quasi neutrality with constant momentum distribution

[floresv299]

Hello all,

I'm trying to replicate some findings that I got using OSIRIS. My simulation is simple: it is a 1D3V simulation involving a laser interacting with a solid target (10s to 100s of times the critical density). My density profile has an exponential scale length to model the pre-plasma. The 1D3V aspect of my simulation comes from a constant momentum drift in the -y direction (p =-$\hat{y}$ mc tan$\theta$). This sets up a "boosted frame", essentially allowing interactions to be oblique (with other spatial and temporal transformations). Currently, when I apply the momentum drift I get different densities for the ions and electrons. I was thinking that since both species have the same constant momentum then quasi-neutrality will be maintained like in my previous simulations. However, this seems not to be the case. Any advice ? Below I have attached my inputs and a plot of the densities. Thank you all !
Warpx_Momentum_Drift.pdf
Warpx_mom_drift.txt

[RemiLehe]

Hi,
Sometimes, in PIC simulation, a neutral plasma can become numerically unstable when it is drifting (even though it should be physically stable). The numerical stability usually depends on many parameters, including the density of the plasma, the drift velocity, the grid resolution and the type of solver (e.g. Yee vs other solvers).
I am not too familiar with the stability properties of the the 1D case with a drift in y, but do you have a way to check whether this should in principle be numerically stable. For instance, have you run the exact same simulation (i.e. same density, same resolution) with Osiris?

[floresv299]

Hi Remi,

Thank you so much for the help. Turns out it was a resolution problem. Increasing the resolution did the job ! Thanks for your insight.

best,

Victor