Solvent Fractionation of Girard T Derivatives of Carbonyl Compounds Using Dimethyl Sulfoxide S. F. Osman and J. L. Barson Eastern Utilization Research and Development Division, Agricultural Research Service, U.S . Department of Agriculture, Philadelphia, Pa. 19118
IN THE COURSE of our investigation of the chemical composition of cigar smoke we recently isolated a neutral fraction by silicic acid column chromatography which was shown to contain at least 30 compounds by gas chromatographic analysis. Attempts to obtain the pure components of this fraction by further chromatographic methods were unsuccessful. Therefore, we attempted to separate this fraction into a carbonyl and noncarbonyl subfraction. Isolation of the carbonyl compounds via the corresponding 2,4-dinitrophenyl hydrazones was ruled out because of the difficulties encountered in regenerating the free carbonyl compound. Reaction of the carbonyl compounds with Girard T reagent (tetramethyl ammonium acetyl hydrazide) to give a polar hydrazone (Equation l),which can be separated from relatively nonpolar unreacted compounds by solvent partitioning, appeared to be a desirable alternate route for this separation: 0
I/
RI-C-Rz
+ + H2N-N-C--CH2N(CH3)3
e
+
Ri
\
C=N-N-C--CHzN(CHs)a
/
+ HzO
(1)
R2 We used this method to separate farnesyl acetone (6,10,14trimethyl-5,9,13-pentadecatriene-2-one) from the Czo-terpene alcohol, phytol [these two compounds were previously identified in this fraction (I)]. We did not obtain encouraging results using reaction conditions described in the literature (2). When the two compounds were subjected to treatment with Girard T reagent in absolute ethanol followed by the addition of water (giving a 15 % aqueous ethanol solution) and then extraction of the unreacted phytol with hexane, an appreciable amount of farnesyl acetone was also recovered (1) S. Osman and J. Barson, Tob. Sci.,X, 85 (1966). (2) 0.H.Wheeler, Chern. Reo., 62, 205 (1962).
in the hexane layer. That the free farnesyl acetone found in the hexane was a product of hydrolysis of the already formed hydrazone and not the result of the incomplete reaction of the Girard reagent with the ketone was demonstrated in the following experiment. Farnesyl acetone was reacted with the Girard reagent in absolute ethanol, the unreacted Girard reagent was precipitated by adding hexane to the solution, the precipitate was filtered, and the filtrate was analyzed by gas chromatography. No farnesyl acetone could be detected in the hexane. To delineate the effect of hydrolysis, a standard mixture (containing compounds similar to those we believed to be present in our tobacco sample) was subjected to a wide variety of reaction and separation conditions described in the literature. The results of these experiments are given in Table I. In all cases hydrolysis of the hydrazone of farnesyl acetone and octanone is evident even when the amount of water introduced to produce an immiscible polar layer (Case 1) is kept at a minimum. To avoid this problem we decided to explore reaction conditions using a nonaqueous system. We reacted our known mixture with Girard T reagent in dimethyl sulfoxide (DMSO) for 1 hour, followed by extraction of the unreacted compounds with hexane. The separation, although not complete, was better than we had obtained using aqueous systems. The presence of carbonyl compounds in the noncarbonyl fraction could be avoided completely. With six extractions, only 65 of the geraniol was recovered (which is still the best separation obtained). The cigar smoke fraction was reacted with the Girard T reagent as follows: The condensate (dry weight approximately 1.0 gram) was dissolved in 10 ml of DMSO, 0.5 gram of Girard T reagent was added, and the solution stirred for 3 hours at room temperature. The DMSO solution was then extracted with 6 X 10 ml of petroleum ether. The petroleum ether was concentrated and analyzed by gas chromatography. At least 20 peaks were apparent in the chromatogram of which phytol (a methyl ethyl phenol) and dihydrocinnamyl nitrile were identified by spectroscopic examination of the corresponding peak eluates. The DMSO solution was diluted with an excess of water, heated to 50' C for 1 hour and then extracted with hexane. Farnesyl acetone, solanone (2-methyl-5-isopropyl-l,3-nonadiene-8-one), and a
Table I. Summary of Girard T Reaction in Various Solvent Systems
H10 added in Reaction solvent Extraction solvent isolation step, % Absolute ethanol Petroleum ether 15 Absolute ethanol Petroleum ether 30 Absolute ethanol Methylene chloride 15 r-Butanol Petroleum ether >lo0 Petroleum ether Dimethyl sulfoxide 0 Free carbonyl compounds that are extracted along with geraniol. * Farnesyl acetone. c By extraction with equal volumes of hexane 6 X. 0
530
ANALYTICAL CHEMISTRY
x Unreacted carbonyl" Citral 0 0 0 55 0
2-Octanone 10 40 40
F.a.b 15 49
90
43 93
0
0
Geraniol recovered, %" 50 70 82 55 65
C-17 methyl ketone (not completely characterized) were identified in this fraction. There was no evidence of compounds such as phytol in this fraction, and none of the carbonyl compounds were apparent in the gas chromatogram of the hexane extract containing the phytol. Until both hexane fractions are completely characterized, however, we cannot state with certainty the degree of separation. In conclusion, many Girard T hydrazones are susceptible to hydrolysis even under mild conditions. This problem limits the use of the Girard reagent for many separations of carbonyl from noncarbonyl compounds. A nonaqueous system has been developed to overcome this problem. Al-
though poor partitioning factors may arise in the nonaqueous system these can be minimized by increasing the number of extractions. ACKNOWLEDGMENT
The authors thank J. Showell and E. Rothman for their helpful discussions during the course of this work. RECEIVED for review September 27, 1966. Accepted January 18, 1967. Work supported by the Cigar Manufacturers Association of America.
Determination of Ephedrine and Certain Related Compounds by Ultraviolet Spectrophotometry Jack E. Wallace Forensic Toxicology Branch, USAF Epidemiological Laboratory, Lackland Air Force Base, Texas THEALKALOID EPHEDRINE (1-phenyl-2-methylamino-propanol) and the related compounds pseudoephedrine and phenylpropanolamine have similar medical uses. These drugs have a low toxicity but their identification and quantitative determination in biologic specimens is, nevertheless, frequently required. Each of the compounds has a structure that includes a benzyl alcohol group and this functional group is utilized as a basis for their determination by a spectrophotometric method described in this report. Many pharmacologic and physiologic investigations have been concerned with the occurrence and biochemical action of this chemical class of compounds and the literature contains numerous descriptions of isolation and determination procedures. The many methods attest not only to the importance of the drugs but also to the difficulty in achieving analytical methods which are sensitive and selective. For example, many alkaloids give a positive reaction with the colorimetric procedures which have been described (1-3) for determining ephedrine. Fluorometric techniques ( 4 ) are extremely sensitive but these procedures are often time-consuming. Nonaqueous titration systems (5, 6) exhibit a low order of sensitivity and specific;ty when applied to biologic specimens. The ion exchange chromatographic methods (7, 8) are excellent for separation (of closely related compounds but they often fail to yield reproducible quantitative results. Paper (9, IO) and gas (11-13) chromatographic procedures are (1) L. G . Chatten and L, Levi, ANAL.CHEM., 31,1581 (1959). (2) F. Feigl and E. Silva, J. Am. Pharm. Assoc., 47,460 (1958). (3) G . N. Thomis and A. Z . Kotionis, Anal. Chim. Acta, 16, 201 (1957). (4) P. Laugel, Compt. Rend.,225,692 (1962). (5) R. Reiss, 2.Anal. Chem., 167, 16 (1959). (6) M. Rink and R. Lux,Arch. Pharm., 294,117 (1961). (7) S. Blaug and L. Zopl', J. Am. Pharm. Assoc., 45, 9 (1956). (8) L. Sanders, P. H. Elworthy, and R. Fleming, J . Pharm. Pharmacol., 6, 32 (1954). (9) K. Hiller, Phapmazie, 16, 600 (1961). (10) E. Vidic and J. Schuette, Arch. Pharm., 295, 342 (1962). (11) M. Anders and G. Mannering, ANAL.CHEM., 34, 730 (1962). (12) L. Kazyak and E. Knoblock, Ibid.,35,1448 (1963). (13) K. Parker, C. Fontan, and P. Kirk, Zbid.,34, 1345 (1962).
sensitive, but they lack the ability to provide the positive identification which is required by the forensic scientist. Ephedrine and its related compounds have a low order of molar absorptivity; therefore, direct ultraviolet spectrophotometric assays do not exhibit sufficient sensitivity for determination of therapeutic concentrations of these drugs in biologic materials (14). Chafetz (15) described a sensitive ultraviolR spectrophotometric method utilizing periodate oxidation for determining several phenethanolamine compounds in pharmaceutical preparations but the procedure is not applicable to biologic specimens. The need for a specific and sensitive spectrophotometric method for determining ephedrine in biologic fluids is apparent. The method described in this report is rapid, sensitive, and is group specific for ephedrine and those related compounds which extract as a base and have a benzyl alcohol functional group. Preliminary separation from other alkaline drugs is not required. EXPERIMENTAL
Apparatus. A Beckman DK-2A ratio-recording spectrophotometer with linear presentation, settings and operation routine, was used for the ultraviolet absorption measurements. The sample path length was 10 mm throughout. Infrared spectrograms were prepared on a Beckman IR-4 spectrophotometer. A Barber-Colman Model 5000 gas chromatograph with a coiled 6-foot 1 S.E. 30 column was used for gas chromatographic analysis. Procedure. Ten-milliliter amounts of blood, serum, or urine are placed in a 250-ml separatory funnel to which 1 ml of 1N NaOH and 100 ml of ether are added. This mixture is shaken vigorously for 3 minutes. The ether is removed and filtered through Whatman No. 541 filter paper. Complete recovery of the ether should not be attempted, but the volume of ether recovered is recorded and is included in the final calculations. Five milliliters of 1M acetic acid are added to the filtered ether and the mixture is shaken for 5 minutes. Four milliliters of the aqueous layer are transferred to a dry 1000-ml round-bottom flask, and 10 ml of (14) H. Thies and Z . Oezbilici, Arch. Pharm., 295, 715 (1962). (15) L. Chafetz, J . Pharm. Sci., 52, 1193 (1963). VOL. 39, NO. 4, APRIL 1967
531
