I just had an idea and wanted to run it by the experts:
I see many pH calibration buffers in packs which user supposed to dissolve in water. Measured amount of powder goes into measured amount of water to yield desired pH. This got me thinking about errors:
1. Amount of power needs to be accurately measured
2. Quantity of water needs to be accurately measured.
So i wonder: why not dissolve excess amount of solvent in arbitrary amount of water and use solubility of ingredients in solute to get a known pH value at saturation point in water? This way temperature of water is the only driving factor. Filter the excess of solute and you get some arbitrary, but known pH value?