Protien Skimming: How It Works

This article is intended to give detailed background information about how skimmers work in conjunction with marine aquaria. Since skimmers vary considerably in design and represent an evolving technology, it will not attempt to show that one design is best. Instead, it will describe some of the physical principles behind skimming in a more detailed fashion than is presented in marine hobby books. The greater goal is to help people understand how their tanks work. If it has that effect, then it is a success.

Basic Principles

Before getting into the details of skimmers, it is useful to nail down some critical chemical definitions.


Molecules, such as the organic molecules found in marine tanks, are often described as being either hydrophobic or hydrophilic. Hydrophobic simply means water hating (hydro meaning water, phobic meaning hating). Likewise, hydrophilic means water loving. Examples of hydrophobic molecules are methane, nitrogen gas, oil, fat, cholesterol, most of the molecules in gasoline (e.g., hexane), lighter fluid (butane), ether, some vitamins (e.g., A, D, E, K), and most of the refrigerants (e.g., chlorinated fluorocarbons (CFC’s)). These do not mix with or dissolve in water to any appreciable extent. Examples of hydrophilic molecules are water, salt, sugar, ethyl alcohol, ethylene glycol, glycerin, glucose, ammonia, most amino acids (e.g., glycine), some vitamins (B6, B12, Biotin, C, Niacin) and almost all inorganic compounds. These molecules are all much more soluble in water than in oil. There is, in fact, a continuum of molecules from the most hydrophobic to the most hydrophilic, so it is rarely correct to state that a molecule must be either completely hydrophobic or completely hydrophilic. Some molecules that fall into the middle of this continuum include formaldehyde, aspirin, phenol, many fragrances, rubbing alcohol (isopropanol), and acetone. Some large molecules can have portions which are hydrophilic, and other portions which are hydrophobic. Fatty acids, many proteins, soaps and detergents, and a wide variety of biological molecules fall into this category. These are often called amphipathic or amphiphilic. [Note: don’t confuse amphipathic with amphoteric. The latter describes something with both acid and base properties, like bicarbonate.]


As almost everyone knows, skimmers function by generating a large amount of air/water interface. All commercial aquarium skimmers do this in the form of air bubbles suspended in water, though the line between air bubbles in water and water droplets in air is a fuzzy one in some downdraft skimmers. Other configurations, such as flat surfaces, are also possible. Organic molecules which are hydrophobic and those which are amphipathic will collect at this interface, for reasons explained later. An oil scum seen floating on water is a perfect example of absorption at the air/water interface. As the bubbles in a skimmer start to collect together (simply under the influence of gravity forcing them to the top of the collection chamber), they begin to interact and form a foam. Foams form when bubbles approach closely, and the water trapped between them is allowed to drain. The more draining that takes place, the dryer the foam. This dry foam, which still contains some water along with the organic molecules, can then be collected and discarded.

For a skimmer to function maximally, the following things must take place:

1. A large amount of air/water interface must be generated.
2. Organic molecules must be driven to and/or allowed to collect at the air/water interface.
3. The bubbles forming this air/water interface must come together to form a foam.
4. The water in the foam must drain without the bubbles popping prematurely.
5. The drained foam must be separated from the bulk water, collected, and discarded.

Anything that alters skimming efficiency must be impacting one of these things. I’ll try to explain each of these requirements in turn, and what things impact their efficiency.

The First Step: Air/Water Interfacial Area

Why is a large amount of surface area required? This question goes to the root of why organic molecules absorb at this interface. The fundamental reason is that the interaction between two water molecules is much stronger than that between a water molecule and a hydrophobic organic molecule, like oil. Water forms hydrogen bonds to other water molecules and certain other hydrophilic molecules, but not to oil. This interaction between water molecules is very strong, and has a large impact on the properties of water. Thus, if an oil molecule is buried down inside of water (i.e., dissolved), it is essentially “getting in the way” of water molecules that want to interact with each other. Squeezing the oil up to, and out of the surface of the water eliminates this interference, since the water molecules at the surface of the water do no have anything above them to hydrogen bond with (air isn’t any good for this, it is too ‘thin” in that there is hardly anything there to interact with). This effect is called the “hydrophobic effect”, even though it is really driven by hydrogen bonds in water, not by the hydrophobe at all.
If the hydrophobe under discussion is oil, all of the squeezed out oil molecules can ball up, forming a second phase of oil, as is observed on mixing olive oil and water. If one is talking about amphipathic molecules, however, their hydrophilic ends still want to interact with the water (because these ends can form hydrogen bonds, or other types of strong interactions with water). Thus, the best that these molecules can do is squeeze their hydrophobic portions out of the water, leaving the hydrophilic portion in contact with the water.

In actual practice, most organic molecules found in a marine tank will be amphipathic, with the bulk of the remainder being hydrophilic. There will be relatively few purely hydrophobic molecules (e.g., fat) in the tank. Most very hydrophilic molecules will not be removed by a skimmer, so understanding how amphipathic molecules react in a skimmer is the key to understanding how a skimmer works. One reason that skimmers are often referred to as protein skimmers is that most proteins are amphipathic. They often have an interior made from hydrophobic amino acids, and an exterior made of hydrophilic ones. When dissolved in water, only the hydrophilic exterior portions contact the water molecules. When placed in contact with an air interface (or something that is hydrophobic) the proteins will alter their shape, and present the hydrophobic portion to the interface. In this fashion they are readily attracted to an air/water interface.

How Much Absorbs at the Interface?

So what does all this mean for a skimmer? It means that only a monolayer of amphipathic molecules can form at the air/water interface. In other words, only a single layer of molecules can form at the air/water interface which have their hydrophilic tails in the water and their hydrophobic heads exposed to the air. Unfortunately for marine aquarists, a monolayer is hardly anything. A monolayer of soap comprises something like 5 x 10(14) molecules per sq cm, which corresponds to about 0.0025 g/m2. In order to take out 1 g of soap as a monolayer, one would need to generate over 3500 sq ft of surface area. Certain things can change this number significantly, but in general, this is why we need to generate so much surface area. One way to think of this is to look at the surface area of a typical tank. A four foot by 2 foot tank has 8 sq feet of surface area (0.7 sq meters). If you had a monolayer of organic molecules at this interface, and suddenly removed all of them, you would have only removed 0.002 grams of organics. Since feeding a teaspoon of shrimp to a tank will add several thousand times this amount, one can quickly see that the need for generating large amounts of surface area is called for.

How To Generate Air/Water Interfacial Area

The name of the game in recent improvements in skimmers has been to develop improved ways to generate large amounts of air/water interfacial area. Downdraft skimmers and needle pinwheel types are mostly designed to increase this interfacial area. Any process that breaks up water and air into fine bubbles will work. In terms of bubbles in water, the smaller the bubble, the greater will be the surface area. In fact, for a sphere, surface area goes as the square of the radius (S = 4*pi*r2) while volume goes as the cube (V= (4/3)*pi*r3). Consequently, one bubble that is 1 mm in diameter contains 0.52 cubic millimeters of gas and has a surface area of 3.1 square millimeters. Alternatively, if we have 1,000 bubbles one tenth the size (0.1 mm) then the volume of gas is still 0.52 cubic millimeters, but the surface area is now a whopping 31 square millimeters.

In practice, a lower limit to bubble size is reached in skimmers where making the bubbles smaller precludes them from rising by gravity to be collected. One can easily see this in a marine tank. Swishing an object through the water will result in some large bubbles that rapidly rise, and some smaller ones that are much slower to rise. A small enough bubble may take hours to rise to the top of a collection unit. An analogy is dust in the wind. Big objects (rocks, etc.) will quickly drop out of air, but fine dust may stay suspended for days. Designing a skimmer is thus a trade off between bubble size and collection time. The only other way to win the game is to generate large numbers of bubbles. As an academic consideration, it is not essential for one to generate the interface as bubbles in water. Drops of water in air (which may, in fact, occur in portions of some downdraft skimmers), or even a rapidly turned over flat surface could be just as effective. For practical reasons, mostly relating to gathering and removal of the collected organics, air bubbles in water seems to work best.

What Collects at Air/Water Interfaces and Why?

An obvious question about skimmers is what is collected, and why. Let’s start with the why, as in why molecules absorb at this interface. As stated earlier, hydrophobic molecules are squeezed out of the water because of the hydrogen bonds formed between water molecules. But some obvious questions remain:

1. Why does skimming work better in salt water than in fresh?

It is not that bubbles form more readily in salt water. In fact, in highly purified salt water (no organics at all) the surface tension is greater than in pure water, and bubbles will be harder to make. Any “evidence” that bubbles form more readily in salt water is due to organics present. remember, only 0.002 g of an amphipathic molecule would be sufficient to completely cover a large aquarium. Thus, only a tiny bit of contaminating organic can make salt water appear to form bubbles readily. One of the main reasons for better skimming in salt water is actually the reduced solubility of organics, especially hydrophobic ones. Since organics are typically less soluble in salt water, they are more easily squeezed out of it to an air/water interface, and collected as foam. This is the basis for the well known salting out effect of proteins. Quoting from a basic biochemistry text: “At sufficiently high ionic strength a protein may be almost completely precipitated from solution, an effect called salting-out.” It is not impossible, however, to skim fresh water. Rivers from certain areas of the northeast US often have foam on them which comes from tree sap and other natural things that enter the water. They have a low solubility in water, and are easily collected as a foam. Other organics, with some solubility in fresh water, just have less attraction for the air water interface, and are consequently harder to skim.

2. Are inorganics removed?

There are few, if any, natural inorganic molecules that will absorb at an air/water interface on their own. Nearly all inorganics in a marine tank are highly polar, charged ions, which will actually be excluded from the interface for the same reason that hydrophobes are attracted there. These inorganics interact more strongly with water than even water does with itself. Thus, to expose these at the water surface would require a severe energy penalty. Inorganics can, however, be complexed to organics that are skimmed out. Metal ions, like iron, for example, can be complexed by humic acids and other organics (citrate, EDTA, etc.), which themselves are skimmed. They will also be skimmed if they are contained inside of a microorganism that has a hydrophobic exterior (many do) and are skimmed out. It is unlikely that iodine, in any natural aqueous form, will be removed by skimming, either alone or in a complex with organic molecules, at a relative rate higher than chloride is removed. Iodate complexes with certain organics might be an exception, but even the formation of such complexes has not been established. However, microorganisms which themselves may take up significant amounts of iodine, may subsequently be skimmed. If there is any basis to the widely held (and probably incorrect) idea that iodine is removed significantly by skimming, it is most likely in the form of microorganisms. Of other ions of concern to aquarists, neither nitrite, nitrate, nor phosphate will be removed directly. Phosphate might be incorporated into certain inorganic particulates, like CaCO3/MgCO3, which could be skimmed if coated with organics. Of course, calcium and magnesium in these particulates are also removed. Ammonia might be blown off since it is always in equilibrium with atmospheric ammonia gas, and strong aeration will eliminate some of it.

3. What else is removed?

Nearly any hydrophobic or amphipathic molecule can be skimmed. This includes amino acids, vitamins, proteins, carbohydrates, fats, many of the combination biomolecules (e.g., lipoproteins), RNA, DNA, etc. . This list includes most, but certainly not all organics. The removal of microorganisms by skimming was mentioned previously. This might have positive effects in the sense of nutrient transport from the tank. The reduction of undesirably high levels of bacteria, pathogens, and dissolved algae might also be a benefit. On the other hand, skimming almost certainly removes many micro- and even macroorganisms from the water column that might otherwise become food for tank inhabitants. It is not clear how large of an impact this has, but it will certainly depend upon the type of tank inhabitant that is being considered. I would also expect that many toxins and slimes produced by tank organisms will be removed to varying degrees by skimming. Some will be readily removed, and others more slowly.

4. What organics would not be removed?

All of the highly polar organics will not be removed. Simple sugars, acetate, oxalate, methyl alcohol, choline, citrate, etc. will remain behind. They simply are not attracted to the air water interface. Most charged species are, in fact, repelled from the air water interface, so they will not be collected.

Allowing Time for Absorption

Once the skimmer has generated a large amount of surface area, the next issue involves allowing enough time for organics to actually diffuse to the interface. How long does this take? That’s an important question without a perfect answer. Diffusion of molecules in water can be slow. For very large molecules, like proteins and carbohydrates, it can be very slow. It might take hours for a protein to diffuse a few inches in water. Fortunately, we do not need to rely on pure random diffusion to carry organics to the surface. Nearly all skimmers have bubbles in a turbulent environment, where they can be carried around by water flow as well as by diffusion. As they approach the bubble surface, however, movement of water relative to the bubble will be greatly reduced, and diffusion will be necessary for the final travel to the interface. The amount of time necessary for complete accumulation of organics at the surface will also depend upon the concentrations of organics in the water, and even on the chemical nature of the organics present. It makes perfect sense that in water with high levels of organics, the interfacial area will be rapidly occupied by organics. That is because there are enough in the local area around the bubble to saturate the interface. When the concentrations are lower, organics have to diffuse from farther and farther away from the bubble to saturate it. Additionally, different organics have different strengths of binding to the air/water interface. Things which have a strong preference will slowly replace those already at the interface which have a lower binding strength. Thus, a bubble which is completely occupied with organics might still be changing with time on exposure to tank water. It will not, however, go on increasing its organic load indefinitely. For these reasons, one cannot readily state that a certain amount of time is necessary for organics to fully saturate bubbles. Further, it is incorrect to claim that it is always better to increase the contact time between bubbles and the tank water. Likewise, the way in which the bubbles move relative to the water is important. If the bubbles are moving against the water flow, or are in a turbulent environment, the required absorption time will be lower (because the water flow helps bring organics to the interface) than if the bubbles are moving with the water flow.

Foam Formation and Draining

Once a skimmer has a large number of bubbles coated with organics, it is necessary to somehow remove the bubble surfaces, but not the majority of the water nearby. This is most easily accomplished by allowing the bubbles to form a foam. Foam formation takes place when bubbles come together. The froth of bubbles begins to drain under gravity, removing much of the water between the bubbles. Some of the bubbles merge into larger bubbles. As long as the bubbles do not pop before significant draining occurs, then the organics will be left behind, along with some residual water. Eventually, the concentration of organics on the top of the foam becomes great enough that they exceed the solubility limit, and small particulates of organics form. These particulates are what is collected from a skimmer, along with some water and organics that are still present in solution or at the air/water interface. Foam draining is a critical stage for most skimmers. One problem with draining, in my opinion, is that some organics are washed away with the draining water. There is always an equilibrium between organics in solution, and those actually attached to the interface. As water continues to drain, some of the organics are lost. Further, as some bubbles pop and organics are redistributed into the nearby water, the local concentration of organics in the water between bubbles in the foam can rise to concentrations far higher than are present in the tank. For this reason, I believe that the greatest skimming will come from removing a relatively wet foam, rather than waiting for this same wet foam to drain prior to removal. The only difference between a wet foam, and one that has drained more to form a dry foam, is that additional water, and some organics, have drained away. I believe that this important point is often neglected.

Bubble Popping

Other critical things can happen at this stage, and they usually impact skimming negatively. One is the addition of things that cause bubbles to pop prematurely. That is, things that cause bubbles to pop before they have drained and can form particulate organics or be removed. Oils, for example, cause this to happen. When oil droplets are added to a tank, they quickly arrive at the skimmer. An oil droplet is hydrophobic on all sides. Oil drops work their devilish tricks by spanning across the water between two air bubbles in a foam. Once an oil droplet completely spans the water, it causes an instability. Bubbles are a remarkable balancing act between surface tension and the “hydrophobic effect”. All molecules have some attraction for each other, but water molecules form an especially strong interaction (hydrogen bonds). Water molecules at the air/water interface have nothing above them (only air) and are thus only able to hydrogen bond to things below them. Since they cannot form such good interactions, they are less “happy”. Surface tension is thus the effect of all of these hydrogen bonds pulling at the water molecules on the surface. The net effect is that the water minimizes its surface area. Surface tension is why water drops are nearly always spherical: a sphere has the lowest surface area for a given water volume. So what does this have to do with a foam? A foam has a very high surface area, and the surface tension of the water is always trying to reduce the surface area. Popping of bubbles is one way to quickly reduce the surface area. This is the reason that bubbles formed in pure water pop almost instantly. Try it with tap water: your water will probably not be able to support bubbles for more than a second or two.
In water with organics, the organics greatly reduce the surface tension, and thus the tendency for popping, but do not eliminate it completely. [Note that a low surface tension is not the only requirement for stable bubbles. Pure organic solvents also do not form stable bubbles, even though they have very low surface tension. That’s another story. In water with organics, the tendency for the organics to want to come out at the air/water interface opposes the surface tension, and bubbles become a balancing act between surface tension that wants to pop or otherwise decrease bubbles, and the spreading pressure of organics that want to spread out across the air/water interface. So back to the oil drops. Once an oil droplet spans the water gap between bubbles, its all over. What happens is that the amphipathic molecules on both of the bubble surfaces spread along the interface between the oil and the water (if they were not there already) and connect both of the air gaps with a continuous line of amphipathic molecules along this oil/water interface. Once these amphipathic molecules are in place, the system is unstable. The surface tension pulls at the oil drop, and it simply comes apart. The bubble ruptures from the site of the oil drop, and the effect is that the bubble pops. The reason that this does not happen in the absence of an oil drop is that to cause a rupture requires the water present between the air bubbles to become exposed as fresh air/water interface. In fact, it requires a continuous line of water molecules to become exposed all at once. Because that would require a large number of hydrogen bonds to be broken simultaneously, it simply requires too much energy to actually take place. When the oil drop is there, one is no longer exposing water molecules, but rather oil or amphipathic molecules, which are much “happier” to be exposed to air. In any case, if the oil drop explanation makes sense to you, great. If not, don’t worry, as it is a very subtle and complicated concept.

Bubble Popping in Marine Tanks

In a real marine tank, many things have this bubble popping effect. One that most aquarists encounter is oil from their hands. On reaching into a tank, skimming action often comes to a near halt as bubble popping dominates over foam draining and collection. The popping will proceed until the oil is somehow removed. Among other ways, oil can be removed by splattering it above the foam height in the skimmer, being foamed out bit by bit, being emulsified into the general foam as very, very tiny droplets which no longer span air bubbles, becoming attached to solid objects and removed, being consumed by tank microorganisms, and by eventually dissolving into the bulk tank water. As an aside, the bubble popping action of hydrophobic oils is exactly how most antigas medications for humans function. Simethicone is really polydimethylsiloxane, which is a hydrophobic polymer liquid. It pops bubbles in your stomach or intestine, and permits the gas to be eliminated. Antifoaming agents are also the basis of a large number of industrial products that work by the same principle. Other things also cause bubble popping. One of these is the fatty acid supplement Selcon. It causes bubble popping in the same fashion as skin oil droplets. Hydrophobic solid objects can also cause popping. Sand coated with organics, inorganic precipitates from salt mixes covered with organics, food particles, etc. all function in a similar fashion. They cause bubble popping just like hydrophobic oils, except that they are solid.

Collection of Drained Foam

After a foam has drained to an extent desired, it must be collected and removed from the system. Most skimmers perform this simply by permitting the foam to be created at a rate that pushes the drained foam over a certain threshold, where it is irreversible collected and discarded. This process is relatively straightforward, and is mostly an engineering issue, as opposed to a chemical issue. The tricky thing is to balance foam creation, draining, and collection. If any of these are out of whack, the skimmer will suffer in efficiency.

What About Ozone?

For years, many people have suggested that ozone improves skimming efficiency. But does it really? Whether one uses a skimmer or not, the introduction of ozone into water that contains organics will have a big impact on those organic molecules. Nearly any organic molecule can be oxidized in the presence of ozone. One common effect of oxidation is the elimination of the types of molecules that lead to the adsorption of visible light. Compounds containing conjugated double bonds are one of the most common natural chromophores. These molecules are, however, very susceptible to oxidation to uncolored species.
It is thus highly probable than any elimination of yellow coloration in a marine tank on using ozone is due to the simple conversion of the organics from a light absorbing form, to a nonabsorbing form. The organics are not removed by a skimmer, and are not completely destroyed. They are just in a different form. Whether that is desirable or undesirable will depend entirely on the specific organic compounds in question.

The follow-up question is whether these oxidized compounds are more susceptible to being removed by a skimmer than before oxidation. At present, I am not aware of any study which shows that they are, or even of any physical reason why the would be. In general, oxidation makes organic molecules more hydrophilic. In some cases, it also breaks molecules into smaller pieces. Neither of these actions should lead to greater skimming. Hydrophobic molecules (in the presence of amphipathic molecules) are easily skimmed. Converting them to amphipathic molecules through introduction of a hydrophilic group will mean that they are still skimmed, but not that they are easier to skim. Oxidizing amphipathic molecules is also unlikely to increase skimming, and if they are oxidized so much that they become completely hydrophilic, then they will not be skimmed at all. I cannot think of a single molecule which becomes easier to skim by oxidation. On balance, there does not appear to be any evidence that the use of ozone increases skimming efficiency per se. That is not to say that ozone has no effect. The use of ozone can certainly lead to fewer yellowing compounds in the water, and might make many organics more susceptible to biodegradation. It can also sterilize water if used in sufficient concentrations. Are these things desirable? That’s up to each aquarist to decide.


The advent of skimmers has gone a long way to reduce the quantity of dissolved organics in marine aquaria. They have also been suggested to have a number of other effects, from aeration to elimination of certain inorganic species. Hopefully, this article will help hobbyists understand how skimming works, and then be able to use that information to critically evaluate claims about what skimmers can and cannot do.