Following the sodium leak and fire at Monju fast breeder reactor in December 1995², Japan switched the focus of its nuclear fuel cycle policy from fast breeder reactor development to MOX (mixed oxides of plutonium and uranium) fuel use at light water reactors. The MOX plan is commonly called the plu-thermal program in Japan. While fast breeder reactor (FBR) development has been hindered mostly by technical problems, the plu-thermal program has been met with great difficulty due to strong local opposition. As such, there are many people, even among the nuclear promoters, who back a "once through" approach.³ However, neither the utilities, the Japanese government, nor the country's Atomic Energy Commission have plans to end their promotion of the plu-thermal program.

In fact, as part of the plu-thermal program, several facilities are planned or already in operation. The Tokai Reprocessing Plant began full operation on 20 November 2000 after being shut down for over three-and-a-half years following the fire and explosion at the facility's bituminization plant in March 1997.⁴ In December 2000, the operator of Monju, shut down since the 1995 incident, asked Fukui Prefecture and Tsuruga City to agree that it should apply to the Ministry of Economy, Trade and Industry for the safety review of its remodeling plans for Monju. The construction of Rokkasho Reprocessing Plant, to be completed by July 2005, continues at an accelerated pace.

JAPAN

The Rokkasho Reprocessing Plant

The Rokkasho Reprocessing Plant is being constructed in Rokkasho Village, Aomori Prefecture, by Japan Nuclear Fuel Ltd., and financed by major Japanese electric power companies and the nuclear industry⁵, for the purpose of processing spent fuel from Japanese light water reactors. About 35 large and small concrete buildings will be built on about 3.8 million square meters of land. Each building will consist of four underground floors and four floors above ground, and thus half of the plant is being built underground. The total distance of pipes which connect the facilities will amount to about 1,500 km. As of the end of March 2001, 64% of the plant's construction had been initiated. Because of the subcontracting system particular to the Japanese construction industry, there are about 1,000 companies involved in the construction, and about 7,000 construction workers working around the clock.

Though the plant is still under construction, the spent fuel storage pool has already been completed. The maximum storage capacity of the pool totals 3,000 metric tons of uranium, or tU (1,500tU each for boiling water reactor and pressurized water reactor spent fuel). The transportation and storage of spent fuel at the pool began in December 1999. About 1,600tU of spent fuel is expected to be stored at the pool by the time of the completion of the plant.

Rokkasho Reprocessing Plant will use the PUREX method, which dissolves spent fuel in nitric acid and separates uranium, plutonium and high-level waste. The plant's annual capacity is 800tU, with a daily maximum capacity of 4.8tU, and will annually separate about 5 metric tons of fissile plutonium. Maximum burn up of the spent fuel to be reprocessed at this plant is 55,000 megawatt days thermal per metric ton of uranium (MWdth/tU). The average burn up of spent fuel reprocessed within a day will be under 45,000MWdth/tU. The spent fuel will be cooled for more than a year before it arrives at the plant, and must be cooled for more than four years before it is sheared.

As with the Tokai Reprocessing Plant, operated by the Japan Nuclear Cycle Development Institute (JNC), the main process of the Rokkasho Reprocessing Plant is based on technology imported from France and is modeled after the French company COGEMA's UP-3 Plant located in La Hague, France. Other parts are based on technology adopted from various countries.

As shown in Figure 1, the plant consists of the following processes: receiving, storage, chopping (shearing), dissolving, separation, refining (purification), denitration, storage of product uranium and plutonium, and solidification (vitrification) of high-level radioactive waste. In most cases there is one building for each process. Technology for main processes like chopping, dissolving, separation, and refining is provided by COGEMA's subsidiary SGN.

Technology from:
1. SGN (COGEMA's subsidiary)
2. BNFL
3. KEWA
4. JNC

The plant is a mosaic of technologies from overseas and domestic companies. Technology for high-level liquid waste treatment and acid recovery is provided by British Nuclear Fuels, plc (BNFL), iodine removal technology by Germany's KEWA, uranium-plutonium denitration technology by JNC, Mitubishi Materials, and Toshiba, high-level liquid waste vitrification technology by JNC and Ishikawajima-Harima Heavy Industries, and spent fuel storage pool technology by Hitachi, Toshiba and Mitsubishi.

The basic blueprint for the main processes adopted from SGN was prepared by SGN itself, but blueprints for the processes with technology from other overseas companies and domestic companies were prepared by Japanese companies. Due to Japan's earthquake prone nature, it was necessary to also add anti-seismic designs to the blueprint. Thus Japanese companies made the necessary alterations and additions to the blueprints for aseismic purposes, and were also in charge of detailed designing, manufacturing and installing of aseismic equipment. This complex and confused process resulted in mis-transcribing of designs, and a number of missing or faulty parts have been discovered as the construction progresses (described below).

The plant differs from the reprocessing plants in France and England in that it will produce mixed 50% plutonium and 50% uranium oxide (MOX) as the end product, whereas French and British plants yield uranium oxide and plutonium oxide separately. As a non-proliferation measure, Japan is forbidden under the U.S.-Japan Nuclear Agreement⁶ to extract plutonium from uranium supplied by the U.S. Most of Japan's spent fuel includes uranium from the U.S.

Partial test operations using water and vapor began in April 2001at completed parts of the Rokkasho Reprocessing Plant to identify cracks, holes and problems with welding and connections of pipes. Tests and test operations using uranyl nitrate solution and subsequently spent fuel dissolved in nitrate solution will continue until the planned completion of the plant in 2005. For example, the confirmation tests for the chopping and dissolving treatment building alone involves literally millions of check items to confirm whether that part of the plant is precisely built according to the blueprint.

The inexperience of Japan Nuclear Fuel Ltd. (JNFL) and the fact that the plant's technology has been adopted from various companies pose serious concerns about the construction and operational safety of the plant. On 26 February 2000, it was reported by the Daily Tohoku and the To-o Nippo Newspaper that a storage tank for low-level radioactive liquid waste and two temporary storage tanks for high-level radioactive concentrated liquid waste brought into the plant were lacking important parts due to Hitachi staff's mis-transcribing the blueprint.⁷ For example, the aseismic support for the inside of the high-level concentrated liquid waste storage tanks were inversely installed. Such defects are a result of the confusion between SGN and Japanese companies following the drastic alterations that were made to the original blueprint in order to lower cost after construction began in 1993. JNFL's incompetence regarding quality control is clear, and it is highly possible that there are various problems with other parts of the plant and equipment.⁸

In 1987, JNFL signed the Technology Transformation Agreement with the French company SGN, and the General Framework Agreement with U.K.'s BNFL. JNFL has requested COGEMA to send about 50 and BNFL to send a couple technical assistants during the test operations until the plant's completion. Since last year, COGEMA technicians and their families have begun to arrive in Rokkasho, and a "French Village" is being constructed in the vicinity of Rokkasho Village where houses are being built especially for the French technicians.

Together with technology transfer, plant operator training is proving to be a serious task. JNFL expected to accomplish technology transfer by having Japanese workers trained at facilities of COGEMA and BNFL in addition to JNFL's mock reprocessing facility, JNC's Tokai Reprocessing Plant and test facilities of the domestic nuclear industry. Training at domestic facilities is limited since it is conducted at mock or test facilites or at the Tokai plant which does not use the technology with which the Rokkasho plant will be equipped. JNFL is negotiating with COGEMA to get its operators trained at COGEMA's UP-3 plant, but the negotiations have been met with difficulty. As a condition for such an arrangement, COGEMA is requesting training fees of about 100 million yen⁹ per trainee, in addition to new contracts for the reprocessing of Japanese spent fuel. The two have not finalized this plan.

According to the original licensing application, submitted in 1989, the Rokkasho plant was to be completed in December 1997, but the completion date has been postponed four times. When construction began in 1993, the completion date was set at January 200 and the total construction cost was estimated at 760 billion yen. In 1996, due to construction delays, the projected completion date was moved to January 2003, and the construction cost re-estimated at 1.88 trillion yen . Then in 1999, the completion was projected to be July 2005, and estimated construction costs soared to 2.14 trillion yen (about US$20 billion) -- three times the original estimate. The construction cost is expected to rise further by the time of the plant's completion.

JNFL has not released an estimate of the plant's projected reprocessing cost which takes into account this rise in construction cost. It is estimated by the Japanese Agency for Natural Resources and Energy that reprocessing at Rokkasho will cost around 351 million yen per metric ton of spent fuel¹⁰, which would be about one and a half times the reprocessing costs of BNFL and COGEMA.¹¹ There are other estimates with higher cost projections, the highest being 500 million yen per metric ton of spent fuel.¹² Because of Germany's decision not to continue reprocessing, COGEMA and BNFL might be forced to lower their reprocessing costs and thus force Rokkasho's expensive reprocessing cost to be even more non-competitive. [Editor's note: This MOX fuel will be at least 20 times more expensive than LEU (low enriched uranium) fuel. French MOX fuel is about 5 times more expensive than LEU fuel.]

According to its financial report for fiscal year 1999 (April 1999 to March 2000), released in June 2000, JNFL was 500 million yen in the red after taxes. This financial report included the utilities' annual allotment payment of 12.5 billion yen for the construction of the Rokkasho Reprocessing Plant, but this allotment is not covering the rising construction cost. JNFL is going further into debt by continuing the construction of the reprocessing plant, and it is very likely that Rokkasho-manufactured plutonium will be the world's most expensive.

While the construction of Rokkasho Reprocessing Plant progresses, concerns are intensifying over the possibility that the plutonium separated at the plant will become excess. To amend this situation, Japan's Federation of Electric Power Companies and JNFL decided to build Japan's first commercial MOX plant.

Plans call for the MOX fuel fabrication plant to be constructed in the vicinity of the Rokkasho Reprocessing Plant. The two plants would be connected to each other with an underground trench through which MOX powder will be transferred. This large-scale plant is projected to have an annual processing capacity of 130 metric tons of heavy metal (t-HM), and to begin operation in 2008 or 2009. Like plants in Belgium and France, the MIMAS method¹³ was chosen as the manufacturing process. The total construction cost is estimated at 120 billion yen.

The plutonium enrichment level of the MOX fuel manufactured at the proposed plant would range from 5 to 10%. The plant would use the 50:50 uranium/plutonium MOX powder manufactured at the Rokkasho Reprocessing Plant as raw material. The mix would be diluted by adding depleted uranium stored at the Rokkasho Enrichment Plant. Fuel pellets would then be manufactured and fuel rods prepared.

Because plutonium is involved with MOX fuel manufacturing, neutron emissions increase by about 10,000 times and gamma-ray emissions increase by about 20 times compared to the manufacturing process of uranium fuel. Thus, strict safety control is necessary, especially to set up measures for shielding and trapping of radioactive materials, heat management, and criticality control. The 1999 criticality accident at JCO's Tokai plant, which claimed two lives, exposed local resident to neutrons and forced residents within a 350 meter radius of the plant to evacuate, is still fresh in the minds of the Japanese public.

In addition since this raw MOX will include retrieved uranium, unlike MOX plants in England and France, the protection of workers and the public from strong gamma-rays from daughter nuclides of uranium-232 and uranium-236 will pose a considerable challenge. The Rokkasho MOX fuel fabrication plant is burdened with the necessity to have fortified shielding structures and stricter control over worker exposure than other MOX plants.

As if tagging behind the industry's plans, the Nuclear Safety Commission (or NSC, Japan's counterpart to the U.S. Nuclear Regulatory Commission) is hastily preparing safety review standards for the licensing of the commercial MOX plant in order to set standards by the end of this year. Moreover, discussions on the review standards include talk of possibly using plutonium oxide as raw material in addition to 50:50 MOX. Should this become a reality, it may be that Japanese-owned plutonium extracted and stored overseas would be returned, since under the U.S.-Japan Nuclear Agreement, Japan cannot extract pure plutonium in the country from most of its spent fuel. The use of plutonium oxide at a large-scale plant in Japan poses serious safety and proliferation risks and is not something to be taken lightly. Citizens must keep a watchful eye on the NSC as it prepares the safety review standards.

Meanwhile, plans to use MOX fuel in Japan are reaching a dead end. As of May 2001, none of the MOX fuel transported since Fall 1999 from Europe to Japan has been used due to BNFL data falsification scandals and strong local opposition.¹⁴ Utilities had planned to begin irradiating MOX at light water reactors in Fall 1999. At the moment, none of the country's reactors have fixed dates for the loading of MOX fuel. Even if the plans had progressed according to original targets, the projected demand for plutonium until 2010 would be 30 metric tons, in comparison to the projected supply which is 55 metric tons -- a serious surplus of plutonium. (See Figure 2).

There is strong criticism inside and outside Japan over the country's plans for domestic reprocessing and MOX manufacturing. The 1999 JCO criticality accident, MOX fuel data falsification scandal, and other scandals involving concealment and manipulation of information concerning the 1995 Monju breeder reactor incident and the 1997 Tokai incident have made the Japanese public increasingly skeptical of nuclear technology and the nuclear industry. The government's blind promotion of plutonium use, despite such skepticism and local opposition, will meet further delays and difficulties.