WHAT could be the worlds most expensive clean energy project is under way in a quiet corner of southern France.
China, the United States, Japan and the European Union have committed billions of dollars to construction of the International Thermonuclear Experimental Reactor, or ITER, in a heavily forested corner of Provence called Cadarache that is a center for atomic research.
The goal is to prove that energy can be generated through nuclear fusion a process akin to how light and heat are produced by the sun. The promise is virtually unlimited amounts of energy from abundantly available sources.
Fusion creates no greenhouse gases and produces far less hazardous waste than fission, the current nuclear process, although fusion reactors do become radioactive and waste would still require special disposal.
If successful, the concept is not expected to be commercially viable until midcentury. There has already been a two-year delay because of difficulties setting up an international organization for the project.
Rising costs for equipment could further complicate relations among the participants, which include South Korea, India and Russia.
Even so, scientists say an international approach is critical.
No one nation can develop fusion alone, said Pascal Garin, the project leader at the International Fusion Material Irradiation Facility in Japan, which is helping to develop materials for the reactor. The technical and economic challenges are enormous compared with other low-emissions technologies like solar power or conventional nuclear power. The original budget, set in 2001, was about $10 billion, to be spent over 30 years. About half that amount was to be spent by participating governments on national projects to build components for the reactor.
That budget is being revised because prices of crucial commodities like steel and copper have soared, and because features are being added to the original design to help to ensure that the experiment works.
The seven parties are expected to decide on a new amount, which could be significantly higher, and on how to share the burden of paying for it, at meetings in June and November, said Neil Calder, a spokesman for the ITER Organization.
The parties are also expected to decide this year whether Kazakhstan, which offers nuclear expertise, and mineral and oil wealth, will join the organization, Mr. Calder said.
While adding another country could help defray the costs borne by each member, it also could make the project even more complicated.
The United States has already spent about $27 million to finance the ITER Organization, which is responsible for the construction and safe operation of the reactor, according to the Department of Energy. The countrys total contribution to the project will be $1.45 billion to $2.2 billion over the next decade, the department said.
The European Commission, which is the unions executive body here in Brussels, said the 27 member nations and Switzerland had committed to spend about 2 billion euros ($2.6 billion) on ITER over the next five years, but that sum could rise sharply.
Over all, the European Union is supposed to contribute most of the costs, or more than 45 percent of the total, in exchange for locating the project in France. The agreement consists of funds contributed by the union and Japan totaling about 700 million euros.
So far, experimental fusion reactors have required more energy to operate than they have produced. ITER and the project in Japan are supposed to prove that fusion could be much more than a drain on the grid.
Nuclear plants now generate power through fission, splitting heavy uranium atoms. Fusion involves heating very lightweight hydrogen atoms to about 100 million degrees Celsius or about 10 times the temperature of the sun.
This process creates a so-called plasma gas in which normally repelling particles combine and yield vast amounts of additional energy. By confining the hot plasma with the use of powerful magnets, the scientists aim to keep the process going in much the same way that the sun, confined by gravity, continues to burn.
Scientists say the heat from fusion could be used to drive electrical turbines with steam.
Scientists already know how to make a reactor like ITER work, but they say they need to conduct about 15 years of experiments once the machine is built to learn how to keep it running and to test materials under extreme conditions.
Keeping the financing going over that period could be a challenge as governments and priorities change. Experts also warn that nations like China, keen on developing their own fusion industries, could eventually split from the group.
The project is a vast undertaking. In March, the authorities at Cadarache finished clearing and flattening a forested area the size of 70 football fields where the reactor will be built. The foundation will be laid in September and shortly after construction should begin on dozens of other buildings where huge components like magnets, each weighing several tons, will be assembled.
By 2018, the machine should be ready.
Scientists then would spend the next decade or so trying to create bursts of power of up to 500 megawatts for several minutes at a time, and experimenting with the plasma gas.
Critics say that if the primary purpose of the ITER venture is to create large supplies of emissions-free electricity to mitigate climate change, then the money would be better invested in existing technologies like wind and solar power, which have been proven and are ready for deployment.
But defenders of the project say the total that will be spent on ITER and its associated projects is small when compared with what some companies would spend on research and development for consumer goods, like new car models, over the same time frame.
Defenders also liken the project to the NASA space program, which yielded spinoff technologies that benefit other industries.
European Union officials said that work on the reactor could lead to a host of breakthroughs, including the development of new coatings that perform well under very high temperatures, as well as new designs for microchips.