Improving Particle Capture Predictions with COMSOL

[latexpage]

In Episode 3 of this series, I have begun working towards improving my COMSOL modeling for particle capture. In addition to running simulations, I need to run experiments to try and correlate real world solutions with simulated results. One of the major focus points of my work has been trying to explain the shape of a capture curve for different supply () and ultrafiltration () rates as seen in Figure 1: 

Figure 1: Comparison of plug flow (yellow) capture results with parabolic (gray, orange, blue) capture results from COMSOL particle tracking.

In these simulations, I have varied the flow rate ratio (by varying the ultrafiltration rate) to observe the expected percentage of particles captured. Quite simply, the flow ratio on the x-axis is the fraction of the flow that goes through the membrane. It is therefore expected that as the rate of ultrafiltration decreases, so will the particle capture percentage. Intuitively we may think that this obeys a linear relationship, however in practice we can see that this is not the case. For a parabolic flow system, we observe that the particle capture rates are worse than expected at higher ultrafiltration rates, while at lower ultrafiltration rates the capture is significantly better than predicted. The only scenario where we observe a 1:1 flow ratio to capture correlation is when we have full slip in the system or “plug flow”. In the plug flow case, we see that if we have 75% of the flow going through the membrane then we have ~75% particle capture. However, with parabolic flow conditions we see that we only expect 55 – 60% capture.

In addition to this, I have also generated higher temporal resolution videos of the particle capture process, which can be seen below:

Video 1: 100 μL/min supply and 0μL/min ultrafiltration

Video 2: 70 μL/min supply and 60 μL/min ultrafiltration

Video 3: 70 μL/min supply and 60 μL/min ultrafiltration side view.