What is the impact of feed concentration on ultrafiltration cassettes?

When I first started working with ultrafiltration (UF) cassettes back in my early days as a filter supplier, I’ll admit I didn’t give feed concentration much thought beyond checking if the solution I was processing was “thick enough.” That changed fast—when a biotech client hit a wall with their antibody concentration step, blaming frequent clogging and weirdly inconsistent flux rates. Turns out, they were pushing way more than their UF cassette could handle, and it cost them days of downtime and a batch of expensive product. If you’re working with UF cassettes, this isn’t just a tiny detail—it’s one of the biggest levers that makes or breaks your process, from yield to run time. Let’s break down how feed concentration impacts every part of UF, and what I’ve learned over the years to keep things running smooth.

First, let’s ground this in what UF cassettes actually do, just to make sure we’re on the same page. If you’re new to the game, a UF cassette is a self-contained unit with flat membranes (check out our Flat Sheet Cassette page for the full breakdown) that separates molecules by size—small stuff passes through the membrane, while bigger ones get “retained” and concentrated on the feed side. What makes them so popular in biopharma, food and beverage, and wastewater? They’re compact, reusable (sometimes, if cleaned right), and way more consistent than older, hand-cut membrane setups. The Cassettes Ultrafiltration line we offer is built for scalable runs, from small lab batches to full manufacturing, which is why we’ve worked with clients ranging from startup biotechs to global food brands.

Feed concentration, for those who haven’t deep-dived the terminology, is basically the amount of target (or unwanted) solute in the solution you’re pumping into the cassette. For example, if you’re concentrating monoclonal antibodies (mAbs) from a cell culture broth, feed concentration here is how many grams of mAb are per liter of that broth. It’s not just the target, either—you’ve got impurities like cell debris, media components, and leftover buffer salts mixed in too. The sweet spot for feed concentration is different depending on your membrane’s molecular weight cutoff (MWCO), right? A 30KD UF cassette works great for small proteins, while a 500KD one is for larger molecules like viral vectors. I’ve seen clients grab the first cassette they look at, regardless of their feed concentration, and that’s where the mistakes start.

Let’s start with the most immediate impact: flux rate. Flux is the volume of solution that passes through the membrane per unit area per hour—it’s the backbone of UF productivity. When your feed concentration is too high, flux drops fast, and not in a linear way. Here’s why: as you process the solution, the retained solutes build up near the membrane surface, forming what’s called a “concentration polarization layer.” That layer acts like an extra barrier, so water and small solutes have to work harder to get through. If the feed concentration is extra high to begin with, that polarization layer gets thick fast, and soon you’re barely pushing any permeate through. I had a client last year running a 30KD Ultrafiltration Cassette on a dense cell culture supernatant—they were feeding at 25 g/L of protein, when our recommendation was 10–15 g/L. Their flux dropped from 60 L/m²/hr to 12 L/m²/hr in 2 hours, and they had to shut down to clean the cassette prematurely, wasting 12 hours of run time. Even worse, the high concentration polarization led to some proteins getting stuck in that layer, so their yield was 18% lower than expected.

But it’s not just flux—feed concentration also directly ties to membrane fouling, which is the big boogeyman of UF. Fouling is when solutes, debris, or aggregates stick permanently to the membrane surface or pore walls, and it’s way worse with over-concentrated feeds. When solutes are crammed together in the feed, they’re more likely to collide with the membrane and get trapped. I’ve cleaned hundreds of UF cassettes over the years, and 80% of the time, heavy fouling traces back to feed concentration that was 2–3x the recommended range. For example, a food client processing fruit juice wanted to concentrate to 30° Brix in one pass, instead of 15° Brix (which is the sweet spot for their 10KD cassette). The juice’s pectin and sugar started building up on the membrane within 30 minutes, and by the end of the run, the membrane pores were partially blocked, so they only got a quarter of the concentration they expected, and had to replace the cassette after three runs instead of 10. That added up to way higher costs per liter than if they’d split the concentration into two passes with a lower feed concentration each time.

Yield is another area where feed concentration hits hard, especially in biopharma where every gram of product matters. When your feed is too concentrated, not only do you get more fouling, but you also start seeing product loss in two ways. First, the concentration polarization layer can catch target molecules, and if it gets too thick, those molecules might be carried away with the retentate when you drain the cassette, instead of being concentrated. Second, some target molecules (like proteins) can start aggregating at high concentrations, and those aggregates are often small enough to slip through the membrane or get stuck in the permeate, so you lose them. A viral vector client I worked with last year was using our 500KD Ultrafiltration Cassette to concentrate AAV vectors—they pushed the feed concentration to 1e13 vector genomes per mL, which is way over our recommendation of 2e12. Their yield was only 45%, vs. the 82% they got when they dropped the feed concentration to the right range. When we tested the retentate, we found that 30% of their target vector was stuck in the concentration polarization layer, and another 15% was aggregated and lost. That’s the kind of hit that can make a whole batch worthless, especially when you’re dealing with GMP-compliant processes where you can’t just make more product on a whim.

Wait, but before you think “lower feed concentration is always better”—that’s not true either. If your feed is way too dilute, you’re wasting time and cassette capacity. UF is most efficient when you’re concentrating the solution to a reasonable level, but if you start with a feed that’s too thin, you’re moving way more volume than you need to, which increases energy costs for pumping and takes longer to get to your target concentration. For example, a lab client concentrating a small protein sample was using a 10KD cassette with a feed concentration of 0.1 mg/mL—they were trying to get to 10 mg/mL, so they were processing 100 mL of solution to get 10 mL of retentate. It took them 8 hours, vs. 2 hours when they concentrated the feed first via centrifugation to 0.5 mg/mL, so they only processed 20 mL. That’s a huge difference in labor and time, especially if you’re running multiple batches a day. The key is hitting that middle range—what I call the “concentration window” for your specific cassette and application.

So how do you figure out that concentration window? It’s not one-size-fits-all, but there are a few rules of thumb we share with all our clients. First, check your cassette’s MWCO: a cassette with a higher MWCO (like 500KD) can usually handle a higher feed concentration than a lower MWCO one (like 30KD), because the pores are bigger, so polarization is less of an issue. Our Cassettes-TFF-filtration Series page has specific concentration ranges for each cassette, but as a general rule, never go above 10–15% total solids for most protein applications, and 5–10% for small molecules. Second, test your feed for impurities: if your feed has a lot of debris or aggregates, you need a lower feed concentration to avoid fouling, no matter what your target molecule is. Third, run small-scale tests first—don’t just jump into full production. A 100 mL or 1 L test run will tell you exactly what flux you get, how fast fouling sets in, and what yield you can expect, so you don’t waste valuable product.

Another thing I’ve noticed over the years is that feed concentration interacts with other process parameters, like crossflow rate and transmembrane pressure (TMP). It’s not like adjusting concentration in a vacuum—if you have a high feed concentration, you can compensate a bit by increasing crossflow rate, which sweeps the polarization layer away from the membrane surface, and increasing TMP (but not too much, because TMP can cause compaction of the polarization layer). But this is a balancing act: higher crossflow means more energy, and higher TMP can lead to membrane damage if you’re not careful. We always advise clients to run full parameter tests that include feed concentration, because changing one can throw the others off. For example, a client once tried to fix a flux drop from high feed concentration by cranking up TMP, and ended up compacting the membrane, so their next runs had even lower flux. That’s why at our company, we don’t just sell cassettes—we help our clients work through these process kinks, because we’ve seen almost every mistake you can make with feed concentration.

Let’s talk about long-term effects too, because feed concentration doesn’t just impact a single run—it impacts the cassette’s lifespan. If you’re running with over-concentrated feeds, you’re creating way more fouling, which means you have to clean the cassette more often, and more aggressive cleaning agents (like higher pH or temperature) to get rid of the buildup. Over time, that aggressive cleaning degrades the membrane, so the cassette has fewer runs in it before it needs to be replaced. I’ve had a client that was running our Flat Sheet Cassette, and by the 5th run, they had to replace it because of heavy fouling, while another client running the same cassette at the right feed concentration was getting 15+ runs. That’s a huge cost difference when you’re going through cassettes every week vs. every month.

Now, let’s get practical—what should you actually do if your feed concentration is off? First, measure it properly. A lot of clients go by “visual thickness” of the solution, which is super inaccurate. Use a Brix meter for food products, or absorbance (like A280 for proteins) to get a precise concentration. Then, adjust your feed either by diluting it (if it’s too high) or concentrating it further with a smaller pre-processing step (like tangential flow filtration with a different cassette, or centrifugation) if it’s too low. If you’re dealing with a one-of-a-kind feed, like a custom biologic, reach out to our team—we have application specialists who can help you design a process tailored to your feed concentration, whether that’s picking the right MWCO (like 30KD vs. 500KD), adjusting crossflow, or figuring out the best run time.

I can’t stress enough how much small tweaks to feed concentration can change your whole UF process. It’s not a boring detail—It’s the difference between a run that hits all your yield and timeline goals, and a run that leaves you troubleshooting for days, wasting product and money. We’ve built our Cassettes Ultrafiltration line to be flexible, with options for different MWCOs (like our 30kd Ultrafiltration Cassettes and 500KD Ultrafiltration Cassettes), flat sheet designs, and TFF series setups, because we know that every client’s feed is different. If you’re struggling with feed concentration issues, or just want to make your UF process more efficient, we’re here to help—we’ve worked with thousands of clients to get their concentration windows right, and we can do the same for you. Don’t let a small detail like feed concentration ruin your runs—reach out to our team to learn more about optimizing your process today.

References

1. Belfort, G., Davis, R. H., & Zydney, A. L. (1994). The behavior of suspensions and macromolecular solutions in crossflow microfiltration. Journal of Membrane Science, 96(1–2), 1–58.
2. Charcosset, C. (2007). Hydrodynamic and concentration polarization in ultrafiltration: A review. Desalination, 211(1–3), 108–122.
3. Zydney, A. L. (2018). Tangential flow filtration for biopharmaceutical processing. Current Opinion in Chemical Engineering, 20, 109–115.
4. Mulder, M. (1996). Basic Principles of Membrane Technology (2nd ed.). Kluwer Academic Publishers.