Green LEDs are the least efficient of all the visible LEDs. If you need really bright green LEDs it is probably better to use a green phosphor on a Blue LED similar to the way that white LED's are made.
Another option is to silver the substrate yourself by reducing silver from a solution of silver nitrate. If you use a glass tray with a light underneath you can monitor the glass getting more opaque. Your substrate in the tray will be getting coating. By pulling the substrate and rinsing at...
What looks to be leakage in the IR if you look solely at the absorption spectra isn't an increase in transmission, but an increase in reflectivity. Similarly much of the decrease in reflectivity around 400-500 nm is really an increase in absorbance.
As to getting such a filter, I would be...
My experience is that silver becomes as reflective as its going to get at about 200 nm of thickness. I suggest a thickness of 100 nm as a test. Perhaps you could find data on the reflectivity of silver as a function of thickness. Then you could make a better estimate of the thickness you want.
https://en.wikipedia.org/wiki/File:Image-Metal-reflectance.png
You might need this as well. So as long as you want UV light at about 350 nm, you have a solution. Longer wavelength will be absorbed or reflected.
I hope that this helps.
Have you considered a very thin silver film? Silver has significant transmission in the UV. You might not be able to afford the loss in sensitivity depending on how bright your source is.