Focusing by plasma lenses requires small enough timestep

[RemiLehe]

The plasma lenses require the timestep to fine enough so that the motion of the particles through the lens is resolved, in order to get converged. This is not necessarily the case in boosted-frame simulation, where the timestep is dilated, and the lens is shrunk.

The following plots illustrate the issue as a function of c*dt_boost/L_{lens,boost} and its impact on the final emittance:

These plots were obtained with the following script:
inputs_3d.txt

[ax3l]

Discussed today: we can throw a warning based on c*dt_boost/L_{lens,boost}

[RemiLehe]

As a side-note, here is also a version of the input script in RZ:
inputs_rz.txt