You could be onto something, but I don’t know. The system is not being actively mixed. Since we can watch the ppm of H20 increase at the valve over time and flushing fixes it, it seems to be a local effect. But I’m going to calculate the dead volume in the pipe today and see if basically we are emptying the entire contents of the standing pipe and then the fresh supply from the keg is replacing it.
Our testing apparatus and methodology are so excessively anal-retentive that I do not suspect it at all. You’ll have to trust me on that one. I definitely believe the results we are getting from our testing methods. Besides, we are seeing this phenomena consistently: ppm is low, it stands, it gets high, and goes back down with flushing.
If I have a problem with the process line, that gets back to my original confusion. I’m using 316 SS all around (pipes, fittings, valves). Acetonitrile does not react with it. We’ve been doing this for years at this facility. The only reason it caught my attention is that I personally installed this whole new run of plumbing and noticed the water content issue. After asking around, it has been observed by others that this happens and no one has really ever cared to investigate it too much.
Some further info: the same issue has been observed with a pipe containing acetone instead of acetonitrile. The spec on acetone is <5000ppm. I’m currently unaware if the acetone will exceed this spec from sitting in the pipe, only that it increases.
You’ve got me thinking about this separation angle. I don’t know enough about fluid dynamics to know how some H20 in the system will behave in acetonitrile (or acetone). If it is pressurized, does the water “slip” by the solvent in the keg and accumulate in the pipe? I don’t know. Maybe when I flow from a valve, there is some inertia that remains after closing the valve, and the water is still moving in the direction of the pipe, but faster than the solvent?