Exporting Disaster

~ The Cost of Selling CANDU Reactors ~

Chapter 3

Argentina, India, Pakistan, Taiwan

written by David Martin
of Nuclear Awareness Project
for the Campaign for Nuclear Phaseout

November 1996

For a copy of the complete report ($10 + shipping) write
Campaign for Nuclear Phaseout
1 Nicholas Street   Suite 412
Ottawa Ontario
K1N 7B7 Canada

. . . back to [ Table of Contents ]

3. Past CANDU Sales

3.1. Argentina

. . . back to [ Table of Contents ~ Argentina ]

3.1.1. The Cordoba CANDU

In 1972, AECL submitted a bid to Argentina's Comision Nacional de Energia Atomica (CNEA), in partnership with the Italian company Italimpianti, for a 600 MW CANDU. Italimpianti handled marketing and the plant's conventional equipment, and AECL was responsible for the nuclear side. The bid was selected in early 1973, and despite a change of government with the election of Isabel Peron, the contract was signed in December 1973. The total estimated cost of the reactor was $420 million. Of this, about $270 million went to Italimpianti, and $150 million to AECL. Of the $150 million, $61 million was for Canadian goods and services, $56 million was for Argentinean goods and services, $30 million was for heavy water; and $3 million was for Canadian uranium (if Argentina decided to buy it). [267] Canada provided a $129.45 million concessionary loan from the Export Development Corporation, payable over 25 years, with repayment starting only when the reactor entered service. [268] A further $25 million loan was made by the EDC in 1978. The original contract also apparently had a 25% ceiling on inflation, and with the 1973 OPEC oil embargo and a period of high inflation, it became apparent by 1975 that AECL was heading for a $200 million loss.

Subsequent attempts to renegotiate the contract were interrupted in March 1976 by a bloody military coup led by General Videla. By May 1977, AECL had made provision for a loss of $130 million on the deal. [269] , [270] The Cordoba deal was not just complicated by inflation and underpricing - defective boilers costing $15 million were also supplied by Babcock & Wilcox Canada. [271] Repairs delayed the project for over a year.

The sale of the Cordoba reactor was also controversial because bribes were paid to secure the contract. As part of the deal, AECL and Italimpianti agreed to split an "agent fee", which was used as a bribe. In April 1974, AECL President Lorne Gray approved the deposit of $2.4 million in a Swiss bank account. [272] An Argentinean investigation in 1985 revealed that José Ber Galbard, then Argentine Minister of Economic Affairs, was the recipient of the $2.4 million, plus another $1.1 million in May 1974, and an additional $300,000 two years later. [273]

. . . back to [ Table of Contents ~ Argentina ]

3.2. India

. . . back to [ Table of Contents ~ India ]

3.2.1. CANDUs for India The CIRUS Research Reactor:

The first Canadian reactor export took place in 1956. It was a "research" and plutonium production reactor modelled on the 40 MW NRX (National Research X-metal or X-perimental) reactor that began operation at Chalk River in 1947. The NRX was a heavy water moderated reactor that was built to produce plutonium for the American nuclear weapons program. It was well known that heavy water moderation results in very efficient plutonium production.

The Indian reactor was part of an aid program organized under the Colombo Plan Administration. The total cost of the reactor was about $17 million, of which the Canadian government provided $9.5 million as foreign aid under the Colombo Plan. [274] The reactor was known as CIRUS (Canada-India-Reactor-United States). The "US" was added because the United States supplied the heavy water for the reactor. The reactor went critical in July 1960, and became infamous as the source of plutonium used by India to manufacture the nuclear bomb it exploded in May 1974. It was still in operation in 1996.

. . . back to [ Table of Contents ~ India ] The RAPP-1 Power Reactor:

After the start-up of the CIRUS reactor in 1960, AECL negotiated another reactor export to India. In 1963 an agreement was signed for export of a 200 MWe power reactor modelled on the Douglas Point reactor at the Bruce site in Ontario, even though Douglas Point was not in service until January 7, 1967. The Rajasthan Atomic Power Project (RAPP) was built at Rawatbata, 64 km southwest of Kota in the northwestern Indian state of Rajasthan.

The total estimated cost of RAPP-1   [ now known as RAPS-1, for "Rajasthan Atomic Power Station Number 1" ]   was $79 million, of which $35 million was to be spent in Canada, [275] financed by the Export Credit Insurance Corporation (ECIC - predecessor of the Export Development Corporation). The plant began operation in 1972. It should be noted that AECL also gave India, free of charge, the technical and design information for construction of the 200 MW reactor. [276]

A serious problem with cracking of the reactor's end shield led to the reactor being shut down for long periods, and the reactor was finally downrated to 100 MW. [277]

. . . back to [ Table of Contents ~ India ] The RAPP-2 Power Reactor:

In 1966, another agreement was signed by Canada and India for construction of a second 200 MW reactor RAPP-2   [ now known as RAPS-2, for "Rajasthan Atomic Power Station Number 2" ]   with some improvements, at the same site as RAPP-1. The ECIC provided $38.5 million financing for the project's Canadian services and equipment. [278] Perhaps because of concessionary financing terms, or because of the earlier provision of the CIRUS reactor, the contract was signed (like the first reactor) without international tenders.

The reactor was still under construction in 1974, when India exploded a nuclear bomb, and Canadian personnel stopped work on the project. The lack of Canadian support delayed start-up of RAPP-2 until 1981. Canadian government financing for RAPP-1 and RAPP-2 totalled at least $140 million, payable over fifteen years with about six years' grace at 6%. [279] The Indian Department of Atomic Energy (DAE) was involved with construction and commissioning of the two RAPP reactors, and also fabricated some fuel. Indian content in RAPP-1 was 55% and 75% of RAPP 2. [280]

Canada's nuclear non-proliferation safeguards were strengthened following India's 1974 explosion of a nuclear bomb fuelled with plutonium made in the CIRUS reactor. Following the explosion of the weapon, and the failure of subsequent negotiations, Canada ended nuclear assistance to India.

RAPP-2 (now known as RAPS-2 for Rajasthan Atomic Power Station) was shut down in August 1994, and RAPS-1 was shut down in February 1994, both for maintenance and for detailed examination of pressure tubes. [281] By May 1995, it was announced that RAPS-2 would be shut down for at least three years to conduct a full retubing, after hydrogen blisters and elongation were found in several pressure tubes. [282] It has been reported that RAPS-1 may restart in January 1997, following repairs to the pressure relief system, although a total retubing for that reactor is also apparently planned. [283]

At issue is the fact that pressure tubes in the RAPS reactors are made of Zircalloy-2, a zirconium-steel alloy used only at early CANDU reactors. This alloy is a problem because of its susceptibility to embrittlement and elongation. Embrittlement in 1983 of a pressure tube in Ontario Hydro's Pickering reactor-2 led to a metre-long rupture and a major loss of coolant accident. All of India's PHWRs prior to Kakrapar-2 have used Zircalloy tubes. Kakrapar-2 has used the zirconium-niobium alloy that has been used in later CANDU reactors. [284]

. . . back to [ Table of Contents ~ India ]

3.2.2. The Indian Nuclear Program Current Status

Nuclear power and nuclear R&D in India are under direct control of the Prime Minister through the Indian Atomic Energy Commission (AEC), below which is the Department of Atomic Energy (DAE). In an inappropriate linkage, the nuclear regulatory agency, the Atomic Energy Regulatory Board (AERB), also reports to the AEC. The DAE has a number of subdivisions and companies.

The Nuclear Power Corporation (NPC) was created in 1986, and given responsibility for design, construction and operation of nuclear power plants (with the exception of RAPS-1 [285] , which because of serious problems was left separate as a "non commercial" plant). The NPC was intended to operate as a private company, however, as one industry commentator said, because of "NPC's low level of profitability . . . investors could have earned more return on their investment in a standard deposit account". [286]

In addition to   [ the Canadian-supplied CANDUs ]   RAPS-1 and RAPS-2, India now has six Pressurised Heavy Water Reactors (PHWRs) [287] in operation, all 200 MW reactors modelled on the   [ Canadian-built CANDU ]   Douglas Point and  [ the Canadian-supplied ]   RAPS reactors. They are (in-service dates noted):

In addition, India has two 160 MWe Boiling Water Reactors (BWRs) known as Tarapur (TAPS) 1 & 2, that went in service in 1969. Other than RAPS 1 & 2, these are India's only foreign-supplied reactors. They were turn-key projects built by General Electric. [288] Four more 200 MW PHWRs - Kaiga-1 & 2, and Rajasthan-3 & 4 - are expected to start up in the 1998-99 fiscal year.

India has five research/production reactors:

India's PHWRs have suffered from a number of problems that have resulted in very poor performance. For example MAPS 1 & 2 (the Madras Atomic Power Station in the state of Tamil Nadu) were ordered in 1967 and 1971, and took 17 and 15 years to bring into service. Problems with their heat transport systems have resulted in their downrating to 175 MWe.

In a potentially disastrous accident, NAPS-1 was shut down in 1993 after turbine blades broke, igniting a hydrogen fire in the turbine coolant that destroyed the turbine hall, and blacked out the station. With safety systems out of action, reactor cooling was maintained only by diverting an emergency source. [290]

In a stinging rebuke for the Indian nuclear industry, the outgoing Chairperson of India's Atomic Energy Regulatory Board (AERB), A. Gopalakrishnan, stated,

"Many of our nuclear installations have aged with time and have serious problems which are characteristic of the early vintage designs. Our own efforts to find indigenous solutions to these problems are not well organised or focused. The country had an inherent capability to tackle these problems, but the Department of Atomic Energy (DAE) has uniquely failed in bringing together these national strengths and coordinating them for the benefit of the nuclear sector." [291]

Gopalakrishnan noted that the AERB is only a "subordinate body" of the DAE, although he had strived for independence. He also noted that safety concerns are increasing, with more than 130 safety problems at various sites that need urgent action. [292]

The Indian nuclear industry is subsidized by the state. The NPC alone received a subsidy of 9.7524 billion rupees (about US $282.678) for the 1996-97 fiscal year. However, this represented a cutback of about Rs 1.5 billion (about US $43 million) from the previous year. [293] Attempts by the NPC to attract investment have also remained unsuccessful, and despite artificially low electricity rates, State Electricity Boards have remained delinquent in their payments, being in arrears a total of Rs 2.5 billion in March 1996. [294]

. . . back to [ Table of Contents ~ India ] Fuel Cycle Activities (Plutonium et al.)

India is committed to several advanced fuel cycles, notably a thorium fuel cycle, and thorium fuel bundles have been produced and used in Kakrapar-1. The Advanced Fuel Fabrication Facility at Tarapur has also produced MOX fuel bundles   [ a mixture of plutonium and uranium oxides ]   for PHWRs and is working on bundles for BWRs. The Fast Breeder Test Reactor (FBTR) is located at the Indira Ghandi Centre for Atomic Research (IGCAR) on the MAPS site at Kalpakkam near Madras. Ten years after start-up in October 1985, the FBTR is still operating only at low power levels. Design work is also taking place on a 500 MWe fast breeder prototype.

March 1996 saw the start-up of India's third reprocessing facility at IGCAR (the other two are at Tarapur and Trombay). The BARC site at Trombay was built between 1961 and 1964, and has capacity of about 30 tonnes per year. It uses the Purex process with solvent extraction columns   [ to extract plutonium from spent nuclear fuel ]   , and reprocesses fuel from the Dhruva and CIRUS reactors. India's main reprocessing facility at Tarapur started up in 1969. Although it has a nominal capacity of 100 tonnes per year, it has not functioned at full capacity. It reprocesses the spent fuel from the Tarapur, Rajasthan and Madras reactors. [295]

. . . back to [ Table of Contents ~ India ]

3.2.3. India and the Bomb

On May 18, 1974, India exploded a 12 kiloton fission weapon at the Pokhran site in the Rajasthan desert. The land for the test had been acquired in 1966, when 200,000 people had been forcibly displaced. [296]

The weapon was made with plutonium from the CIRUS reactor. Canada had transferred the CIRUS to India on the condition that it be used for peaceful purposes, but did not require external inspections. Although Canada advised India that plutonium from the reactor could not be used for nuclear explosives, India claimed that the agreement did not exclude nuclear explosions "for peaceful purposes". Following the 1974 explosion, India stuck to its story that the test was a "peaceful nuclear explosion" (PNE). [297] Technically, Canada considers the agreement to be still in effect, although it is not clear if India does. For the last twenty years, India has denied having nuclear weapons, but it has admitted the capability of making them.

The Indian nuclear test was a crippling blow to the myth of the peaceful Canadian nuclear program. Canada had received many warnings that India was moving towards nuclear weapons production, but had done little about it. Pierre Trudeau had even visited India in 1971 to discuss the matter personally with Indira Ghandi. However, in the wake of the test, on May 22, 1974, then Secretary of State for Foreign Affairs Mitchell Sharp stated,

"The Canadian government has suspended shipments to India of nuclear equipment and material and has instructed the Atomic Energy of Canada Limited, pending clarification of the situation [it took two years to "clarify"], to suspend its cooperation with India regarding nuclear reactor projects and the more general technological exchange arrangements which it has with the Indian Energy Commission." [298]

In December of 1974, a new proliferation policy was partially revealed. It required a binding assurance that Canadian nuclear material, equipment and technology would not be used for a nuclear explosive device, and rejected the excuse of "peaceful nuclear explosions". The policy also prohibited in the "contamination clause" the use of Canadian-supplied technology for nuclear explosive devices. Thirdly, exporters of nuclear technology and materials had to go through the Department of Industry Trade and Commerce, and the Atomic Energy Control Board. [299] These three points were actually part of an 8-point program approved by Cabinet in December 1974, but not made public until 1975. The other safeguards included:

"...a binding recognition of Canada's right of prior consent over: transfers to third countries; reprocessing of Canadian-origin material; and enrichment of uranium beyond 20%. As well, Canada maintained the right to apply fall-back safeguards on reprocessing and enrichment should IAEA safeguards cease to be applied for any reason. Finally a binding commitment required the provision of adequate physical protection for Canadian-origin material." [300]

For two years after the Indian nuclear test, Canada continued with negotiations in order to convince India to accept the new safeguards. However, India remained obdurate, and on May 18, 1976, Canada formally terminated nuclear co-operation. The nominal reason was that India agreed to safeguard only the two Rajasthan reactors. [301]

In December 1976, the Canadian government announced a further upgrading of its non-proliferation policy when its decided that nuclear trade would be restricted to those countries which had signed the Non Proliferation Treaty. [302] The fundamental problem of course with paper safeguards, is that agreements can simply be ignored. This was clearly seen in the original transgression of India in the use of plutonium from CIRUS despite a bilateral agreement. The ineffectiveness of safeguards was also seen more recently in the case of Iraq, which developed a nuclear weapons program despite being a signatory of the Non-Proliferation Treaty.

Regardless of whether India has continued to use plutonium from the CIRUS reactor for nuclear weapons, it now has other sources which it can use. In 1985, the 100 MW Dhruva reactor went critical, and it is theoretically capable of making 25 kilograms of plutonium per year. It is usually estimated that 5 to 8 kg of plutonium is the minimum required for a basic fission weapon.

A 200 MW CANDU is capable of making 50 to 60 kg of weapons grade plutonium annually. [303] Indian now has six reactors of this size operating, in addition to the two RAPS reactors that operated sporadically since 1972 and 1981. All of these reactors were built using the CANDU technology that Canada provided. India's decision to opt for an advanced fuel cycle with a breeder reactor involves the reprocessing of fuel to extract plutonium, which provides the infrastructure and effective camouflage for military production and reprocessing.

. . . back to [ Table of Contents ~ India ]

3.2.4. Indian Politics and the Bomb

The initial motivation for the Indian nuclear weapons program was the first nuclear weapons test by China in 1964. However, in the 1970s, nuclear competition shifted to Pakistan. Hostilities run deep between Pakistan and India after the post-colonial partition, three wars, and an ongoing bitter conflict over the possession of Kashmir, aggravated by the Sikh separatist movement in India.

Recent events have made the escalation of nuclear tensions between India and Pakistan a distinct possibility. In 1995, it was widely reported, after satellite surveillance reports by US intelligence, that India was preparing for another nuclear test in Rajasthan. [304] It is very likely that if India tested another nuclear weapon, Pakistan would respond in kind.

Both India and Pakistan have refused to sign the Nuclear Non-Proliferation Treaty. India also blocked consensus for adoption of the Comprehensive Test Ban Treaty in Geneva in August 1996, meaning that it could not be forwarded to the United Nations General Assembly and approved with immediate legal force. While Pakistan voted for the Treaty at the General Assembly, all 44 nations with nuclear reactors (including India) must sign the treaty to bring it into legal force, and Pakistan has said that it will not sign unless India does. [305]

The nuclear bomb is popular in India. In the May 1996 national election, the Congress Party, dominant in India since independence, received a lower vote than ever before. The Hindu fundamentalist Bharatiya Janata Party (BJP) received the largest share of the vote. The BJP is a longtime and consistent supporter of an open nuclear weapons program, as well as the deployment of short and intermediate-range missile delivery systems. [306] Although the BJP only exercised power briefly before being replaced by a coalition government, the political situation is very unstable. Support for the BJP is a measure of the popularity of the 'Hindu bomb'.

Over the last 20 years, India has certainly had the opportunity to produce enough plutonium for several hundred nuclear weapons. [307] A test would allow it to try a so-called 'boosted' fission weapon, or even a thermonuclear (hydrogen) bomb. Nuclear weaponeers in Pakistan would like to see another Indian test, in order to justify bringing their nuclear program out of the closet. It remains to be seen if the CTBT agreement will exercise any constraint on India and Pakistan.

. . . back to [ Table of Contents ~ India ]

3.2.5. The Downgrading of Canadian Safeguards - India

In October 1996, Foreign Affairs spokesperson John Bell denied a report from India that Canada would end its 22 year boycott on nuclear trade with India. [308] However, Bell was less than honest about the real situation. Canada has been backing off from its boycott since 1989.

In the 1988 Annual Report of the CANDU Owners Group (COG), Chairperson Hugh Irvine stated that "...discussions were opened with the Canadian government on the possible participation of India and Pakistan in the COG Information Exchange Program." [309] In the following year's Annual Report, it was noted that "India and Pakistan joined this program [the Information Exchange Program] mid-year...". [310] There was no public consultation or parliamentary debate on the re-start of Indo-Canadian nuclear relations.

In June 1994, former AECL CANDU President Donald Lawson wrote to the International Atomic Energy Agency (IAEA) to draw their attention to safety problems at the RAPS reactors. At that time he said, "The conditions are such that there is a real potential for a ... pressure tube rupture to occur at any time". [311] AECL's motivation was certainly not altruistic, however. In 1993, S.K. Chatterjee, managing director of India's Nuclear Power Corporation, had invited western nuclear companies to submit "pre-qualification documents" for the retubing of India's PHWRs. [312] AECL would undoubtedly like to bid on these contracts, and has been pressuring the federal government to remove or loosen its restrictions on nuclear relations with India.

In October 1994, an Ontario Hydro Nuclear employee on secondment to the International Atomic Energy Agency in Vienna, Frank King, reported that he had arranged a meeting between "nine Canadians from AECL and Ontario Hydro and nine Indian experts" in July 1994 to discuss the safety of pressure tubes in the two RAPP reactors sold to them by Canada. King stated that he would be making his second trip to India in December 1994. [313] In February 1995, it was reported that "In the interest of safety, Canada has now reinstated limited contacts with India". [314]

In April 1995, the Canadian television station CTV reported that AECL staff visited India in March 1995 to discuss Canadian assistance on the retubing of the RAPP reactors, and have also been discussing the sale of CANDU reactors to India. [315] In fact, the Managing Director of the CANDU Owners Group (COG), Barry Collingwood, led a four-person group to India in March 1995 to meet with officials of the Nuclear Power Corporation. He stated, "We have reached informal agreement on a broad scope of possible cooperation on core assessment, in-service inspection, and safety related quality assurance associated with retubing." [316] The nuclear industry, led by COG, has clearly been pressing the federal government to downgrade its non-proliferation policy for both India and Pakistan. In the COG 1994-95 Annual Report, Collingwood stated,

"Another impediment to the flow of information between CANDU stations is the Canadian export policy restrictions currently being placed on two of our members; India (with nine operating CANDU stations) and Pakistan (with one operating station). We will continue to encourage the Canadian Government to review their export policy with the ultimate objective of reducing the current restrictions on cooperation among CANDU stations" [317]

In May 1995, Collingwood formally asked the Canadian government to allow the restart of commercial nuclear relations with India. [318] There has been no public response on this matter from the government. Clearly, however, without any public consultation or parliamentary debate, the government has quietly been eroding its non-proliferation restrictions for India and Pakistan over a period of years. This process began in 1989 by allowing the provision of nuclear technical information through the CANDU Owners Group. Information exchange and provision of various services may be a prelude to the re-introduction of full-scale commercial relations.

. . . back to [ Table of Contents ~ India ]

3.3. Pakistan

. . . back to [ Table of Contents ~ Pakistan ]

3.3.1. KANUPP

In 1964 an agreement was made between Canadian General Electric and Pakistan to build a 137 MWe CANDU reactor on the coast, 30 km west of Karachi. The reactor, known as the KANUPP (Karachi Nuclear Power Project), became operational in 1971. Aid was provided in the form of concessional financing through the Canadian International Development Agency (CIDA). [319] The total cost of KANUPP was $63 million, $51 million of which was financed by Canada, half as external aid at 3/4% interest over 40 years, with 10 years' grace; the other half at 6% over 15 years with 5 years' grace. [320] Aid was provided through the ECIC, the External Aid Organization (EAO) and CIDA. The Export Development Corporation (EDC) took over from the ECIC in 1969. The financing arrangements have been described as follows:

"Between 1966 and 1978 a total of $12.4 million was provided in export credits, and $29.4 million was loaned through the EAO/CIDA account. (Grants of about $1.5 million, which covered supervision, training and the financing of spare parts, were also extended by CIDA.) The EAO/CIDA loan was concessionary. Its terms included a 10-year period of grace followed by a 50-year repayment schedule with no interest charges." [321]

Following the Indian nuclear explosion in May 1974, Canada turned its attention to Pakistan as a proliferator. It is widely accepted today that Pakistan is a 'threshold' nuclear weapons state, and that it was on its way to being one even in 1976. Following the December 1974 upgrade of Canadian proliferation safeguards, negotiations ensued to convince Pakistan to accept the new standards. Pakistan argued that the retroactive application of safeguards was unfair. With Korean and Argentinean export sales in the offing, Canada ended nuclear cooperation with Pakistan on January 1, 1977, shortly after its December proliferation decision that nuclear trading partners with Canada must sign the NPT.

3.3.2. The Downgrading of Canadian Safeguards - Pakistan

As noted in the discussion above on downgrading of Canadian proliferation safeguards for India, discussions took place with Pakistan in 1988 about participation in the Information Exchange Program of the CANDU Owners Group. [322] The following year, "Pakistan, along with India joined the Information Exchange, ending 12 years of isolation from the Canadian nuclear industry". [323] There was no public consultation or parliamentary debate on this reversal of Canadian proliferation policy.

In 1989, Barry Collingwood, Manager of the CANDU Owners Group (COG) noted that "provision of fuel channel inspection services, and design reviews of critical safety components and systems" had been proposed for KANUPP. [324] The following year, Collingwood noted,

"A second major development in 1990 was the initiative taken by COG in providing assistance to the Pakistan Atomic Energy Commission in assessing the state of the KANUPP reactor. The assistance is provided under the auspices of an IAEA Steering Committee and with the approval of the Canadian Government. This initiative is consistent with the COG mandate of promoting co-operation and mutual assistance in the safe operation of CANDU reactors and is a positive step towards re establishing technical co-operation with Pakistan." [325]

In 1992, the Pakistan Atomic Energy Commission (PAEC) requested US $40 million from the Pakistani government to upgrade the reactor. [326] PAEC initiated negotiations with AECL, Ontario Hydro, New Brunswick Power and Babcock & Wilcox Canada for a contract on fuel channel assessment, boiler and balance of plant inspection, and a safety analysis. The CANDU Owners Croup served as a project manager for the Canadian companies. [327] A Canadian team visited KANUPP in 1989 to initiate the project. Problems identified with KANUPP included: bellows leaks, condenser tube leaks, heat exchanger tube leaks, sticking and rupture in the refuelling machine, problems with the shield plug, flow blockage in one channel, and problems with computer software and hardware. [328]

By 1993, based on the promise of Canadian aid, PAEC announced that the Pakistan government had approved a US $32 million package for the "Safe Operation of KANUPP" (SOK) project. Pakistan expects that the Canadian work will allow a ten year extension of the 30 year life expectancy of the KANUPP plant. Paul Barton of External Affairs tried to justify the project by separating "safety" work from "efficiency" work on KANUPP - i.e. arguing that safety improvements by Canadians are acceptable, from the point of view of Canada's proliferation policy, while efficiency improvements are not. Barton noted however, that "you cannot clearly delineate between operational efficiency and safety with a nuclear reactor". [329] Safety improvements by Canadian companies on KANUPP were taking place despite the fact that Canada has no Nuclear Cooperation Agreement with Pakistan, and the fact that Pakistan has not signed the Non-Proliferation Treaty. In March 1993, the Atomic Energy Control Board commented on a May 1992 IAEA report on KANUPP, and concluded that on safety grounds, continued operation of the KANUPP reactor could not be justified. [330]

Despite a plethora of other problems at KANUPP that have resulted in a lifetime load factor of only 25% [331] it was confirmed in 1994 after investigation by the Canadian team, that re-tubing of the reactor will not be necessary. It was reported in 1994 that permission was being sought to export various hardware items from Canada to Pakistan, including control computers, fuel channel bellows, spare valves, neutron power monitors and power regulation instruments. [332] AECL confirmed in 1994 that it is doing fuel channel maintenance work on the KANUPP reactor. [333] As noted above with regards to India, the Canadian nuclear industry has been pressing the federal government to lift its export restrictions for Pakistan, as well as India [334]

3.4. Taiwan

3.4.1. Taiwan Research Reactor (TRR)

In 1969, AECL sold a 40 MWt 'research' reactor to Taiwan. Like CIRUS it was based on the NRX. The TRR was a turnkey project with a total cost of $35 million, of which 80% or $28 million was sourced in Canada. [335] AECL clearly anticipated possible future sales to Taiwan, but relations were terminated when Canada recognized the People's Republic of China in 1970. In 1976, US intelligence sources confirmed that Taiwan had begun operation of a small clandestine plutonium reprocessing plant in 1975, used irradiated fuel from the TRR. Only after intense political pressure was the reprocessing facility dismantled in 1977. [336]

Despite Taiwan's attempt to make nuclear weapons, AECL complained bitterly that it was not allowed to subsequently bid on nuclear power reactors for Taiwan. Despite the lack of diplomatic relations, AECL officials met with the Taiwan Power Corp. in 1983, and were clearly lobbying the government for the right to do business in Taiwan. [337]

Taiwan apparently only shelved its program to make nuclear weapons temporarily. In 1988, it had begun work on another reprocessing plant. Upon discovery of this, the US government insisted that the TRR be shut down. The US government had earlier insisted that Taiwan ship all of the irradiated fuel from the TRR to the USA. [338] Canada had come close to helping yet another country to make a nuclear bomb.

4. Other CANDU Sales Attempts

4.1. Australia

In the late 1960s, the Australian Atomic Energy Commission (AAEC) requested proposals by June 15, 1970 for a proposed 500 MW reactor. However, before a final selection was made, Australia decided that it could not afford a nuclear power program. [339]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.2. Chile

AECL has proposed the sale of a CANDU-6 to Chile, and has entered into discussions for cooperation with three Chilean companies, Chilquinta (an electrical utility currently acting as Ontario Hydro's partner in a Peruvian venture); Chilgener, and the state corporation Colbun Machicura. In January 1995, an AECL agent, Ernesto Puig, met with Mr. Bobadilla, Executive Director of the Chilean commission of Nuclear Energy (CCHEN) to present the project. [340] The $1.7 billion reactor is proposed for the Antofagasta region, 1,400 km north of Santiago. [341] Jean Chrétien apparently raised the possible sale of a CANDU during his visit to Chile and meeting with Chilean President Eduardo Frei on January 24, 1995.

. . . back to [ Table of Contents ~ Sales Attempts ]

4.3. Egypt

In 1989, AECL announced that along with the American corporation Overseas Bechtel Inc., it had received a contract in Egypt for a "CANDU localization program" [342] - actually a reactor fuel manufacturing program. In 1993, AECL said that the Egyptian Energy Ministry and the Nuclear Plans Authority were continuing their "ongoing review of CANDU technology to determine its suitability...". [343] In 1994, AECL reported that sales efforts were continuing. [344]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.4. France

As in the case of the United Kingdom, AECL was unsuccessful in selling CANDU in France, in the face of a strong, publicly owned military and civilian French nuclear industry. Like the United Kingdom, however, in 1968 an agreement was signed with the French Commissariat l'Energie Atomique (CEA) transferring CANDU technology to them for a rumoured $1 million. [345] Another agreement in the same year transferred radioactive waste to France for plutonium extraction. [346]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.5. Greece

Following the withdrawal of Canadian General Electric from the reactor export market, Greece was one of AECL's first sales targets. However, AECL's 1969 bid was rejected in favour of a British bid, which reportedly was supported because of related (non-nuclear) trade deals. [347]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.6. Hungary

In 1988, AECL signed a Memorandum of Understanding with three Hungarian companies to study the marketing and construction of Slowpoke reactors in Eastern Europe. [348] In 1989, AECL and Ontario Hydro's New Business Ventures signed an agreement with the Hungarian state utility to study the feasibility of building a CANDU reactor in Hungary dedicated to the export of electricity to western Europe. [349] In 1991, AECL reported that

"A market feasibility review and analysis for Czechoslovakia and Hungary has been completed and presented to marketing management. The purpose of the review is to confirm preliminary assessments of the long-term opportunity for CANDU in those countries". [350]

In 1996, Canada supported a Hungarian radioactive waste program. In May 1996, it was announced that the federal government had already given Hungary $230,000 to "assist with drilling underground tunnels and to develop the disposal research program". The Canadian International Development Agency (CIDA) was providing another $500,000 to complete the project. Monies for the funding were committed in the 1996 federal budget. [351] David Iftody, MP Provencher stated, "Hungary needs Canada's know-how as well as capital to expand its nuclear power generation capacity and at the same time safely dispose of spent fuel". [352]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.7. Italy

In 1970 a contract was signed with the Comitato Nazionale per l'Energia Nucleare (CNEN) (an R&D body) and the Ente Nazionale per l'Energia Elettrica (ENEL, the national electric agency) for transfer of CANDU-BLW (Boiling Light Water reactor) technology. AECL never revealed how much it was paid for the reactor design, but it has been speculated that it was a complete or virtual gift, designed to prepare the way for a reactor bid. [353] Despite this incentive, and its connection with the Italian company Italimpianti (its partner in the Cordoba deal in Argentina), AECL failed to win an ENEL contract for four reactors in the early 1970s. The contract went to an Italian licensee of an American LWR.

AECL next decided to promote a licensing strategy itself, with a December 1976 agreement with Progettazione Meccaniche Nucleari (PMN). In a controversial move, for an undisclosed fee, the agreement gave AECL a consulting role for design of the nuclear system and the supply of heavy water, while PMN would design and contract for the balance of the plant. The agreement stipulated certain levels of Canadian content, but also gave PMN the right to export CANDU itself under certain conditions. [354] Despite AECL's willingness to give away CANDU technology, the strategy could not overcome the American competition and the effect of a weak Italian economy.

. . . back to [ Table of Contents ~ Sales Attempts ]

4.8. Japan

In September 1971, AECL completed a five year agreement with Power Reactor and Nuclear Fuel Development Corporation (PNC) of Japan, again for technology transfer. The Japanese were interested in the CANDU-BLW (Boiling Light Water reactor), although the abject failure of Gentilly-1 BLW undoubtedly reduced their interest. [355]

In 1976, AECL signed a $1.7 million contract with the government owned Japanese company, Electric Power Development Company (EPDC), for a study on the introduction of the CANDU to Japan by 1981. AECL proposed the construction of four CANDU reactors as an "intermediate step" between their existing light water reactors, and the fast breeder reactor that Japan was planning even at that time. In June 1979, Prime Minister Joe Clark met with Japanese Prime Minister Masayoshi Ohira, and later with his trade minister to promote a CANDU sale. [356] In order to sweeten the deal, Clark offered a 30-year guaranteed supply of uranium. [357] As part of the negotiation, the Japanese were also pressing for favourable terms in access to Alberta tar sands. [358] However, after an initial announcement in August, the Japanese Atomic Energy Commission turned down the AECL proposal for four CANDUs in October 1979 [359], supporting instead the construction of their own Advance Thermal Reactor (ATR) - also a heavy-water moderated reactor. [360]

In 1983, AECL maintained a full-time representative in Tokyo, [361] but the chances of a CANDU sale were negligible. The Japanese utilities are closely associated with the three Japanese nuclear suppliers whose reactor systems are each licensed by either General Electric or Westinghouse. [362]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.9. Mexico

In 1969, Mexico requested bids on a reactor to be located at Laguna Verde in the state of Veracruz. AECL lost out in the bidding to General Electric, but this was probably a blessing in disguise. When construction began in 1973, it was estimated that GE's Laguna Verde reactor would be completed in three years at a cost of $263 million [363] - when it finally began operation in 1988, it had escalated to a final cost of about $3.5 billion. [364]

This disastrous experience did not stop Mexico's Instituto Nacional de Energia Nuclear (INEN) from issuing a call for bids in October 1980 for a second station at Laguna Verde with a capacity of 2,300 MW. The contract was worth more than $2 billion. [365] It was to be part of a massive nuclear expansion program of 20,000 MWe by 2000, estimated to cost $32 billion (US). The same companies that had bid on Laguna Verde, also put in bids by February 1982.

Cheap financing and technology transfer, were as usual the key bargaining chips in the competition, but Canada's marketing effort, dubbed "CANDU Diplomacy" was truly stupendous. It included a free $2 million study performed by AECL looking at Canadian importation of Mexican oil. [366] As a official "bribe", in 1980 the Canadian-Mexican Agreement on Industrial and Energy Cooperation was signed under which Ottawa said it would buy Mexican crude oil if Mexico considered buying a CANDU. In 1983, a million dollars worth of Mexican crude oil was coming into Montréal every day, despite the fact that Mexico had killed the whole nuclear deal in 1982. [367]

Mexico pressured Canada to supply generous concessionary financing for 100% of the cost of the package, estimated at as much as $6 billion for 4 CANDU-6 reactors. The Canadian government was apparently willing to borrow billions of dollars at commercial rates (then about 16%) to finance the project for Mexico at 7% or 8% - in effect providing a subsidy for as much as 40% to 60% of the station. [368] It was also suggested that Mexico was pressuring for a "guaranteed cost agreement" putting the risk of cost overruns onto AECL. [369]

In January 1982, two weeks before the bid deadline, Pierre Trudeau visited Mexico for the third time in a year, along with MP Roy MacLaren. MacLaren was parliamentary assistant to Energy Minister Marc Lalonde, and was known at that time as Canada's "unofficial CANDU salesman". [370] A number of government agencies had also been involved in the bidding process, including the Export Development Corporation, the Atomic Energy Control Board, Ontario Hydro and various private corporations. AECL even established an office in Mexico City, and one source reported that $50 million (CDN) had been spent on the Canadian bid. [371] However, under the weight of falling oil prices, and its $70 billion debt, the Mexican peso had collapsed in February, 1982. Mexico announced the cancellation of its grandiose nuclear expansion plans in June 1982. [372]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.10. The Netherlands

In 1987, AECL noted that discussions were continuing with Netherlands, providing information for a technical evaluation of nuclear power. [373] At this time there was already an AECL office in The Hague. In 1989, Dutch media reported that as part of its CANDU campaign, AECL's Marketing Director Bill Wemmers had suggested that Canada would take back the high level radioactive waste from CANDU reactors sold to the Netherlands. [374] In 1992, AECL reported that the Dutch technical assessment was continuing; that a team of Dutch scientists were visiting Canada in June 1992; and that France, Germany and Canada were "finalists" in the selection of a reactor system by the Dutch utility SEP. [375] In 1994, AECL reported that sales efforts were continuing. [376]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.11. The Philippines

In June 1981 Canada signed a Nuclear Cooperation Agreement with the dictatorial Marcos regime, opening the way for the sale of reactors, nuclear fuel and other technology. However, the following year, a Canadian embassy official said "Our chances of selling a reactor here are as slim as anywhere in the world. The Philippines won't go that route again. With the interest rates they are paying on their only nuclear plant, they could almost buy a thermal plant". [377] This did not stop Prime Minister Pierre Trudeau from promoting the sale of CANDU reactors in Manila in January 1983. [378] AECL continues to promote the sale of CANDUs to the Philippines. [379]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.12. Russia

In 1992, AECL reported that in conjunction with the Kurchatov Institute in Moscow, it was "assessing the possibility of locating a CANDU-6 power reactor in the St. Petersburg region". [380] Also in 1992, Canada signed a three-year nuclear safety initiative with Russia, sending Canadian nuclear personnel to Russia, and bringing Russians to Canada. [381] In May 1992, Minister of Energy, Mines & Resources Jake Epp went to Russia and Ukraine, visiting Chernobyl and the Sosnovy Bor nuclear station near St. Petersburg. While calling Chernobyl one of the world's "greatest human tragedies", he promoted the sale of CANDUs. [382] In June 1992, Brian Mulroney and Boris Yeltsin signed an accord that sent $30 million in Canadian aid to Russia. [383] The program, headed by the Roger Lucy of the Canadian International Development Agency (CIDA) is known as the Canadian Nuclear Safety Initiative (CNSI). At the end of 1995, about 75% of the $30 million had been spent on work including a $12 million study to improve safety at RBMK reactors. [384] Most prudent nuclear analysts have simply supported shutdown of the dangerous nuclear plants. AECL is using this aid as the first step in a marketing campaign aimed at future CANDU sales in Russia, Ukraine, Belarus and the Baltic states of Lithuania, Estonia, and Latvia. [385]

In October 1995, AECL met with officials of the Ministry of Atomic Energy for the Russian Federation (MINATOM) to discuss the possibility of building two CANDU-6 reactors near Vladivostok. Given the fact that the Russian nuclear industry is itself looking for business [386], it is highly unlikely that any sale could be made in the absence of extravagant financial support from the Canadian government. Nevertheless, AECL is apparently doing a feasibility study on "the energy problems of the region". [387]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.13. Thailand

AECL has reported that it is trying to sell the CANDU in Thailand. [388] In December 1995, Thailand's Office of Atomic Energy for Peace (OAEP) invited bids to supply a research reactor; an isotope production facility; and a waste processing and storage facility. Vendors bidding include AECL, as well as Chinese, American, Argentinean, German, and French companies. [389] The AECL bid is for the MAPLE reactor, which has not even been built in Canada, although the Korean Atomic Energy Research Institute (KAERI) has built one.

The Canadian Atomic Energy Control Board has trained members of the Nuclear Facility Regulatory Centre of Thailand for "several years". [390] In the Fall of 1995, the Thai Deputy Prime Minister Samak Sundaravej visited Canada and toured AECL's Chalk River Nuclear Laboratories. AECL also cosponsors a Human Resources Development Linkage Program in Thailand with the Canadian International Development Agency (CIDA). [391] Canada is sending a trade mission to Thailand in early 1997, in which AECL will undoubtedly participate. [392]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.14. United Kingdom

In 1963, a UK-Canada Nuclear Power Agreement was concluded, but instead of selling a CANDU, AECL sold the CANDU design to the United Kingdom Atomic Energy Authority for $750,000. [393] In the mid-1970s, AECL again tried to interest the United Kingdom in CANDU, when the British nuclear industry was trying to determine which type of reactor system to support. However, the effort failed. [394]

. . . back to [ Table of Contents ~ Sales Attempts ]

4.15. United States

While Canadian producers of plutonium (for the American weapons program) and uranium (for both military and civilian use) have been successful in penetrating the American market, the CANDU has been a non-starter. The American reactor market has been dominated by indigenous private-sector companies - mainly General Electric and Westinghouse. However, AECL established an American subsidiary (AECL Technologies- AECLT) in Rockland, Maryland, which headed up what it calls the "U.S. CANDU Initiative", mainly aimed at marketing engineering services to American nuclear utilities.

In 1993, AECL began the process of application to the US Nuclear Regulatory Commission for Standard Design Certification of the CANDU-3 reactor system. [395] However, this attempt to break into the US market also seems doomed to failure. In March 1995, AECL called an "indefinite" halt to its attempt to license the CANDU-3 in the US. AECLT has estimated that the certification review could cost US $50 million. [396] The US Nuclear Regulatory Commission performs a much more detailed comprehensive review of reactor design than Canada's Atomic Energy Control Board, and requires the reactor designer to pay for the review. The CANDU-3 is a 450 MW reactor which AECL began designing in 1987. As a relatively small reactor, it is billed as suitable for developing countries, or to reduce the risk of overbuilding capacity. However, faced with the failure of its proposals to build a CANDU-3 in Canada, [397] it seems unlikely that AECL will be successful in selling the reactor in the US or in the developing world.

. . . back to [ Table of Contents ~ Sales Attempts ]

4.16. Yugoslavia

In June 1986, AECL submitted a proposal to Yugoslavia for a CANDU nuclear program, and an office was maintained for some years in Zagreb. [398] Yugoslavia's international call for tenders attracted 10 bids including AECL's. The tender was for four plants to be built by the year 2000 at an estimated cost of $10 billion (US). [399] Yugoslavia had one nuclear plant at the time in northwestern Krsko, built by Westinghouse in 1981. The nuclear expansion program eventually collapsed.

. . . back to [ Table of Contents ~ Sales Attempts ]

5. Payment of Bribes & 'Agent Fees' by AECL

While 'Agent Fees' undoubtedly include fees for legitimate public relations and promotional activities, monies allocated as 'agent fees' by Atomic Energy of Canada Ltd. (AECL) have also been used in the past for bribery. The payment of bribes by AECL to secure CANDU sales has been well documented in the cases of Argentina and South Korea. The details of these cases have been provided in sections 2 and 3 above.

In the case of Argentina, almost $4 million was reported to have been spent on bribes in the 1973 to 1975 period. This was prior to the time when AECL was required to report 'agent fees', so these monies are not reflected in Table 2. In the case of South Korea, over $18 million was paid to an agent, who in turn bribed Korean officials to secure the initial CANDU sale to South Korea. The bulk of this bribe is reflected in the 'agent fee' in Table 2, for the year 1976-77. Despite the scandal around this incident, AECL's bribery of South Korean officials almost certainly continued at least until 1991.

Following a report by Canada's Auditor General in 1976 on AECL financial irregularities [400], the federal government authorized a review of AECL by the House of Commons Committee on Public Accounts (1976 77). The Committee found AECL witnesses to be uncooperative - in particular AECL President Lorne Gray. Subsequently, the government launched an inquiry by the Royal Canadian Mounted Police (1977-78) on AECL use of bribery. The reviews did little to restrict AECL in its marketing practices. However, beginning in 1977, AECL was required to report 'agent fees' in its Annual Reports. The Annual Reports name the companies of the agents, but only provide a total for the monies paid out, i.e. the amounts paid to each individual agent is not provided.

From 1977 to 1994, AECL paid almost $60 million - an average of about $3 million per year (see Table 2).

Table 2 : "Agents' Fees" Paid by AECL 1977-1996

Agents' Fees   ($ of the year)
Agents' Fees   ($ of the year)
$17,400,000.00   [$17.4 million]
$2,000,000.00   [$2.0 million]
$2,200,000.00   [$2.2 million]
$2,400,000.00   [$2.4 million]
$1,600,000.00   [$1.6 million]
$2,500,000.00   [$2.5 million]
$800,000.00   [$0.8 million]
$1,000,000.00   [$1.0 million]
$600,000.00   [$0.6 million]
$3,100,000.00   [$3.1 million]
$800,000.00   [$0.8 million]
$900,000.00   [$0.9 million]
$3,400,000.00   [$3.4 million]
$1,700,000.00   [$1.7 million]
$1,900,000.00   [$1.9 million]
$3,800,000.00   [$3.8 million]
$1,600,000.00   [$1.6 million]
$7,300,000.00   [$7.3 million]
$2,300,000.00   [$2.3 million]
$2,000,000.00   [$2.0 million]
Total 'Agent Fees' Paid by AECL 1977-1996
$59,300,000.00   [$59.3 million]
Annual Average 'Agent Fees' 1977-1994
$2,970,000.00   [$2.97 million]

Source: AECL Annual Reports 1976-77 to 1995-96

6. Conclusion

Since construction was completed on Ontario Hydro's $14 billion Darlington Nuclear Generating Station in 1993, no utility in Canada has agreed to further nuclear construction.

Internationally, nuclear power is experiencing a brown-out. For the first time since commercial nuclear programs began in the 1950s, the tide is clearly turning against nuclear power. As of January 1996, only 34 nuclear plants worldwide were under construction - the smallest number in 25 years. A total of 84 reactors had been shut down, and 434 reactors were operating. [401] Nuclear power's share of world electricity supply probably peaked in 1994-95 at about 17%, and is expected to decline to 13% by 2015. [402]

Countries which are targeted by AECL marketing efforts typically have dictatorial regimes, or have only recently democratized. The same countries tend to have serious problems with human rights and corruption. Public consultation on energy futures in CANDU client countries is virtually non-existent. Small elite groups are responsible for promoting nuclear power programs despite their high cost, environmental and safety risks, and technological complexity.

The potential for reactor sales to contribute to nuclear weapons capability cannot be avoided.

This author of this report recommends:

  1. Canada should terminate its subsidization of Atomic Energy of Canada Ltd. (AECL). If reactor exports are as promising as AECL claims, then the industry should be able to support itself.

  2. Because AECL is a crown corporation, supported by public funds, the Canadian government should require full public disclosure of the financial structure and terms of CANDU export agreements.

  3. Canada should end governmental financial backing for nuclear exports. Financing by the Export Development Corporation (EDC) should be limited to its 'Corporate Account', and aid from the Canadian International Development Agency (CIDA) for exports of nuclear technology should be prohibited.

  4. Canada should strengthen rather than weaken its non-proliferation policy. No nuclear cooperation should be permitted with known or suspected proliferators.

  5. The Canadian government should require detailed public disclosure of the amounts and the purposes of fees paid to each foreign agent of AECL. Marketing costs should also be itemized and reported by AECL.

  6. Canada should abandon its policy of 'constructive engagement' on human rights, and be prepared to impose trade sanctions in response to serious and sustained human rights violations.
[ . . . MORE . . . (Footnotes) ]

Appendix A

Table 3 : Canada's Nuclear Cooperation Agreements

In Force
1. Argentina a
June 1994 July 1996
2. Australia
October 1959 March 1981
3. Brazil
May 1996 NIF
4. China
November 1994 November 1994
5. Columbia
1986 June 1988
6. Czech Republic
February 1995 February 1995
7. Egypt
May 1982 November 1982
November 1959 see note
9. Hungary
November 1987 January 1988
10. Indonesia
July 1982 July 1983
11. Japan c
July 1960 July 1960
12. Lithuania
November 1994 May 1995
13. Mexico
November 1994 March 1995
14. Philippines
June 1981 April 1983
15. Republic of Korea d
January 1976 January 1976
16. Romania
October 1977 June 1978
17. Russia
November 1989 November 1989
18. Slovak Republic
1996 ?
19. Slovenia
May 1995 April 1996
20. Switzerland
December 1987 June 1989
21. Turkey
June 1985 July 1986
22. Ukraine
December 1995 NIF
23. United States of America e
June 1955 see note