Finding a better scipy curve_fit! — exponential function?

Question

I am trying to fit data using scipy curve_fit. I believe that negative exponential is probably best, as this works well for some of my other (similarly generated) d