ok - so I undrstand the the issue is not leak rate, but success (or failure) rates.
With high volume applications, if there's a certain chance of the procedure not sealing, it translates directly into yield numbers.
For this design, if the procedure fails, the experiment won't work, so it...