Downstream Processing (DSP): Membrane Operations

All right, well, welcome everyone to the Laney Biomanufacturing Program. This is the Scientific Instrumentation course, and tonight we are having a guest lecture on the topic of membrane operations and membrane training. And our guest is Chris Dahl, who is the principal, I would say, of Maverick Separation Solutions. And I met Chris over the summer while working at PowBio in Cell Valley Labs, and he was teaching me all about the DISTAC centrifuge. We had a talk on that recently. So Chris is also an expert on membrane separations for all kinds of industries. He's going to get into the different areas he works in in his talk now. So I'll just let him take it away. Chris. Thanks for the introduction. As mentioned, my name is Chris Dahl. I'm a consultant with Maverick Separation Solutions, and today I will be providing the basic membrane training. Some of the things that we're going to be covering today, basic membrane theory is what is membrane filtration, why we use membranes, membranes versus filters. We'll go into different types of membranes, such as molecular weight cutoff, materials of construction, membrane types. Then we'll go ahead and go into applications, such as process types, potential process locations for membranes. Then we'll go ahead and go into CIP and storage of membranes before we go into a very basic testing procedure that I like to use for membranes before we actually start running them. So first, basic main membrane theory. Membrane filtration is a physical separation where the media with different particle sizes are separated along a membrane surface by the use of flow and pressure. The size of particles separated from the media is dependent on the pore size of the membrane surface. Separation occurs through a semi-permeable barrier where the driving force is pressure, and the feed stream is divided into two different streams, the permeate stream and the retentate stream. The permeate stream is a stream that is clarified and should have particles that are smaller than your target particle size, while the retentate stream will have particles that are larger than your target particle size. Why do we use membranes? Well, it's easy to separate a target size particle with membranes. They can remove smaller particles and centrifugation, and they're often used after a decanter or a distac to polish off the centrate. They're also used before evaporators because they're more energy efficient than evaporation, and typically you will see these before an evaporator to pre-concentrate the feed material. They're also used in instances where you have temperature sensitive products that cannot be used in an evaporation system, so then they are used with a membrane filtration to pre-concentrate. Cross-flow and dead-end filtration are both filtration systems, but they do differ slightly. Dead-end filtration is when the flow goes perpendicular into a filter paper or a membrane, and you get a differential pressure, then you just shut off the filtration skid, whereas cross-flow is constantly pushing the filter layer across the membrane itself, and we like to use cross-flow over them because it's still easier to target your particle size versus dead-end filtration, because as dead-end filtration builds up the layer, the particle size that can pass through the filter layer is to get smaller and smaller until your differential pressure is too high. In cross-flow filtration, the filter layer is pushed along the surface, which keeps it constantly cleaner and allows the target particle size to pass through the membrane. It's almost self-cleaning, but every so often it still does require a CIP when your flux rates get too low. Because you're constantly pushing a filter layer across the membrane, it reduces fouling, which then extends your processing time, and it also gives you higher flux rates throughout the course of a run. Now we're going to talk a little bit about membrane pore size. There are four basic membrane pore sizes, starting with the Titus, which is reverse osmosis, then going into nanofiltration, ultrafiltration, and then the largest pore size would be microfiltration. Microfiltration is typically between 0.05 and 5 micron pore size, which is usually greater than a 200,000 molecular weight cutoff. Microfiltration is typically used for debris or cell removal, and proteins, sugars, and salts will typically pass through the membrane into your permeate. Ultrafiltration is the next size up. It has between 1,000 and 200,000 molecular weight cutoff, which is 0.05 to 0.002 micron pore size. This is typically used for protein separation, enzymes, stuff like that. It'll allow sugars and salts to pass through. You will see diafiltration in this step. It's often used to rinse the proteins and pass all the sugars and salts out the permeate so you have a nice clean protein stream. Nanofiltration is between 100 and 1,000 molecular weight cutoff, which is usually smaller than 0.002 micron pore size. This is used to retain sugars but allow salts to pass through. Again, you can use diafiltration to rinse off any kind of salts that you would want removed from your proteins or your sugars that are caught in the retentate of the nanofiltration membrane. And then finally, reverse osmosis. It's typically just a little bit tighter than nanofiltration, usually a little bit smaller than 100 molecular weight cutoff. It is around 0.001 micron pore size. Typically, these membranes have a thin film composite over the top of them to allow total retention of all particulates, allowing just the permeate to be clean going through. You will see these membranes in dealkalization of beer and wine as well, where you can add diafiltration water to the feed stream, which then rinses out the alcohol, out the permeate, so you have all your flavonoids and stuff still in your retentate side of your membrane. Now, not all membranes are created equal. Here's an example of the nominal pore size for three different membranes. Now, if these were all UF membranes with a nominal pore size of 100,000 molecular weight cutoff, you can see that the membrane on the far left has most of the membrane pores distributed right around that 100,000 kilodalt cutoff. The membrane in the center has a slightly wider distribution, while the membrane on the right has a very wide distribution. The reason you want a membrane with a very narrow distribution of your membrane's pore size is so that you don't have larger particles carrying over in your permeate. So, the membrane on the left would be a very high-quality membrane, where the membrane on the right would be a less expensive, cheaper membrane that's probably used in a wastewater application, where it's not as important to get the target particle size that you're looking for. So, just remember, when you purchase a membrane, you're going to look for one that has a nice nominal pore size or fits your process better. Don't always go by price. All right, next we're going to talk about the materials constructions of most membranes. Membranes are usually made of two materials, either ceramic or a polymeric membrane. Ceramic membranes are a ceramic support with alumina, silica, titania, or zirconia coating. They're usually created by casting, extrusion, or pressing, and they're typically a tubular or a disc-style membrane. Disc-style membranes are usually like a hollow fiber-style membrane, just in a disc form. These type of membranes tend to have a smaller surface area versus the footprint of the membrane. They're great for MF and UF applications. They do have a high resistance to solvents. They have a very high mechanical strength, and the great thing about them is you can back flush them, and you can also steam sterilize them. They do have a long lifetime. One of the downsides are they are an expensive membrane, which they try to counter with a long lifetime on them, so that way the capital cost up front is high, but the consumable cost down the line is less expensive. Polymeric membranes are a variety of different pore sizes. They come in MF, UF, NF, and ROs, and typically you'll see these in plate and frame membranes, hollow fibers, spirals, and MBRs, or membrane bioreactors. Some of the different materials that polymeric membranes are made out of, you have your cellulose acetates and your regenerated cellulose membranes, which have a good hydrophilicity, which means that they have a lower absorption for less fouling. The downside to these membranes is they do have a limited chemical resistance, and they typically have a lower pH range compared to most membranes. When you get into the polysulfone and the polyethersulfone membranes, they do have a very high chemical and mechanical resistance, but they do tend to absorb, which means they foul a little quicker, and you'll see lower flux rates on your permeates. And then there's combinations of membrane materials in them, whether they're PVDF or PSO, PA membranes, which kind of take the best from both worlds. They have a high chemical resistance. Some of them might be a little better with solvents, and they have a good mechanical strength with a little bit lower absorption than a polysulfone membrane or a polyethersulfone membrane. They may have certain sensitivities, such as chlorine, and always refer to your membranes cut sheet that tells you what the pH ranges are, the temperature ranges, and also any kind of sensitivities to them. So we've discussed what membranes are made out of. We discussed the tightness of them. Now we're going to talk about the different types of membranes. The basic types of membranes are tubular, hollow fiber, plate and frame, spiral wound, and MBR, or membrane bioreactors. Tubular membranes are great for MF and UF applications. They have a very high mechanical strength just because of the build of them. They're typically either ceramic, but there are some polymeric membranes as well. They can handle very viscous materials, high speed solids, and because of the design of them, they can be back flushed. You can either send the feed through the center of the membrane or from the outside and let the permeate go towards the center. They do have a relatively small surface area, but due to the diameter of the holes or the passages going through the membranes, they can handle very high solids and you don't usually need a pre-screen or anything like that for them. Hollow fiber membranes are similar to a tubular type membrane. They have extruded fibers through the center of them, which are much smaller, and they can still process feed with high suspended solids and viscous liquids. They're typically made of polymer, but there are some that are made of ceramic. You can back flush them from the permeate side of the membrane if necessary when they start becoming plugged. However, because the tubular diameter is much smaller than a tubular membrane, you need a pre-screen or a strainer in front of them to remove any kind of large solids that may plug them off. Plate and frame membranes are typically always polymeric type membranes. These ones are flat sheets that run on support plates. They have good for high solids loading and viscous liquids due to the channel length and the height of the plate and frame. However, these ones are polymeric membranes on some sort of backing material, such as polypropylene, and you cannot back flush them. Otherwise, the membrane gets unseated from the backing material, and it'll rip the membrane. The great thing is plate and frame membranes are a very low capital cost compared to any kind of ceramic, tubular, or hollow fiber membrane. And typically, you can order these membranes in one square meter sheets. You can order them for the size or shape of your plate and frame skid, or you can order them in the and just cut them yourself. The flow path for different plate and frames, they vary a little bit from skid to skid. Typically, they're going to be like the picture on your right, where feed comes in through the one side of the membrane. You have uniform flow down each channel, and then the retentate comes out through the other side of the skid, and your permeate is collected along tubes on the skid itself. Whenever the pressure drop becomes too great for these, they just add another loop to them, and you have another supply pump, so you always have a nice uniform flow without too much of a differential pressure or drop across these skids. The unit on the left is a pilot unit, and these are typically not scalable to production size units like the one on the right. And the reason is, is flow comes in through the top and hits a plate, which you can see in the bottom left side of the picture here. The flow then goes outwards on the plate to the on the peripheral, where it hits the next membrane plate, and then the fluid goes in towards the center, where it then hits another orifice, flows out to the next membrane, down towards the center of the next membrane. Because of this, there's a very high pressure drop across the membrane skid, and that's one of the reasons that we can't use this for direct scale-up. But the great thing about the pilot unit on the left is you can test multiple membranes. So you can test several different 10,000 kilodalton membranes at the same time and find out which one has the best flux rates, which one cleans the best, if there's any kind of reaction or retention towards certain particles. But you cannot set up the membrane skid, just membrane one, two, three, and four, because there's such a high pressure drop. So what you have to do is you have to get an average pressure drop across for each membrane. So you would stack the first type of membrane on the stack, and then one at the bottom of the stack. So the average is equal to whatever membrane is second, which is second from the top and second from the bottom. Third membrane would be third from the top, third from the bottom, and the fourth membrane would be two in the center. This way you have an average pressure across the entire disk stack for each type of membrane. Even though the membrane on the top membranes might have a little higher flux, the ones on the bottom, since they're lower pressure, will have a little bit lower flux. That way you can kind of get an idea of what the flux rates would be, as well as retentions and stuff like that. The last type of membrane is a spiral round membrane. They come in standard sizes. Each one is roughly 36 inches long, and they are 2.5 inches diameter, 3.8 inches diameter, 6.3, 8, and 8.3 inches in diameter. The 2.5 inch spiral membranes are a little bit different in construction. They're not always scaled up flow rates as the other membranes, but all the other ones are considered production membranes. The great thing about spiral round membranes is they have a very large surface area for a very small footprint. Again, these are very similar to the membranes that are in plate and frames. They cannot be back flushed because they are a polymeric membrane over a backing material, but the great thing is if you accidentally wreck one, you can easily do membrane replacement by pulling one out and dropping another one into the housing. Typically, these have a 1.1 bar pressure drop across them, and you can usually put three membranes in a housing. So this is basically showing the design of a spiral membrane. In the very center of the membrane, you have a permeate tube, and then you have some corrugated spacers that run down the membrane itself that separate out each of the membrane sheets. These spacers can either run parallel for more viscous liquids, or they can be crisscross pattern as shown in the picture here. The point of having them be in a crisscross pattern is it creates more turbulence, which prevents solids from settling on the membrane surface itself, and it allows them to run longer without plugging up. Now, if you look in the top left side of the picture, you will see several different types of plugs. Basically, each of these spiral membranes are kind of like a paper towel roll, where you have a center tube and then several membranes wrapped around it. If you hit one side with pressure, it kind of telescopes out. So we put these plugs on both sides of the membrane, so that way it prevents it from telescoping out and tearing the membrane. This is another diagram of the shape of the membranes. You have a tube in the very center with a membrane support plate, or a membrane support paper with membranes on both sides of it that create like an envelope, and then each of the spacers that kind of go in between it, and then all it is is just spun and wrapped into. And finally, MBRs, membrane bioreactors, are typically used in wastewater applications. They're a very low energy type of membrane. What they are is these packs of plates that are submerged into a tank, and then use air bubbles that are pushed from the bottom and filter and force the liquid upwards. As the feed material is rubbed up against these membranes, the permeate goes through the membranes and out with the effluent, and then you have a thick retention material inside the tank. They typically require 18 to 36 inches of head pressure to allow for better scouring of the air bubbles up the membranes. And whenever they become plugged, you can back flush them, and you can always drain out your tank or do a CIP through chemicals. So now we're going to go ahead and talk about some different applications for membranes. They can be used in a variety of different locations. You can use MF or UF membranes after a centrifuge for clarifying the effluent. In some situations, you can actually use an MF or UF membrane instead of a centrifuge. And of course, you can use them for an RO membrane before an evaporator to pre-concentrate your feed. And you can even use an RO membrane to purify wastewater streams, or in this case, the evaporator condensate. And then any of the permeate coming out of these ROs is usually clean enough that you can use upstream in the process for dye filtration or rinse water. There are three different ways of processing materials with membranes. They are batch, semi-batch, and continuous operation. Batch operation is exactly what it sounds like. You have a feed tank with a certain volume. You feed it to a membrane system. You have the retentate recirculate back to the feed, and the permeate is pulled out. In this type of application, you're just concentrating the feed tank down until you hit a certain point of concentration. Whenever you hit that point, you stop the system, you collect your retentate, you collect your permeate, and you're done with the run. This type of operation, you're going to see a declining flux rate as the system runs. And that's because when the tank is 100% full, it's at 1x concentration. When the feed tank drops down to 50% and you've removed 50% of the product through the permeate, you're now at 2x concentration. It's a lot thicker, so your flux rates will drop down. And as you continue to suck down in the tank, it's getting more and more concentrated, and your flux rates will reflect that as they continue to drop. Semi-batch is a kind of a play between batch and continuous operation. You have a feed tank that you're pulling into a membrane skid. The retentate recirculates back to the beginning of the membranes, and not back to the feed tank. The permeate is pulled out, and then you also pull out a little bit of retentate. The reason you run it this way is to get a true concentration pulled out of the retentate instead of having it drop down as you do in batch. Because of this, your flux rate will be a lot more stable throughout the run, although it would be lower than if you were to run 1x concentration through a membrane system. And then continuous operation is kind of like semi-batch, but with more loops onto it. You're pulling in from one membrane, the retentate is recycling back to the front of that membrane, while some of the retentate is going into another loop, where that retentate is recycled back to the second loop, and then some of it goes onto the third loop. The reason you do this would be to have a different size footprint for membranes on each side. You'd have different surface areas and still be able to achieve a very high concentration by the time the product actually comes out. And then of course, we do die filtration, which I touched on a little bit earlier. Basically what we do for die filtration is it's used to rinse sugars, salts, or any kind of material out of a feed stream or a retentate stream for a more pure product. And usually we do die filtration as a continuous fill into a feed tank at the same rate that we remove the permeate. This provides the cleanest type of purity. You can do a batch type method where you pull out 50 liters and add 50 liters of die filtration, but typically that rinsing isn't as effective as adding DI water at the same rate that you're pulling out your permeate feed. So where do the membranes fit in the process? In evaporators, you can use UF membranes to remove any kind of heat sensitive components. You can also pre-concentrate into the evaporator itself and then remove and polish off any of the condensate from the evaporator. In distac centrifuges, you can use them to pre-concentrate any of your feed. And you can also clarify the centrate when it comes out and then pre-concentrate any of the permeate from that centrate before you do any kind of actual concentration with an evaporator. And it's the same idea with the decanter. You can pre-concentrate the feed to a decanter or you can clarify the centrate afterwards. Now decanters typically have a dirtier centrate than distac centrifuges due to the g-force and the separating capacities of them. So you may have to have a pre-strainer or something like that before a UF membrane if you're going directly from the decanter into a membrane system. Some of the different applications that we use membranes in. This is just a limited selection. This isn't all of them, but we use them for juice clarification and concentrate. We use them for leachate and the paper industry, food coloring, protein concentration, enzymes, amino acids, algae extracts, yeast extracts, dextrose, evaporator condensate, blood plasma for both animal and human blood, dairy, carrageenan, vegetable proteins, gelatin, vinegar, beer clarification, wine clarification, dealkalization, stick water, tea clarification, coffee concentration, wastewater. There's multiple applications, so don't limit it to these ones, but they can be used in many applications. Next we're going to go ahead and talk about some membrane storage and CIP cycles. So for membrane storage, any unused membranes that you use, you want to seal them up and store them in a cool dry location. Polymeric membranes are shipped with a glycerin layer over the top of them to keep them from cracking and breaking. So whenever you have a used membrane, it should be cleaned and then stored in a slightly caustic solution to prevent biological growth. Once you clean or once you get a polymeric membrane wet, remove that glycerin layer from it. You have to keep it wet, otherwise it'll crack and then you just have to throw it away. And the other thing that you should know is if you're dealing with plate and frame membranes, try to keep the plate stacks together to prevent the membranes from tearing. So if you're changing from a plate and membrane stack to a spiral membrane stack, you're going to want to keep the plate stack together, since they're usually compressed very tightly together, and because of that, the polymers tend to stick to each other. So when you try to pull apart the plate stacks, you'll usually rip the membranes. All right, so some cleaning basics. The very first time you get a plate and frame or a polymeric membrane, you're going to want to condition it prior to that first use to remove the glycerin layer. After that, to do that, all you got to do is a basic CIP. It can be 15 to 20 minutes long at a, you know, 50, 60 C or whatever the membrane might be rated for, and close to the upper limits of its pH range. After you get done doing a very quick CIP, you do a clean water flush, and then you would do a clean water flux on the machine to verify what your membrane's flux rate actually is when it's in its cleanest situation. And you would record all the parameters for that, the flow rates, the permeate flow, the pertentate cross flow, the temperature, any kind of information, the pressures. So that way, whenever you do a CIP and you do a clean water flux after that, you can always compare it to that very first clean water flux to verify if your membrane is coming clean or if your membrane is getting close to the end of its life. You're allowed to lose 20 percent of flux rates after each CIP. However, if that occurs, you will have a very short membrane life. So you're going to try to get it cleaned back to the initial clean water flux each time you CIP the machine. You can use straight sodium hydroxides or acids to clean, but I do recommend that you use a membrane cleaning solution. There's a couple by Ecolab and Diversity that make great cleaning solutions for membranes that are made just for polymeric membranes. And the reason I recommend these is because they have detergents in them that help clean the membranes a little bit better than straight sodium hydroxide or acid. And they typically have buffers as well, so that way you don't overshoot your pH range for a membrane. And make sure whenever you CIP a membrane that you do it at a lower pressure, so you get a good cross flow across the membrane to help clean it better. And then also try to stay in the pH range that's listed for each type of membrane. Now a basic CIP cycle that we use, and this can occur whenever the membrane gets fouled up and the flux rates drop too low. The standard CIP cycle is based on the dairy industry, which is eight hours of runtime, four hours of CIP. However, you may find that your membrane systems can go 12 hours of running and only need a two hour CIP, or might only go four hours of running eight hour CIP. So some of these CIP cycles will have to be adjusted a little bit to get better results. But the first thing you're always going to do is displace your product. Now you do have to be aware when you displace your product, if you have a product that has a material that comes out, if you drop your pH or increase your pH, then you're going to want to flush it out with the same pH and temperature type of material or a buffer that you are feeding the product. You want to do this so that way when you put in DI water, RO water to flush out the system, this stuff doesn't precipitate out and bug up your membranes. So once you displace the product fully, then you can go ahead and flush the system with water, whether it's RO water or purified water. You want to stay away from city water because there are a lot of minerals and stuff in there that can help foul up your membrane. Once you flush out the system and your permeate, your retentate lines look fairly clear, you can then go into a hot caustic circulation. Typically this is a 30 to 60 minutes and you get as close to as hot as the membranes can handle at a very low pressure and close to the highest pH that the membranes can handle for CIP. Remember each of these are listed on all the membrane cut sheets that you would have, so don't exceed them. You will destroy the membranes. You want to keep them just below or right at it. After the 30 to 60 minute recirculation, you're going to do a water flush and a flux test. If your flux looks great, hey, you have a very easy product to clean and you can just store the membrane, get it ready for the next run. However, if you have any minerals or anything that precipitate out into the membranes, you might have to do an acid recirculation for 30 minutes. This is typically done around 40C and whatever the lower limit range of your membrane is. After your acid recirculation, you would do a water flush and another flux test. If your flux goes up higher, then more than likely you still have a lot of product in your membrane and you'd have to CIP with another caustic. Either way, you would probably do another caustic rinse and that's because acid tends to tighten up membranes. Actually, your flux rate should go down a little bit if your membrane was fully clean after the first caustic wash. This caustic wash after the acid doesn't have to be as long. It's 20 to 30 minutes. It doesn't have to be a super high pH or anything like that. If your membrane's rated to a pH 11, you can do it at pH 10. Then once you're done with that, you do a water flush and a final flux test to just verify if your membrane is fully clean or not. Finally, I'm going to talk about a basic testing procedure. The first thing you ever want to do with a testing procedure is just determine what the objective is. What kind of process you're going to be running, whether it's batch, semi-batch, continuous. If you're going to be doing dye filtration or not, what the size of the batch is, what the objectives are for concentration. Then you're going to go ahead and install your membranes once you decide if it's MF or UF or RO or NF that you're using. You'll clean the membranes and gather your initial water flux. Then you'll go ahead and drain the system and add your feed material and then start your pump. You're going to want to start it slowly so that way you don't get a big water hammer or anything like that into your membranes. Then allow the system to recirculate with the retentate and permeate lines back to the feed tank for 10 to 15 minutes. The reason you wait 10 to 15 minutes is to build up a little bit of a filter cake on top of the membranes. If you were to take a sample immediately after starting the system, you're going to have very high flux rates, which are not accurate because the membrane doesn't have that filter layer over it that starts and no fouling. After the 10 to 15 minutes, you begin your concentration. Then depending on the style that you're running, you're going to record your sample. You're going to sample and record your data. If you're running a batch type process, I recommend doing it at every volumetric concentration. If you're running a continuous operation, then I would go ahead and record every 15 minutes to every half hour because the concentration should be the same throughout the runs. Some of the different data points that you want to collect are the pressures, the transmembrane pressure, inlet outlet pressures, the temperature, the cross flow, the feed volume, the permeate volume that you've pulled out, any dye filtration volume that you've added, concentration factors, and then also any lab data such as pH, bricks if you're doing sugar type products, or if you're going to be doing alcohol, you take your alcohol data. And then once you're done running your test, at the end you would stop the system. You would collect your retentate, weigh both your retentate and your permeate so you can close up your mass balance, and then go ahead and flush out the system. All right. Thank you for listening to this. If you have any questions, feel free to contact myself or any other membrane person out there. Thank you. Great. Thank you, Chris. That was awesome. And how should people get a hold of you if they're interested in contacting you with questions? Is LinkedIn the best way? You can get a hold of me at LinkedIn, or you can get a hold of me at my email, which is Okay. Great. Thank you. And I did have a couple of questions for you since we have, looks like a little bit of time. You mentioned that dead end filtration at the beginning, is still used in industry. There's still some application for it. I was wondering what you have experienced with the use of filter aids like diatomaceous earth, and what are some applications and potential even hazards of those? Because I understand some of that stuff can be difficult to work with. Yeah. So diatomaceous earth is still used actually in a lot of food applications. And there's a certain volume where customers will use DE. You'll still see it in brewery and food applications. Now you do have to worry because I believe that diatomaceous earth is actually a carcinogen. So you don't want to have any of that stuff pass into your food products. So it's kind of ironic that it's used in so many food industries. It's still used because it has a very cheap capital cost, but they do have, you know, consumable costs for diatomaceous earth. And I think that's why they tend to use those over a lot of cross-flow technologies. Right. Right. And one of the questions I had was you talked a little bit about diatomaceous earth filtration in different scenarios. And another term I've heard a lot of industry is when they're doing what's called a buffer exchange. And I assume that's pretty similar, except instead of using DI, would you just use the replacement buffer in a diatomaceous earth filtration? Yeah. Yeah. You would just use the buffer. So because RO water, DI water is not always compatible products. So you're going to want to use whatever they use for a buffer solution in that situation. Right. Right. And that was leading to my next question, which was you brought up the issue of precipitation and that being a problem when potentially when you're CIPing, but it could be a problem anytime, I suppose. What are some potential ways precipitation could occur? I know pH change could be one, but what is it and also what are some ways to fix that? In other words, you know, how can you bring a membrane back from the brink of death, I guess? Yeah. Yeah. So precipitation, it occurs either usually pH or temperature wise. The easiest way with temperature is use the heat exchanger on a membrane system to prevent it from happening in the first place. However, if you're flushing out the system and you've been running at 50 C or something on your membranes and you go to flush it out with 30 C water and something precipitates out, sometimes if the products go back into solution, it's not a big deal if you raise your pH or raise your temperature back. If they don't, you may crash your membrane. And the problem with crashing a membrane is once it's done, it's done. And you just have to change it out. Now, ceramic membranes, you could probably, and tubular and hollow fiber membranes, you can back flush them to save them. But the way that the pores are designed on a lot of plate and frame membranes and spiral membranes, once they become, the solids get kind of impregnated in the pores, they don't really come out. And the good thing is the membranes are cheap, so you can just replace them, relatively cheap, so you can replace them. Whereas you can't really do that with the hollow fiber or ceramic because they're so much more expensive. Right, right. So assuming you are able to bring it back, you mentioned performance testing of membranes. It sounds like that needs to be a pretty regular process that you do with your membranes to make sure they're still performing. Are there other ways to performance test other than the clean water tests that you mentioned? The clean water flux test is the most common one that I've actually come across. You usually do it after every CIP and after every CIP step you do. And typically you can see when you first do your very first caustic wash after a flush, you'll see that the, you'll see a lot of product actually turn color in the in the feed tank or in the recirculation lines. And normally if you don't come all the way back on your fluxes after the first caustic wash, you would do your acid wash. Or maybe you have to remove that caustic and then restart another caustic wash because you fouled up, you know, too much of the of the CIP solution. And different cycles will be a little bit different. Like a coffee concentration is one of the hardest things for us to clean on polymeric membranes. And sometimes it's not just a CIP solution. You might have to add chlorine or something like that as well. Now certain membranes can handle chlorine, other membranes cannot. Like you don't want to put chlorine to reverse osmosis, but some UF membranes can handle it. And if you're really having a hard time getting a membrane clean, I know Ecolab, and I think Diversity does too, they actually have pilot labs where you can send in fouled membranes or product to fouled membranes and they'll actually come up with a cleaning solution that they can make for your process. Oh, interesting. Interesting. All right. Yeah, one question I had to, I don't know if you can go back in your slides a little bit, but you had a slide, if they're down it's okay, but you had a slide on continuous operation of membranes. And I was curious with the systems that are out there for like, I guess pilot scale and larger, you have a lot of automation possible. And when you have say multiple membranes, let's say in a series, I was just curious, like what, what's possible with automation with some of these systems and, and how, how much of these different, you know, process considerations can be automated away from the user. So there's less, you know, interventions and things like that. Yeah. So the, unfortunately pilot labs don't have a lot of automation. Pilot lab membranes are typically valves or pumps that you increase VFD frequency or open the valve a little bit more. Now, when you get up, so like say Alfa Laval, they have an M20 unit that's fully a manual machine, but when you get up into their first scalable piece of equipment, it is fully automated. They actually have automated back pressure control valves for your transmembrane pressure. They actually have a VFD is controlled by HMI. So it'll increase or decrease your cross flow as necessary to, to maintain a certain cross flow through your membranes or a pressure through your membranes. Most of the systems above a certain size are fully automated. So it's literally just push start. And then when you're done running, you push CIP button, and then it goes into a CIP and adds all the chemicals and everything automatically. When you're in the small pilot lab size, almost everything is going to be manual, including the dosage of chemicals or anything like that. Okay. And one of the things I was wondering about, as you mentioned, de-alkalization of beer, and I was really curious about the point you made about how, what they would do is use a, it sounds like use a nano filter, but also use di-filtrate to bring back in some of the flavor. Can you talk more about that? Yeah. So de-alkalization of, of wine and beer, it's starting to gain a little bit of traction in the industry. Basically they use either a very loose reverse osmosis membrane or a nano filtration membrane. And the idea of this is to retain all the flavonoids from the beer or the wine, use RO water, di-filtration water to rinse out. Basically what you're doing is you pre-concentrate by using an RO membrane to cut your volume in half. What that does is send all, half the alcohol or half the volume is alcohol that goes out with the herniate. And then you re-dilute with, with RO water, di water until you, and you do this until you basically get alcohol that, or a beer or a wine that's below that 0.5% or whatever it is threshold. And you usually do this after you've already removed all yeast. And that way you don't have any extra fermentation occurring after de-alkalization. Now I must say if you don't keep the temperature consistent, you do kind of lose some flavor and you can kind of counter that by brewing stronger beer. So that way you don't lose as much of the flavoring, you know, cause it does kind of have a little bit of a watered down flavor. And there, there's some ideas out in the industry on how to better de-alcohol the, the beer or the wine using membranes. But as of right now, for the most part, they're just using DI water to rinse out any, any alcohol in the stream. And then then you have your low alcohol beverage. Great. Thank you. My last question is one of the products you mentioned on your product list, there was yeast extract and you had a couple others on there that are pretty low value products. And I was wondering what, what typically when you have a very low value product like yeast extract, you know, the cost per gram or kilograms pretty low, what do you recommend for considering membranes in terms of costs? Like what's the most cost effective way to approach membrane infiltration? Well, typically there's a range when it comes to what the cost is per flux rate. So when a member typically on like a yeast extract or so on most polymeric membranes, I would say it's when the flux rate drops below eight liters per an hour per square meter of membrane surface area, that's kind of the cost cutoff. Now different areas, you have to look at the, you know, what the energy consumption is. If you're in California, of course, energy consumption is expensive and so is water. So they may not be as profitable as it is running it out and say Iowa or someplace where a lot of those manufacturing facilities are located at. That makes sense. That makes sense. Great. Well, that's all my questions, Chris. I really appreciate you doing this for us. And it was great talk. And if you have any final comments, feel free. I appreciate it for the opportunity to share a little bit of my knowledge to everybody. All right. Well, thank you, Chris. All right. Thank you.