The Unrecognized Role of Mixing During Tangential-Flow Filtration (Webcast Summary)

Tangential-flow filtration (TFF) membrane performance is difficult to maintain between batches and during scale-up. Lucas Smith (product manager at Repligen) discussed the benefits of good mixing during TFF in a recent BPI Ask the Expert webinar. In laboratory-scale processes, recirculation flow into the vessel may provide adequate mixing. However, in large-scale tanks, issues such as inconsistent feeds and heterogeneity can cause poor membrane performance. Using a method for evaluating mixing efficiency via its impact on membrane performance, Repligen presented a study comparing traditional mixers with their Tulip Tank technology.

Smith’s Presentation

Many TFF vessels provide inadequate mixing, allowing material to stratify, localize, or short-circuit. Poorly mixed material enter the filter nearly untouched, creating a portion of feed with an abnormally high concentration that leads to compromised membrane performance. Drug manufacturers must adopt a method to analyze feed homogeneity during early stage process development and, in turn, select an appropriate mixing vessel.

TFF must be optimized to minimize process deviations at larger scales. If a homogeneous feed solution is maintained, then filtration will be more efficient, predictable, and controllable. With Repligen’s TFF systems, membrane performance can be monitored during TFF to help correct feed inconsistencies.

Diafiltration Case Study: Diafiltration (DF) is used to exchange one buffer for another while retaining target proteins. This process is tracked by the number of vessel volumes that are transferred through a TFF membrane before buffers are changed. Smith presented a graph showing the theoretical buffer-exchange rate curve as a function of transferred diavolumes (DVs). The curve represented mixing performance when supplying an equal amount of new buffer against removed DVs to maintain a constant volume within the retentate vessel. Low DVs correlated with low buffer-exchange rates. If a retentate vessel provides the membrane with a consistently homogeneous feed, then mixing performance should match the outcomes predicted in the theoretical curve.

Smith’s team executed a control run using a traditional feed vessel, in which they delivered new buffer directly below the liquid level. Mixing was performed by a single impeller that was positioned toward the bottom of the tank and the resulting recirculation flow. One portion of the control profile showed a lower degree of exchange than predicted by the theoretical curve, meaning that the membrane feed had a higher concentration than would be expected from a homogeneous solution. A second portion of the control profile exceeded the theoretical values, suggesting a higher buffer concentration and inconsistencies in the feed solution. Such factors increased circulation time, which can result in permanent membrane fouling and poor product quality. Those examples demonstrated the importance of good mixing during filtration.

Tulip Tank: Repligen’s Tulip reservoir is a single-use mixing tank for optimizing TFF performance. The square self-baffling design has two independently controlled impellers, which allow for mixing in both the sump and upper volume of the tank. The Tulip reservoir also has multiple return lines to distribute retentate throughout the tank. As fluid levels drop, the system uses load cells to slow or stop the impellers before they reach the fluid surface to prevent high churning and foam creation, both of which can compromise target proteins. The valves also adapt the route of retentate flow to prevent air introduction and foaming from above the fluid surface level.

Buffer additions are not brought directly into the tank. Instead, buffer enters the retentate return flow to mix with solution before entering the tank and evenly distributing. The sump features baffled walls and a flow diverter, which routes flow along the walls rather than directly to the center above the outlet. The sump has an antivortexing outlet design to enable high flow rates at minimum working volumes. The Tulip tank used in this study has a working range from 3 L to 150 L without slowing filtration.

Questions and Answers

How do you recommend analyzing run-to-run consistency? Our fully automated TFF system can analyze parameters such as pH, conductivity, and UV absorbance. You can monitor membrane performance per batch for deviations.

Should the same mixing study be run at multiple scales? In small scales, active mixing isn’t often necessary because flow rates in conical bottles provide adequate recirculation. When scaling up, that recirculation flow isn’t enough. Mixing should be analyzed with scaling to ensure that TFF is consistent across all levels.

Find the full webinar online at www.bioprocessintl.com/category/webinars.