Anal. Chem. 1980, 52, 572-573
572
multiplet overlap, poor S / K , and second-order distortions, as mentioned in the introduction. ACKNOWLEDGMENT Thanks are due to U. Obenius for technical assistance and stimulating discussions and t o M. E. Moseley for linguistic corrections and stimulating discussions. We thank G. Bergson for making the spectrometer available to us. LITERATURE CITED (1) Ernst, R. R. J . Chem. Phys. 1966, 4 5 , 3845. (2) Freeman, R.; Hill, H. D. W. J . Chem. Phys. 1971, 5 4 , 3367. (3) Birdsall, B.; Birdsall, N. J. M.; Feeney, J. Chem. Commun. 1972, 316.
(4) Wehrli, F. W.; Wirthlin, T. "Interpretation of Carbon-I3 NMR Spectra", Heyden: London, 1976: pp 66-77. (5) Chandler, J. P. Program 66, Quantum Chemistry Program Exchange. Chemistry Dept., Indiana University, Bloomington, Ind. 47401 (6) Johannesen, R. B.; Ferretti, J. A , ; Harris, R. K . J . Magn. Reson. 1970, 3 , 84
P e t e r Stilbs Institute Of Chemistry Uppsala University Box 532, 751 21 Uppsala, Sweden
RECEIVEDfor review August 16, 1979. Accepted December 12,1979. This research was supported by the Swedish Natural Sciences Research Council.
N-Methylimidazole as a Catalyst for Acetylation of Hydroxyl Terminated Polymers Sir: A number of prepolyers used in solid rocket propellant formulations contain hydroxyl end groups which are reacted with organic isocyanates to provide a cross-linked rubber binder. In the past, the equivalent weight of such polymers has been determined by esterification of the hydroxyl groups with acetic anhydride (AA) catalyzed with pyridine (PY) ( 2 ) . When this technique is used, polymers often require as much as 4 h for quantitative esterification (2). More recently phthalic anhydride (PA) has been substituted as the esterifying reagent (3). In addition t o requiring no more than 2 h esterification time, small quantities of phenolic antioxidant often used with unsaturated polymers do not interfere. The mechanism of this selectivity was thoroughly discussed by Cohen and Fong ( 4 ) . Equivalent weight measurement of polymers with limited thermal stability may be adversely affected by the vigorous esterification conditions described above. Connors and Pandit ( 5 )demonstrated that quantitative esterification of alcohols occurs with acetic anhydride a t 45 "C in 10 min when catalyzed with N-methylimidazole (NMIM) instead of pyridine. Use of NMIM to catalyze the esterification of hydroxyl end groups on polymers could significantly reduce the possibility of polymer degradation and also reduce the analysis time. In the following work this method was modified slightly so that homogeneous polymer solutions would result during esterification and subsequent titration. EXPERIMENTAL Apparatus. A steam bath (Precision Scientific Co.) was used t o heat the reaction vessels, and titration end points of highly colored PA/PY polymer solutions were determined with an Orion Model 801 digital pH meter equipped with a Beckman combination electrode (39030). Phthalic anhydride esterification was accomplished in 500-mL aerosol compatibility bottles (FisherPorter Co.) equipped with polytetrafluoroethylene closure seals. Iodine flasks, 500-mL capacity, were used for all other esterifications. Materials. All solvents and chemicals were AR grade except where noted. N-Methylimidazole (Aldrich Chemical Co.) was used directly. Benzoic acid and potassium acid phthalate used for standardization of the methanolic KOH and aqueous NaOH, respectively, were primary standard grade. Reagents. Acetic anhydride was diluted 1:6 by volume with either pyridine or 1,2-dichloroethane. Phthalic anhydride (112 g) was dissolved in 800 mL of pyridine. Standard Solutions. Aqueous 0.5 N NaOH was standardized with potassium acid phthalate. Methanolic 0.5 N KOH for the
Table I. Comparison of Equivalent Weight Values of Selected Polymers method/polymer HTPB, HTPB: PEG PECH DECH AAIPY .Y
1270 16.5
S
1233 2 3 4 4 11.0
-16.5
1235 11.6
113b
3.2
A/ NMIM -
s
1403 1407
S
11.9
15.0
23-17 1253 1 1 9 6 18.5 19.5 2 . 6
PA/PY -
X
1373 2355
1565 1 4 0 7
S
8.5
20.6
11.5
1267 3.5
-
heat -
X S
8.0
AXiNMIM
AA/NMIM method was standardized by dissolving the benzoic acid in 1 mL H20 and 7 mL methyl alcohol followed by addition of 1 mL NMIM and 50 mL chloroform. The titration end point was detected with thymol blue. This latter standardization method was used because it duplicates the sample solutions and results in normality values approximately 0.001 lower than those from aqueous potassium acid phthalate. Analytical Procedures. The AA/PY and PAIPY methods are described elsewhere ( I , 3). The following AA/NMIM method was developed because it could be used without alteration on all of the polymers tested. Three- to four-milliequivalentsamples of polymer were weighed into separate flasks containing 20 mL of 1,2-dichloroethane. Exactly 4.00 mL of acetic anhydride/l,2-dichloroethanereagent and 4 mL of NMIM were added to the mixture and the polymer was dissolved. Each flask including reagent blanks was briefly purged with gaseous nitrogen and placed on the steam bath for 15 min. The excess acetic anhydride was hydrolyzed with 3 mL H20 and the flasks were heated an additional 5 min. After cooling, 200 mL chloroform and 35 mL methanol were added and the solutions were titrated with the standard 0.5 N KOH to the th>mol blue end point. RESULTS AND DISCUSSION Four hydroxyl terminated polymers were selected for comparison; two batches of hydroxyl terminated polybutadiene (HTPB, Atlantic Richfield Co.), polyethylene glycol 4000 (PEG, Dow Chemical Co.), polyepichlorohydrin (PECH,
This article not subject to U.S. Copyright. Published 1980 by the American Chemical Society
Anal. Chem. 1980, 52, 573-575
Hercules Chemical Co.), and a derivatized epichlorohydrin polymer (DECH, North American Rockwell). izll of the polymers were used direct15 and none contained significant levels of antioxidant. Table I shows the equivalent weight values obtained using the three analysis methods described. Each value is the mean of a t least three replicate determinations and differences were considered statistically significant a t the 9 5 7 ~confidence level. Analysis of the Table I data indicates that for HTPB2 the AA/PY vs. AA/NMIM and the AA/PY vs. PA/PY values are different whereas the P A / P Y vs. AA/NMIM values are not different. Samples of HTPBLunder gaseous nitrogen were heated on the steam bath for 4 h after which the equiLalent weight was determined using the AA/NMIM method. The Table I data indicate that the 4-h esterification period required for the AA/PY method causes significant degradation of this polymer. No significant bias is apparent for a linear saturated polymer such as PEG; however. the PA/PY data for PECH and DECH are obviously different compared to the other methods. Both of the latter polymers are initiated from a trifunctional alcohol (e.g., glycerol) and the chain length of a specific branch is not controllable. Cohen and Fong ( 4 ) concluded that, in addition to the steric nature of the anhydride, variation in equivalent weight values between methods may also be due to steric or electronic properties of the hydroxyl compound. T h e presence of the chlorine atom on PECH or a similar group adjacent to the hydroxyl group may alter the properties of the polymer sufficiently to cause the observed differences. N-Methylimidazole was also used t o
573
catalyze the esterification of HTPB, with phthalic anhydride but no significant reaction occurred. Although the interference of phenols with equivalent weight determinations using acetic anhydride is recognized, the AA/NMIM method described herein is more generally applicable to different hydroxyl terminated polymers than either the AA/PY or PA/PY methods which were compared.
LITERATURE CITED Joint Army, Navy, NASA, Air Force (JANNAF) Propellant Characterization Working Group Method No. 528.1 (June 1978), Chemical Propulsion Information Aqency, The Johns Hopkins University Applied Physics Laboratory, S i i e r Spring, Md. R. D. Law, private communication, Thiokol Chemical Co., Wasatch Division, Brigham City, Utah Joint Army, Navy, NASA, Air Force (JANNAF) Propellant Characterization Working Group Method No. 529.1 [June 1978), Chemical Propulsion Information Agency, The Johns Hopkins University Applied Physics Laboratory, Silver Spring. Md. J L. Cohen and G.P. Fong, Anal. Chem., 47, 313-16 (1975). K . A. Connors and N. K. Pandit, Anal. Chem., 50, 1542-5 (1978).
Louis A. Dee* Bridget L. Biggers Mary E. Fiske Air Force Rocket Propulsion Laboratory Chemical Branch LKLR Edwards. California 93523
RECEIVED for review October 15, 1979. Accepted November 30, 1979. This work was supported by the United States Air Force under Air Force Rocket Propulsion Laboratory Project Number 2303MlRE.
Circular Dichroism Spectra of Opium Alkaloids in the Solid State Sir: We wish to demonstrate that analytical distinction among compounds of similar molecular structure is possible using a method which is not usually considered capable of such a level of individuation. The experimental procedure itself is not new. The novelty was to take two well established and conceptually unconnected procedures and to combine them into what appears to be a definitive method of identification, albeit limited a t this time tu only seven members of the family of opium alkaloids in their pure states. T h e method applies the technique of circular dichroism (CD) ultraviolet spectropolarimetry to the study of compounds in the solid state. More specifically, the compounds are pressed in KBr pellets formed in the usual way. This use of a KBr matrix in CD studies is not commonly encountered ( I , 2 ) , although other solid state media have been investigated ( 3 ) using either CD or ORD. None have been used for analytical purposes of the kind described here. Other innovative experimental uses of polarized spectroscopy include magnetocircular dichroism ( 4 )(MCD) and liquid crystal CD ( 5 ) . Our interest in the analytical distinction of the opiates by the latter method led us directly into an investigation of the spectra of the compounds pressed into dry KBr pellets. This correspondence contains our preliminary results, which hold considerable promise for qualitative, and perhaps with development of the technique, even quantitative analyses. EXPERIMENTAL All samples and the analytical grade KBr were dried at 140 "C for at least 48 h prior to their use in preparing specimens. The opiates studied were morphine, morphine sulfate, codeine, 30003-2700/80/0352-0573501 OO/O
acetylmorphine, 6-acetylmorphine, 3,6-diacetylmorphine (heroin) hydrochloride, thebaine, and hydrocodone. Salts and free bases can be investigated with equal ease. Spectra were taken on a Cary 61 spectropolarimeter operating in the normal mode. Samples consisted of an approximate 1:100 dilution with 80 mg KBr. Specimens must be thin (ca. 0.2 mm) for adequate u\:transmission. The sample holder was constructed to f i t both the pellet press and the carriage in the sample compartment of the Cary 61 so that damage to the fragile specimen could be avoided. Tests for the existence of linear dichroism components to the signal were made by rotating the specimen 90" and 120' and repeating the spectrum. Samples were dissolved in ethanol or water and the UV and isotropic solution CD spectra obtained for comparison.
RESULTS AND DISCUSSION The solid state CD spectra of the seven compounds are shown in Figures 1 and 2. There are two morphine spectra, one for the free base, the other for the sulfate. Similarities exist in the spectra of codeine, 3-acet.ylmorphine, and morphine sulfate, all of which show negative ellipticities a t the 'Lb absorption band around 286 nm, and positive ellipticities a t the 'La band around 248 nm. These are qualitatively consistent with the more familiar ethanol CD spectra of these compounds (6). Semiquantitatively, the ellipticities increase in the order 3-acetylmorphine, morphine sulfate, and codeine, respectively. Spectra of the other compounds are uniquely different. The morphine free base spectrum bears no resemblance even to that of morphine sulfate, which is perhaps the most startling result. Liberation of the free base brings about the spectral change anticipated from these data. This is an interesting distinction from the viewpoint of the forensic analyst in that d 1980 American Chemical Society
