The goal of fusion research is a "burning plasma" - a fully ionised
gas self-sustained in an extreme state by power released from the fusion reactions
of its atomic nuclei. The burning plasma would then provide a new powerful,
clean and safe source of energy. To achieve this, we need to overcome two
major challenges. Firstly, to ignite the plasma, temperatures in the order
of hundreds of millions of degrees centigrade (ten times the temperature of the sun - Fig. 1) must be attained i.e. the
plasma must be heated sufficiently. The second, more difficult challenge, is to sustain the plasma at these temperatures by confining and controlling it in order to maintain its density and ensure that it does not suffer excessive heat losses.
Tokamaks (a family of fusion research devices, which
includes both JET and the future burning plasma experiment ITER) utilise an ingenious
scheme that addresses both challenges at the same time: a huge electric
current is induced in the plasma to heat it and to complement the confining
magnetic field. The electric current produces heat due to the 'Joule Effect', a
phenomenon familiar to us in everyday items such as electric ovens, irons or light bulbs. In
these household appliances the electric current usually does not exceed a
few Amperes. Electric currents can also produce strong magnetic fields,
an effect which is used in, for example, magnetic cranes, and in fact in all electric motors (Fig. 2).
Hundreds or even thousands of Amperes of electric current can flow in
industrial electromagnets. However, in a large tokamak like JET, we may induce millions
of Amperes into a plasma in order to heat and confine it.
Ohmic Heating
The tokamak concept is a breakthrough in plasma
research, but not a complete solution. At millions of degrees and above,
plasma is conducting electricity far too well, with very little resistance - which also means with not enough heat produced by the Joule Effect.
The unit of electric resistance is the Ohm, so plasma physicists usually say
'Ohmic heating is ineffective at high temperatures' where the word 'high' refers to the hundreds of millions of degrees required for burning plasmas. In
order to attain the target temperatures some sort of 'additional
heating' is required to supplement the 'Ohmic heating' (as a
matter of fact, the 'additional heating' eventually plays a dominant
role). Neutral particle beams ('NB Heating') and resonant electromagnetic waves ('RF Heating') can do
this job (Fig. 3).
Furthermore, tokamaks cannot maintain a continuous electrical current in the plasma and this limits the concept of complementing the magnetic
field (Fig. 4). Tokamaks have a transformer-like electrical setup, with plasma that
acts as a single secondary loop - and no transformer can provide
continuous direct electric current in its secondary circuit. An additional 'current drive' needs to be provided if we wish to confine burning
plasma continuously. Electromagnetic wave current drive offers a possible
solution.
Nowadays at JET, the role of the additional heating and current drive
systems has been considerably broadened compared to their original
tasks. Neutral beams and resonant electromagnetic waves are at present the
key tools in optimising the plasma performance.
The most important spatial characteristics of a tokamak plasma are its
"profiles" which show how physical quantities changes along the plasma radius,
from the plasma centre to the plasma edge. For example, we measure and
study plasma temperature profile, plasma density profile, magnetic field
profile etc. These days, neutral beams as well as electromagnetic waves are used to control and modify the plasma profiles by proper targeting of the
additional energy deposition. This technique is sometimes referred to as
"plasma tailoring" and proves to be extremely efficient in improving
plasma performance. The technique can also create completely new
regime conditions, for example by generating a so called "Internal Transport Barrier" which provides improved plasma confinement.
The heating and current drive facilities have an even greater mission when applied as "actuators" (acting powers) in JET's Real Time Control. Powerful actuators can be used to automatically counteract plasma instabilities or to safeguard an intended change in plasma parameters, see Focus On : Real Time Control of JET Plasmas. In this respect, additional heating and current drive will almost certainly be used in future reactors with burning plasmas.
Neutral Beam Injection (NBI)
A widespread technique for additional plasma heating is based on
the injection of powerful beams of neutral atoms into ohmically pre-heated
plasma (see Fig. 5 & 6). The beam atoms carry a large uni-directional kinetic (motional) energy. In the plasma, beam atoms lose electrons due to collisions,
i.e. they get ionised (electrically charged) and as a consequence are captured
by the magnetic field of the tokamak. These new ions are much faster then
average plasma particles. In a series of subsequent ion-ion, ion-electron
and electron-electron collisions, the group velocity of beam atoms is
transferred into an increased mean velocity of the chaotic motion of all plasma
particles. The action is similar to the opening break in the game of pool,
when a fast motion of one ball can cause the seemingly chaotic motion of
all balls. However, the world of plasma particles is
inconceivably small, and many billions of particles are in play. We sense an increase in their chaotic motion as an increase in
temperature. In other words, a neutral beam heats the plasma - and that is
what we desire!
In fusion experiments, the neutral beams are usually
formed by atoms of hydrogen isotopes (hydrogen, deuterium or even tritium
at JET). Notice that we always speak about a 'Neutral beam' and its
'atoms'. Indeed, the beam needs to consist of neutral atoms (as opposed to
electrically charged ions) otherwise it could not penetrate the strong
magnetic field that confines fully ionised plasmas. The energy of the beam
(corresponding to the velocity of its atoms) must be sufficient to reach
the plasma centre - if the beam atoms were too slow, they would get
ionised immediately at the plasma edge. At the same time, the beam is supposed
to have power sufficient to deliver significant amounts of fast atoms into
the plasma, otherwise the heating effect would not be noticeable. At JET, the
beam energy is 80 or 140 keV, corresponding in the case of a deuterium beam to
2800 or 3600 km/s which is approximately five times faster than
the mean velocity of the ions in a JET deuterium plasma. The total power of beam heating at
JET is as much as 23 MW (million Watts). With this power, the number of beam
atoms per second corresponds approximately to 10% of the total number of JET
plasma ions.
It is not at all straightforward to generate powerful
neutral beams of very fast atoms (see Fig. 7). The only way to form the neutral beam is
to produce large amounts of ions first, then to accelerate the ions in a
strong high-voltage electric field and finally to neutralise the
accelerated beam (Fig. 8). The accelerated ions get neutralised in charge-exchange
interactions with a gas cloud, however, some leave the cloud still in a
charged state. These residual fast ions must be deflected by a dedicated
electromagnet to a cooled ion dump that can withstand heavy ion
bombardment. Last but not least, powerful vacuum pumping must assure that
practically no slow atoms from the neutralising gas cloud can diffuse as
far as to the plasma chamber, so that the fast neutral atoms have free access
to burst into the plasma. This technology works well but is still being refined in order to increase the
reliability, purity and efficiency of the neutral
beam.
Plasma and electromagnetic waves
Plasma is an intriguing state of matter. Being formed by charged
particles (ions and electrons) it is affected by long-range electric and
magnetic forces. As a consequence, plasma - and specifically magnetically
confined plasma - can host an extremely rich mix of oscillations and
plasma waves, covering sound, electrostatic, magnetic and electromagnetic
waves. Depending on local plasma parameters, plasma waves can propagate,
get dumped (absorbed), be reflected or even converted to different plasma
waves.
In general, plasma waves carry energy, so that wave absorption involves
energy transfer (compare Fig. 9 & 10). Their energy is then, in most cases, converted to an increased
mean velocity of the chaotic motion of particles, i.e. to a higher temperature
of the absorbing medium. Wave absorption is extremely efficient if the wave frequency is resonant with some of the fundamental oscillations
of the medium. However, significant heating can occur even at non-resonant frequencies - witness the widespread everyday use in microwave ovens where
magnetron devices produce electromagnetic waves which heat by cyclically turning over the water molecules in food, rather than resonating with them.
Ion Cyclotron Resonant Heating (ICRH, also known as RF Heating)
In magnetically confined plasmas, particles (ions and electrons) rotate
around magnetic field lines (Fig. 11) with frequencies that depend only on three
quantities: charge and mass of the particle, and magnetic field strength.
Other parameters like temperature or density play no role at this
'cyclotron' frequency. Therefore, if an electromagnetic wave with
cyclotron resonant frequency is launched into the plasma, all the targeted
particles (defined by mass and charge) are heated, provided that the
magnetic field complements the resonant condition. In tokamaks, the
magnetic field decreases with distance from the tokamak major axis. This
allocates the resonant region to a narrow vertical layer, thus giving us a
simple control over deposition of the cyclotron resonant
wave.
To accommodate complicated wave propagation rules, multiples of the base cyclotron frequency, called 'higher harmonics' are mostly applied in practice. The effect of higher harmonic resonance relies on space variations in the wave intensity, so that such a resonance is stronger for particles with larger orbits. That is, higher harmonic heating is more significant for fast particles than for slow particles, which introduces temperature dependancies as well as distortion in thermal distribution due to the heating.
Ion cyclotron resonant heating (ICRH) is routinely applied on JET. It is resonant with the second harmonic (i.e. double) frequency of ion gyration of main JET plasma ions (deuterium) or with a base frequency of gyration of a minority species (e.g. tritium, helium...). The available resonant frequencies at JET are in the range of 23-57 MHz (megahertz, or million of oscillations per seconds) which correspond to the length of the vacuum electromagnetic wave from 13m (at 23 MHz) down to 5m (at 57 MHz). This is a "shortwave" frequency, which is not very popular in the air due to many fades, blackouts and interferences (FM radio frequencies are just a bit higher, around 100 MHz). In total, the installed power of the JET ICRH system is as much as 32 MW (megawatts = million watts), and in practice only part of this potential is sufficient for the JET experiments. This is a huge power compared to radio or TV broadcast, where a 50 kW (kilowatts = thousand watts) transmitter is already considered powerful.
Figure 12 show a schematic of the route taken by the wave from the point of its excitation to its launch into the plasma. Amplifier chains generate the ICRH electromagnetic waves, each chain with a powerful (2MW output) tetrode tube in its final stage (Fig. 13). The transmission lines which conduct ICRH waves from the generators to the JET tokamak are low loss coaxial cables which consist of a conducting outer metal tube with an insulated central conducting core, with overall insulation. Such cables are generally used in any high-frequency transmission - e.g. the signal from a TV aerial or satellite dish is transferred to the TV set by a coaxial cable. However at JET, due to high powers involved, the ICRH output coaxial cables look rather like 'pipelines' with a 20cm diameter overall diameter. Several hundred metres of these transmission lines are installed at JET. The transmission lines terminate in 4 ICRH antennas that are installed within the JET inner wall and are slotted in the front. Each antenna consists of four conductors called straps (Fig. 14), and each strap is fed by a separate generator. The ICRH electromagnetic waves cannot propagate in the JET vessel vacuum (their wavelength being too long) so the antenna must be as close to the plasma as possible.
For information, there is a similar technique called electron cyclotron resonant heating (ECRH) but we do not use it at JET. The principle is based on the fact that electrons, being several thousand times lighter than ions, have much higher cyclotron frequencies. In tokamak plasmas the required ECRH frequencies are in the order of 100 GHz (gigahertz = billions of cycles per second, corresponding to vacuum wavelength in the order of a few millimeters only) which is more challenging for the wave generation and transmission. These frequencies are also used in some modern radar applications. However, the power of radar devices is negligible compared to ECRH requirements. ECRH targets plasma electrons only, and the heat transfer from electron to ions is relatively slow. The advantages of ECRH are that its waves can propagate in vacuum and that they can be steered with high precision. ECRH is installed on some other large tokamaks such as JT-60U in Japan, DIII-D in USA and ASDEX-U in Germany.
Lower Hybrid Current Drive (LHCD)
There are many other resonant frequencies in tokamak plasmas but experiments have found some to be inefficient or impractical while others simply cannot penetrate through the plasma edge region. Two of the candidate frequencies are "hybrid", so called because they result from force interplay between ions and electrons, so that their frequencies lie between ion cyclotron and electron cyclotron ones. Although the lower hybrid frequency can get into the plasma, unfortunately it has an inefficient heating effect. Nevertheless another significant application of lower hybrid frequency has evolved: the corresponding lower hybrid wave can drive electric current thanks to the fact that it has an electric component parallel to magnetic field lines.
One would perhaps expect that the very rapidly alternating electric field of electromagnetic waves could not generate a constant electric current, but this common sense proves to be false. Plasma electrons with thermal velocities slightly slower than the wave propagation velocity can actually "surf" on the uprising electric potential and thus increase their velocity in the direction of the wave (Fig. 15). It is also true that any electrons which are slightly faster than the wave will be slowed down. However, the thermal distribution of velocities causes there to be fewer faster particles (Fig. 16). Consequently, there are more electrons which are accelerated rather than decelerated so that in total a net electric current appears. Though the effect looks minute on the electron velocity distribution, in terms of electric drag it is significant.
At JET, the Lower Hybrid Current Drive system (Fig. 17) works at a frequency of 3.7 GHz (gigahertz = billions of cycles per second) which corresponds to a wavelength of 0.1 m in vacuum. The frequency belongs to the "L-band" used (for example) by satellite broadcasting. The LHCD installed capacity at JET is 12 MW (million watts) of additional power. Thanks to this system, electric current of several MA (million Amperes) can be driven. The electromagnetic wave is generated in klystrons - tubes that can produce the above frequencies by resonant modulation of an electron beam. At JET, 24 klystrons are installed in 6 independent modules. The electromagnetic wave is then transmitted to the LHCD antenna by a complex system of waveguides (Fig. 18). Waveguides are hollow rectangular metallic conductors with a cross-section size that corresponds to the transmitted wavelength. The LHCD antenna is of a very sophisticated design, called a "multijunction grill" in order to allow for a correct phasing of the wave before it is launched into the plasma (Fig. 19). The correct phasing of LHCD waves is hampered by propagation in a vacuum, therefore it is required that the LHCD antenna is mounted directly in the JET inner wall, as close to the plasma as possible.
For the first burning experimental reactor, ITER, a complete portfolio of all efficient methods of plasma heating and current drive is likely to be adopted, with an expected total output power of over 100 MW. JET is the closest tokamak in plasma size and shape to ITER, so it is natural that JET's heating and current drive facilities are widely involved in experiments relevant to ITER. The proposed ITER "plasma scenarios" are optimised on JET by accurate profile tailoring as explained in the introduction. In 2007 a new "ITER-like" ICRH antenna is to be installed as a major JET enhancement in order to validate the new concept of robust, stable ion cyclotron wave emission suitable for the harsh conditions of the future burning plasmas of ITER. |
Fig. 1 Nuclear fusion is the driving force of all stars including our Sun
Fig. 2 Electricity can produce heat and magnetic forces
Fig. 3 Heating of JET plasmas
Fig. 4 Magnetic fields in a tokamak - the toroidal field is generated by external coils, poloidal by electric current in the plasma
Fig. 5 One of two identical NBI systems at JET
Fig. 6 Installation of one of JET's NBI systems
Fig. 7 Scheme of the NBI principle: ions in red, neutral atoms in green
Fig. 8 Assembly of one of the sixteen ion sources of JET's NBI system
Fig. 9 & 10 Leaves and solar panels convert the energy from electromagnetic waves into other forms of energy
Fig. 11 Cyclotron motion of a plasma ion around a magnetic field line
Fig. 12 Schematic of the ICRH system at JET
Fig. 13 ICRH wave generators at JET
Fig. 14 LHCD and ICRH antennas in JET vaccum vessel (LHCD grill in frame in left, next to it four slotted launchers of ICRH
Fig. 15 A charged particle can increase its velocity by "surfing" on an electromagnetic wave
Fig. 16 As a consequence of the particles "surfing" on the wave, the thermal distribution of the particles changes
as the red line shows. The assymetry in velocity distribution causes a net electric current to appear
Fig. 17 Schematic of the JET LHCD system
Fig. 18 Connection of the LHCD waveguides to the JET vacuum vessel
Fig. 19 Completion of the LHCD antenna known as "multijunction grill"
|
