Troubleshooting Quality Issues in UPW Systems

Ultrapure water (UPW) systems integrate many unit operations and can be very complex.  Understanding what each of the unit operations are designed to do and how they affect downstream unit operations is critical for routine operation and maintenance. However, even with this understanding, there will come a time when a water quality issue or event will arise, taking you on a journey of cause-and-effect analysis and require a good bit of sleuthing to solve.

One key takeaway from the roundtable discussion Troubleshooting Quality Issues in UPW Systems was the importance of maintaining historical data and records for the purpose of validating either grab samples or on-line metrology. Have we seen similar data before? What was the cause? How did we fix it?

Another common discussion topic at the roundtable revolved around the reliability of metrology. UPW systems rely heavily on metrology, but can it be trusted? Key discussion takeaways included the importance of understanding how these instruments operate; what their limitations are; and how to perform the required routine instrument maintenance (calibration, reagents, etc.).  

The question was posed - if an analyzer is reporting a value that is outside of your control limit, what would you do? Many good questions were raised at the roundtable to help identify if the issue is real or not: Is the reported value a trend or an outlier? Is there any other data that could corroborate the reported value? Have there been any maintenance activities in the system such as mixed bed regenerations, resin replacement, and/or filter changeouts? Has the instrument been maintained as required? Does the instrument have the proper sample flow and pressure? All are important to consider. 

If the water quality issue is real, i.e., not due to faulty metrology, then a deeper dive into the UPW system is required – which some roundtable participants referred to as the moment ‘where the fun starts.’  

There was a discussion regarding the reoccurring short-term elevation in UPW Total Organic Carbon (TOC) when bringing a new mixed bed on-line after rinsing to the prescribed TOC endpoint. After ascertaining the TOC analyzer was providing reliable data, the next suggestions from the group discussion were the potential for dead leg(s) in the manifold; the rinse flow rate being higher than the service flow rate; and lowering the rinse TOC endpoint value (as a short-term “fix”).

One participant reported an example in which they were troubleshooting uncharacteristically short primary mixed bed life. In reviewing historical data on the unit operations upstream of the mixed beds, they discovered the pH control of reverse osmosis (RO) system was operating below the set point. At pH > 9 boron exists as a salt and is well rejected by RO membranes. Below this pH boron exists as an acid and is not well rejected by the RO membranes which placed a higher ionic load, namely boron, to the mixed beds. The takeaway here is that it was important to keep track of critical operating parameters that could affect the performance of downstream unit operations.