Reasonably stable free radicals were prepared by chemical methods at the turn of the century, but techniques for coping with the more interesting and reactive types are a recent development. In one ingenious procedure raw material in the form of gas at low pressure is piped through a high-frequency electromagnetic field that breaks up the molecules; the free radicals then migrate into an evacuated vessel and condense on a cold surface [see "Frozen Free Radicals," by Charles M. Herzfeld and Arnold M. Bass; SCIENTIFIC AMERICAN, March, 1957]. An apparatus based on this principle was constructed last year by Fred Swift of Maquoketa, Iowa, now a student at the State University Iowa. Instead of using a single condensing surface Swift equipped his system with a series of three traps: U-shaped glass tubes that are maintained at progressively lower temperatures. The apparatus was constructed primarily for decomposing carbon tetrachloride into free radicals and collecting the more less stable reaction products according to the temperature at which each condenses.

"Essentially," writes Swift, "the apparatus consists of five parts: a vessel from which carbon tetrachloride or some other compound is introduced into the system at low pressure, a tube in which the vapor is bombarded by a high frequency electromagnetic field, the radio-frequency generator, a series of three traps respectively maintained at temperatures of 0, -79 and -197 degrees centigrade and a pair of vacuum pumps for exhausting the system. To make an experimental run the storage vessel is first charged with a few milliliters of carbon tetrachloride and the Dewar flasks that surround the traps are filled, in the order of their temperature, with ice water, a mixture of dry ice and acetone, and liquid nitrogen. The system is then evacuated, the power is applied and the vapor of carbon tetrachloride is decomposed at controlled pressure. After the materials have reacted and condensed, all three traps are immersed in liquid nitrogen to freeze the products, and helium is admitted until the pressure of the system is in equilibrium with atmospheric pressure. The traps are then disconnected and the products are removed for analysis.

"The radio-frequency generator develops a maximum output of 600 watts at a fixed frequency of 14 megacyces. It consists of a seven-megacycle crystal-controlled oscillator, a frequency doubler, final amplifier and appropriate power supplies. The output can be continuously adjusted from O to 600 watts by means of a variable transformer that energizes the high-voltage power supply for the final amplifier and can be varied through the full range of power without influencing the frequency to which the unit is tuned. The final amplifier is also designed to operate under a wide range of loads without affecting the tuning. The alternating-current output of the generator flows through a coil of copper tubing that surrounds the glass tube conducting carbon tetrachloride from the storage vessel to the traps. The ends of the coil are connected to the amplifier through capacitors for insulation from the dangerously high voltage that energizes the final stage. The electronic apparatus was made entirely from parts for amateur radio transmitters.

"The vacuum system includes a mechanical fore pump and a single-stage mercury diffusion pump. The fore pump was not easily acquired. I first tried to reverse a small compressor of the piston type but could not make it produce a vacuum better than 40 millimeters of mercury. The sealed rotary compressor from a Frigidaire was then converted for operation in reverse by sealing the oil-bypass line and removing the intake check valve and screens. At its best this unit produced a vacuum of two millimeters. Three Norge Rollator compressors were then converted. Two of the units had been operated with sulfur dioxide and contained a large amount of sludge. The third, in which Freon 12 had been used, was much cleaner. The oil seals on the shafts were polished with fine emery paper, all screens and intake check valves were removed and new gaskets were installed. The three pumps ere then filled with Welch Duo-Seal vacuum-pump oil and tested. The unit that had used Freon performed best, producing a vacuum higher than .1 millimeter at 600 revolutions per min. I found that refrigerator compress are inferior to conventional fore pumps in two respects: their pumping speed at pressures below one millimeter relatively low, and lubricant is drawn into the system when vacuum is maintained at the input of the idle pump.

"My diffusion pump is of the water-cooled type. Running water could not be used, however, because the apparatus was designed for portable operation. The pump was therefore equipped with a refrigeration unit removed from a discarded soft-drink cooler. The evaporator coil of this particular unit was manufactured as an integral part of a heat exchanger that included a second coil for cooling water. The heat exchanger was wrapped with insulation to prevent the condensation of water vapor; alcohol was circulated through it and the jackets of the diffusion pump by a small rotary pump from a washing machine.

"Pressure in the system is measured by a sealed-end manometer calibrated from 50 millimeters to one millimeter of mercury and a McLeod gauge calibrated from one millimeter to .001 millimeter. The McLeod gauge is of the Stokes type, in which mercury is forced into the capillary tubes by tilting the unit 90 degrees [see "The Amateur Scientist," SCIENTIFIC AMERICAN, January, 1939, and March, 1960]. Only the glass part of the gauge was bought. The supporting framework and bearing were improvised from aluminum strip and pipe. Pressure is determined by the McLeod gauge by first trapping part of the residual gas in a vessel of known volume, then compressing it by means of a mercury column into a capillary tube closed at the outer end. Simultaneously the mercury is allowed to rise in an identical capillary that is connected to the vacuum system under measurement. Pressure is indicated by the relative height to which the mercury rises against the gas trapped in the sealed capillary. The gauge is calibrated by determining the volume of the trapping vessel and the cross-sectional area and length of the capillaries. I made the volume determinations by measuring the amount of water required to fill the respective parts of the gauge.

"The evacuated portion of the apparatus is constructed of sections of Pyrex glass that are interconnected by tapered joints. It includes stopcocks for controlling the flow of raw material from the dispensing vessel, for connecting the vacuum gauges and pumps and for admitting air or helium to the system as desired. All joints and stopcocks are lubricated with Apiezon vacuum grease. The discharge region in which the raw material is dissociated by high-frequency bombardment consists of a straight section of tubing with an outside diameter of 20 millimeters supported near the ends so that the excitation coil of the generator can be moved to any position along the tube.

"I experienced no particular difficulty in handling the cold fluids that are used for refrigerating the traps. The liquid nitrogen was received and kept in a 15-liter Dewar flask. A simple dispensing apparatus was improvised out of copper tubing and a small air compressor for transferring liquid nitrogen from the shipping container to the appropriate trap. A long section of tubing, bent over at the top for a spout, and a shorter length connected to the air compressor were inserted in a rubber stopper that fitted the shipping container. Compressed air entering the short tube forced liquid nitrogen up through the spout as needed. I found that some care must be taken in handling dry ice and acetone. When the mixture absorbs heat too quickly, carbon dioxide forces acetone out of the container, damaging the equipment and presenting a fire hazard.

"To make the apparatus portable and self-contained, all the components were mounted according to function on a three-tiered tool cart 32 inches high, 28 inches wide and 20 inches deep. The cart was modified by bolting a quarter-inch steel plate to the underside of the bottom tray for extra strength and by adding a set of rubber casters. The fore pump, refrigerator system and variable transformer occupy the bottom tray. The radio-frequency unit was placed on the middle tray. The reaction system, together with the diffusion pump and necessary apparatus supports, was on top [see Figure 2].

"A small laboratory was set up in which to analyze the reaction products. The problem of analysis consists in determining the atomic ratios of two-element compounds that can be dissolved in water and whose constituent elements are known. The major items of equipment include an analytical balance, a centrifuge and a drying oven. A used balance was obtained and renovated by cleaning the knife edges and replacing the level, which was badly corroded, by a hand level. Weights were then made up and calibrated against a set of standards. The drying oven is merely a box of sand heated by an infrared lamp. A test tube containing the sample is placed in the hot sand and its temperature is determined by a thermometer. In making an analysis a three-milliliter test tube is weighed after drying; the sample, purified by vacuum recondensation if necessary, is added and the tube is weighed again. The sample is then dissolved in water and one of the ions is precipitated.

"After being washed and centrifuged the sample is dried in the temperature-controlled oven. Because the precipitate is known, the highest temperature at which it can be dried is determined by its heat of formation. The precipitate and test tube are weighed. Since the total weight of the compound and the weight of one of its elements are determined by this analysis, the atomic ratio can be calculated. A simple formula can then be written and the substance identified. The melting and boiling points of the sample are determined by a method that requires only minute amounts of material. The sample is placed in a small capillary tube, the ends of which are sealed in a gas flame. The tube is then placed in a beaker of mineral oil and heated gradually over a hot plate con trolled by a variable transformer. An v electric stirrer keeps the temperature of the oil uniform. The temperature indicated by a thermometer immersed in the oil; changes in the state of the sample are observed visually.

"A spectrophotometer has also been constructed for identifying products by their characteristic spectral absorption. Light from a tungsten lamp passes through the sample, is diffracted by a replica transmission grating and is analyzed by a photoelectric cell. The mechanical parts of the instrument were salvaged from an old spectrometer of the Bunsen type. All lenses were replaced by adjustable slits made from safety razor blades and the optical sys, tem was lined with black felt to minimize internal reflection. The spectrum is scanned by rotating the tube that l holds the photocell through an appropriate angle with a motor-driven jack; screw. The output of the photocell, which varies between a tenth and a hundredth of a microampere, drives a three-stage direct-current amplifier that includes a negative-feedback circuit for minimizing both drift and distortion. Output, proportional to the spectral characteristics of the sample, is indicated by a microammeter. The characteristics of the instrument were determined by scanning the spectrum without a sample and plotting the amplified photocurrent against the wavelengths of the spectrum. This standard graph is then added to graphs subsequently made of samples in order to correct for variations in the spectral emission of the light source and the spectral sensitivity of the photocell. In addition to these instruments I had access to an infrared spectrophotometer through the co-operation of a local organization. The resulting graphs made by professional workers have been most useful in identifying some of the reaction products.

"Although the apparatus was constructed for experimenting with chemical reactions, I made a number of preliminary tests with air, mercury, helium and vaporized compounds in order to learn how the various components of the system worked physically and, in particular, to become familiar with the effects that occur in the section exposed to the high-frequency field. When an ionizing reaction is initiated in this region, at pressures that enable a gas or vapor to emit light, effects of two distinctly different types are observed. High power applied to the vapor of carbon tetrachloride at a pressure between one micron (.001 millimeter) and 50 microns of mercury causes an intense glow, rich in ultraviolet light, that fills most of the available space. The tube of the final amplifier draws a high current but the over-all strength of the high-frequency field is low. When the power is lowered, the character of the glow changes abruptly to that of a conventional high-voltage discharge and is confined to the region near the coil. The current to the final amplifier decreases and the intensity of the field increases. A discharge of the low-power type can be maintained in the case of carbon tetrachloride at pressures between .1 micron and five millimeters. Both types of discharge appear to be effective in decomposing test materials. I used the low-power type in most experiments, however, because less heat is produced by the discharge and better temperature control is possible.

"Tests were also made of various substances before I selected carbon tetrachloride for extensive experimentation. A series was run on aluminum chloride, for example. The sample had to be heated to 50 degrees C. to develop a vapor pressure high enough to sustain a glow in the high-frequency field. As the reaction progressed a yellow deposit condensed in the first trap that could not be analyzed because it decomposed on contact with the air. In addition a coarse, silver-gray powder accumulated in the first and second traps. This material appeared to be aluminum chloride that had recrystallized before entering the discharge region and had then acquired a coating of metallic aluminum. This would explain why the heavy particles were swept through the system without condensing.

"Similar difficulties were experienced with other compounds that required heating for the development of adequate vapor pressure. All recrystallized or changed in some other undesirable way when entering the lower temperature of the discharge region. The glass tubing of these sections could not be warmed conveniently because expansion induced by the heat could break the tapered I joints. This limits the apparatus to experiments with substances that develop adequate vapor pressure at room temperature.

"Other considerations also restrict the choice of specimen materials. When a compound of very low vapor pressure at room temperature is used, decomposition proceeds slowly and the yield is correspondingly small. If the vapor pressure is substantially below one micron-too low for the discharge to exist-the total pressure can be increased and decomposition carried out by admitting helium as a carrier gas. Conversely, test substances must not develop a higher pressure at the temperature of the system than will sustain ionization. Finally, the vapor pressure of the most volatile product expected must be negligible at the lowest trap temperature to prevent harmful products from entering the vacuum gauges and the pumps.

"Carbon tetrachloride meets all these requirements. In a typical experimental run the supply vial is filled, connected to the system and periodically immersed in a bath of dry ice and acetone as required to maintain a vapor pressure between two and five millimeters. The material is decomposed at a rate of about 10 milliliters per hour. The first product forms on the wall of the discharge tube, its composition depending on the temperature of the glass. At room temperature or higher the product is dark brown and nonvolatile. When it forms below 10 degrees C., it is a colorless solid that can be sublimed in high vacuum. I did not succeed in collecting enough of either substance for analysis, but they appear to be polymers of carbon and chlorine.

"A clear viscous oil collects in the first trap. The infrared spectrum of this oil is characterized by six medium-sized to high peaks at 11.8, 12, 12.3, 13, 13.4 and 13.6 microns. When the oil is allowed to stand, it crystallizes into colorless crystals that melt at between 65 and 67 degrees C.

"Two fractions collect in the second trap. The least volatile was identified as hexachloroethane from its sealed-tube melting point and infrared spectrum. The second was established to be tetrachloroethylene by comparing its boiling point and infrared spectrum with those of a known sample.

"Two products collected in the third trap, one condensing on the wall of the entering arm of the U and the other at the bottom. The first substance breaks it down into a variety of substances when evaporated at room temperature followed by recondensation at -190 degrees C. The principal constituents of the decomposed mixture are tetrachloroethylene and carbon tetrachloride. A small amount of hexachloroethane is also present, along with a highly volatile constituent that appears to be dichloroacetylene. The product collected at the bottom of the third trap is a pale yellow crystalline substance that melts at slightly below-100 degrees C. and contains chlorine plus a small amount of dichloroacetylene. The mixture appears to be identical with material previously identified as dichlorocarbene. It reacts readily with air to yield phosgene, and on being stored at room temperature it yields a mixture of compounds, chiefly hexachloroethane.

"The mixture of all fractions collected in the third trap was aged for two months in a sealed ampoule free of oxygen. Infrared spectra of the decomposed residue indicated the presence of hexachloroethane and possibly a small quantity of hexachlorobenzene. No tetrachloroethylene could be detected in spite of the fact that it is one of the products of the material that condenses in the entering arm of the third trap.

"The positions of the traps and relative variations in the temperatures of the interconnecting tubing have a major effect on reactions. Condensation in sharply defined bands is encouraged by uniform temperature in a specific section and the composition and yield of products both by the velocity at which the dissociated material migrates from the discharge section to th traps and by the temperature of intervening zones.

"The technique of decomposing carbon tetrachloride by a high-frequency field appears to have several advantages over the conventional procedure of passing vapors at a pressure of one micron or less over tungsten or carburized tungsten heated to temperatures ranging from 1,200 to 2,000 degrees C. When vapors are decomposed by a high-frequency field, pressures of several millimeters and a high rate of flow can be maintained. All molecules are fragmented and the yields are much higher. In addition manipulation is greatly simplified and the process requires only a few watts of power.

"One characteristic of the radio-frequency generator was responsible for a side effect of possible interest to anyone who undertakes this experiment. The final amplifier radiates harmonics that interfere with television reception. Accordingly our neighbors firmly suggest that I find something else to do except between 1:00 A.M. and 7:00 A.M."

Bibliography

QUALITATIVE TESTING AND INORGANIC CHEMISTRY. Joseph Nordmann. John Wiley & Sons, Inc., 1957.

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.