The basic Commission position [with the Joint Committee] should be that AEC had no evidence or suspicion that diversion had occurred; neither could the Commission say unequivocally that the material had not been diverted. Staff did, though, have a theory to support its lack of suspicion. Specifically, staff had determined during its two surveys of NUMEC that the company had consistently underestimated its actual process losses. Additionally the difference between actual and estimated losses appeared to have been passed on from completed jobs to new jobs. . . .
[AEC’s material control expert, Dr. Sam McDowell,] agreed with Mr. Brown’s conclusion that NUMEC simply had never taken the time and trouble to develop methods adequate to determine the amounts of material being lost through fabricating processes. . . .
[At the time the material was lost from Apollo] there had been a real [AEC] desire to accelerate the development of peaceful uses of nuclear energy both at home and abroad. Particularly with regard to the latter, the “peaceful atom” had been seen as an important, even vital, element of U.S. foreign policy. And most significantly, proliferation of a military nuclear capability had not been a decisive consideration. . . .
There was further brief discussion of [interviews of NUMEC employees] during which Mr. Brown indicated he would work out a suitable procedure with Mr. Hollingsworth to conduct the interviews. . . .
Mr. Brown said a second procedure the Commission might want to consider would be to request NUMEC to allow the Commission to examine the Company’s confidential financial records. . . . An objection to this procedure was that the impact on NUMEC itself and the nuclear industry in general would, to say the least, be traumatic. . . .
Commissioner Palfrey suggested that Mr. Brown might informally suggest to Mr. Shapiro, President of NUMEC, that if the Company offered to make its financial records available to the AEC, the entire situation might be put in a more favorable light. The General Manager agreed. Mr. Brown said he would telephone Mr. Shapiro. . . .
Mr. Brown summarized staffs views on the NUMEC situation. The theory under which it appeared the losses could be accounted for made it, in Mr. Brown’s opinion, unnecessary to involve the FBI formally in the matter. . . .
There was also brief discussion of the possibility that a premature leak of the NUMEC situation could lead to sensational and probably inaccurate press reports. The Chairman suggested the desirability of a prepared statement for contingency use in the public authorization hearings [before the Joint Committee] should the NUMEC matter arise. The General Manager observed he had been informed by Mr. Conway that the staff of the Joint Committee had placed strict limitations on the correspondence regarding NUMEC.163
These minutes reveal several oddities in AEC’s handling of the missing HEU. First, the AEC did not discuss the elephant in the room, i.e., the record does not indicate that the commission and staff considered whether his allegiance to Israel and his ties to senior Israeli nuclear experts like Bergmann might have influenced Shapiro’s behavior. The AEC knew by this time that Bergmann and Shapiro were collaborating in a joint venture between NUMEC and the Israeli AEC. However, there is no indication in the record before 1965 that the AEC knew of Shapiro’s association with various Israeli intelligence officers. Perhaps the commission was embarrassed to admit that its safeguards system was highly vulnerable to anyone with a compelling motive to steal and could not conceive of such an unprecedented crime being committed on its watch. AEC Chairman Seaborg later said as much, “On the international level, what sense did it make to try to prove that the United States had failed to prevent the theft of enriched uranium from a facility under AEC supervision for use by Israel to make a bomb . . . [so] the commissioners eventually found a way to avoid confrontation altogether.”164
Second, Brown was unequivocal in his statement that Shapiro could have diverted HEU to Israel had he been inclined to do so, despite the fact that Brown was similarly unequivocal in judging that the FBI need not be involved.
Third, the reason AEC staff gave for its conclusion that the material had not been diverted actually proved the opposite; i.e., the material could have been diverted. If someone at NUMEC had been underestimating the losses intentionally, they could have been spiriting away a little HEU every now and then and either storing it in the vaults or taking it offsite. The company had to pay a nominal amount for losses that were not reconciled on each contract, but the HEU was priceless if it served to jump-start the Israeli nuclear weapons capability thereby affording Israel time to build its own uranium enrichment facility as an adjunct or backup to the plutonium it expected to produce at Dimona. Although this conclusion is opposite from the one reached by the AEC staff, it is based on the same facts.
Fourth, if Shapiro or someone else diverted HEU to Israel, they did not do so for money as the commission presumed.
Fifth, the minutes demonstrate the commissioners’ promotional attitude toward nuclear power: they were quick to deny public access to the unpleasant news that the staff’s explanation rested solely on the presumption of Shapiro’s honesty.
Finally, the commissioners and staff scorned the press and conspired to hide unclassified information about the missing HEU.
After the briefing, AEC’s general manager sent a confidential report to the Joint Committee to describe what the commission had learned.165 He said AEC’s inventory concluded that the total cumulative loss between plant startup in 1957 and October 1965 was 178 kilograms of U-235. The report continued,
NUMEC’s cumulative losses from time of plant start-up in 1957 have been higher than those determined by other companies having comparable operations [an enclosed table showed the losses at NUMEC were more than double the average losses for similar plants]. NUMEC underestimated its process losses. Adequate documentation of internal plant transfers was not maintained. Losses on individual contracts as they occurred were not established. As a result, this accumulation of unrecorded and unreported losses from prior contracts continued and became a recognized loss when the inventory was at a low level following completion of the WANL [Astronuclear] contract. It should be noted, however, that in their scrap recovery operation, as distinguished from their fabrication processes, the very nature of the operation results in the loss of contract identity.
The survey team concluded that a major contributing factor to these circumstances was that NUMEC management had not assigned the necessary caliber of full-time professional talent to the complex job of materials management.
The general manager did not tell the JCAE that his New York office had data showing that Shapiro violated contractual requirements by intentionally blurring the distinction between U-235 used on government contracts and U-235 used on commercial contracts. The New York office was concerned because NUMEC’s liability for losses was less on government contracts than on commercial contracts.166
The AEC told the JCAE that AEC staff and NUMEC management had agreed on how the company came to lose 84.2 kilograms of the 178 kilograms of U-235 in the form of HEU that were missing from Apollo. The following table lists the loss mechanisms identified by AEC staff and NUMEC for those 84.2 kilograms.
Loss Mechanism
Kilograms
Accidental fire in storage vault (1963)
3.0
Liquid waste discharged to river
58.0
Measured discards to burial pit
2.2
Discharges to air
14.0
Liquid waste in storage drums
2.0
Laundry, shoe covers, track-out
5.0
Subtotal
84.2
Unknown loss mechanism(s)
93.8
Total Inventory Difference (ID)
178
The AEC concluded that 93.8 kilograms of the 178 kilograms of U-235 could not be explained, i.e., it was due to unknown causes. If there are no other possible loss mechanisms, it is prudent to examine whether someone could have stolen material in the unknown cause cate
gory. Neither AEC nor NUMEC had incentives to underestimate the known loss mechanisms that were part of the ID; no one wanted a high number for the unknown cause category.
The AEC’s experts used NUMEC’s calculations based on plant data to estimate the known material losses at Apollo. When addressing the unknown causes, the AEC did not distinguish between HEU that might have been lost through unidentifiable mechanisms and HEU that might have been stolen; i.e., the AEC could not tell if the missing HEU was stolen or was missing in some other way that went undetected by the experts from AEC, Oak Ridge and NUMEC.
If other loss mechanisms were plausible, such as more HEU returned to NUMEC’s scrap-recovery clients or more HEU held up in cracks and crevices of the plant, AEC’s experts would have named the other mechanisms and estimated the amount of U-235 associated with them. The people involved were the world’s leading experts in uranium accountability. They labored for months trying to explain the missing 93.8 kilograms by any mechanism they could imagine other than theft. They failed.
***
In 1979, NRC experts reviewed the analyses performed under AEC auspices in 1965 and 1966. The NRC concluded, “Although the AEC survey team estimate for cumulative ID was probably the best that could be made under the circumstances, it is quite possible that the quantity of material loss which was not attributable to known loss mechanisms was, in fact, less than their 93.8 kilogram estimate.”167 However, NRC did not address whether NUMEC’s estimates of the loss mechanisms (e.g., air and water effluents) were biased high to account for measurement uncertainties and to minimize its liability for so much missing HEU. George Murphy of the Joint Committee staff thought this was the case. He said that the estimate for liquid discharges to the Kiski River could only have been achieved if the plant had “run seven days a week, 24 hours a day since before the Revolutionary War.”168 In addition, discards were exhumed and measured, not estimated, thus removing another significant source of error from NRC’s concern that known loss mechanisms may have been underestimated in 1965. The fact remains that AEC told Congress in 1965 that 93.8 kilograms of U-235 in the form of HEU were unaccounted for.
***
The AEC did not tell Congress that the amount could have been even larger. A more pessimistic contemporary estimate of the amount of missing U-235 did not make its way to the Joint Committee or to the commission. Douglas George, AEC’s Director of Nuclear Materials Management (i.e., AEC’s leading in-house expert on the matter), wrote a memorandum to Assistant General Manager Brown on February 9, 1966 to describe uncertainties in the conclusions of the AEC survey team. He said that NUMEC’s losses were higher than other plants doing similar work because NUMEC failed to maintain adequate documentation of internal plant transfers of uranium and to establish losses on an individual contract basis. He went on to say, “The 93.8 kilograms estimate of unaccounted for material assumes that NUMEC presented accurate data with regard to the amount of uranium received in conjunction with its scrap recovery operations.” He described how the AEC survey team had accepted NUMEC’s claim that its scrap recovery clients were consistently overestimating the amount of U-235 shipped to NUMEC by 10 percent. However, after 1965, the AEC began to monitor shipper-receiver differences and found that differences were typically 2 percent not 10 percent as claimed by NUMEC. George concluded, “What this means is that the 93.8 kilogram MUF [ID of unknown cause] at NUMEC could exceed 250 kilograms.”169 No document has been found to show that Brown responded to George or conveyed this uncertainty to the commissioners. In the remainder of this account, no allowance is made for George’s prognostication. If he was correct, the amount of U-235 in the form of HEU that could have gone to Israel was understated by subsequent investigations.
The Joint Committee held a hearing on the NUMEC affair after AEC completed its investigation. At the hearing, AEC staff testified that the 93.8 kilograms of missing U-235 had benign explanations, along the lines of the letter that Brown sent earlier to the Committee’s Executive Director.
***
It is a mystery how Dr. Zalman Shapiro, president and founder of NUMEC and a leading expert in uranium metallurgy, while striving to make a name for his company and a profit for his investors, could run his plant so carelessly and with so little concern for the value of the uranium he was losing. He could have hired any one of a number of qualified material controls and accounting professionals to keep his company out of the mess it was in. His actions cost him and NUMEC their prestige in the industry and a lot of money to reimburse clients for lost uranium. Why was he so careless?
In February 1966, AEC Chairman Seaborg described this conundrum in a letter to Chairman Chet Holifield of the Joint Committee. “It appears that the losses for which NUMEC is now paying under its financial responsibility requirement are primarily the result of inadequate attention by NUMEC to generally recognized materials management methods.”170 When Seaborg used the phrase “generally recognized” with Holifield, they both understood the words to refer to materials accounting methods that were routinely practiced at the time in both private and government nuclear facilities. All it would have taken was for Shapiro to hire one person that understood the generally recognized methods.
Seaborg’s admission begs a larger question. Did Shapiro mismanage those first nine years of operations at Apollo or were his actions intentional, designed to hide more calculated objectives? As Charles Keller, the leading expert from Oak Ridge in AEC’s investigation of Apollo, told the FBI, if he was planning to steal nuclear material he would use exactly this kind of operation, i.e., sloppy handling and accounting procedures.171
***
Based on the results of its investigation, AEC billed NUMEC $1,134,849.34 for the U-235 that had not been returned to Westinghouse under the Astronuclear contract. AEC reduced the payment to $764,000 by giving credit for Astronuclear material still in inventory at Apollo that would eventually be returned to Westinghouse.
The $764,000 charge was the value of the 61 kilograms of U-235 that were associated with the Astronuclear contract and were still missing. Thus, NUMEC paid $12,500 per kilogram. Any prospective nuclear weapons state in those days would have paid that much; it was a pittance.
In addition to the $764,000 payment, AEC staff told the commissioners that NUMEC had already paid earlier clients for “much of” the 149 kilograms of U-235 it had lost in contracts that were completed before the Astronuclear contract.172 These two amounts, 61 kilograms and 149 kilograms, add up to 210 kilograms. Thus, when AEC said “much of” the 149 kilograms, it must have meant 117 out of 149 kilograms since AEC staff told the commissioners in that same briefing that the total amount lost was 178 kilograms (i.e., 61 + 117).
Contemporary documents show that AEC required NUMEC to pay for all of the lost U-235 (the ID of 178 kilograms), not just the 93.8-kilogram portion of the ID for which there was no identified loss mechanism. Using the same unit price charged by AEC for the 61 kilograms lost on the Astronuclear contract and applying it to the 117 kilograms lost before that contract, shows that the total amount paid by NUMEC for the HEU that it lost through October 1965 was about $2.2 million. On September 30, 1966, the AEC press office distributed a list of approved questions and answers concerning NUMEC which said the “losses over many contracts have totaled approximately 178 kilograms of uranium-235 having a value of about two million dollars.”173
Shapiro confirmed that the total amount paid by NUMEC for missing HEU, up to and through the Astronuclear contract, was in the neighborhood of $2 million. He told the commissioners on August 10, 1965, “in earlier years NUMEC had paid AEC up to $1 million for material losses,” that is, losses that occurred before the Astronuclear contract.174 The following table summarizes the HEU inventory differences and NUMEC’s payments through 1965.
The $2.2 million in total charges was more than four times AEC’s 1967 budget for research and development on nuclear material safeguards. The $2.2 million exceeded the initial capital used to form NUMEC. The $2.2 million was nearl
y ten times the company’s net income in its fiscal year ending September 30, 1965. The $2.2 million was more than 50 percent of the company’s long-term debt of $4.3 million at the end of its fiscal year 1965, i.e., before the final $764,000 of the $2.2 million in fines for missing U-235 was levied.175 However, $2.2 million was a bargain for securing the material needed for a fledgling nation to build its first atom bombs.
***
There were at least two reasons for Israel to take a strong interest in obtaining HEU from Apollo in the early 1960s. First, the Israelis could not have been sure that the plutonium production complex at Dimona would work as designed and on their desired schedule. Thus, they would have wanted the uranium as a backup for use in uranium-fueled weapons in lieu of plutonium-fueled weapons.
Second, the HEU might have been used to boost the reactivity of the Dimona reactor. Higher reactivity would have been desirable for a variety of reasons. For example, in power reactors fueled by natural uranium and moderated by heavy water, such as the Canadian CANDU reactors, booster rods holding HEU are used to overcome the effect of “xenon poisoning” due to the buildup in the fuel of xenon gas following shutdowns. Xenon absorbs neutrons and acts an impediment to reactor restarts. Although this approach typically was not used in research or plutonium production reactors where timely restarts were not as important as in power reactors, it was an alternative for Dimona if needed to facilitate returns to power after unanticipated shutdowns. Another possible reason to use HEU booster rods in Dimona would have been to compensate for any neutron absorbers that were put into the core for irradiation purposes. For example, Vanunu said that in the 1970s and 1980s, Dimona produced tritium by neutron irradiation of lithium-6 targets in the reactor core.176 Such a reason for the use of HEU booster rods was confirmed by an early report by Atomic Energy of Canada, Ltd. It described the use of HEU booster rods to compensate for the loss of reactivity due to the insertion of experiment rods in the Canadian NRX reactor, which produced plutonium and medical isotopes.177 Consistent with the idea of using Apollo HEU in booster fuel to increase reactivity at Dimona, journalists and authors Raviv and Melman said the increased production capacity of Dimona was related to “The facility [being] fed by the extra uranium obtained from Zalman Shapiro’s NUMEC company in America. . . .”178, 11
Stealing the Atom Bomb: How Denial and Deception Armed Israel Page 10