Hello Zoeyy,
you don't say anything about the background (physics) of your data, so I have no idea of a reasonable fit. What do you think of this?
fff = FindFit[data1, a (1 - b Erf[x - c]), {{a, 2200}, {b, 3}, {c, 12}}, x]
h = a (1 - b Erf[x - c]) /. fff
Plot[h, {x, 0, 18}, PlotRange -> {0, 2500}, Epilog -> {Red, PointSize[.02], Point /@ data1}]