Fix pulse slice initialization

[k-wolfinger]

Takes too many slices to converge for number of photons in pulse vs number of laser slices.

[k-wolfinger]

Also:

• Look at other dependencies for gain, like mesh resolution (nx, ny)
• Initialization types for the crystal at some point: flat, exponential decay (1 vs 2 sides), outside data

[k-wolfinger]

Not doing the above yet (mesh resolution and crystal init types). Are changing alpha to cm, adjusting init of excited states.

Adding trap rule