Consider a spherical drop of water, initial temp 40C, radius 3mm, mass 0.1g
To get it down to 0C, you need to remove 4.18 (J/gK) * 0.1 g * 40 K = 17 J
then, to freeze it solid, you need to remove latent heat of fusion 333 (J/g) * 0.1 g = 33 J
for a total of 50 J.
The heat conductivity equation is
$H=\frac{\Delta Q}{\Delta t} = k A\frac{\Delta T}{x}$
where $k$ is the thermal conductivity of water ($0.6 W/m\cdot K$), $A$ is the surface area, and $x$ is the thickness. Take $A=4\pi R^2$ and $x=R=3$mm, and you find that it would take 27 sec to freeze the drop of water in -40C.
Now in practice, the drop will be elongated, increasing $A/x$, and really only the surface layer needs to freeze, possibly eliminating 50-90% of the required latent heat of fusion, so in practice, I think it should be possible to freeze in about a second.
For a real answer, I think we need to go to Mythbusters!!