116 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Typically, a hair spray formulation contains several volatile ingredi- ents, e.g., ethanol, a denaturant, and propellants. Of the latter, Pro- pellants 11 (trichlorofluoromethane), 12 (dichlorodifluoromethane),* and isobutane are used singly or in mixtures. Gas-liquid chromatography was first reported for separating chloro- fluoromethanes and a solvent in 1956, when Green (2) reported the separations of Propellants 11, 12, and 13, and carbon tetrachloride. The column was packed with Celite 545* impregnated with approximately 30% dibutyl phthalate. The gaseous sample was introduced into the column which was jacketed and held at 24øC by circulating trichloro- monofiuoromethane (Propellant 11) vapor through the jacket. Twenty- five minutes later the analysis was complete and calculation of the com- ponents could be made from the curve obtained on a 5-mv recorder. In 1956, Root and Maury (3, 4) described the analysis of aerosol products. Included in their examples was the gas chromatographic sep- aration of Propellants 11, 12, methylene chloride, and ethanol. The method required the volatilization of the components under vacuum and the injection of a large (9.15 ml) sample into the gc. The alcohol, which eluted last, was off the column after 18 min. Errors in the analysis were attributed to the condensation o•: alcohol at the sampling pressure of 150 mm at room temperature. Higher temperatures or lower pres- sures complicated the analyses at higher alcohol concentrations. These early gc methods disclosed some problems associated with the analysis of mixtures of solvents and aerosol propellants. The three most evident were: obtaining a representative sample from the container transferring the sample quantitatively to the instrument and reducing the time on the gc column. Jenkins and Amburgey in 1959 (5, 6) reported dissolving, in cold cyclohexane, the chilled contents of an aerosol can containing Propellants 11, 12, and 114, methylene chloride, and ethanol. This solution was then injected into the gc with a hypodermic syringe. Quantitative sep- arations within 2% of the theoretical value were achieved. Brook and Joynet (7) reported using a special adaptor to fit the valve of an aerosol can and a hypodermic needle in order that a liquid sample could be transferred to an evacuated glass flask equipped with a rubber serum cap. The flask was then attached to a gas sampling valve * "Genetron," Allied Chem. Corp., Industrial Chem. Div., PO Box 70, Morristown, N.J., 07960 "Freon," E. I. du Pont de Nemours g: Co., Inc., Freon Products Div., Wilmington, Del. 19898 "Ucon," Union Carbide Chem. and Plastics Div., 207 Park Ave., N.Y. 10017. ?Johns-Manville, 22 E. 40 St., N.Y. 10016.

GAS CHROMATOGRAPHY OF METHYLENE CHLORIDE 117 of agc and the volatilized gas was chromatographed. The discussion following the presentation of the paper suggested that the method could be applied to the determination of methylene chloride. Several workers reported using can-piercing devices and liquid sam- pling valves to inject samples directly into the gas chromatograph (8-10). This eliminated determination of temperatures and pressures, vapor pressures of various components in aerosol systems when gaseous sam- ples were taken, and eliminated the use of elaborate and expensive vacuum sampling devices such as previously discussed. However, liquid sampling valves are expensive, and the aerosol can must be retained with the puncturing device in place, or discarded. Cohen (11) eliminated the liquid sampling valve by using a Pre- cision Sampling Corporation Pressure-Flo©* high-pressure liquid sam- pling syringe. He modified the can-piercing device to sample the liquid in the can directly through a fitted filling adaptor. This reduces the cost of sampling to below $100 but, again, it does destroy the sample container by piercing. Bourne and Murphy (10) separated ethanol and methylene chloride from a number of propellants, both fiuorocarbons and hydrocarbons, on a series Carbowax 20M* and porous polymer column. However, they never attempted the separation of methylene chloride, ethanol, and Pro- pellant 11 in the same system. Cannizaro and Lewis (9) separated these three components on long (9.2 m) columns packed with 20% Halloomid M-18.* Under the conditions given in their procedure, methylene chlo- ride eluted in 13.5 min. By increasing the temperature of the column oven the retention time of the methylene chloride was shortened. In our work, we chose to use the porous polymer columns because the work of Bourne and Murphy (10) and our own experiments had shown we could achieve quick, clean separations of the propellants and solvents found in hair sprays. Because Par II©,õ the column previously used (10), is no longer available, we used Porapak©ll instead. Our re- sults showed incomplete separation of the Propellant 11 and methylene chloride. Increasing the length of either the Carbowax 20M or the porous polymer column or reducing the mesh size of the porous polymer * Precision Sampling Corporation, P.O. Box 15119, Baton Rouge, La. 70815. • Polyethylene glycol, tool wt ca. 20,000, Union Carbide Chemicals Co., 270 Park Ave., N.Y. 10017. $ Applied Science Laboratories, Inc., P.O. Box 440, State College, Pa. 16801. õ Hewlett-Packard Co., Route 41, Avondale, Pa. 19311. II Waters Associates Inc., 61 Fountain Street, Framingham, Mass. 01701.