GAS CHROMATOGRAPHY OF AEROSOLS 361 Figure 4. Resolution of P-152a, P-142b, and P-22 (speed: 0.5 in./min) A. Nitrogen D. P-22 B. P-152a E. Vinylchloride C. P-142b sampling and bleed valves provided a point of observation of gross frac- rionation. With an excessive bleed rate, there was a noticeable loss in the propellants with the higher vapor pressures. The system also provided for the analysis of nonaqueous aerosol con- centrates consisting of mixtures of methylene chloride, resins, perfume, and alcohol, prior to filling. Increasing the column temperature to 87øC shortened the analysis time considerably, but still provided a baseline separation. Concentrate samples were injected through the B side injection port with a 5-/•1 syringe, and the peak areas were deter- mined. Determination of concentrate/propellant ratios was also pos- sible by a comparison of the peak area of one of the propellant peaks to the alcohol peak in finished containers. Sample size for concentrate analysis was 2.0-3.0 ul. Finally, a relatively fast, qualitative analysis of headspace vapors and volatiles in the liquid phase was also possible by attaching a sample- piercing valve at a point above the liquid-vapor interface and attaching a nut with a self-sealing septum to the threaded outlet of the piercing

362 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table III Typical Sample Analysis Utilizing Internal Standard Sample % Theoretical % Found Code Component (W/W) (W/W) A P-12 20.43 20.38 P-114 12.66 12.72 P-11 15.20 15.32 P-114B2 10.57 10.49 Methylene chloride 31.14 31.02 Ethanol 10.00 10.07 B P-115 15.21 15.32 P-152a 9.77 9.70 P-11 2.43 2.39 Methylene chloride 23.05 23.21 Ethanol 21.88 21.76 Isopropanol 27.66 27.62 C P-12 35.00 35.09 Isobutane 10.31 10.43 n-Butane 5.04 4.95 P-11 21.68 21.57 P-21 16.74 16.63 Isopropanol 11.23 11.33 D P-12 10.96 10.89 P-114 62.49 62.60 Methyl chloride 6.88 6.81 Ethyl chloride 12.60 12.52 P-21 7.07 7.18 E Propane 4.11 4.02 Isobutane 58.29 58.37 n-Butane 5.56 5.49 P-11 32.04 32.12 device. A vapor sample was then withdrawn through the septurn with a 1.0-ml gas-tight syringe, and allowed to equilibrate briefly at one atmo- sphere finally, a 0.3-0.5 ml sample was injected. Warming the con- tainer to 43-54 øC prior to sampling aided in the analysis. This pro- cedure was particularly useful in qualitative analysis of unknown sam- ples, especially in aerosol powder formulations and aqueous systems, which are not directly applicable to pressurized liquid sampling. The 2-ul precision sampling valve proved to be more suitable for this type of analysis than the 10-ul valve used previously. The latter ap- peared to monitor the pressurized sample at the interface of the Teflon core and the steel body, so that some tailing of all components had been