An improved mixed dextran rejection test for characterization of UF membranes in an R&D environment Original Research Article

Ultrafiltration performance is often characterized using solutes of known molecular size that emulate protein molecules. At Millipore we currently use a mixed dextran rejection test for release of UF membranes. We have recently developed and validated a new mixed dextran test that shows improved precision and sensitivity. The improved precision results from changes in the test format, changes in and standardization of all operating conditions used, improvements in general lab practices that resulted in reduction of bioburden, as well as better management of the SEC calibration procedures. The %CV for R90 (MW corresponding to 90% rejection) has been reduced by about 50% compared to previous practices. The improved sensitivity results from the use of more sensitive detection equipment, better integration and processing algorithms and better understanding of chromatographic artifacts. Sensitivity has been improved by 10–100x and it is now possible to robustly report R99.9 (MW corresponding to 99.9% rejection or sieving of 0.001 of a dextran molecule) for membranes that have that capability. The improved test is presently being used for membrane characterization within an R&D environment. The capability has provided a valuable guide to new membrane development projects. More specifically, the consequence of the precision improvement is the ability to quantify the effect of small changes in manufacturing parameters and thus establish the trends in the process. The consequence of the sensitivity improvement is that we have a finer lens to study the effect of various treatments on the membrane. Changes in membrane properties that are not reflected in R90 can now be detected. The improved test could be used as a tool to explore the correlation between dextran and protein sieving performance, based on the expectation that R99.9 might be a more sensitive parameter compared to R90. Other potential benefits include the ability to map and better understand UF membrane process capability. Ultimately, better membranes can only be characterized to their full potential if a better test is available.