ITER - PDF by wulinqing



      With its big political hurdle behind it,
      the make-or-break project must run a gantlet of
      technical challenges to see whether fusion can
      fulfill its promise of almost limitless energy

      ITER’s $12 Billion

                                                                                                                                                                           Downloaded from on December 4, 2007
      ABINGDON, U.K., AND GARCHING, GERMANY—            seawater in virtually limitless
      Several times a year, hunters gather in the       quantities. Such a plant,
      forests around Saint Paul lez Durance in          unlike alternative sources,
      southern France to shoot wild boar. Over the      would produce the steady,
      coming decade, however, a portion of their        reliable baseload power that
      hunting ground will be cleared, and the town’s    cities need. And the eco-
      cafés will gradually fill up with newcomers       nomics are astounding: A
      from San Diego and Seoul, Moscow and              1-gigawatt coal-f ired plant
      Munich, Naka and New Delhi. Rising out of         burns about 10,000 tonne
      that forest clearing will be a 20,000-tonne       of coal per day, whereas a
      experiment that just might point a way out of     1-gigawatt fusion plant would
      the world’s looming energy crisis.                need roughly 1 kilogram of
          In November, politicians representing         deuterium-tritium fuel.
      more than half the world’s population will            We’re not even close yet,
      sign an agreement that fires the starting pis-    however. Indeed, skeptics joke
      tol for the International Thermonuclear           that “Fusion is the power of the
      Experimental Reactor (ITER). Although             future and always will be.” The sun
      f irst mooted in 1985, ITER has so far            is a gigantic fusion reactor, but recre-
      existed only on paper. The governments of         ating the conditions here on Earth in
      China, the European Union (E.U.), India,          which atomic nuclei collide with such force      have to keep on doing it steadily for years
      Japan, South Korea, Russia, and the United        that they fuse together has proved fiendishly    without a break.
      States are now ready to hand over a $6 bil-       difficult. A few dozen examples of the cur-          ITER needs to show such performance is
      lion check for ITER’s construction, fol-          rently favored reactor design—a doughnut-        at least possible. But it faces many chal-
      lowed by a similarly sized one for 20 years’      shaped vessel known as a tokamak—have            lenges: Scientists and engineers need to find
      operation. Then it is up to an international      been built since the 1950s, but only a handful   a lining for its inner walls that can withstand
      team of scientists and engineers to show that     have managed to get fusion in their plasma.      the intense heat; they must tame the plasma
      the thing will work.                              In 1997, the Joint European Torus (JET) in       instabilities that plague existing reactors; and
          If it does, the rewards could be huge. With   Abingdon, U.K., the biggest existing toka-       they must find a way to run the reactor in a
      the global population due to climb from           mak, managed to produce 16 megawatts, but        steady state rather than the short pulses of
      6.5 billion to 8.1 billion by 2030 and the        that was only 65% of the power used to keep      existing reactors. ITER must do all of this
      economies of China, India, and others hun-        the reaction running.                            and, for the first time, maintain the plasma
      gry for power, many new generating plants             By studying those earlier reactors,          temperature with heat from the fusion reac-
      will have to be built. The choices are stark:     plasma physicists have derived scaling laws      tion itself rather than an external source.
      Burn more coal, with the inevitable impact        that predict that a bigger tokamak (ITER is          “There’s no doubt that it’s an experiment.
      on climate; build new nuclear fission plants      twice the size of JET in linear dimensions)      But it’s absolutely necessary. We have to
      and deal with the radioactive waste and risk      would overcome many of the problems. But         build something like ITER,” says Lorne
      of terrorism; or try alternative sources such     ITER is not a prototype power plant; it is an    Horton of the Max Planck Institute for
      as solar power, although this option remains      experiment designed to f inally decide           Plasma Physics (IPP) in Garching, Germany.
      expensive and lacks efficiency.                   whether taming the sun’s energy to generate      Researchers are reasonably confident that
          But there is an outside bet: fusion. If it    electricity is even viable. ITER aims to pro-    ITER can achieve the basic goals laid out in
      can be built, a fusion power station would        duce 500 megawatts of power, 10 times the        the project’s plans, but there is less certainty
      emit no greenhouse gases and produce little       amount needed to keep it running. But a
                                                                                                                                                            CREDIT: ITER

                                                                                                         about what comes after that. “I’m pretty
      radioactive waste, it cannot explode in a         moneymaking energy utility would need            confident ITER will work as advertised, but
      runaway reaction, and its fuel is found in        several times that amount, and it would          you can’t be 100% sure,” says Christopher

238                                       13 OCTOBER 2006         VOL 314     SCIENCE
                                                                       Published by AAAS

                   Hotter than the sun. ITER’s interior must endure     they’re itching to have some input. Fusion              boundary conditions. We must respect the
                   colossal heat loads and neutron bombardment.         science has moved on in the 5 years since               agreement and keep within the set cost. They
                                                                        ITER’s design was completed, and many                   can’t expect miracles.”
                   Llewellyn-Smith, director of the U.K. Atomic         want to make changes in the light of recent                Holtkamp has heeded calls for a design
                   Energy Authority’s Culham Laboratory in              results. This generates a creative tension              review and will convene a meeting in
                   Abingdon, home of JET. IPP’s Hartmut                 between the wider fusion community,                     December. But, as Chuyanov points out, not
                   Zohm agrees: “Certainly there’s an element           which would like an adaptable machine to                everything has to be set in stone now. Con-
                   of risk. I’m very confident, 90-something            test as many scenarios as possible, and the             tracts for big-ticket items such as the build-
                   percent, that we can produce a plasma domi-          ITER staff, who want the machine built on               ing, the vacuum vessel, and the super-
                   nated by fusion. But I’m much more uncer-            time, on budget, and ready for the next                 conducting magnets must be signed almost
                   tain that it will make a viable fusion power         step: a power plant prototype.                          immediately, but other systems are years
                   reactor.” German physicist Norbert Holtkamp              “The physics community still wants mod-             away from procurement. “The design review
                   says that ITER’s goal of generating excess           ifications, all the bells and whistles. We’ll           is not a moment in time but a continuous
                   power is clear: “Either it can do it, or it can’t.   always keep asking. It’s healthy,” says Horton.         process,” Chuyanov says.
                   If it fails, the tokamak is out.”                    Valery Chuyanov, head of ITER’s work site in               Above all, researchers want the ITER
                                                                        Garching and now the nominee deputy direc-              design to be flexible. Since it was fired up in

                                                                                                                                                                                            Downloaded from on December 4, 2007
                   The waiting game                                     tor general for fusion science and technology,          1983, JET has had numerous transformations
                   The ITER project is currently in a state of          counters that physicists “must understand the           including the retrofitting of a divertor, a
                   limbo. Researchers nailed down the design of
                   the reactor in 2001 after a 13-year effort cost-
                   ing about $1 billion. Since then, governments
                   have been in charge. The ITER partners at
                                                                          How to Squeeze a Plasma
                   that time—the E.U., Japan, and Russia (the             After numerous attempts during the 1940s and 1950s to find an arrangement of magnets to confine
                   United States had pulled out in 1999)—                 a plasma—an ionized gas—Soviet physicists Igor Tamm and Andrei Sakharov came up with the toka-
                   began negotiating who would construct                  mak. The name derives from the Russian words for “toroidal chamber in magnetic coils.”
                   which parts of the reactor. By December                     The searingly hot plasma is kept in place by the combined effects of two magnetic fields. The first,
                   2003, China and South Korea had joined the             known as the toroidal field, is generated by vertical magnetic coils ringing the vacuum vessel—in the
                   team, the United States had rejoined, a divi-          case of the International Thermonuclear Experimental Reactor (ITER), 18 of them made from a niobium-
                   sion of labor had been agreed upon, and the            tin superconductor. These create a field that loops horizontally through the tokamak’s “doughnut.”
                   list of sites had been whittled down from four              The second, poloidal field forms vertical loops. It is generated by the plasma flowing around the
                   to two: Rokkasho in northern Japan and                 torus in a current of 15 million amps. This current is itself created by electromagnetic induction: The
                   Cadarache, near Saint Paul lez Durance.                plasma current acts as the secondary windings of a transformer, with superconducting coils in the
                   Politicians gathered in Washington, D.C., to           middle of the torus acting as the primary windings. A rising current in the primary coils induces the
                   close the deal but failed to decide between the        plasma current to flow around the torus.
                   two sites, and the initialing of the agreement              The combined magnetic field carves a slow spiral around the whole of the torus, and plasma par-
                   was put off (Science, 2 January 2004, p. 22).          ticles zip around the ring in tight orbits around the spiraling magnetic field lines. The configuration
                       Acrimonious negotiations continued for             keeps the particles clear of the walls and maintains a pressure in the plasma that is key to fusion.
                   18 months. Finally, in June 2005, a deal was                The tokamak is not the only way to confine a plasma. Physicists are actively pursuing other
                   struck: Japan agreed to support Cadarache,             schemes, such as stellarators and reverse-field pinch machines. But the tokamak is the most success-
                   and in exchange, the E.U. will place some of           ful design so far and forms the basis of ITER and, most likely, the commercial power reactors that will
                   its contracts with Japanese companies and              come after it.                                                                                    –D.C.
                   will share the cost of extra research facilities
                   in Japan (Science, 1 July 2005, p. 28).
                       Since then, negotiators have reworked the
                   international agreement ahead of the signing
                   next month. India has also joined, and key
                   appointments have been made. Kaname
                   Ikeda, a Japanese diplomat with experience
                   of nuclear engineering, will be ITER’s direc-
                   tor general; its principal deputy director gen-
                   eral will be Holtkamp, who managed acceler-                                                                                            In a twist. The huge
                   ator building on the Spallation Neutron                                                                                                superconducting
                   Source at Oak Ridge National Laboratory in                                                                                             magnets required to
                   Tennessee. Six other deputies—one from                                                                                                 contain ITER’s plasma
                   each partner apart from Japan—were                                                                                                     are a major engineer-
                   appointed in July.                                                                                                                     ing challenge.
                       By the end of this year, the ITER organ-

                   ization will employ no more than 200 peo-
                   ple. But across the globe, as many as
                   4000 researchers are already working
                   directly or indirectly on the project, and

                                                          SCIENCE        VOL 314       13 OCTOBER 2006                                                    239
                                                                                         Published by AAAS

        The Long Road to Fusion                                         1968                                        1985
                                                                        Soviet T-3 tokamak,                         Urged by Soviet physicists,
           1951                                                         designed by Igor Tamm                       Mikhail Gorbachev
           Lyman Spitzer founds Princeton
                                                 1958                   and Andrei Sakharov,                        proposes international
                                                Research declassified   makes huge improve-                         fusion project to Ronald
           Plasma Physics Laboratory and
                                                as part of Atoms for    ments in confinement.                       Reagan at Geneva summit.
           begins working on stellarator.
                                                Peace conference in

                        1950                           1960                         1970                           1980                                1990
                                                                                           1976                                                   1988
                                                                                        Joint European     1978
                              Early 1950s                                                                JET given the
                                                                                                                                          ITER conceptual       1997
                              U.S., Soviet, and
                                                                                        Torus (JET)
                                                                                                         go ahead.
                                                                                                                         1983             design work           JET achieves
                                                                                        design work                      JET achieves     begins.               16 megawatts
        1947                  British doughnut-                                         begins.
                              shaped fusion devices,                                                                     first plasma.                          of fusion
        First kiloampere
                              such as ZETA at                                                                                                                   power.
        plasma created

                                                                                                                                                                                                                                                                                                                                        Downloaded from on December 4, 2007
        at Imperial College   Harwell, U.K., fail to                                                                                              1992
        London.               generate fusion. Most                                                                                               ITER engineering
                              work classified.         ZETA                                                              JET                      design work begins.

      structure at the bottom of the tokamak                  bles and strikes the surface, it would set loose     But Kimihiro Ioki, who heads the vacuum
      designed to siphon off waste heat and parti-            tungsten ions. Because tungsten has a high       vessel and blanket division in the existing
      cles that is now considered essential for a             atomic number, once it is stripped of all its    ITER organization, warns that changing the
      tokamak. Researchers worry that if ITER’s               electrons it has a huge positive charge, so      700 square meters of the first wall would be

                                                                                                                                                                               CREDITS: (TOP ROW, L TO R) AP PHOTO; AFP/GETTY IMAGES; (2ND ROW, L TO R) THE SCIENCE MUSEUM/SCIENCE & SOCIETY PICTURE LIBRARY; EFDA-JET; (BOTTOM) ITER
      design is too fixed and the current best con-           even a few ions would severely dilute the        no mean feat. The 421 panels of the main wall
      figuration turns out not to work, they will             plasma. As a compromise, some tokamaks           (excluding the divertor) together weigh more
      have little room to maneuver. “We need to               have experimented with beryllium, the metal      than 1 ton, and technicians would have to
      maximize the flexibility of the machine. It             with the lowest atomic number (4, compared       extract them one by one through a small port
      must give enough information to build a first           with tungsten’s 74). But beryllium has a low     using a many-jointed mechanical arm. Ioki
      electricity-generating reactor. Society can’t           melting point, so it cannot be used in areas     estimates it would take at least a year.
      afford another intermediate                                                                                              To run a power reactor with an
      step,” says IPP’s Harald Bolt.                                                                                       all-tungsten first wall, operators

      Skin deep
                                                                                     “Either [ITER] can would have to be sure that the
                                                                                                                           plasma will behave and not touch
      One area in which researchers                                                  do it, or it can’t.                   the sides. Today, it’s far from clear
      would particularly like some                                                   If it fails, the                      that researchers will be able to
      wiggle room is the lining of
      the inner surface of the vac-                                                  tokamak is out.” guaranteeexperiment you can do.
                                                                                                                           worst lab
                                                                                                                                       that. “A tokamak is the

      uum vessel. This so-called                                                                  —Norbert Holtkamp, It’s an extremely hostile environ-
      first wall must be able to with-                                                        Principal Deputy Director ment, and there are too many
      stand huge heat loads: Parts of                                                                     General, ITER variables. It’s a very diff icult
      the divertor will weather as                                                                                         process to understand,” says
      much as 20 megawatts per                                                       with a large heat load, plasma physicist Steven Lisgo of the Culham
      square meter for up to 20 sec-                                                 it erodes easily, and lab. Phenomena at work in a tokamak range
      onds. In an ideal world,                                                       neutrons can trans- in size from micrometers to meters across,
      researchers would line their                                                   mute it into hydrogen operate over time periods ranging from
      reactor with carbon: It can                                                    or helium.                microseconds to years, and interact in com-
      stand the heat and doesn’t                                                         ITER’s designers plex nonlinear ways. It’s the antithesis of a
      erode and pollute the plasma.                                                  have opted for a nice, clean controlled experiment, and theo-
      But tritium in the fuel readily reacts with             compromise: The f irst wall will be 80% rists still struggle to understand everything
      carbon, and the resulting radioactive hydro-            beryllium with 5% to 7% carbon and about that is going on.
      carbons can be hard to shift. Nuclear licens-           12% tungsten, both concentrated in the
      ing authorities require that all tritium must be        divertor region (see diagram, p. 241). But Feeling the heat
      rigorously accounted for because any                    researchers know that beryllium is just not In the beginning, during the 1950s,
      released in an accident would soon enter the            tough enough for a generating plant and researchers thought it was going to be easy.
      food chain. So tritium retention in the vessel          is highly toxic. “I’m convinced that at a Theorists made a calculation, based on stan-
      is a major worry.                                       late stage we need to convert ITER to full dard random diffusion of particles, about the
          Many believe the answer to be tungsten: It          tungsten coverage to learn if this scenario is transport of energy and particles from the
      has a very high melting point and low ero-              compatible with a power reactor,” says Bolt, burning center of the plasma to the edge and
      sion. The problem is that if the plasma wob-            head of materials research at IPP.               concluded that a fusion reactor would only

240                                          13 OCTOBER 2006             VOL 314      SCIENCE
                                                                               Published by AAAS

                                                                                                                                                                             2001 Revised ITER design completed. Canada joins.

                                                                                                                                                                      1999             2006 ITER implementation agreement signed.
                                                                                                                                                                      from ITER.                    2007                  2016                                              DEMO begins
                                                                                                                                                                                                    ITER construction      ITER completed and first                         operation.         ASDEX Upgrade
                                                                                                                                                                                                    begins.                plasma achieved.

                                                                                                                                                                             2000                             2010                   2020                        2030                         2040                      2050
                                                                                                                                                                          2003                                                              2022                                                          2050
                                                                                                                                                                          China, South                                                      Construction of DEMO, prototype power plant, begins.          First commercial
                                                                                                                                                                          Korea, and U.S.             2005                                                                                                fusion power plant
                                                                                                                                                                          join ITER. Canada           Cadarache                                                                                           begins operation.
                                                                                                                                                                          withdraws. Stale-           chosen as
                                                                                                                                                                          mate over site.             ITER site.

                                                                                                                                                                                                                                                                                                                                              Downloaded from on December 4, 2007
                                                                                                                                                                                                      India joins.                                                 Are we there yet? Political as well as technical trials have
                                                                                                                                                                         1998                                                                                      dogged the footsteps of fusion. This largest of international
                                                                                                                                                                      ITER engineering design completed,but
                                                                                                                                                                                                                                                 Cadarache         collaborations will likely hit some more bumps before it is done.
                                                                                                                                                                      partners ask for a 50% cost cut.

                                                                                                                                                                      need to be a half-meter across. The early            they must squeeze the balloon full of plasma keep plasma pressure confined when they
                                                                                                                                                                      machines revealed, however, that transport           as much as possible, but more pressure were trying to solve a different problem: how
                                                                                                                                                                      was actually five orders of magnitude higher         breeds more instabilities, which ultimately to siphon off waste particles and heat from
                                                                                                                                                                      because of turbulent fluctuations in the             doom the fusion.                                       the edge of the plasma. When a deuterium
                                                                                                                                                                      plasma. “That’s why ITER has an 8-meter                  Researchers are fighting instabilities in a nucleus fuses with a tritium nucleus, they
                                                                                                                                                                      radius,” says Zohm. Ever since those early           number of ways. One is to tweak the distribu- produce a fast-moving helium nucleus, or
                                                                                                                                                                      days, the trend has been toward ever-bigger          tion of the plasma current flowing around the alpha particle, and a speedy neutron. The
                                                                                                                                                                      tokamaks, on the simple understanding that a         tokamak ring. Looking at a cross section of neutrons are unaffected by the tokamak’s
                                                                                                                                                                      larger body “cools” more slowly than a small         the ring, if the heating beams are used to give magnetic fields, so they zip straight out and
                                                                                                                                                                      one. And researchers have been wrestling             the current a boost in one spot here and bury themselves in the surrounding “blanket”
                                                                                                                                                                      with turbulence. “It’s really difficult physics.     another spot there, this can calm instabilities material, where their energy can raise steam
                                                                                                                                                                      We’re only now starting to understand it,”           and allow the plasma to reach a higher pres- to drive an electricity-generating turbine.
                                                                                                                                                                      says Zohm.                                           sure. “You can make very small changes in                  The charged alpha particles are held
                                                                                                                                                                          In the center of the sun, the heat and pres-     the internal current distribution, and instabil- inside the tokamak and heat up the plasma.
                                                                                                                                                                      sure necessary to spark fusion come from the         ities can go away,” says Zohm.                         But once they’ve imparted their energy, these
                                                                                                                                                                      mass of material pressing down on the core.              Another method is to change the shape of alphas become waste and must be removed
                                                                                                                                                                      On Earth, we don’t have the benefit of all that      the plasma. In early tokamaks, the plasma from the plasma before they quell the fusion.
                                                                                                                                                                      gravity, and a tokamak is not a strong vise          was usually circular in                                In the late 1980s, researchers decided to try
                                                                                                                                                                      (see sidebar, p. 239); ITER’s plasma pressure        cross section, but more                                   reshaping the magnetic field toward the
                                                                                                                                                                      will reach only about 5 atmospheres in the           modern machines                                                bottom of the plasma vessel so that
                                                                                                                                                                      center. To compensate, the plasma has to be          have D-shaped plas-                                               some of the outer magnetic sur-
                                                                                                                                                                      very hot, about 100 million kelvin. Researchers      mas or almost trian-                                                  faces, instead of bending round
                                                                                                                                                                      don’t yet know which heating method will             gular ones. That helps                                                   and up again, actually diverge
                                                                                                                                                                                                                                                              Transport barrier
                                                                                                                                                                      work best in a power reactor, so ITER will be        because the magnetic                                                      at the bottom and pass
                                                                                                                                                                      equipped to try several, including electron          s u r f a c e s t h a t yo u                                                through the vessel wall. The
                                                                                                                                                                      beams, ion beams, neutral particle beams,            cross as you move                                                            result is that any particles
                                                                                                                                                                      and microwaves.                                      outward from the cen-                                                         that stray out near the edge
                                                                                                                                                                                                                                                                       Beryllium first wall
                                                                                                                                                                          But as Aladdin discovered, once you fire         ter of the plasma keep                                                        of the plasma eventually
                                                                                                                                                                      up such a genie in a bottle, it’s devilishly hard    changing direction                                                            get swept down to the bot-
                                                                                                                                                                      to control. Researchers have found a plethora        slightly, an effect                                                          tom and dumped into the
                                                                                                                                                                      of instabilities that cause plasma to wobble,        known as “magnetic                                                          divertor, a heat-resistant
                                                                                                                                                                      bulge, vibrate, and generally misbehave. It’s        shear.” “Shear sup-                 Magnetic field lines                   target where particles are
                                                                                                                                                                      like trying to squeeze a balloon full of water:      presses [turbulent]                                                      cooled and then pumped out
                                                                                                                                                                      The more you squeeze, the more it bulges             eddies, and so trans-                    Tungsten                      of the vessel.
                                                                                                                                                                      between your fingers. These troublesome              port is less efficient. It          Carbon
                                                                                                                                                                                                                                                                                                   JET was f irst f itted with a
                                                                                                                                                                      behaviors sport exotic names such as Alfvén          keeps energy in,” says                                           divertor in 1991. The devices are
                                                                                                                                                                      waves, neoclassical tearing modes, sawtooth          Richard Buttery of
                                                                                                                                                                      instabilities, and resonant magnetic perturba-       the Culham lab.                                              Black art. Researchers hope that by nudging,
                                                                                                                                                                      tions. The bind for researchers is that to get           Researchers also            Divertor                    squeezing, and bombarding the plasma, they
                                                                                                                                                                      the most fusion power out of their tokamak,          found another way to                                       can get it to burn hot without instabilities.

                                                                                                                                                                                                             SCIENCE        VOL 314      13 OCTOBER 2006                                                    241
                                                                                                                                                                                                                                            Published by AAAS

      now considered indispensable because they                                                                   former. But the current in the coils can’t
      not only remove waste but also help confine                                                                 keep rising forever, so the length of any
      the plasma. Although researchers don’t yet                                                                  fusion run is limited. The French tokamak
      understand why, these open, diverging mag-                                                                  at Cadarache, Tore Supra, holds the record
      netic surfaces create a “transport barrier”                                                                 with 6-minute pulses.
      inside the bulk of the plasma, near the edge.                                                                   But pulsed operation would put intolera-
      The pressure increases very steeply across                                                                  ble stresses on a power plant that must keep
      this barrier so that the core of the plasma can                                                             working for decades, so researchers are
      be maintained at a significantly higher pres-                                                               looking for other ways to drive the plasma
      sure—a configuration that plasma physicists                                                                 current. Firing the heating beams in a partic-
      call H-mode.                                                                                                ular direction will push plasma around the
          H-mode has been so successful that it is                                                                ring, but this will never provide all the nec-
      now part of ITER’s baseline scenario, but it                                                                essary current. In the 1980s, theorists pre-
      does have a downside: Running the plasma in                                                                 dicted another way: If the pressure gradient
      H-mode can lead to the mother of all instabil-                                                              in the plasma is high enough, particles,
      ities, known as edge-localized modes                                                                        which move through the plasma by spiraling

                                                                                                                                                                                    Downloaded from on December 4, 2007
      (ELMs). These happen because the transport                                                                  around magnetic field lines, will interfere
      barrier doesn’t let out excess energy gradu-                                                                with each other in such a way as to produce
      ally but bottles it up until it’s finally released                                                          a net current around the ring. This “boot-
      all at once. “ELMs are not fully understood.                                                                strap” current was demonstrated in the
      They are bursts of power, like earthquakes,”                                                                1990s, and the Asdex Upgrade, for example,
      says Jerome Pamela, head of the JET project.                                                                has produced as much as 30% to 40% of its
      ELMs can damage the first wall or send a                                                                    current from the bootstrap effect.
      blast of energy down to the divertor. Few                                                                       Getting more bootstrap is hard because
      believe that they will be able to banish ELMs                                                               of the usual problem: It needs higher pres-
      altogether, but if they can be made small and        Testing, testing. Engineers have already built         sure gradients in the plasma, which mean
      regular, they are manageable. “The name of           pieces of ITER, such as this slice of vacuum vessel,   more instabilities. Nevertheless, once
      the game is to let the energy out smoothly,”         to test construction.                                  ITER has demonstrated its baseline sce-
      says Buttery.                                                                                               nario, researchers will be aiming for an
          Such is the value of H-mode that even at            Even if one of these techniques does tame           “advanced” scenario in which the induction
      this late stage ITER’s designers are consider-       ELMs, no one knows what will happen when               coils are switched off and 80% to 90% of
      ing design changes to cope with ELMs. One            ITER’s self-heating regime kicks in. The fast-         the plasma current is generated by boot-
      scheme investigated at JET involves injecting        moving alpha particles created by fusion will          strap with the remaining push provided by
      impurities such as nitrogen into the transport       have much more energy than the bulk of the             heating beams. “At the very least, we will
      barrier to make it a bit more leaky, but this        particles in the plasma, and these could open          want long pulses,” says Hor ton. But
      also degrades H-mode, so it is not popular.          up a whole hornets’ nest. “This is the first           researchers don’t expect the advanced sce-
      Another tactic, tested at IPP with its Asdex         time a plasma has been heated by alphas. It            nario to be easy. “It will be a real pain to get
      Upgrade tokamak, is to regularly fire pellets        could create new instabilities. Experts don’t          to this,” says Zohm.
      of frozen deuterium into the barrier. This           think it will, but we cannot logically exclude
      sparks an ELM every time, keeping them               that possibility,” says Llewellyn-Smith.               Ready when you are
      steady and small. This system “will be               “That’s why we need ITER,” adds Zohm.                  With a total price tag of about $12 billion,
      installed” on ITER, says Chuyanov, “but is it        “We can’t simulate internal heating. It’s the          ITER is the most expensive experiment in
      enough? We don’t know.”                              part we know least about.”                             the world apart from the international space
          A late entrant into the race is a system                                                                station. Some plasma physicists are skepti-
      developed in the United States using the             Seeking steady state                                   cal that fusion will ever be a power source on
      DIII-D tokamak at General Atomics in San             Although there may be surprises along the              Earth and argue that we shouldn’t be wasting
      Diego, California. Extra magnetic coils              way and whole new scenarios may have to                our money on ITER. After 50 years of
      added to the tokamak create a sort of chaotic        be developed, few doubt that ITER will                 research, even fusion’s flag-wavers concede
      static in the transport barrier, making it           reach its goal of generating large amounts             that it may still be another half-century until
      leaky enough to avoid large ELMs. “It’s              of excess power. But power is not much use             we have a workable fusion power plant, but
      much simpler than pellets, more reliable,”           commercially in bursts a few minutes long              ITER researchers are undaunted. “By the
      says Pamela. The problem is where to put the         followed by a long wait while the reactor is           middle of the century, we’ll know how to do
      coils. Ideally, they would be inside the reac-       reconfigured. Tokamaks are by their nature             it. Then it’s up to the world community to
      tor vessel, close to the plasma, but that sort       pulsed devices. Some of the magnetic field             decide if they want it,” says Zohm. Soviet
      of reconfiguration would be one step too far         that conf ines the plasma is provided by               fusion pioneer Lev Artsimovich, speaking
      for ITER’s designers. “We’re working very            plasma particles flowing around the toka-              more than 3 decades ago, had the same mes-
      actively to find a solution for ITER, but it’s       mak—a current of some 15 million amps.                 sage. Asked when fusion power would be
      impossible to put the coils inside,” says            This current is induced by a rising current            available, he answered, “Fusion will be
      Chuyanov. Researchers at JET are consider-           in coils in the central hole of the tokamak            ready when society needs it.” That time may
                                                                                                                                                                     CREDIT: ITER

      ing fitting them outside their vessel to see         ring, the coils and plasma acting like the             be fast approaching.
      whether that might work for ITER.                    primary and secondary windings of a trans-                                            –DANIEL CLERY

242                                         13 OCTOBER 2006           VOL 314       SCIENCE
                                                                            Published by AAAS

To top