Ionization types

[cbontoiu]

Below is a comparison of all ionization types for the same model. These are independent simulations where the speciesDefinition.param file was modified accordingly for the Carbon atoms. Looking at the ratio between the total number of "free" electrons at a given time and the number at the beginning I am surprised to see that collisional ionization (TF) yields the same line as all the ionization models applied simultaneously. Is this right?

I checked the file and I have: with all types

and with collision only

This is how I collected the number of electrons

Thank you, Cristian

[sbastrakov]

@n01r could you have a look?

[n01r]

In principle I find it unlikely that all the models give such similar results. Especially TF, as a collisional model, and the field ionization models should be different. Could you share more information on the setup? Have you possibly tried out estimating the ionization with the Python modules given in the docs

[bussmann]

To me this looks like a case of 'highest ionization rate possible' achieved. I am not sure PIConGPU can (and should) get in such a state. But if you assume that in one time step you have an upper limit on how many electrons can be stripped away, then if this limit is universally met, all ionization models should look alike.

[n01r]

You're right, this could happen for the field ionization models which create at maximum a single electron per ion and step. Chaining them together allows for more. In our typical simulation setups, the steps are generally needed to be small enough to resolve the laser wave, so that this does not normally occur. The TF model is not limited by this and will calculate a new charge state based on ion density and electron average energy under the assumption of local thermal equilibrium. The model will still create only one new macro-electron per ion and step but its weighting will account for multiple ionization.

[cbontoiu]

Thank you all for your replies. I produced a better picture adding the laser periods on the top

The peculiar parameters are:

I0 = 1e21 W/cm^2 lambda = 50 nm a0 = 1.35 T = 0.167 fs Dt(FWHM) = T full pulse length = 3 T = 0.5 fs base_density for the target = 1e23 cm^-3 DELTA_T_SI = 1.93e-19;

A lot happens just when the pulse reaches the target due to the high energy compressed in a short time scale. Maybe this explains why the lines are nearly the same for short times t/T < 3. As the laser pulse is depleted the ADK mechanism slows down more than the BSI mechanism which itself is weaker than the TF mechanism. Some electrons are pushed out of the simulation domain such that the lines never reach 2, the maximum possible. At least this is my interpretation for what is going on. I will try to estimate the ionization using the documentation.

I also attach my param files (all ionization models on) if you would need to run this setup at some time for checking the ionization models.

param.zip

[cbontoiu]

I also forgot to mention that my simulations start with 3 "free" electrons for each Carbon atom and this explains why the TF mechanism causes ionization even when the other two mechanisms are switched off.

[BrianMarre]

@cbontoiu Is there still an open question? If not please close the issue. ;)