Two small calculators for planning grazing-incidence pump-probe experiments,
ported from calx.pumpprobe (pure
numpy/scipy, no facility-specific dependencies, so reimplemented directly in
bernina_prep.timing rather than depending on the full calx package).
Each parameter row has a small live plot of the output versus that parameter, swept over whatever range you've actually explored so far (it grows as you move the slider/field further); a side-view diagram below each calculator shows the beam(s) hitting the sample surface at the current angles, both updating automatically.
Pulse energy spread over the beam's footprint on the sample surface. At
grazing incidence (angle well below 90°) the same beam energy is
spread over a much larger illuminated area than the beam's own
cross-section, reducing the fluence per unit area accordingly.
Sample surface as a horizontal line; the beam (shaded, collimated/parallel -- not converging) hits it at the current incidence angle. Since this calculator only takes a beam area, the diagram assumes a circular beam and uses its diameter as the effective linear size (labeled explicitly below) -- purely a drawing convenience, it doesn't affect the fluence number above.
Independent of the pump/probe pulse durations themselves, a purely geometric timing smear arises whenever pump and probe cross the sample surface at an angle to each other with a finite footprint: different points across that footprint see the pump wavefront arrive at slightly different times, because of the path-length difference across the (often strongly elongated, at grazing incidence) illuminated area.
The sign of the result indicates which beam's wavefront leads across the footprint (pump vs. probe); the magnitude is the relevant smearing contribution to add (in quadrature with the pulse durations and other jitter sources) to get the overall time resolution.
Both beams (shaded, collimated) on the same surface line: the probe (blue, typically much more grazing) and the pump (orange, typically steeper). Their footprints rarely line up end-to-end -- that mismatch is exactly the geometric smear computed above.