One last point about inventory differences is that they tend to grow over time, i.e., the cumulative ID might go up and down in the short term due to statistical variations or changes in plant operations, but over time the cumulative ID trends upwards due to the holdup category of ID or to a continuing unidentified loss of material from the facility. However, if a facility operated well and if there were no theft, then when the facility is decommissioned and the material in holdup returns to inventory, the cumulative ID should trend back to zero, or near zero. The importance of this last point will become clear as this NUMEC account unfolds.
***
In mid 1965, the AEC set about to find the missing HEU identified by the Oak Ridge team. On July 21, 1965, Assistant General Manager Howard Brown and others from AEC met with NUMEC officials in Apollo. Shapiro and his staff argued that buried wastes and air filters from the uranium plant contained enough uranium to more than make up for prior losses and “result in the Commission owing them money.” They asked the AEC representatives to accept their word for the missing material and not press for its recovery in order to avoid embarrassing the company in its ongoing expansion efforts. The AEC staffers responded that they “were nonplussed by the disclosure of information which had been known [by NUMEC] for some time.” They directed NUMEC to get its house in order and propose a financial settlement.149
Charles Keller attended that July 21, 1965 meeting with NUMEC in Apollo. His notes from that meeting, based on statements by S. A. Weber of NUMEC, show that between October 1, 1959 and September 30, 1964, NUMEC received 3,982 kilograms of HEU. Although Keller’s notes did not say so, it is safe to assume Weber reported the amounts in kilograms of U-235 in the form of HEU because the receipts covered a variety of enrichments in the HEU category.
Keller also kept notes of an August 10, 1965 meeting wherein Shapiro told the AEC commissioners that NUMEC processed 5,068 kilograms of HEU with enrichment greater than 75 percent during the years 1961 to 1965. Again, this amount would have been expressed as kilograms of U-235 because of the variable enrichments. Adding Weber’s FY 1960 tally to Shapiro’s total for 1961 to 1965 brings the total from start of HEU processing in October 1959 through December 31, 1965 to about 5.5 metric tons of U-235 contained in HEU with greater than 75 percent enrichment.150
***
The commissioners summoned Shapiro and Forscher to the August 10 meeting at AEC Headquarters in Germantown, Maryland to make certain that NUMEC’s most senior managers understood the gravity of the situation. At the meeting, NUMEC supplemented its earlier supposition that the missing HEU was contained in buried wastes by stating that some of the missing material was also held up in the scrap associated with its uranium carbide manufacturing process. The AEC staff tested that hypothesis by directing Oak Ridge to perform radiochemical analysis of samples of the scrap. That analysis failed to confirm NUMEC’s estimate or to account for the missing material. In November, because of Oak Ridge’s findings, AEC sent its own material accounting experts to Apollo. They supervised another plant-wide inventory supported by independent laboratory studies to try to account for the lost uranium.
AEC Chairman Seaborg was involved directly in AEC’s investigation of the missing uranium. He did not delegate his decision authority to others. For example, in a meeting between the AEC commissioners and their staff on August 9, 1965, he recommended billing NUMEC for $650,000 based on the first audit by Oak Ridge. Later information proved that amount insufficient to pay for the missing uranium, but the fact that Seaborg could make a dollar estimate shows that he knew the details of the case.
In the August 10 meeting between the AEC commissioners and the two senior NUMEC officials, Seaborg told Shapiro and the others that he had “extreme concern” about the missing material. The switch in emphasis from money to missing material reflected the pressure Seaborg was under from the Joint Committee, which was more concerned with where the material went than who was going to pay for it. Seaborg went on to show a special interest in helping Shapiro explain away the missing uranium. “The Chairman said it certainly seemed desirable to give consideration to an investigation of the matter. He asked if NUMEC felt that the 52 kilograms of material under discussion might already have been delivered to the Commission in the NERVA element batches. Mr. Shapiro replied that it was possible.”151
Shapiro told the commissioners that the Astronuclear contract was NUMEC’s first production experience with uranium carbide and its first large contract. He apparently forgot Shippingport, which was a much larger contract. He explained that uranium carbide is pyrophoric; i.e., it reacts with oxygen in air or water vapor and burns. Because of the pyrophoricity, special precautions were taken inside the glove boxes where the work was performed. His lengthy explanation emphasized that no one had ever produced reactor fuel of this type.
Shapiro also told the commissioners that the Astronuclear project didn’t go as well as he had expected based on previous work at laboratory scale, and the work dragged on longer than he expected. He said he undertook an investigation after learning that losses during processing exceeded his estimates (i.e., the yield was lower than expected). He found that the high amount of recycle in the process had tied up his scrap recovery facility; thus, the plant could not achieve its normal scrap recovery rate. Shapiro then provided a complex description of the chemistry of scrap recovery. He also said that his only mistake was hiring the wrong people to account for his nuclear materials, and he claimed that NUMEC had been honest in its dealings. He noted that in earlier years NUMEC paid AEC up to $1 million in fines (cost reimbursements) for other losses of enriched uranium. He concluded his presentation by saying NUMEC was responsible for finding the missing material.
One of the details in Shapiro’s presentation was the fact that the Astronuclear contract required his technicians to use a plasma torch inside a sealed glove box to make spherical particles of uranium carbide. During production, the particles acquired an electrostatic charge and stuck to the inner surfaces of the glove box, blocking the windows through which the technicians viewed and controlled the torch. Technicians had to wipe the inner surface of the windows with wood-based fiber towels known as Kimwipes. NUMEC disposed of the Kimwipes as low-level radioactive waste because they were contaminated with uranium-carbide dust. Shapiro told AEC he thought the Kimwipes must contain much of the missing HEU. The AEC told him to prove it by exhuming the low-level radioactive wastes that NUMEC had buried in its nearby disposal site.
On August 26, 1965, Keller sent a telex to Douglas George, AEC’s Director of Nuclear Materials Management, and others at AEC headquarters to express his concerns about Shapiro’s proposal.152
There is a serious question in my mind that the buried materials will contain the quantities postulated by NUMEC. . . . NUMEC has been less than candid in their dealings with us on nuclear materials management. Their internal controls and records on uranium are of such a nature that I will not repeat not state that they have satisfactorily accounted for material entrusted to them and that there is no cause for concern. It is my concern that prompted my statement at the meeting that if it were in my province to do so I would, as a prudent business man, stop all further deliveries of enriched uranium to NUMEC. . . .
Keller was not the only AEC official worried about the implications of the missing uranium. A September 2, 1965 memo from Lawton Geiger in AEC’s Pittsburgh Naval Reactors Office to AEC’s Oak Ridge Field Office indicated that officials in the naval reactors program were concerned about NUMEC. “In view of a NUMEC declared loss of 20 kilograms under a recent Bettis purchase order, the survey results (i.e., the Oak Ridge Survey of NUMEC covering the period July 1, 1963 – April 30, 1965) are of particular concern to me. The purpose of this memorandum is therefore to advise you of the circumstances of this loss and to solicit any additional information and comments you may have pertaining to this loss.” NUMEC had neither returned the Navy’s HEU nor reported discards of unrecoverable materials. NUMEC only provided a certifica
tion that it did not possess the missing material.153 Finding HEU in Israel that could only have come from the naval reactors program would later be the basis for CIA’s conclusion that it came from NUMEC.
Keller’s concerns were confirmed a few weeks later when NUMEC’s exhumation of the burial pits produced only 5.5 kilograms of U-235.154 NUMEC eventually revised its estimate of the U-235 recovered from the pits to 7.4 kilograms.155 The burial pits were in a waste disposal site approved by the AEC and called the shallow land disposal area. It was located a few miles north of Apollo, near NUMEC’s Parks Township facility. The recovered uranium was added back into the Apollo plant inventory and did not figure in the AEC’s final tabulation of inventory difference.
About the same time as the exhumation of the burial pits, Shapiro fired the radiation protection manager who approved the shipments of Kimwipes to the pits. He apparently took this action to show AEC that he was a strict enforcer of the rules or to deflect attention from the missing uranium. Ironically, he fired the person whose judgment had just been vindicated; i.e., there was not much HEU in the burial pits.
***
Years later, Keller recounted his assessment of the situation in 1965 in an interview with the FBI.156
His gut feeling is that NUMEC probably lost a major part of the material through mishandling and sloppy operations. He said that he could not swear to this if he were asked where the material was but his opinion was that he just did not feel that they were smart enough to divert the material. Additionally, to divert the enriched uranium it would have to have been taken out of the plant and NUMEC had a guard force who were familiar with procedures for securing special nuclear material. Mr. Keller felt that a great deal of collusion would have been required to remove 50 kilograms of enriched uranium. It would also be difficult to ship this amount of material to another company with forged documents because this would require collusion with someone in another plant, which would be even more difficult.
Keller overestimated the difficulty of removing the uranium. As NRC would conclude later, it would have been relatively easy to remove the material from the Apollo plant. The volume of uranium is relatively small because its density is high, so its theft would be easy to accomplish and to disguise. The density of uranium is 19 grams per cubic centimeter. Thus, for the purpose of illustration, 100 kilograms of uranium has a volume of about five quarts. For criticality protection, smaller containers would actually be used. For example, 10 containers of one-half-quart each, weighing about 10 kilograms apiece, would suffice to transport 100 kilograms out of the building. Such containers would have to be separated to prevent a criticality, but bare hands would suffice to guarantee the safety of the worker. Carrying one container in each hand, one person could remove 100 kilograms of HEU from the building in five trips, or five people could do it in one.
After opining on peripheral matters, Keller summarized for the FBI his assessment of the NUMEC situation in 1965.
He said essentially the problem in a nutshell is that the material was not there that the books said should have been there but there is absolutely no way to say how or where it went. His opinion is that sloppy plant operations, lack of records, and improper sampling probably [were] the reason for the loss. He indicated, however, that if he [were] planning to steal nuclear material he would use exactly this kind of operation, i.e., sloppy handling and accounting procedures.
About the same time as his FBI interview, Keller commented on an NRC review of the losses at NUMEC. “I found only one statement to which I might take exception and that [is] where it is stated that the AEC concluded that the ‘loss’ was not real. There were a number of us in the AEC who did consider the loss to be real but who were frustrated in our attempts to determine where or how the loss did in fact occur.”157 Keller was referring to the unexplained part of the inventory difference discovered in 1965. Reading between the lines, it appears there was more going on among the participants in the Apollo investigation than official records indicate. Either the embarrassing circumstances pressured the AEC hierarchy to end the investigation or someone pressured Keller to curtail his work.
***
In December 1965, Shapiro offered another explanation for the missing uranium. After failing to find it in the waste burial pits, he said it only appeared to be missing because NUMEC had consistently underestimated its losses of uranium in the scrap recovery process. Thus, more material had been returned to clients than they were due. This time, because the Joint Committee was pressuring the commission to conclude its investigation, AEC staff did not require Shapiro to prove his theory, as it had done in ordering exhumation of the burial pits and radiochemical analysis of the filters. Perhaps the staff was reluctant to challenge this latest assertion because the chairman had suggested it.
Instead, the AEC decided to let Shapiro pay for the missing HEU. Thus, in late January 1966, when AEC General Manager Robert Hollingsworth wrote to John Conway, executive director of the Joint Committee, to update him on the AEC’s investigation, he told Conway that NUMEC owed $1,134,849.34 to AEC to pay for the missing HEU. Hollingsworth noted that $500,000 of that amount had already been paid.158
Hollingsworth went on to summarize the various ways that NUMEC and AEC had tried to find the missing uranium, including exhuming waste burial pits and shipping a backlog of contaminated filters to Oak Ridge for analysis. Without saying why, he accepted Shapiro’s latest theory that NUMEC had underestimated losses on a contract-by-contract basis. He described the relationship between the HEU missing on the Astronuclear contract and the total amount missing since the beginning of operations. He blamed inadequate material accounting procedures for the large amount of missing uranium. He concluded, “No evidence has been developed that would suggest that the losses occurred under circumstances that would indicate possible diversion.”
On February 14, 1966, after testing its conclusions with the Joint Committee, the AEC senior staff, led by Assistant General Manager Howard Brown, briefed the AEC commissioners on the outcome of their investigation of NUMEC.159 The staff told the commissioners how the Astronuclear loss compared to the cumulative eight-year loss of U-235 from Apollo.
During this period NUMEC had reported the loss of 149 kg and had made appropriate financial restitution for much of it. Following the April 1965 surveys, AEC staff had determined that NUMEC had lost an additional 29 kg over the 8-year period, which had not been reported, and possibly not realized by NUMEC. . . .
Mr. Brown stressed there was no way specifically to relate the losses ascribed to the WANL [Astronuclear] contract to the total plant figures over the period. He reviewed the WANL data for the Commissioners noting the company had been billed $1.1348 [million] for the 61 kg of unaccounted-for material plus the 32 kg in inventory. He reiterated that although the 61-kilogram loss under the WANL contract was part of the total plant loss of 178 kg [since the beginning of operations], it was impossible precisely to establish the relationship between the two sets of data.
The AEC estimated when the losses occurred that eventually added up to the cumulative inventory difference of 178 kilograms, as shown in the following table, which attributes losses to fiscal years before the AEC’s inventory ended in November 1965.160 The row in the table for 1966 is for the first month of fiscal year 1966, i.e., October 1965. These estimates show that more than half of the 178 kilograms that were missing by late 1965 had gone missing before the end of 1963. Although discussions between AEC and Shapiro in 1965 centered on the complexity and difficulty of the Astronuclear contract, earlier contracts, which were not claimed to be novel or difficult, including naval reactor fuel contracts, were the source of a sizeable portion of the HEU that was missing by late 1965. The units in the table are kilograms of U-235 in the form of HEU. In 1977, ERDA issued a report on the history of IDs for strategic quantities of SNM, including AEC-licensed facilities such as Apollo.161 The results ERDA presented for Apollo are somewhat different than those presented by the earlier estimate compiled by A
EC staffers James Lovett and Cal Solem. The ERDA results also show that a significant fraction of the total ID by the end of calendar year 1965 occurred by the end of 1964. This conclusion is confirmed by the January 25, 1966, Howard Brown memo to Chairman Seaborg, saying in part, “The WANL contract to which the losses are charged was dated September 14, 1962, and delivery of product was made on October 31, 1964.”162 The significant IDs attributed in the table to 1965 and 1966 must have been corrections to prior estimates of inventory differences because NUMEC curtailed operations at Apollo while the Oak Ridge and AEC audits were underway in calendar year 1965.
***
Other items of interest in the AEC staff briefing of the Commissioners on February 14, 1966, as recorded by the Secretary of the Commission, were as follows:
[The NUMEC] case had convincingly demonstrated that fulfillment of a financial responsibility requirement might not really satisfy the AEC’s interest in special nuclear materials unaccounted for. Although the criterion remained valid in the sense that good SNM management was also good business, an accountability system based wholly on this criterion was, pari passu [at the same time], also based on a presumption of honesty. If the presumption of honesty were removed, the system did not present itself in the most credible light. . . .
[Mr. Brown, who did most of the talking] noted in this regard that if collusion between a shipper and a foreign government were assumed it would be theoretically possible to ship material abroad in excess of the amounts indicated in the company’s records. Because it was based upon a presumption of honesty and financial responsibility, the AEC materials accountability system might not reveal a deliberate and systematic attempt to divert materials in this manner. In addressing the system, however, Mr. Brown said it was important to bear in mind that the presumption of honesty was not a mindless assumption. Specifically, the Atomic Energy Act provided severe criminal penalties for violation of accountability procedures. The deterrent value of these penalties had been considered fundamental to the entire system of domestic safeguards. Analogously, the international safeguards system relied upon formal sovereign guarantees of foreign governments. . . .
Stealing the Atom Bomb: How Denial and Deception Armed Israel Page 9