Background
Well-designed #WFI Storage and Distribution Systems are isolated from the atmosphere and use sanitary or inert materials of construction for all product contact surfaces. Given its purity, #WFI readily absorbs and dissolves both ionic and organic compounds, and degradation of water quality in closed WFI Distribution Systems is expected. The effect is acceptable, as long as the quality does not result in excursions beyond the standards prescribed in the USP WFI monograph. As the water is consumed at use-points and the WFI Storage Tank is replenished with fresh water from the generation system, maintaining low #TotalOrganicCarbon (TOC) and conductivity levels are easily achieved for most systems.
Quality excursions for chemical parameters including TOC and conductivity are infrequent in WFI Storage and Distribution Systems. Sources of contamination include:
The feed water or generation system,
Back contamination for points-of-use, or
The product contact components.
Product contamination as a result of a feed water upset or generation system performance issue can readily be detected by sampling and analysis of the water throughout the generation system. Contamination from use-points is improbable for mechanically sound designs. Introduction of impurities to the system from external sources, however, can be more difficult to detect.
WFI Quality Excursions
A recently commissioned and validated WFI System had operated well within the USP quality specifications for WFI when an increase in both product water TOC and conductivity was suddenly observed. Although the excursions did not exceed the specifications of 500 ppb and 1.3μS/cm at 25°C for TOC and conductivity respectively, there was a definitive deviation from the normal operating range for these critical attributes. Interestingly, the slow, gradual increase in both parameters was observed over the course of several days and the corresponding fluctuations had remarkable correlation. The product water from the distillation unit used to generate the WFI was well within specification, so the troubleshooting focused on possible sources of contamination in the distribution loop. In addition, the accuracy of the online instrumentation was verified to ensure this was not a detection issue instead of a contamination issue.
Troubleshooting
The WFI storage and distribution consists of a 316L stainless steel tank, sanitary centrifugal recirculation pump, and in-line double tube sheet shell and tube heat exchanger. WFI is distributed to points-of-use through a 316L Stainless Steel, orbital welded distribution loop tubing. Most connections are welded, and there are minimal elastomers or gaskets in the system. Potential sources of contamination are minimal.
Since an increase in both TOC and conductivity was observed, a compromised heat exchanger that could be introducing plant steam into the WFI was thought to be the most likely cause. However, the integrity of the heat exchanger was verified and there was no indication of any type of mechanical failure. In addition, the materials of the diaphragms and elastomers, all of which were brand new, were verified for compatibility with the elevated temperature.
Next, WFI was drained from the system and the TOC and conductivity decreased back down to historical levels. Unfortunately with as the water was recirculated, a slow and steady deterioration of water quality began once again. The use-points were manually isolated ensuring that no back contamination was possible. All the while the total viable bacteria and endotoxin levels remained remarkably low. The system was definitely being contaminated, but the source was not evident.
Root Cause
After a couple days of troubleshooting, all the while still observing a decrease in quality, the source of contamination was identified. When the investigation extended beyond the logical contamination sources, the WFI Storage Tank Vent Filter was examined, initially on the basis that maybe it was compromised and that atmospheric impurities were being introduced into the system. Upon opening the vent filter housing, the source of the problems became clear.
The WFI is generated hot and the WFI Tank is held at 80°C. The tank vent filter housing is equipped with an electric tracing to prevent condensation in the housing that could possibly “wet out” the hydrophobic vent filter. The electric jacket is equipped with a temperature controller that allows for adjustment of the jacket temperature. In this case, the temperature set point was arbitrarily raised above the default set point of 95°C and maximum operating temperature of the filter element of 145°C. The result was thermal decomposition of the vent filter hardware (see picture). The PTFE filter media remained relatively intact while the polypropylene hardware was singed. The decomposition products found their way into the WFI and subsequently affected both the TOC and conductivity of the water.
The set-point for the jacket was lowered, the filter was replaced, and the WFI Storage and Distribution System decontaminated and cleaned. Since returning to normal operation, all water quality parameters have returned to their normal operating conditions and the system continues to reliably produce water well within specification.