Focus On : Plasma Edge

Introduction

In order to protect the vital fusion processes in a burning plasma from the cold reactor components (and vice versa), considerable effort is invested into researching the plasma edge.

To obtain energy from fusion reactions sufficient to recover the energy spent in keeping these reactions alive and, in a not too far future, even produce net electrical energy, temperatures of hundreds of millions of degrees must be well confined at a sufficiently high fuel density. At the Joint European Torus (JET), these conditions have nearly been achieved. However the fusion power is still lower than the heating power required for operation.

JET is an experimental tokamak in which an extremely hot gas (called plasma, the fourth state of matter- see Fig. 1) is confined by strong magnetic fields in a torus - a doughnut-shaped chamber. The fusion processes take place inside the doughnut - in the so called plasma core, and the heat of this extremely hot gas needs to be well confined and isolated from the vessel in which it is placed. At JET, the distance from the plasma core (at hundreds of millions degrees Centigrade) to the first wall (i.e. to the plasma-facing tiles at several hundreds of degrees) is about one metre. In order to increase the volume in which fusion reactions take place the geometry of the plasma has been designed in such a way that in reality most of the temperature drop from the core plasma to the walls of the vessel occurs over the last few centimetres. This means that in this region the temperature may decrease by several tens of million degrees per centimetre. By comparison, the gradient within a candle flame, from wick (cool) to flame outer (hot), is of the order one thousand degrees per centimetre.

The plasma is composed of electrically charged particles, electrons and ions. Such electrically charged particles have the natural property of following magnetic field lines as shown in Fig. 2. The field lines may be imagined as strings along and around which charged particles move. Some of these magnetic field lines intersect the solid materials of the vessel at some location. Charged plasma particles that happen to be on such field lines are therefore guided into collisions with the first wall and deposit their energy onto the plasma-facing material. In plasma physics, such a terminal field line is called "open". Open field lines are found at the edge of the plasma close to the walls. In contrast, deeper inside the JET torus, the field lines run around the "doughnut" in never-ending loops, without ever encountering any solid material. They form so called "closed" field lines. In an idealized scenario, plasma particles are safe from collisions with the first wall as long as they are guided along these closed field lines - see Fig. 3. However, there are processes that force plasma particles to leak out from the confined volume, which is the volume entirely filled by closed field lines: particles diffuse across the magnetic field. As can be seen from Fig. 2, particles may leave the confined volume simply due to the fact that their orbit around each field line has a finite radius. Furthermore they can "jump" from one guiding field line to another due to collisions with other plasma particles (Fig. 2) or due to fluctuating electric fields causing so called turbulent cross-field transport (turbulent transport is a hot research topic across different disciplines in physics, see Focus On: Computer Modelling of Fusion Plasmas ).

Because of complex instabilities, in addition to the above mentioned mechanisms, plasma particles can be ejected out of the region of closed field lines in big quantities during bursts, commonly called Edge Localised Modes (ELMs) (see article). However, as the mechanisms that are responsible for transport across the field lines are slower than the particle mobility along the lines, the torus, with its doughnut-shaped geometry and closed field lines, is currently the most successful design for plasma confinement.

Introducing the plasma edge

It may seem ideal to totally avoid the presence of open field lines by building a fusion reactor with a wall perfectly aligned with the closed magnetic field lines. As shown previously, transport across the magnetic field exists by default and this therefore wouldn't prevent plasma reaching these walls. Furthermore, such a simple solution is technologically not realistic. Any imperfection in the shape of the plasma facing wall and/or in the geometry of the magnetic field lines would result in some field lines intersecting the solid walls - they'd break "open". In addition, as the plasma itself contributes to the establishment of the total magnetic field (due to its nature of comprising moving charged particles) it is not possible to keep the field geometry under stringent control. Not to mention that one wants to diagnose the plasma and needs observation ports along the wall. In brief, plasma particles eventually collide with the first wall at arbitrary locations and the particle and energy confinement inside the closed field lines is not perfect. In order to overcome this problem the researchers have opted for a design with a set of well defined open field lines between the first wall and the confined plasma. This constraint resulted in the constitution of a whole branch of fusion research, the physics of the plasma edge, which studies phenomena related to the existence of the open field lines at the edge of the confined plasma volume.

The edge of the plasma is a region between solid materials in the vessel walls, and the main plasma volume, called the core region with closed field lines. In a fusion reactor the plasma edge may be imagined as a protective skin: its properties control the power and particle exchange between the burning plasma (the plasma core) and the vessel walls. It must be pointed out that there is a strong interplay between the behaviour of the plasma edge and the interactions of this plasma with the first wall. They affect each other through various processes that occur on the walls themselves due to contact with the plasma or inside the volume of the plasma edge as a result of these plasma-wall interactions. See plasma-wall interaction (below).

The concept of limiters and divertors

As we have learned so far, particles are confined to a certain degree within the volume composed of closed field lines. Those that escape this region are called plasma exhaust. The border of the confined region is known as the Last Closed Flux Surface (LCFS) or separatrix, while the term Scrape-Off Layer (SOL) designates a narrow region (usually only a few cm wide) outside this border. The SOL may be imagined as the region where the plasma is essentially scraped off from the core plasma. Here the magnetic field lines are open, and direct the plasma exhaust into a defined region where the exhaust particles are allowed to collide with the wall and much colder neutral gas (see also divertor detachment).

There are two ways by which the last closed field line can be delimited, see Fig. 4. In the simplest and historically earlier option the confined region is "limited" by inserting a barrier a few cm into the plasma. This is called a limiter and essentially it was there to protect the walls from the hot core plasma. Though successful to some extent, it had two major disadvantages: Firstly, material released by impact of the plasma on the limiter could penetrate straight into the core and degrade its properties. Secondly, in a reactor it would not be possible to pump away the "ash" (Helium resulting from fusion reactions and diluting the core plasma) in a sufficiently efficient manner.

Therefore a more sophisticated solution was developed about 20 years ago, using a modification of the magnetic field lines at the plasma edge, so that the field lines of the SOL are diverted into a dedicated region where the plasma exhaust ends up in collisions with the wall (the target plates) or with gas. This is shown in Fig. 4 (right hand side) and Fig. 5, with the diversion of the field lines at the bottom. This latter configuration, called a divertor, has proven in experiments to be significantly more advantageous.

Pros and cons of the divertor concept

Originally the main purpose of limiters and divertors was to separate plasma from the first wall and improve the performance of the tokamak. Particles enter the SOL only by cross-field transport which, as we have learned, is small compared to transport along the field lines. Therefore as a particle moves radially outward from the SOL towards the wall the number of particles "running" along each field line diminishes, as there are less and less that can diffuse into the field lines radially, being transported away along the field to the "targets". In its simplest form this results in an exponential decay of the temperature and density of the plasma. The result of this is that the heat and particle fluxes onto the walls become sustainable for the wall materials. (Similar conditions exist in neon light bulbs, which everybody knows work well.) Most particles and nearly all of the power entering the SOL are immediately guided along the open magnetic field lines to the limiter or the targets of the divertor. Wherever the plasma impinges onto material wall surfaces impurities from these walls are released.

However, divertors have several important advantages over limiters:

• The materials facing the exhaust plasma are not in any direct contact with the main (confined) plasma. Consequently tokamaks with divertor plasmas have lower levels of impurities in the core plasma. As a result they tend to achieve much higher temperatures in the core, increasing the probability for fusion reactions.

• So called high confinement modes (or H-modes) can be achieved nearly exclusively in the presence of divertors. In the H-mode a barrier against cross-field transport is created that significantly reduces the diffusion of particles into the open field lines thereby increasing the density and temperature of the core plasma. The H-mode was discovered by pure serendipity in 1982, while operating the German ASDEX tokamak with a divertor configuration. See our World Year of Physics article. Despite much progress having been made over the past two decades in the description of the H-mode, an understanding of the basic mechanisms that lead to an H-mode, which are likely to include phenomena of the edge plasma, still remains unclear and is a major topic of fusion research around the world.

• As will be explained below, the path of the exhaust particles along a field line from when they enter the SOL along the separatrix to the divertor targets, known as the connection length, can be very long (~30 m in JET and ~150 m in ITER). Depending on the plasma conditions along the separatrix this can be long enough for the plasma to cool down so far that the plasma electrons and ions recombine to neutral atoms before even reaching the solid surfaces. These neutral particles create a "cloud" of gas in the divertor region (see "divertor detachment)".

• With a neutral gas developing in the divertor region, high enough gas pressures can be achieved such that pumps (at JET, powerful divertor cryopumps, see Focus On: Pumping Systems) are able to remove the now cold plasma exhaust from the tokamak. Such removal of the exhaust is crucial for the functioning of a reactor as this exhaust contains the fusion "ash" Helium, which, if not pumped away, would dilute the fuel to such an extent that a burning plasma would not be sustainable anymore.

As the most expensive units of a tokamak are the magnetic field coils, the volume inside such coils in which fusion reactions can occur needs to be maximised to make fusion cost efficient. In this respect the introduction of a volume for the divertor in which no fusion reactions can occur is a major drawback as it increases the costs for a reactor compared to a limiter (but at least it works!).

Why is the connection length so long?

For reasons of stability, in a tokamak the closed magnetic field lines (i.e. field lines entirely immersed in the plasma, see above) do not wind around the torus in simple circles. Instead they have to be imagined as follows: each field line is a very long and thin string that covers, at certain distance from the plasma centre, the surface of the doughnut entirely. This is achieved by tilting the field lines by a small angle as shown in Fig. 3. This angle is called the pitch, Fig. 6. When the pitch angle is set properly, the field line winds around the doughnut without ever reaching its point of departure, so that after many revolutions a single field line can cover almost the entire doughnut-shaped surface. Open field lines (i.e. magnetic field lines in the SOL) must be tilted as well, indeed, the pitch angle cannot change abruptly.  Therefore, each particle entering the SOL may have to do several revolutions around the torus before reaching the targets of the divertor. The path is longest for particles residing in the immediate proximity of the separatrix. In fact the X-point in Fig. 5 represents the perpendicular projection of a field line with a zero pitch (a horizontal circular field line) so that charged particles in its vicinity follow trajectories with an extremely low pitch. Therefore these particles can undergo many collisions before reaching the target plates, strongly altering the plasma properties along these field lines.

Modes of Divertor Operation

• Sheath Limited. Let us define the region of the SOL adjacent to the confined region upstream and the target plates downstream (following the flux of plasma exhaust along the open magnetic field lines). When the connection length between the upstream and downstream locations is rather short and/or the plasma density in the SOL is low (e.g. when the core plasma density is low), then the temperature drop along a field line is negligible. In this case, all the power entering the SOL reaches the solid surfaces, namely the divertor target plates. The power deposition is highly localised close to the divertor strike points (i.e., the intersection of the separatrix with the divertor plates, see Fig. 5).

  When plasma is in contact with a solid surface then a "Debye sheath" forms. Heat transfer across this sheath between the plasma and the wall must be proportional to the product of the particle flux and plasma temperature. Therefore the heat that can be transported along the field lines is limited by the heat that can cross the sheath. That is why this regime is named "sheath limited". As a consequence, the plasma temperature and particle flux in front of the target will increase until the sheath can transport the entire power that enters the SOL, resulting in high temperatures and heat loads on the target. In practice this mechanism excludes the Sheath limited regime from being relevant for a future fusion reactor.

• High Recycling. When the density in the SOL is increased then the plasma flux to the targets increases. As charged particles of a plasma impinge on the solid first wall they recombine on its surface or inside the bulk material to form neutrals which may subsequently be released from the first wall back into the plasma. This process is called recycling. With increasing plasma density more and more neutrals are released from the target plates and penetrate into the SOL – the recycling becomes high. The neutrals are ionized in the plasma of the SOL which removes energy from the SOL in the volume not far from the target. Consequently the temperature along the field line drops. The temperature is further decreased if impurities are present in the SOL, enhancing radiative losses and so cooling down the divertor plasma. The difference between upstream and downstream temperatures may be also increased by extending the connection length between the two regions.
  
  In the high recycling regime the pressure, being proportional to the product of plasma temperature and density, remains constant along any given magnetic field line connecting upstream and downstream locations. That is, the density of the plasma gets high in front of the targets where the temperature is low. If the plasma and the target material are composed of chemically active species, these may even react with each other leading to the further release of impurities into the SOL, enhancing the above cycle with increasing core plasma density.

  The volumetric radiation in front of the targets reduces the power flux to the material surfaces and thereby increases the lifetime of the divertor plates - see Fig 7. With the neutrals being ionized in front of the targets, the main particle sources of the SOL are now the target plates and no longer the particle flux across the separatrix as it would be in the case of the sheath limited regime – the difference denotes the high recycling regime. However, the only source of power for the SOL remains the plasma power losses across the separatrix!

• Divertor Detachment. With further increase of the plasma density the amount of charged particles that reach the divertor plates falls to negligible levels. As the density is increased more impurities are released by plasma facing components that raise the radiation levels. For tokamaks where the walls of the divertor are made of materials that do not radiate efficiently enough, impurities can be puffed i.e. "seeded" into the divertor for obtaining the required radiation and thus cooling of the divertor volume. This method is known as impurity seeding. As the temperature in the divertor decreases over a large volume, electrons and ions can recombine to form neutrals volumetrically. This process is amplified by the presence of those neutrals that, recycled at solid surfaces, now act as a "break" for the plasma that flows towards the targets through friction. They increase the time that the charged particles have for recombining, making this process more likely to happen. When this occurs in large quantities the measured particle flux at the target plates drops by more than an order of magnitude. Neutral atoms transport the residual power and as they are not bound by magnetic field lines, they can deposit power and particles over broad areas reducing the peak values to acceptable levels for materials to sustain the bombardment - see Fig. 7.

  This regime is known as the plasma (or divertor) detachment as ideally the plasma becomes completely detached (separated) from any solid surface. Plasma detachment allows higher operating temperatures upstream. Due to the high neutral particle densities/pressures established in the divertor volume in front of the pump ducts, the pumping of the helium ash becomes more efficient. And due to the negligible plasma influx onto material walls the production of impurities may be reduced.

Expected divertor operation in fusion reactors

As we have seen, divertor detachment is very advantageous for handling the exhaust power and fusion ash, sparing the divertor targets from unacceptable localised power loads and removing the Helium exhaust. However, in experiments it can sometimes be challenging to stabilize detachment on both targets when plasmas become fully detached. Here fully detached means that the plasma detaches along the entire target surface. Under these conditions large volumes of the divertor are cold and neutrals have a long mean free path allowing them to penetrate into the confined region. The influx of neutral particles and, in particular, impurities into the confined plasma causes high radiation levels from this region, which may result in the thermal instability of the whole plasma. The phenomenon that leads to such instabilities is known as MARFE (Multi-faceted Asymmetric Radiation From the Edge) and needs to be avoided - see Figs 8 and 9.

In current normal tokamak operation (and also for the future ITER machine) it is planned to run the divertor in the so called partially detached regime, Fig. 10. In this regime usually the plasma at the inner target (the target on the left side of the images at smaller radii) is still completely detached; whilst at the outer (right) target it is only partially detached. This means that it is not detached along the entire target but only in those regions where the connection length is longest, thus close to the strike point. Further out it remains in the high recycling regime such that neutrals cannot leak in large quantities into the SOL outside of the divertor. It has been found and extrapolated for ITER that such a degree of detachment is sufficient for handling the power load and it is best for the performance of the SOL. It also reduces the risk for a MARFE, by limiting the size of the cold cloud of neutrals in front of the target plates, as shown in Fig. 10. Computer models are used to predict and interpret the complex behaviour of the SOL and divertor plasma, thus also those of JET and ITER. These models are very complex and include many different processes. Whilst providing a reasonably reliable qualitative interpretation of the plasma edge behaviour they often fail to predict dependable numerical values (see Focus On: Computer Modelling of Fusion Plasmas). For example, it is not definite whether ITER will achieve plasma detachment naturally or whether impurity seeding will be required. It is thus one of the tasks of the researchers concerned with edge physics to further develop the models for the edge and improve their accuracy against existing experiments such as JET.

Plasma-wall interaction

In magnetic confinement fusion devices, plasma facing components are subject to heat and particle fluxes that strike the first wall either continuously or in bursts. The effect on the wall surface is usually tolerable in present facilities but in future fusion power reactors the power load will be much higher and the duration of the plasma discharges much longer (current machines a few tens of seconds, future machines several tens of minutes if not continuous). The potential scale of the damage to the first wall challenges fusion research and technology, particularly for the development of the divertor. Even when most of the power of the plasma is exhausted in volumetric processes, some plasma facing components will have to withstand peak temperatures of more than 1000 degrees, despite being actively cooled!

The particle fluxes and heat fluxes onto solid surfaces lead to erosion and release surface material into the plasma where it acts as an impurity. Some of the released impurities can migrate to very remote locations inside the machine before they stick to a plasma facing component so forming a layer of an amorphous material. The studies of the processes responsible for erosion, migration and deposition of materials in fusion facilities constitute a significant fraction of the present fusion research program, see Fig. 11 and, for example, Focus On: Computer Modelling of Fusion Plasmas.

The migrating particles can also make their way to the confined plasma volume, diluting the fusion fuel and cooling the plasma through increased radiation losses. Impurities can reduce the fusion gain to unacceptably low levels. Therefore, the choice of the first wall materials and control of the power fluxes set important boundary conditions for the performance of the future reactor.

The main mechanism for material erosion is sputtering (definition) by which atoms from solid walls are ejected due to bombardment by the energetic plasma ions. One may imagine this in a similar way as firing cannon balls on stone walls in order to slowly destroy the fortress. For some materials an energy threshold for sputtering exists and no particles are released when the surface is bombarded by particles that have less than the threshold energy. Another sputtering mechanism is through chemical reactions if the target material and the plasma form chemically active combinations, such as deuterium and carbon. This is called chemical sputtering. The yields of these sputtering processes are subject to intensive research as carbon has excellent thermal properties and is therefore an interesting candidate for target materials but is also chemically active in the presence of deuterium.

If the wall material reaches a certain temperature (i.e. a certain power flux, given the thermal properties of the first wall) melting and blistering of solid material may occur causing a very rapid erosion of the surface. Therefore, engineers and physicists have to design the reactor first wall and its cooling system so that the wall temperature is always safely below a critical temperature - see Fig. 12. To be successful, reliable predictions of the peak power fluxes that may arise are needed, particularly during plasma instabilities. This is a crucial task for researchers who, in parallel, continue their search for improved plasma operation scenarios that would further alleviate the peak power loads onto the first wall. A particular issue is again the appearance of ELMs (mentioned above) that can deposit very high power levels in a very short time interval on plasma facing components, significantly decreasing their life expectancy. The control and suppression of these ELMs is again a major field of research as they occur on a regular basis in the very preferable H-mode (see pros and cons).

Plasma textbooks mention another potential source of local wall erosion - arcing - that may occur when the electric potential between a plasma and first wall materials exceeds a critical level. At present, this phenomenon is well under control in all standard fusion experiments.

Role of JET

JET acts as a bridge to ITER in many respects: It is currently the largest tokamak and, therefore, the closest facility to ITER in size; its shape and configuration is quite similar to ITER; and it is currently the only facility capable of operating with tritium, a fuel component of future fusion reactors. Many of the current JET experiments are devoted to the development of operating scenarios for ITER, including studies of the divertor physics as presented above. Due to its large size JET produces Edge Localised Modes (ELMs) of high amplitudes allowing appropriate scaling to ITER, and studies have clearly highlighted that certain types of ELMs must be avoided in ITER. In the recent past, JET has played a leading role in divertor design optimisation, which is documented in Fig. 13 as a series of historical photographs from inside the vessel.

In the near future, JET will assume the key responsibility as a test bed for the first wall materials that have been chosen under the current design for ITER. In particular, JET will assess the effect of the metallic and Beryllium first wall on divertor plasma operations. For details on this topic, see ITER-like Wall Project. Last but not least, JET exploits several diagnostic tools for plasma edge observations and contributes to their further development (see Focus On: Diagnostics and Focus On: Enhancing Capabilities .)

"The boundary edge is where the stellar world of hot plasmas meets the earthly world of cold solids.  Understanding the complex interaction of these two worlds is essential for operating a fusion reactor successfully."
Wojtek Fundamenski, Deputy Task Force Leader, Exhaust Task Force.

Co-Author: Jan Mlynar, Public Relations Officer

Co-Author: Marco Wischmeier, SOL & Divertor Modelling, IPP Garching