Start simulation with neutral atoms #768
Conversation
If ionization is chosen to be enabled by the parameter set then no electrons will be initialized but only ions. Immediately after initializing them their attribute `boundElectrons_` will be set to the number of protons chosen for the respective ion species.
The ionization energies and the necessary ionization field strengths are now stored in `PMACC_CONST_VECTOR()` arrays and can be accessed by the ionization algorithm according to the current charge state. Also with recent changes to the paramSets by @psychocoderHPC the cmake flag for the gas shape has been removed.
|
great job 👍 |
|
Like it! |
|
OK, to clarify things a bit. I did not pay so much attention, yet as to how the simulation parameters fit to each other. The ionization model still uses only the z - component of the electric field and the laser was circularly polarized. I now tried a bigger simulation area and a linear polarization in z direction. The structures are still there. I could change the sign of the amplitude of the laser field to see if the regions of high and low density get mirrored. (Btw: 5fps look horrible 😛 ... but a 42,9MB GIF was no viable option) |
|
well done @n01r, let's review the changes! :) |
| @@ -41,7 +41,7 @@ flags[8]="-DCUDA_ARCH=sm_20 -DPARAM_OVERWRITES:LIST=-DPARAM_DIMENSION=DIM2" | |||
| flags[9]="-DCUDA_ARCH=sm_20 -DPARAM_OVERWRITES:LIST=-DPARAM_CURRENTSOLVER=Esirkepov;-DPARAM_PARTICLESHAPE=CIC;-DPARAM_DIMENSION=DIM2" | |||
| flags[10]="-DCUDA_ARCH=sm_20 -DPARAM_OVERWRITES:LIST=-DPARAM_PRECISION=precision64Bit;-DPARAM_DIMENSION=DIM2" | |||
| flags[11]="-DCUDA_ARCH=sm_35 -DPARAM_OVERWRITES:LIST=-DENABLE_CURRENT=0;-DPARAM_DIMENSION=DIM2" | |||
| flags[12]="-DCUDA_ARCH=sm_35 -DPARAM_OVERWRITES:LIST=-DPARAM_GASPROFILE=gasSphereFlanks;-DPARAM_IONS=1;-DPARAM_IONIZATION=1" | |||
| flags[12]="-DCUDA_ARCH=sm_35 -DPARAM_OVERWRITES:LIST=-DPARAM_IONS=1;-DPARAM_IONIZATION=1" | |||
There was a problem hiding this comment.
hm, why did you change that? do you mind reverting that change (or moving gasSphereFlanks up to DIM2 would also be ok).
There was a problem hiding this comment.
I removed it because @psychocoderHPC 's big pull request #730 removed all of the gas configuration via cmake flags. Test case 12 was originally introduced by me anyway. I plan to make gas droplet studies later but for now this -DPARAM_GASPROFILE=gasSphereFlanks was totally without effect for all of the structure was changed.
|
looks great to me! ✨ just some inline questions left. |
Start simulation with neutral atoms
|
great work, well done! 👍 |
|
Yay that was fast ⏩ |
It will be possible from now on to start the simulation neutrally with un-ionized atoms.
Therefore the number of electrons bound to the core is set to the number of protons of the ion species via the
manipulatorconcept during initialization.Furthermore ionization energies and the fields necessary to overcome them are now stored in
PMACC_CONST_VECTOR()arrays and can be accessed by the ionization algorithm according to the current charge state.This allows now for any chemical element to be implemented. With it also comes for free that the effective charge of the initialized atoms is equal to zero because it is calculated from the
numberOfProtonsand the number ofboundElectrons.If you happen to implement different elements as species, remember to change the weight of the atom / ion accordingly. A more generic approach should follow in due time.
In the picture above a simulation result of Hydrogen atoms being ionized can be seen (electron density is depicted). I guess that these funny "filament-structures" are being caused by the step-like nature of the BSI algorithm.