1530 Anal. Chem. 1984, zyxwvut 56, 1530-1533 also supported our view. The contamination level was about 17 pg/sample with the standard deviation of 1.5 pg of sele- nium. The working detection limit was, therefore, calculated to be 3 pg/sample, taking twice the standard deviation of blank value as the detection limit. Table I1 summarizes the detection limits for selenium by various methods. The present method gave the lowest de- tection limit compared with other values reported. It should be also pointed out that a fairly large amount of the extract can be injected into the HPLC system without loss of selec- tivity. This is one of the advantages of HPLC over other extraction-detection systems that use TLC or GC as the separation step (9-11, zyxwvutsrq 14). The injection volume can be in- creased to 80 pL without loss of selectivity. Furthermore, the volume of cyclohexane can be reduced considerably with little loss of extraction efficiency. For example, the recovery was 94% when zyxwvutsrqp 5 mL of the reaction mixture containing 100 pg of selenium was extracted with 1 mL of cyclohexane and 86% when the same solution was extracted with 100 pL of cyclo- hexane. Thus we should obtain a potentially lower detection limit (0.19 pg of Se/sample) by extracting with a 100-pL volume and injecting 80 pL of the extract, provided that the blank peak could be completely removed. Our system has now reached a level where the working detection limit seems to be determined by the contamination level of selenium(1V) in the reagents. Registry No. HzO, 7732-18-5;Se, 7782-49-2;NSD, 269-20-5; DAN, 771-97-1. LITERATURE CITED Burk, R. F.; Pearson, W. N.; Wood, R. P.; Viteri, F. Am. J. Clin. Nub. 1967, 20, 723. Levine, R. J.; Olson, R. E. Roc. SOC. Exp. Biol. Med. 1970, 134, 1030. Money, D. F. L. N. Z. Med. J. 1970, 71, 32. Yamamoto, Y.; Kumamaru, T. Fresenius’ zyxw 2. Anal. Chem. 1976, 287, 353. Thompson, K. C. Analyst (London) 1975, 100, 307. Goulden, P. D.; Anthony, D. H. J.; Austen, K. D. Anal. Chem. 1981. 53, 2027. Watklnson, J. H. Anal. Chim. Acta 1982, 134, 417. Parker, C. A.; Harvey, L. G. Analyst (London) 1962, 87, 558. Funk, W.; Kerler, R.; Schiller, J. Th.; Dammann, V. HRC CC, J. High Resolut. Chromatogr. Chromatogr. Commun. 1962, 5, 534. Moreno-Domlnguez, T.; Garcia-Moreno, C.; Marine-Font, A. Analyst (London) 1983, 108, 505 Stahr, H. M.; Kinker, J.; Nicholson, D.; Hyde, W. J. Liq. Chromatogr. 1982, 5, 1191. Watkinson, J. H. Anal. Chem. 1966, 38, 92. Watkinson, J. H. Anal. Chlm. Acta 1979, 705, 319. Toei, K.; Shimoishi, Y. Talanta 1981, 28, 967. RECEIVED for review December 19, 1983. Accepted March zy 7, 1984. Determination of Oxirane Ring Position in Epoxides at the Nanogram Level by Reaction Gas Chromatography Athula B. Attygalle and E. David Morgan* Department of Chemistry, University of Keele, Staffordshire zyxwvut ST5 5BG, United Kingdom Epoxide groups are present in many biologically important natural products. The juvenile hormones (I-IV) are one group of important epoxides encountered in insects. Disparlure (V), (Z,Z)-cis-9,1O-epoxyheneicosa-3,6-diene, and cis-9,lO-epoxy- tricosane (VI) found in the gypsy moth (I), the saltmarsh caterpillar moth (2), and the house fly (3), respectively, are some examples of epoxide sex pheromones. Therefore the techniques to locate the epoxide position are of great im- portance to the natural products chemist. When mass spectrometry (MS) facilities are available, EI-MS (4) and particularly CI-MS (5) are useful to locate the position of the oxirane ring. As an alternative to MS, simple microchemical methods are often employed to determine the epoxide posi- tions. Bier1 et al. (6) performed this by the cleavage of 1-100 pg samples with periodic acid in a chlorinated solvent and subsequent examination of the carbonyl products by GC. A column of periodic acid on calcium sulfate has been used by Schwartz et al. (7) to cleave micromole amounts of epoxides to aldehydes. Similarly, Mizuno et al. (8) used HIOl in an- hydrous ether and subsequently analyzed the carbonyl products formed by GC. We have recently described some reaction gas chromatog- raphy methods, without solvent, for the identification of na- nogram quantities of natural products (9). In the present study, a simple reaction gas chromatographic technique was developed to locate the oxirane position in nanogram quan- tities of unknown epoxides by cleavage of the epoxide to corresponding carbonyl compounds with a periodic acid precolumn. EXPERIMENTAL SECTION Apparatus, A Pye-Unicam PU 4500 gas chromatograph with a flame ionization detector (FID) was used for GLC, using one of the following columns: (A) 2.75 m X 4 mm (Ld.) glass column packed with 10% PEG 20M on Chromosorb W, 100-120 mesh; 0003-2700/84/0356-1530$0 1.50/0 (11) R’=Et, Rz=R3=Me JH I1 R’ R3 (HI) R’=R’=R’=MR JH 111 (IV) R’=R*=R’=Et JH 0 and (B) 1.5 m X 4 mm (i.d,), glass column packed with Porapak Q (Waters Associates, Milford, MA), 100-120 mesh. Nitrogen was used as the carrier gas at a flow rate of 50 mL/min. Reagents. Periodic acid supplied as H,IO, by Fluka AG (Buchs, Switzerland) was ground to a fine powder and dried to a constant weight in an evacuated drying pistol at 100 OC. m- Chloroperbenzoic acid was purchased from Aldrich Chemical Co. (Milwaukee, WI). Test Compounds. (7S,8R)-cis-7,8-Epoxy-P-methyloctadecane (disparlure) was a gift of B. A. Bierl-Leonhardt (USDA). The juvenile hormones were purchased from Sigma Chemical Co. (St. Louis, MO). The noncommercially available epoxides were synthesized from the corresponding alkenes by reacting with m-chloroperbenzoic acid (IO). Precolumn Preparation. The precolumn packing (IO%, w/w) was prepared by evaporating a solution of anhydrous periodic acid (100 mg) in absolute ethanol in contact with 5% OV-101 on 0 1984 American Chemical Society