AS RECENTLY AS 1950 PERFUMES were graded by the noses of connoisseurs and wines were judged by the palates of wine tasters. Even chemists occasionally relied on taste and smell in some of their procedures. Today these arts have been outmoded by the gas chromatograph. This simple apparatus can quickly detect and measure each of the scores of fractions in a complex mixture, even though some components may amount to only a few trillionths of the total bulk. The gas chromatograph contains no complex parts, and it can be built easily and inexpensively by the amateur.

George V. Downing, Jr., a specialist in gas chromatography, described the operating principle of the device in the October 1964 issue of The Oregon Science Teacher, official publication of the Oregon Science Teachers' Association. "The principle of gas chromatography," he wrote, "is illustrated by a stream of inert gas such as nitrogen passing through a tube packed with granules of an inert substance coated with a thin film of liquid such as silicone oil that is essentially nonvolatile at the temperature of the tube. The nitrogen is called the carrier gas and the oily granules are known as the chromatographic column.

"Suppose we now inject at the entrance to the tube one milliliter of a mixture consisting of three gases such as propane, butane and argon. Propane ' and butane are soluble in the silicone oil, and hence some atoms of each will dissolve in the liquid. Propane is the more volatile of the two and for this reason it has a lesser tendency than the butane to dissolve in the oil. Some atoms of both gases, however, continuously dissolve in the oil and evaporate out of it only to be redissolved and reevaporated again and again during their transit through the oily granules.

"The propane will spend more time during the transit in the moving gas than the butane will. Conversely, the butane will spend more time in the stationary oil than the propane will. Hence the propane will outdistance the butane as the two substances are urged downstream by the carrier gas; the two will separate and the propane will arrive first at the end of the tube. Even so, the propane will have spent some time in the stationary oil. For this reason it will not keep pace with the nitrogen gas that urges it along. Argon is not soluble in silicone oil. This gas would move through the tube at the same velocity as nitrogen and would have arrived at the end before either the propane or the butane."

The chromatographic separation of the fractions and their serial discharge at the outlet of the column depend on the affinity between silicone oil and other substances. In the case of generally inert gases such as nitrogen, helium and argon the affinity is insignificant and chromatographic separation does not occur. Affinity is significant, however, in the case of chemically active gases. Moreover, it varies distinctively from substance to substance.

Materials other than silicone oil can be used for the chromatographic column. Detergents, for example, are particularly effective for separating petroleum products. Materials such as the polyglycols are used for the analysis of plant vapors.

Having made the separation, the gas chromatograph must also detect the fractions as they emerge serially from the outlet of the column. Although many detection schemes are possible, three have gained favor. One is based on the relative effectiveness of gases as conductors of heat. The other two are based on the extent to which gases can be made to conduct electricity.

The detector based on conduction of heat makes use of a heated wire immersed in the stream of gases that flows from the column. Each of the gases cools the wire at a characteristic rate. The electrical resistance of the wire varies with temperature and can be continuously measured by a sensitive Wheatstone bridge. Variations of the galvanometer in the bridge reflect changes in the composition of the gas fractions passing out of the chromatographic column. In advanced versions of the instrument variations in the resistance of the detector operate a pen and are recorded as a graph.

Detectors based on the electrical conductivity of gases take the form of an ionization chamber. This enclosure, through which the exhaust gases of the column flow, is fitted with a device that ionizes the gases and with a pair of electrodes that set up a potential field through the gases. A sensitive ammeter connected to the electrodes measures the amplitude of the resulting current.

In one form of the conductivity detector the gases entering the ionization chamber are exposed to a small piece of radioactive metallic foil. High-energy nuclear particles from the foil bombard and thus ionize the gas. In an alternate scheme the radioactive foil is replaced by a burner that is fed with hydrogen and the exhaust gases from the column. The hydrogen burns in the chamber, resulting in pyrolysis: the electrification of the gases by ionization arising from heat. Each gas becomes more or less electrified, depending on its nature. The electric currents developed in the gases are quite small, amounting to less than a millionth of an ampere. Even so, the most sensitive detectors are of the electrical-conductivity type.

Each of the detection schemes has certain advantages and disadvantages. The heated-wire system, known as the katharometer, is reliable and easy to construct but not suitable for use with all carrier gases. The katharometer works best with gases (such as helium) that are good conductors of heat. It can detect differences in gas fractions amounting to about one part in10,000.

This is adequate sensitivity for the analysis of mixtures in which each component is present in fairly substantial quantity.

Some investigations, however, require more sensitivity. Examples are the analysis of atmospheric pollutants and the separation of compounds that are unstable at the high temperature necessary to achieve a vapor pressure adequate for detection by the hot-wire method. Many biological materials such as alkaloids and steroids are in this category; the analysis of such substances requires sensitivities on the order of one part in 10 trillion. Sensitivities of this magnitude are achieved by the ionization detectors.

The sensitivity and efficiency of the gas chromatograph vary with the specimen, the composition of the column, the kind of carrier gas and its temperature, the rate at which the carrier gas flows through the column and the detection scheme. Many carrier gases have been tried. Nitrogen was used by A. T. James and A. J. P. Martin, the British chemists who first proposed gas chromatography [see "Gas Chromatography, by Roy A. Keller; SCIENTIFIC AMERICAN, October, 1961]. Helium has the highest thermal conductivity of any inert gas and is a popular carrier in the U.S., particularly for use with a katharometer. Although hydrogen has high thermal conductivity and is inexpensive, it is also reactive and potentially explosive. Air, the least expensive carrier, is usually avoided because it can oxidize the specimen. Argon is used with certain radioactive detectors because it acts as an integral part of the ionizing mechanism.

"The rate at which the carrier gas flows through the column is usually controlled by maintaining constant pressure with a regulator of the diaphragm type at the input. The flow can be retarded by placing finely drawn capillary tubes of glass between the regulator and the head of the column. Traces of water vapor and oil are normally removed from the carrier gas by a drying tube, a molecular sieve, sintered glass filters or silica gel.

The rate of flow that yields optimum separation of the fractions must be determined experimentally. If the rate is too high, the components do not have time to separate before they are carried out of the column. If the rate is too low, diffusion can mix the fractions partially and thus prevent sharp separations. In columns with a diameter of six millimeters practical rates of flow vary between 30 and 100 milliliters per minute.

According to Downing, a good column when operated properly will have a useful life of several months. Most columns consist of a copper, a glass or an aluminum tube that has an inside diameter of about six millimeters and is packed with a suitable material. The tube, some three meters long, is coiled or otherwise bent into a package of convenient size.

Metal tubing conducts heat better than glass tubing and therefore tends to maintain the column at a more uniform temperature. Metal columns are also more rugged than glass and easier to package for use in ovens. On the other hand, glass assemblies are chemically more inert. They are also transparent and so enable the experimenter to observe the packing procedure in detail and ensure uniformity.

The packing material should provide the maximum possible surface for exchanges between the liquid and the gases. For this reason it usually consists of an inert granular substance coated with a liquid. Widely used granules include diatomaceous earth, pulverized firebrick, glass beads, salt pellets and Teflon pellets. Some commercial detergents act as efficient column substances without a fluid coating.

Particles should be screened to remove dust that might otherwise plug the column. Particle sizes of 40 to 50 mesh (a measure of screen fineness per square inch) are frequently used. Column efficiency varies inversely with particle sizes; 80 to 100 mesh or even 100 to 140 mesh have been used. The finer the particles, however, the higher the pressure required to drive the gas through the column.

An introductory experiment in gas chromatography conducted last year by Steve Langhoff and Glen Martin, who were then seniors at the South Salem High School in Salem, Ore., illustrates the procedure. Langhoff and Martin's instrument is properly known as a "vapor phase" chromatograph because the specimen (white gasoline), which exists as a fluid at room temperature, must for analysis be converted to vapor by heating the chromatographic column. The envelope of this column consists of copper tubing eight feet long and a quarter of an inch in outside diameter.

A short T connection was first soldered into the discharge end of the tube. The space just beyond the T was packed with a plug of rock wool about half an inch long. The tubing was then suspended vertically, with the T at the bottom, and completely filled with the commercial detergent Tide. The granules were slowly poured into the tube through a small funnel and packed by holding a vibrator against the metal. When solidly packed, the tube was wrapped around a cylindrical form to make a coil six inches in diameter, a convenient size for use in an oven improvised of plywood, Transite and an electrical heating element of the radiant type [see Figure 1].

The carrier gas for the system was helium, which Langhoff and Martin obtained from a local welding shop. A regulator of the diaphragm type maintains the pressure of the carrier gas at nine centimeters of mercury, which is equivalent to 1.6 pounds per square inch. Carrier gas enters the column through the leg of a glass T. One end of the crossarm of the T is closed by a plastic cap of the kind used for sealing biological preparations such as serum in bottles. The other end of the T is attached to the column through a compression coupling as illustrated in Figure 2. The space in the copper tube between the granules of detergent and the compression coupling is plugged with a wad of rock wool. A second T fitting that is inserted in the hose connection between the pressure regulator and the column connects to the mercury manometer used for observing the input pressure of the carrier gas.

The T connection at the exhaust end of the column leads to an ingenious flowmeter suggested to Langhoff and Martin by Don F. Weinhart, professor of chemistry at the University of Oregon. The device consists of a glass tube calibrated in milliliters and fitted at the lower end by a side arm for admitting gas and by a rubber squeeze bulb partly filled with soap solution [see Figure 4]. Squeezing the bulb deposits a soap film across the calibrated tube. The soap film is then forced up the tube by gas entering the side arm from the chromatographic column. By timing the passage of the soap film along the graduations the rate of flow in milliliters per minute is easily calculated. The flowmeter was made of a broken buret by cutting off the jagged top, sealing in a side arm at the bottom and pulling the soft glass beyond the side arm into a taper that fits the squeeze bulb.

"The oven is simply a plywood box of convenient size lined with Transite and insulated with rock wool. The construction is not critical because the column operates at a temperature of only 130 degrees centigrade. (The temperature of the column need be only high enough to vaporize the specimen, in this case white gasoline.) Heat is supplied by a 600-watt unit from a radian electric heater. The temperature controlled by energizing the heated through a variable autotransformer.

The essential element of the detector is the filament assembly of a 6.3-volt 25-ampere incandescent lamp of the type used in a four-cell flashlight. Such bulbs are available in most hardware stores. The filament is removed from the lamp by filing a nick completely around the bulb at the point where the glass joins the metal base. Score the glass with a corner of a fine, sharp file. Use light pressure, not more than five ounces. Having extended the nick completely around the envelope, keep filing until the glass cracks. Do not grow impatient and strike the bulb if it fails to break when you think it should. Avoid striking and thus bending the filament assembly when the glass falls away.

Without removing the filament assembly from its base, extend the filament leads by soldering a pair of 24-gauge copper wires about six inches long to the leads just below the glass bead that supports the filament. Use a good liquid soldering flux. Clip off the portion of the filament leads just below the soldered joints and remove the filament from the base. A new base is made by spiraling asbestos tape between the leads and winding it into a cylinder as illustrated above. The assembly is then slipped into the discharge end of the column.

Measure carefully the distance between the end of the tube and the wad of rock wool that retains the packing. Slide the filament assembly into the tubing just far enough so that a space of about a quarter of an inch remains between the filament and the rock wool. Before installing the filament assembly try it for size in a short length of the copper tubing and simultaneously adjust the leads to prevent contact between the filament and the tubing. When all is ready, slide the assembly into the column, brace the rear of the asbestos base with a tuft of rock wool and secure the leads in place with a plug of putty.

The Wheatstone bridge can consist of composition resistors of half-watt size. Never apply voltage to the bridge until air in the chromatographic column has been displaced by carrier gas. The filament will burn out if it is heated to incandescence in air. Use a fully charged storage battery for energizing the bridge, because accuracy of analysis requires constant voltage across the filament.

To analyze a specimen of white gasoline turn on the heater and adjust the autotransformer until the temperature of the oven remains constant at 130 degrees C. for 30 minutes. When the temperature has stabilized, turn on the carrier gas and adjust the regulator to a pressure of nine centimeters as indicated by the manometer. After about a minute test the system to see if the carrier gas has displaced air from the column. One makes this test by holding a lighted match or a glowing splint immediately above the discharge end of the flowmeter. If the gas quenches the flame, the probability is that the system has filled with helium.

Close the switches that connect the two-ohm and 22-ohm resistors across the microammeter of the Wheatstone bridge. (Incidentally, the most sensitive microammeter available should be used-in the bridge, preferably one that requires not more than 50 microamperes for full-scale deflection.) Now clip the power leads to the battery and adjust the potentiometer in the arm of the bridge opposite the filament until the microammeter indicates minimum current. The resistance of the potentiometer at this setting should approximate that of the heated filament. The setting of this potentiometer should need no additional adjustment.

The fine adjustment is made by altering the setting of the potentiometer that is connected between the 22-ohm arms of the bridge. This is the lower potentiometer in the accompanying illustration in Figure 6. The switches that connect the two-ohm and 22-ohm protective resistors across the meter should be closed during the initial adjustments. As balance is approached the two-ohm resistor is cut out of the circuit. This usually increases the excursion of the meter. When the excursion has been minimized by further adjustment of the potentiometer, the 22-ohm resistor is cut out, after which the bridge is balanced as completely as possible.

The temperature of the oven, the pressure of the carrier gas and the rate of carrier-gas flow are then measured and recorded. Incidentally, by dividing the known volume of the column by the rate of flow of the carrier gas the interval required for displacing air with inert gas can be found. This is the interval between the time the carrier gas is turned on and the time voltage can be safely applied to the bridge.

Finally, about a twenty-fifth of a milliliter of clear gasoline is taken up in a small hypodermic syringe. Insert the needle through the serum cap and squirt the specimen on the detergent at the top of the column. According to Langhoff and Martin, the sharpness of separations appears to vary inversely with the size of the specimen. Readings are most conveniently made by two people, one observing the current amplitude and tabulating the results while the other calls the time. Most runs can be completed within 20 minutes. The constituents of mixtures appear as peaks when the results are plotted in the form of a graph.

A typical graph of clear gasoline made by Langhoff and Martin is reproduced above. Having made the analysis, the experimenters immediately injected a small specimen of pure octane in the column. Every effort was made to maintain all conditions of the apparatus constant, including the temperature of the column and the pressure of the carrier gas. The results of the analysis of the octane specimen appear as the broken curve below the seventh peak of the graph that represents white gasoline. Langhoff and Martin do not seriously doubt that the seventh peak of the graph was made by the octane fraction in their specimen of white gas, but they cannot explain the slight discrepancy in position between the two peaks. The pure octane appears to have made its way through the column some 20 seconds faster than the octane fraction of the gasoline. Langhoff and Martin plan to make another experiment in which the octane will be added to the gasoline before the run. The combination should result in a single, higher seventh peak. Incidentally, this experiment demonstrates how the gas chromatograph can be calibrated: graphs derived from known fractions are simply compared with the peaks made by unknown mixtures.

Bibliography

CHROMATOGRAPHY. Roy A. Keller, George H. Stewart and J. Calvin Giddings in Annual Review of Physical Chemistry: Volume Xl, edited by H. Eyring. Annual Reviews, Inc., 1960.

PRINCIPLES AND PRACTICE OF GAS CHROMATOGRAPHY. Robert L. Pecsok. John Wiley & Sons, Inc., 1959.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.