Store plasma species in separate PlasmaParticleContainer

[AlexanderSinn]

Time to calculate the loop over slices in calculate_wakefields() for one time step, with numba cache already set up:

Wake-T/examples/track_plasma_qs2d.py:

dev:

first iteration: 66.8 ms
all other: ~20.5 ms

With multiple species, commit 93ba49b54be54f9a2dab69a1552adad68e8fc32d:

first iteration: 103.5 ms
all other: ~32.3 ms

Multiple species, everything in one JIT function:

first iteration: 50.8 ms
all other: ~11.8 ms

10 TeV Wake-T example:

dev:

first iteration: 160 ms
all other: ~98 ms
total runtime:  4m9s

With multiple species, commit 93ba49b54be54f9a2dab69a1552adad68e8fc32d:

first iteration: 236.3 ms
all other: ~147 ms
total runtime:  5m24s

Multiple species, everything in one JIT function:

first iteration: 109.1 ms
all other: ~58 ms
total runtime:  3m18s