7. EPOCH results are in Cartesian or Cylindrical?
The EPOCH domain is Cartesian and EPOCH is not capable of R-Z simulations. In 2D there are pathologies in any choice that you make and we chose this one as having the fewest. While you are correct that it corresponds to an infinitely extended plane wave, going to an R-Z geometry imposes m=0 symmetry on the simulation. So for example a laser with an Ey field component in the Cartesian model corresponds to a laser with radial polarisation in the R-Z model which is also not physical for the vast majority of lasers.
After careful consideration we decided that a Cartesian model was the best general purpose model since it's limitations are immediately clear (nothing varies in the direction that you are not simulating). We believed that the explicit labelling of all of the axes as x,y etc. and similarly for the field components made it clear that this was how EPOCH works. There are PIC codes that use an R-Z geometry in the wild, but I'm afraid that I can't really recommend any of them specifically.
6. Hot electrons appearing on rear surface of plasma
In reality materials have strength due to the chemical bonding between materials. No such bonds exist in PIC simulations. I'm fairly sure what you're seeing is the electrons free-streaming out of the target rear surface due to:1) the electrons enhanced mobility with respect to the ions and 2) the pressure imbalance cause by the disparity in the number density at the target rear surface|vacuum interface. Remember P=nkT for an ideal gas; in the vacuum n is a LOT lower. One way to reduce this effect is to increase the temperature in the quasi-vacuum region to create a pressure balance, although this may cause problems of its own depending on what you're doing... this is a 'known problem' with PIC methods.
5. If you're trying to focus the laser onto the target, you can make the phase a function of y and time according to
https://en.wikipedia.org/wiki/Gaussian_beam#Beam_parameters
to get the focus positioned on your target.
4. EPOCH parameters initialization:
In large simulations, the simulation geometry is broken up into smaller units or cells. To resolve c/ωpe and to be stable, the cells must be smaller, by a factor of about 3 or more in practice, than c/ωpe.
The electric and magnetic fields in a particular cell should not influence the distribution of particles in an adjacent cell, a cell should thus be greater in size that the Debye length ve /ωpe , where ve is the electron thermal velocity.
3. 备注 dist_fn模块输出选项：
data = sdf.read("../Data/0009.sdf",dict=True)
denden = data['dist_fn/theta_en/electron'].data[:] / ...data[:,:]
print denden.shape 输出（400，400，1）
denden = data['dist_fn/theta_en/electron'].data[:,:,0]
print denden.shape 输出（400，400)
2. laser electric field will have Ex, Ey and Ez component in EPOCH calculation, although you defined the laser travelling in x direction with electric field in y direction.
EPOCH propagates the laser using Maxwell's equations, which will cause the beam to diffract and expand. As your beam waist is small, the laser will expand with a small F number giving you some significant components to the laser Ex field. The notion that a linearly polarized laser pulse only has E field components in a single transverse direction is not correct for any real beam, only for infinite plane waves.
1. Using 'dist_fn' will get the npart/cell~energy relation. To get the energy spectrum...
Since the plot is in particles per cell, simply divide by the width of the cell in MeV and you'll get number of electrons per MeV.
网友评论