Towards the end of the CDA, its success let to preparations for a more detailed engineering design phase. This involved negotiations between the ITER Parties on the terms of the agreement between them and on the site of the design work. Such an agreement was needed due to

• the much larger scope of the design work,
• the fact that considerable R&D would have to be carried out on the design,
• the wish that each Party had to use the information generated on its own if necessary to build the next step,
• and therefore the need to have a more explicit agreement about the sharing of intellectual property and costs.

In particular the agreement would enshrine the equal contribution each Party was expected to make to the work, the equal status each would have in the project, and the fact that the information gained would be fairly shared between all the Parties. Furthermore, there was no commitment to go on in the same frame to construct ITER - rather that each Party should be able to build it on their own if necessary.

It was initially thought that agreement would be quickly reached, but three of the Parties offered comparable sites for the design work - in Garching, Germany, Naka, Japan, and San Diego, USA - and were not prepared to back down on their wish to host the design work. Quadripartite negotiations took 2 years to reach agreement on this point, with the result that the design team for the EDA would be split over the three sites (named "Co-centres" and later "Joint Work Sites"). Vessel and In-vessel components would be designed in Garching, Ex-vessel components (e.g. magnets) would be designed in Naka, and utilities design, project engineering, and design integration would be carrried out in San Diiego.

This "Joint Central Team"or JCT would be supported by "Home Teams", one per ITER Party, which would consist of all those working on ITER activities. These Home Teams would conduct detailed design tasks and R&D as agreed with the JCT and documented in Task Agreements and their Final Reports. The project would have no central funding (although a small Joint Fund was later created for common project purchases), and the work of the Parties would be carried out for "credit" so that a check could be kept on the size of the contribution from each Party. General oversight of the project would be under the control of the ITER Council.

Signature of the ITER EDA Agreement in 1992

The ITER EDA Agreement, under the auspices of the IAEA, was finally signed on July 21st 1992 in Vienna, with the Russian Federation replacing the Soviet Union. Canada and Kazakhstan became involved in the Project by association with Euratom and Russia respectively. The EDA was defined initially for a six year period during which the Parties agreed jointly (and on a basis of equality) to produce a detailed, complete and fully integrated engineering design of ITER and all technical data necessary for future decisions on the construction of ITER. The Agreement called for the EDA to be implemented through 2 Protocols. Protocol 1 was accepted and signed by the four Parties at the same time as the Agreement. One of the aims of the EDA was to produce a design close in cost to that proposed during the CDA.

Paul-Henri Rebut, previous Director (and chief engineer) of the JET Joint Undertaking, was appointed ITER Director, and pursued an ambitious programme to develop and convince the Parties of his vision of a "thermonuclear furnace", larger than the concept for ITER developed in the CDA. Such a design was in some respects more ambitious (using innovative designs as well as more power-reactor-relevant materials such as vanadium as structural material), and in other respects more conservative (for instance the larger size bestowing a greater margin for obtaining plasma ignition.

Initial Concept of "Thermonuclear Furnace"

An outline design (above) was developed by the end of 1993. This served as a point of departure for proceeding into Protocol 2, the second and last step in the implementation of the EDA Agreement. However the reaction of the Parties to the design itself was cautious. Criticisms were made that it was overly ambitious, and would require extensive R&D which the Parties could not afford and which would take too long to accomplish before ITER could be licensed. But Dr. Rebut continued to pursue his vision in the face of the strengthening opposition, and in July 1994 he was replaced as ITER Director by Robert Aymar. Dr. Aymar continued to seek a design with similar plasma parameters (see below), but involving more conservative engineeering.

The design development was not realised overnight, and the design went through several changes documented in an Interim Design Report in June 1995, a Detailed Design Report in January 1997, leading up to a Final Design Report on in July 1998. This work was regularly overseen by a Technical Advisory Committee, and project resources were managed through a Management Advisory Committee. All these reports were published by the IAEA. By July 1998 the JCT consisted of 150 professionals, which had contributed 740 pmy of design effort, and the Parties' Home Teams had contributed a further 730 pmy (EU-28%, JA-25%, RF-21%, US-26%). This was about 25% more than had been envisaged by the CDA assessment, and was in line wiith the expectation that splitting the design team into three parts would need extra staff for management and liaison.

ITER 1998 Design

Evolution of ITER design parameters

In the first two years of the EDA the R&D for the project had addressed many issues, but there was considerable duplication between the four Parties, and some considerable work carried out on what would turn out to be low priority issues. The R&D continued to be organised by each Party to emphasise those items each Party thought would be important, and the design of ITER had not developed enough, nor had the central team been established long enough, to allow proper direction of the work. In 1994 the TAC recommended a refocussing of the work onto a few large R&D projects, addressing the key components and technologies to prove the viablility of ITER construction. These seven projects would each be multi-Party, already rehearsing the cooperation that would later be essential for ITER construction. By July 1998 the total volume of expenditure by the Parties on technology R&D during the EDA had risen to $565M (1989 values) shared EU-33%, JA-32%, RF-15%, US-20%.

Six years of international collaborative work within the framework of the ITER EDA Agreement culminated in the approval by the ITER Council in June 1998 of the ITER Final Design Report. This provided the first comprehensive design of a fusion reactor based on well established physics and technology. Its design fulfilled the overall programmatic objective of ITER and complied with the detailed technical objectives and technical approaches adopted by the ITER Parties at the start of the EDA.

After mid-1997, doubts began to be expressed inside the US physics community about ITER's ability to ignite. Meanwhile the ITER design had evolved in sufficient detail that reasonably detailed and accurate estimates could be made of the overall machine cost. At the same time oil prices had reached a low and with the Middle-East relative stability there were few concerns for future energy demands.

At the time of the FDR’s acceptance, therefore, the ITER Parties, recognising the possibility that they might be unable, for financial reasons, to proceed to the construction of the then foreseen device, deemed it prudent to make available option(s) for ITER whose costs would be reduced by reducing ITER’s detailed technical objectives and possibly decreasing the technical margins while maintaining the overall programmatic objective of ITER. Following the recommendations of a Special Working Group (SWG), the ITER Council asked the Director "to continue efforts with high priority toward establishing, with the assistance of the JCT and Home Teams, option(s) of minimum cost aimed at a target of approximately 50% of the direct capital cost of the present design with reduced detailed technical objectives, which would still satisfy the overall programmatic objective of ITER. The work should follow the adopted technical guidelines and make the most cost-effective use of existing design solutions and their associated R&D." The EDA was subsequently extended to July 2001 to cover this work, in spite of the US decision to leave the project by the end of its 1999 fiscal year.

The SWG was also asked to give the basis for its rationale for the proposed guidelines and to examine the effect of alternative development paths towards fusion energy. They concluded that:

• Successful development of fusion energy requires demonstrating scientific and technological feasibility as well as environmental attractiveness and economic viability.
• The long-term goal is a convincing demonstration of this in a demonstration power plant. (DEMO).
• A critical prior step is to integrate high energy gain plasmas at or near steady-state conditions with power-plant-prototypical technologies, and demonstrate safe operation of a fusion power system.
• ITER will provide the conditions required for these critical tests, and it has focused the attention of the world fusion community on key scientific and technical issues.
• Through this collaboration, the cost and benefits of this important step can be shared by the ITER Parties.
• A device, which would achieve energy gain of at least 10 and explore steady-state operation, at a direct capital cost of approximately 50% of the 1998 ITER design, would still satisfy the ITER overall programmatic objective.
• Successful operation of ITER would provide experience of broad, generic value to fusion energy development.
• A strategy of constructing a new generation of separate specialized facilities addressing selected critical issues would delay by 10 years or more the key fusion demonstration and integration step, and would increase the total cost of fusion development substantially.
• The world program is scientifically and technically ready to take the important ITER step.

Work focused on the key issues involved in the revised detailed technical objectives, drawing on the database already established through the collaboration, and on understanding of the choices and trade-offs that needed to be made between performance margins and cost reductions. In parallel the complete experimental physics community was mobilised to document a consensus view on the physics basis of ITER.

Complete and fully integrated documentation of the new ITER design, initially known as ITER-FEAT to distinguish it from the earlier design, and now just known as the ITER design, was produced to a level suitable for the Parties to make a decision on construction, including a full cost analysis and a safety assessment for ITER sited at a generic site. This was approved by the ITER Council in July 2001.

The ITER Final Design

This design redefinition was accomplished in the EDA extension by the expenditure of an additional 280 pmy of JCT effort, plus 210 ppy of Home Team design effort (EU-37%, JA-38%, RF-25%). Further R&D worth $92M (1989 values, EU-46%, JA-44%, RF-10%) was carried out by the Parties to support the developing design.

The next logical step was to start construction of ITER. The way the project had been conducted during the EDA, in accordance with the terms of the EDA Agreement it was in principle possible for any one Party to now go and build ITER. However, all the remaining ITER Parties wished to continue the collaboration, so negotiations began on "joint implementation".