I too have struggled with this. The best I've found is to create the standard in a glass bottle with a teflon-backed septum on top. I use 120 mL serum vials. Just search "serum bottle" with google and focus on images. Something like toluene is pretty volatile and will stay in the vapor phase, even at RT. If you make a concentrate in one serum vial by adding a small amount of liquid toluene to it, you can dilute it by taking a known volume of gas from the vial and then adding it to a sealed second vial, etc.
My biggest problem with this is that my additions generally don't match up with what I calculate. For instance, 0.1 µL of toluene at RT is roughly 0.94 µmole (use the face-value of the density from Sigma-Aldrich). In a 120 mL serum vial (4.90E-3 mole gas at 1 atm) that's about 192 ppm toluene (µmole/mole). A 1:120 dilution of that should be 1.6 ppm. It never seems to work out the way I expect. I have even struggled to make the same thing twice with this sort of procedure. Very frustrating. Since I usually have a certified gas calibration standard to which I can compare my standards, I can "certify" my preps that way. I'll prep something and then check it to my known standard. Actually, since I compare against a known standard anyway, I will usually just take a bit of the headspace of the reagent bottle and add it to the sealed serum vial. Saves a couple of dilution steps.
Seems to me that all of those tedlar-type bag things will absorb your analytes over time. I've never really had good experiences with them - especially for long-term type studies. Perhaps others have had good experiences with them?
