TL8W stated in the picture of
j-frog's Philips F32T8/TL735 T8 lamp , that the end banding on several types of fluorescent lamps (I never seen end banding in T8s, at best not in the european krypton energy saver T8s), "supported by argon buffering rather than krypton". You can see in his comment of the picture i posted in this topic.
Probably the american F32T8s have less krypton, so the starting voltage is matched to the OCV of the F32T8 rapidstarts. I think that the ignition voltage of the european T8s are in the 750 volts or similar to the voltage that the preheat ballast makes when the starter opens. Even thermal and glow starters have difficulties with T8 starting, which is expressed by more blink happy and failing to start an EOL lamp (Strobing, like HID lamps reignitions), instead of successfully strike the lamp but then being triggered by a rectification (Flashing).
There is no way to build efficiently a rapidstart with an OCV of 750V, and the F32T8 rapidstart ballasts have at least in the 400Vs range. So an internal electrically conductive coating will help reducing the starting voltage. But if the F32T8 can develop an end banding similar to short T5s, T12 and T9 circlines, while the end banding don't occurs in the european T8s with krypton, probably the argon is the dominant gas in the F32T8.
Also, if the end banding can occurs also with the T12 VHO, which have neon as the primary gas in the buffer, probably the buffer of the F32T8 is an argon+neon+krypton.
Perhaps Max or James knows the chemistry of the american energy saver T8s.