Trying to understand why the best fitting function is not a erf function (given the hypothesis that the beam is cut at constant speed): maybe the exponential decay we see in the data is dominated by the electronics ? one can also try to fit with a function erf + exp.