«Association EURATOM / IPP.CR I N S T I T U T E O F P L A S M A P H Y S I C S, v.v. i. ACADEMY OF SCIENCES OF THE CZECH REPUBLIC ANNUAL REPORT ...»
An example of the EEDF for the radial position r = 56 mm in stationary phase of the plasma current for the shot #26402 is shown in Figure 1. The green line is a sum of the red one and the blue one. It is clearly seen that the EEDF is bi-Maxwellian. In Figure 2 the radial profiles of the two electron temperatures are presented. The fit with the experimental EEDF was obtained with an accuracy of 5%. It has to be noted that in the limiter shadow, r 85 mm, the EEDF is Maxwellian with a temperature of about 8 eV. In the same figure the results obtained by Stangeby method  (squares) are also presented. We must point out that the Stangeby method assumes a Maxwellian EDF of the electrons. Only the temperature of the high energetic fraction of electrons dominates [6,7] and can be evaluated. We estimate the density of the hot population of electrons at about 10% of the bulk electron density. Figure 3 shows the total electron densities at different radial positions. We compare the first derivative method results (dots) with the Stangeby method (squares) using the cold electron temperature only. The uncertainty in the first derivative method values evaluated does not exceed ± 30%. We see that the two methods show good agreement. Moreover, the first derivative method allows one to obtain in addition the “real” EEDF and the plasma potential values as well (Figure 4).
We have to note that, for those results, the raw data are not filtered. By using an advanced method based on an adaptive choice of the filtering and differentiating instrument functions, the 4x10
-3 electron density m
Part II - PHYSICS plasma parameters thus calculated are similar and we foresee that the advanced method will be useful for identifying the main modes of plasma turbulence.
 Arslanbekov R. R. et al., Plasma Sources Sci Technol. 3 (1994) 528 –538.
 Demidov V. I. et al., Phys Plasma 6 (1999) 350-358.
 Popov Tsv. K. et al., Journal of Physics: Conference Series 63 (2007) 012002.
 Martines E. et al., Plasma Phys. Control. Fusion 44 (2002) 351.
 Stangeby P. C., McCracken G. M., Nuclear Fusion 30 (1990) 1225-1379.
 Shoucri et al., Contrib. Plasma Phys. 38 (1998) 225.
 Batishchev O. V. et al., Phys. Plasma 4 (1997) 1672.
We present in this study the power loads in the ITER castellated PFCs by means of kinetic calculations during ELMs. The code used for this purpose has been developed at the IPP Prague and is adapted to such a tile gap geometry  by taking into account the specific geometry of the components, the inclination of the magnetic field lines and the gyration of the incoming particles.
Part II - PHYSICS TG and Ldep = 0.45 mm (-22%) for the PG, which is much greater than the geometric projection (Lgeo =0.018 mm for α = 2.1o).
For high plasma conditions, the effects of Te and Ti on the plasma deposition length inside the gap are not significant and the density plays a little role by increasing Ldep marginally (+15% when we double the density).
An important parameter is the inclination angle of the magnetic field lines with respect to the gap. Figure 2 shows the normalized power loads profiles in a TG for two angles, α = 2.1o (dotted line) & α = 1.2o (full line), and for the same plasma conditions (ne = 5.1019 m-3, Ti = Te = 2.5 keV, Bt = 5.9 T). The fluxes falling to the tile surface are in a ratio of a factor of 2 due to the different angles. However, the decrease of the wetted area inside the gap is only by 30%. We have a deposition of the power in the gap Ldep = 0.50 mm for α = 2.1o Figure 2: Normalized power deposition in a 0.5 mm whereas the deposition is Ldep = 0.35 mm for TG for an inclination angle of 2.1o (dotted line) and o α = 1.2o. It has to be noted that the peak value 1.2 (full line) at identical plasma conditions.
increases relatively to the incoming power to the tile with the smaller angle, but remains nevertheless lower in absolute value. We can note that the integrals of the curves shown in Figure 2 are identical due to this feature. This means that the total power inside the gaps is only dependant on the incoming power to the tiles. However, thanks to the change of the slope, by reducing α, the power loads inside the gap are spread differently and the wetted area reduced.
The same conclusions concerning this effect of the incident angle apply to the PGs. The curves are similar, the only difference between the types of gaps coming from the geometric effect described previously in Figure 1.
This work was performed in the frame of the EFDA Task: TW6-TPP-DAMTRAN "Report(ing) on the expected power deposition profiles onto ITER PFCs during steady-state and transient loads with realistic PFC geometry by PIC modeling".
 R. Dejarnac and J.P. Gunn, J. of Nucl. Mater. 363-365 (2007) 560-564.
Particle and power loads in tile gaps are of high interested since ITER plasma facing components will be castellated . Previous experimental studies all show shot- or campaignaveraged results of deuterium and impurities deposition inside the gaps [2,3], but giving no information on how the plasma flows between the tiles. In order to investigate more deeply the physics of the plasma deposition, we have developed a special probe that can measure the ion saturation current profile along the gap (see Figure 1). The experimental data are compared to the results of self-consistent kinetic simulations performed by a 2D particle-in-cell code developed here at the IPP Prague and described in .
Figure 1: Picture of the 2 components of the dismantled "sandwich probe" for measuring the plasma deposition into a gap between tiles.
Part II - PHYSICS phenomenon is due to an asymmetry of the electric potential inside the gap as explained in  and is in good agreement with the numerical predictions.
2-Poloidal gaps Figure 3-a shows the ion saturation current distribution along a poloidal gap measured for different inclinations of the probe with the magnetic field lines. We observe exponential decays only for low angles (α≤32.8o) whereas the plasma seems to penetrate deeper and linearly for angles greater than 32.8o. The deposition shown here corresponds to the plasma facing side of the probe. The profiles calculated by our kinetic code are shown in Figure 3-b and we can observe that a similar behavior is well reproduced, with a non-exponential decay for high inclinations.
The absolute values are in the same range as in the experiment, however we notice that the deposition is in reality deeper, especially for the larger angles. More investigations must be undertaken to understand this result which is reproducible over all the angle scans we performed.
On the other side of the probe (plasma shadowed side), we do observe some signals but with an intensity 10 times lower. We explain this deposit by a "bump" on the potential profile inside Figure 3: Ion saturation profile along a poloidal gap measured by the probe (a) and calculated by the 2D self-consistent PIC code (b) for different inclinations with respect to the magnetic field lines (from 5o to 45o) and for the plasma facing side.
the gap, strong enough to repel the incoming ions towards the surface that is not directly wetted by the plasma . Our simulations also reproduce the same asymmetry, plasma facing side/shadowed side. The absolute values at the entrance of the probe tend to be lower than the predicted ones but the radial deposition is in good agreement for all angles.
Conclusion We have developed a unique tool to measure the plasma deposition into a gap between tiles.
The experimental data confirm the numerical predictions we made with our 2D self-consistent numerical code. In the case of toroidal gaps, we have a quantitative and qualitative agreement.
However, in the case of poloidal gaps, it is only qualitatively acceptable. The 2-sided deposition is confirmed, with the good order of magnitude. This set of experiments confirms nevertheless the understanding of the plasma deposition in tile gaps presented in .
 W. Daener et al., Fusion Eng. Des. 61&62 (2002) 61.
 K. Krieger et al., J. of Nucl. Mater. 363-365 (2007) 870-876.
 A. Litnovsky et al., Phys. Scr. T128 (2007) 45-49.
 R. Dejarnac, J.P. Gunn, J. of Nucl. Mater. 363-365 (2007) 560-564.
In collaboration with:
H.P. Laqua, Association EURATOM – Max-Planck-Institute für Plasmaphysik, Greifswald We have adapted our EBW code to obtain a crude estimate of the profile of the current driven by
2.45 GHz wave. We investigated the dependence of the current profile on the central magnetic field and the temperature of the hot component.
WEGA plasmas are sustained by electron Bernstein wave heating. This produces a considerable suprathermal electron population. The fraction of this populations ranges up to ~20% and the temperature reaches 200 – 300 eV (compared to ~50 eV of the bulk electron population). These data are estimated from probe measurements and are only approximate. We have extended our EBW code to include the effects of the suprathermal electrons. Significantly different behaviour of the EBWs, compared to single component plasma, has been demonstrated by the simulations.
Even a modes fraction of suprathermal electrons enables very efficient absorption.
Current density profile was, for the first time, measured by a small Rogowski coil during our visit . Therefore, we have concentrated on current drive studies. We have simulated the absorbed power corresponding to individual resonances (and components) and determined the corresponding direction of the driven current, supposing Fish-Boozer current drive mechanism.
The parallel wave vector direction of the EBWs can be reversed during the propagation through the plasma. The driven current direction has to be reversed analogously. Such behaviour has been observed experimentally . We have investigated the effects of the central magnetic field and of the hot component on the EBW current drive.
Fig. 1. Distribution of the absorbed power between components and harmonics. Individual contributions are multiplied by the sign of vres=(ω-nωce)/k||. B(R0)=0.65Bce, T1=300eV.
Fig. 2. Hot component temperature scans. Summed Fig. 3. Central magnetic field scans. Summed components components of the absorbed power. Individual of the absorbed power. Individual contributions are contributions are multiplied by the sign of vres=(ω- multiplied by the sign of vres=(ω-nωce)/k||. T1=300eV nωce)/k||. B(R0)=0.65Bce.
 Laqua, H.P., et al., Bulletin of the American Physical Society, vol. 52, no. 16 (2007) 280
EFIT2006 is a modern version of the magnetic equilibrium reconstruction code EFIT, developed at UKAEA Culham. A parallel version of the code was desirable to accelerate the calculations.
EFIT2006 will be also used at IPP Prague for the COMPASS tokamak.
Parallelization of EFIT2006  has started from scratch. It has been decided to implement a coarse grained parallel version of EFIT2006 , i.e. to parallelize the high level interface, so that a master process will serve other processes with input data for individual time slices. These processes (slaves) will perform the computation and send the results back to the master, where all the outputs will be written to a file or sent to a database. The parallel version should run on computer grids or clusters using the MPI interface.
EFIT2006 is an object oriented code in C++, which is very beneficial for understanding and enhancements, but parallelization using MPI requires rather complicated transmission of C++ objects. Fortunately, the Boost C++ libraries have recently been extended by an MPI library, which allows transmitting complex C++ objects very conveniently. A functional version of parallel EFIT2006 has been implemented and its performance has been tested. The results are shown in Fig. 1. A linear growth can be clearly seen, even though the benchmarks have been run on serial computers interconnected by rather slow Ethernet network.
EFIT2006 and all the necessary libraries have already been installed at IPP Prague.
Benchmarking results for the same inputs shows that the two installations are equivalent.
 Appel, L.C., et al., 33rd EPS Conference on Plasma Physics, Rome (2006)  Havlíček, J. and Urban, J., WDS'07 Proceedings of Contributed Papers: Part II - Physics of Plasmas and Ionized Media (2007) 234-239  K. Tani, M. Azumi and R. S. Devoto, J. Comput. Phys. 98 (1992) 332 We consider the the quasi-neutral self-consistent particle-in-cell (QPIC) simulation of a tokamak plasma scrape-off layer (SOL) bounded by two material walls – possibly divertor plates, limiters, probes, or a combination of these. At every time step of the simulation some ions and electrons reach the material walls, but suffer very different fates. The ions are absorbed by the walls and are lost from the simulation. A QPIC simulation satisfies local plasma quasineutrality so that during the simulation the plasma must also be globally neutral. It follows that not all electrons travelling to a wall are allowed to reach it: the total number of electrons lost to the walls must equal the total wall ion charge. If there is global charge balance at each wall, i.e if the ion and electron fluxes to a wall are equal, then the wall potential equals the floating potential. By definition, this floating potential is actually the energy of the most energetic electron which is reflected back into the plasma from it.
Introducing secondary electron emission changes the wall potentials and the wall charge balance.