That's really cool. There's clearly an effect, but it is small enough we can ignore it. I'm interested in why the trend is negative for the 5ppm sample.
That is an interesting question....I might consider moving out beyond 5ppm and see what happens...