEC’s regulatory staff monitored the security and the safety of the plant. Then the troubles began.
* * *
8 The ZOA was a member of the American Zionist Council, progenitor of the American Israel Public Affairs Committee (AIPAC), which is widely understood today to be the most influential organization affecting America’s relationship with Israel.
9 A June 13, 1968 internal FBI memorandum identified Falk as a multimillionaire who served 10 other organizations at that time as an officer or director.
10 In the early days of the nuclear program, the term MUF was used to denote differences between book inventory (what the accounting system said should be in a plant or an area of a plant) and physical inventory (what was actually there). Later the MUF term was replaced with the term inventory difference or ID. Both DOE and NRC use the ID terminology today.
Part Two: Missing, Not Lost
Chapter 4
Where Oh Where (1965)
In early 1965, AEC began to watch operations at Apollo more closely because of NUMEC’s mounting difficulty with uranium control and accountability. Before 1965, AEC thought the accounting problems owed to the plant’s novelty and complexity, but significant additional losses of HEU occurred in early 1965 on a contract with Westinghouse Astronuclear Laboratory. The Astronuclear contract began in 1962, ended in 1964 and accumulated losses much larger than the norm.
The Astronuclear contract presented particular difficulty by requiring the production of small, spherical, coated pellets of 93 percent enriched uranium carbide, something no one had attempted before. Even though NUMEC took the novelty of the project into account when it bid for the contract with Westinghouse, its losses of HEU during production were higher than expected.
The uranium carbide was to fuel a prototype nuclear rocket engine called NERVA that Westinghouse was developing for the AEC. The uranium carbide was more difficult to handle than the uranium dioxide normally produced at Apollo. At the start of the contract, Westinghouse provided NUMEC with 1012 kilograms of HEU, which was the total amount needed to produce the amount of uranium carbide that was specified in the contract. Westinghouse leased the HEU from AEC.146
NUMEC experienced very poor yield in working with the uranium carbide; i.e., less than half of the HEU that entered the manufacturing process came out as useable product, while the rest ended up as scrap. The HEU in the scrap was difficult to extract or assay because of its novel chemistry and physical form. In addition, NUMEC could not borrow from the next contract in line to make up the shortfall of uranium, as it had been doing with preceding contracts, because the AEC stopped production at Apollo when it found out about the missing uranium on the Astronuclear contract. Thus, there were no new projects in the plant to make up for uranium losses on the Astronuclear contract.
NUMEC was the first to detect and report a significant shortfall of uranium on the Astronuclear contract. When AEC learned of the problem, it ordered the company to cease operations and commissioned an independent inventory of all the uranium in the plant. The AEC’s Oak Ridge Operations Office began the inventory on April 30, 1965 and soon confirmed that there was a significant shortage.147 A one-page summary of the results of that inventory survives. It records the total throughput of HEU from July 1, 1959 to June 30, 1965. It shows that in that six-year period the plant processed 4,068 kilograms of U-235 contained in 93 percent enriched HEU and 643 kilograms of U-235 contained in 97.7 percent enriched HEU, with an inventory difference of 99 kilograms of U-235. The document does not indicate whether it was prepared by NUMEC or Oak Ridge.148
***
Inventory difference (ID) for a uranium processing facility like Apollo is defined as book inventory minus physical inventory minus removals, where removals include product shipped to clients, measured effluents and measured discard
s during a particular period. In equation form this definition is written as follows:
ID = Material on the books – material in the plant - removals
Typically, when a significant ID occurs, personnel at the facility stop what they are doing and analyze and explain the ID to the best of their ability. Clients and regulatory authorities oversee that process. This process occurs whenever inventory differences are outside statistical control limits. One motive for careful review of IDs is that the facility has to pay the client for the missing material no matter the cause. Typical sources of inventory differences are as follows:
ID = measurement error + holdup + sneak discharges + unknown
Measurement errors might occur due to a faulty instrument, statistical variation in instrument readings or human error in recording a reading from an instrument. Holdup refers to material that could stick to the pipes, structures or equipment in a plant and not be accessible for measurement. Sneak discharges are effluents and disposals that were not monitored and recorded when they occurred. For example, a filter in a glove box could have developed an undetected hole that allowed particles in a gaseous effluent to escape monitoring.
Every uranium-processing facility tries to avoid the “unknown” category of inventory difference because if no other explanation can be found for material that remains in that category, then the personnel at the facility and their overseers must consider whether the material has been stolen.
As noted earlier, the term “material unaccounted for” or MUF was used in the early days of the nuclear program in the United States. The term inventory difference has replaced the term MUF in today’s parlance, but in the transition between terms some people and facilities continued to use the old terminology. Sometimes, people used the two terms to mean different things, which were not in keeping with the definitions. Where misuse of these terms causes potential confusion in this history of material accounting at NUMEC, and its oversight by others, parenthetical statements are provided to help the reader understand what was being said.
Stealing the Atom Bomb: How Denial and Deception Armed Israel Page 8