Author: Andrzej Sadanowicz
USP<645> details requirements for inline conductivity measurement of purified water. In this post, I am only focused on calibration/verification method in the field which in my opinion can render the best results for accuracy. The measuring system consists of the meter itself and conductivity probe installed in RO distribution loop piping. Output signals from the meter and PLC control configurations are important, but not discussed in this article.
There is an ongoing discussion in pharmaceutical industry how conductivity meter should be calibrated in the pure water range. Ideal typical reverse osmosis (RO) water has conductivity of 0.055 µS/cm² or 18.2 MΩ*cm at 25.0 °C. Standard solutions do not exist in that low range.
After pure water leaves EDI unit of producing part of the system, it is pumped in to RO storage tank then distributed to points of use and recirculated. Conductivity increases due to impurities picked up from the system and carbon dioxide absorbed. In the well maintained system, readings could be somewhere in the range of 0.2 to 0.5 µS/cm² (uncompensated) and system temperature range of 21 to 24°C. So still, there is no conductivity standard solution in that range. The best solutions in the market as low as 1.29 µS/cm² at 25.0 °C have +/-1% uncertainty (Hamilton).
That’s why one test point direct comparison with a least 1:1 accuracy ratio calibrated NIST traceable STD meter with a known accuracy and known cell constants connected in line next to UUT is the best approach. RO water circulating in the loop can serve as a source solution for calibration (the same way like temperature bath serves as a source of temperature). Its temperature and conductivity is accurately measured by STD meter. Make sure that both measuring systems are configured in uncompensated mode and temperature has stabilized, before readings are taken.
If “As found” observation meets the user tolerance requirements, we are done. Results are recorded and documented. If the results are not satisfactory due to elevated UUT conductivity readings or reported earlier problems, calibration can be performed. For specific instructions refer to OEM manual.
The other method which utilizes immersing sensor in standard conductivity solution of value within meter’s operating range and then adjusting cell constant is not recommended to perform onsite due to greater total uncertainty of measurement caused by environmental conditions. Open container of the solution quickly absorbs carbon dioxide from the air while operator is waiting for temperature stabilization. Additional error can be introduced by contaminating both the standard solution and conductivity probe during the test. Just touching any of these would increase readings. Temperature stability is critical.
As much as low conductivity calibration is very difficult, there are practical ways to achieve results accurate enough and sufficient degree of confidence to the method used.