ANALYSIS OF 1,4-DIOXANE 285 RESULTS AND DISCUSSION A rapid procedure that could be applied to very small experimental samples was desired for the analysis of 1,4-dioxane in alcohol ethoxylates. Early attempts at direct injection gas chromatographic procedures using various single columns were unsuc- cessful due to interferences from other components eluting with the dioxane. The nature of the two required separations--l)that of separating volatile components from the heavy ethoxylates and 2) separation of polar volatile components from each other, suggested that a dual-column system with column switching might be applicable. This is the type of column-valve configuration described in ASTM Method D-3606-77 (8) for the analysis of benzene in gasolines. Several modifications were made to this method. A flame ionization detector was used instead of the prescribed thermal conductivity detector. To achieve baseline resolution of 1,4-dioxane and the chloro- benzene solvent, a 3:1 mixture of 10% methyl silicone stationary phase (SP2100 or OV-101) and 6% Carbowax 20M was required as the packing in the stripper column (Column A, Figure 1). The alcohol ethoxylates dissolved in chlorobenzene are injected directly into the gas chromatograph. When the light components including dioxane and the chlorobenzene solvent have eluted from the SP2100/Carbowax 20M column, the six port valve is switched, and the heavy components are backflushed from the first column by reversing the flow. The use of an auxiliary carrier gas allows continued analysis on the polar analytical column while the first column is being backflushed. It was found that o o I mg/kg 1,4-Dioxane Std. in Chlorobenzene 2 min Back Flush Dioxane ectrometer Range: 10-•2 nuation: x 4 • I • I 16 20 24 28 Time, min Figure 3. Chromatogram for a 1 mg/kg standard. 32

256 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS • [ • Chlorobkg/mg5 1,4-Dioxane, i•2minBackFlush l/ / • o Sample Size: 2 -- Electrometer Range: 10-12 AFS -- Attenuation: x 4 -- 4 8 12 16 20 24 28 32 36 Time, min Figure 4. Chromatogram of an ethoxylate spiked with 5 mg/kg dioxane. variation of the flow rates given in Table I causes the dioxane peak to co-elute with the solvent peak or to be backflushed. A glass injection port insert, which reduces peak tailing, is replaced daily to prevent polymer migration onto the column head. Even though the total analysis cycle time is 2.0 hr, only 15 min of technician time are required and four to five analyses can be performed in an eight-hour day. The column switching and backflushing can be automated for installation in a quality control laboratory. Since the ethoxylates are dissolved in chlorobenzene and injected directly, the method is applicable to alcohol ethoxylates ranging from low molecular weight liquids to high molecular weight solids. The limit of detection for 1,4-dioxane in ethoxylates using this method is 0.5 mg/kg with a linear range of up to 10 g/kg (1%). Quantification is accomplished by relating peak height measurements to an external standard curve, which is checked daily by a standard analysis. The coefficient of variation of the standards was 4.2% based on 80 standard solution analyses. Figure 3 shows a typical chromatogram obtained for the analysis of a 1 mg/kg 1,4-dioxane in chlorobenzene standard solution. The recovery of the method was determined by analyzing an ethoxylate sample spiked with 1, 3, and 5 mg/kg, respectively, of 1,4-dioxane using the above procedure (Figure 4). In all cases the recovery was 100% at these levels. The repeatability of the method,