1. Introduction
  2. Properties of Colloids
  3. Colloids and Light
  4. Aerosols
  5. Emulsions
  6. Foams
  7. Gels
  8. Colloids and Membranes
  9. Colloids and Adsorbents
  10. References


There should be a more exciting name for the phenomena to be discussed here. The dull term colloid that reminds us of glue is, nevertheless, the accepted word. It was coined in the early 19th century by the Father of Physical Chemistry, Thomas Graham (1805-1869), to distinguish those materials in aqueous solution that would not pass through a parchment membrane from those that would. Glue was indeed a material that would not, and the Greek for glue is kolla, from which we also get "protocol" and "collagen." Those that would pass through were things like salt, and other soluble crystalline substances, which Graham called crystalloids. As we shall see, the field is much, much richer than this.

Colloids received little attention until the end of the century, when van't Hoff, Oswald and Nernst founded modern physical chemistry and they, and others, became fascinated by colloid phenomena. There had been famous observations by Tyndall and others in the meantime, but chemists could not get excited over glue. Then, in the 1920's and 1930's, the importance of colloids to industrial processes and biochemistry changed everything. Colloids became a hot field, and soon every elementary textbook said something about them. In writing some of the other articles on the site recently, I became gradually aware of the fascination of colloids, and recognized that my knowledge of them was very deficient. This article is the result, and I hope it will serve as an introduction to what colloids are all about, and demonstrate how interesting and useful they are.

An interesting philosophical point was suggested by this study. Colloids are often called a "fourth state of matter," and I wondered just how meaningful this concept is. We shall find that it is very difficult to encapsulate any concept concerning colloids in a word, though heaven knows chemists have tried, and many words have been coined. It is necessary to name things to think about them efficiently, and one thing scientists have done assiduously is to assign names. Biology comes to mind, with endless terms and names based only on surface appearances, at least until recently. Naming gives the appearance of knowledge, where there is no real knowledge at all. The antithesis to mere naming is mathematical analysis, which gives real conclusions and effective knowledge. The danger of names comes when they are regarded as real things and are used to delimit instead of simply to denote and describe.

It is easy to recognize the three conventional states of matter in ice, water and steam. The names solid, liquid and gas can be attached to certain suites of properties, and makes a useful distinction. In a gas, particles of the substance move freely and have to be stopped by walls. In a liquid, the particles are sometimes associated, sometimes not, but always occupy a certain volume. In a solid, the particles cannot move far relatively, and can only vibrate. Many substances can be classified by these properties, but the terms do not separate matter into three mutually exclusive boxes, and may not be descriptive enough. Where is tar, for example, or jelly, or a substance above its critical point? Colloids will give many examples of substances for which the simple classification into three states is wholly inadequate.

The properties of matter depend almost completely on its structure. All metals are alike, to a good approximation. They are shiny, soft, tough substances that conduct electricity. All ionic crystals are alike (granted differences in crystal symmetry). They are hard, transparent and do not conduct electricity in the solid state. They can usually be crushed into white powders. The variations between metals, or between ionic crystals, are very much less important than their similarities. Saying that a substance is a metal, or an ionic crystal, says much more than simply that it is a solid. Solidity is only a macroscopic appearance, of no fundamental significance, like being green.

I have seen the definition of matter as "that which occupies space." But what about gases? They occupy space, of course, but two or more gases can occupy the same space, as far as appearances go. The important thing is to use terms like solid, liquid, gas only as far as they are useful descriptions, and not consider them as exclusive classifications into which everything must fit. To see that this is not trivial, consider the many sciences (not generally chemistry or physics) in which there are bitter controversies about which named category to assign to some object or process. We should not be limited by the arbitrary names we give our concepts.

Properties of Colloids

I recalled that colloids were particles larger than molecules, but smaller than grains of sand. This is true, and colloidal dimensions can be considered to be from about 10 nm up to 1000 nm, or 1 μm, but mere size is not the important thing about colloids. The overwhelmingly important property of colloids is that they have very large surface area. To some degree, they are all surface and their properties are those of their surfaces. I do not remember appreciating this properly before, but I can assure you of its significance. Incidentally, 1 μm is about the size of a bacterium. I shall use the word "colloid" to refer to a substance of colloidal dimensions, or to a colloidal system, indifferently.

To see the significance of this observation, consider the cubic centimeter in the diagram at the right. In this form, it has an area of 0.0006 m2. We could say that it is almost all volume. Most of its molecules are safely resident behind its surface, secure from disturbance or attack. Let us now divide it into thin laminae, 10 nm thick, a colloidal distance. The cube becomes a million laminae, with a total surface area of 200 m2. Every molecule is now only a short distance from the cold outdoors, and the material is all surface. We have turned the mass cube into a laminated colloid by this delicate slicing alone.

Continuing, we now slice each of the million laminae into a million fibers, and the surface area doubles. We still have a colloid, of course, with two dimensions colloidal, but have not increased the area greatly, not as we did in the first slicing. We can expect fibrillar colloids as well as laminar ones. Finally, each fiber is chopped into a million bits, giving a corpuscular colloid. This increases the surface area only by 50%, to 600 m2. From the mass to the corpuscle, the surface area has been increased by a factor of a million, which is typical of a colloid. Note that most of this increase came with the first dimension to "go colloidal," so we can call anything with any least dimension of colloidal size to be a colloid. This was another thing that I did not appreciate in my ignorance.

The large area emphasizes surface effects relative to volume effects, giving colloids different properties than those of bulk matter. The surface tension of a liquid is the free energy required per unit area to create new surface. For water against air, its value at 20°C is 72.75 dyne/cm. For mercury against air, it is 450 dyne/cm. For mercury against water, it is 375 dyne/cm, not surprisingly close to the difference of the other two values. The ratio of surface to volume for a sphere of diameter d, or a cube of side d, is 6/d. The surface energy of small drops will strongly affect their properties. If a cubic centimeter of water were divided into 10 nm cubes as above, the coalescence of the cubes would release enough surface energy to heat the water by about 10°C. A cubic centimeter of mercury would be warmed by 143°C by the same procedure.

It is better to define a colloid as a system in which the surface area is large and in which surface effects are predominant, rather than simply in terms of particle size. Indeed, in foams there are no colloidal particles at all--it is the thinness of the films that creates the colloidal behavior. Similarly, in a gel the fibrous structure is what is colloidal. In any colloidal system, there must be at least one structural dimension of colloidal size in order for the large surface area to exist in a limited volume, however. This broadened definition of colloid is not only reasonable, it is useful. A colloid is a material system that is mainly surface.

The next important characteristic is that a colloid is a two-phase (at least) system. A phase is a homogeneous component of a system, in the sense of the Phase Rule. The Gibbs Phase Rule applies to systems in which the phases have negligible surface energy, which is perfectly applicable to phases that are "all volume" as our centimeter cube was, or even to phases of microscopic dimensions (larger than 1μm). It does not apply to the system of colloid phases, in which surface energy predominates. Therefore, be careful when applying the Phase Rule to colloidal systems.

A colloidal system consists of an internal phase, which is the material of colloidal dimensions, and an external phase, which is the material in which the colloid is dispersed. These designations are analagous to the terms solute and solvent used for simple solutions (which form a single Gibbs phase). As the particles of a corpuscular colloid become smaller and smaller, we go over imperceptibly from a two-phase colloid to a single-phase solution, and there is no definite boundary. This gives a hint as to why I discussed names and their significance in the introductory paragraph.

The colloidal system that is most similar to a simple solution is a dispersion of corpuscles, or particles, in a liquid. This is called a sol, and the liquid is the external phase. This is the classical colloid as described by Graham. If the external phase is a solid instead of a liquid, the system is called a solid sol. The only difference is the mobility of the molecules. In a solid sol, they can move only by diffusion. If the external phase is a gas, usually air, instead of a liquid, we have an aerosol. There is no definite boundary between a sol and a solution, but still they are significantly different. There is also no definite boundary between a sol and a coarse suspension. A coarse suspension will settle out rapidly, while a sol may be permanent.

The particles that appear in a sol may be wetted by the liquid, or may not. Wetting is a typical surface effect, and so is of paramount importance in a colloid system. In the first case, the liquid is adsorbed on the surface of the particle. The terms adsorb and absorb sound alike, but are quite different. A substance that is absorbed is taken into the volume of the absorbing substance, like water into sand. If is is adsorbed, it attaches itself only to the surface. Since colloids are all surface, as we have pointed out, adsorption is what is important with them. If the particle adsorbs the external phase, it is called lyophilic, or hydrophilic, if the external phase is water. The Greek verb "luo" means to dissolve or destroy, and philic is from "philos," love. A lyophilic colloid "loves the external phase." On the other hand, if the particle does not adsorb the external phase, is is said to be lyophobic, or "fears the external phase."

Sols that contain inorganic particles, such as metals, are mostly lyophobic, as are most aerosols and solid sols. Lyophobic hydrosols are a very common kind of colloid, and deserve detailed description. For example, consider the hydrosol of gold with particles about 4 nm in size. This was one of the first sols studied extensively, and has interesting properties. With about 0.1% gold, the sol is a rich ruby red. The similar solid sol in glass makes ruby glass. The gold particles absorb strongly in the green and blue, so the transmitted light is red. There is a little yellow-green scattered light, but mostly it is a case of absorption by the gold metal. If the gold particles clump together, which they may do as time passes, the color of the solution changes. When the particles are about 40 nm in diameter, the solution is blue, with considerable scattered light. If the particles agglomerate further, the color disappears and gold flakes settle out.

Bacteria, which are about 1 μm in diameter, can be suspended in water to form a sol, which has all the classic properties. The Brownian motion, the Tyndall effect (turbidity), and even electrophoresis are seen. The bacteria act as a hydrophobic sol, peptized by their electrostatic charge. The properties of a sol are largely independent of the nature of the internal phase.

There are two interesting questions here about the stability of the sol. First, what keeps the gold suspended in the red solution so that the tiny particles do not settle out? Second, how are the particles kept from agglomerating? Let's take the first question first. The gold particles fall under gravity through the water. The terminal velocity v of their fall is given by Stokes's equation, (mg - m'g) = 6πηav, where m is the mass of the particle, a its radius, m' the mass of the water displaced, and η is the viscosity of water (1.002 centipoise at 20°C). A correction factor (1 + Kλ/a) must be applied for the small particles, where K is a constant, λ is the mean free path of the liquid molecules (this factor really applies better to aerosols), and a is the radius of the particle. On the other hand, the particles are subject to the bombardment of the molecules of the liquid, which produces the Brownian movement. For a sufficiently small particle, the upward diffusion produced by the Brownian movement overcomes the gravitational fall, and an equilibrium is reached, much like the equilibrium of gases in the atmosphere. Perrin first verified this effect, although it is quite complicated in this case, the gas-like effect only occurring close to the upper surface of the sol. There is a critical size for a particle, below which it will not settle out. At any rate, this is why colloidal-size particles do not settle out of a sol.

Now for the second question. Generally, if two colloidal particles collide, they will stick together and make a bigger particle, because it is usually favored by energy. Eventually, the particles get larger than the critical size to be suspended by the Brownian movement, and they settle out. There must be some good reason if this is not to happen. In most lyophobic colloids, the particles are electrically charged with the same sign, and this keeps them apart, since they repel one another. The particles are charged mainly because they adsorb certain ions in the environment. In water, they may be OH- ions, which are generally present, and give the particles a negative charge. The H+ ions are hydrated, so are not as easy to adsorb, but apparently some particles like them, and become positively charged. If you try to make a hydrosol with particles of opposite charges, they neutralize each other and the sol collapses. Since lyophobic sols are stabilized by electric charge, adding electrolytes generally destroys the sol. When rivers reach the sea with their loads of colloidal sediment, the ions in sea water coagulate the sol, and the load is deposited in the delta. There are other ways for a particle to become charged. A zinc particle, for example, may become negative when some zinc dissolves according to Zn → Zn++ = 2e-, and the electrons remain on the particle.

How the charges are distributed is an interesting question. The sol appears electrically neutral on the large scale. The particle with its adsorbed charges is called a granule. The charges are in a thin layer on the surface. They attract an atmosphere of opposite charges from the external phase, just as an electron in a plasma surrounds itself with a shielding positive charge by attracting positive ions and repelling electrons. The whole neutral structure, granule plus mobile external charge, is called a micelle. This, then, is what moves around, the charged particle and its cloud of opposite charge. When we apply an electric field to the sol, the cloud of charge is moved in one direction, and the granule moves in the other. There is a local viscous flow about the granule, and the micelle moves toward the anode, if the granule is positive, or toward the cathode, if it is negative. This movement is called electrophoresis, and can be practically useful. The particles of a sol do not repel one another until they come quite close, and their micelles overlap, because of the shielding.

The mobility of the colloidal particle is its velocity in a unit field. Helmholtz developed a formula for the mobility, M = ζκ/4πη. Here, κ is the dielectric constant of the external phase, η its viscosity, and ζ is the potential difference across the micelle from the outside to the adsorbed charges on the particle, that is, through the fluid that is sliding around the granule. From measurements of M, it is possible to find ζ, which becomes a kind of "fudge factor" since it is hard to calculate. For the small gold particles in the red sol, M = 4 x 10-4 cm2/s-V, and ζ = -0.058 V. The negative sign indicates that the gold migrates to the anode, and so has a negative charge. The Helmholtz equation gives M = 4.15 x 10-4, which doesn't prove much except that we know how to use the equation. Since the equation is written in esu, we must divide by (300)2 to convert volt to statvolt.

In order to create a lyophobic sol, we must either reduce a mass to colloidal size, called dispersion, or we must build the colloidal particles from molecules, called condensation. In either case, a third substance, a peptizing agent, may have to be added to stabilize the sol. This agent can supply ions that will be adsorbed on the particles resulting from dispersion or condensation to give them a stabilizing charge. For clays, the OH- ion is a peptizing agent, which can be supplied by alkalis. Dispersion can be done mechanically, in a colloid mill that grinds the substance into small, equal particles. Another method is with an electric arc. Metal electrodes are used, at a current of 5-10A and voltage of 30-40V. Bredig made particles of about 40 nm by this method, and it was improved by Svedburg to obtain sols of many metals down to 5 nm particle size. Ultrasonics can also be used to disperse sols.

When a sol is created in a nonpolar solvent, the particles may not be charged (and must be stabilized by some other means). They do not then exhibit electrophoresis or other electrical phenomena depending on granule charges.

Colloids and Light

A homogeneous phase or solution does not disturb the propagation of light, except to change its phase velocity to v = c/n, where c is the speed of light in vacuum and n is the index of refraction. In a gas, density fluctuations that are a natural result of the free movement of the molecules can scatter light. Scattering is the emission of light in all directions, which decreases the intensity of the ordered beam. This Rayleigh scattering is proportional to the inverse fourth power of the wavelength, so blue light is scattered more than red, giving the blue color of the clear sky. Scattering should be distinguished from absorption, which is the conversion of the energy of the light to other forms. Both scattering and absorption cause attenuation of the light beam. The transmitted beam and the scattered light may be colored if the scattering or absorption is not constant with wavelength. The blue sky and the orange sunset colors have the same cause, Rayleigh scattering by density fluctuations.

A colloidal system contains particles that affect a light beam by scattering and absorption. If the particles are of a size comparable to the wavelength of light or larger, they scatter or absorb light independently. The same thing happens if they are separated by distances comparable to or greater than the wavelength of light. The wavelength of visible light is 400-700 nm, with the maximum sensitivity at 555 nm. This is in the middle of the range of colloidal dimensions, so colloids can be expected to have significant effects on a light beam.

One common effect of colloids is turbidity, an effect like that of stirring up mud in water. Slight turbidity may not be noticed until a beam of light passes through the colloid. Colloidal systems need not be turbid: a gel may be quite transparent when the particles are small. Solutions, as homogeneous phases, are not turbid. The turbidity causes scattering so that the path of the light beam can be clearly seen. This is called the Tyndall effect, and the observed scattered light is called the Tyndall cone. John Tyndall (1820-1893) was Faraday's successor at the Royal Institution. He suggested that the blue sky was caused by scattering by dust particles, but Rayleigh later found the true cause, showing that the sky would be blue even if the air were pure. The scattered light is polarized perpendicularly to the direction of the beam if the particles are approximately spherical and small. If the particles are nonspherical, or larger than a wavelength, the scattered light will be partially polarized.

The Tyndall effect is a common atmospheric phenomenon. Searchlights produce Tyndall cones in slightly hazy air. Smoke is often blue when seen in scattered light, orange in transmitted light. This was quite clear in the summer of 2002, when forest fires near Denver put smoke in the air, and the sunlight became a strange reddish color. The crepuscular rays seen at sunset, which seem to radiate from the position of the sun, are parallel Tyndall cones. A laser beam may make a distinct Tyndall cone in dusty air. If you have a piece of polarizing filter, try to determine the polarization of any Tyndall cones you may observe. All the Tyndall cones that you see are evidence of lyophobic sols.

If the sol is composed of transparent particles with an index of refraction considerably different from that of the external phase, and they are present in sufficient concentration, the sol wil be come opaque and white, the limit of turbidity. If the particles have the same index of refraction, the Tyndall effect will be small. This property is used in enamels, which are opaque glasses fused onto a metal substrate, and in pottery glazes. Stannic oxide, for example, gives a white enamel or glaze.

A beam of bright sunlight entering through a window may be marked with moving bright specks that are light scattered from colloidal dust grains that are always present in the air. The grains themselves, which may be submicroscopic, are not seen, only the light they scatter. This principle is used in the ultramicroscope that allows individual, submicroscopic particles to be observed. Only the light from them is detected against a dark background, so that they can be counted and their motion observed; no image of the particles can be formed. The instrument consists of a bright light source, a slit and optics to focus the slit on the sample, and a microscope. A thin slice of the sample is illuminated by the slit. By turning the slit 90° the depth of the area viewed can be determined. The direction of viewing is at right angles to the illumination. Some ultramicroscopes are coaxial and use a different illumination method. If the number of particles per unit volume is found in this way, by counting using a squared graticule, and the weight of colloid per unit volume is known, then the size of the particles can be determined.

Colloids can produce color. The red of a gold sol or ruby glass has already been mentioned, where the color is due to the wavelength dependence of scattering and absorption. Color can also be produced by the interference of white light, especially if there are thin films or a periodic regularity in the density. The colors in thin oil films are familiar, and the films are, of course, colloidal in thickness. Color produced by such means is called structural. Color produced by absorption by colored pigments is called pigmental. The blue color seen in blue eyes and birds' feathers is structural, caused by scattering by fibers. Brown pigments in the iris modify the blue color to green, then overwhelm it to make brown eyes. Much color in the insect world is structural, such as the colors of a butterfly's wings, often combined with pigmental color to produce a great variety. The greenish colors of crude oil and its products are a result of colloidal suspensions.


An aerosol of colloidal solid particles may be called a smoke, while if the particles are liquid, it is a fog. Sometimes the two are combined, in a suspension of solid particles with an adsorbed liquid film on the surface. The original smog was a smoke with a liquid film of sulphuric acid, which made it excessively unpleasant to breathe. The Great Smog occurred in London on 5 December 1952, killing nearly 4000 people who had respiratory problems, and stimulating clean air legislation. Now it usually lacks the smoke, and is a fog of some unpleasant liquid coming from motorcars instead. Aerosols have the usual characteristics of lyophobic sols: a strong Brownian movement, the scattered blue light of the Tyndall effect, and stabilization with particle charges.

Clouds are aerosols of water particles, supported by the Brownian motion like any sol. The droplets are produced by condensation of water vapor on condensation nuclei, which are usually hygroscopic particles of dust, or positive ions. The mist above splashing water is positively charged, with compensating negative charge in the form of unhydrated negative ions. For condensation to occur, the air must be supersaturated for water vapor. The radius of the droplets is large for colloidal particles, and they are often supported by updrafts more than by Brownian motion. The radius of cirrus cloud droplets may be 2 μm. When they reach a radius of about 0.04 mm, the droplets fall as rain, often coalescing with others. The largest raindrops have a radius of about 3.6 mm, and such large drops are rare.

Raindrops can be blown upwards into freezing air by updrafts, gathering water by coalescence with others, and freezing to a solid ball. This can happen repeatedly, forming large hailstones that eventually fall. The largest hail reported may be the 3" diameter hail that fell in Bloomington, Ind. in 1917. A farmer was killed by hail near Lubbock, Texas in the 1930's, but this is the only reported fatality.

Saturated air can be expanded adiabatically to cool it and achieve supersaturation. This can be done in the laboratory with the Wilson cloud chamber with an adiabatic expansion against a piston or the equivalent, which does work that cools the air. The cooling is given by T2/T1 = (V1/V2)k - 1, where k is the ratio of the specific heat at constant pressure to the specific heat at constant volume (1.4 for air). In clean air, condensation occurs for a volume ratio greater than 1.25. From 1.25 to 1.34, the condensation falls as fine rain. For 1.25 to 1.28 the condensation is on negative ions, and the particle radius is 200μm. For 1.28 to 1.34 the condensation is on the positive ions, and the particle radius is 20μm. For larger expansion ratios, the mists can be colored: 1.408 gives green, 1.422 purple, and 1.429 red.

At the lower expansion ratios, condensation is difficult enough that it may occur along the tracks of alpha particles, which ionize strongly, making the tracks visible. The condensation soon settles out, since the droplets are large. The chamber can work with alcohol vapor as well as water vapor.

The vapor pressure p' of a small drop of radius r is greater than the vapor pressure p of a plane surface of the liquid. The vapor pressure p' is given by ln (p'/p) = (1/RTd)(2γ/r - q2/8πr4), if q is the charge density on the surface. γ is the surface tension of the liquid. This shows that a charge of either sign will stabilize a small droplet against evaporation. In the absence of a charge, small droplets will evaporate in favor of large drops. Electric charges may explain the stability of clouds, and the fact that rain may not fall from them.

An aerosol of starch grains, with a density of 100g/m3, has a very large surface area, and adsorbs O2 from the atmosphere. Starch grains in flour are from 5 nm to 200 nm in diameter. It is no wonder that it is a violet explosive. Coal ground to pass through a 200-300 mesh sieve can be blown directly into a fire as a convenient fuel. It can also make a sol with fuel oil to form a colloidal fuel that can be used exactly like fuel oil. Coal dust is reputed to be a fire hazard, but it is not as dangerous as starch dust. Many "coal dust" explosions may have another cause.

Smudge pots are used to protect agricultural plants from frost. They are small fires producing dense smoke. The heat produced by the fires is probably one of the most important effects. The smoke may provide condensation nuclei for the dews of evening, increasing the heat by the latent heat of the condensed vapor. The aerosol blanket slows loss of heat by radiation, which can be very important into a clear night sky.

Smoke is also used for military purposes, for concealment or signalling. White phosphorus burns to hygroscopic P2O5, which forms a dense white fog in humid air. Silicon tetrachloride mixed with ammonium hydroxide gives a dense white smoke of ammonium chloride and metasilicic acid. This smoke is used in skywriting, since it is easily made from liquids when needed. The navy's smoke for concealment was made by restricting the air supply to the boilers, producing thick black smoke from the funnels. Radar and the end of gunnery has rendered smoke screens useless. Colored smoke may be used for signalling.


An emulsion is a colloidal system in which both phases are liquid. If the liquids were miscible, they would form a solution, so emulsions are lyophobic colloids. The typical example is water and oil. The internal phase is determined by which component has the higher surface tension. This component will form spherical bubbles immersed in the other, which will be a continuous phase. The granules of an emulsion may be large, even microscopic. An emulsifying agent is usually required to form a stable emulsion. The emulsifying agent, or protective colloid, is surface-active, meaning that it reduces the surface tension of the liquid, and so tends to concentrate in boundary films. In the case of water and oil, sodium oleate, a soap reduces the water surface tension and raises that of the oil, so that the emulsion will be oil droplets in water, and quite stable. There are many other detergents, but sodium oleate will serve as a good example.

Oleic acid is a fatty acid with the formula CH3(CH2)7CH=CH(CH2)7COOH, a monounsaturated carboxylic acid with a long hydrocarbon chain. Fats are glyceryl esters of this acid. Glycerol has three OH groups, each of which can take the H off the end of the oleic acid and stick the rest to the glycerol framework, making a triglyceride. Boiling the fat with NaOH makes three molecules of sodium oleate, a liquid soap. The hydrocarbon chain nestles up to the oil, the sodium to the water, and peptizes the oil. The emulsion can be made by simply shaking oil, water and soap together, but this will not make droplets of a uniform size. The larger droplets may "cream off" in this case, and float to the surface of the emulsion. The emulsion may be homogenized by blowing it through small orifices under a pressure of 4000 to 5000 psi. The small, uniform drops that result will make a stable, long-lasting emulsion.

Milk is an emulsion of oil ("butterfat") in a watery sol of the hydrophilic protein casein in which the external phase is a solution of lactose and various salts. Milk will "cream" unless homogenized, since the fat globules range from 100 nm to 22 μm in diameter. Cream can contain from 29% to 56% fat, in packed globules stabilized by casein. Human milk contains albumin in addition to casein. Mayonnaise is an emulsion of oil in water, using egg yolks as an emulsifying agent. Hollandaise sauce is another emulsion, of butter in lemon juice, again using egg yolks as an emulsifying agent. Butter itself is an emulsion, this time of water droplets in oil. The cosmetic "vanishing cream" is an oil-in-water emulsion. If this is reversed, "nourishing cream" is a water-in-oil emulsion. Crude oil as it comes from the well is often emulsified with water. In this case, heating is sufficient to "break" the emulsion, and separate the oil from the salt water. Many insecticides are oil-in-water emulsions for spraying. The oil wets the oily leaf surfaces and sticks, while the water carrying the poison evaporates. Lubricating grease is a water-in-oil emulsion. The emulsifying agent, calcium oleate, is soluble in oil, not in water, and so makes the water the internal phase. The idea here is to make the lubricant stiff, so it will not drip off.

Whether a component of an emulsion is the internal or external phase is determined by the relative surface tensions, not be the amounts of the components. Equal-sized globules can close-pack like spheres to occupy 74% of the volume. The internal phase can be even higher in concentration if the globules are not all the same size, so the smaller can huddle in the voids left by the larger. Because of the presence of an emulsifying agent, emulsions are very stable, and "breaking" them when required can be difficult.

To find out whether an emulsion is oil-in-water or water-in-oil, the effect of adding a small amount of either oil or water to a sample of the emulsion on a microscope slide is observed. If you add the external phase, it will mix easily and quickly, but the internal phase will not mix and remain a drop. This is called the dilution test.


A foam may be an internal phase of gas in an external phase of liquid or solid. In a liquid foam, a colloidal adsorptive agent forms a film that bounds the gas bubble. Bubbles blown with soap solution are related to foams, but are quite large and have an independent existence. Smaller bubbles in a mass form the more usual foam. People washing dishes or clothes are reassured by a thick layer of soap foam on the water, showing that there is still detergent left to emulsify additional oil. Such foams are mainly air, with very little liquid. The colloidal dimension in a foam is the thickness of the film, not the size of the bubble. The bubble is lighter than its surroundings, and will rise to the top, where it joins the foam. In beer, the foam is stabilized by albumin and by the hop resins added to the beer. When carbon dioxide is released, it uses the adsorptive agent it finds to make the foam. This "head" on Guinness stout is famous and creamy--and carefully engineered. Meringue is a dried foam using egg albumin. Marshmallows use sugared gelatin for the same purpose.

Ore flotation depends on the property of the adsorptive agent to wet the valuable metallic sulphides or other ore, but not the silicates, which are preferentially wetted by water. Note again the central role played by surface chemistry. A froth is made with water and the adsorptive agent, and mixed with the ore to be beneficiated. The froth floats to the surface, where it is skimmed off, together with the enriched ore. In a few special cases, the valuable mineral is wetted by the water, while the gangue sticks to the oil. We use the buoyancy of the bubbles to effect the separation.

A fire-fighting foam is made from mixing water, aluminium sulphate and sodium bicarbonate with an adsorptive agent. The carbon dioxide that is released makes a dry foam, while the other ingredients form a kind of gel. This foam can be used on all kinds of fires, including burning oils.

Examples of solid foams are pumice, meerschaum, and Ivory® soap. The white soap has colloidal air beaten into it so that it will float. Meerschaum, "sea foam" in German, is a light-colored metamorphic rock associated with serpentine, and is a magnesium silicate, also known as sepiolite from its resemblance to light cuttlefish (sepia) bone. It is soft, smooth, light and translucent, used mainly for carving smoking pipes. It is fibrous and porous, with gaseous inclusions, apparently a dried gel. It is so porous that it floats on water, in spite of its mass density of 2.0 g/cc. Pumice, an extrusive igneous rock, is a solidifed foam of volcanic glass, usually obsidian. It also floats on water, and makes a good gentle abrasive. Diatomaceous earth consists of the microscopic shells of diatoms, very common marine plants, made of opaline silica and very porous. It adsorbs nitroglycerine to make dynamite, rendering it much less sensitive and much safer to handle. Bread is, of course, a dried foam as well. The protein gluten makes the film that surrounds the CO2 bubbles produced by the enzyme zymase secreted by the yeast. Zymase acts on the sugars (hexose) produced from starch by other enzymes, and also makes ethyl alcohol at the same time as the carbon dioxide. The alcohol is often the desired product! Hydrophilic gluten and water make a good gel in "strong" flours that include as much CO2 as possible and hold the starch granules. Rye flour has little gluten, and will not make light bread by itself. Foams aid digestion by providing as large a surface area as possible.


So far we have looked at lyophobic colloids, which will have nothing to do with the other phase. A lyophilic colloid, by contrast, actively seeks out the other phase and adsorbs it strongly. In most cases, the other phase is water, so we are dealing with hydrophilic colloids. Aluminum hydroxide and orthosilicic acid are inorganic examples. Most organic colloids are hydrophilic. Gelatin is of animal origin, derived from protein, made by boiling bones and horny parts. Gums are of vegetable origin. Gums are branching polysaccharides, soluble in water. They include the mucilages carrageenan (from Irish moss, a seaweed) and agar-agar (also from seaweed), which are sulphate esters. Gum arabic is from the acacia tree, and gum mastic is from the Pistacia lentiscus. Gums are secreted by trees to protect and seal wounds. When you find any of these in a list of food ingredients, they are for the purpose of making a gel. Ice cream may contain guar gum, cellulose gum, locust bean gum and carrageenan. These gums are not digestible.

A resin is similar to a gum in purpose, but is insoluble in water. Resins are notably secreted by evergreens, and are terpene derivatives, soluble in turpentine and similar solvents. Rosin is the residue when turpentine is distilled. Amber and copal are other natural resins, while alkyd and phenolic resins are artificial, used in plastics and paint.

Gelatin or gums dissolve in hot water. As the clear colloidal system cools, its viscosity increases steadily. The viscosity of the external phase is not affected in hydrophobic sols, by distinction. At some point, the system gels, forming a wobbly but definitely solid body. This is really an extraordinary thing to happen. If you heat the gel, it will melt and form a viscous liquid. On cooling, it will gel again. If you dry it out, it will shrink and look horrible. On adding water, it will plump up again into a wiggly gel. These colloids are called reversible or elastic. The gels formed from inorganic hydroxides will not reform a gel once they have dried out, and the dry form will be brittle. The pore spaces will still be there, however, and will absorb moisture and other substances. Silica gel is a widely used substance. Though it can be renewed by heating and live to absorb again, it will never again be wobbly and gelatinous. I say absorb, since it will appear to be this, but on a microscopic scale it is still adsorption, of course.

Gels are used as culture media for microorganisms. Gelatin was originally used, but it melts at 37°C, and so cannot be used to study microorganisms at body temperature. Agar-agar makes a gel that can stand higher temperatures, so it is used in preference as a culture medium, poured into the familiar Petri dish. It was first used by Robert Koch.

The colloidal phase in a gelatin is fibrillar, composed of fibers of colloidal cross-section. When a gel sets, these fibers form a tangled mass like a pile of brush, that holds the system together. Droplets in a gel are lens-shaped, showing the packing. There has been quite a bit of controversy over the structure of gels, but the fibrillar structure seems correct. The original idea of a cellular structure is untenable. The fibers adsorb large quantities of water, and there are also droplets of water, so that the gel is mainly water, given a doubtful rigidity by the stacks of fibers.

Pectin is a gelatin-like protein substance found in the rind of citrus fruit, in apples and generally in fruits. If a slightly acid solution of pectin is made 65% or 70% sucrose, it will gel. This is the reaction used to make jams. The verb pectize is used to describe the creation of a gel, as peptize is to create a sol.

Gels have some curious properties. As they age, syneresis may occur, which is the loss of liquid. This is a result of the closer agglomeration of the colloid. The reduced active area requires less water, so the excess water is eliminated. It does not mean that the gel is deteriorating. Another property is thixotropy. On agitation, the gel becomes fluid, but reverts to the gel when left alone. Fresh gelatin gels are quite thixotropic, and advantage can be taken of this to add ingredients to a jelly. The "thixo-" comes from the future of the Greek verb for "touch," thixomai, and a thixotropic substance would be one whose state changes with touching.

Colloids and Membranes

Let's review the phenomenon of osmosis. In the diagram, we have put some solution into a container closed off by a semipermeable membrane, and put the container into a vessel containing the pure solvent. Consider the dots as representing the dissolved substance. This applies to a sol, as well as to a solution, but there will be more "solute" particles in the solution. The membrane can be cellophane, but ordinary wrapping cellophane cannot be used, since it has been lacquered to close the pores. The pores must be small enough to prevent the solute from passing through. The solvent is found to cross the membrane and enter the solution, making it more dilute. The level of the solution rises to some height h in equilibrium. If d is the density, then π = dgh is the osmotic pressure. This is the pressure required on the solution side to make the rate of movement of the solvent the same in both directions across the membrane. If the solution is dilute, van't Hoff's equation gives the osmotic pressure: πV = nRT, where the volume V of the solution contains n moles of solute.

The osmotic pressure of a solution containing 1 mole of solute per 1 kg of solvent, called a 1 molal solution, is quite large, about 25 atmospheres (as we can estimate from van't Hoff's equation). Not only is this hard on cellophane membranes, but would require a liquid column 825 feet high, which is inconvenient. Nevertheless, osmotic pressures are high enough to push fluids to the tops of high trees (this is actually a complex subject about which there is controversy). Strong semipermeable membranes can be made by precipitating insoluble salts in the pores of unglazed porcelain. Sols will exert quite modest osmotic pressures, since there are many fewer particles per unit volume. In fact, the pressures are so low that they cannot be used in most cases to find the number of particles per unit volume. However, the molecular weight of haemoglobin was determined in this way. Cell walls are semipermeable membranes. If an erythrocyte (red blood cell) is put into pure water, it will swell and burst because it contains a solution of electrolytes in a semipermeable membrane. If put into strong salt solution, it will become dehydrated and shrivel. Solutions of the same osmotic pressure are called isotonic. Isotonic solutions are used to prevent damage to biological systems. Erythrocytes are about 8.6 μm in diameter and 2.6 μm thick, and slightly concave on the faces. They are a bit large for colloidal particles, though the blood plasma is a colloid that can gel under certain circumstances. Like platelets, which are definitely colloidal, they are not living cells like the leucocytes that accompany them.

The separation of colloids and crystalloids is called dialysis, and can be carried out with a suitable semipermeable membrane, just as in osmosis. Chemists actually distingush dialysis, in which the components diffuse at different rates across a membrane, from ultrafiltration, in which larger particles are mechanically stopped while smaller ones are allowed to pass. Ultrafiltration can occur through gels, while dialysis uses dialytic membranes. There is no fundamental difference between the processes, however.

Dialysis is carried out in the body by the kidneys, which separate the crystalloids urea, uric acid, hippuric acid and ammonia compounds from the colloidal albumin and other proteins of the blood. The crystalloids diffuse more rapidly across the interface than the larger particles. In the laboratory, we can put the sample to be dialyzed into a container like that used for osmosis, and change the solvent as it becomes concentrated in the crystalloids. Parchment paper can be used for the dialytic membrane. A bag of parchment paper can be filled with the sample to be dialyzed, and it can be suspended in a bath of moving warm water for rapid dialysis. Membranes can also be prepared from other animal membranes, or collodion (nitrocellulose dissolved in alcohol and ether), or artificial sheet polymers. Electrodialysis can also be used, taking advantage of ion migration in an electric field.

Dialysis has many industrial applications. Sugar is extracted from sugar beets by using the cell walls as dialytic membranes, washing cut beets in warm water. Dialysis is used in the artificial fiber industry to separate alkali from the colloidal fiber material, and in the pharmaceutical industry for the purification of colloidal medicines. Dialysis is used in artificial kidney machines to simulate the action of the kidneys.

Colloids and Adsorbents

As has been pointed out, absorption is a volume effect, while adsorption is a surface effect. Colloids, having large surface areas for a given volume, are excellent at adsorption. A cubic centimeter of charcoal can have a surface area of 1000 m2, so a little charcoal can do a lot of adsorbing. Adsorption may be specific. For example, Ni, Pt and Pd in colloidal form adsorb H2 up to 1000 or 3000 times their volume. Palladium is the best at this. At red heat, palladium metal will absorb hydrogen readily, and release it at even higher temperature. Black palladium powder can be made by reducing PdCl2 in solution. Because of the adsorption, these metals, in colloidal form, make excellent catalysts for hydrogenation.

Usually, adsorption is not very specific, and a wide variety of substances can be adsorbed on a certain medium. Adsorption is temperature-sensitive, being much more effective at low temperatures, and evolving the adsorbed substances at higher temperatures. Carbon at room temperature does not adsorb oxygen and nitrogen, but does so at liquid-nitrogen temperatures. Adsorption is generally accompanied by a negative change in enthalpy, so it is exothermic. The temperature dependence is a consequence of LeChatelier's Principle.

The three most commonly used adsorbents are carbon as charcoal, alumina and silica. When specially prepared as adsorbents, they are called activated. Activation is usually a matter mainly of heating, and perhaps some chemical cleaning. A saturated adsorbent may be re-activated by heating, say at 175°C for 6-8 hours in air. Charcoal may be prepared from wood, bone, blood and sugar. The charcoal from different sources has different impurities and therefore somewhat different characteristics. Charcoal is a nonpolar adsorbent that is good for organic vapors and nonpolar substances in general. Alumina and Silica are polar adsorbents and are best for polar substances, like water.

Charcoal is used in gas masks, since it readily adsorbs toxic organic vapors rather indiscriminately. It does not adsorb carbon monoxide or ammonia well, so specal adsorbents for these must be included. Carbon monoxide is usually oxidized to the dioxide, and ammonia by silica gel. Therefore, a complete gas mask canister contains a mechanical filter, charcoal, silica gel, and an oxidant for CO. Charcoal can be used to produce ultra-high vacuum by cooling it to liquid nitrogen temperatures. If charcoal in a stout glass tube in a U-shape is saturated with chlorine or sulphur dioxide, and then sealed off, the gas is liquefied when the charcoal end of the tube is heated, and the other end is cooled. Natural gas may be filtered through charcoal to remove higher alkanes. Pentane and hexane are more strongly adsorbed than the lighter methane and ethane. Charcoal is, in general, used when organic substances are to be removed from a gas.

Alumina gel and silica gel are produced by drying the gel produced by precipitating the hydroxides in water. Alumina gel is good for water vapor, carbon dioxide and alcohol. It makes an excellent laboratory dessicant. A small amount of colorimetric indicator is added that is blue when the alumina is dry, but turns pink when it has adsorbed water and is alkaline. The dessicant can then be regenerated by heating in an oven. Alumina gel is generally used for drying organic liquids, which it also decolorizes, and for removing acids from oils. It also removes oil vapor from compressed air. Alumina gel is a good example of the nonspecifity of adsorption.

Silica gel is the most generally useful adsorbent. It is especially good at adsorbing benzene, for example from coke-oven gas. Silica gel is used to dry carbon dioxide, hydrogen and oxygen before they are liquefied or solidified. It is often found in small bags in sealed packages that must be protected from humidity. These bags are easily re-activated by heating. Silica gel can be used to dry natural gas for pipeline transport, to prevent the formation of ice clathrates at low temperatures. These clathrates can form above the freezing point and easily clog pipelines.

Other colloidal adsorbents found naturally are diatomite, also known as diatomaceous earth or kieselguhr, which is composed of diatom shells made from opaline silica, and bentonite, a strongly hydrophilic colloidal clay consisting mainly of montmorillonite. Bentonite can clarify and deodorize petroleum, purify and soften water, make drilling mud, plug leaks as a grout, improve cleaning powders, and destroy building foundations by swelling. Diatomite has perhaps even more uses than bentonite. Both are valuable colloidal minerals.

Adsorbents are imporant in dyeing fabrics. Often the fabric will not adsorb the dye directly, since the dye may be polar and the fiber nonpolar. However, gels like amphoteric metal hydroxides, especially aluminium hydroxide, may cling to the fibers while strongly adsorbing the dyes. Such intermediates are called mordants, which are usually colloids. The dye and the mordant together, without the fiber, is called a lake (from the same word that gave "lacquer"). Purple of Cassius is a famous lake, formed on stannic hydroxide gel by colloidal gold.


Most textbooks of Elementary Chemistry will include a chapter on colloids that makes a good introduction to the subject.

R. J. Hartman, Colloid Chemistry (London: Pitman & Sons, 1949). A classic reference with a great deal of description of colloid phenomena of all types.

Return to Physics Index

Composed by J. B. Calvert
Created 5 December 2002
Last revised 9 December 2002