Acid Anhydride - ACS Publications - American Chemical Society

Department of Pharmacy, Punjab Univei situ, Chandigarh-3,. India. Received ...... Ida-Hydroxymethyl-3~-hydroxypree;n-5-en-2&one. (VII). -To a solution...
1 downloads 0 Views 9MB Size




amine in methanol were oxidized by 30% aqueous hydrogen peroxide solution a t room temperature. The excess peroxide was destroyed by the addition of platinum black, the mixture was filtered and concentrated to a sirup. The N-oxide was decomposed in the apparatus described above a t 10-15 mm. pressure. The products were isolated in the manner described for the pyrolysis of the quaternary hydroxides. In

two consecutive experimente methylenecyclopentane was obtained as the sole product in 43% and 56% yields, respectively. In one experiment one equivalent of potassium hydroxide in methanol was added to the N-oxide solution which was then concentrated and pyrolyzed in the usual manner. Methylenecyclopentane and 1-methylcyclopentene were formed in 54% and 9% yields, respectively.

Notes 4,5-Diethylphenanthrene-9,lO-dicarboxylic

Acid Anhydride D.


9,lO-dicarboxylic acid anhydride (IV). Desulfurization of IV in aqueous sodium hydroxide solution using Raney nickels proceeded smoothly to yield 4,5-diethylphenanthrene-9,l0-dicarboxylic acid anhydride (V) in 34% yield.

Department of Chemistry, West Virginia Unicersity, Morgantoum, West Virginia Received February 93, 1966

Although phenanthrene derivatives with substituents in the 4- and 5-positions have been relatively inaccessible until recently, the literature records several examples of the synthesis of 4,5-dimethylphenanthrenel and some of its derivatives2 and other related molecules with substituents in hindered position^.^ The successful synthesis of naphthalenes and phenanthrenes containing substituent ethyl groups by desulfurization of the corresponding naphthoand phenanthro [b]thiophenes4 suggested a parallel route to phenanthrene derivatives containing ethyl groups in the 4- and 5-positions. The synthesis of 4,5-diethylphenanthrene-9,10-dicarboxylicacid anhydride has been accomplished by synthesizing a phenanthrene derivative containing two fused thiophene rings which were cleaved during the final step of the synthesis by desulfurization with Raney nickel. Pinacolic reduction6 of 4-oxo-4,5,6,7-tetrahydrobenzo [blthiophene (I)afforded 6,7,6’,7’-tetrahydro4,4’-bibenzo [blthiophene (11) in 67% yield. Addition of maleic anhydride to the diene I1 in refluxing xylene solution gave a 69%-yield of the adduct, 1,2,7,8,9,10,11,12- octahydro[3,4 - b:6,5-b’]dithienophenanthrene - 9,lO - dicarboxylic acid anhydride (111). Aromatization of I11 was accomplished using :6,54’]dithienophenanthrenesulfur to yield [3,4-b (1) M. 9. Newman and H. S. Whitehouse, 3. Am. Chcm. SOC.,‘71, 3664 (1049).G.M. Badger, J. E. Campbell, J. W. Cook, R. A. Raphael, and A. I. Scott, J . Chum. Sac., 2326 (1950),G. Wittig and H. Zimmerman, Ber., 86, 629 (1953). (2) M. 9. Newman and A. S. Hussey, 3 . Am. Chem. Soc., 69, 3023 (1947). (3) M.S. Newman, W. C. &,gar, and M. V. George, (bid., SI, 2376 (1960). (4) E.J. Modest and J. S~musekovicr,ibid., 78, 577 (1950). (5) M.8. Newman, J . Or& Chem.. 86, 588 (1961).



Ld V

The assignment of the above structure to compound V is in accord with its ultraviolet spectrum which is similar to that of phenanthrene. The alternative formation of the corresponding diethyl dihydropyrene or diethylpyrene derivative is excluded on the basis of elemental analysis of compound V and the dissimilarity of its ultraviolet spectrum to those of pyrene derivatives of the type mentioned. Experimental’

6,7,6’,7’-Tetrahydro-4,4‘-bibenzo[b]thiophene(II).-To a mixture of absolute ethanol (75 ml.) and dry benzene (50 (6) E. Campaigne and W. E. Krieghbaum. ibid., 26, 359 (1961). (7) Melting Bointa are uncorrected. Microanalyses were performed by Galbraith Laboratoriea, Knoxville, Tennesaee. Infrared spectra were determined a i listed using a Perkin Elmer-Model 21 double beam recording spectrophotometer. Ultraviolet spectra were obtained wing a Beokman DU speotrophotometer using 96% ethanol sa iolvent.

JULY, 1962


263 1

ml.) containing freshly sandpapered aluminum weighing Anal. Calcd. for CnoHls03:C, 78.93; H, 5.30. Found: pans (Fisher Scientific Company) cut into half-inch squares (3 g.) and mercuric chloride (0.12 g.) was added 10 g. of the C, 78.82; H, 5.35. ketone 1.8 The mixture was heated under reflux on a steam bath causing immediate reaction. After 3 hr. the mixture was almost solid with a gray material. An additional portion Structure of N- Acyl-s-triazoles' of absolute ethanol (50 ml.) was added and the mixture was heated under reflux for 5 more hours. The mixture was then poured into ice and dilute hydrochloric acid and the orK. T. POTTS AND T. H. CRAWFORD ganic material was thoroughly extracted with benzene. Removal of the benzene left a solid whirh was recrystallized Department of Chemistry, College of Arts and Sciences, from 95% ethanol as light tan colored needles, 6.0 g. (67%), University of Louisville, Louisville, K y . m.p. 136.5-138'. An analytical sample, recrystallized from the same solvent melted a t 136.5-138'. Received February id, 196s Anal. Calcd. for ClaH1,S2: C, 71.06; H, 5.22. Found: C, 70.83; H, 5.33. 1,2,7,8,9,10,11,12-0ctahydro[3,4-b:6,5-b'] dithienophenThe recent application of N,N'-carbonyldi-s-tnaanthrene-9,lO-dicarboxylicAcid Anhydride (III).--When a zole to peptide synthesis2 makes our results that mixture of I1 (5.3 g., 0.019 mole) and maleic anhydride (3.0 g., 0.031 mole) in technical xylene (100 ml.) waa heated establish the structure of N-acyl-s-triazoles of conunder reflux for 4 hr. and then allowed to cool over a 12-hr. siderable theoretical interest. period, a white solid separated, 4.9 g. (69%), m.p. 255Acylation of s-triazole with an acid anhydride or 257'. Recrystallization from xylene gave an analytical acyl chloride8 readily gives the corresponding acyl sample, as white feathery needles, m.p. 259.5260'. An infrared spectrum of the adduct in Nujol mull showed strong derivative which is extremely sensitive to moisture, bands a t 1780 em.-' and 1855 cm.-1 characteristic of the being hydrolyzed to s-triazole and the acid (or in the case of carbonyldi-s-triazole, carbon dioxide). The anhydride grouping. Anal. Calcd. for C*OH160&: C, 65.19; H, 4.38; S, lability of this acyl group can be attributed to the 17.41; mol. wt. 368. Found: C, 65.15; H, 4.44; S, 17.28; aromatic nature of the s-triazole nucleus4 and has mol. wt. (Rast) 358, 365. [3,4-b:6,5-btJ Dithienophenanthrene-9,lO-dicarboxylic precluded the usual chemical methods for deterAcid Anhydride (IV).-Dehydrogenation of the adduct IV mining the position of substitution of the acyl was accomplished by heating a mixture of IV (0.5 g., 0.0014 group. Thus, lithium aluminum hydride reduction mole) with sulfur (0.18 g., 0.0056 mole) in a nitrogen atmos- of N-acetyltriazole gave s-triazole and acetaldephere under a pressure of slightly less than 1 atm. in a salt bath at a bath temperature of 300' for 45 min. The crude hyde.s As alkylation occurs predominantly a t product was crystallized from acetic anhydride and melted position 1 of the nucleus,sal6 it is to be expected that at 29C-303'. This material was found to be satisfactory acylation would occur a t the same position. for use in the subsequent step; no percentage yields are reExperimental verification of substitution occurported as purification was not carried out except to prepare ring a t position 1has nom been found from the NMR an analytical sample by sublimatinn followed by recrystallization from acetic anhydride affording a sample of an orange spectra of the acyl derivatives and suitable model colored solid, m.p. 30&-307". The purified sample exhibited substances. A 1-substituted s-triazole (I) has no x strong green fluorescence in both acetic anhydride and plane of symmetry, whereas a 4-substituted s-triachloroform solutions. The infrared spectrum of IV in Nujol zole (11)has one as shown. Hence in I the deshieldmull absorbed strongly a t 1770 em.-' and 1830 cm.-I. The ultraviolet spectrum in 957, ethanol solution showed A, 264 mp (log E 4.55), Amsx 298 mp (log c 4.70). Anal. Calcd. for czo&osf&: C, 68.65; H, 2.24; S, 17.80. Found: C, 66.79; H, 2.37; 6, 17.90. I I 4,5-Diethylphenanthrene-9,lO-dicarboxylicAcid AnhyCO-R CO--K dride (V).-To n solution of 400 ml. nf 10yc sodium h\-droxide solution was added 2.85 g. of unpurified IV from the I I1 dehydrogenation of 111. The mivture was boiled until a clear solution was obtained. The solution was allowed t o cool and Raney nickel alloy (25 9.) was added in 0.5-g. por- ing effect of the carbonyl function' of the acyl group tions over a period of 1 hr. The solution was stirred and would influence only one proton (that' at C-5), maintained a t a temperature just below the boiling point for causing it t'o shift to lower field, whereas in I1 it 1 hr. It was then filtered while hot. Acidification of the would have an equal effect on both the protons a t cooled filtrate with dilute hydrochloric acid followed by extraction a i t h chloroform afforded a dark red oil on removal C-3 and C-5. of the chloroform. Crystallization of the oil from glacial (1) Part V in this series. Support of this investigation by PHS aretic arid gave vellow crystals, 0.80 g. (3475), m.p. 16,i- research grant CY-5973 from the National Cancer Institute, Publio 172'. An analvticnl sample was prepared by chroma- Health Service, is gratefully acknowledged. tography of a solution of the yellow anhydride in petroleum (2) H. C. Beyerman and W. Maassen van Den Brink, Rec. traa. ether solution (b.p. 65-67') over alumina followed by re- chim., 80, 1372 (1961); H.A. Staab, Annalan, 609,76 (1967). (3) (a) M. R. Atkinson and J. B. Polys, J . Chem. Soc.. 141 (1954); crystallization from glacial acetic acid and melted at 170(b) H. A. Staab, Cham. Ber.. 89, 1927 (19Ij6). lil'. The infrared Rpectrum of a chloroform-solution of V (4) K. T. Potts. Chem. Rev., 61, 87 (1961). showed bands at 1770 cm.+ and 1845 em.-'. The ultra(5) K. T. Potts. unpublished results. violet spectrum in 95% ethanol showed Amax 263 mp (log c (6) C. Ainswortb and R. G. Jones, J. A m . Chem. Soe., 11, 621 4.52), ,A, 320 mp (log E 4.02). (1955). (8) P. Cagniant and P. Cagniant, Bull. aoc. chim. Francs. 62 (1963).

(7) L. M. Jackman, "Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry," Pergamon Press, N. Y.. 1959. pp. 121-125.





values for ring protons a t position 3 5

s-Triazole 4-Amino-s-triazole N - Acetyl-s-triazole h',N'-Carbonyldi-s-triazole 1,3-Diphenyl-s-triazole 1,5-Diphenyl-s-triazole One intense peak only observed. intensity observed.

1.82 1.73 1.97 1.82

1.W 1.73' 1.05b 0. 63b 1.65

2.08 Two peaks of equal

Application of this symmetry principle to the striazole derivatives listed in the table has enabled the assignments shown to be made and establishes that these N-acyl-s-triazoles are best represented as 1-acetyl-s-triazole (I, R = CHa) and 1,l'-carbonyldi-s-triazole (I, R = 1-s-triazolyl). Removal of the carbonyl group of (I, R = l-striazolyl) by hydrolysis (addition of a small amount of water to the sample tube) resulted in the disappearance of the low-field peak and an increase in intensity of the peak at 7 value 1.82. This corresponded to the formation of s-triazole. The low r value for the C-5 proton in (I, R = 1-s-triazolyl) is attributed to a double deshielding effect: (i) that exerted by the carbonyl group and (ii) that exerted by the second aromatic s-triazole ring which is able to orientate itself with respect to the C-5 proton so that it is in the position of maximum deshielding.* Experimental The compounds used in this investigation were prepared by methods described in the literaturesg The spectra were recorded from a Varian V-4302 Dual Purpose, 60 Me., NMR spectrometer, equipped with a flux stabilizer and field homogeneity controls. The calibrations were made by the side-band technique10 and chemical shift values are reported in T units." Samples were examined as saturated solutions in anhydrous methyl cyanide solvent or a t concentrations not exceeding lo%, depending on solubility. 1,3-Diphenyl-s-triazole and s-triazole were examined in carbon tetrachloride solvent and preliminary studies indicated that there were only minor solvent effects by carbon tetrachloride and methyl cyanide on the 7 values. (8) Ref. 7, pp. 125-129. (9) See ref. 4 and ref. 2. (10) J. T. Arnold and M. G. Packard, J. Chem. Phzls., 19, 1608 (1951). (11) G. V. D. Tiers, J. Phys. Chem., 68, 1151 (1958).


School of Chemistry of the University of Minnesota, Minneapolis 14, Minnesota

Received Dsceinber 11, 1961

Numerous cases have been cited in which cyclopropanes result when pyrazolines are decomposed


thermally in the presence of platinum. Simple pyrazolines often produce olefinic products, while more complex pyrazolines generally give cycloprop a n e ~ . The ~ availability5 of a number of sulfonylpyrazolines has prompted us to investigate this reaction as a possible source of cyclopropyl sulfones. The sulfonylpyrazolines were decomposed thermally (CQ. 140') in the presence of polished platinum, and in an atmosphere of nitrogen. The results are summarized in Table I. TABLE I Pyrazoline




CH-CeHs 1


Product (yield) thermal decomposition

4-Phenylpyrazole (58%)

I11 C4HsS0~-CH-CH-C~H~ I 1 HZC\~+S IV CaHSS02-CH-CH-CH3 I


C "H

3-Phenylpyraaole (63%)

4Methylpyrazole (64% as the picrate)


It can be seen that the principal product from each reaction was the pyrazole, formed by elimination of alkyl- or arylsulfinic acid.6 Furthermore, there was no signjficant difference in behavior of the A'-pyrazolines (I1 and 111) and the isomeric 42-pyrazolines (I and IV). I n no case was any cyclopropane noted; in this respect the sulfonyl pyrazolines behave in a manner similar to the nitropyrazolines which lose nitrous acid to give pyrazoles.' Rinehart and Van Auken* recent'ly reported that the light-induced decomposit'ion of certain stereoisomeric pyrazolines leads to the formation of cyclopropanes in which the stereochemist'ry of the (1) Supported in part by the Office of Ordnance Research, U.S. Brmy, Contract No. DA-ORD-31-124-61-013, and by the Smith Kline and French Reaearch Foundation. (2) In part from the P1i.D. thesis of H . G. Braxton, Jr., the University of Minnesota. (3) (a) I n part from the Ph.D. thesis of D. R. Theissen, the University of Minnesota, 1961; (b) Sinclair Research Fellow, 1959-61. (4) Cf. K. von Auwers and F. Kbnig. Ann.,496, 252 (1932). ( 5 ) W. E. Parham, F. D. Blake, and D. R. Theissen, J . Ore. Chem., ~ 7 , 2 4 1 5(1962). (6) The same pyrazoles result in higher yield when the pyrazolinea are treated with methanolic potassium hydroxide: cf. ref. 5. (71 (a) W. E. Parham and .J. 1., Bleaadale, .I. Am. Cbem. S o c . , 78, 4664 (1951): (b) I n oert4in eases pyrazolines derived from nitroolefins and diphenyldiazomethane give nitrocyclopropanes when decomposed thermally; cf. w. E . Parham, H. G. Braxton, J r . , and C . Serres, Jr., J. Org. Chem., 26, 1831 (1961). ( 8 ) K. L. Binehart, .Jr., and T. V. Van Auken, .I. Am. Chem. Yoc.. 82, ,5251 ( 1 9 6 0 ) .

JULY, 1962



(1.0 g., m.p. 151-155') was recrystallized from ethanol. The white solid (m.p. 22%230', 0.95 g., 81.9% yield) was identified5 (melting point and mixture melting point) as 4-phenylpyraeole. Additional 4-phenylpyrazole (0.12 g.) was obtained from the ether wash; the total yield was (1.07 g., 92.3%). I n the case of 3-phenyl-4-benzenesulfonylpyrazole(11) no gas was evolved. The product (m.p. 151-153') was water soluble and assumed to be a salt of 3-phenylpyrazole and benzenesulfinic acid. The action of aqueous sodium hydroxide on this product generated 3-phenylpyrazole (90% yield). Irradiation of Pyrazolines. 1. 3-Phenyl-4-benzenesulfonylpyrazolene (II).-A sample of 1 1 6 (0.20 g., 0.7 mmole) wm irradiated with a 125-watt ultraviolet lamp at 30-40' for 1 day. The reaction mixture was chromatographed on alumina using petroleum ether (b.p. 30-60') and mixtures of this ether and benzene as eluent. The solids obtained were combined and recrystallized from benzene-petroleum ether (b.p. 30-60'). The white powder (0.07 g., 70% yield, m.p. 72-73') was 3-phenylpyrazole (m.p. and mixture m.p.672-73'). 2 . 4-Phenyl-5-benzenesulfonylpyra~oline (I) .-The solid R-C,H,C H- SOdR' obtained when 1 6 (0.98 g., 3.4 mmoles) was treated as described in (l), above, was recrystallized from ethanol. The white solid (0.28 g., 58% yield, m.p. 227-229') was + RCH=CHSOJt'+ 4-phenylpyrazole.6 VIa.R=H; R'=CeHB The ethanol filtrate was concentrated to a white solid VIIa-c Nz This material melted at 138(0.27 g., m.p. 134-136'). b. R=CsHe, 1lrC4He 139' after recrystallization from ethanol, and was not idenV C. R=R'=CsH6 tified. The infrared spectrum was similar to starting material, but showed no absorption a t 3340 cm.-l (N-HY (a) phenyl vinyl sulfone (VIa) gave VIIa in 98% and a new band at 1600 cm.-l, characteristic of the N-N yield; (b) cis-n-butyl 0-styryl sulfone (VIb) gave group.6 VIIb (70.2'%)and bis(bipheny1ene)ethylene (19%); Anal. Found: C, 59.88; H, 4.62; N , 8.88; S, 10.77. and (e) phenyl 8-styryl sulfone (VIc) gave VIIc (ca. Calcd. for ClsH1402S (cyclopropane): C, 69.74; H, 5.46; 70%), arid a red impurity assumed t o be bis(bi- N, 0.00; S, 12.41: for CI5Hl4N202S(isomeric pyrazoline): H, 4.93; N, 9.78; S, 11.20. pheny1ene)ethylene. Mustafa and Harbash9 have C,3.62.91; 3-Phenyl-4-butylsulfonylpyrazoline(IV) .-The green reported that the reaction of 9-diazofluorene with oil obtained when IV (0.65 g., 2.44 mmoles) was irradiated nitroolefins of the type RCH=CHN02 gives 2- was chromatographed on alumina as described in (l), alkyl-3-nitrospiro-( cyclopropane-l,9'-fluorenes) di- above. 3-Phenylpyrazole (m.p. and mixture6 m.p. i 3 ' , rectly, and have suggested that these result by loss 0.15 g., 42.970 yield), and a pale green fluorescent liquid obtained. The infrared spectrum of the oil showed no of nitrogen from the intermediate nitropyraxoline were absorption near 1020 cm.-1, characteristic10 of the cycloVI11 (X = Nos). The pyraxolines VI11 have not, propane structure. Reactions of 9-Diazofluorene with Vinvl Sulfones. The Synthesis of Cyclopropyl Sulfones. 1. With cis-n-Butyl 8-Styryl Sulfone (VIb).--A solution of VIbS (1.30 g., 5.8 R-CH N mmoles) in anhydrous benzene (15 ml.) was combined with a solution of diazofluorenell (1.11 g., 5.8 mmoles, m.p. I I/ 98-100') in anhydrous benzene (10 ml,), and the resulting X-CH-N solution was heated a t the reflux temperature for 3 days. The benzene was removed and the red solid was dissolved in IX VI11 ethanol. The solution was cooled and fine red needles of however, been isolated. This fact, together with bis(bipheny1ene)ethylene" (0.18 g., 19%, m.p. 195-197') the quantities of bis-(bipheny1ene)ethylene formed, were separated. The filtrate was concentrated and cooled, and 2-phenyl-3-butylsulfonylspiro(cyclopropane-l,9'-fluosuggests that these reactions may involve combina rene) (VIIb, 1.58 g., 70.2% yield, white powder, m.p. tion of the carbene IX with the olefins. 161-164', m.p. 166' from ethanol) was obtained. Anal. Calcd. for C26H,,02S: C, 77.28; H, 6.23; 5, 8.25. Found: C, 77.01; H, 6.43; S, 8.09. Experimental The infrared spectrum of I1 showed characteristic10cycloThermal Degradation of Pyrazo1ines.-The general pro- propane absorption at 1020 cm.-l. 2. With Phenyl Vinyl Sulfone (Via).-A sample of VI8 cedure used is illustrated below. The results for other (2.52 g., 0.015 mole) was treated with diazofluorene (2.88 pyrazolines are summarized in Table I. Thermal Decomposition af 4-Phenyl-5-benzenesulfonyl- g., 0.015 mole) in anhydrous benzene (65 ml.) as described pyrazoline (I).-A mixture of I (2.30 g., 8.04 mmoles) and polished platinum (5.2 g.) was heated slowly to 140' in an (IO) I,. J. Bellamy, "The Infrared Gpeotrs, of Complex Mdeoules," atmosphere of nitrogen. Heating was continued until Methuen and Co., Ltd., London, 1956. (11) A. %ahonberg, W. Avad, and N Letlb. J Cksm Sac 1368 evolution of gasas w e d . The cooled solid (tan-orange) (1951). was triturated with ether and the resulting yellow powder

pyrazoline is maintained. In pursuit of this approach to cyclopropyl sulfones, three sulfonylpyrazolines were irradiated a t room temperature, using the 125-watt high pressure mercury vapor lamp. These reactions were not as clean as those conducted thermally, but in each case moderate to good yields of substituted pyraxoles resulted. In no case was a cyclopropane detected: (a) 3-phenyl4-benxenesulfonylpyrazoline (11) gave 3-phenylpyraxde (70% yield) : (b) 4phenyl-5-benaenesulfonylpyrazoline (I) gave 4-phenylpyraxoline (580j0). together with an almost equal weight of an unidentified nitrogen containing product (m.p. 138-139') ; and (c) 3-phenyl-Pbutylsulfonylpyrazoline (IV) gave 3-phenylpyrazole (43%). Acyclic vinyl sulfones were found to react readily with 9-diaxofluorene (V) to give sulfonylspiro (cyclopropane-1,9-fluorenes)(VII) :


8 ,

(1% (9) A. Muetafaand 4. Hnrhaah, J Am Clrem Soc , '76, 1.383 (1954).

Authentic sample supplied b y C. F. Koelsch, t h e Uni! ersity of





above. The solid product obtained directly from the benzene solution (pale yellow, 4.90 g., 98% yield, m.p. 139-140') was recrystallized from benzene to give %benzenesulfonylspiro(cyclopropane-l,9'-fluorene) (VIIa) aa a white powder (m.p. 140-141'). Anal. Calcd. for CZ1Hl102S: C, 75.88; H, 4.85; S, 9.65. Found: C, 75.70; H, 4.99; S, 9.56. The infrared spectrum showed absorption a t 1025 cm. -1, characteristic of the cyclopropane structure. 3. With Phenyl 8-Styryl Sulfone (VIc).-A solution of VIc (2.24 g., 0.01 mole) and diazofluorene (1.80 g., 0.0094 mole) in benzene (50 ml.) was heated at the reflux temperature for 3 days. The benzene solution was allowed to stand for 7 days at 32' and the resulting red solid (2.72 g., 72%, m.p. 190-220') was washed wit8hether and recrystallized from benzene. 2-Phenyl-3-benzenesulfonylspiro(cyclopropane-1 ,g-fluorene) (VIIc) waa obtained as a white solid, m.p. 238-240'. The benzene solution contained additional VIIc, but this was not processed. Anal. Calcd. for Cz~H200$3: C, 80.97; H, 5.03; S, 7.85. Found: C, 81.23; H,5.24; S,7.70.

to another based upon infrared absorption from a Nujol mull a t 2.84, 5.69, 5.75, 5.80, 5.93, and 6.12 p is supported by ultraviolet absorption in chloroform A,, 213 mp (log e 3.97); 255 mp (log E 3.87). Infrared absorption a t 2.84 p could represent the OH absorption of carboxyls while the four peaks, 5.69, 5.75,5.80, and 5.93 p are characteristic of carbonyls; the first three corresponding to unconjugnted and the last to conjugated carbonyl groups.* In addition, the peak at 6.12 p corresponds to the double and the ultrabond of the system -COCH=CHviolet spectrum agrees, with a maximum at 255 mp14a wave length too low for a benzenoid system while the absorption a t 213 mp compares favorably with that for a,@-unsaturated acids.6 Heating an ethanolic (10% hydrochloric acid) solution of the dimer causes rapid conversion to a more stable phenolic material, IV (R = H), which gives an intense ferric chloride reaction indicating a phenol nucleus. The infrared spectrum shows new Dimerization of 3-Carboxybenzoquinone-1,2 peaks a t 2.81 (hydroxyl), 5.57 (y-lactone), 6.23 and 6.67 p (benzene ring) and there is a shift in the ultraI,. R.MORGAN, JR. 270 mp (log e 4.01).6 violet absorption, Treatment of IV (R = H) with acetic anhydride Department of Pharmacology, Louisiana State University, School of Medicine, New Orleans, Louisiana and pyridine gives a monoacetate, IV (R = Ac), and with o-phenylenediamine a phenazine. ReReceived February 6,196.9 fluxing IV (R = H) with 10% sodium hydroxide in methanol results in V (R = H), which when reDuring the course of a program designed to study fluxed with acetic anhydride and pyridine gives a 1he chemistry of the intermediates in the enzymatic diacetate, V (R = Ac). The diacetate has a strong transformation of 3-hydroxyanthranilic acid (I) infrared absorption a t 8.04 p (-O.CO-), 6.03 p into quinolinic (11. R = R' = C02H), nicotinic (-CH:CH-), and 5.80 p (-COCO-). (11. R = H, R' = C02H), and picolinic acids'(I1. Pyrolysis at 200-210' causes I11 to undergo a smooth reverse Diels-Alder reaction with reduction t,o two molecules of 2,3-dihydroxybenzoic acid. Under similar conditions V (R = H) is converted to NH: N ,co*H VI. Oxidation of VI with silver oxide in anhydrous OH methanol gave a good yield of 3,8-dicarboxynaphI I1 thoquinone-1,2 (VII). With a-phenylenediamine in acetic acid VI1 gave a yellow phenazine. the abR = C02H, R' = H),l the possibility of 3-car- sorption spectrum of which was similar to those boxybenzoquinone-1,2 or a dimer being an interreported for henxo [a]phenazines.' mediate was investigated. While neither 3-carIn accordance with the stereochemical factors boxybenxoquinone-1.2 nor a dimer have been previously reported, some have postulated them as (3) These peaks are practically identical to those found in substituted and unsubstituted o-benzoquinone dimers [cf., J. Harleyproducts in the above degradations.2 Mason and A. H. Laird, J. Cham. Soc., 1718 (1958) and E. Adler, R. We now wish to report the successful oxidation of Msgnusson, B. Berggren. and H. Thomelius, Acto Chem. Scond., 14, 2,3-dihydroxybenzoic acid leading to 3-carboxy- 515 (196O)l. This value agree8 favorably with a value of 252 (log c 3.95) for benzoquinone-1,2 and the isolation of a dimer of the(4)anologous dimer of 4.5-dimethyl-o-quinone [cf., L. Homer and K. the quinone. Sturm, Ann,, 697, 1 (1955)l and a value 254 mM (log c 3.87) for the dimer of +benzoquinone [cf., A. A. Patchett and B. Witkop, J. OW. Treatment of 2,3-dihydroxyhenzoic acid with Cham., 2 2 , 1477 (195711. silver oxide in anhydrous methanol containing a (5) H. E. Ungnade and I. Ortega, J . Am. Chem. Soc., 7 8 , 1664 trace of 98% formic acid afforded 3-carboxybenzo- (1951). (6) The presence of the laotone in IV allows elimination of the other quinone-1,2, which dimerized to IT1 upon refluxing possible structure for 111, &e., in anhydrous ether for forty-eight hours. 0 An assignment of I11 as a Diels-Alder adduct formed by the addition of one molecule of quinone




(1) A. H. Mehler, Proc. Intern. Conpr. Biochsm. 4th. Vienna. 18, 184 (1958). (2) A. Miyake, A. € I . Bokman. and B. 8. Schweigert. J . Bio2. Cham., 211, 391 (1954).

COzH COaH (7) H. J. Teuber and N. Gotr, Chsm. Bar., 87, 1236 (1054).

JULY, 1962



A d . Calcd. for CI4H8OI:C, 55.27; H, 2.64; 0,42.07; mol.&., 304. Found: C, 55.12; H, 2.51; 0,42.37; mol. O RO * wt., 302.1° Pyrolysis of 111.-In a small distillation apparatus, 3 g. of 0W C 0 2 H -co,H I11 waa heated for 30 m'n. a t 200-220". The dark residue VI V after cooling to room temperature waa recrystallized from HOzC 0 ethanol affording colorless crystals of 2,3-dihydroxybenzoic acid, m.p. and mixture m.p. 201-202', 2.1 g. (68%). Rearrangement of I11 with Acid.-When I11 waa warmed VI1 with ethanolic hydrochloric acid (10%) for 15 min. on tt steam bath, and taken to dryness in vacuo, the rearranged phenol, IV, could be isolated. Recrystallization from methylene chloride afforded colorless microcrystals of IV, 270 mp (log e 4.01), X:".:"' 2.81,2.87. attendant upon the Diels-Alder reaction,8 the di- m.p. 232-234', 5.57,5.80,5.93,6.03,6.23, and 6.67 p. meric product (111) is assumed to occur through a Anal. Calcd. for CuHB07: C, 58.75; H , 2.11. Found: complex (VIII) rendering the endo product (IX). C, 58.79; H, 2.01. It i s interesting to note that only one isomer could be A phenazine derivative of IV waa prepared by heating IV in glacial acetic acid with an equivalent amount of sodium isolated. acetate and o-phenylenediamine dihydrochloride on a steam bath for 30 min. Cooling afforded the crude phenazine, 0.021 g., which waa crystallized from ethanol as brick red 0 needles, m.p. 310-312' dec. COzH U Anal. Calcd. for CmHloNzO~:C, 67.04; H, 2.81; N, 7.81. Found: C,67.21; H,2.61; N,7.59. Acetylation of IV (R = H).-The rearranged isomer, IV (R = H ) (0.1 g.), waa allowed to stand for 36 hr. with excess acetic anhydride and 6 drops of dry pyridine. The excess acetic anhydride and pyridine were removed in u a w and the \\ residue recrystallized from ethyl acetatepetroleum ether as HOzC' 0 yellow prisms of the monoacetate, IV ( R = Ac), m.p. 149HOzC 151', 0.08 g. (69.5%). IX VI11 Anal. Calcd. for ClsHsO~:C, 58.54; H, 2.45. Found: C, 58.41; H, 2.57. Hydrolysis of IV (R = H).-A solution of IV (0.3 g.) in Experimental0 50 ml. of 10% methanolic sodium hydroxide waa gently re3-Carboxybenzoquinone-l,2.-2,3-Dihydroxybensoic acid fluxed for 1.5 hr. Neutralization with 10% hydrochloric (3 9.) waa dissolved in 100 ml. of anhydrous methanol con- acid and extraction with ethyl acetate afforded a residue taining 1 ml. of 98% formic acid and added at 0' to a mix- which recrystallized from ether-cyclohexane, aa yellow ture of anhydrous sodium sulfate (10 g.) and freshly prepared microcrystals of V ( R = H), m.p. 181-182', 0.11 g. (31%). dry silver oxide (7 g.). The mixture waa vigorously Anal. Calcd. for C1,HsOs: C, 55.27; H, 2.64. Found: shaken for 15 min. and filtered. The red solution was C, 55.32; H, 2.71. allowed to stand at -10' for 2 hr. and the red needles of 3Acetylation of V (R = H).-V (0.1 g.) waa boiled with cafbozybenzopuinone-1,dthat formed collected and dried, 1.5 excess acetic anhydride and 7 drops of pyridine for 1 hr. g. (49%), m.p. 195-197" dec.; X2?OH 420 mp (log e 3.10). The excess anhydride waa removed in vacuo and the residue Anal. Calcd. for C&&04: C, 55.27; H, 2.65; 0, 42.07. recrystallized from ethyl acetatepetroleum ether m yellow Found: C, 55.20; H, 2.41; 0,41.89. prisms of the diacetate, V (R = Ac), 0.087 g. (68.5%). m.p. Refluxing the quinone (700 mg.) in 150 ml. of methanol 192.194'. with a solution o-phenylenediamine dihydrochloride (900 A d . Calcd. for C18HlnOlo:C, 55.67; H, 3.11. Found: mg.) and sodium acetate (1.3 g.) in 5 ml. of water for 30 C, 55.52; H, 3.23. min. afforded 900 mg. of 1-carboxyphenazine (recrystallized Pyrolysis of V (R = H).-In a small distillation apparatus from ethanol aa yellow needles), m.p. and mixture m.p. 3 g. of V was heated for 30 min. a t 195-200' and the result242-244'. ing dark mass dissolved in benzene. Chromatography over o-Quinone Dimer (1~).-3-Carboxybenzoquinone-l,2 silicic acid eluting with benzene afforded 1,2-dihydroxy-3,8(3.0 g.) waa refluxed in 1 1. of dry ether with anhydrous dicarboxy naphthalene (VI) as yellow microcrystals, m.p. magnesium sulfate for 8 hr. The magnesium sulfate and the 213-214', 0.9 g. (34.570). precipitated material were filtered off and extracted with Anal. Calcd. for C12H806:C, 58.07; H, 3.24. Found: about 500 ml. of hot benzene. The extract waa concentrated C, 58.31; H, 3.16. until turbid, cooled to O', and the flocculent precipitate reShaking VI with silver oxide and anhydrous sodium sulcrystallized from petroleum ether (b.p. 65-70') and benzene fate in methanol for 5 min. afforded 3,s-dicarboxynaphthotu yellow microcrystals of the dimer I11 0.13 g. (4.3 %), m.p. quinone-1,2 (VII), brick red needles, m.p. 273-275' dec., 216-218' dec., X":z 213 Q (log e 3.97), 255 mp (log e X%'OH 267 mp (log E 4.31), 395 mM (log E 3.56). 3.87). Xi?' 2.84, 5.69, 5.75, 5.80, 5.93, and 6.12 p. Anal. Calcd. for ClZH,Os: C, 58.54; H, 2.45; 0, 38.99. Found: C, 58.67; H, 2.31; 0,39.12. The benzo[a]phenazine of VI1 waa prepared by heating VI1 in acetic acid with an equivalent amount of o-phenylene(8) R. B. Woodward and T. J. Kata. Tetrahedron, 6, 70 (1959). diamine for 30 min. on a steam bath. CooIing t o room tem(9) Semimicro analyses by Alfred Bernhardt, Max Planck Institut perature resulted in reddish microcrystals of 1,G-dicarboxyMicroanalytisches Laboratorium, Mulheim (Ruhr), Germany. M e l t benzo[a]phenazine, m.p. 261-204', A$:"' 235 m Ir (log 8 ing points are uncorrected. Ultraviolet absorption spectra were ob4.50); 247 mp (log c 4.56); 285 mp (log e 4.36); 430 mp tained on a Unicam S.P. 600 spectrophotometer. Infrared spectra (log e 4.02). were determined with a Baird Associates infrared spectrophotometer. (10) Ebullioscopic determination in benzene by Huffman MicroAnal. Calcd. for C18Hlo04Ng: C, 67.92; H , 3.16; N, analytical Laboratories, Wheatridge, Colorado. 8.80. Found: C, 68.11; H, 3.09; N, 8.62. HOX







F l a v o n o i d s of Madhuca butyracea Nut- Shell

Y. C.AWASTHIA N D C. R. MITRA National Botanic Gardens Utilizatian Research Laboratory, Luclcnow, India Received January 4, 1 3 W

VOL. 27

tained from the drastic alkali degradation products of the tetramethyl ether. In another experiment, the methylated product of the flavanonol on alkaline hydrolysis yielded a cream-white crystalline neutral product, m.p. 119', which is being further investigated. Reduction of madhuca flavanonol with zinc and hydrochloric acid yielded a colorless crystalline compound, m.p. 268-270' dec., apparently identical with eriodyctiol. Racemization of the flavanonol yielded the optically inactive isomer, m.p. and mixed m.p. 241' dec. The characteristics of madhuca flavanonol, its derivatives, and degradation products pointedly show its similarity with dihydroquercetin and it is further substantiated by its iodine oxidation to quercetin in almost quantitative yield. The identity of the flavanonol with the optically active dihydroquercetin isolated from Douglas fir-wood (Pseudotsuga douglasii, Carr. ; Coniferae) was confirmed by mixed melting point and their comparable infrared and ultraviolet spectra, very kindly made available by Dr. J. C. Pew to whom our grateful thanks are due. The following table shows the comparative data of the two natural flavanonols : It is significant, to note that dihydroquercetin has until now been isolated only from the stem (wood and bark) of certain temperate zone plants; it is for the first time isolated in appreciable yield along with quercetin from a nut-shell.

The flavonoid (m.p. 222-224' dec.) isolated from Madhuca butyracea nut-shell is a mixture of quercetin and the flavanonol dihydroquercetin, the latter being 90% of the crystallizate. Paper chromatography (n-butyl alcohol-acetic acid-water, 4 : 1:5) also revealed only two constituents. The identity of quercetin was established by elemental analysis, coupled with comparison with an authentic sample in regard to its mixed melting point, ultraviolet spectra, and paper chromatography. The madhuca flavanonol is soluble in common organic solvents and water, and insoluble in benzene. The characteristics are noted in the accompanying table; Wilson's boric acid test' with it is positive. The flavanonol obtained on crystallization from dilute alcohol (30%) is not solvated, whereas the dihydroquercetin, taxifolin12was isolated with one or more molecules of water of hydration when crystallized from alcohol of various dilutions. The ultraviolet spectra of the flavanonol is substantially different from those of common flavones and flavonols, though the bathochromic Experimentale shift in alkaline solution is somewhat similar to Isolation.-Coarsely powdered M. butyracea nut-shell to that in case of polyhydroxy flavones; the peak (2 kg.) was percolated a t room temperature (20-35') with at 250-260 mp (cf., flavones3) is conspicuously 95% alcohol (5 X 3 1.). The semisolid dark brown extracabsent and to that extent the curve resembles tive ( c 300 g.) obtained after removal of solvent, under those of polyhydroxy i~oflavones.~The combina- reduced pressure and finally in vacuo, was successively tion peaks at 1620, 1591, 1456, and 1513 cm.-l extracted with petroleum ether (2 l.), ether (2 l.), and ethyl acetate (500 mi.). The ether extract on concentration in its infrared spectra show the absence of absorp- yielded a pale yellow crystallizate, m.p. 222-224' (c 40 g.) tion in the lower region, 1640-1680 em.-', re- responding to the color tests for flavonoids, and on repeated ported as characteristic of flavonols and flavanones. fractional crystallizations from dilute alcohol (30%) it When treated with dilute mineral acids the yielded quercetin, m.p. and mixed m.p. 310-312" dec., (4 9.) the flavanonol as dull white needles, m.p. 241' dec. flavanonol is slowly converted into quercetin. and (35 g.). Ultraviolet (quercetin) A$:* 257 mp (log c 4.32) and While methylation with methyl iodide did not 375 mG (loa E 4.30). Zits: 258 m,u (log . - e 4.32) and 375 mp yield any well defined product, on prolonged treat- (log E 4.34r Anal. quercetin; Calcd. for ClaHloO,: C, 59.7; H, 3.3. ment with dimethyl sulfate both penta- (m.p. Found: C. 59.9; H, 3.4. 149-150') and tetramethyl (m.p. 159-160') Anal. $abanonol; Calcd. for C15H1207; C, 59.2; H, 3.95. ethers of quercetin were obtained in comparatively Found: C, 59.1; H, 4.0. poor yields (cf., ref. 5). The flavanonol pentaFlavanonol pentaacetate was prepared by heating (8 hr., acetate, m.p. 130') [ c x ] ~ ~ ' D 60' (c 1.0, alcohol), steam bath) the substance (500 mg.) in acetic anhydride (1.0 ml.) with fused sodium acetate (5 g.). After the usual is different from quercetin pentaacetate in con- work-up and crystallization (alcohol-charcoal) the acetat? tradistinction to the formation of the methyl (150 mg.) was obtained as silky white needles, m.p. 130". ethers. The ferric chloride test with the acetate was negative and The alkaline hydrolysis of the flavanonol yielded the hydroxyl band was absent in its infrared spectra. 280 mp (log c 3.47) and 330 m p (log e protocatechuic acid, and veratric acid was ob- Ultraviolet:



(1) C. W.Wdson, J . A m . Chem. Soc, 6 1 , 2303 (1939). (2) J. C Pew, t b z d , TO, 3031 (1948). ( 3 ) L. H. Briggs and R. H Locker, J . Chem. Soc., 3136 (1951). (4) K Venkataraman (Review) Fort. Chem. O r p iVatW#tofe, Syrmgrr-\'erlag, 17, be (1868).

(5) W.E. Hills, A w l . J . 8 c i . Red., 879 (1952). ( 6 ) Melting points uncorrected: petroleum ether, b.p. 40-t30° ultraviolet data recorded b y Mr. B. P. ,Mittal; microanalyoed by M r J. saran, Central Drug Relearoh Inntitute, Luckaow.


JULY, 1962



TABLE I Madhuca flavanonol

M.p. Optical rotntion bltraviolet spectra: ~2'.mp k Z x .mp (0.001 N alkali) Infrared bands, p (KBr) Mg or Zn and hydrochloric acid Ferric chloride Caustic alkali

Dihydroquercetin from D o u g h 8 fir-wood

240-242' dec. f46" (acetone-water) + I 3 (absolute alc.)

241' dec.

4-36'' (alc.) 291 (log e 4.18) and 330 (loge 4.04) 213 (log E 4.40); 330 (log E 4.38) and 410 (log e 3.89) 2.98, 3.05, 6.17, 6.28, 6.61, 6.87, 7.33, 7.69, 7.93, 8.58, 8.81, 8.93, 9.24, 9.80, 10.28, 10.57, 11.65 Deep purple-red Violet-brown-blueblack Pink-yellow

Anal. Calcd. for C Z ~ H Z ~ OC, I Z ;58.3; H, 4.3. Found: C, 57.92; H, 4.46. Methyl Ethers.-Methylation w a carried ~ out by refluxing (36 hr. ; steam bath) the flavanonol(1 g.) in acetone (100 ml.) with dimethyl sulfate (1.0 ml.) and anhydrous potassium carbonate ( 5 g.). After work-up and crystallization (alcohol), methylated products were obtained, one as a white crystalline product (50 mg.), m.p. 149-15O0, and the other aa pale-yellow needles, m.p. 159-160" (300 mg.). The former is identical with pentamethylquercetin (m.p. and mixed m.p. 149-150') while the latter was found to be the tetramethyl ether of quercetin (Zit., m.p. 159-160' of 3,7,3',4'tetramethylquercetin). A ferric chloride test with both methyl ethers was negative but the tetramethyl ether showed the OH band (2.98 p ) in its infrared spectra. Anal. Calcd. for CI~HSOZ (OCH,),: C, 64.6; H, 5.4. Found: C, 64.9; H, 6.0. Anal. Calcd. for Cl&t603 (OCH&: C, 63.7; H, 5.3. Found: C, 63.67; H, 5.8. Conversion of the Flavanonol to Quercetin.-(a) The flavanonol (300 mg.) was heated (24 hr.; steambath) with dilute sulfuric acid (7%) and the reaction mixture on fractional crystallization with dilute alcohol (30%) yielded quercetin (24 mg.), m.p. and mixed m.p. 310-312' dec., while the major fraction of the flavanonol (250 mg.) was recovered unchanged. (b) The flavanonol (300 mg.) was heated (1 hr.; steam bath) with iodine (100 mg.) in glacial acetic acid ( 3 ml.) and fused sodium acetate (500 mg.). The yellow precipitate obtained on dilution and cooling was worked up and crystallized (glacial acetic acid) to give yellow needles of quercetin (270 mg.), m.p. and mixed m.p. 310-312' dec. Reduction of the Flavanonol with Zinc and Hydrochloric Acid.-The flavanonol (200 mg.) was treated in methanolic solution (10 ml.) with zinc powder (2 g.) and concd. hydrochloric acid (4 ml. added dropwise) with vigorous shaking (1.5 hr.; room temperature). The reaction mixture was kept overnight and the clear filtrate after separation of the unchanged zinc was diluted with ice-cold water (20 ml.) and the resultant white precipitate collected and dried. On repeated crystallizations (water) it yielded crystalline needles (50 mg.), m.p. 268-270' (eriodictyol, lit.,z m.p. 272"). Its characteristic color reactions are same as those reported for eriodictyol. Anal. Calcd. for ClaH1206: C, 62.5; H, 4.2. Found: C, 62.3; H, 4.5. Alkali Degradation.-(a) The flavanonol (300 mg.) was refluxed ( 1 hr.) with caustic potash (2.5 g. in 25 ml. water and 25 ml. alcohol) and worked up as usual to give a cream-white acidic product (30 mg.), m.p. 190" after purification through sublimation ( 150'/3 mm.). There was no depression in its melting point in admixture with authentic sample of protocatechuic acid. (b) The total methylated product of the flavanonol (m.p. 150-155'; 1 9.) was boiled (3 hr.) with a potassium hydroxide solution (aqueous, 5y0; 50 ml.) and the neutral precipitate obtained on cooling and dilution yielded after

290 and 330 (inflexion) 2.96, 3.07, 6.10, 6.60, 6.80, 6.90, 7.37, 7.83, 7.95, 8.60, 8.81, 8.96, 9.25, 9.81, 10.04, 10.30, 10.62, 11.70 Deep purple-red Emerald green-black Brown

crystallization (alcohol) rhombic plat,es (250 mg.), m.p. 119', CtTHlsOe. Anal. Calcd. for C1,Hla06: C, 64.1;, H,. 5.6. Found: C, 63.7; H, 5.76. The filtrate on acidification (hydrochloric acid, 5 ml.) vielded a solid residue ( 100 mp.) which on sublimation (130"/ 3 mm.) gave veratric acid, m.i.' and mixed m.p. with authentic sample, 179-180'. Anal. Calcd. for C~Hlo0,: C, 59.3; H, 5.49. Found: C, 59.5; H, 5.7.

Acknowledgment.-Authors' thanks are due to Professor K. N. Kaul, Director, National Botanic Gardens, Lucknow, for his keen interest in this investigation.

Crystalline Complexes of 1,3,5-Trinitrobenzene and Alkali Sulfites RONALD A. HENRY Organic Chemistry Branch, Chemistry Division, U S .Naval Ordnance Test Station, China Lake, California Received January SO, 1969

Muraourl reported that aqueous solutions of sodium sulfite ( 3 4 % ) would dissolve 1,3,5-trinitrobenzene (TNB) and lesser quantities of 2,4,6trinitrotoluene (TNT) through the formation of red-colored addition compounds but would not dissolve 2,4,6-trinitro-m-xylene. From these aqueous solutions the TNB or T N T could be recovered by acidification or in the case of T K T by dilution with water. Although Muraour stated that this complex of TNB and sodium sulfite would be studied further, no other information has appeared. Dark red crystalline compounds, in which the mole ratio of TNB to sulfite is 1:2, can be isolated if TNB is dissolved in warm aqueous solutions of sodium or potassium sulfite (20% by weight) and the resulting solutions quickly cooled to ambient temperature. If these solutions stand for twentyfour hours, the complex disappears completely and the TNB is destroyed (see below). When dry, these compounds are very stable; samples have (1) Ha Muraour, Bull, eoc. chim. Francr, 81, a67 (1924).



been stored a t room temperature for six years without change. By way of contrast, the addition complexes obtained from TNB and sodium alkoxides or potassium cyanide2-' are both unstable and explosive. These TNB-sulfite addition compounds show absorption maxima at 462 mp (peak) and 525 mp (shoulder) in either water or 40% methanol solutions which are lo-' M in complex and 0.01 to 0.09 M in sodium sulfite. The extinction coefficients are 17,300and 8700, respectively. In 40% acetone, 0.01 M sodium sulfite, an absorptioi maxima occurs at 510 mp (E 15,300); color fading in this system is also much more rapid than that in the corresponding methanol solution. In the absence of excess sulfite, a M solution of the complex in water has a weak absorption a t 462 mp (e 500) and a strong absorption between 230-240 mp; the latter absorption is attributed to the 1,3,5-trinitrobenzene which arises from the dissociation of the complex. The location of the absorption maxima and the magnitJudeof the molar extinction coefficients place the TNB-sodium sulfite compound in the Group I complexes as categorized by Miller and WynneJones.6 These complexes, which are the primary reaction product of aliphatic amines, pyridine, alkoxide ion, etc. Qith TNB, probably arise through an e!ectron transfer process to give an ionic structure (MeisenheimeP addition compounds) and undergo in solution slow secondary reactions which lead to aromatic nucleophilic substitution. The behavior of the TNB-sulfite complexes is also consistent with this latter point.


N*S0ion of unsaturation t o the 10,ll-position.

48.8 0.7

18.1 24.9 12.0

4.3 45.0 0.7

Experimental The techniques involved in exposure of the compounds t o tritium gas and subsequent purification, degradation, and assay were as described previously.lJ Gram samples of methyl stearolate and methyl linolenate were irradiated with 2.6 curies (1 ml. STP) of tritium gas. By means of the “mixed chromatograph” technique in which inactive known esters are added to unknown mixtures, the position of the major radiochemical component of methyl stearolate is shown to coincide in elution v d h that of inactive methyl oleate and the radiochemical shoutder u-ith inactive methyl stearate. Only a trace of radioactivity coincided with the position of inactive methyl stearolittc. C:as chromatographic and osidrttive c*leitvagt>reeults n r v given in Table I. The configuration of double bond in the 9,lO-octadecen02tte-HS waa studied in the following manner: .4 portion ( I mg. and 21 pc.) of the irradiated product was added t o 7 g. of a methyl oleate-methyl elaidate mixture. This solution was countercurrently distributed in a 200-tube automatic instrument between pentanehexane and 0.1 M silver nitrate in 90% methanol.* Recycling waa continued until, as shown by the monitoring curve of the recording refractometer, the head of the elaidate band nearly caught up to the tail of the oleate band. Upper layers were then withdrawn into the collector according to the single withdrawal principle. Radioactivity peaks were coincident i n position with those for methyl oleate and elaidate, and the integrated counts under the respective peaks indicated that the ratio of cis- to trans-g,lO-octadecenoate-R*was 2: 1 . This ratio is approximately the equilibrium attained by selenium or nitrous oxide isomerization of methyl oleate.

(4) C. R. Scholfield, E. P. Jones. and H. J. Dutton, Chem. Ind. (London), 1874 (1961).



Separation of the Diastereoisomers of Methyl 2,4-Dimethylheptauoate1 STEVEN R. TANNENBAUM~ AND EMILYL. WICK

Contribution No. 461 from the Department of Nutrition, Food Science and Technology,Massachusetts Institute of Technology, Cambedge, Mass. Received November 27, 1961

Although the resolution of diastereoisomers by gas chromatography has been reported, the small quantity thus separated has often prevented their isolation and the accumulation of direct evidence that separation had indeed occurred. During the course of a routine check of the purity of 2,4dimethylheptanoic acid by gas chromatographic analysis of its methyl ester, it was noted that two components were present in comparable quantity. Evidence is presented below that these components were diastereoisomers of methyl 2,4-dimethylheptanoate. Experimental6 Apparatus.-Gas chromatographic analyses were done on an instrument constructed in these laboratories employing a thermistor katharometer aa detector. Two columns were used: (1) 1 m., 4 mm. i.d., 10% (w./w.) Carbowax 4000 on 60-80 mesh diatomaceous earth'and (2) 2 m., 4 mm. i.d., 10% (w./w.) diethyleneglycol succinate (DEGS) on 60-65 mesh, acid-washed diatomaceous earth. Helium flow rate waa 50 ml./min. Column temperatures were 50 or 100". Infrared spectra were obtained by means of a Beckman IR-5 spectrophotometer equipped with a KBr beam condenser. Materials.-2-Methylpentanol-l (I) (Matheson, Coleman and Bell) waa dried and distilled twice before use, n% 1.4185, b.p. 93-95' (92 mm.). Gas chromatographic analysis of I, after its purification, showed one peak on the Carbowax column at 100'. Diethyl methylmalonate (111)(Matheson, Coleman and Bell) was distilled before use, 12% 1.4129, b.p. 194-196' (760 mm.). The purity of IT1 was studied by its saponifiration, decarboxylation, and esterification of the resulting free acids with diaeomethane.' Gas chromatographic analysis of the esters on the DEGS column at 50' showed the presence of more than 99% methyl propionate and slightly less than 1% of methyl acetate. It was concluded that I11 had contained more than 99% diethyl methylmalonate and slightly less than 1%diethyl malonate. Diethyl malonate (Eastman, White Label) was more than 99% pure as determined by gas chromatography on the (1) This work was supported by Contract No. AF 33(616)-6008 with the U.S. Air Force. Mention of names of firms or trade products does not imply endorsement or recommendation by the Department of the Air Force. The authors are indebted to Dr. William R. Moore, Dept. of Chemistry, M.I.T., for his intereat and advice. (2) General Foods Fellow, 1981-1962. (3) P. S.Frederick8 and J. M. Tedder, Proc. Chcm. SOC.,9 (1959). (4) M. C. Simmons, D . B. Richardson, and I. Dvoretzky, "Gas Chromatography 1960." R . P. W. Scott, ed., Butterworth Inc., Washington, D . C., 1960, p. 211. (5) J. R. Smith, ibid., p. 222. (6) We are indebted to Dr. 8. M. Nagy and his associates for the microanalyses. (7) Johns-hlanville Cbromosorb P. (8) R. Roper and T. 9. Ma, Microchem. J . , 1, 245 (1957).


DEGS column at 100'. A single, minor, unidentified peak was present. 2,4-Dimethylheptanoic Acid (IV).-I waa treated with phosphorus tribromide according to the procedure of Noller and Dinsmore.9 1-Bromo-2-methylpentane (11) waa obtained in 55.6% yield, b.p. 42-43' (15.5 mm.), nlOD 1.4480, (n% 1.4495).1° Gas chromatography of I1 on the Carbowax column at 100' showed one peak. Its infrared spectrum contained no OH bands. Anal. Calcd. for CsHlaBr: C, 43.65; HI 7.94; Br, 48.44. Found: C, 43.47; H, 7.82; Br, 48.73. Sodium t-butoxide (1.05 moles, 24.15 g. of sodium) in 800 ml. of t-butyl alcohol was treated with I11 (1.10 moles, 191.5 g,) followed by I1 (1.05 moles, 173.2 g.). The resulting mixture was worked up in the conventional manner1* to give the disubstituted malonic ester. The ester was saponified for 48 hr. with 50% potassium hydroxide. The free acid was decarboxylated by refluxing with constant boiling hydrochloric acid for 24 hr. The free acid IV waa obtained in 50% yield, b.p. 82-83' (0.6 mm.), n% 1.4302. Anal. Calcd. for CsHla02: C, 68.35; H, 11.39; neut. equiv., 158.0. Found: C, 68.34; H, 11.15; neut. equiv., 158.0 f 0.5. Esterifications of about 50 mg. of IV yielded methyl 2,4dimethylheptanoate (V). 4-Methylheptanoic Acid (VIII).l*-Sodium ethoxide (0.32 mole, 7.36 g. of sodium) in 150 ml. of absolute ethanol, diethyl malonate (0.34 mole, 54.5 g.), and l-bromo-2methylpentane (0.33 mole, 54.5 9.) reacted in the usual manner to give diethyl 2-methylpentylmalonate (VI), n% 1.4287, b.p. 110-112' (3.2 mm.), in 73% yield. Gas chromatography of VI on the Carbowax column a t 100' showed one peak. Saponification of 5 g. of VI yielded 2methylpentylmdonic acid (VII) from which was formed its bis-S-benzylthiouronium salt. Anal. Calcd. for C ~ S H ~ ~ N ~C, O ~57.66; S ~ : HI 6.98; N, 10.78. Found: C, 58.16; H, 7.17; N, 10.42. Decarboxylation of about 50 mg. of VI1 and esterification of the resulting 4-methylheptanoic acid (VIII) yielded methyl 4-methylheptanoate (IX). Gas chromatography of IX on the DEGS column a t 100' showed one peak. Separation of Diastereoisomers of V.-Chromatography of methyl 2,4dimethylheptanoate (V) on the DEGS" column a t 100' (flow rate 50 ml./min.) resulted in elution of two well separated peaks (A and B ) within 10 min. The separation factor, p (the ratio of the retention times of peak A and peak B), was 1-09. Resolution14 of the components was 1.07. The possibility that one of the components waa methyl 4-methylheptanoate (VIII), methyl propionate, or 2,4dimethylheptanoic arid was eliminated on the bwis of the retention volumes of reference samples at, the above conditioris. Peak A was collected aa it mas eluted from the DEGS column and rechromatographed under the same conditions. Only one peak was obtained. The infrared spectrum of peak A, as a pure liquid and in carbon tetrachloride solution, was essentially identical to the spectrum of methyl 2,4dimethylheptanoate (V). Normalization of the area under peaks A and B showed that peak A represented 53.2% of the total mixture. Epimerization. (a) Pure Peak A, collected from the DEGS column, was heated in a sealed tube at 110' for 18 hr. in (9) C. R . Noller snd R. Dinsmora. Org. Suntheaes, CoU. Vol. 11, 358 (1943). (10) C. E. Rahberg and H. R. Henle, J . Am. Chem. Soc., 68, 2785 (1941). (11) C. S. Marvel, Org. Sunthesea. Coll. Vol. 111, 485 (1955). (12) G. Weitzel and J. Wojohn, Z . phyeiol. Chem., 887, 65 (1951). (13) Although a number of liquid phases were investigated (Carbowas 4000, Carbowax 1500, Ucon I,B-1715, Silicone SE-30, and Silicone QF-l), only DEGS effected a separation of the diastereoisomers of methyl 2,4-dimethylheptanoate. (14) As recommended by D . Ambrose in "Gas Chromatography 1960," R. P. W. Scott, ed.. Butterworth Inc., Washington, D . C., 1960, p. 429.


JULY, 1962 the presence of 1 M sodium methoxide in anhydrous methmol. The reaction solution was neutralized with anhydrous methanolic hydrochloric acid, and the resulting sodium chloride removed by centrifugation. The supernatant liquid wm evaporated to a small volume and chromatographed directly on the DEGS column. The resulting chromatogram showed the presence of both peak A and peak B. The area of peak A represented 41.8% of the mixture. (b) Methyl 2,4dmethylheptanoate was heated under sodium methoxide in reflux for 21 hr. in the presence of anhydrous methanol. The infrared spectrum of the resulting ester, as a pure liquid or in ,carbon tetrachloride was essentially identical with the spectrum of untreated methyl 2,4dimethylheptanoate. Chromatography on the DEGS column showed the presence of peaks A and B. The area of peak A had decreased to 41.8% of the mixture. (c) An approximately 10% aqueous solution of sodium 2,4dimethylheptanoate was heated 18 hr. in a sealed tube at 115'. The free acid was isolated and esterified with diazomethane. Separation of the ester on the DEGS column that peak A had increased to 56.2% of the total mixture of components A and B.

Discussion The synthesis of 2,4-dimethylheptanoic acid Was accomplished in a straightforward manner. The fact that its ester Was separated gas chromatography into two almost equal components (A and B) suggested that diastereoisomers had been resolved. The possible presence of impurities such as methyl 4-methylhe~tanoate~ methyl Propionate, and free 2,4-dimethylheptanoic acid was eliminated by "parison of the retention volumes of cornponents A and B with those of the authentic reference compounds. A and B were shown to be diastereoisomers by the fact that the infrared spectrum of A was essentially identical to the infrared spectrum of 214-dimethylhePtanoate1 that A was epimerized to a mixture of A and B, that the relative quantities of A and B were changed by ePberiZation, and that the infrared spectrum of the epimerized mixture was the Same as that of the original methyl 2,4-dimethylheptanoate.

Addition of Selenium and Sulfur Tetrachlorides to Alkenes and Alkynes REEDF. RILEY,J. FLATO,' AND D. BENGELS D e p a r t m t of Chemistry, Polytechnic Instilute of Brooklyn, Brooklyn I, New York Received December 4, 1961

Preparation of organic selenides and their dihalides by addition of selenium halides to an unsaturated linkage is interesting in that it produces compounds having a reactive halogen on the beta carbon atom.216 This note describes further work (1) Ahatraoted in part from the B.S. thesis (1961) of J. Flato. (2) H, Brintzinger. I(. Pfannstiel, and H. Vogel, 2. an010. albsm. C h . ,366,75 (1948).


on addition to olefinic linkages and an extension of the reaction to acetylegc compounds. Although the reaction of selenium tetrachloride with unsaturated linkages to produce selenium dichlorides is quite general, our results show that a selenide is produced with acrylonitrile. Subsequent of this selenide yields the expected selenium dichloride which reverts to the selenide on recrystallization from acrylonitrile. As the latter is not chlorinated with particular ease: the presence of the nitrile g o u p beta to the selenium atom must labifiing influence on the selenium-chlorine linkage in this compound. Thus, selenides are to be expected whenever the selenium-chloride linkage is labilized and/or the unsaturated starting material is easily and rapidly chlorinated. lnan attempt to isolate the dichloride from the addition reaction, the stoichiometric amount of acrylonitrile was reacted with selenium tetrachloride in ether. This leads to a highly colored solution, typical of solutions of alkylselenium trihalides (an almost certain intermediate of the addition reaction), from which no solid product be obtained. hi^ suggests that excess alkene is needed to force the reaction towards completion. Lack of addition of selenium tetrachloride to trans-stilbene, even at temperatures substantially above those required for the other additions to ocis probably due to the well documented a-electron delocaEzation of the carbon double bond in this type of compound. Addition with diphenylacetylene, but subsequent cyclization takes place. This reaction will be reported in a separate placeinthe nearfuture. AS the nearly quantitative reaction between sulfur tetrachloride and cyclohexene occurs more rapidly and at temperatures substantially lower than reported for the analogous reaction using SUIfur dichloride,? the tetrachloride is a more active addition reagent than the dichloride. It is also a more active chlorinating agent, however, and it was found impossible to separate the sulfides from other reaction products without nearly total decomposition of the desired product in the other cases studied (see Experimental). All attempts to isolate bisorgano sulfur dichlorides of cyclohexene and hexene by separating the products from impurities a t Dry Ice temperature resulted in failure. Experimental Sulfur and Selenium Tetrahalidea.-The selenium compound waa prepared by condensing dried tank chlorine onto elementary repurified selenium in a 4ssk protected from atmospheric moisture and warming to room temperature while simultaneously passing through a slow stream of nitrogen. The sulfur compound was prepared in the same (3) H. L. Riley and N. A. C. Friend, J . Cham. Sac., 2342 (1932).

(4) H. Funk and W. Weka, J . prakt. Chem., 1,33 (1854). ( 5 ) H. Funk and W. Popenroth, {bad., 8, 256 (1959). (6) D. A. Evans and P. W. Robertson, J . Chem. Sac., 2834 (1950). (7) H. W. Moll, U. 9. Patent 2,472,755 (1949); Chsm. Ab8t?., IS, 6558 (1949).





TABLE I Substance

2-Chloropentyl selenoxide 2-Chlorohexyl selenoxide 2-Chloro-2-phenylethenyl selenoxide 2-Chloro-2-phenylethyl selenoxide 0 See ref. 12.



103-104' 103-104' 142-143' 91-91.5'

93% 94% 89% 82%

way except that excess chlorine was removed a t -78". Yields are essentially 1 0 o ~ o . Bis-2-chloropentylselenium Dichloride.-1-Pentene, 7 g. ( 1 mole), was added dropwise over B period of 15 min. to 22.1 g. (0.1 mole) of selenium tetrachloride cooled to -78'. The flask was gradually warmed to room temperature and the solvent stripped off leaving a crystalline solid. Recrystallization from pentene yielded a white solid melting at 74-75'. The recrystallized yield was 52%. A n d 8 Calcd. for CloH20C14Se:C, 33.64; H, 5.55; C1, 39.31; Se, 21.83. Found: C, 34.07; H, 5.72; C1, 39.18; Se, 21.85. Bis-2-chlorohexylselenium Dichloride.-1-Hexene, 84 g. (1 mole), was added dropwise over a period of 15 min. to 22.1 g. (0.1 mole) of selenium tetrachloride cooled to 78'. The flask was gradually warn--ed to room temperature and the solvent stripping off leaving an oil. The oil crystallized after standing overnight at -30' and was recrystallized from pentane, yielding a white solid melting at 45-46' (lit.S 42-43'). The recrystallized yield was 86%. Anal. Calcd. for C12HZ4C&Se: C, 37.01; H, 6.18; C1, 36.50; Se, 19.80. Found: C, 37.31; H, 6.32; C1, 36.04; Se, 19.90. Bis-2-chlorocyclohexylseleniumDichloride .-Cyclohexene, 82 g. (1.0 mole), wag added dropwise over a period of 15 min. to 22.1 g. (0.1 mole) of selenium tetrachloride cooled to -78'. The flask was gradually warmed to room temperature, the solid filtered off, and recrystallized from benzene. Recrystallization gave a white solid melting at 141-142' (lit.4 144') in 74% yield. Anal. Calcd. for ClzHzoC14Se:C, 37.52; H, 5.22; C1, 36.90; Se, 20.56. Found: C, 37.64; H, 5.39; C1, 36.40; Se, 20.57. Bis-2-chloro-2-phenylethylseleniumDichloride .-Styrene, 26 g. (0.25 mole), was dissolved in 100 ml. of ether and the solution added dropwise to 22.1 g. (0.1 mole) of selenium tetrachloride at room temperature yielding a dark red solution. Vacuum evaporation of solvent and excess styrene gave a solid which, after recrystallization from ether, melted at 100-lO1° (lit.691'). The recrystallized yield was 83%. Anal. Calcd. for Cl6H&1&: c , 44.70; H, 3.73; c1, 33.11; Se, 18.35. Found: C, 44.68; H, 3.96; C1, 33.,59; Se, 18.12. 2-Chloro-2-cyanoethyl Selenide.-Acrylonitrile, 80 g. (1.5 moles), was added dropwise over 15 min. to 22.1 g. (0.1 mole) of selenium tetrachloride a t room temperature. The color became dark red and considerable heat ww evolved, but on standing for 0.75 hr. the color changed again to light yellow. Evaporation of solvent and excess reactant yielded an oil which solidified on trituration with ether. Recrystallization from ether gave a white solid melting a t 94-94.5' in 77y0 yield. Infrared analysis showed a nitrile peak a t 2280 cm.-l. The reaction was repeated in ether using a 2: 1 olefin to halide ratio, but on evaporation of solvent only an intractable oil resulted. Anal. Calcd. for CBHaN&12Se: C, 29.72; H, 2.49; C1, 29.35; Se, 32.62. Found: C, 29.17; H, 2.65; Se, 32.70. Mol. Wt. Calc: 256. Found: 258.9 (8) All carbon, hydrogen, and chlorine analyses were done by the Schwarzkopf Microanalytical Laboratories, Woodside, New York. Selenium analyses were done gravimetrically after decomposing the compound with fuming nitric acid and precipitating the elementary selenium with sulfur dioxide. (9) Molecular weights were measured by the vapor pressure equilibrium method,


C, 39.20; H, C, 43.13; H, C, 51.99; H, C, 51.40; H,


6.54 7.19 3.24 4.29

Ca, 37.55; H, 6.54 Ca, 41.60; H, 7.43 C, 51.63;H,3.16 C, 51.43;H,4.53

Bis-2-chloro-2-cyanoethylseleniumDichloride.-Chlorine was bubbled through an ethereal solution of 10 g. of 2-chloro2-cyanoethyl selenide held a t room temperature until material ceased separating from solution. The filtrate was then chilled and the precipitate filtered and recrystallized from ether yielding a white solid in over 90% yield which melted at 71-72'. The compound is unstable and slowly yields a mixture of chlorine and hydrogen chloride a t room temperature. When dissolved in acrylonitrile, the dichloride loses its halogen to give an almost, quantitative yield of 2chloro-2-cyanoethyl selenide (ident. by m.p.). Anal. Calcd. for C&,N2C14Se: C1, 43.67; Se, 24.16. Found: C1,42.10; Se, 25.30. Bis-2-chloroethenylseleniumDichloride .-Acetylene, 7.8 g. (0.3 mole), was bubbled into a cooled ( -45') ether slurry of selenium tetrachloride made from 100 ml. of ether and 22.1 g. (0.1 mole) of selenium tetrachloride with slow stirring. The dark yellow solution was stirred 1 hr., warmed to room temperature, and vacuum evaporated. The resulting solid w&s recrystallized from ether yielding a white solid in 9% yield, melting a t 95-96' (lit.4 104-105'). A =C-H absorption was found in the infrared at 3100 cm-l. Anal. Calcd. for C4HICl$e: C, 17.69; H , 1.46; CI, 52.00; Se, 28.93. Found: C, 18.97; H, 1.58; C1, 52.13; Se, 29.60. Bis-2-chloro-2-phenylethenylselenium Dichloride .Phenylacetylene, 25.5 g. (0.25 mole), was dissolved in 100 ml. of ether and the solution added dropwise over 30 minutes t o 22.1 g (0.1 mole) of selenium tetrachloride a t room temperature. h the halide dissolved, the solution became dark red subsequently lightening to pale yellow on standing. Overnight chilling at -78' produced a solid which, when The yield of recrystallized from ether, melted at 89-90'. white solid was 73%. A=C-H absomtion was found in the infrared 3100 cm.-l. Anal. Calcd. for C18H12Cl4Se: C, 45.60; H, 2.82; C1, 33.34: Se, 18.50. Found: C, 46.39; H, 2.97; C1, 32.78; Se, 1S.10. 2-Chlorohexylsulfide.-l-Hexene, 18.5 g. (0.22 mole), dissolved in 100 ml. of ether was added dropwise to 17.4 g. Removal of the (0.1 mole) of sulfur tetrachloride a t -78'. solvent and excess reactant yielded a yellow solid which gave on recrystallization from petroleum ether a 12% yield of yellow solid melting at 52' (lit.10 46-50'). Anal. Calcd. for C12HZ4C12S: C1, 25.62; S, 11.81. Found: CL25.34; S, 11.68. 2-Chlorocyclohexyl Sulfide.-Cyclohexene, 16.4 g. (0.2 mole), dissolved in 100 ml. of ether was added dropwise to 17.4 g. (0.1 mole) of sulfur tetrachloride cooled to -78". A white solid immediately formed which was filtered off and recrystallized from ether. The recrystallized product melts at 72-73" (lit.' 72-73.5') and is formed in 84% yield. Anal. Calcd. for CpaH2&lnS: C1, 26.63; S, 11.98. Found: CL25.61,; 27.23; S, 12.00. Other Alkenes with Sulfur and Selenium Tetrachlorides. -Pentene, styrene, acrylonitrile, 1,2-dicyanoethylene, and phenylacetylene reacted (-78") in various mole ratios with sulfur tetrachloride in ether and pentane solvents. Reaction occurred a t the low temperature, but in each case evaporation of solvent left an oil which was intractable in our hands. Butadiene and selenium tetrachloride in ether behaved similarly. trans-Stilbene on the other hand would (IO) Ger. Patent 846,397 (1861); Chem. Ab&., 60, 7847 (1956).

JCLY, 1962



alkyl P-naphthyl ethers,? p-alko~ydiphenylamines,~ alkyl 2,4dinitrophenyl ~ulfides,~ alkyl 6-nitro-2mercaptobenzothiazolyl sulfides,lo and S-alkylmercaptosuccinic acids.“ In terms of suitability, each of these derivatives has certain limitations, except the alkyl 2,4-dinitrophenyl sulfides, which appear to have general applicability. The preparation of the p-alkoxybenzoic acids involves a lengthy experimental procedure, and the alkyl triiodophenyl ethers, p-alkoxydiphenylamines, and (11) T. Smedsland, Finska Kemistamfundets Medd., 41, 13 (1932); alkyl 6-nitrobenzothiazoly1 sulfides are low melting Chem. Bbstr.. 2 6 , 5905 (1932). (12) Too low a combustion temperature was used resulting in the derivatives. The alkyl P-naphthyl ethers usually low carbon results observed. were separated as oils and had to be isolated as picrates. The usefulness of S-alkylmercaptosuccinic acids has been demonstrated only for alkyl Potassium p-Phenylazophenoxideas a bromides which require long reaction times (four to twenty-four hours) for preparing derivatives. Reagent for the Identification of Organic We wish to report the use of p-phenylazophenol Halogen Compounds as a reagent for the identification and chromatographic separation of organic halogen comE. 0. WOOLFOLK, ELWOOD DONALDSON,’ pounds. Similarly, Hurd and c o w ~ r k e r s ~have ~J~ AND MUERLPAYNE‘ reported the use of pphenylazophenol as a reagent Chemistry Department, Central State College, Wilberforce, Ohio for the preparation and chromatographic separation of p-phenylazophenyl polyacetylglycosides. In Received December 11, 1961 several instances,14-16 the reaction of an alkyl halide with p-phenylazophenol under alkaline The literature records many reagents which conditions to form an ether has been reported. have been suggested for the identification of alkyl The potassium salt of p-phenylazophenol (m.p. halides.2J No one of these is entirely satisfactory, 155-156’) reacted readily with organic halogen either because of difficulty in preparation or be- compounds in dimethylformamide or dimethcause their application is not general. It has been oxyethane to form crystalline p-phenylazophenyl suggested4 that thiourea and p-bromobenzene- ethers in good yields. In a few cases of these sulfon-p-anisidide appear to be the most generally derivatives similar melting points were obtained; useful reagents. However, the 8-alkyl isothio- however, a mixture melting point showed a marked ureas have to be isolated as picrates or styphnates depression in all such cases. and the n-alkyl-p-bromobenzenesulfon-p-anisidides Of fifty halides investigated, all the primary have low melting points. More recently it has and secondary halides reacted, in many instances been stated3 that the best method for making with only mixing and no heating. The reagent also derivatives of alkyl halides was their conversion reacted with straight- and branched-chain halointo the corresponding Grignard reagents and the gen esters, halohydrins, halo ketones, halo ethers, reaction of these with phenyl-, p-tolyl-, or cy- and chloroformates. Good yields of the derivatives naphthyl isocyanate to give the anilides, toluides, of the chloroformates were obtained only by allowor a-naphthalides. However, this is a rather ing their reaction mixtures to stand for sixteen lengthy experimental procedure for preparing hours at room temperature. The vicinal halide, derivatives, and its application is limited only to l,Zdibromoethane, gave constant melting point organic halogen compounds which form Grignard material (196.0-198.0’) ; however, analytical rereagents. sults indicated that this material was a mixture of Among the many other derivatives suggested the mono- and diderivatives. A well defined for the identification of alkyl halides are the p( 5 ) W. M. Lauer, P. A. Sanders, R. M.Leekley, and H. E. Unguade, alkoxybenzoic acids,6 alkyl triiodophenyl ethers,6 ibia., 61,3050(1939).

not react with selenium tetrachloride even on extended refluxing in benzene. Selenoxide Preparations.-All selenoxides were prepared by the same procedure. A slurry made from several gram of the selenium dichloride and 100 ml. of water wm stirred for 2 hr. at room temperature. The resulting material was recrystallized from acetone yielding a white solid. The results are shown in Table I. Stirring for longer periods (up to 2 days) did not result in hydrolysis of the chlorine-chlorine linkages as is reported by Smedslandll for bis-2-chloroethylselenium dichloride.

(1) This paper is based on work presented b y Elwood Donaldson and Muerl Payne in partial fulfillment of requirements for an undergraduate research course offered in the Department of Chemistry of Central State College. (2) A comprehensive review of the reagents, which have been suggested for the identification of alkyl halides, has been published b y Hopkins and Williams Research Staff, “Organic Reagents for Organic Analysis,” 2nd ed., Chemical Publishing Co., Inc., Brooklyn, New York, 1950,PP. 32-33. (3) R.L.Shriner, R. C. Fuson,and D . Y. Curtin, “The Systematic Identification of Organic Compounds,” John Wiley and Sons. New York, N. Y., 1956,p. 242. (4) L. L. Merritt, Jr., S. Levey, and H. B. Cutter, J. Am. Chrmd Soc., 61,15 (l9aQ).

( 6 ) R. D. Drew and J. M. Sturtevant, ibid., 61,2666 (1939). (7) V. H.Dermer and 0. C. Dermer. J. O w . Chem., 3 , 289 (1938). (8) D. F. Houston, J. A m . Chem. Soc., 11,395 (1949). (9) R. W. Bost, P. K. Starnes, and E. L. Wood, ibid., 73, 1968 (1951). (10) H.B. Cutter and H. R . Golden, ibid., 69,831 (1947). (11) J. G. Hendrickson and L. F. Hatch, J . Org. Chem., 26, 1747 (1960). (12) C. D. Hurd and W. 8. Bonner, ibid., 11,50 (1946). (131 C. D. Hurd and R. P. Zelinski, J. Am. Chem. Soc., 69, 243 (1947). (14) Sa Scichilone, Gazd. chim. ital., la, 110 (1882). (15) P. Jacobsen and w. Fischer, Ber., 26, 994 (1892). (16) L. Claisen and Qi Eisleb, Ann., 401, [See Chem. Absfr., 8, 64 tlQl4)l.



O N = N O O R -Halide



% M.P.. 'Ga Methylc I 93 51.5-52.0 Methylened I 81 183.0-184.5 I 86 71.0-72.0 C: Ethyl8 Hydroxyethylf C1 80 99.0-100.0 Br 68 59.5-60.0 C* 1-Propyl0 Br 76 77.0-87.0+ 3-Chloropropyl %Propenylh Br 77 50.5-51.5 Br 77 115.5-116.5 Carbomethoxymethyl Carboethoxy C1 95(s) 84.0-84.5 Br 90 60.0-61.5 CC l-Butyl' 1-Butyl' C1 30 60.0-61.5 I-Butyl' I 90 60.0-61.5 2-Butyl Br 72 62.0-63.0 %Butyl C1 58 62.0-63.0 %Butyl I 49 62.0-63.0 Carboethoxymethyl' Br 95 75.5-76.0 C1 95(s) 57.0-59.0 Carbo-7-chloropropoxy Br 81 36.5-37.0 3-Methylbutyl c b Br 90 57.4-57.7 CS 1-Hexyl I-Hexyl C1 70 57.4-57.7 1-Hexyl I 91 57.4-57.7 2,4Dinitrophenyl Br 73 137.0-138.0 CNitrophenyl Br 93 138.0-140.0 Br 95 68.05-69.0 CI 1-Heptyl 1-Heptyl C1 91 68.05-69.0 I 92 68.05-69.0 1-Heptyl l-Carboethoxy-2methylpropyl Br 64 58.0-59.0 1-Octyl Br 93 72.0-73.0 c 8 1-Phenylethyl Br 87 108.0-109.0 Rr 40 84.5-85.0 ZPhenylethyl Br 90 128.0-128.5 Benzoylmethyl Benzoylmethyl C1 71 128.0-128.5 Br 94 92.5-93.0 CP 3-Phenoxypropyl Br 94 63.0-64.0 ClO I-Decyl 1-Decyl C1 94 63.0-64.0 I 97 63.0-64.0 1-Decyl Br 96 79.0-80.0 C1a 1-Hexadecyl Br 96 83.0-84.0 C18 1-Octadecyl Yields are reported on crude products from reactions run in dimethylformamide. Reactions were carried out under reflux unless indicated by (8) = shaking and standing. Melting pointa are on derivatives purified by chromatography then recrystallized. The found analyses agreed with the calculated values to within &0.24/0 nitrogen except in six cases and the difference was not greater than &0.5% except in one case. Microanalyses were performed by the Du Good Chemical Laboratory, St. Louis, Missouri. *Nitrogen and chlorine analyseB were performed on this compound and both agreed with the calculated value within k0.273,. Melting pointa which have been, reported for the methyl ether prepared by various synthetic methods were: 56' [A. Colombano, Gam. chim. ital., 37, 11, 474 (1907)l; 54' [H. Gorke, E. Koppe, and F. Staiger, Ber., 41, 1157 (1908)l; 55.5' [P.Pfeiffer and 0. Angern, 2.angm. Chem., 39, 258 (1926)l; 64' [J. Burns, H. McCombie, and H. A. Scarborough, J. Chem. SOC.,2934 (1928)l; 54' [E. Bergmann and A. Weizmann, Trans. Faraday SOC.,32, 1321 (1936)l; 56' [A. Burawoy and I. Markowitsch-Burawoy, J . Chem. SCC.,39 (1936)l; 53.5" [R. P. Zelinski and W. A. Bonner, J. Am. Chem. SOC.,71, 1792 (1949)l; 52-53' iJ. N. Ospenson, Acta Chem. Scand., 5, 491 (1951)l. Bisether structure in which R group is connected to two p phenylazophenoxy radicals. e Melting points which have been reported for the ethyl ether prepared by various synthetic methods were: 77-78" [see ref. 151; 78" [G. Charrier and G. Ferreri, Gam. chim. itul., 44, 175 (1914)J; 85' R



VOL. 27

[E. Naegeli, Bull. SOC. chim. 131, 11, 897 (1894)]; 85" [H. Gorke, E. Koppe, and F. Staiger, Ber., 41, 1157 (1908)l. I Melting point reported for the hydroxyethyl ether obtained by reaction of p-phenylazophenol with ethylene oxide was 100' [D. It. Boyd and E. R. Marle, J. Chem. SOC.,105, 2123, 2137 (1914)l. ''The melting point of n-propyl ether has been reported to be 61' [H. Gorke, E. Koppe, and F. Staiger, Ber., 41, 1157 (1908)l. Ir The 2-propenyl ether has been reported to melt a t 52' [see ref. 161. The n-butyl ether has been reported to melt a t 67' [H. Gorke, E. Koppe, and F. Staiger, Bet., 41, 1157 (1908)l. i The carboethoxymethyl ether has been reported to melt a t 70" [J. Mai and F. Schwabacher, Bet., 34,3937 (1901)l.

product also was not obtained with l-bromo-2chloropropane. The disjunctive halide, l-bromo3-chloropropane, gave a well defined product in which the bromo group had been displaced as shown by analysis for nitrogen and chlorine. The reagent also reacted with the aryl halides, l-bromo4-nitrobenzene and l-bromo-2,4-dinitrobenzene. The reagent did not react with cyclohexyl bromide, tertiary alkyl halides, bromo or iodo aromatics, and 1,2- and 1,3-bromonitrobenzene. The various p-phenylazophenyl ethers prepared during this investigation and their physical properties are given in Table I. The brilliantly colored ethers can be eluted from columns of alumina, silicic acid, and mixtures of each of these adsorbents with Celite. Our chromatographic studies indicated that there is not a large enough difference in the adsorption affinities of homologous ethers differing by less than four carbon atoms to permit separation of mixtures of them. Experimental Reagents.-The organic halogen compounds were commercially available grades, and were used without further purification. p-Phenylazophenol is manufactured by Distillation Product Industries. The solvent Skellysolve B (b.p. 60-71') waa redistilled and ethyl alcohol, dimethylformamide, and dimethoxyethane were used aa purchased. The adsorbents used in preparing the chromatographic columns were silicic acid (Mallinckrodt, prepared by the method of Ramsey and Patterson), alumina (Aluminum Company of America, Grade F-20) and Celite-535 (JohnsManville). Potassium p-Phenylazophenoxide .-Potassium hydroxide pellets (1.5 9.) were added to a solution of 5 g. of p-phenylazophenol in 20 ml. of absolute alcohol. The mixture was refluxed until complete dissolution of the potassium hydroxide occurred. The solution waa concentrated and the last traces of solvent were removed in uucuo under a, stream of nitrogen. The salt obtained t u a deep red powder was dried in a vacuum desiccator. Preparation of p-Phenylazophenyl Ethers.-A mixture of approximately 0.5 g. (0.002 mole) of potassium p-phenylazophenoxide and an equal molar quantity of the halide was dissolved in 10 ml. of dimethylformamide. In many case8 there was an immediate reaction as indicated by the formation of salt. The mixture was then heated under reflux for I-hr. Better results were obtained with the chloroformates if the reaction was allowed to stand at room temperrLture for about 16 hr. When the refluxing waa completed, the mixture was cooled to room temperature and poured into a mixture of ice and water. The inorganic salt dissolved. The crude reaction product was filtered.

JULY, 1962



washed with water, and dried. The crude product waa dissolved in Skellysolve B or dimethoxyethane or a mixture of these two solvents and chromatographed on silicic acid on which a small amount of gummy material was strongly adsorbed. The derivative waa then recrystallized from alcohol or alcohol-water mixtures. The crude products could be recrystallized to give substantially constant melting points over a range of two degrees. However, for good analysis it waa necessary to remove a small amount of impurity by chromatography.

was dependent on the rate of heating; melting was accompanied by decomposition and resolidification took place upon further heating. Pyrolysis of I1 at 170' (below 1 mm.) yielded 1-naphthyl isocyanate. No other product of the pyrolysis (which might include the 1-naphthylurethane of I) was isolated. Single attempts to reduce the keto group of I1 with either sodium borohydride or lithium tri-tert Acknowledgment.-We are indebted to the butoxyaluminohydride and attempted reaction of National Science Foundation for financial aid in the I1 with 2,4-dinitrophenylhydrazine or semicarbapursuance of this work. zide failed to give isolable products. Sodium hypoiodite did not yield a positive iodoform test with 11, although a reaction did occur. In contrast, a color test for methyl or methylene ketones6 was positive Reaction of 1-Naphthyl Isocyanate with for I1 and reaction between I1 and hydroxylamine 3-Hydroxymethyl-3-methoxymethyl-Zyielded the oxime 111. butanone. A Reinvestigation Substituted allophanates are distinguishable from urethanes in that the latter, but not the former, S. E. ELLZEY, JR., AND CHARLES H. MACK possess an absorption band in the region of 6750 cm.-', attributable to the first overtone of the NH Souulenz Regional Research Laboratory,' New Orleans 19, stretching vibration at 3500 cm.-1.6 In substituted Louisiana allophanates the band at 6750 cm.-l is replaced Received December 12, 1061 by one at 4831-4877 cm. -I which is characteristic of Upon reaction of 1-naphthyl isocyanate with 3hydroxymet h yl-3-met hoxymet h yl-Zbu tanone (I) in the presence of pyridine, Tieman and Gold2 obtained a compound (no experimental details or yield given), m.p. 162+163', the analysis of which indicated the empirical formula CzsH2sNz05.3 They suggested reaction of the second equivalent of isocyanate with the hydroxyl group formed by enolization of the acetyl group, although they noted that the acetate of I did not form a urethane with the isocyanate. The present report identifies by spectral and chemical data the r e a d o n product obtained by Tieman and Gold as (2-methoxymethyl2 -methyl - 3-oxo)butyl 2,4 - bis(1- naphthy1)allophanate (11). Several unsuccessful attempts were made to prepare I1 using the usual methods for the preparation of urethanes. The method of Kogon' for the preparation of allophanates waa successfully used to prepare I1 in good yield. The melting point of I1 0 CHa


/I I O--....--.H


these compounds.' The near infrared spectrum of I1 in benzene shows a strong band at 4819 cm.-' but absorption in the 6667-7042-cm.-' region is lacking. The infrared spectrum of I1 obtained at high resolution possesses twin absorptions at 1725 and 1734 cm.-', which may be assigned to the two allophanate C=O group^.^ A third band at 1701 cm.-' is probably due to the keto C=O absorption .' In the oxime I11 the characteristicallophanate C=O bands are well resolved peaks a t 1727 and 1733 cm.-'. A band at 1690 cm.-' may be assigned to the C=N vibration of the oxime.8 All the above evidence is in good agreement with the structures postulated for I1 and 111. Experimental0

3-Hydroxymethyl-3-methoxymethyl-2-butanone (I) was prepared according to the procedure of Tieman and Gold.' (2-Methoxymethyl-2-methyl-3-oxo)butyl2,4-bis(l-naph-


~ H ~ O H


the hydrogen-bonded grouping -CNCON-






**.H-( I-Naphthyl) 11. =0 111. X = NOH


(1) One of the laboratories of the Southern Utilization Research and Development Division, Agricultural Research Service, U.S. DeparC ment of Agriculture. Mention of trade names and firms does not imply their endorsement by the Department of Agriculture over similar products or firms not mentioned. (2) C. H. Tieman and M. H. Gold, J . Oro. Chem.. PS, 1856 (1958). (3) Tieman and Gold erroneously reported the formula CrsH*&Os: their aotual analysis ia oorrect for CwHzsNzOs. ( 4 ) I. C. Kogon, J . Am. Chem. Soc., 78, 4911 (1956).

thy1)allophanate (11) waa prepared essentially according to the procedure of Kogon.4 A mixture of 8.7 g. (0.06 mole) of the alcohol I and 60.8 g. (0.36 mole) of 1-naphthyl isocyanate (protected by a calcium chloride tube) WM heated in an oil bath at 130' for 24 hr. Distillation of the reaction mix(5) F. Feigl, "Spot Teats in Organic Analysis," 6th English ed., Elsevier Publishing Co.. New York, 1960, p. 236. (6) I. C. Kogon, J . Am. Chsm. Soc., 79,2253 (1957). (7) L. J. Bellamy, "The Infra-red Speotra of Complex Molecules," 2nd ed., John Wiley and Sons, Inc., New York, 1958, p. 132. (8) Ref. 7, p. 268. (9) Melting points and boiling points are uncorrected. Infrared speotrs were obtained from potassium bromide disks with the PerkinElmer spectrophotometers, Modela 21 and 221. Near infrared spectra were obtained in benzene solution with the Cary Model 14 spectrophotometer.




ture at 0.05 mm. afforded 7970 of the excess isocyanate, b.p. 71-72'. T o the residue waa added 50 ml. of carbon tetrachloride and the insoluble solid, m.p. 284", 3.7 g., waa filtered from the mixture. To the filtrate waa added 100 ml. of petroleum ether (b.p. 30-60') and after chilling, the crude product 11, 14.0 g., waa obtained. The filtrate waa reduced in volume and 4.4 g. (total crude yield 63y0) of solid, m.p. 150-160°, waa obtained. Recrystallization of the two crops of crystals from acetone allowed the isolation of 11, m.p. 160-161" (lit.,2 m.p. 162-163'). The total weight of the acetone-insoluble solid (probably N,N'-bis( 1-naphthy1)urea or the isocyanate dimer) was 6.0 g. An analytical sample of I1 waa recrystallized from aqueous acetone and yielded small, colorless needles, m.p. 159-160'. Ana2. Calcd. for C20H~8N205: C, 71.88; H, 5.82; N, 5.78; OCH3, 6.40. Found: C, 72.54; H, 5.88; N , 5.94; OCHs, 6.02. The product I1 gave a positive test for a methyl ketone using the alkaline sodium nitroprusside reagent .5 The infrared spectrum of I1 had bands at 3322 and 3413 cm.-l (free and bonded NH, respectivelylo) and at 1701, 1725, and 1734 cm.-l. The near infrared spectrum of a solution of I1 in benzene showed a strong band a t 4826 cm.-l but no band in the range 6667-7042 cm.-l. A 2-g. sample of I1 was heated at 170' below 1 mm. preesure and a small amount of liquid was distilled from the melt by heating to 195' at 0.1 mm. Reaction of this liquid with ethanol gave ethyl 1naphthylurethane, identified by melting point and mixture melting point with an authentic sample. (2-Methoxymethyl-2-methyl-3-oximino)butyl 2,4-bis( 1naphthy1)allophanate (111) w&g prepared by refluxing for 2 hr. a solution of 11, hydroxylamine hydrochloride, and pyridine in absolute ethanol. The oxime (80y0yield), after three recrystallizations from aqueous ethanol, had m.p. 168-169". Anal. Calcd. for CZ8H~N805: C, 69.72; H, 5.85; N, 8.41. Found: C, 69.76; H, 5.78; N, 8.34. The infrared spectrum of I11 showed a broad band at about 3268 cm.-l (free and bonded N H and OH'O) and three well resolved bands at 1690, 1727, and 1733 cm.-'.

Acknowledgment.-The authors wish to thank Sylvia H. Miles, Gordon J. Boudreaux, and Elizabeth R. McCall for the spectral data, and Lawrence E. Brown, Alva F. Cucullu, and Marian C. Willis for the analyses. (10) Ref. 7,p. 5.

Mannich Bases and Aromatic Amines in Amine Exchange Reactions NAZAR SINGHAND SCKHDEV SINGH

Department of Pharmacy, Punjab Univei situ, Chandigarh-3, India Received December 18, 1961

It was of interest to determine if Schiff bases could be obtained by reaction of hydrochlorides of Mannich bases with primary arylamines in the presence of a condensation agent. After a mixture of p-dimethylamino- (I) or 0piperidinopropiophenone hydrochloride (11), aniline, zinc chloride-aniline complex, and ethanol had (1) V. M. Tbaker and R. G Shah, J , Indian C h r m Soc,, 26, 261 (1949)h

been refluxed, the reaction product isolated in each instance proved to be p - phenylaminopropiophenone. It was evident, therefore, that amine exchange reactions, instead of Schiff base formation, had taken place: these reactions are similar to those which have been reported2-8 for Mannich bases or their salts and aliphatic amines. When I or I1 was allowed to react with other aromatic amines such as p-chloroaniline, p-toluidine, p-anisidine, or p-phenetidine, the products, p-arylaminopropiophenones, obtained from interaction of either I or I1 with a specific amine were identical. In order to establish identity, the p-arylaminopropiophenones were synthesized from p-chloropropiophenone and the required arylamine. P-pToluidinopropiophenone was also obtained when I was heated under reduced pressure and the phenyl vinyl ketone formed was allowed to react with ptoluidine. In an attempt to obtain a Schiff base of the desired type in a different manner, a mixture of equimolar amounts of acetophenoneanisil, dimethylamine hydrochloride, paraformaldehyde, and ethanol was heated for two hours on a water bath; only oily intractable material was obtained. Experimental

(3-Arylaminopropiophenones(Table I).-A. A mixture of the hydrochloride of the Mannich base (I or 11) (0.02mole) the required amine (0.02 mole), and 25 ml. of absolute ethanol was refluxed for 6 hr. After 12 hr. at room temperature, the precipitate was recrystallized from absolute ethanol. B. A mixture of P-chloropr~piophenone~ (0.0065 mole), the required amine (0.015 mole), and water waa heated for 30 min. at loo", cooled, and the precipitate was recrystallized from ethanol. The melting points and mixed melting points of corresponding products prepared by methods A and B were identical. The products are insoluble in water, slightly soluble in benzene and ether, and soluble in chloroform. The hydrochlorides, which are insoluble in water, were prepared by mixing a solution of 8-arylaminopropiophenone in dry acetone with dry acetone made acid to congo red by passing in dry hydrogen chloride, and recrystallizing the precipitate from alcohol. p-Anilinopropiophenone hydrochloride and p-p-phenetidinopropiophenone hydrochloride separated only on trituration after addition of ether. The former WM recrystallized from alcohol-acetone mixture. The hydrochloride of p-p-toluidinopropiophenonewas recrystallized from alcoholether mixture. 8-p-To1uidinopropiophenone.-Compound I (4.0 g.) waa heated in an oil bath at 160-170O under 18 mm. pressure. The phenyl vinyl ketone which distilled (1.8 g.), b.p. 114116"/18 mm., was dissolved in chloroform and refluxed with

(2) E. E. Howe, A. J. Zambito, H. R. Snyder, and M. Tishler, J . A m . Chem. Soc., 67, 38 (1945). (3) H.R. Snyder and J. H.Brewster, i b i d . , TO, 4230 (1948). (4) H.R. Snyder and E . L. Eliel, ibid., TO, 4233 (1948). (5) H.R. Snyder and J. H.Brewster, ibid., 71, 1058 (1949). (6) Ha R. Snyder and W. E. Hamlin, ibid., 7 8 , 5082 (1950). (7) Jb Kenner end F,Sa Ststhem, J . Chem. Sot,, 999 (1986).



JULY, 1962






-Carbon, Calcd.


-Hydrogen, %- -Nitrogen, %- -Chlorine. Calcd. Found Calod. Found Calcd.


CsH6 111-112" 55' 66' C16H160N 80.00 80.01 6.66 6.70 6.22 6.20 5.35 5.62 13.57 13.74 150-15Id Ci6HieONCl CsHs 736-138 39 58 C16HlaONCl 69.36 69.41 5.40 6.03 5.40 5.11 P-CICSH, 160d CisH150NClt 4.73 5.07 Z3.98 24.39 p-ClC66 113-114 62 66 Ci6H1iON 80.34 79.73 7.11 6.73 5.85 5.50 p-CHaCsHa CieHiaONCl 5.08 5.45 12.90 13.65 p-CHaCsHa 137-13Sd p-CHzOCeHa 114-115 40 53 C ~ ~ H I ' O ~ N 75.29 74.86 6.66 6.38 5.49 5.91 4.80 5.05 12.18 12.43 p-CHaOCsH4 149-150d CisHisOzNC1 37 52 C1iH190zN 75.70 75.46 7.06 7.29 5.40 5.23 P - C ~ H ~ O C S H ~105-106 154-155d Ci'HzoOzNCl 4.58 4.83 11.62 11.91 p-C&.OCsH( " Reported m.p. 111-112', J. Kenner and F. S. Statham, J . Chem. floc., 299 (1935). The yield reported in this column is that obtained by the use of 8-dimethylaminopropiophenone. The yield reported in this column was obtained when @-piperidinopropiophenonewas employed. Hydrochloride.

1.4 g. of p-toluidine for 30 min. The solvent was removed, and the residue was recrystallized from ethanol; m.p. and mixed m.p. 113-115'.

The reaction with the silicon-containing amine was extended to other chloro-s-triazines to form 2,4diamino - 6 - trimethylsilylmethylamino - s - triazine (V) and 2,4,6-tris(trimet hylsilylme t hylamho)s-triazine (VI). It seems likely that a large Silicon-Containing s-Triazine Derivatives number of silicon-containing derivatives of striazine are accessible through reactions of chloroJ. C. GOAN,'S. H. SMITH, JR., AND R. R. MILLER s-triazines with carbon-functional silanes. Because of the extremely low water solubility of chemistry Division, U.S.Naval Research Laboratory, the silanes their reactions with chloro-s-triazines Washington, D. C. were carried out in nonaqueous solvents. The amino-s-triazines I, V, and VI were prepared in Received December 18, 1961 benzene from the chloro-s-triazines and trimethylsilylmethylamine in the presence of triethylamine. Most reactions involving organic substituent in Attempts to obtain I1 and I11 using triethylamine silanes, particularly those in which the functional as the hydrogen chloride acceptor were not successgroup is separated from the silicon atom by only ful, and these compounds were synthesized through one methylene group, are only slightly affected by the interaction of the chloro-s-triazines and the the presence of silicon and are not very different sodium alcoholate or thiolate. It is probably that from those of analogous purely organic compounds.2 the slight electron-releasing effect of silicon, Since many derivatives of s-triazine can be obtained which makes trimethylsilylmethylamine a strong by nucleophilic displacement reactions with chloro- b a ~ e ,favors ~ , ~ the reaction with chloro-s-triazines; s-triazines,3 it seemed to us that one route to the however, the low reactivity of alcohols and thiols preparation of s-triazines having silicon-containing toward chloro-s-triazines6 does not seem to be substituents would lie in the treatment of chloro-s- greatly affected by the presence of silicon. triazines with various nucleophilic carbon-functional silanes. I n this paper we wish to report the success Experimental of this approach, which has led to the preparation of the s-triazine derivatives 1-111 by the reaction of 2,4-Diphenyl-6-trimethylsilylmethylamino-s-triazine the appropriate silicon-containing amine, alcohol, (I).-A solution of 0.10 mole of trimethyl~ilylmethylamine~ or thiol with 2,4-diphenyl-6-chloro-s-triazine (IV) . (b.p. 94-95'/756 mm.) and 0.13 mole of freshly distilled


I. X = NHCH2Si(CH& 11. X = OCHzSi(CH& 111. X = SCHZS~(CHS.)~ (1) Present address: Melpar, Ino., Falls Church, Va. (2) P. D. George, M. Prober, and J. R. Elliott, Chem. Rev.,66, 1065 (1956). (3) E. M. Smolin and L. Rapoport, "8-Triazines and Derivatives," Intericience Publithera, Inca,New York, N. Y., 1969. pp. 63-62.

triethylamine in 200 ml. of dry benzene was placed in a 500-ml., three-necked flask equipped with a mechanical stirrer, reflux condenser, and dropping funnel. Provision was made for keeping the entire system under an atmosphere of dry nitrogen. Stirring waa begun and 0.10 mole of 2,4diphenyl-6-chloro-s-triazine' (IV) in 100 ml. of benzene was added over a period of 1 hr. The mixture was then heated to reflux and stirring continued for 6 hr. After cooling, the precipitate of triethylamine hydrochloride (0.074mole, 74%) was filtered off and the filtrate was evaporated to dryness under reduced pressure. Upon recrystalliaation of (4) J. E. Noll, B. F. Daubert, and J. L. Speier, J . Am. Chem. Soc.. 78, 3781 (1961). (5) L. H. Sommer and J. Rockett, ibid., 73, 5130 (19iil). (6) H.Schroeder, ibid., 81, 5659 (1959). (7) R. Ostrogovich, Chem. Z t p . , 16, 738 (1912).




TABLE I Reaotgnta


+ trimethylailylmethylamine (TMSMA)







Analyak forb





Calcd. 68.22 6.64 16.75 Found 67.82 6.47 16.89 I V -t Na alcoholate of trimethylsilylmethanolc I1 43 94-95 Ci$H?1NaOSi Calcd. 68.02 6.32 12.53 Found 68.14 6.51 12.10 I V -t Na thiolate of trimethylsilylmethylthiold I11 50 103-104 C19Ht1NsSSi Calcd. 64.91 6.04 11.95 Found 65.07 6.35 11.19 2,4-Diunino-6-chloro-s-triazinee TMSMA V 34 153-155 CtHleNeSi Calcd. 39.60 7.60 39.58 Found 39.40 7.30 39.39 2,4,6-Trichloro-s-triasine~ TMSMA VI 48 136-137 ClsHssN& Calcd. 46.85 9.43 21.83 Found 47.20 9.92 21.59 a FisherJohns-uncorrected. a Analyses were performed by Schwarzkopf Microanalytical Laboratory, Woodside, N. Y., and by Dr. Mary Aldridge, American University, Washington, D. C. ’ J. L. Speier, B. F. Daubert, and R. R. McGregor, J. Am. Chem. SOC.,70,1119 (1948). D. C. Noller and H. W. Post, J . Org. Chem., 17,1395 (1952). e J. T. Thurston, J. R. Dudley, D. W. Kaiser, I. Hechenbleikner, F. C. Schaefer, and D. Holm-Hansen, J . Am. Chcm. SOC.,73,2981 (1951). 1 Eastman Practical grade cyanuric chloride.



the residue from 95% ethanol, I waa obtained in 48y0yield. In general, all preparations were conducted in a manner similar to that used for compound I. Compounds I1 and I11 were prepared directly from the products of the reaction of sodium and the alcohol or thiol without isolating the intermediate. All products were white needles except I and 111; thwe were light yellow. Ethanol (95%) was the recrystalhation solvent except for compound I11 which ww recrystallized from methanol, and VI, for which 60:40 mixture of benzene and n-hexane w a ~employed. Table I shows the reactants, yields, analysis, etc. Each of the compounds reported here shows strong infrared absorption at 860 cm.-1 and 1243 cm.-l, frequencies considered characteristic of the (CHa)a Si-CH2 - group.*

boxylic acid with a diphenylamine (I) and zinc chloride a t 200-270’ for about twenty hours.2 The wide use of polyphosphoric acida as a reagent in organic chemistry has led to this study of the use of polyphosphoric acid as a catalyst in the Bernthsen reaction. In a previous study of this type Elston4 has reported that the reaction of diphenylamine and an excess of benzoic or p-toluic acid, in the presence of polyphosphoric acid at 160’ for fifteen minutes, led to the formation of a diacyldiphenylamine (probably bis-p-) in a 12-15% yield, rather than a 9-substituted acridine. It was felt that perhaps by changing the reaction conditions, particularly the ratio of amine (8) L. J. Eellamy, “The Infrared Spectra of Complex Molecules,’’ John Wiley and Sons, New York, N. Y.,1954, pp. 274-281. and acid, one could successfully apply polyphosphoric acid to the Bernthsen reaction. From the results of the reaction of benzoic acid and diphenylamine in polyphosphoric acid under a Polyphosphoric Acid in the Bernthsen variety of conditions it appears that a ratio of one Reaction’ mole of acid to two moles of amine and a temperature of about 200’ for fifteen minutes are the best FRANK D. POPP conditions. Use of benzoic anhydride and ethyl benzoate gave results similar to benzoic acid. It is Department of Chemistry, University of Miami, Coral Gables, of interest to note that in one run the reaction was Florida scaled up from 0.03 to 0.1 mole and the yield increased from 18 to 48%. Received December SO, 1961 Table I lists the results obtained. No attempts were made to find the ideal reaction conditions in The Bernthsen reaction, one of the earliest used these cases. The use of p-nitro- and p-aminobenfor the synthesis of acridines, (11) consists in heating a mixture of an aromatic or aliphatic car- zoic acids did not lead to acridines by the use of zinc chloride.6 With polyphosphoric acid the latter gave 9-(p-aminophenyl)acridine while pR nitrobenzoic acid did not lead to any acridine. The use of polyphosphoric acid, rather than zinc chloride, is more convenient, particularly in regard to reaction times, and in view of the ready availI H ability of starting materials this may be a worthI I1 while procedure in many cases, despite the poorer yields. (1) Presented in part at the Combined Southeastern-Southwestern Regional American Chemical Society Meeting, New Orleans, Louisiana, December, 1961. (2) For reforences me: (a) R. M. Acheson, “Acridines,” A. Weis.+ berger, ed., Interscience Publishers, New York, N. Y., 1956. pp. 19-25; (b) A. Albert, “Heterocyclic Compounds,” Vol. 4, R. C. Elderfield. ed., J. Wiley and Sons,Inc., New York, p7. y,, 1952, pp. 502-594.

(3) (a) F. D. Popp and W. E. McEwen, Chem. Re%, 56, 321 (1958); (b) F. D. Popp and W. E. McEmen. PTOC. Kansas Acad. Sci., 69, 189 (1960). (4) C. T. Elston, Ph.D. thesis, University of Illinois, 1954; quoted in ref. 38. (5) W. Heae and A. Bernthsen Bw., 16, 689 (1885). e

JULY, 1962






Apparent meta Rearrangement of Benzoylsulfuric Acid STEPHENJ. TAUBER AND NANCY N. LOWRY Department of



Reported M.P.~

ported yieldsd

184' 48 18' 184-185'f CeHS 173" 50 11 170-173' CsHsCHz 8 101-103' 104' 25 CeHsCHzCH&Hz 66-68' 69-70' ? CHdCHzha Trace p-BrCeHI 6 239-240°g 234' ? p-CHGK 18 188-189' 189-190" 40 ... 0 p-NHzCeHd 24 270-2720h ... 0 p-NOpCsH4 0 ... a All reaction were 0.03 mole acid to 0.06 mole amine with 230-250 g. PPA a t 195-205' for 15 min. Yield of material purified to within 2-3' of reported m.p. M.p. of purified See ref. 2a product (from ethanol or aqueous ethanol). for original literature reference. 48% in a 0.1 mole run. I Phosphate salt, m.p. 194-196'. Calcd. for C19HlsNP0.: C, 64.58; H, 4.56; N, 3.97; P, 8.77. Found: C, 64.43, 64.64; H, 4.65, 4.63; N, 3.94, 4.04; P, 8.72. Calcd. for C19HlzNBr: C, 68.28; H, 3.62; N, 4.19. Found: C, 68.15; H, 3.81; N, 4.43. Calcd. for CleH14N2:C, 84.41; 11, 5.22; N, 10.37. Found: C, 84.07; H, 5.40; N, 10.38.


In several instances a small amount of ketonic material, which is probably similar to the diacyldiphenylamine of Elstonj4 was noted. Experimentale Preparation of 9-Substituted Acridines.-A mixture of diphenylamine, a carboxylic acid, and polyphosphoric acid' was heated, with stirring, to the desired temperature and held at this temperature for the stated time. The reaction mixture was then poured onto ice and filtered or decanted. Examination of the infrared spectra of the gummy solid indicated that it contained mainly diphenylamine, a small amount of ketonic product,* and possibly some 9-substituted acridine.0 Treatment of the solution with 25% sodium hydroxide solution caused the precipitation of 8 solid (while the solution waa still acidic) which waa apparently the phosphate srtlt of the 9-substituted acridine.IO After filtration the solution waa made strongly basic with sodium hydroxide and extracted with chloroform. Concentration of the chloroform generally gave a trace of the 9-substituted acridine. The phosphate salt was shaken with 25% sodium hydroxide and chloroform and the chloroform concentrated to give the 9-substituted acridine as listed in Table I. (6) AI1 melting points are uncorrected. Analysis by Spang Microanalytical Laboratory, Ann Arbor, Michigan. (7) Victor Chemical Co. (8) IQ the case of the benzoic acid reaction, this fraction had peake identical t o those reported by Elston4 for his dibenzoyldiphenylamine. In addition, in the case of stearic acid treatment of this gum with sodium hydroxide and chloroform gave, OQ evaporation of the chloroform, a solid, m.p. 107-108.5" (from absolute ethanol), which was a p parently distearyldiphenylamine (Calcd. for ClsHisNOn: C, 82.11: H , 11.34: N, 2.00. Found: C, 82.41: H, 11.26: N. 1.98, 2.11). (9) Since we were interested in the convenience of the method, no attempt was made t o isolate this acridine and include i t in the yield. (10) In the case of 9-phenylacridine this salt was washed with water and alcohol t o give analytically pure material (see Table I). The same a a l t could be obtained b y reacting 9-phenylacridine with 85% phosphoric acid.

Chemistrql, Smith College, Northampton, Massachusetts Received December 26,1961

Benzoylsulfuric acid was reported by Oppenheim' to rearrange to m-sulfobenzoic acid. Elliott, et d.,2 described the rearrangement as yielding chiefly o-sulfobenzoic acid. They suggested that the difference in products is due to direct sulfonation under Oppenheim's conditions as opposed to chiefly rearrangement under their own. Oppenheim either let benzoylsulfuric acid stand in the presence of sulfuric acid or warmed sulfuric acid with an excess of benzoyl chloride; details of time, temperature,3 and amounts are lacking. Elliott, et al., heated benzoylsulfuric acid per se a t 100'. We have established that if the reaction mixture from sulfuric acid and a large or small excess of benzoyl chloride is heated at 100' then the predominant product is m-sulfobenzoic acid. Our evidence indicates that this product is not due to direct sulfonation. The presence of unchanged sulfuric acid was avoided by the use of an excess of benzoyl chloride, the size of the excess evidently being immaterial, and by drawing out with dry air the hydrogen chloride produced in the initial reaction. Even should a trace of unchanged sulfuric acid have been present we could argue against direct sulfonation by sulfuric acid either of benzoyl chloride or of benzoylsulfuric acid from the observation that neither ethyl benzoate nor benzoic acid is sulfonated by sulfuric acid a t our temperature. We consider the comparison applicable because of the similarity of the electronic environment in the several compounds of the carbon atom attached to the benzene ring. Ethyl benzoate reacts with sulfuric acid only by cleavage to benzoic acid, in consonance with the scheme of Newman, et aL4 The difference between our results and those of Elliott, et al., is presumably to be sought in the difference in conditions. They heated purified benzoylsulfuric acid, which might undergo an intramolecular rearrangement-a self-sulfonation constrained to reaction a t the ortho position by steric requirements-as indeed suggested by Elliott, et aL2 Our conditions evidently cause intermolecular sulfonation (involving a sulfonating agent other (1) (a) A. Oppenheim, Ber., 8 , 735 (1870); (b) Ador and Oppenheim. i b i d . , 738. (2) G. A. Elliott, L. L. Kleist. F. J. Wilkins. and H. W. Webb, J . Chem. Soc.. 1219 (1926). (3) Elliott, at d.,ascribed a temperature of 150°, which Oppenheim did not mention. ( 4 ) M. S. Newman, R. A. Craig, and A. B. Garrett, J. Am. Chem. Soc., 71, 869 (1949).



than sulfuric acid). Since alkanoylsulfuric acids act as sulfonating agent^,^ it is reasonable to ascribe a similar power to aroylsulfuric acids. The greater electronegativity of the aroyl group than of a proton suggests progressively greater sulfonating power in the series sulfuric acid, aroylsulfuric acids, diaroyl sulfates. Identification of m-sulfobenzoic acid rests on two types of evidence: (a) negative results for attempted reactions known to occur with o-sulfobenzoic acid, L e . , formation of an anhydride2s6and of Phenol and (b) direct evidence: identity of its bis(S-benzylisothiouronium) salt with an authentic sample, alkaline fusion to m-hydroxybenzoic acid, and identity of the infrared spectrum with that of authentic m-sulfobenzoic acid. Under our conditions rearrangement during alkaline fusions does not occur.9 9 1 0 Minor impurities of o-sulfobenzoic acid probably are present; traces of p-sulfobenzoic acid may also be present. Under the influence of the initial hypothesis of an ortho rearrangement we wished to determine the effect of blocking groups in the ortho positions. Neither 2,6-dimethylbenzoyl chloride nor 2,4,6-trimethylbenzoyl chloride gave with cold sulfuric acid visible evidence of reaction comparable to the precipitate formed by benzoyl chloride. Furthermore in each of these cases the formation of a complex or of a mixed anhydride is belied by the fact that 110 major portion of the acyl moiety is extracted into the sulfuric acid phase nor of the sulfuric acid into the organic phase. (Cloudiness of the carbon tetrachloride solutions indicates that a t least some of the sulfuric acid titrated for in that layer was mechanically carried over.) The reaction mixture from 2,4,6-trimethylbenzoyl chloride and sulfuric acid upon heating under the conditions med for benzoyl chloride once yielded a material which remailled unidentified aiid could not be recarystallized, but which, on the basis of solubility (*h;lracteristics, obvioudy was iiot a monomeric sulfonic acid. The absence of mi absorptioii inaximum near 3.8 p indicates the 103s of thc rurboxyl group. Experimental

For the reactions with sulfuric acid all glassware was ovendried. Infrared spectra were determined of material contained in potassium iodide pellets, using a Perkin-Elmer Infracord. Reaction of Benzoyl Chloride with Sulfraric Acid. (A) Equimolar Amounts.-Typically 10.0 ml. (0.087 mole) of benzoyl chloride [b.p. 104-105’ (66 mm.)] was placed with 10 ml. of carbon tetrachloride*l (distilled from phosphorus ( 5 ) A. J. van Peski. Rec. trau. chim., 40, 103, 736 (1921). (6) I. Remsen and A. R. L. Dohme, Am. Chem. J., 11, 332 (1889). (7) A. I. Vogel, “A Textbook of Practical Orgsnia Chemistry.” 3rd ed., Longmans, Green and eo., Ltd.. London, 1956, p. 990. (8) Cf.,e a , Paul Karrer, “Organic Chemistry.” 4th English ed.. Elsevier Publishing Co.. Inc., Amsterdam, 1980, p. @9. (9) J. N. RAY and M. L. Dey, J . Chem. Soc., 117, 1405 (1920). i10) J. S. Reese, J . Am. Chcm. Soc., S4, 2009 (1932). (11) The carbon tetrachloride can be omitted withouk affrcting the rewlt.



pentoxide) into a three-necked round-bottomed flask fitted with a stirrer and a compensating dropping funnel and connected to a drying tower. Then 4.6 ml. (0.086 mole) of sulfuric acid (assay 99.3%) was dripped in. The reaction mixture gradually became yellow, then yellow-green, and, when about one half of the acid had been added, a white precipitate formed. h stream of dry air was sucked through the flask for 2 to 4 hr. The carbon tetrachloride (if any) wm distilled, then the reaction mixture heated for 24 hr. a t 100’; a dark red sirup remained. Warm water, 125 ml., was added and any undissolved benzoic acid filtered off and washed; the combined filtrate and washings were evaporated to 60 ml., let cool, and any additional benzoic acid was filtered off and washed. Sulfuric acid was precipitated with excess aqueous saturated barium hydroxide, then the barium removed by an exactly equivalent amount of 6 N sulfuric acid. The filtered solution was extracted with ether, then evaporated to dryness, and the off-white powder remaining was dried in a vacuum desiccator over concentrated sulfuric acid.12 The product, obtained in a variable yield in the vicinity of 45%, melted over a wide range, the upper limit of which w m not over 140’. (B) Excess Benzoyl Chloride.-By the same procedure (without carbon tetrachloride) 10.0 ml. (0.087 mole) of benzoyl chloride reacted with 2.3 ml. (0.043 mole) of sulfuric acid. The product, which was water-soluble, showed the same infrared spectrum as the product from procedure A. Attempts at Sulfobenzoic Anhydride. ( A ) Method of ElliotLLSulfobenzoic acid, 2.8 g. (0.014 mole), was refluxed with 3.3 g. (0.042 mole) of acetyl chloride for 2 hr., then the volatile material was distilled. The residue was thrice washed with dry benzene, the residual benzene distilled, then the residue exposed to moist air. A white film, too slight to collect, formed. (B) Method of Remsen and Dohme.6-Sulfobenzoic acid, 1.0 g. (0.005 mole), was heated by a wax bath in a three-necked round-bottomed flask in a stream of dry air at 115-130’ for 5 hr. With a 20-em. long tube 2 cm. in diameter, topped by a drying tube, extending into the flask to the level of the external wax bath, the acid was then heated with phosphorus pentoxide for 2 days at 130’. -4very slight film formed at the top of the flask; nothing collected in the tube. Attempt at Phenolsulfonephthalein.2~7-Sulfobenzoicacid, 0 . i g. (0.0035 mole), was heated with 0.7 g. (0.007 mole) of phenol and 1.0 g. (0.007 mole) of zinc chloride. Insufficient dye formed to be collected, but one drop of aqueous 20% sodium hydrovdc added to the reaction mixture produced a red color. S-Benzylisothiouronium Salts of In-Sulfobenzoic Acid.-Sulfobeneoic acid, 1.0 g. (0.005 mole), was dissolved in 10 in!. of water w d heated to boiling. To the hot solution 10 nil. of aqueous 10L/, S-benzylisothiouronium chloride wa5 added. Upon gradual cooling to room temperature nothing separated; at 0’ a finely dispersed oil separated and resisted attempts a t crystallization.13 Upon first attempt the method of VeibeP4also gave only an oil. The latter method applied to authentic ni-sulfobenzoic acid again produced an oil, which began crystallizing after 4 days in a refrigerator and, after being left a t room temperature for 7 hr. and returned to the refrigerator, could be filtered off crystalline after a t o t d of 8 days. Recrystallization from aqueous ethanol again produced an oil which eventually crystallized, but which had a very wide melting range. Subsequent treatment of sulfobenzoic acid (from reaction of benzoyl chloride with sulfuric, acid) gave the same results, except that the original oil began (12) The method here described of isolating sulfobeneoio acid, despite the necessity of repeated filtration in order to see the end point of the barium sulfate precipitation, is less tedious than either of two alternative methods, which furthermore gave smaller yields. (13) Cf.E. Canlpalgne and C. 11. Suter, J . Am. Chem. Sor.. 64, 3040 (1942). (14) S. V e h r l , J. Am. Chem. Soe., 67, 1867 (1945).

JULY, 1962



crystallizing within 1 day and could be filtered off after 3 mediately mixed with 292 g. (2.45 mobs) of potassium brodays. mide and pyrolyzed by being added in portions to a threeAuthentic m-sulfobenzoic acid provided, by the method of necked round-bottomed flask fitted with a reflux condenser. Friediger and Pedersen,lS the bis(S-benzylisothiouronium) The reaction mixture was well extracted with benzene; the salt as an oil which crystallized partly within 2 days and txoduct was collected with b.0. 127' (88 mm.); yield 11.8 completely within 4 days, m.p. 136-137' (cor., lit.,16 130g. (0.064 mole). 2.6-Dimethvlbenzoic Acid.-The method of BowenZswaa 133'). Subsequently sulfobenzoic acid by the same method yielded an oil which crystallized completely within 8 hr., used. The carbonated reaction mixture never became "granular" and was acidified while sticky. Based on 2m.p. 135-138'; mixed m.p. 135-138'. m-Hydroxybenzoic Acid.-In a nickel crucible 1.O g. bromo-l,3-dimethylbenzenethe yield was 5&58y0 of white (0.005 mole) of sulfobenzoic acid was fused with 9.0 g. (0.16 powder, m.p. 115-116' (cor., lit.,2B116'). 2,6-Dimethylbenzoyl Chloride.-2,6-Dimethylbenzoic mole) of potassium hydroxide; the gently heated melt was stirred with an iron file for 10 min., during which time a gas acid was refluxedfor 2 hr. with excess thionyl chloride. Prodwaa evolved. The melt was cooled, dissolved in water, and uct was collected with a b.p. 108-110' (30 mm., lit.,27 the solution acidified with 50% sulfuric acid. Potassium 217') in 767, yield. Attempted Reaction of 2,6-Dimethylbenzoyl Chloride with sulfate was filtered off.18 The filtrate was extracted with ether, and an acid isolated, m.p. 170-180'. Recrystalliza- Sulfuric Acid.-In the apparatus described for the reaction tion from water gave a 30% yield of off-white granular crys- of benzoyl chloride with sulfuric acid, 1.15 g. (0.0069 mole) tals, m.p. 197' (cor., lit.,'? 200.8'), with a spectrum corre- of 2,6-dimethylbenzoyl chloride was dissolved in 10 ml. of sponding between 5 and 10 p to that published1*for m-hy- carbon tetrachloride, then 0.36 ml. (0.0067 mole) of 99.3% droxybenzoic acid in this region. The infrared spectra of sulfuric acid was slowly added. After 30 min. of additional known salicylic acid and of known p-hydroxybenzoic acid stirring the reaction mixture was let stand for 1.5 hr. The differ at a number of wave lengths. not quite clear carbon tetrachloride layer was pushed by dry m-Sulfobenzoic Acid.-By the method of Reeselo m- air pressure through a glass tube into another flask, fitted sulfobenzoic acid was obtained in 70% yield as a light tan with a drying tube. The yellow sulfuric acid phase was solid with a wide melting range, the upper limit of which was washed three times with carbon tetrachloride, which was 140'. The infrared spectrum was identical with that of the removed in the same manner. Evaporation of the crtrbon sulfobenzoic acid obtained by reaction of benzoyl chloride tetrachloride while exposed to air left' a white solid which, with sulfuric acid. after thorough washing with water and drying, weighed 0.63 Attempted Sulfonation of Ethyl Benzoate with Sulfuric g. (0.0042 mole), m.p. 114' (cor.). The water washings Acid. (A) Equimolar Amounts.-In the apparatus described indicated 1.47 meq. of acid. From the sulfuric acid phase for the reaction of benzoyl chloride with sulfuric acid 10.0 ml. 0.20 g. (0.0013 mole) of 2,6-dimethylbenzoic acid, m.p. 113', was precipitated by dilution with water. Titration of (0.070 mole) of ethyl benzoate [~PD 1.5069 1.50682)] was heated with 3.6 ml. (0.068 mole) of 99.3% the filtrate showed 11.43 meq. of acid. 2,4,6-Trimethylbenzoic Acid.-By the method of Bowenz5 sulfuric acid at 100" for 24 hr. After adding water 4.3 g. (0.036 mole) of benzoic acid and 4.5 g. (0.030 mole) of ethyl yields of 5041% were obtained, m.p. 152.4-153.4' (cor., benzoate were isolated. lit.,z5 153.4-154.4'). 2,4,6-Trimethylbenzoyl Chloride.-By the method of (B) Excess Sulfuric Acid.-In the same apparatus 12.51 g. (0.0833 mole) of ethyl benzoate (C.P., The Coleman and Barnesz8 yields of 27-86y0 were obtained, b.p. 128-130' Bell Co.) and 42.0 g. (0.425 mole) of 99.3y0 sulfuric acid [40 mm., lit.,28143-146' (60 mm.)]. were heated a t 99-104' for 42 hr. After pouring the reaction Attempted Reaction of 2,4,6-Trimethylbenzoyl Chloride mixture into 150 ml. of water 10.02 g. (0.0820 mole) of with Sulfuric Acid.-By the same method as for the attempt benzoic acid was isolated. with 2,6-dimethylbenzoyl chloride, 1.10 g. (0.0060 mole) of 2-Bromo-1,J-dimethylbenzene. (A) Sandmeyer Reac- 2,4,6-trimethylbenzoyl chloride was treated with 0.31 ml. tion.-By a modification of the method of Hartwell20 20.0 (0.0058 mole) of 99.37, sulfuric acid. 2,4,6-Trimethylbenml. (0.16 mole) of 2,6-dimethylaniline (Eastman Kodak) was zoic acid was obtained, 0.8 g. (0.0049 mole) from the carbon diazotized, then subjected to the Sandmeyer reaction: The tetrachloride phase and 0.1 g. (0.0006 mole) from the sulfuric diazonium salt, which separated as a yellow foamy solid, acid phase. The wash water from the former portion of was spooned in portions into the solution of cuprous bromide, 2,4,6-trimethylbenzoic acid contained 1.86 meq. of acid. then the residual solution added. Product was collected Reaction of 2,4,6-Trimethylbenzoyl Chloride with Sulfuric with b.p. 200-205" M . 206'); yield 4.7 a. (0.025 Acid. (A).-In a manner analogous t o the reaction of mole ). benzoyl chloride with sulfuric acid 0.58 g. (0.0032 mole) of (B) Via Mercuric Bromide Comple~.~~-2,6-Dimethyl-2,4,6-trimethylbenzoyl chloride, dissolved in carbon tetraaniline, 37.5 g. (0.310 mole), was diazotized as above, then chloride, was treated with 0.19 ml. (0.0034 mole) of 99.3y0 54.1 g. (0.150 mole) of mercuric bromide added either as an sulfuric acid. Distillation of the carbon tetrachloride left a aqueous suspensionz3or dissolved in aqueous potassium bro- homogeneous brassy yellow liquid, which was heated for 24 mide.24 The pale yellow precipitate was filtered off and hr. at 100'. A chunk of brown solid remained. This solid washed with water, acetone, and ether. The dry complex dissolved in concentrated sulfuric acid and was recovered, obtained by the latter method in a yield of 78.9 g. was im- evidently unchanged, upon dilution with water; it dissolved readily in cold cyclohexene; it was insoluble in the following (15) A. Friediger and C. Pedersen. Acta Chem. Scand., 9, 1425 solvents a t their respective boiling points: water, aqueous (1955). 10% sodium hydroxide, absolute ethanol, methanol, ether, (16) Identified b y flame test and formation of an acid-insoluble precipitate with aqueous barium hydroxide. (17) H. Offermann, Ann., 280, 1 (1894). (18) R. B. Barnes, V. Liddel, and V. Z. Williams, Ind. Eng. Chem., Anal. E d . , 16, 605 (1943). (19) K.Auwera and F. Eisenlohr, J . Prakt. Chem., 121 82, 65 (1910). (20) J. L. Hartwell, Org. Syntheses, Coll. Vol. 111, 185 (1955). (21) 0.Jacobean and W. Deike, Ber., 20, 903 (1887). (22) H. C. Brown and M. Grayson, J . A m . Chem. SOC.,16, 20 (1953). (23) M. S. Newman and P. H . Wise, J. A m . Chem. Soc., 65, 2887 (1941). (24) H.-W. Schwechten, Ber., 66, 1605 (1932). This niethod proved more successful.

(25) D. M.Bowen, Org. Syntheses, Coll. Vol. 111, 553 (1965). 126) W.A. Noyes, Am. Chsm. J . , 20, 789 (1898). (27) I. &I. Heilbron and H. M. Bunbury, eds.-in-chief. "Dictionary of Organic Compounds," Vol. 1. Oxford University Press, New York, 1943, p. 895. No primary source for this value was found, either among the references cited by Heilbron and Bunbury for 2.6-dimethylbenaoia acid or by searching Beilstein's "Handbuoh der organischen Chemie," 4th ed. plus two supplements, Springer-Verlag. Berlin, Chem. Abstr. indexes until June, 1960, and E. H. Huntress, "The Preparation, Properties, Chemical Behavior, and Identification of Organic Chlorine Compound.%,"J. Wiley and Sons, Inc.. New York, 1948. (28) R . P. Barnes, Org. Synthesea, Coll. Vol. 111, 556 (1955).



dioxane, benzene, toluene, cyclohexane, carbon tetrachloride, chloroform, dichloromethane, glacial acetic acid, and nitromethane. After removal of nitromethane by boiling to dryness, the recovered brown solid no longer dissolved in cyclohexene. No material could be recovered from cyclohexene solution by dilution with chloroform or with carbon tetrachloride. The spectrum between 2.5 and 15 p of the brown solid showed absorption maxima at 2.85 (weak), 3.3, 6.2, 6.35 (shoulder), 6.85, 7.2, 7.9, 8.4 (broad), 9.2 (very broad), 9.6 (broad), 11.4 (weak, broad), 11.65, 12.25 (broad), 13.0-13.2 (shouldex), and 14.2-14.5 p (weak, very broad). (B).-The same reaction was run with 5.5 g. (0.030 mole) of 2,4,6-trimethylbenzoyl chloride and 1.7 ml. (0.031 mole) of 99.3% sulfuric acid. The residue, after heating, wm dissolved in water, and the sulfuric acid removed as after reaction of benzoyl chloride with sulfuric acid. Evaporation of the water left a brown tacky sirup, which did not solidify even after several weeks in a vacuum desiccator over concentrated sulfuric acid.

Selective Etherification of p-Hydroxybenzyl Alcohol

VOL.27 Experimental

p-Methosymethylphenol. Zeo-Karb 225 exchange resin in the hydrogen form was thoroughly washed with water,and then methanol. A solution of 1.7 g. of p-hydroxybenzyl alcohol' in 9 ml. of pure methanol was allowed to stand with 3.4 g. of the resin for 12 hr. at room temperature. The resin was separated, washed with methanol, and the combined filtrate and washings evaporated in vacuo. The oily residue soon crystallized, and was washed with cold water and dried. Recrystallization from benzene gave the phenol as compact prisms, yield 1.1 g. (60%), m.p. 81.582.5', ( k e m . p . 82.5-83.5"). When the methanol solution was refluxed with the exchange resin for 1 hr., and the mixture worked up as before, a white, water-insoluble, resinous material was obtained. p-Ethoxymethylpheno1.-Treatment of 1 g. of p-hydroxybenzyl alcohol with ethanol and Zeo-Karb 225 in the same way, followed by distillation, gave the phenol aa an oil, yield 0.45 g. (37%), b.p. 118-120°/2 mm., which solidified on standing. It was recrystallized from benzene-petroleum ether, map. 50-51" (lit. m.p. 50-51';' 56.5-57.502). (4) p-Hydroxybenzyl alcohol wa8 obtained by the lithium aluminum hydride reduction of ethyl p-hydroxybeneoate in the usual way.6 (5) R. F. Nystrom and W. G . Brown, J . A m . Chem. Soc., 69, 1197 (1947). ( 6 ) H. Mitawa, Bull. Chem. Soc. Japan, 17, 53 (1954).

F. H.C. STEW ART^ Chemical Research Laboratol.ies, Commonwealth Seientijk and Industrial Research Organization, Melboume, Awrtralia Received December $8, 1981

Dehydrogenation of Alcohols by Lithium Metal-Ethylenediamine System SHELBERT SMITH, RICHARD THOMPSON,~ AND E. 0. WOOLFOLK

Preparation of the simple p(and &)-alkoxyDepartment of Chemistry, Central State College, methylphenols has generally been attended by Wilberforce, Ohio difficulties on account of their strong tendency to resinify on heating, particularly in the presence of Received December 26,1961 acids or bases. Substituted alkoxymethylphenols are rather more stable in this respect, and it was Previous workers293 have shown that lithium noted early by Auwers and Bauma that these metal in ethylenediamine and in aliphatic monoethers were formed with great ease by merely amines can reduce aromatic compounds to monoheating a substituted hydroxybenzyl alcohol with olefins. It has been shom2J also that phenol can the required etherifying alcohol. This procedure be reduced to cyclohexanone in the presence of was extended to unsubstituted 0- and p-hydroxylithium metal in amines. I n unpublished work benzyl alcohols by de Jonge and Bibo12who showed from our laboratories, me have found that lithium that the reaction proceeded without appreciable metal in ethylenediamine will reduce a variety of resinification by heating the mixture in a sealed phenols to saturated and unsaturated ketones ancl tube at 150' for several hours. alcohols. It has now been found that selective etherificaThe proposed mechanism2b for this reduction tion of phydroxybenzyl alcohol can be carried reaction suggests a stepwise 1,&addition of two out by treating its solution in an alcohol with a moles of hydrogen to the aromatic ring. The addistrong acid cation-exchange resin a t room tempera- tion of a third mole of hydrogen to the isolated ture. In this way p-methoxymethylphenol and double bond is so slow that the monoolefin may be the corresponding ethyl ether have been conven- isolated. It is reasonable to predict from this iently prepared in moderate yield on R small proposed mechanism that the ketones from the scale. When a methanol solution of p-hydroxy- reduction of phenols arise from the isomerization benzyl alcohol was refluxed with the exchange resin, of an enol intermediate. The alcohol products however, rapid polymerization occurred with formation of a resinous product. (1) American Chemical Society-Petroleum Research Foundation (1) Present address: Department of Organic Chemistry, Weizmann Institute of Soienoe, Rebovoth, Iarael. (2) J. de Jonge and B. H. Bibo, Re. trap. ohim., 74, 1448 (1955). (3) K. Auwers and F. Baum, Ber., 19, 2329 (1896).

Undergraduate Fellow, 1961-1962. (2) (a) R. A. Benkeser, R. E. Robinson, D. M. Sauve, and 0. H. Thomas. J . Am. Chem. Soc.. 76, 631 (1954); (b) R. A. Benkeaer. C. Arnold, Jr., R. F. Lambert, and 0. H. Thomas, kbid.. 77, 6042 (1956). (3) L. Reggel, R. A. Friedel, and I. Wender. J . Ow. Chem., 11, 891 (1957).


JULY, 1962


TABLEI ~ - 1 L 1 . p . OC., . 2,4-D---

Carbonyl product


% yield



Cyclohexanone 54 157-158.5 (162) Cyclohexanolb 2-Methylcyclohexanone 38 134.5-135.5 ( 137) 2-Methylcyclohexanol n-Valeraldehyde 66 103-105 (106) n-Pentyl alcohol ZHexanol %-Butyl methyl ketone 45 103-106 (106) Phenylacetaldehyde 25 118-120 (121) 2-Phenylethanol Acetophenone 24 243-244 (250) 1-Phenylethanol 80 234-236 (237) Benzyl alcohol Benzaldehyde a R. A. Shriner, R. C. Fuson, and D. Y. Curtin, “Systematic Identification of Organic Compounds,” 4th ed., John Wiley and Sons, Inc., New York, N. Y., 1956. Another run in which 0.01 mole of lithium ethoxide was added to the reaction increaaed the yield of carbonyl compound to 76%.

could arise from the reduction of the ketones or by the direct reduction of the aromatic ring of the phenol. This reducing system has been utilized in our studies on the synthesis of terpene compounds. I n our studies, we have found that primary and secondary aliphatic and aromatic alcohols can be dehydrogenated to carbonyl compounds in the presence of lithium metal in ethylenediamine. The results of our investigation are shown in Table I. Verification of the products of the dehydrogenation reactions was by gas chromatography. Confirmatory evidence was obtained by the preparation and determination of the melting points of the 2,P dinitrophenylhydrazone derivatives of the carbonyl products. Many references may be found in the literature on the dehydrogenation of alcohols; however, most of these reactions involve the vapor phase dehydrogenation of alcohols over a metal oxide catalyst such as copper oxide. Recently, a similar use of the lithium metal-ethylenediamine system has been reported for the dehydrogenation of certain dienes to yield aromatic hydrocarbon^.^ These authors propose that N-lithioethylenediamine is the active dehydrogenation agent. To our knowledge, this is the first reported use of the lithium metal-ethylenediamine system for the dehydrogenation of alcohols. Experimental

All alcohols used in this work were purchased aa C.P. grade chemicals and were used without further purification. General Procedure.-Lithium Ribbon (0.8 g.-atom) cut in 1-in. strips waa added to 77-100 ml. of ethylenediamine in a three-neck flask fitted with a stirrer and reflux condenser. A slow stream of nitrogen was admitted during the addition of lithium metal. The lithium metal-ethylenediamine solution waa heated and stirred until all the lithium metal had dissolved, and no more gas evolution waa observed. The appropriate alcohol (0.3 mole) was added slowly to the solution. After all the alcohol had been added the solution was refluxed 2-3 hr. The reaction mixture was cooled to room temperature and poured over ice. The solution was neutralized with solid carbon dioxide or hydrochloric acid. The organic material waa extracted with ether. The products were then recovered from the ether extract by distillation. The analysis of the reaction products WBB by gaa chromatography a t 150’ using a 6 ft. X ‘/,-in. column packed (4)

L. Regeel, 8. Friedman,and I. Wander, J . Or#. Chum., 88,



with UCON 50 HB 2000 aa the partitioning agent. The bands were identified by comparison of retention times with those of authentic samples. Quantitative determinations were made by compming the area of the band of the carbonyl product and unchanged alcohol from the reaction mixture with the area of the band given by a known amount of an authentic sample of the respective carbonyl compound and alcohol.

Acknowledgment.-This research was supported by a Grant from the Petroleum Research Fund administered by the American Chemical Society. Grateful acknowledgment is hereby made to the donors of said fund.

Cyclization of N-Phenylcarbamates of Ethynylcarbinols KEIITISISIDO,KEISTJKE HTJKUOKA, MINORU TUDA, AND HITOSI NOZAKI

Departmen’ of Industrial Chemistry, Faculty of Engineering, Kyato University, KyGto, Japan Received December B7, 1951

In an attempt t o dehydrate ethynylcarbinols (I) to enyne compounds (11), carbinols IB and I C were treated with phenyl is0cyanate.l Under conditions as specified in the experimental part, however, crystalline products were isolated in both cases. The analyses agreed with the values calculated for the N-phenylcarbamates of these carbinols. Although the product similarly obtained from propargyl alcohol (IA) was found actually to be the corresponding N-phenylcarbamate (IIIA), those obtained from 1-ethynylcyclohexanol (IB) and ethynylcarbinol of dihydro-p-ionone (IC) did not show the characteristic infrared absorptions of the acetylenic and imide hydrogens. The constitution IVB was assigned to the reaction product from IB, as the ozonolysis of IVB afforded a crystalline imide which was identical with 3 - phenyl - 5,5 - pentamethylene - 2,4 - oxazolidinedione (VI) prepared independently from ethyl 1-hydroxycyclohexanecarboxylate (VII) as shown in the accompanying flow sheet. The (1) For the dehydration with phenyl isocyanate, see W. J. Bailey and F. C a r e , Angow. Chsm., 71,470 (1959).






R' I1 IA, IIIA, IVA: R-CHz-=R'=H

/ \







R- CH2- C-C 'CH, IV




OH R-CHz-&--C=CH RI ' I

VOL. 27

\ /

CHz- CH2/"\

V IC, IVC: R = H , R'= &H2-cHz-


o("-T=" C-N-CBHs





~ C O , , , VI11








NUR spectrum2 of IVB had two apparent singlets a t r 8.4 and 6.0 in an intensity ratio of 5 : l . The former absorption is ascribed to the pentamethylene protons and the latter to the olefinic protons. A possibility of the cyclization product being a six-membered product (V) is excluded on the basis of the negligible coupling constant of olefinic protons3 as well as of behavior on the ozonolysis. We would be safe to assume the constitution IVC for the product from ethynylcarbinol of dihydro-pionone in view of the analyses and infrared absorptions which are quite similar to IVB. When propargyl N-phenylcarbamate (IIIA) was treated with potassium hydroxide, isomerization occurred to afford a cyclization product similarly lacking the infrared absorptions due to acetylenic and imide hydrogens. The S M R spectrum of this compound afforded an evidence supporting the constitution IX instead of the expected IVA. h doublet and a quartet were observed at r 8.1 and 3.4, respectively, in an intensity ratio of 3 : 1 and with a coupling constant of 2.0 C.P.S. The former absorption is ascribed to methyl protons and the latter to olefinic proton. This product (IX) was subsequently hydrogenated over palladium to 3-phenyl-4-methyl2-oxazolidinone (X), whose NMR spectrum could again be reasonably interpreted on the basis of the structure given. Methyl absorption appeared as a doublet at 8.9 with a coupling constant of 5.0 (2) For N M R measurements and chemical shift presentation, see Experimental. (3) For examples of low coupling constants of vinyl methylene protons, see J. N. Shoolery in the NMR-EPR Staff of Varian Associates, "NMR and E P R Spectroscopv," Pergamon Press, Oxford, 1960, p. 114.



c.P.s., methine proton as a multiplet at r 6.3. methylene protons as a polyplet at r 5.7, and aromatic protons as a multiplet at r 2.8, all these absorptions being observed in an intensity ratio of 3 : 1:2:5 . Regarding the mechanism of the formation of IX, it would be judicious to assume the cyclization of IIIA to IVA and the following transposition of (sxo-cyclic double bond to an endo-cyclic one. Experimental6 NMR Measurements and Presentation.-Spectra were obtained for IVB on a 40-Mc. instrument in chloroform solution, for I X on a 56.4-hfc. instrument in chloroform solution, for X on a 56.4-Mc. instrument in carbon tetrachloride solution with water as an external reference, and for impure IVA containing I X on a 40-Mc. instrument in carbon tetrachloride solution with cyclohexane as an external reference. Chemical shifts actually observed are: for IVB at 5.65 (wt. 5 ) and 3.30 p.p.m. (wt. l), higher fields from chloroform peak; for I X at 5.41 (wt. 3) and 0.64 p.p.m. (wt. l ) , higher fields from chloroform peak; for X at 3.57 (wt. 3), 0.96 (wt. l ) ,and0.36p.p.m. ( a t . 2 ) , higherfieldsandat 2.55 p.p.m. (wt. 5), lower field from external water reference; and for impure IVA a t 0.66 (wt. 0.44), 2.89 (wt. 2), 3.73 (wt. 2), and 5.40 p.p.m. (wt. 0.15), lower fields from external cyclohexane peak. These chemical shifts were converted (4) Attempted isolation of IVA afforded in one experiment a crystalline produot, m.p. 97-9S0, whose NMR spectrum showed two multiplets of equal intensity centering a t T 5.9 and 5.0, respeotively. besides absorptions at T 8.1 e n d 3.4 indicating the contamination with 13% of IX. The former absorptions should be ascribed to olefinic protons of em-cyclic double bond and allylic protons of IVA. This product, however, could not be obtained in numerous experiments repeatedly carried out thereafter, supposedly due to the facile rearrangement of IVA t o I X and to some unknown factors catalyzing the transposition of the double bond. ( 5 ) AI1 temperatures are uncorrected. Elenieutal analyses werc carried out by Miss Kenko Ogawa.



into 7 values given above, assuming T values of 2.73 for internal (solvent) chloroform peak, 5.35 for external water reference, and 8.77 for external cyclohexane reference, for the sake of convenience in comparing shift values. 4-M ethylene-5, S-pentamethylene-3-phenyl-2-oxazolidinone (IVE3) from 1-Ethynylcyclohexanol (IB).-To a solution of 3.5 g. of IB6in 40 ml. of toluene, 3.7 g. of phenyl isocyanate was added and the mixture was heated under reflux for 4 hr. and then the solvent was removed under reduced pressure. The residue was diluted with about 40 ml. of petroleum ether (b.p. 30-60') and this solution was extracted with 5% methanolic potassium hydroxide solution. The methanol layer was separated, neutralized with dilute hydrochloric acid, and extracted again with petroleum ether (b.p. 30-60'). Concentrating this petroleum ether solution, followed by two recrystallizations of the residue from petroleum ether (b.13. 30-60"), gave 6.4 g- . .(88%) - of the oxazolidinone, m.p. 1-67-167.5'. Anal. Calcd. for C16H1&02: C, 74.05; H , 7.12; N, 5.76. Found: C, 74.05: H, 7.04: N , 5.83. Infrared absorptions (Nujol): 1750 (-0-CO-KC=C 100; IIa (DED - salt) > 1001. Experimental 6-Chloropenicillanic Acid (Ia).-A solution of 6.5 g. of 6-APA in 100 ml. of 1 N hydrochloric acid waa cooled to 0-2" and a solution of 2.5 g. of sodium nitrite in 20 ml. of water waa added dropwise. One hour after the completion of the addition, the temperature was allowed to rise to 15-18' and the separated oil extracted with ethyl ether. The organic layer waa dried over sodium sulfate and the ether evaporated in vacuo at room temperature. The oily residue, 4.5 g., showed in the infrared spectrum bands a t 1725 cm.-l (C=O stretching of carboxylic group) and at 1770 cm.-l ( C 4 stretching of B-lactam fused to thiazolidine ring) in agreement with the structure of Ia. This oil, which decomposed under distillation, was dissolved in ethyl ether and treated with an ether solution of dibenzylethylenediamine; 5.5 g. of Ia dibenzylethylenediamine salt was obtained, which, after several crystallizations from aqueous ethanol, melted at 159-160'; [CY]D +154.4 (c0.5%, in methanol). The infrared spectrum is reported in Fig. 1. Anal. Calcd. for C32Hi10C12N40BS~: C, 54.00; H, 5.67; N, 7.88; S, 9.01; C1, 9.97. Found: C, 53.75; H, 6.19; N, 7.65; S,8.99; C1, 10.24. 6Bromopenicillanic Acid (Ib).-was prepared essentially in the same way described for Ia, starting from 4.32 g. of 6-APA difiEQlVedin 50 ml, of 2.5 N sulfuric acid containing

A~kylfluorophenylcarbinols with Choleretic Activity N. P. Bwu-HOT,N. D. XUONG, B . K. D I ~ PAND , N. N. QUANG Institut de Chimie des Substances Naturelles du C.N.R.S., Cif-sur-Yvette (8.-et-0), Francs Received January 16, 1962

p-Tolylmethylcarbinol, a component of the essential oil of Curcuma domestica, is known to show potent choleretic activity,l and the same properties have been found in several of its homologs2 and in similar heterocyclic carbinol^.^ It is also known that methyl groups can be replaced by halogens without qualitatively affecting pharmacological activity. Hence, it was of interest to investigate whether alkylfluorophenylcarbinols would likewise display choleretic activity; fluorine radicals were preferred to other halogens since the solidity of the C-F bonds in aromatic molecules renders biochemical dehalogenation more difficult. We now report the preparation, for biological evaluation, of a wide series of such fluorinated carbinols, by reaction of the appropriate alkylmagnesium bromide or iodide with o-, m-, and p-fluorobenzaldehyde; in this series, only methyl-p-fluorophenylcarbinol was known.4 All the carbinols thus obtained in 80-90% yield, were colorless oils with an aromatic odor that was pronounced for the lower terms and less marked for the higher terms. Several arylalkyl(1) H. Dieterle and P. Kaiser, Arch. Pharm., 270, 413 (1932); 271, 237 (1933). (2) H. Langecker, A. Harwart, and K. Junkmann. Arch. e z p t l . Pothol. Pharnzakol., 226, 303 (1955); R. Engelhorn. ArzneimzttelForech., 10, 255 (1960): N. P. Buu-Hoiand N. D. Xuong, Compf.rend., 249, 970 (1959). (3) N. P. Buu-HOY, N. D . Xuong, and B. K. DiCp, J. Org. Chem., 26, 1673 (1961). (4) G. 01&h,A. Pavlhth, and I. Kuhn, Acto Chtm, Aced, Sci, [email protected],, 71 06 (1966).




TABLE I Carbinol

Ethyl-p-fluorophenyl Propyl-a-fluorophenyl Isopropyl-p-fluorophenyl Butyl-p-fluorophenyl Isobutyl-p-fluorophenyl Amyl-p-fluorophenyl Isoamyl-p-fluorophenyl Isohexyl-p-fluorophenyl 8-Phenethyl-p-fluorophenyl 7-Phenylpropyl-p-fluorophenyl Ethyl-m-fluorophenyl Butyl-m-fluorophen yl Isohexylm-fluoropheny1 Octyl-m-fluorophenyl Dodecyl-m-fluorophenyl 7-Phenylpropyl-m-fluorophenyl Methyl-o-fluorophenyl Propyl-a-fluorophenyl Amyl-a-fluorophenyl 8-Phenethy 1-a-fluorophenyl

B.P.. .C./mm.


116/17 124/17 115/17 132/17 128/17 148/17 143/17 149/17 177/15 185/15 115/17 131/17 149/17 176/17 214/17 188/17 105/17 118/16 144/16 177/15

1.5004/28' 1.4958/28' 1.4963/28' 1.4901/21.5' 1.4938/28' 1.4953/19' 1.4937/24O 1.4862/26.5' 1.5732/24' 1.5637/25" 1.5039/25' 1.4975/25' 1.4913/25' 1,4830/25' 1.4753/24' 1.5649/25' 1.5087/25' 1.5013/25' 1.5+30/25'

fluorophenylcarbinols, prepared from arylalkylmagnesium chlorides, were also included in our investigation. The carbinols readily underwent dehydration by means of formic acid to give the corresponding ethylenes, including the hitherto unknown 2,3 '-difluorostilbene (I) and 3,4'-difluorostilbene (11). .F F.



In biological tests in rats, the lower terms of the series of carbinols reported herein displayed pronounced choleretic activity. Experimental Preparation of Intermediates.-0-, m-,and p-fluorobenzaldehyde were prepared by the Sommelet reaction from the corresponding fluorobenzyl bromides, themselves obtained from a-, m-,and p-fluorotoluene by side-chain bromination with N-bromosuccinimide in the presence of benzoyl peroxide. Grignard Reactions.-An ethereal solution of a Grignard reagent prepared from the appropriate alkyl iodide or bromide (1.15 moles) was treated portionwise at 0 ' with the aldehyde (1 mole, dissolved in anhydrous ether). The mixture was refluxed for a few minutes on the water bath, and after cooling, treated with an ice-cold aqueous solution of ammonium chloride. The organic layer waa then collected, washed with water, and dried over sodium sulfate, the ether was distilled, and the residue vacuum-fractionated twice. The carbinols thus obtained in 80-90% yield aa colorless liquids, are listed in the table. Dehydration of the Carbinols.-A solution of one part of the carbinol in 5 parts of anhydrous formic acid waa heated for 1 hr. on the water bath. The reaction product waa poured into water, taken up in benzene, and dried over sodium sulfate, the solvent wm removed, and the residue vacuum-fractionated. 1-Fluorophenyl-1-propene,obtained from ethyl-p-fluoroDhenvlcarbinol. wm a colorless fluid liauid. . . b.D. 9Q0/17 mm., n i 4 1.5690. ~ Anal. Calcd. for CoHoF: C, 79.4; H, 6.6. Found: C, 79.6; H, 6.6.

. "



-Carbon, Calcd.


70.1 71.4 71.4 72.5 72.5 73.5 73.5 74.3 78.3 78.7 70.1 76.5 74.3 75.6 77.6 78.7 68.6 71.4 73.5 78.3

70.0 71.1 71.1 72.3 72.2 73.4 73.3 74.0 78.5 78.6 70.0 76.2 74.0 75.5 77.3 78.9 68.5 71.2 73.4 78.5

-Hydrogen, Calod.

7.1 7.7 7.7 8.2 8.2 8.7 8.7 9.0 6.5 7.0 7.1 8.2

9.0 9.6 10.5 7.0 6.4 7.7 8.7 6.5


7.0 7.7 7.8 8.4 8.5 8.7 8.6 9.0 6.8 7.3 7.0 8.3 9.2 9.5 10.8 7.3 6.6 7.9 8.7 6.7

2,3'-Diiluorostilbene (I). a-Fluorobenzyl-m-fluorophenylcarbinol, prepared from o-fluorobenzylmagnesium chloride and m-fluorobenzaldehyde, was a pale yellow oil, b.p. 164'/15 mm.; this underwent dehydration to the stilbene (I), which crystallized from pentane in colorless needles, m.p. 43'. Anal. Calcd. for ClrH1iFg: C, 77.8; H, 4.6. Found: C. 77.7: H. 4.7. . 3,4'--Difluorostilbene (II). p-Fluorobenzyl-m-fluorophenylcarbiiol, prepared aa above from p-fluorobenzylmagnesium chloride, gave on dehydration the stilbene (11) which crystallized from hexane in colorless prisms, m.p. 99'. Anal. Calcd. for C14Hd?2: C, 77.8; H, 4.6. Found: C, 77.5; H, 4.6.

Insect Sex Attractants. 111. The Optical Resolution of dl-10-Acetoxy-cis-7-hexadecen1-01 MARTINJACOBSON Enlomology Research Division, Agricultural Research Service, U.S. Department of Agriculture, Beltsville, Md. Received January 16, 1968

The extremely potent sex attractant secreted by the female gypsy moth (Porthetria dispar) to lure the male has been identified previously as dextrorotatory l0-acetoxy-cis-7-hexadecen-l-ol(I),and the equally attractive optically inactive (dZ-) form has been synthesized.' The dl-form has now been successfully resolved by treating its acid succinate with cbrucine, separating the brucine salts by fractional crystallization from acetone, decomposing the salts, and saponifying the acid succinates with ethanolic alkali. The resulting d- and &forms of I were identical in all respects save optical activity with one another, with the dl-form, and with the (1) M. Jacobson. M. Beroza, and W. A. Jones. Science, 182, 1011 (1960): J . Am. Chsm. Soc., 88,4819 (1961).

JULY, 1903






7-hexadecen-1-01 aa a colorless liquid, [aInD -7.1" ( c 2.0, chloroform), that solidified in the cold but melted again at room temperature. The infrared spectrum* waa identical with those of d- and dl-I.

I natural attractant, and all forma were equally attractive to male gypsy moths when tested in the field atclo-' microgram/trap.z

The Bromination of Ethyl 4-Phenylphenyl Carbonate



Preparation of the Acid Succinate of d2-I.-A mixture of 500 mg. of d2-I, 0.3 g. of recrystallized succinic anhydride, and 0.5 ml. of dry pyridine wm heated at 100" in a sealed tube for 20 hr., cooled, poured into excess 5% hydrochloric acid, and extracted with several small portions of ether. The combined ethereal extract wm washed free of minerd acid with water, extracted with 5T0sodium carbonate, and the alkaline extract was acidified with 20% hydrochloric acid and extracted with ether. The ethereal extract was washed free of mineral acid with water, dried over sodium sulfate, and freed of solvent, giving 530 mg. of orange, viscous liquid that resisted repeated attempts at crystallization. Preparation of the Brucine Salts.-A solution of the acid succinate in 4.5 ml. of acetone was heated to boiling on the steam bath, treated with 1 g. of L-brucine, heated an additional 30 sec., filtered, and cooled in an ice bath with scratching. The white solid that rapidly separated was recrystallized from acetone to give 273 mg. of insoluble brucine salt A, m.p. above 220', [a]*'D -4.7' ( c 2.0, chloroform). The combined acetone filtrate and mother liquors were evaporated to dryness and the residue waa taken up in ethyl acetate, washed with water, and dried over sodium sulfate. Removal of the solvent and crystallization of the white residue from 0.5 ml. of acetone a t -50" gave 245 mg. of brucine salt B, [a]"D -5.3" ( c 2.0, chloroform). Decomposition of the Brucine Salts.-Brucine salt A (270 mg.) was stirred at room temperature with 15 ml. of 20% hydrochloric acid, and the mixture was allowed to stand for 5 hr. and then extracted with ether. The ether extract waa washed free of mineral acid with water, dried over sodium sulfate, and freed of solvent to give 246 mg. of the acid succinate of d-I aa a pale yellow, viscous oil, [a]*'D +22.4" ( c 2.0, chloroform), that failed to crystallize. Decomposition of brucine salt B (240 mg.) in the same manner gave 219 mg. of the acid succinate of 1-1 aa a pale yellow, viscous oil, [a]"D -17.6" (c 2.0, chloroform). Saponification of the Acid Succinates.-A solution containing 246 mg. of the acid succinate of d-I, 10 ml. of 10% sodium hydroxide, and 1 ml. of ethyl alcohol waa heated on the steam bath for 2 hr., cooled, and extracted with ether, and the extract waa washed with water, dried (sodium sulfate), and freed of solvent to give a yellow oil. This oil waa dissolved in petroleum ether (b.p. 40-60") and the solution filtered through a column of 10 g. of cellulose acetate and evaporated to dryness, giving 212 mg. of (+)-loacetoxy-cis-7-hexadecen-1-01 aa a colorless liquid, [a]p ' ~ +7.8" ( c 2.0, chloroform), identical with the natural gypsy moth attractant.' In a similar manner, the saponification of 219 mg. of the acid succinate of 1-1 gave 179 mg. of (-)-10-acetoxy-cis(2) These tests were carried out as described b y J. M. Corliss, Yearbook Aor. (U.S.Dept. Agr.), 694 (1952). The assistance of E. C. Paszek, U.S. Department of Agriculture, Nashua, N. H., in carrying

out these tests is gratefully acknowledged. (3) Determined with a Perkin-Elmer Model 21 spectrophotometer, by means of sodium chloride optics and a 1% solution of the sample in carbon disulfide. Mention of trade names or proprietary products does not necessarily constitute endorsement by the Department of Agrioulture.

Department of Chemistry, Washington State University, Pullman, Washington Received January 16,1968

A number of esters of 4-phenylphenol have been brominated, and the results have been reported previously.' The chlorination2 of 4-phenylphenyl acetate resulted in the formation of 4-(4-chloropheny1)phenyl acetate, although bromination had given a product with the halogen atom a t a position in the molecule ortho to the acyloxy group.' I n an attempt to determine the behavior of another ester of 4-phenylphenol in which the acyloxy group is rather small and should have little steric effect a t ortho positions, the bromination of ethyl 4-phenylphenyl carbonate4has been studied. The results, which are shown in the Experimental, did not indicate that any bromination of the ester occurred ortho to the acyloxy group. Rather, substitution occurred in the ester a t the most remote position in the biphenyl group, as had been observed in most of the previous investigations; there was evidence that some ortho substitution had occurred in free phenols, which were isolated from reaction mixtures. Compounds isolated from reaction mixtures were identified by mixed melting point procedures. The bromination of ethyl 4-phenylphenyl carbonate proceeds more smoothly in carbon tetrachloride than in glacial acetic acid solution. An improved method for the preparation of 2,6-dibromo-4-(4bromophenyl)phenol has been developed. Several new esters of bromine substituted 4-phenylphenols are reported. Experimental Bromine Substituted 4Phenylphenols.-Thee compounds were prepared by methods that have been reported previously, except that 2,6-dibrom64-(4-bromophenyl)phenols was prepared by a modification of a procedure1 that had been used for the preparation of 2-bromo-4-( 4-bromopheny1)phenol. 4-Phenylphenol (15 g., 0.088 mole) waa suspended in 120 ml. of carbon tetrachloride, a trace of iron powder was added, and while the mixture was stirred and (1) 9. E. Hazlet and L. C. Hensley, J . Am. Chem. SOC.,69, 708 (1947) and earlier papers. (2) C. M. 9. Savoy and J. L. Abernethy, ibid.. 64, 2219 (1942). (3) 9.E. Haclet and H. A. Romberg, ibid., 61,3037 (1939). (4) E. Baumgarten, H. G. Walker, and C. R. Hauser, as ~ woiild be predicted for II,4 and desulfurization with Raney nickel5 afforded 3-phenyl-3 H-quinazoline-4one in good yield. We were led to the tentative conclusion that the product of this reaction has the benzo [dlthiazolo[2,3-b]quiriazoline-lI-one structure (VII). This 0


&,xJyJ N





compound has previously been prepared by Bose and PathakGa(from anthranilic acid (VIIIa) and 2chlorobenzothiazole) and by Katz6b (from ethyl anthranilate (T'IIIh) and 2-chlorobenzothiazole) .



C I ( D / KH,




VI11 a. R = H b. R=C*HS

KatzGbhas also described the basic hydrolysis of the amide linkage present in VII to form X-(2-benzothiazoly1)anthranilic acid (IX). VI1


1 0H';heat

H,O L,


Benzo [d]thiazolo[2,3-b]quinazoline-1 I-one (VI) Method 1. This compound was prepared by the method of Ghosh' in a 46y0 yield, m.p. 190-19lo, reported1 184-185'. :3-Phenyl-IH,3H-quina~oline-2,4-dione, m.p. 276-4277" (rrported,*275-277'), was also isolated in a 41% yield. Method 2. The procedure followed was essentially that of Katz.Ob To 5.0 g. of methyl anthranilate, 5.0 g. of 2chlorobenaothiaeole was added, and the mixture was heated cautiously until a vigorous exothermic reaction was initiated. Thereafter the reaction vessel was cooled with a jet of air as necessary to keep the reaction under control. When the reaction had moderated, the crude product was powdered and washed with dilute sodium bicarbonate solution. To ensure complete reaction, the air-dried powder was heated above 200" for 10 min. Two crystallizations from isopropyl alcohol gave 6.5 g. of white needles, m.p. 190-19lo, reported 189°6aand 193°.6b A mixed melting point of this compound prepared by method 1 with that prepared by method 2 was 189.5-191'. The infrared spectra of these two were identical, each showing maxima at 1695, 1590, 767, and 750 ~ m . - l . ~ Ar-( 2-Benzothiazoly1)anthranilicAcid (IX).-This compound was prepared by the method of Katz6b in a 72% yield, m.p. 194-195" (reported, 187'88 and 195-1960sb). Raney Nickel Desulfurization of Benzo[d]thiazolo[2,3-b]quinazoline-11-one (VII).-To a solution of 2.5 g. of VI1 in 200 ml. of absolute ethanol, 10 g. of W-6 Raney nickella was added. The resulting mixture was heated under reflux for 2 hr. The nickel was removed by filtration, and the ethanol by evaporation, to give 1.6 g. of white needles. Two crystallizations (aqueous ethanol) gave 1.1 g. of 3-phenyl3H-quinaaoline-4-one (VI), m.p. 138-139", reported 139011 and 142-142.5" . 3 b

Acknowledgment.-The author is greatly indebted t o Dr. J. Allinger for determining the infrared spectra, and to Drs. J. and N. L. Allinger for helpful discussions. The author also wishes to thank the National Science Foundation for financial support of part of this work. (7) All melting points were determined in capillary melting point tribes and are uncorrected. (8) B. Pawleski, Chem. Ber., 38,130 (1905). (9) These spectra were determined in carbon disulfide solution and recorded on a Beckman Model I R 4 spectrophotometer. (10) H. R. Billica and H. Adkins, Org. Sun., Coll. Vol. 111, 176 (1955). (11) C. Paul and M. Busch, Chem. Ber., 22, 2863 (1899).


The Kat'z preparationBaof VI1 was repeated in a The Preparation of slight'ly modified form and the product was indis16a-Hydroxymethylprogesterone t'inguishable from the compound prepared by t,he method of Ghosh.' The melting points and mixed R ~ I L T O NHELLER,STEPHEN R.I. STOLAR, AND melting points were identical, and the infrared SEYMOUR BERKSTEIN absorption spectra were the same. The hydrolysis Chemical Research Section, Lederle Laboratories, a described by KatzGbwas performed on the two Organic Division of American Cyanamid Company, Pearl River, New (4) F. A. Miller in H. Gilman, "Organic Chemistry, An Advanced Treatise," Vol. 111, John Wiley and Sons, Inc., New York, N. Y., 1953, p. 122. (%(a) R. Mozingo, D. E. Wolf, W. A. Harris, and K. Folkers, J . Am. Chem. Soc., 66, 1013 (1943); (b) R. 0. Robbins, Jr., J. 0. Lampen, J. P. English, Q. P. Cole, a n d J. R. Vaughn, Jr., ibid., 67, 290 (1945); (c) J. A. Zderic, W. A. Bonner, a n d T. W. Greenlee, ibid.. 79, 1696 (1957). (6)(a) P. K. Bose a n d K. B. Pathak, J . I n d i a n Chem. Soc., 11, 463 (1934); (b) L. Kate, J. Am, Chom. Soc., 76, 712 (1958).


Received January 34, 1963

In view of the recent disclosure' of the preparation of 16a-substituted methyl corticoids we wish (1) P. F. B e d and J. E. Pike, J . O7ud Chrm,, 96, 3887 (1961).



VOL. 27

to record here the use of a similar set of reactions to prepare authentic 16a-hydroxymethylprogesterone CHJ I @a). For preparing pregnanes with the desired 178acetyl-16a-substituent it was decided to proceed (I) from 16a-cyano-3~-hydroxyregn-5-en-20-one protected from the possibilities of inversion3 by an ethylene ketal grouping at C-20. Accordingly, H U 1 reaction of I with ethylene glycol and p-toluenesulfonic acid afforded 16a-cyano-20-ethylenedioxypregn-5-en-3,!3-01 (11). Strong alkaline hydrolysis of the latter gave 16a-carboxy-20-ethylenedioxyCH3 pregn-5-en-3/3-01 (111), which upon hydrolysis with &OCTO dilute sulfuric acid was converted into 16a-carboxy-3fi-hydroxypregn-5-en-20-one (IV) , markedly different in properties from the compound synthesized by Romoz with this assigned structure. This latter compound must, presumably, be the HO IV 16&carboxy-17a-acetyl isomer. The correctness of the stereochemistry of the l6a-carboxylic acid IV was further confirmed by treatment of IV with sodium borohydride followed by acetylation to give 3/3,20/3-diacetoxypregn-5-ene-l6a-carboxylic acid (Vb), identical to an authentic sample.4 Lithium aluminum hydride treatment of 16acarboxy - 20 ethylenedioxypregn 5 en 38 01 ( I n ) gave the 16a-hydroxymethyl compound VI as a gelatinous solid. Acid hydrolysis of the latter Va. R = H b. R = A c afforded 16a-hydroxymethyl-3,3-hydroxypregn-5en-20-one (VII). Oppenauer oxidation of either the 20-ketal VI or the 20-one VI1 was not fruitful, and, consequently, an alternate pathway was investigated. Oppenauer oxidation of the ketal nitrile I1 gave 16a cyano 20 ethylenedioxypregn 4 en 3 one (VIII) which was converted into the 3,20bisketal IX. The latter compound was also prepared directly from 16a cyanoprogesterone.2 Strong alkaline saponification of I X gave a crude mixture which was treated with lithium aluminum hydride followed by acid to yield 16a-hydroxymethylpregn-4-ene-3,20-dione (Xa). This compound possessed an optical rotation which wm considerably more dextrorotatory than that of the compound assigned this structure by Romo.2 Moreover, it did not exhibit hemiketal ring formaIX tion in its infrared spectrum as shown by Romo's compound. Indeed, this type of ring formation is most difficult to visualize from molecular models for the assigned structure since the C-20 ketone


I1 I

& i.



- -COaH



1 ('H3

I Pl

- - - -




- -



- - - -


(2) J. Romo, Tetrahedron, S, 37 (1958). (3) Mazur and Cella,' have clearly demonstrated that alkaline hydrolysis of a 16a-cyano-2O-ketone i y t e m , &ut prepared and studied by Romo,*actually resulted in a double inversion of configuration to afford a 17a-acetyl-16/3-carboxylia acid system. This finding wa8 in contraat to the assignment of the 17,9-acetyl-l6a-~arboxylio acid configuration8 by Romo,: and later, by B. Ellis, V. Petrow, and D. Wedlake, J . Chsm. Soc., 3748 (1958). Thin occurrence of a double inversion has been confumed both chemically1 and physically ( o p t i d robtory dispersion analysis).c (4) R. H. Mazur and J. A. Cella,Tstrahsdron. 7 , 130 (1959). (5) W. A. Struck and R. L. Houtman, J . Ow. Chem., 26, 3883 (1961).

-r--YY " Xa. R = OH b. R = 0 3 S oCH3 C. R = I

and the hydroxyl functions are at a considerable distance from each other in the trans geometry of a 17/3-acetyl-16a-hydroxymethyl molecule. Finally, 16a-hydroxymethylprogesterone (Xa), was converted into the tosylate Xb, which, in turn, was treated with sodium iodide in acetone to afford 16aiodomethylpregn-4-ene-3,20-dione(Xc). In the Clauberg progestational assay (subcutaneous route) 16a-hydroxymethylprogesterone (Xa)


JULY, 1962


yield 80 mg. of Vb, m.p. 187-188'. The infrared spectrum waa identical to that of an authentic sample.'J Ida-Hydroxymethyl-3~-hydroxypree;n-5-en-2&one (VII). -To a solution of 14.1 g. of 16~~carboxy-20-ethylenedio~YExperimental pregn-5-en-3fl-ol(III) in 1.5 1. of tetrahydrofuran was added Melting Points.-All melting points are uncorrected. 14 g. of lithium aluminum hydride, and the mixture waa reAbsorption Spectra.-The ultraviolet spectra were deter- fluxed for 2 hr. Ethyl acetate was added cautiously, the mined in methanol; the infrared spectra were determined in solids were filtered off, and the filter cake was washed a potassium bromide disk. thoroughly with hot acetone and ethyl acetate. The comPetroleum Ether.-The fraction used had a bop.60-70'. bined filtrates and washings were evaporated to give the 16a-Cyano-20-ethylenedioxypregn-S-en-3~-d (II).-To a 16a-hydrosymethyl 20-ketal VI as a gelatinous solid. Apsolution of 2 g. of 16~-cyano-3~-hydroxypregn-5-en-20-oneproximately 5 g. of the latter waa dissolved in 300 ml. of (I) in 100 ml. of benzene and 3 nil. of ethylene glycol waa methanol to which was added 30 ml. of 8% (v./v.) sulfuric added 50 mg. of p-toluenesulfonic acid monohydrate. Tho acid, and the solution waa refluxed for 1 hr. Evaporation mixture was refluxed for 5-6 hr. with continuous water re- gave a solid which waa separated by filtration, and was moval, coolcd, and 50 ml. of ethyl acetate were added. The washed well with water to yield 2.18 g. of VII, m.p. 225organic layer was separated and washed with aqueous so- 227'. Two crystallizations from acetonepetroleum ether dium bicarbonate, and then with water until neutrd. After gave m.D. 226-228'; umar 3380 and 1703 cm.-*; [ollU~ being dried, the solution WLES evaporated to give an oil G17' (methanol). which crystallized from acetone-petroleum ether to afford Anal. Calcd. for CBHXO) (346.49): C, 76.26; H, 9.89. Several crystallizations from Found: C, 76.32; H, 10.02. 1.6 g. of 11, m.p. 184-189'. acetono-petroleum ether raised the m.p. to 191-194'; 16a-Cyano-20-ethylenedioxypregn-4-en-3-one (VIII) A umnx 3560, 2235, and 1050 cnl.-l; [aIuD -70.5' (chloro- solution of 4.3 g. of 16a-cyano-20-ethylenedioxypregn-5-enform). 3p-01 (11) in 500 ml. of toluene and 45 mi. of cyclohexanone Anal. Cdcd. for CZ4H,bOsN (385.53): C, 74.76; H, was distilled until 25 ml. of distillate was collected. Alu9.15; N,3.Vd. Found: C, 74.40; H, 8.79; N,3.67. minum isopropoxide solution in toluene (40 ml.; 0.2 g./ml.) 16a-Carboxy-20-ethylenedioxypregn-5-en-3p-01(III).-To W R then ~ added, and the reaction mixture waa refluxed for 40 a solution of 0.5 g. of the 16a-cyano compound I1 in 25 ml. of ndn. Rochelle e d t (40 9.) in water was added and the mixethylene glycol, was added a solution of 5 g. of potassium ture waa steam distilled until a11 the cyclohexanone was rehydroxide in 15 ml. of ethylene glycol, and the mixture was moved. The cooled aqueous mixture waa extracted with refluxed for 24 hr. Addition of water gave a small amount ethyl acetate; the combined extracts were washed with of amorphous solid which wae discarded. The filtrate was water, dried, and evaporated to give 2.85 g. of VIII. Two acidified with dilute hydrochloric acid and extracted with crystallizations from acetone gave the analytical sample, ethyl acetate and methylene chloride. The combined ex- m.p. 248-250": A,, 240 mp (e 18,700); vmax 2240, 1668, tracts were washed with water until neutral, dried, and 1620, and 1044 cm.-'; [ ( Y ] ~ D4-44' (chloroform). evaporated to a semisolid which waa crystallized from aceAnal. Calcd. for C24H~aOsN (383.51): C, 75.16; tone-petroleum ether to afford 0.24 g. of 111, m.p. 239-242". 8.67; N, 3.65. Found: C,74.86; H, 8.70; N,3.68. Two crystnllizations from acetone-petroleum ether followed 1da-Cyano-3,20-bisethylenedioxypregn-S-ene(IX).--tt. by two crystallizations from acetone gavo a m.p. of 245To a solution of 18.2 g. of 16a-cyanopregn-4-ene-3,2O-dion~~ 248'; u , , ~ 3300, 2670, 1710, and 1040 cm.-l; [a]% -30' in 1200 nil. of benzene and 35 mi. of ethylene glycol was (chloroform). added 1 g. of p-toluenesulfonic acid monohydrate. Thc Anal. Citlcd. for C&Z& (404.53): C, 71.25; H, 8.97. reaction mixture waa refluxed for 6 hr. (continuous watcr rcFound: C, 71.3G; H, 9.18. moval), cooled, washed with aqueous sodium bicarbonate, 16c~-Carboxy-3p-hydroxypregn-5-en-2O-one (IV).-To (I and then with water until neutral. Evaporation gave a solid solution of 300 mg. of 16a-carboxy-20-ethylenedioxypregn- which waa recrystallized from acetone to yield 10.3 g. of IX, 5-en-38-01(111) in 25 ml. of methanol was added 3 ml. of 8% m.p. 265-270'. Concentration of the mother liquor afforded (v./v.) sulfuric acid, and the mixture waa refluxed for 5 min. an additional 2.2 g., map. 263-265'. Two crystallizations Evaporation a t 40-45' gave a solid which, after the addition from acetone gave pure IX, m.p. 267-270'; umax 2235 and of water, was collected by filtration and was washed well 1100 cm.-l; [CY]*~D-60' (chloroform). with watcr to afford 220 m g . of IV. Three crystdizations A n d . Calcd. for C%HJ7O~N (427.56): C, 73.03; 11, from acetonepetroleum ether yielded a m.p. 235-238": S.72; N,3.28. Found: C,73.08; H,8.76: N , 3.34. vmax 3460 and 1710 (broad) cm.-'; [ ~ ] Z ' D f14" (chloroform). U. To a solution of 2.85 g. of the 20-ketal VI11 in 1 5 0 A t d . Calcd. for ClzHazO~(360.48): C, 73.30; H, 8.95. nil. of bcnacne and 4 ml. of ethylene glycol waa added 80 m g . Found: C, 73.44; H, 9.16. of p-toluenesulfonic acid monohydrate. The nlixtwre was Admixture melting point determination with 16b-carboxy- refluxed for 5-6 hr. (continuous water removal) cooled, 3p-hydroxy-17a-pregn-5-en-20-onee(m.p. 233-236') gave washed with aqueous sodium bicarbonate, and then with a m.p. of 223-225'. water until neutral. Evaporation gave 2.7 g. of solid whose 3~,20p-Diacetoxypregn-5-ene-16~-carboxylic Acid (Vb) infrared spectrum showed an appreciable amount of a& To a solution of 0.41 g. of 16a-carboxy-3p-hydroxypregn-5- unsaturated ketone. Consequently, the reaction in benzene en-20-one (IV) in 35 ml. of methanol, was added 1 g. of and glycol was repeated, and worked up as described above. sodium borohydride and the solution wae allowed to stand This gave 2.4 g. of material wliich still showed the presence overnight at room temperaturo. Water was added and the of some a,p-unsat,urated ketone. Recrystallization from solution waa acidified with hydrocliloric acid. The resulting acetone yielded 1.1 g. of IX; m.p. 260-264'; and whose insolid was separated by filtration and W'W washed well with frared spectrum WRS identical to that of IX prepared in A water to give 0.24 g. of Va, m.p. 271-5277', Recrystulliaa- above. tion from ethanol afforded 0.10 g., n1.p. 283-286". This l&-Hydroxymethylpregn-4-ene-3,20-dione ( 16a-Hydroxymaterial was dissolved in 3 ml. of pyridine to which was methylprogesterone) (Xa).-A mixture of 12.2 g. of the added 0.3 ml. of acetic anhydride, and the solution wa9 cyanobisketal I X in 280 ml. of ethylene glycol, was treated heated on the steam bath for 1 hr. It was then poured into with 100 g. of potassium hydroxide in 100 ml. of ethylene water and the product was extracted with ethyl acetate. glycol. The mixture waa refluxed for 24 hr. under nitrogen. The extract was washed with water, dried, and evaporated It waa cooled, diluted with water, and dilute hydrochloric to an oil which Crystallized from ether-petroleum ether to

mas inactive at a total dose of 1 mg. at which level progesterone exhibits a pronounced response.6



(6) This amay was carried aut by the Endoorine Laboratory, Madison, Wieoonain.

(7) We thank Dr. E. W. Cantrall of these laboratories for this sample.



acid was added until the mixture was just barely alkaline. The resulting solid was filtered, washed well with water, and waa then refluxed in 2.5 1. of acetone, and filtered. The insoluble residue so collected consisted of silicates as shown by infrared analysis. The filtrate was evaporated to a solid which waa slurried with acetone to give 3.68 g. of solid. This material was dissolved in 300 ml. of tetrahydrofuran, 3.5 g. of lithium aluminum hydride waa added, and the mixture was refluxed for 2 hr. Ethyl acetate was added cautiously and when the lithium aluminum hydride had decomposed, the mixture was filtered and the filter cake was washed thoroughly with boiling acetone and ethyl acetate. The combined filtrates and washings were evaporated to give a white solid which was refluxed in 11. of methylene chloride and filtered. The insoluble residue was discarded, and the filtrate waa evaporated to a solid which waa slurried with petroleum ether to afford 2.6 g. of solid. Several crystallizations from acetonepetroleum ether gave material melting over a broad and inconsistent range (m.p. 176-194'). A portion (200 mg.) of this material was dissolved in 30 ml. of methanol to which wm added 3 ml. of 8% (v./v.) sulfuric acid. The solution waa refluxed for 0.5 hr., water was added and the turbid mixture waa evaporated until solid formed. This was filtered off and washed well with water to yield 100 mg. of Xa. Four crystallizations from acetone241 petroleum ether gave 40 mg. of Xa, m.p. 163-164"; .A,, nip (e 16,800); vmsx 3510, 1717, 1667, and 1613 cm. -I; [a]as^ f160" (chloroform). Anal. Calcd. for CZ2Ha2O3 (344.48): C, 76.70; H, 9.36. Found: C, 76.32; H, 9.41. 16~-p-Toluenesulfonyloxypregn-4-ene-3,2O-dione (Xb).To a solution of 250 mg. of Xa in 1.5 ml. of pyridine was added 250 mg. of p-toluenesulfonyl chloride at -5", and the mixture waa allowed to stand a t -5' for 24 hr. Water was added, and the crude solid obtained waa collected and recrystallized from acetone-petroleum ether to give 240 mg. of 228 mp ( e 21,600), 240 mp (e Xb, m.p. 203-205"; ,X, 16,400); vmax 1708, 1670, 1625, 1603, 1178, and 945 cm.-l; [CY]%D$98" (chloroform). Anal. Calcd. for CzSHasO,S (498.65): C, 69.85; H, 7.68; S,6.41. Found: C, 69.39; H, 7.80; S, 6.71. 16~~-Iodomethylpregn-4-ene-3,20-dione (Xc).-To a solution of 210 mg. of the tosyl compound Xb in 25 ml. of acetone was added 400 mg. of sodium iodide. The mixture was refluxed for 9 hr., concentrated, and water was added. The resulting solid wm collected and washed with water to give 130 mg. of Xc; m.p. 11&120". Recrystallization from acetonepetroleum ether afforded 50 nig., m.p. 120-135"; 1700, 1670, and 1615 cni.-l. Anal. Calcd. for CP2Ha1021 (454.37). C, 58.15; IT, 6.88; 1, 27.93. Found: C, 57.62; H, 7.05; I, 37.35.

VOL. 27

elaborated by Streptomyces antibioticus4s6 a selectively labeled hydroxyproline was needed. It has been shown previously that sodium borohydride reduction of N-carbobenzyloxy-4-keto-~proline (I) yields exclusively N-carbobenzyloxy-4allohydroxy-L-proline (IT') .6 However, in marked



0" I.R=Cbz II.R=H



contrast to results from another laborai,ory,6a wc have now found that a mixture of V and 4-hydroxyL-proline (111) is obtained by the action of sodium borohydride (or in this study by tritium-labeled sodium borohydride) on 4-keto-~-proline (11). This effect of the ring nitrogen on the stereochemist'ry of reduction by complex met'al hydrides is also operative in the reductions of S-carbobenzyloxy-5keto-L-pipecolic, 5-ket'o-nL-pipecolic acid and 4keto-L-pipecolic acid (Table I), and may be rationalized in terms of part'icipation of t'he nitrogen6in an intermediate complex of t'he metal hydride with the carbonyl groups7 The quantitative assay of this mixture of 4tritiated diast,ereoisomeric*hydroxyprolines (111 and V) 011 a coluinn of ion exchange resinsv9showed the preseiice of three part8sof nllohydroxy- and onc pa,rt of h?.dros;l-r,-pioline. 011it prepariat scale the diast~reoisomers \\'PIT separated hy cliit'ion \vit,h t.thaiiolic.--aqueoiis ammonium ac*ettatc hiifffer Acknowledgment.--We wish to thaiik Louis Jl. from a c:oliiniii of .hnberlite C!C;-1'20.'" Brancone and associates for the analyses, and WilBeyoiicl its iise in actinoniyciii studies 4-H liam Fulnior and associates for the infrared and hydroxy-L-proline will permit the study of the reultraviolet absorption spectra and optical rotation versibility of the reduction of 4-keto-L-proline by a data. special hydrogenase present in rat-kidney homotrmax



The Preparation of 4-H3-Hydroxy-and 4-R3-Allohydroxy-L-proline 21. 1-.ROBERTS OX,^ E.

h h T Z , ' AND

B. W I T R O P

.Vationid Ilwtztute of Arthritzs and Metabolzc Dzseasea, and the National Heart Institute, National Institutes of Health, Public Health Service, U.8.Department of Health, Education,

and Welfare, Bethesda, Md. Received January 24, 1962

For the study of the incorporation of aniino acids into the peptide part of the actinomycins3

(1) Associate in the Visiting Program of the USPHY, present address: Dept. of Chemistry, University of Sydney, Australia. (2) Special Fellow, USPHS. (3) E. Kate, D. Prookop, and 8. Udenfriend, Precursors of the €I!#droxyproline and Ketoproline in the Actinomycins, J . B i d . Chem.. in preas. (4) CJ. E. Kate and L. H. Pugh, A p p l m d Micrubiology, 9, 868 (1961). ( 5 ) H. Brookmann and J. H. Manegold, Chem. Ber., 9S, 2971 (1960). (6) A. A. Patchett and B. Witkop, J . Am. Chem. Soc., 79, 185 (1957).

(ea) H. C. Beyerman, Rec. Lrar. c h i n . 80, 556 (1961). (7) Cf.D. M. S. Wheeler end J. W. Huffman, Ezperienlia, 16, 516 (1960). ( 8 ) K. .4. Pies, J . Bid. Chem., 207, 77 (1954). (9) Cf.B. Witkop and T. Beiler, J . A m . Chen. Soc., 78, 2882 (1956) (10) We are greatly obliged t o Dr. S. M.Birnbauni, N C I . for directions on this extremely useful unpubliahed procedure.


JULY, 1962


TABLEI IXFLUENCE OF THE HASIC NITROOEN ON THE STEREOCHEMISTRY OFREDUCTION O F ~ E T O . 4 b l I N OhCIDS TO Product of reduction Keto-amino acid with sodium borohydride Stereospecificity

HIDROXYAMINO A(:lI)s Reference

100% cis-N-carbobenzyloxy- B. Witkop and C. A I . Foltz, J . A m . Chem. SOC.,79, 192 5-hydroxy-~-pipecolicacid (1957)


K-Carbobenayloxy-5keto-L-pipecolic acid


lliiiost exclusively truns-5hydroxy-nL-pipecolic acid

5-Keto-uL-pipecolic acid

H. C. Beyenilan and P. Boeker, &ec. traa. chim., 78,648 (1959)



80 yocis-4-hydroxy-~pipecolic acid

4-Keto-L-pipecolic acid HO

J. W. Clark-Lewis and P. L. Mortimer, J. Chem. Soc., 189 (1961)


100% ci! (or a h ) 4-hydroxy-L-proline

X-Carbobenzyloxy-4keto-L-proline B

O o I Cbz

e C O O H N H



A. A. Plttchett and B. Witkop, J. Am. Chein. SOC., 79, 185 (1957)




25% trans- and 757; cis4-hy droxy-~proline

V. Robertson, E. Katz, and B. Witkop, this paper


genate.l' The loss of tritium in such a case would be an accurate measure of reversion to 4-keto-~proline. Experimental 4-Keto-L-proline hydrobromide (35 nig.)6s12 in 0.5 i d . of water was added drop by drop to a solution of 5 nig. of sodium borohydride-H3 (100 mC).13 When the addition was complete the solution was slightly acid, and sufficient cold sodium borohydride (ca. 1 mg.) was added to obtain an alkaline solution and hence complete reduction. After 30 min. the reaction mixture was transferred to the top of a column of Amberlite CG-120, type 2 ion exchange resiu (70 ml. of 300-400 mesh) prepared in 0.2 ;M ammonium acet,ate solution in water-ethanol (60:40, v/v). The column was eluted with t8hesame buffer solution and 4-niI. frartions were collected. Aliquots (10 A) from each tulw were spott,ed on paper and sprayed with isatin reagent. Tubes 13-16 contained 4-hydroxyproline (ident,ified in :t cold run by paper electrophoresis). These fractions were combined and the solvent was removed in umuo. The ammonium acetate was then sublimed during 3 hr. at 60°/1 mm. The residue was dissolved in 0.1 ml. of water and 3 ml. of hot ethanol was added. The solution was boiled and acetone was added dropwise until crystallization commenced. After 2 hr. a t O", 4-H3-hydroxy-~-proline ( 5 . 2 mg.) was collected as colorless needles. Its specific activity was 2.86+107c.p.m./pmole. Tubes 20-24 contained 4-H3-4allohydroxy-~prolinewhich was worked up in a similar way to give 10.8 mg. of colorless recrystallized product of specific activity 3.24.107 c.p.m./ rmole. The total recovery of the initial activity present in the sodium borotritide (100 mC) amounts to 1.4%.

(11) C. Mitoma, T. E. Smith, F. M. DaCosta, S. Udenfriend, A. .4. Patchett and B. Witkop, Science, lBS, 95 (1959). (12) We are greatly indebted to Dr. Sydney Archer, The Sterling\Vinthrop Research Institute, for the preparation of this compound. (13) Obtained from New England Nuclear Corp., Boston, hlass.

Anomalous Hydrolytic Behakior of Some

Basically- Substituted Phthalimides. A Novel Rearrangement of a 4-Aminoquinoline Side Chain1 R I C H ~ RM. D PECK Diuision of Chemotherapy, The Institcite for Cancer Research, Philadelphia 11, Pa.

During the preparatioii of ,V-(2'-niethylaminocthy1)phthalimide ( S o . 1 in Table I) for iise as an intermediate, it was foiuid that its hydrochloridr 011 careful iieutralizatiori with alkali in cold water did not yield the free base, but the phthalamic: acid; further, the phthalimide obtained by heating the phthalamic acid and distilling gave only a transient alkalinity with water, hydrolyzing to regenerate the phthalamic acid. Investigation showed that this unexpected ease of hydrolysis persisted with homologs containing two-, three-, and four-carbon chains between the two nitrogens, although to a lessening degree. In one other case (No. 4 in Table I) the phthalamic acid also crystallized; three other cases showed a reversal of hydrolysis, in which the base was regenerated on steam cone evaporation of the solution. All of these phthalimides have been previously reported, either as such or, more generally, as the (1) Supported in part by research grant CY-2975 froiii the National Cancer Institute, U. 8. Public Health Service.

VOL. 27




I1 0 Alkyl chain



Pbthalamic acid, m.p.


187.5-189"dec? Instantaneous 1. 4 C H z ) r CHI H CnHfi cas Unstable Instantaneous 2. -(CH$)nInstantaneous 3. 4 C H Z k CHs H Unstable 4. -(C&)a-CHI CH, 107-108.5" dec? Intermediate Slow CiHs CIHS Unstable 5. -CH(CHa)(CHz)aa Qualitative; based on zero titration of a solution of the first three compounds in aqueous ethanol; No. 4 required about SO% of the theoretical acid after a solution time of two minutes; No. 5's basicity decremed 50% only after several hours. With loss of water to form the phthalimide.


hydrochloride, but search of the many references2-7 available revealed no mention of this property, Rearranoes other than Shriner's recommendation of neutraliA c1 c1 \ H * O I OHzation of the hydrochloride below 25" to isolate I11 I1 the base of the diethylaminopropyl homolog and Kanevskaya's mention of relatively easy acid cleavage of some analogs with nuclear substituents. This would refer to a different phenomenon, since with a sample of the previously unreported 111, the hydrochlorides are stable to water in contrast and by the nonidentity of the phthalamic acid derivatives produced (1) by the action of phthalic to the free bases. The free bases No. 1 and 3 were obtainable only anhydride on I11 and (2) by the partial hydrolysis by the dehydration of the phthalamic acid and of compound I (n = 2). vacuum distillation; they both formed glasses Experimental in contrast to the mobility of the distilled tertiary bases. They were both successfully condensed The aminoalkylphthalimides were synthesized by addiwith 4,'i-dichloroquinoline to give products (I) tion of phthalic anhydride to a stirred solution of the diamine which still possessed the ease of hydrolysis of the in water and evaporation of the solution. If the phthalamio


acid wm stable, it W M isolated at this point by acetone dilution and recrystallization. If not, the residue was heated to achieve dehydration and the phthalimide WM distilled in vacuo. Physical constants of the base and/or hydrochloride checked literature values in every cme. N-Methylaminoethylphthalamic Acid.-This compound was isolated from the concentrated reaction mixture by acetone dilution in 41% yield. It wm recrystallized by diusolving in a small amount of water and precipitating with ethanol, map. 187.5-189' dec. Crystallization of this conipound wm extremely slow. Anal. Calcd. for CllH14N~03:C, 59.44; H, 6.35; N , 12.61. Found: C, 58.86, 58.73; H, 6.39, 6.43; N, 12.91. N-Dimethylaminopropylphthalamic Acid.-This compound was obtained by concentration of the initial reaction mixture; the analytical sample was prepared by evaporating a water solution of the distilled imide and recrystallization fromethanol, m.p. 107-108.5". A n d . Calcd. for ClsHl&zOs~11/2H~O:C, 56.4; H, 7.62; N, 10.1. Found: C,56.18,56.21; H,7.71,7.54; N, 10.01, 10.02.

parent conipound, ie., they dissolved over a period of several hours a t room temperature in a dilute alcoholic solution containing one excess equivalent of sodium hydroxide. I n unsuccessful attempts to synthesize N(7 - chloro - 4 - quinolyl) - N - methylethylenediamine : (a) The corresponding phthalimide (I, n = 2) was hydrolyzed with ethanolic hydrazine. (b) Its derivative phthalamic acid was hydrolyzed with hydrochloric acid. I n each case, the product obtained was 7-chloro-4-(2-methylN - [N'-(7-Chloro4quinolyl)-N'-methylaminoethyl] :tminoethyl)aminoquinoline (111). This was con- phthaliide Hydrochloride.-A mixture of 42 g. of methylfirmed by melting point and mixed melting point aminoethylphthalimide (distilled from crude phthalamic (2) 8. I. Kanevskaya and V. B. Brasyunas, Zhur. Obeclasi. Khim.,

as,1930 (1959).

(3) W. Voigtlhder, Chem. Tech., 11, 523 (1959). (4) P. C.Jocelyn, J . Chsm. Soc., 3305 (1967). ( 5 ) L. A. Rice, et at., J . Am. Chsm. SOC., 71, 4911 (1953). (6) M.B. Moore and R. T. &pala, {bid., 68, 1667 (1946). (7) K. Miura, J . Pharm. SOC.Japan, 61, 224 (1942).

acid), 50 g. of 4,7-diohloroquinoline, and 35 g. of diethrtnolamine was condensed at 120' for 15 hr., taken up with excess dilute acetio acid, 50 ml. of saturated sodium chloride added, cooled, and filtered. The crude sparingly soluble monohydrochloride weighed 65.7 g. (79%). Recrystallization f i s t from water and then from alcohol gave an analytical sample, m.p. 282-285".

JULY, 1962


Anal. Calcd. for C20H&1NB02.HC1.11/zH~O: C, 56.00;

H, 4.70; N, 9.79; C1-, 8.28. Found: C, 55.49, 55.28; H, 4.79,4.91; N,9.72,9.79; C1-, 8.47,8.42.

A New Method of Preparation of 2,2'-Biwinoxalines

The free base melted at 153-154.5'. Anal. Calcd.: C, 65.65; H, 4.41; N, 11.49. Found: C, 64.89; H, 4.42; H. SMITHBROADBENT AND RICHARD C. ANDERSON N,11.39. N-[N'-(7-Chloro-4-quinolyl)-N'-methylaminopropyl] Department of Chemistry, Brigham Young University, Prwo, phthalimide Hydrochloride.-This homolog was synthesized Utah in the same way and isolated as the sparingly soluble hydrochloride. After recrystallization from hot water, it Recewed January 96, 1968 amounted to 8270 of the theoretical yield and melted at 139141'. Recrystallization from ethanol gave 65% recovery The synthesis of some 2,2'-biquinolines and 2,2'of pure material, m.p. 246-248', a . 150'. bipyridines by catalytic dehydrogenation of the Anal. Calcd. for C ~ I H ~ ~ C ~ N ~ O Z ~ H C I . C, ~ -56.35; ~/~HZO : H , 5.06; N, 9.21; C1-, 7.76. Found: C, 56.36, 56.27; corresponding quinolines and pyridines has been H, 5.58,5.49; N, 9.29,9.03; C1-, 7.79,7.90. described by prolonged heating of a mixture of the The free base melted a t 118.5-120.5'. heterocyclic base and 10% by weight of 5% pallaAnal. Calcd.: C, 66.35; H, 4.78; N, 11.06. Found: dium-on-carbon catalyst, usually at a reflux temC, 66.65, 66.41; H,4.73,4.78; N,11.03, 10.94. perature of the base.* Modest yields (2-210/0) of 7-Chloro-4-(2-methylaminoethy1)aminoquinoline(In).This compound wm produced by the condensation (2 hr. coupled product were formed by such treatment. 2,2-Biquinoxalines have not been studied exa t 115') of 4,i-dichloroquinoline and 2-methylaminoethylamine.8 It was purified by distillation (115'/20 p ) , tensively, and no generally applicable unambiguous crystallization from benzene, and vacuum sublimation. syntheses of compounds of this class have been The yield was approximately 5070, m.p. 110-111.5'. reported. We have encountered a new method of Anal. Calcd. for C12Hd3X;a: C, 61.18; H , 5.98; N, 17.83. Found: C, 61.13, 61.25; H. 6.41, 6.45; N, 17.72, preparation of 2,2'-biquinoxalines as an outgrowth of an attempt to prepare 2-vinylquinoxaline by 17.98. N-(2-[Methyl-(7-chloro-4-quinolyl)]aminoethyl]phthal~c heating 2-methylquinoxaline with paraformaldeAcid.-A mixture of 60 ml. of water, 72 ml. of N sodium hyde a t 160-165' in an autoclave. Under these hydroxide, 30 ml. of ethanol, and 15 g. (36 mmoles) of N- conditions, a high-melting red-brown crystalline [N'-(7-chloro-4- quinolyl) -N'-methylaminoethyllphthal- compound was obtained in about 5% yield. This imide hydrochloride (I, n = 2) was stirred for 1 hr. while it became homogeneous. Acidification with acetic acid pre- compound was shown to have the molecular forby elemental analysis and moleccipitated the product. It weighed 14.5 g. and melted at mula, ClaH14N4, 1741'7'7' dec. An analytical sample obtained by reprecipi- ular weight determination. The structure, 3,3'tating from an alkaline ethanolic solution with acetic acid dimethyl-2,2'-biquinoxaline, was assigned on the melted at 172-174' dec. basis of the following observations: (1) QuinoxaAnal. Calcd. for C2oHjsClNa03.H20: C, 59.75; H , 5.01; N, 10.46. Found: C, 59.25, 59.01; H, 5.17, 5.14; N, line was found to react under the same conditions to give an almost quantitative yield of 2,2'-bi10.52, 10.42. Acid Hydrolysis of Phthalamic Acid.-A 16.8-g. sample of quinoxaline (identical with authentic sample). N [2 [methyl - (7 chloro 4 quinolyl)]aminoethyl]- (2) 2,3-Dimethylquinoxaline failed to form a phthalamic acid was hydrolyzed with 100 ml. of 6 N hydro- coupled product even upon prolonged heating a t chloric acid by heating on the steam cone overnight, con- 200'. (3) The same products are formed in the centrated in vacuo, diluted to 100 ml., and 5.5 g. of phthalic of paraformaldehyde under otherwise absence acid was removed by filtration. The filtrate was concentrated and diluted with ethanol and ether to precipitate similar conditions. (4) The infrared absorption 11.3 g. of base hydrochloride. The crystalline base was spectrum of the product from 2-methylquinoxaline isolated in 97y0 yield and recrystallized several times from exhibits absorption bands in the region characterhydrocarbon solvents. Both recrystallization and vacuum sublimation gave a melting point of 110-112' not depressed istic of four adjacent aromatic hydrogens. ( 5 ) The by a sample of 7-chloro-4-(2-methylaminoethyl)amino- ultraviolet absorption spectrum of this compound quinoline (111). exhibits absorption maxima a t much longer wave N - [ 2-(7-Chloro-4-quinolylamino)ethyl] -N-methylphthala- length than does the ultraviolet absorption specmic Acid.-A 5.9-g. sample of the above material WBS al- trum of 2-methylquinoxaline indicating conjugalowed to react with an equimolar quantity of phthalic anhydride in ethanolic solution. The precipitated product tion between aromatic rings. Yields of the coupled product from 2-methylquinwas filtered and reprecipitated from alkaline solution with oxaline were enhanced by heating a t higher temacetic acid. It weighed 8.3 g. and melted at 250-255'. Anal. Calcd. for C&,sC1NaOd.2Hz0: C, 57.20; H, peratures. For example a 39% yield of coupled 6.27; N, 10.01. Found: C, 57.98; H, 5.88; N,9.32. product was obtained by heating a t 200' for six Hydrazine Hydrolysis.--A sample of the phthalimide hours, whereas only a 5% yield was realized by free base (I, n = 2) was refluxed for 2 hr. with an equiva- heating a t 160-165O for four hours. Slight catalytic lent amount of molar ethanolic hydrazine and gave a product which on recrystallization and sublimation proved to be activity was exhibited by 5% palladium-on-carbon identical with that from the acid hydrolysis.




- -

( 8 ) The usual method of attachment of an amine to a 4-chloroquinoline; cf., R. M. Peck, R. K. Preston, and H. J. Creeoh, J . Am. Chem. SOC.,81,3988(1959).

(1) Finanoial eupport of thia work by a grant (CY-3751)from the National Canoer Institute, U.S. Public Health Service, is gratefully acknowledged. ( 2 ) H. Rapoport, R. Iwamoto, and J. R. Tretter, J . Ore. Chem., 26, 373 (1960).



since a 46% yield of the coupled product was obtained by heating for six hours at 200' in the presence of 2% by weight of the catalyst. With a related compound, 4-methylquinazoline, we failed to obtain any coupled product upon heating a t 200' for four hours in the presence of 2% by weight of 5% palladium-on-carbon. Experimental 3,3'-Dimethyl-2,2 '-biquinoxaline .-2-Methylquinoxaline (50.0 g., 0.347 mole) was placed in a glass liner for a Pendaclave high pressure hydrogenation apparatus, and 1.O g. of 5% palladium-on-carbon catalyst was added. The mixture was agitated and heated at 200' for 6 hr. The solid reaction mixture waa taken up in 200 ml. of hot ethyl acetate and filtered, using suction. The filtrate waa cooled in an ice bath and the product waa collected by suction filtration. Yield was 16.3 g. of brown-red crystalline material having m.p. 25C-258.5'. The filtrate was evaporated and the residue was distilled at reduced pressure, leaving an additional 5.8 g. of product having m.p. 245-252' as residue. Total yield waa 22.1 g. (46%). Twenty-three grams of 2-methylquinoxaline was recovered from the vacuum distillation. An analytical sample aa long brownish red needles having m.p. 256.5-258.5' was prepared by vacuum sublimation. Anal. Calcd. for Cl8HI4N4:C, 75.5; H, 4,9; N, 19.6. Found: C, 76.0; H, 4.6; N, 19.6. Infrared (KBr): 3.20 (w), 3.24(w), 6.20(m), 6.45 (m,sh), 6.52(s), 6.80(s), 7.04(s), 7.30(w), 7.67(m), 7.93(m), 11.76 (m), 12.33 (m), 13.20 (s), 15.32 (m)p. Ultraviolet (CHC13): Xm,,277mp(~38,5O0),3 6 7 (t12,900), ~ 383.5mp(e15,000), 468 mw ( e 5600).

VOL. 27

butanol under a variety of conditions.3 In no case was one isomer obtained exclusively. A desire for a convenient source of pure 3-chloro-1-butanol stimulated a reinvestigation of the cyclic sulfatehydrogen chloride reaction. Reaction of 1,3-butanediol cyclic sulfate with hydrochloric acid was carried out as described.' The product, a distillable liquid, exhibited physical properties which were quite similar to those reported for 3-chloro-l-butano11~4;however, vapor phase chromatography indicated that the product was a mixture of two materials in approximately a three to one ratio. Syntheses of 3-chloro-1-butanol and 4-chloro-2-butanol were performed by known method^.^ Vapor phase chromatographic retention times of these materials demonstrated that the product from the cyclic sulfate reaction was indeed a mixture of 73% 3-chloro-1-butanol and 27% 4chloro-2-butanol rather than pure 3-chloro-l-b~tanol. The corresponding reaction with hydrobromic acid has provided a three component mixture comprised of the isomeric l,&bromobutanols and a lower boiling material. Experimental6

Instrument.-A Beckman GC-2 Vapor Fractometer with a 6-ft. column of lac 446 polyester on Chromosorb-W waa employed for gas chromatography at 130' with a helium flow of 60 ml./min. 1,j-Butanediol Cyclic Sulfate (I).-The method of LichCleavage of 1,3-Butanediol Cyclic Sulfate with tenbergerl was employed; however, 20% oleum was used in place of 47% material. Hydrogen Chloiide A solution of 104 g. of 1,3-butanediol in 600 ml. of chloroform was placed in a 2-1. flask equipped with a stirrer, condenser, thermometer, and dropping funnel. The solution JOHN J. MILLER was cooled in an ice bath and the temperature maintained at 0-10' as 1060 g. of 20% oleum was slowly added. After Rohm & Haas Company, Research Laboratories, Bristol, addition, the reaction mixture was kept a t ice-bath temperaPennsyloania ture for 1 hr. and then poured onto ice. The layers were separated and the aqueous layer was extracted with chloroReceived January 29, 1962 form. The chloroform portions were combined, washed with water, 10% aqueous sodium bicarbonate, and water again. Activated charcoal was added to the chloroform solution Reactions of 1,3-diol cyclic sulfates with hydrogen and then the solution was filtered, dried, and evaporated halides have been shown to give 1,3-halohydrin~.'~~ leaving 99.5 g. (56% yield) of a clear oil which solidified on Symmetrical sulfates can provide only one product, cooling. Recrystallization from ether provided white crysh t unsymmetrical materials could lead to isomers. tals of I, m.p. 44-46' (reported,' m.p. 43-44'), Reaction of I and Hydrochloric Acid.-A mixture of 20 g . of the cyclic sulfate of 1,3-butanediol and 50 ml. of concenn X trated hydrochloric acid was heated on the steam bath for 40 min. During this time the two layer mixture became homogeneous. The mixture was cooled, 200 ml. of water was added and the solution was extracted with three 100-ml. I portions of ether. Washing of the combined ether extracts R-CH- -CH,CH,X with 10% aqueous sodium bicarbonate and then with water was followed by drying over magnesium sulfate. EvaporaThe reactions of l,&butanediol cyclic sulfate tion of the ether left 9 g . of an oil which was distilled in with hydrogen halides have been described to Vacuo to give 7.5 g. (53%) of a clear liquid, b.p. 70-71"/ furnish only 3-halo-l-b~tanols,1~~ while similar 17mm., ~ * S D1.4397. Anal. Calcd. for C4H&10: C, 44.25; H, 8.34; C1, reactions of hydrogen chloride with 1,3-butanediol 32.66. Found: C, 44.34; H, 8.16; C1,32.25.

or 2-methyloxetane have been shown to provide mixtures of 3-chloro-1-butanol and 4-chloro-2-

(1) J. Lichtenherger and R. Lichtenberger, Bull. aoc. chzm. France,

1002 (1948). (2) J. Lichtenberger and L. Diirr, ibid., 664 (1956).

(3) S. Searles, Jr., K. A. Pollart, and F. Block, J . A n . Clrcm. Soc., 79,952 (1957). (4) J. Velhulst, Bull. aoc. c h i n . Belgea, 40,85 (1931). (5) 411 boiling and melting points are uncorrected.


JULY, 1962 Vapor phase chromatography showed two peaks representing 27% and 73% of the product, respectively. 3-Chloro-l-butanol.-This material was prepared by lithium aluminum hydride reduction of 3-chlorobutyric acid as described by Searles, et a1.,8 b.p. 70-72.5'/17 mm., n% 1.4399. (Reported, b.p. 67-68'/15 mm.' and 74'/16 mm.;* n% 1.4446' and 1.439P). Vapor phase chromatography showed only one peak with the same retention time as that of the large peak from the above mixture. 4-Chloro-2-butanol.-Lithium aluminum hydride reduction of 4-chloro-2-butanone gave this material, b.p. 68-70 "/ 20 mm., n% 1.4395 (reported, b.p. 67O/20 ".,e 12% 1.44088 and 1.44403). Gas chromatography showed a single peak with a retention time equal t o that of the small peak from the cyclic sulfate product. (6) F. Sondheimer and R. B. Woodward, J. A m . Chem. Soc., 76, 5438 (1953).

Vinylogy in Alkylation of t-Butyl Esters KEIITI SISIDO,KAZUO SEI, AND HITOSINOZAKI

Department of Industrial Chemistry, Faculty of Engineering, Xy&to University, Ktbto, Japan Received January SO, 1961

I n an extention of studies on alkylation of t-butyl acetate with organic halides,l the reaction of 1,4dibromo-trans-2-butene and t-butyl acetate has been investigated. When a mixture of t-butyl acetate with an equivalent amount of lithium amide suspended in liquid ammonia was treated with 1,4dibromo-trans-2-butene dissolved in anhydrous tetrahydrofuran, a crystalline diester was obtained in 6 4 4 7 % yields based on the bromide. The ester was found to be di-t-butyl trans-4-octene-1,S-dioate, as hydrolysis with ethanolic potassium hydroxide afforded trans-4-octene-l,8-dioicacid in an 88% yield and catalytic hydrogenation of the ester over Raney nickel gave crystalline di-l-butyl suberate, which was subsequently hydrolyzed to suberic acid. The condensation of 1,Pdibrom0-2-butene with sodiomalonate has already been described as resulting in the exclusive formation of 2-vinylcyclopropane l,l-dicarboxylate.2 An analogous reaction of t-butyl acetate according to equation 1had been anticipated, but no cyclopropanecarboxylate could be isolated from the condensation product. A pre-




+ Me[CH~==C-OBu-11e or MXHI



h f e [IC H sE-0Bu-tle

+ CH~COOBU-t ( 2 ) or


requisite to the ayclixatioii should be the deprotoiiation of monoalkylated ester (equation 2>.3 A h described previously, the second alkylation of the t-butyl monoalkylacetates is scarcely perceptible with lithium amide, while dialkylated esters along with monoderivatives are produced in condensation by means of sodium or potassium amide. Accordingly the present reaction was repeatedly examined for possible formation of cyclized products in the presence of the latter amides, but the sole reaction product isolated in reduced yields was di-t-butyl trans-Poctenedioate, as before. a-Dialkylation of ester by means of lithium amide was observed in the benzylation of t-butyl crotonate. When a mixture of t-butyl crotonatc and lithium amide suspended in liquid ammonia was treated with an equivalent amount of benzyl chloride and the condensation products were hydrolyzed by boiling hydrochloric acid dissolved in dioxane, two crystalline acids melting at 101-102° and 98.5-99.5', respectively, were isolated. Analyses of the higher melting acid agreed with a monobenzylated crotonic acid, while the lower melting acid with a dibenzylated product. Infrared spectrum of the dibenzylated acid indicated the presence of a vinyl group and a carbonyl group which is not in conjugation with the ethylene bond. The assigned structure of 2,2-dibenzyl-3butenoic acid was confirmed by KMR spectrum as described in the E~perimental.~Infrared spectrum of the monobenzylated crotonic acid revealed the presence of a carboxyl group in conjugation with an ethylene bond, but comparison with authentic 5-phenyl-2-pentenoic acidj exhibited remarkable differences and a mixed melting point showed depression. Evidence supporting the structure of 2benzyl-2-butenoic acid was again obtained by means of NMR spectrum. These benzylated crotonic acids were always accompanied by a considerable amount of insoluble polymeric products, which were not investigated. The yields and relative ratio of mono- and dibenzylated products were variable. In most favorable conditions 2-benzyl-2butenoic acid was obtained in a 31% yield, while 2,2-dibenzyl-3-butenoic acid in a 36% yield.




C H -COOBu-t



(1) K. Sisido, Y. Kazama, H. Kodama, and H. Nozaki, J . Am. Chem. Sac., 81,5817 (1959). (2) (a) R. W. Kierstead, R. P. Linstead, a n d B. C. L. Weedon, J. Chem. Soc., 3610 (19,521; (b) S. F. Birch, R. A. Dean, and N. J. Hunter, J. O r g . Chem., 23, 1390 (1958); (c) J. Nickl, Chem. Ber., 91, 553 (1958).

(3) For a recent publication O R related problems see (a) W. R. Dunnavant and C. R.,Hauser, J. Ore. Chem.. 26, 1693 (1960). (b) 0. Mkrtensson and E. Nilson, Acta Chea. Scand., 16, 1026 (1961). (4) The signal of benzyl methylene protons of 2,2-dibenzyl-3butenoic acid was observed as a quartet of typioal A B coupling pattern. This is 8 token indioatins t h a t two protons of the methylene group are non-equivalent possibly because of the hindered free rotation of benzyl groups. ( 5 ) H. Staudinger and H. Schneider, Bq.,66, 699 (1923).



Condensation of t-butyl crotonate with n-butyl bromide in the presence of lithium amide afTorded similarly a nearly equimolar mixture of mono- and dibutylated crotonates. Fractional saponification of this mixture gave 2-n-butyl-2-butenoic acid in a 15% over-all yield, besides t-butyl2,2-di-n-butyl-3butenoate which was recovered unchanged in a 28% yield, both based on n-butyl bromide. This difference in the hydrolysis rates would be ascribed to the increased steric hindrance of the ester group in the latter dialkylated product. The y-alkylated products could not be identified throughout these alkylations of the crotonate. Analogous a-alkylations of crotonic acida and of alkylidenemalonates' have been mentioned.8 Experimental9 Di-t-butyl trans-4-Octene-l,S-dioate.-To a lithium amide suspension prepared from 1.3 g. (0.19 g.-atom) of metallic lithium in 300 ml. of liquid ammonia and cooled at -40', 21 g. (0.18 mole) of t-butyl acetate was added with continuoua stirring. After an additional 15 min. a solution of 16 g. (0.075 mole) of 1,4aibromo-tl.ans-2-butene10 in 30 ml. of anhydrous tetrahydrofuran waa added dropwise, and the whole mixture waa stirred a t -40' for 3 hr. The mixture waa treated in the usual way1 and extracted with ether. Evaporation reaidue (14.2 g. or 67%) solidified and melted a t 54-56'. Recrystallizations from petroleum ether (b.p. 30-60') afforded an analytical sample, m.p. 56-57'. Anal. Calcd. for Q6H%O4: C, 67.57; H, 9.93. Found: C, 67.88; H, 10.01. Infrared absorptions (Nujol mull): 1720 om.-' (nonconjugated ester), 970 cm. -1 (trans-ethylenic hydrogen). trans-4-Octene-1 ,S-dioic Acid.-Hydrolysis of 10 g. of the di-t-butyl ester by refluxing for 3 hr. with 20 g. of potassium hydroxide in 60 ml. of aqueous ethanol afforded 5.3 g. (88%) of the dibasic acid melting a t 171-174'. Single recrystallization from a mixture of petroleum ether (b.p. 3060') and ethanol gave a pure sample, m.p. 175-176'. Anal. Calcd. for CsH120,: C, 55.80; H, 7.03. Found: C, 55.77; H, 7.08. Suberic Acid.-A solution of 5.7 g. of di-t-butyl trans-4octene-1,8-dioate in 50 ml. of anhydrous ethanol waa shaken with 1.0 g. of Raney nickel (W-5 grade)" in an atmosphere of hydrogen a t room temperature. After 1 equivalent of hydrogen wm absorbed, the catalyst was filtered and the solution waa concentrated to afford colorless crystah (5.1 g. or 89%) of di-t-butyl suberate, m.p. 30-31". Anal. Calcd. for C16H8004:C, 67.09; H, 10.56. Found: C, 67.01; H, 10.44. Hydrolysis with hydrochloric acid in dioxane' gave a nearly quantitative yield of suberic acid, m.p. and mixed m.p. 139-140'. Condensation of &Butyl Crotonate with Benzyl Chloride.The reaction was repeatedly examined over ten times, but the yields of products were variable due to the tendency of the olefinic esters to polymerize into a resinous product. In one experiment a lithium amide suspension prepared from (6) A. J. Birch, J . Chum. 900.. 1551 (1950). (7) A. C. Cope, H. C. Holmes, and H. 0. House, O w . Reactionr, 9, 114 (1957). (8) For oondensation of t-butyl crotonate with aoetophenone, which occurs at r-carbon of the ester, see C. R. Hauser and W. H. Puterbaugh, J . Am. Chem. SOC.,76, 1068 (1963). (9) All temperatures are unoorrected. Mioroanalyses were performed by Mian Kenko Ogawa. (10) E. H. Farmer, C. D. Lawrenoe, and J. F. Thorp, J . Chsm Soc.. 729 (1928). ( 1 1 ) H.R. Billica and H. Adkins. Org. Sun., Coll. Vol. 111, 176 (1955).



1.1 g. (0.16 p a t o m ) of lithium and 300 ml. of liquid ammonia waa treated with 21 g. (0.15 mole) of t-butyl crotonate." After stirring for 4 hr. a t -50', a solution of 19 g. (0.15 mole) of benzyl chloride in 50 ml. of anhydrous ether was added and the stirring was continued for an additional 8 hr. a t -60'. The condensation product was hydrolyzed directly by heating with a mixture of 40 ml. of hydrochloric acid and 32 ml. of dioxane and the resulting acid was recrystallized from ligroin (b.p. 60-80') to afford 7.2 g. (36% based on benzyl chloride) of 2,2-dibenzyl-3-butenoic acid, m.p. 98.5-99.5". Anal. Calcd. for C l s H ~ ~ 0C,~ :81.17; H, 6.81. Found: C, 81.33; H, 6.80. The infrared absorptions (potassium bromide tablet); 1724 cm.-l (nonconjugated carboxylic acid), 995 and 920 om.-' (a vinyl group). The NMR spectrum, run on a 5 M carbon tetrachloride solution a t 56.4 Mc. with water aa an external reference, showed peaks: a singlet a t 6 -2.36, weight 10, assigned to protons of twophenyl groups; asextet centering a t 6 -1.25 (X) and two quartets centering at values of 6 -0.46 (B) and -0.36 (A), respectively, each weight 1, assigned to three protons of a vinyl group"; a quartet at 6 +1.70, weight 4, assigned to protons of methylene group which showed a typical AB coupling pattern with JAB 13.6 C.P.S. and AB 8.5 C.P.S. All observations support the structure given. In another run the reaction was carried out as above with 0.7 g. (0.1 g.-atom) of lithium, 300 ml. of liquid ammonia, 14.2 g. (0.1 mole) of t-butyl crotonate, and 12.7 g. (0.1 mole) of benzyl chloride. The halide waa added immediately after the addition of the ester. Upon distillation of the condensation products a fraction boiling a t 112-135'/4 mm. waa obtained and was hydrolyzed by heating with a mixture of hydrochloric acid and dioxane. Fractional recrystallizations of the acidic materials from aqueous ethanol afforded 5.5 g. (31y0) of 2-benzyl-2-butenoic acid, m.p. 96-99'. Recrystallizations from ligroin (b.p. 60-80') yielded a sample, m.p. lOl-lO2',14 which gave correct analyses for carbon and hydrogen. The infrared spectrum (Nujol mull) showed 1665 and 1630 cm.-' (cu,p-unsaturated carboxylic acid). The NMR spectrum, run on 4 M carbon tetrachloride solution at 56.4 hfc. with water as an external reference, showed absorptions: a quartet centering a t 6 -2.31, weight 1, assigned to an olefinic proton adjacent to terminal methyl group; a singlet a t 6 -2.30, weight 5, assigned to protons of phenyl group; a singlet a t 6 f1.18, weight 2, assigned t o protons of a benzylmethylene group; a doublet centering at 6 +2.91, weight 3, assigned to terminal methyl group which is coupled with the olefinic proton with J 7.2 C.P.S. These findings support the structure of 2-benzyl-2-butenoic acid. 2,2-DibenzyI-3-hutenoic acid could not be isolated in this run. Condensation of t-Butyl Crotonate with n-Butyl Bromide. -To a suspension of lithium amide prepared from 0.7 g. (0.1 g.-atom) of lithium and 300 ml. of liquid ammonia, 14.2 g. (0.1 mole) of t-butyl crotonate waa added at -40'. After 1 hr. stirring, 13.7 g. (0.1 mole) of n-butyl bromide was added and the stirring was continued at the same temperature for 10 hr. Upon distillation of the reaction product 8.2 g. of an oil boiling a t 75-90'/3 mm. was obtained. Microanalyses for carbon and hydrogen on this product gave values in agreement with those calculated for a nearly (12) R. IS. Baker and I?. G. Rordwell, OW. Syn., Coll. Vol. 111, 141 (1955). (13) For examples of such an unusual splitting of vinyl proton8 forming an ABX system into fourteen lines with a sextet in the X proton region, see E. Wenkert and P. Beak, J . Am. Chem. SOC.,83, 998 (1961) and also J. D. Roberta, "An Introduction to the Analysis of Spin-Spin Splitting in High-Resolution Nuclear Magnetic Resonanoe Spectra," W. A. Benjamin, Ino., New York, N. Y., 1961, p. 83. Quartets of A and B protons are normal and splitting8 are about 11 and 1.4 a.p.8. for B proton, and about 17 and 1.4 c.p.8. for A proton. (14) Lit., m.p. 99' [E'. Fichter and E. Alber, J . prokt. Chum. [Zl,74, 335 (1QOG)l.

JULY, 1'362


equimolar mixture of mono- and dibutylated &butyl crotonates. Hydrolysis of 7.5 g. of this oil by refluxing with a solution of 10 g. of potassium hydroxide in 40 ml. of aqueous ethanol from 6 hr. afforded 2.0 g. (15% based on n-butyl bromide) of crystalline acid melting a t 22-24'. Single recrystalliaation from petroleum ether (b.p. 30-60') afforded an analytical sample of 2-n-butyl-2-butenoic acid," m.p. 24.5-25.5', which gave correct analyses for carbon and hydrogen. The infrared spectrum (neat liquid) showed absorptions at 1680 and 1640 cm.-' (a,p-unsaturated carboxylic acid). No vinyl absorptions were observed. Catalytic hydrogenation of 0.71 g. of 2-n-butyl-2-butenoic acid was carried out at room temperature in 50 ml. of ethanol and in the presence of 0.5 g. of Raney nickel (W-5 grade).': The resulting 2-butylbutyric acid distilled at 88-95" (bath temperature)/l mm., and had ~ U D1.4280. Chlorination with thionyl chloride, followed by treatment with aqueous ammonia, afforded 2-n-butylbutyramide, m.p. 100-101' (lit.,:@ 101-102'), after recrystallizations from water. The amide gave correct analyses for carbon and hydrogen. Upon distillation of the neutral portion of the hydrolysis products under reduced pressure 3.3 g. (28% baaed on nbutyl bromide) of t-butyl 2,2-di-n-butyl-3-butenoate,b.p. 90-9lo/2.5 mm., nib 1.4385, waa obtained. Anal. Calcd. for CI~HSOO~: C, 75.53; H, 11.89. Found: C, 75.32; H, 11.87. The infrared spectrum (neat liquid) showed absorptions at 1720 cm.-: (nonconjugated ester) and at 1000 and 910 cm.-I (a vinyl group).

. as

Acknowledgment.-The authors are indebted to Mr. N. Hayakawa, Japan Atomic Energy Institute, for NMR measurements. His assistance is gratefully acknowledged. (15) R. Adams and L. J. Roll, J . Am. Clem. SOC.,83, 3469 (1931). These authors recorded the acid as an oil. (16) H. S. Raper, J. Chem. Soc., 91, 1831 (1907).


acceptors under simple reaction conditions. Indeed, the attempted synthesis of I X via the butoxide method a t 0' failed. In no case does the yield exceed 40010, adducts I11 and IV were obtained in highly impure form, while adduct I1 could not be prepared in sufficient yield to permit separation from concomitant impurities.8 In all cases, considerable quantities of quit,e high-boiling and presumably polymeric materials were formed. These difficulties may be obviated if methyllithium is employed rather than the butoxide.6 Adducts 11,V, VI, and X were produced in this manner. The methyllithium benzal chloride reaction may be carried out at 0' and gives little polymer, no impurities boiling in the same range as the desired product and higher yields. For example, X was obtained in only 17% yield by the butoxide reaction, but in 42% yield by the methyllithium route. Adduct 11, which could not be obtained by the butoxide method, resulted in 21% yield by the methyllithium method. The butene adducts V and VI were also preparable by this method only. Structures of adducts V, VI, IX, and X are in accord with NMR data. In all pairs the methyl analog exhibits a signal in the region 7.69-7.82 7 , whereas the remainder of the spectrum is identical in shape to that of the unsubstituted adduct. All adducts pussessed an infrared absorption in the region 1030-1000 cm.-l. In accord with the literature, the band nearest 1021 cm.-' was considered as characteristic of the cyclopropane ring.6 Experimental

Improved Production of Phenylchlorocarbene ROBERT A. Moss Department of Chemistry, University of Chicago, Chieago, Illinois Received February 6,1968

McElvainl and Breslow2generated phenylchlorocarbene by the action of potassium &butoxide on benzal chloride; the carbene, in each case, was trapped via addition to ketene acetals. We have prepared adducts I, 111,IV, VII, VIII, IX, and X in a similar manner (see Table I ) . The preparation of 11,IV, VI, VIII, and X required the prior synthesis of p-methylbenzal chloride, a white, crystalline solid (m.p. 50-51 O, from methanol) obtainable through reaction of phosphorus pentachloride and p-tolualdehyde. Production of 1-chloro-1-phenylcyclopropanesby the butoxide a-elimination method is limited. The reaction does not proceed well a t temperatures much below 70°, thus precluding the use of butenes (1) 8. M. McElvain and P. L. Weyna, J. Am. Cham. Soe.. 81, 2586 (1959). (2) R. Brealow, R. Haynie, and J. Mirra, i b i d . , 81, 247 (1959).

Olefins.-1-Hexene and isobutylene were Phillips Petroleum Cap. Tetramethylethylene was prepared by the method of Whitmore.7 trans-Hexene-3 was prepared by the method of Pomerana .* Methylenecyclohexane waa prepared by pyrolysis of cydohexylmethyl acetate at 500' on a glass-packed column. Benaal chloride waa Matheson technical grade, purified by distillation over a tantalum spiral column at 103'/28 mm. p-Methy1benzalchloride.-An 85-g. sample of phosphorus pentachloride wm placed in a two-neck 500-ml. flask, fitted with a dropping funnel and mechanical stirrer. The temperature was kept a t 25' by means of a water bath. A 50-g. sample of p-tolualdehyde waa added dropwise, with stirring over 1 hr. Stirring waa continued for 6 hr., care being taken to exclude direct light. The reaction mixture wm (3) These impurities absorb at 1700-1680 om.-* in the infrared. Though there is evidence that butoxide can cause elimination in ketene acetal adducts,*d we have treated V for eighteen hours with butoxide in refluxing hexene-1 without effecting any noticeable change. In addition, a prepared mixture of several adducts was not affected by methyllithium at Oo. We conclude that the impurities result from an alternate consumption of benzal chlorides when these are reacted with butoxide in the presence of a poor acceptor olefin. No such impurities were formed when the olefin wan a better acceptor or when methyllithium was employed. (4) IC. B. Wiberg, R. K. Barnes, and J. Albin, J . Am. Cham. SOC.,79, 4994 (1957). (5) G.L. Cloen and L. E. Closs, tbid., 88, 5723 (1960). (6) L. J. Bellamy, "Infrared Spectra of Complex Molecules," J. Wiley and Sons, Inc., New York, 1960,pp. 29f. (7) F. C. Whitmore and L. Black, J . Am. Cham. Soc., 84, 1819 (1942). (8) P.Pomeranz, A. Fookson,T. W. Mears, 5. Rothberg, and F. L. Howard, J . Res. Natl. Bur. Std., 82, 59 (1954).



YOL. 27

TABLE I B.P. (or m.p.), No.


7 L b ’



Infrared.* om. - 1


h ound



I . 5122



C 74.80 H 8.20 C1 I6 99

75.04 8.13 16 71






C1 15.92

14 78’





C 74.80 H 8.20 C1 16.99

75.14 8.19 16 50




\‘ I









C 75.48 H 8.60 C1 15.92

75 75 8.20 15.45





C1 19.62

19 63


1. .5230



c1 18 21






C 75.83 H 8.18 C1 15.99

76.32 7.85 15.87





C 76.41 H 8.55 C1 15.04

76.92 8.30 14.65





C 74.80 H 8.20 C1 16 99

74.86 8.18 17 19



17“ 42e


c 75.49 H 8.60 C1 15.92

75.52 8.97 15.60



\c-c Cn/ / / ‘CH~ Xd CH3

Yields reported are for the purified adducts, except adducts I11 and IT. This is the “cyclopropane” absorption (see text). e These adducts were not separated into their isomeric components. Purified by sublimation. e Made by methyllithium method. f Decomposes upon standing. Attempts to purify this compound by vapor phase chromatography resulted in pyrolysis. Chromatography on alumina destroyed the compound. Distillation under vacuum will not separate the apparent impurity. The 1% error in chlorine analysis is due to the instability of the compound, for a subsequent analysis of the same sample indicated a further 1056 of chlorine to about 13%. Made by butoxide method. a

then poured over water-ice, thoroughly agitated, and the resulting solid was quickly filtered, washed with water, and taken up in ether. The ethereal solution was washed with water, sodium bicarbonate solution, again with aater, and dried over sodium sulfate. The ether was removed under reduced pressure, and the solid material recrystallined from

the smallest suitable volume of hot methanol. The yield of white crystals melting at 50-51’ was 52.0 g. (72%). Anal. Calcd. for CsH8C12: C1,40.51. Found: C1,40.02. p-Methylbenzal chloride slowly hydrolyzes when exposed to the atmosphere and should be stored under nitrogen in a dark place.

JULY. 1962


Butoxide Method for Phenylchlorocarbene (I, 111, IV, 80-mmole sample of potassium tbutoxide was prepared in a 50-cc. three-neck flask, fitted with condenser and dropping funnel. The excess t-butyl alcohol waa removed under reduced pressure, 120 mmolea of the indicated olefin was added, and heat was supplied until the olefin began to reflux. A 40-mmole sample of benzd chloride waa added dropwise (p-methylbenzal chloride was added 89 a solution in 10 ml. of olefin) over a period of 30 min. The reaction mixture was maintained at reflux and stirred magnetically for at least 3 hr. In cases where the olefin boiled below 7 0 ° , the refluxing time was extended to upward of 5 hr. The product was washed three times with water, once with dilute hydrochloric acid, again with water, and dried over sodium sulfate. Olefin was removed under reduced pressure and the adduct vacuum distilled over a short Vigreux column. Redistillation afforded analytical samples. Methyllithium Method for Phenylchlorocarbene (11, V, VI, X.)-A 100-mmole sample of b e n d chloride and 1 mole of the indicated olefin (except adduct 11 where 40 m o l e s of halide and 120 mmoles of olefin were used) were put in a dried nitrogen-filled, three-neck, 250-ml. flask, fitted with a Dry Ice condenser and an addition tube connected to a storage-buret containing methyl lithium (1-2 N in ether). The base was added dropwise until an excess of 20% was obtained; the reaction mixture was maintained a t 0" by an ice bath and was magnetically stirred. After addition wm completed, excess olefin waa allowed to evaporate, and the mixture was then washed three times with water and dried over sodium sulfate. Ether was removed under reduced pressure and several vacuum distillations afforded the p r o d u ~ t . ~

VII, VIII, IX, X).--An


those of the other two. The melting point of I was also closest to those reported for the 1,2,3,4-isomer. A possible mechanism for the formation of 1,2,3,4-tetrachloronaphthaleneis the initiation of a polymer chain of two units of trichloroethylene by a phenyl radical followed by ring closure a t the ortho position with subsequent dehydrohalogenation. The experimental conditions were such that all of these processes could occur to a small extent which is reflected by the low yield. Experimental Benzoyl peroxide (20 g.) was added in 1-g. portions to refluxing trichloroethylene (16.8 moles) over a period of 96 hr. The temperature of the reaction mixture rose from 88 to 100" during this time. The reaction mixture was then flash distilled to remove the unchanged trichloroethylene (9.9 moles). The remainder waa heated to 150-160" at rttmospheric pressure for 12 hr. and hydrogen chloride was slowly evolved. Distillation at 3 mm. pressure then gave pentachlorobutadiene (0.32 mole), hexachlorobutene (3.0 moles), and 210 ml. of a viscous residue. The final temperatures for vapor and residue were 80 and 110". 1,2,3,4-Tetrachloronaphthalene (5 9.) wm recovered from the residue by filtration. Recrystallization from ethyl ether gave oolorless crystals, m.p. 194-200" (lit., 196O,a 198°,4 and 199-2OOo6), Anal. Calcd. for Cl~H4C&: C, 45.15; H, 1.52; C1,53.33. Found: C, 45.14; H, 1.76; C1,53.33. ~~

Acknowledgment.-The author wishes to acknowledge the invaluable counsel of Dr. Gerhard Gloss and the financial stability lent by a National Science Foundation Cooperative Fellowship. (9) Some styrene or p-methylstyrene is always formed in this reaction because of the reaction of methyllithium with the arylchlorooarbene. Styrene formation may be minimized by slow addition of the methyllithium.

1,2,3,4-Tetrachloronaphthalenefrom Trichloroethylene and Benzoyl Peroxide WILLIAM L. HOWARD AND RONALD E. GILBERT Organic Basic Research Laboratory, The Dow Chemical Company, Freeport, Texas Received February 6, 1961

(2) L. Cencelj and D. Had% Spectrochim. Acta, 7 , 274 (1955). (3) E. G. Turner and W. P. Wynne, J. Chem. Sac., 243 (1941); W. P.Wynne, (bid., 61 (1946). (4) J. v . Braun with 0. Braunsdorf, P. Engelbertz, E. Hahn, G. Hahn, 0. Hainbach, W. Kredel, and K. Larbig, Bar., 16B,2332 (1923). (5) A. A. Danish, M. Silverman, and Y. A. Tajima, J. Am. Chem. Sac., 76, 6144 (1954).

Studies on the Leaves of the Family Sdicaceae. I. Populin from the Leaves of Populus grandidentata and Populus tremuloides IRWINA. PEARL, STEPHEN F. DARLINQ, AND OLIVER JVSTMAN The Institute of Paper Chemistry, Appleton, W i s m i n

Received Februury 8, 1968

During a study of the action of benzoyl peroxide on trichloroethylene, a colorless crystalline compound (I) of empirical formula CsHzClz was isolated after the separation of t'he maior products pentachlorobutadiene and hexachlorobutene.' Doubling this formula gives the molecular formula of tetrachloronaphthalene, and a survey of the literature revealed three known tetrachloronaphthalenes with melting points near that of I. The infrared spectrum of I corresponded exactly with that given in the literature for 1,2,3,4-tetrachloronaphthalene2 and was markedly different from

The isolation of a new glucoside, grandidentatin, from the bark of the bigtooth aspen (Populus grandidentata) in small quantities was reported recent1y.l In the course of the determination of the structure of grandidentatin, it was necessary to obtain larger amounts of the new glucoside. Because some of the glucosides of several species of Salix and Populus occur in both the bark and the leaves of these species? the leaves of bigtooth aspen were suggested as a possible source of grandidentatin in larger amounts. The present paper reports the first studies on the isolation of glucosides from the

(1) A. Roedig and R. Moss, Chem. Ber., 90, 2902 (1957);0.Simamura and N. Inamoto, Bull. Chem. Sac. Japan, 27, 152 (1954).

(1) I. A. Pearl and 8. F. Darling, J. Ore. Chem., submitted for publication.



leaves of bigtooth aspen and the related quaking aspen ( P . tremuloides). Freshly obtained leaves from a bigtooth aspen felled in July in Appleton, Wisconsin, were extracted with boiling water, and the hot extract was treated with excess basic lead acetate and allowed to stand overnight. The mixture was filtered, and the clear filtrate was deleaded in the usual manner with hydrogen sulfide. The resulting clear aqueous solution was Concentratedin a laboratory circulating evaporator under reduced pressure. Shortly after evaporation began, crystals separated in the concentrated liquor. These crystals were identified as populin, the 6-monobenzoate of salicin. The filtrate from these crystals was processed by procedures employed previously for aspen barks. Although many unknown compounds were indicated by chromatography, only salicin was isolated as another identified glucoside from the extract. The h d i n g of populi in the leaves of P. grandidentuta was unexpected because earlier studies2 led us to believe that the barks of the American Populus species did not contain populin, and we assumed that the glucoside content of the leaves of a particular species would resemble the glucoside content of the bark of the same species. Subsequently, however, we found populin by Craig machine fractionation of the extractives of P. grandidentata bark,* thus supporting our original assumption. In order to test this assumption further, we subjected to water extraction the July leaves of P. tremuloides, a species whose bark has never yielded p o p ~ l i n . The ~ ~ ~hot water extract of this species, when processed in the manner noted for bigtooth aspen leaf extract, yielded crystalline populin, but in somewhat less yield. As in the case of the bigtooth aspen leaf extract, the filtrate from the populin yielded salicin and a similar mixture of unknown compounds. Thus, the leaves of thi: American quaking aspen (P. tremuloides) provide a good source of the 6-monobenzoate of salicin, populin, while the bark of the same species provides a good source of the 2-monobenzoate of salicin, tremuloidin.6 To investigate the possibility that the presence of populin in the leaves of these two Populus species was a function of the time of collection, leaves of P. grandidentata from the same clone were collected in September and in October. Hot water extracts of these two samples of leaves gave essentially the same results as the earlier July leaf extract, indicating that populin exists in the leaves of bigtooth aspen as a product of biosynthesis and not as an intermediate. Qualitative analysis of the ether-soluble phenolic materials liberated from the lead precipitates from (2) I. A. Pearl, 8. F. Darling, H. DeHaas. B. A. Loving, D. A. Scott, R. H. Turley, and R. E. Wirth, T a p p i , 44,475 (1961). (3) I. A. Pearl, 0. Justman, D. L. Beyer, and D. Whitney, Tappi. submitted for pubdcation. (4) I. A. Pearl and 8. F. Darling, J . Ore. Chsm.. 24, 1616 (1959). (5) I. A. Pearl and 8. F. D~rling,J. Ow. Chsm., 24, 731 (1959).



both the bigtooth and quaking aspen leaf extracts indicated the presence of vanillic acid, p-hydroxybenzoic acid, p-coumaric acid, ferulic acid, vanillin, and acetovanillone in both extracts, and syringic acid and syringaldehyde in the bigtooth extract only. This lack of the two syringyl compounds in the quaking aspen leaf extract may be of taxonomic significance. Experimental8

Isolation of Populin from Bigtooth Aspen Leaves.-An amount of 1490 g. (oven-dry basis) of leaves freshly obtained from a 13-year-old bigtooth aspen (Populus grandidentutu) felled in Appleton, Wisconsin, on JuIy 21, 1960, was digested with 30 1. of boiling water and filtered hot through cloth. The hot filtrate was treated with an excess of basic lead acetate, stirred, and allowed to stand overnight. The yellow precipitate was filtered, and the clear filtrate was saturated with hydrogen sulfide, heated to boiling, and filtered. The almost colorless filtrate was concentrated in a laboratory circulating vacuum evaporator. Shortly after evaporation began, crystals separated in the concentrated liquor. Concentration waa continued until the concentrated liquor became thick with precipitated crystals. The crystals were filtered and washed with water to yield 10.0 g. of white These crystals were recrystalneedles melting a t 180-181'. lized from water to give needles with unchanged melting point and with specific rotation, infrared absorption spectrum, and Rf values identical with those reported earlier5 for synthetic populin prepared by the procedure of Richtmyer and Yeakel' and for natural populin isolated from Populus alba leaves by the procedure of Herberger.8 A mixed melting point with synthetic populin was not depressed. Further concentration of the aqueous filtrate yielded another 1.0 g. of populin melting a t 178-179'. The total yield of populin amounted to0.74% on the basis of the original leaf solids. Further Evaluation of Bigtooth Aspen Leaf Extract.The concentrated aqueous filtrate from the last populin crystals wm evaporated further to 500 ml. and extracted exhaustively with ethyl acetate. The ethyl acetate was filtered to yield 9.7 g. of white crystals melting a t 193-194'. These were recrystallized from ethanol to give white crystah of salicin melting a t 196-197' and having specific rotation and infrared absorption spectrum identical with those of authentic salicin. The ethyl acetate filtrate was chromatographed a t 20' in 10:3: 3 butanol-pyridine-water, and the chromatog r a m were sprayed with the modified silver6 and diazotized p-nitroaniline9 spray reagents. Spots were located at Er's 0.60, 0.72, and 0.85 with the silver reagent and at 0.64 (redviolet), 0.79 (violet), and 0.86 (purple surrounded by redviolet) with the diazo reagent. No spots for either grandidentatin or salireposide were found. The aqueous raffinate from the ethyl acetate extraction was spotted on replicate papers and chromatographed in the butanol-pyridine-water and 8: 2 :1 ethyl acetate-pyridinewater developers. The chromatograms were sprayed with modified silver, diazotized p-nitroaniline, modified aniline hydrogen phthalate,'" and urea11 spray reagents. These chromatogram indicated the presence of salicin, sucrose, (6) All melting points are uncorrected. Infrared abaorption spectra were determined by Mr. Lowell Sell of The Institute of Paper Chemistry Analytical Department. (7) N. K. Richtmyer and E. H. Yeakel, J. Am. Chem. Soc., 56, 2495 (1934).

( 8 ) J. E. Herberger, Buchners Report Pharm., 61, 266 (1835). (9) I. A. Pearl and P. F. McCoy, Anal. Chem., Sa, 1407 (1960). (10) 8. M. Partridge, Nature, 164, 443 (1949). (11) C. 9. Wise, R. J. Dimler, H. A. Davia, and C. E. Rist, Anal. Chsm., 81,33 (1955).


JULY,1962 glucose, fructose, some unknown glucosides and phenolic materials, and trace amounts of several oligosaccharides. Hot Water Extraction of Quaking Aspen Leaves.-Leaves were freshly obtained from a quaking aspen ( P . tremuloides) I 1 years old felled in Appleton, Wisconsin, on July 25,1960. A sample of 870 g. (ovendry basis) was extracted with boiling water and processed in the same manner to yield 5.1 g. of populin identical in all respects with that obtained from the bigtooth aspen leaves. The yield amounted to 0.58% on the basis of the original leaf solids. The ethyl acetate extract yielded all the spots on paper chromatograms noted for the analogous bigtooth aspen leaf extract plus two purple spots with the diazo reagent at R!‘s 0.57 and 0.72. Populin from September and October Leaves of Bigtooth Aspen.-Green leaves from a P. grandidentutu felled on September 22, 1960, and yellow leaves from a P. grandidentutu collected on October 10, 1960, were processed in the manner described. These two bigtooth aspens were from the same clone as the July 21 tree. The September 22 leaves yielded 0.83% populin and the October 10 leaves yielded 0.54yo populin based on the original oven-dry leaves. Further processing of both aqueous extracts gave results entirely similar to those obtained with the earlier leaves of bigtooth aspen. Evaluation of Insoluble Lead Salts.-The precipitated lead salts from the original bigtooth aspen leaf extract were covered with 4 1. of water and treated with hydrogen sulfide with vigorous mechanical stirring. With continued hydrogen sulfide introduction, the mixture was heated to boiling and then boiled without hydrogen sulfide introduction. The hot mixture was filtered, and the cooled filtrate was extracted with ether to yield 0.25% of ether extractives baaed on original leaf solids. The ether extractives were chromatographed qualitatively on paper in 10:3:3 butanol-pyridinewater, butanol saturated with 2% aqueous ammonia, 0.3 N sulfurous acid, and benzene saturated with formic acid developers, and the chromatograms were examined by means of fluorescence under ultraviolet light and 2,4-dinitrophenylhydrazine and diazotized p-nitroaniline spray reagents as outlined earlier.” These qualitative chromatograms indicated the presence of substantial amounts of vanillic acid, syringic acid, p-hydroxybenzoic acid, p-coumaric acid, ferulic acid, vanillin, synngaldehyde, and acetovanillone. In addition, several unidentified phenolic compounds were indicated. Similar processing of the lead salts from the original quaking aspen leaf extract yielded essentially the same results except that syringic acid and syringaldehyde were absent. (12) I. A. Pearl, D. 1., Beyer, B. Johnaon, and 9.Wilkinson, T o p p i , 40,374(1957).

Palladium-Catalyzed Decarbonylation of trans-a- Substituted Cinnamaldehydee NORMAN E. HOFFMAN, A. T. KANAKKANATT, AND R. F. SCHNEIDER

Department of Chemistry, Marquette University, Milwaukee 3, Wisconsin Received February 8, 1962

Palladium has been used as a catalyst for liquid phase decarbonylation of aldehydes’s2 and olefin double bond isomerization.8 The purpose of the (1) H. E. Eschinazi, BUZZ. doc. ehim. France, 967 (1952). (2) J. 0.Hawthorne and M. H. Wilt, J . Org. Chem., 26,2215 (1960).


work presented herein was to determine the stereoselectivity of decarbonylation of a,p-unsaturated aldehydes. The investigation appeared worth while in view of the proximity of the olefinic portion of the aldehyde to the reaction center during decarbonylation and, therefore, the possibility of simultaneous configurational oleiinic isomerization at the active palladium site. trans-Cinnamaldehydes (I) were decarbonylated and the product examined for the normal cis olefin (11) and the isomerized trans olefin (111). Decar-


I1 R = CHa, CIHS,i-CsH7, n-C&

I11 Cas

bonylation was performed in two ways. In method A the product was distilled as rapidly as possible after formation, and in method B the reaction mixture was distilled to remove product only after decarbonylation was complete. Results are shown in Table I. The results show that in the majority of cases decarbonylation by method A gave mainly the normal cis product. If it were possible to remove the product immediately after formation the ratio of cis to trans isomer would be higher, and possibly the cis isomer would be the sole product. When the initial reaction product was allowed to remain in contact with the catalyst until decarbonylation was complete, method B, I11 became the main olefin product. Infrared examination of the unchanged aldehyde showed no isomerization in the reactant. Formation of the trans olefin, therefore, resulted from isomerization of 11. That the particular catalyst used in this work was capable of isomerizing olefins in the absence of aldehydes is shown in Table I1 where the results of isomerization of allylbenzene and cis-p-methylstyrene are presented. cis-trans Isomerization was rapid compared to decarbonylation, and double bond migration was roughly comparable in rate to decarbonylation. On prolonged contact with the catalyst a saturated-side-chain product also appeared in decarbonylation except in the case of a-methyl- and a-phenylcinnamaldehyde. The reduction was slow compared to decarbonylation especially at lower temperatures. Thus a-methylcinnamaldehyde was completely decarbonylated before a detectable amount of n-propylbenzene was formed. However, failure of a-phenylcinnamaldehyde to undergo this reaction cannot be attributed to the same cause and must be related to the structural requirements of the reaction. (3) G. Egloff, G. Hulls, and V. 1. Komarewskg, “Isomerisation of Pure Hydrooarbona,” Reinhold Publishing Corp., New York. N. Y., 1942,pp. 264,366.



VOL. 27


Temp.," OC.


Time,b HI.

-% Ci8

Styrene^ trans

% Alkyl benzene

% YieldC

A CHI 180 23 85 16 0 92 CH3 B 180 24 10 90 0 48 CaH6 A 210 23 61 34 4 85 CIHE B 200 46 10 21 67 49 n-C3H7 A 200 18 35 41 22 48 n-C3H7 B 210 74 8 12 77 49 e i-C3H7d A 205 35 48 5 48 i-CsH7d B 210 45 3 13 62 48 CBHS 4 78 22 0 I A 200 CeH6 B 190 69 64 35 0 42 a Initial temperature. Reactions were run overnight. The times listed show the interval from the start of reaction to the following morning or one or two mornings later, when the reaction was found completed. Based on a distillation fraction of wide enough boiling range to include olefin isomers and saturated side chain product. Two compounds other than those listed were formed, presumably olefin isomers. However, the infrared spectra of the compounds did not appear to have bands expected for the isopropenyl or isopropylidene group. e Not recorded. Reaction run only to 15% completion.


TABLE I1 C3-ALKENYLBENZENES Reflux temp.; 1 wt. % catalyst



Allylbenzene Allylbeneene Allylbenzene cis-p-Methylstyrene cis-p-Methylstyrene

Time, hr.

1 3

8 0.5 2.5

% Allylbeneene

84 72 51 1 1

Methylstyrene cis trans

4 5 9 82 51

12 23 43 15 48

Pure alkenylbenzenes also gave this reaction, and, therefore, the alkylbenzenes arose after decarbonylation, and the reaction is neither an aldehyde reaction nor is it necessarily promoted by an aldehyde. CaHsCHeCHR

+ 2H +CeH$H:CH?R

The hydrogen source in this reaction must be the alkenylbenzene. A product of approximately twice the molecular weight of the alkenylbenzene was formed in the disproportionation. The structure of this product and the nature of the reaction are currently being studied. The formation of alkylbenzenes during decarbonylation is not the first example of an unexpected result in decarbonylation. Hawthorne and Wilt have observed extensive dehydrogenation accompanying the decarbonylation of heptaldehyde. t These results along with the observed cis-trans isomerizations point up the caution that should be exercised in using palladium-catalyzed decarbonylation in structure proof and synthetic work. Because of the ease of preparation of the transcinnamaldehydes and the stereoselectivity of the decarbonylation, a combination of these two (4) The cis-methyl isomer 18 separable from cis rich cis-trans mixtures by e5cient distillation. See R. Y. Mixer, R. F. Heck, S. Winstein, and W. G. Young, J . A m . Chem. Soc., 76, 4094 (1953). We were able to get aome separation of ethyl isomers, as noted by changes in the infrared spectra of the fractions, with only a crude distillation from an 8-in. Vigreux column.

methods appears to be a, simple way of preparing [email protected] and cis-P-ethylstyrene. Although yields in the decarbonylation of the other cinnamaldehydes were not very high, the ratio of cis to trans olefin product was greater than the equilibrium ratio. For this reason decarbonylation could be useful in preparing these cis isomers also. Experimental6 Aldehyde Preparation and Configuration.-a-Substituted cinnamaldehydes were prepared by the base-catalyzed condensation of benzaldehyde with the proper a-unsubstituted aldehyde according to the method of Kraft.6 These methods were used to establish configurational homogeneity of the cinnamaldehydes: ( 1 ) semicarbazones of the aldehydes, prepared in yields of 55-659',,, were repeatedly recrystallized and their infrared spectra examined for changes; (2) cinnamic acids, prepared by silver oxide oxidation of the aldehydes, in yields of 80-90%, were repeatedly recrystallized and their infrared spectra examined for changes; and (3) the aldehydes were gas chromatographed on 20% Silicone Dow 11 on firebrick to separate geometric isomers. All methods showed the aldehydes were homogeneous. The configuration of the aldehydes except a-isopropylcinnamaldehyde was established aa trans through oxidation to known trans-cinnamic acids.?-P a-Isopropylcinnamaldehyde resisted silver oxide oxidation. It was assumed .to be trans by analogy between the product distribution in its decarbonylation and that of the other cinnamaldehydes. Table I11 lists melting points of the cinnamaldehyde derivatives. The semicarbaeone of trans-a-n-propylcinnamaldehyde is a new compound. Anal. Calcd. for C13H17N30: C, 67.53; H, 7.41; N, 18.18. Found: C, 67.37; H, 7.30; N, 18.30. Decarbony1ation.-Aldehydes, 40 to 100 g., were mixed with 1 wt. 7, of a 10% palladium in charcoal catalyst. The catalyst was prepared according t o the method of Mozingolo with Norit SG Extra charcoal.11 Reaction was conducted (5) Microanalyses were performed by Schwarakopf Microanalytical Laboratory. Melting points are uncorrected. (6) W. M. Kraft, J . Am. Chem. Soc., 7 0 , 3569 (1948). (7) R. Stoermer, Ann., 409, 13 (1915). (8) R. Stoermer and G. Voht, ibid., 409, 36 (1915). (9) M. T. Bogert and D. Davidson, J . A m . Chem. Soc., 64, 334 (1932). (10) R.Moeingo, O w . Syn.. 86, 77 (1946). (11) The chsrooal was a gift from the American Norit Co.




-SemioarbasanFound, OC.

-tranr-Cinnamio Found, O C .

Lit., OC.

Reactions of 2H,3H-Thieno[3,2-b]pyrrol-3one. V.lv2 The Reaction of 2,S-Disubstituted Thieno[3,2-b]pyrroles with Oxalyl Chloride


DERIVATIVES Cinnamaldehyde




209-210 207-2OSa 79-80. 5b 81-82‘ Methyl Ethyl 212-214 216-217‘ 103.5-105.5 104‘ +Propyl 179.5-181 e 93-94e 93J i-Propyl 193-195 191-‘&’ Phenyl 199-201 19P195h 173-173.5 172’ a K. von Auwers, Ber., 45, 2764 (1912). Mixed m.p. 79-81.5. Ref. 8. Y. Deux, Compt. rend., 208, 1090 (1939). * Neut. equiv. calcd.: 190.2; found, 191.5. f Ref. 9. P. Shoruigin, V. Isagulyantz, E. Smolyaninova, K. Bogacheva, and S. Skoblinskaya, J. Russ. Phys. Chem. SOC.,62, 2033 (1930); Ckm. Abstr., 25,4247 (1931). H. Ref. 7. Burton, J. Chem. SOC.,748 (1932).

in a Vigreux distillation apparatus under a vacuum chosen to give reflux a t the desired initial temperature. When reaction was allowed to go to completion before distillation, completion was detected by a leveling off of temperature below the initial temperature. Products were redistilled through an 8-in. or 14-in. Vigreux column. Analysis.-The components of the reaction product were separated and isolated by the use of a 5 ft. I/Z in. 20% Ucon polar on fkebrick gas chromatographic preparative column operated about 50’ below the boiling point of the lowest boiling component in the mixture. The reaction products were then analyzed quantitatively by use of a 5 ft. I/, in. Ucon polar on firebrick column. The sample volume-area constants of the components for this column were determined by injecting the pure components previously separated by the preparative column. For qualitative analysis the infrared spectra of the components separated by the preparative column were used. Characteristic olefin bands in the infrared spectral*J* of the pure olefin products were used to identify their geometric configuration. Alkylbenzenes were identified by comparison of their spectra to cataloged infrared spectral‘ or spectra of samples prepared by a combination of Friedel-Crafts acylation and Clemmensen reduction. Isomerization of Alkenyl Benzenes.-Isomerizations were run under nitrogen with 5 cc. alkenyl benzene samples having 0.050 g. of the catalyst mentioned above. Time errors were minimized by inserting the samples in an oil bath maintained a t 180-200° and quenching the reaction mixture in an ice bath after the proper time interval. The samples were then analyzed by gas chromatography employing the I/, in. Ucon polar column previously mentioned. Disproportionation of Alkenyl Benzenes.-When allylbenzene, a-methylstyrene and 8-ethylstyrene were refluxed under nitrogen with 10 wt. % catalyst, amounts detectable by gas chromatography of n-propyl-, isopropyl-, and nbutylbenzene respectively were formed. A fraction, b.p. 160-161’ (3 mm.) was isolated from the 8-ethylstyrene disproportionation. Anal. Calcd. for C ~ O H ZC, ~ : 91.54; H, 8.45. Found: C, 91.90; H, 8.43. Mol. wt. calcd. for CZOHZZ: 262. Found by benzene boiling point elevation in a McCoy apparatus capable of an accuracy within 5 to 10%: 270.

Acknowledgment.-This work was supported by a grant from the National Science Foundation. (12) L. J. Bellamy, “The Infrared Spectra of Complex Molecules,” John Wiley and Sone. Ino.. New York, N. Y.,1958, Chap. 3. (13) E.L. McMurry a n d V . Thornton, A n d . Chem.,24, 318 (1952). (14) “Catalog of Infrared Spectral Data,” American Petroleum Institute Research Project 44, Carnegie Institute of Technolog?.


.4lh~rl A Q V ~ S Laborntor!/,


ersitii o /


Urbana, Illinoas Received February 9, 1962

The recent synthesis of 2-carbethoxy-3-hydroxythieno [3,2-b]pyrrole (11)3 by carbethoxylation of 2H,3H-thieno [3,2-b]pyrrol-3-one (I) has made available a product which might be expected to be of value in the preparation of 6-subst’ituted thieno[3,2-b]pyrrole derivatives. It was shown3 that the dimeric enol ester I1 can be converted readily to the corresponding 3-tosyloxy (111) and 3-acet’oxy (IT’) derivatives; appropriate substitution reactions on these two products, followed by detosylat’ion or deacetylation, decarbethoxylation, and aromatization of the sulfur-containing ring by methods employed earlier,314should therefore provide routes t’o 5- or 6-substituted thieno [3,2-b]pyrroles which are not readily accessible by direct substitution reacti0ns.~-8 I n studies still in progress,8 I11 has been convert’edto the 6-formyl and 6-dimethylamino derivatives under the reaction conditions commonly employed in the preparation of analogous 3-substituted indoles. The present report concerns a study of the reactivity of I11 and IV towards oxalyl chloride, which in the past has found wide application in the synthesis of tryptamine analogs via the corresponding 3-acylindoles.~ 4 - 2 ~ (1) For the preceding paper, see G. W. Michel and H. R. Snyder, J . Org. Chem., 27, 2034 (1962). (2) This investigation was supported in part by a grant [C3969Bio] from the National Cancer Institute, Public Health Service. (3) W. Carpenter and H. R. Snyder, J . Am. Chem. Soc., 82, 2592 (1960). (4) D. S. Matteson and H. R. Snyder, d . Ore. Chem., 22, 1.500 (1957). ( 5 ) J. F. Zack, Jr., thesis, Doctor of Philosophy, University of Illinois, 1956. (6) W. Carpenter, thesis. Doctor of Philosophy, University of Illinois, 1959. (7) A. D. Josey, R. J. Tuite, and H. R. Snyder, J . A m . Chem. Soc., 8 2 , 1997 (1960). (8) J. Witt, Jr., thesis, Doctor of Philosophy, University of Illinois, 1961. (9) B I . Giua, Gam. chim. ital., 64, 593 (1924). (10) M. 8. Kharasch, S. S. Kane, a n d H. C. Brown, J . Am. Chem. Soc., 62, 2242 (1940). (11) M. E. Speeter and W. C. Anthony, i b i d . , 7 6 , 6208 (1954). (12) T. Nogradi, Monotsh. Chsn., 88, 768 (1957). (13) F. V. Brutcher, Jr., and W. D. Vanderwerff, J . Ore. Chem., 23, 146 (1958). (14) X. N. F. Shaw, A. McMilIm, A. G. Gudmondson, and hl. D. Armstrong, J . O w . Chem., 2 3 , 1171 (1958). (15) R. B. Woodward, F. E. Bader, H. Bickel, A. J. Frey, and R. W. Kierstead, Tetrahedron, 2, 1 (1958). (16) A. F. Ames. et al.,J . Chem. Soc., 3388 (1959). (17) F. Benington, R. D. Morin. and L. C. Clark, Jr., J . O w . Chem.. 26, 1 x 2 (1960). (18) G . Frangatos, G. Kohan, and F. L. Chubb, Can. J . Chem., 98, 1484 (1960).

2690 to the a- and 8-protons of the pyrrole ring; because of the greater electronic unshielding effect of the nitrogen btom on the a-position, the lower r-value must be assigned to the 5-proton. In the spectrum of either VI or VI11 there is only one peak in the region attributable to resonance of the 5-proton of the pyrrole ring ( r = 1.67 and 1.62, respectively) and no signal appears in the region where the 6-proton would be expected to absorb. The low r-values of the 5-protons in both VI and VIII, as compared to IV and other 5,6-unsubstituted thieno [3,2-b] pyrroles12*can be explained by the increased unshielding effect caused by the neighboring side chain.29 The combined spectral data are consistent only with structures VI and VIII. Both the acid chloride VI and the ester VI11 VII. R = p-CHjCeHk302 V. R = p-CHaCaHrSOo could be hydrolyzed in the presence of base to form R' = OCeHs vr. R = COCHI the a-keto acid IX in excellent yield. A solution of VIII. R = COCHo dibenzylamine in anhydrous ether very readily R' = OCzHs IX. R = H reacted with VI a t room temperature, affording a R' = OH practically quantitative yield of the corresponding X. R = COCHa R' = N(CH&eHs)a glyoxylamide for which structure X is indicated. XI. R = H It was found that X can be converted to 2-carbethoxy-3-hydroxy-6-( N,N-di benzy1glyoxylamido)thieno[3,2-b]pyrrole (XI) in a yield of 91% when heated with an equimolar amount of sodium hyTreatment of IV2swith an excess of oxalyl chlo- droxide in ethanol, followed by acidification; reride in tetrahydrofuran solution resulted in the moval of the 0-acetyl group with the liberation of formation of a yellow acylation product (VI) in the hydroxyl group was proved by the fact that XI good yield. But whereas indole and substituted gave a positive ferric chloride test. Other evidence indoles are known to react with oxalyl chloride in a for the structure was obtained from chemical analymatter of minutes, compound IV required several sis and spectral data (see Experimental). Attempts hours under comparable conditions for complete t,o decarboxylate XI under conditions previously316 reaction. When VI was heated briefly in absolute applied to other 2-carbethoxy-3-hydroxythienoethanol, the acid chloride was converted to the ethyl [3,2-b]pyrrole derivatives have been unsuccessful so far. Results of further studies will be presented ester VI11 in an over-all yield of 64%. The structure of VI11 was confirmed by the in- a t a later date. frared spectrum, which showed four characteristic 2 Carbethoxy 3 tosyloxythieno [3,2 blpyrrole carbonyl absorptions a t 1780, 1725, 1695, and 1620 (111) was allowed to react with oxalyl chloride, but cm.-l, tentatively assigned to the 0-acetyl group, in order to effect substitution of I11 it was necessary to the two carbethoxyl functions and to the a-keto to increase the length of reaction time (six days) as grouping, respectively, of the substituents. Evi- compared to that found suitable for the acylation of dence for substitution in the 6-position was given IV. The resulting glyoxylyl chloride V was isolated by the nuclear magnetic resonance (NMR) spec- without characterization and immediately treated trumZ6of VI and VI11 and their comparison with the with an excess of 95% ethanol to give the more spectrum of IV. The spectrum of IV exhibits two stable glyoxylic ester VI1 in 83% yield. Treatment doublets of equal intensity centered a t 7-values of of an ethanolic solution of VI1 with an equimolar 2.96 and 3.65, which are attributed," respectively, amount of sodium hydroxide resulted in the hydrolysis of the a-keto ester grouping and the formation (19) A. Buaas, C. Hoffmann, and G. RBgnier, Bull. 8oc. chzm. France, of the corresponding acid; its composition and in643 (1960). (20) H. G. Schlossberger and H. Ruch, Chsm. Bar., 93, 1318 (1960). frared spectrum are in accordance with structure (21) H. Plieninger and W. Mllller, ibid., 93, 2024 (1960). XII. (22) P. F. Rossi and 9. Sorassi, Ann. chim. (Rome), 61, 64 (1961).


(23) G. Domschke and H. Fllrst, Chsm. BeT., 94, 2353 (1961). (24) See aleo F. Lingens and H. Hellmann, A ~ Q S W Chem.. . 69, 97 (1957); F. Weygand and H. J. Bestmann, ibid., 72, 546 (1960). (25) For preliminary studies, see ref. 6. (26) The NMR spectra were recorded by MI. Oliver W. Norton with a Varian Associates high resolution spectrometer (Model V-4300 B with superstabilizer) at a frequency of 60 Mc. per second. Spectra were obtained in 20% solutions of either chloroform or dimethyl sulfoxide (DMSO) with tetramethylsilane a8 an internal standard. Chemical shifts are expressed as shielding values r in parte per million as defined by G. V. D. Tiers [ J . Phys. Chsm., 62,1151 (1958)l.

- -


(27) For a study of NMR spectra of thieno[3,2-b]pyrroles,aee R . J. Tuite, H. R. Snyder, A. L. Porte and H. S. Gutowsky, ibid., 6S, 187 (1961); further referencea under R. J. Tuite, A. D . Josey, and H. R. Snyder, J . A m . Cham. Soc., 82, 4360 (1960); R. J. Tuite, thesis, Doctor of Philosophy, University of Illinois, 1960. and ref. 8. (28) A list of NMR data on a number of 5,6-unsubstituted thieno[3,2-blpyrroles can be found under references 6, 8, and 27. (29) Similarly, the 2-protonyof indole-3-aldehyde shows resonance a t a value which is 1-61 runitu lower than the value of the u-proton in indole: aee d u o ref. 8 for analogous observations.


JULY, 1962 Experimental80

2-Carbethoxy-3-acetoxythieno[3,2-b]pyrrole(IV).-The acetate IV was pwpared in 96% yield from 2-carbethoxy-3hydroxythieno[3,2-b]pyrrole (11) according to the procedure described by W. Carpenter and H. R. Snyder.8 The infrared spectruma' (KBr pellet) of IV exhibits major absorptions a t 3320 (N-H), 1748 (acetoxy C=O), 1694 (carbethoxy C=O), 1280, 1236, 1220, and 1188 cm.-l (C-0 stretch). NMR spectrum (20% CHCb): 7 = 0.76, sa* (N-H); 2.96, d (5proton); 3.65, d (6-proton); 5.71, q and 8.67, t (e.p.); 7.75, s ( S A C ) . 2-Carbethoxy-3-acetoxythieno[3,2-b]pyrrolyl-6-glycoxyly1 Chloride (VI).-To a solution of 0.506 g. (2 mmoles) of IV in 10 ml. of anhydrous ether and 2 ml. of anhydrous tetrahydrofuran (THF) at room temperature, 0.40 ml. (4.7 mmoles) of oxalyl chloride was added dropwise (syringe) over a period of 1 min.; the reaction mixture immediately turned yellow. After standing in a tightly stoppered flask for 1.5 hr. at 30', the solution slowly deposited yellow needles upon scratching. The mixture waa allowed t o stand a t room temperature for another 7 hr. and finally kept a t -7" in the refrigerator overnight. The resulting yellow crystals of VI were collected, rapidly washed with anhydrous ether and dried in vacuo over phosphorus pentoxide and potassium hydroxide; 0.57 g. (83% yield), b.p. 178-180" dec. A sample of VI was analyzed without further purification. Infrared (KBr pellet): 3380 (N-H), 1780 (acetoxy C 4 ) , 1765, and 1660 (CO-CO-Cl), 1695 cm.-1 (carbethoxy C=O). NMR (2001, DMSO): T = -3.13, s (N-H); 1.67, d (5proton); 5.?4, q and8.69, t (e.p.).aa Anal. Calcd. for ClrHloNO23Cl: C, 45.42; H , 2.93; N, 4.07. Found: C, 45:69i-H,3.21; N, 3.98. 2-Carbethoxy-3-acetoxy-6-ethoxyalylthieno [3,2-b]pyrrole (VIII).-To a solution of 0.506 g. (2 mmoles) of IV in 10 ml. of anhydrous ether and 2 ml. of anhydrous T H F was slowly added at room temperature 0.40 ml. (4.7 mmoles) of oxalyl chloride as described for the preparation of VI. The reaction mixture was allowed to stand at room temperature for 2.5 hr., after which time the glyoxylyl chloride VI started to crystallize aa yellow needles. After the reaction mixture had been stored a t 25' for 20 hr., the bright yellow crystals were filtered, rapidly washed with a small amount of anhydrous ether and dried for 2 hr. in vacuo over phosphorus pentoxide and potassium hydroxide. The acid chloride VI thus obtained was converted into the ethyl ester VI11 by dissolving in 30 ml. of hot absolute ethanol and allowing the solution to cool slowly to room temperature over a period of 3 hr. After the mixture had been stored at 0-5' for 10 hr., the yellow crystals were filtered and dried in vucua. The yield of VI11 was 0.40 g. (57%), m.p. 201-203'. An additional amount of the ester VI11 was obtained upon concentration of the ether-THF mother liquor under reduced pressure and treatment of the crude, well dried acid chloride VI with 6 ml. of hot absolute ethanol. On cooling a t 0-5' for 0 hr., the solution deposited 0.05 g. of light yellow crystals, m.p. 19&201', bringing the total yield of ester VI11 to 0.45 g. (64%). An analytical sample waa prepared by recrystallization from absolute ethanol to give bright yellow scales, m.p. 202-204'. Infrared (KBr pellet): 3180 (N-H), 1780 (acetoxy C-O), 1725 and 1620 (COCO&zHe), 1695 cm.-l (carbethoxy C 4 ) . (20% DMSO): T = -2.95, s (N-H); 1.62, d (5-proton); 5.57, q and 8.62, t, 5.71, q, and 8.70, t (e.p.).**

(30) Melting points are uncorrected. Microanalyses were performed by Mr. J. Nemeth and his associates, University of Illinois. (31) The infrared spectra were obtained from a Perkin-Elmer Model 21B spectrophotometer by Mr. D. Johnson and Miss D. Wood. (32) The abbreviations used in describing NMR data are: E, singlet; d. doublet; t, triplet; q, quartet: (e.p.1, ethyl group protons. is not detectable because of (33) The expected band for -CO-CHI the reaonances of the solvent (DMSO) in this region.


AnaE. Calcd. for ClcHlrNO,S: C, 50.98; H, 4.28; N, 3.97. Found: C, 50.80; H, 4.39; N,4.10. 2-Carbethoxy-3-acetoxy-6-(N,N-dibenzylglyoxylamido)thieno[3,2-b]pyrrole (X).-To a magnetically stirred solution of 0.57 g. (2.9 mmoles) of dibenzylamine in 25 ml. of anhydrous ether was added a t room temperature 0.45 g. (1.3 mmoles) of the crude acid chloride VI in small portions. Reaction took place immediately, indicated by loss of the yellow color of the suspension. Vigorous stirring was continued for 3 hr. and the resulting voluminous, colorless precipitate of X was collected, washed with ether and water, rand dried in a vacuum desiccator. There waa obtained 0.66 g. (96%. based on monohydrate) of X m colorless crystals, m.p. 154-156'. An analytical sample, prepared by recrystallization from ethanol-water and by drying over phosphorus pentoxide at 78' (0.025 mm.) for 16 hr., formed colorless needles, m.p. 157-158'. ( I t was found that a purified sample which was dried in vacua at room temperature for 14 hr. had the composition of the monohydrate.) Infrared (KBr pellet): 3340-3400 (broad, N-H), 1785 (acetoxy C=O), 1713 (carbethoxy C=O), 1646 and 1630 cm.-' (CO-CO-N ) Anal. Calcd. for C ~ , H ~ ~ N Z OC,~ S64.26; : H, 4.79; N , 5.56. Found: C,64.07; H,4.76; N, 5.56. 2-Carbethoxy-3-hydroxy-6-( N,N-dibenzflg1yoxylamido)thieno[3,2-b]pyrrole (XI).-To a solution of 252 mg. (0.5 mmole) of the amide X in 6 ml. of 95% ethanol was added 0.2 ml. (1 mmole) of 5 N sodium hydroxide with swirling. The resulting yellow solution was allowed to stand at room temperature for 5.5 hr.; after this time 5 ml. of water waa added, and the mixture was neutralized with ice-cold 6 N hydrochloric acid, whereupon a voluminous, colorless precipitate appeared. After cooling for 1 hr. in an ice-water bath, the product was collected, washed with a small amount of water, and dried in vacuo overnight. The product, m.p. 171-172' weighed 0.21 g. (917,). An analytical sample of X I was prepared by dissolving the product in a minimum amount of hot 95% ethanol, gradually adding water until a slightly turbid solution waa obtained, and finally cooling overnight in a refrigerator at 5'. The pure amide X I crystallized aa colorless, felt-like needles, m.p. 172-173'. A solution of X I in 957, ethanol formed an intense dark blue-green color with aqueous ferric chloride solution.*4 Infrared (KBr pellet): 1645 (carbethoxy C=O), 1620-1635 cm.-l (broad, CO(10% CHCla): 3170 (N-H), 1655 (sh, carCO-N=); bethoxy C = O ) , 1625-1640 cm.-' (broad, CO-CO-N=). Anal. Calcd. for C~SHZN~OSS: C, 64.92; H, 4.80; pu', 6.05. Found: C, 64.89; H, 4.90; N, 6.08. 2-Carbethoxy-3-tosyloxy-6-ethoxylthieno[ 3,241pyrrole(VU).-To a solution of 0.20 g. (0.55 mmole) of 2-carbethoxy-3-tosyloxythieno[3,2-b]-pyrrole (111)' in 25 ml. of anhydrous ether was added a t room temperature 0.1 ml. (1.2 mmoles) of oxalyl chloride. The yellow solution waa allowed to stand in a stoppered flask (exclusion of moisture) a t 25' for 6 days and wm then evaporated under vacuum to a yellow foam-like residue, which was dissolved in 3 ml. of 95% ethanol-water (1:1). After standing at room temperature for 1 hr., the mixture was refrigerated for 2 hr., and the resulting colorless solid was filtered and washed with cold 95% ethanol. After drying i n vacua overnight, the product, m.p. 152-154', weighed 0.21 g. (83%). An analytical specimen, recrystallized from absolute ethanol, formed long colorless needles and had the same melting point. Infrared (KBr pellet): 3290 (N-H), 1752 and 1658 (CO-CO&ZHS), 1675 (carbethoxy C =O), 1390 and 1172 (-O-SOa-), 1260 cm.-1 ( G O stretch). Anal. Calcd. for CzoHloNO&3: C, 51.60; H, 4.12; N, 3.01. Found: C, 51.50; H, 4.07; N, 3.09. 5


(34) R. L. Shriner, R. C. Fuson, and D. Y. Curtin, "The Systematia Identification of Organic Compounds," John Wiley and Sons, Inc.. New York, N . Y., 4th ed., 1956, p. 112.



VOL. 27

2-Carb ethoxy-3-tosyloxythieno [3,2-b]p ynolyl-&glyoxylic The Structure d Xanthinin Acid (XU).-To a suspension of 0.139 g. (0.3 mmole) of the glyoxylic ester VI1 in 5 ml. of 95% ethanol waa added at T. A. GEISSMAN room temperature 0.12 ml. (0.6 mmole) of 5 N sodium hydroxide solution, whereupon B clear, yellow solution wm formed. After standing a t 25' for 1.5 hr., a slightly yellow, Department of Chemistry, University of California, Los Angeles, amorphous precipitate began to separate; the reaction mixCalifornia ture was allowed to stand at room temperature 4 hr., after which time 3 ml. of water waa added. The resulting clear, Received February 19, 1966 orange solution was then neutralized with ice-cold 6 N hydrochloric acid and concentrated under reduced pressure to The structures of xanthinin (I) and xanthatin (11), about 2 ml. The crude acid XII, precipitated aa an oily product, solidified upon cooling and scratching to form light sesquiterpene lactones derived from Xanthium tan crystals. The product waa atered, washed with water, pennsylvanicum (cocklebur), were first proposed by and dried in vacuo over phosphorus pentoxide, affording 0.10 g. (75%) of the crude acid XII. Purification of XI1 waa Deuel and Geissman in 1957.' I n the same year, achieved by dissolving the crude product in 2.5 ml. of hot Dolej S and his collaborators2described their studies 95y0 ethanol and adding 8 ml. of water until the solution on these compounds and agreed that the structure remained faintly turbid. Upon slow cooling over a period of for I1 was correct, but proposed the structure I11 11 hr., the pure acid XI1 crystallized in slightly tan, cotton- for xanthinin. Structures I11 and I differ only in like crystals, m.p. 215-217 dec.; the material contained 0.5 mole of water of crystallization, which was not removed by the position assigned to one of the double bonds. drying overnight a t 78' (0.025mm.). Infrared (KBr pellet) Tn the work of Deuel and Geissman the assignment 3190 (N-H), 1720 and 1645, sh (CO-CO,H), 1663 (carb- was made on the basis of the characteristic infrared 1268 cni.-l ethoxy C d ) , 1390 and 1174 (-O-S02-), I (c-O stretch). absorption associated with the -CH=CgroupAnal. Calcd. for C ~ ~ H ~ & O B S ~ . I / Z HC,~ O48.42; : H, ing, while the Czech workers arrived a t the alterna3.61; N,3.14. Found: C,48.58; H,3.56; N,3.14. 2-Carbethoxy-3-hydroxythieno[ 3,2-bjpyrrolyl-6-glyoxylic tive conclusion from the results of an oxidative Acid (E). (A) From 2-Carbethoxy-3-acetoxy-6-ethoxyalyl- degradation. thieno[3,2-b]pyrrole (VIII).-To a suspension of 0.27 g. (0.76 mmole) of the ester VI11 in 10 ml. of 95% ethanol and OAc , OAc 2 ml. of water wag added at room temperature 0.612 ml. (3.06 m o l e s ) of 5 N sodium hydroxide, and the resulting orange-colored solution wag heated on a steam bath under reflux for 10 min. The solution wag allowed to stand a t room temperature for 13 hr., after which time the precipitated orange solid (Na salt) wm dissolved by addition of 11 ml. of water to the mixture. After neutralization of the excess base with ice-cold 6 N hydrochloric acid the clear solution The structure of xanthinin has now been rewaa poured into 10 ml. of ice water and the colorless, microcrystalline precipitate was collected, washed with water, examined with the aid of NMR. The results are and dried in vacuo. The yield of crude IX wag 0.22 g. clear and unequivocal, and substantiate the struc(96%). An analytical sample, obtained by very slow cooling ture (I) originally put forward by Deuel and Geissof a solution of IX in a minimum amount of ethanol-water (1 :3), formed fluffy, colorless needles, vvhich turned bright man. The relevant portions of the KMR spectrum yellow upon drying in vacuo, m.p. (on a preheated hot stage) are shown in Fig. 1. The presence of the secondary 244-246" dec., darkening a t 239". Two additional recrystallizations from ethanol-water did not affect the melt ing point or infrared spectrum. The product obtained as a monohydrate after drying for 16 hr. over phosphorus pentoxide in vacuo (0.05 mm.) at room temperature. A solution of IX in 9573 ethanol gave a positive ferric chloride test84 (intense dark green color). Infrared (KBr pellet): 3380 (broad, 0-H), 3260 (N-H), 1715 and 1624 (COC02H), 1658 (carbethoxy C=O), 1265 and 1237 cm.-1 (C0 stretch). Anal. Calcd. for CIIHBNO&~.H~O: C, 43.85; H, 3.86; N, 4.65. Found: C, 43.72; H, 3.84; N, 4.62. (B) From 2-Carbethoxy-3-acetoxythieno[3,2-b]pyrrolylI I II I I 6-glyoxylyl Chloride (VI).-To a suspension of 0.343 g. ( 1 4 5 8 9 mmole) of glyoxylyl chloride VI in 20 ml. of water was added T value 1 ml. (5 mmolea) of 5 N sodium hydroxide a t room temperature. The resulting clear orange solution waa heated on a Fig. 1 .--KMR spectrum of xanthinin (60 nic. CDClr), shown for 7 = 4 to 5, 8 to 9 regions. steam bath for 15 min., then allowed to stand at room temperature for 23 hr. and finally neutralized with ice-cold 6 N hydrochloric acid. The mixture was chilled for 30 min. and methyl group is shown by the symmetrical doublet the pale yellow precipitate was collected, washed with 8 small at r = 8.9. S o methyl group singlet of the type amount of water, and dried in vacuo; wt. 0.28 g. (93%). as would be required for structure 111, Purification of the crude product by the method given under CHa-C=C, procedure A afforded yellow needles, m.p. 243-245' dec. (1) P. G. Deuel and T. 5 . Geissman. J . A m . Chem. Soc.. 1 9 , 3778 A mixture melting point determination and the infrared (1957). spectrum (KBr) confirmed the identity of the product with (2) L. IIulelS, Y. Herout, a n d F. h n n , ChenL. Lixlu. 61, 1321 (1957); CoZZ. Czech. Chem. Commun., 23, 504 (1958). the acid IX prepared according to procedure A.

JULY,1962 is present. In addition, the vinyl proton region (near 7 = 4) shows both the pair of doublets associated with the exocyclic methylene group of the lactone and additional vinyl peaks for =CH--, probably affected by the adjacent -CH2-grouping and perhaps by the side chain, in the same region of the spectrum, Were structure I11 the correct one, there would be no additional vinyl hydrogen signals besides those of the =CH2 group. The explanation for the results of the degradation experiments described by the Sorm group is obscure, but there can be no doubt that xanthinin has indeed the structure T, despite the Czech findings.


2693 TABLE I




CcH 19-Methyl l&Methyl 1- or 7-Thioacetyl methyl I- or 7-Hydrogen

of Aoetylthio G r o u m la 18



4.30 8.75 9.08 7.65 5.92

4.23 8.79 9.10 7.68 6.43 6.54

4.25 8.60 9.10 7.68 5.92

4.17 8.70 9.13 7.67 5.74

Jackman’s values for alicyclic acetate^.^ The acetylthio-methyls show a shift of about 0.250.30 p.p.m. to lower T values, while the peaks of the protons on the same carbon atom as the sulfur atom are shifted 1.0-1.2 p.p.m. higher than the corresponding peaks in the oxygen analogs. Configurations of 1- and 7-Acetylthio-4Since the p-isomer had been found at the 7-posiandrostene-3,17-diones tion, we decided to check the addition product from the 1,4-diene as well. A careful reexamination of ROBERTC. TWEIT the preparation of la-acetylthi0-4-androstene-3,17dione also revealed the presence of the p- isomer in Chemical Research Division, G. D. Searle & Co., Chicago 80, 9% yield. In the case of the 1-isomers, the NMR Illinois spectra did not provide conclusive evidence for the assigned configurations. The lp-hydrogen gives a Received February 93, 1964 good singlet, but the la-hydrogen produces a peak I n a previous paper’ the preparation of a number intermediate in shape between a singlet and a multiplet, and the positions of the two peaks are reversed of la- and 7a-acetylthio-A4-3-oxo steroids was recompared to those of the 7-isomers. ported. At that time the configurations at 1 and 7 We then turned to rotations to provide more conwere assigned on the basis of analogy in the absence clusive evidence. The molecular rotatory contriof more definitive evidence. Since NMR spectra have now been used to assign configurations in a bution of the lp-acetylthio group is strongly negasimilar series of steroidal esters,2 we decided to tive as is the value for lp-acetoxy 4-androsteneexamine the curves of some of the sulfur compounds. 3,17-dione -358°,5 while the contributions of the were 1a-acetoxy and acetylthio groupings are positive. The substituted 4-andro~tene-3~17-diones The rotatory dispersion curves6 also confirm the chosen as readily available models whose NMR configurational assignments. The curves for the spectra should not be too complicated. I n the previous paper,’ the 7a-acetylthio com- la-compounds are very similar to each other and pound was described as a glass. This result was to 4-androstene-3,17-diene' (see Table 11), but apparently due to the presence in the starting 4,6- the amplitudes of the peaks of the ID-isomers are dien-3-one of an appreciable amount of I ,4-diene-3- considerably lower than those of the la-isomers. one. The use of pure material gave, by chromatogThe relative yield of the isomers may vary deraphy and fractional crystallization, both the 7a- pending on the conditions of the reaction. In the and 7p- isomers as crystalline solids. The molecular addition of ethanethiolic acid to 17a-oxa-~-homorotatory changes for the two compounds are con- 1,4-androstadiene-3,17dione, total of 97y0 of the sistent with the values reported for the correspond- la-isomer was obtained. This high yield may be ing oxygen esters.2 due to the long heating period or to the fact that A definitive proof of the structures is proved by the product had crystallized from the hot mixture. the NMR spectra.3 The spectra of the thioesters Another compound, previously reported as a (see Table I) are qualitatively quite similar to those glass,’ has now been obtained in crystalline form. of their oxygen analogs, whose structures have been This is 1a-acetylthiotestosterone, whose constants discussed.2 The peaks for the 7a- and 7p-acetoxy- are given in the experimental section. methyls and the 7p-hydrogen are quite close to

(1) R. M. Dodson and R. C. Tweit, J . A m . Chem. Sac., 81, 1224 (1959). (2) R. C. Tweit, A. H. Goldkamp, and R . >I.Dodson, J . Oro. Chem., 28, 2856 (1961). (3) Very kindly determined and interpreted by Dr. N. L. McNiven of the Worcester Foundation for Experimental Biology. They were run a t 60 mc. in deuteroohloroform. The shifts are reported as r values relative t o tetramethylsilane as a n internal standard [see G. V. D. Tiers, J . P h y s . Chem., 62, 1151 il958)l.

(4) L. M . Jackman, “Applications of Nuclear Magtietic Resonance Spectroscopy in Organic Chemistry,” Pergamon Press, New York, 1959, pp. 55-7. (5) R. M. Dodson, S. Kraychy, R. T. Nicholson, and 6. Mizuba, J . Org. Chem., in press. (6) Very kindly determined by Professor W. Klyne of Westfield College (University of London) and Professor C. Djerassi of Stanford University. (7) C. Djerassi. E.W. Foltz, and A . E. Lippman, J . A m . Chem. Soc.,

77, 4H31 (1935).



VOL. 27



No substituent

l o r OAc



385 (1560) 367.5(1170) 320 ( 10,450) Dioxane

400 (1240) 369 (480) 317 ( 13,300) Dioxane




381 (1530) 359 (1250) 310 (14,000) Methanol

Experimental8 7a- and 7p-Acetylthi~-androstene-3,17-dione.4,8Androstadiene-3,17-dione, 3.2 g., was mixed with 4 ml. of ethanethiolic acid and warmed on the steam bath for 0.5 hr. Then part of the excess acid was removed under vacuum and the residue was dissolved in benzene and chromatographed on 400 g. of silica gel. From the early fractions eluted with 10% ethyl acetate-benzene there waa obtained 0.9 g. of 7aacetylthio-4-androstene-3,17-dioneby crystallization from lo,AMD-455O, et&;& m.p. 158.5-160.5', j a ] ~f31 A,, 237.5 mp, e 19,500. A d . Calcd. for C*lHB08S: C, 69.96; H, 7.83. Found: C, 69.94; H, 7.83. From another run a yield of 43% was obtained by direct crystallization. From the later 10% fractions two types of crystals formed, prisms (the 7a-isomer) and needles. The needles were separated by hand and recrystallized several times from ether to yield 0.20 g. of 7B-acetvlthio-4-androstene3,17dione, m.p. 129-132' dec., [a]; +166', AMD +31', 237 mp, 6 17,400. Anal. Found: C , 70,35; H, 7.86. The remaining material was apparently a mixture of the two isomers, but further recovery of pure material waa not attempted.



lg OAo

lg - SA0




315 5750) Methanol


317.5i3580) Methanol

Synthesis of (+)-9-Methyl- trans-1,4,9,10tetrahydronaphthalene HERMAN ZIFFERAND ULRICHWmss Laboratory of Physical Biology,

National Institute of Arthritis and Metabolic Diseases, National Institutes of Health,

U.S. Department of Health, Education, and Welfare, Bethesda 1.6,Maryland

Received February 26,1962

As part of a research program on the optical rotatory dispersion of nonplanar conjugated dienes,' a substance was required on which theoretical predictions of the sign and intensity of the Cotton effect could be tested. The substance had to satisfy the Az*" following requirements: Its absolute configuration had to be known, and it had to contain a nonplanar, conformationally fixed cisoid dienic system; furthermore, its structure had to be sufficiently simple for theoretical calculations to be manageable, and la- and lp-Acetylthio-4-androstene-3,17-dione.-l,4Androstadiene-3,17-dione,10.0 g., was mixed with 10 ml. of the compound had to be available in high optical ethanethiolic acid and heated on the steam bath for 15 purity. min. During this time, the solid dissolved and a new solid These conditions are satisfied by (+)-9-methylformed. Fractional crystallization of the material from trans-1,4,9,1O-tetrahydronaphthalene (I). Dreidmethylene chlorideether gave 10.6 g. (83%) of la-acetylthio-4-androstene-3,17-dione,same m.p. and infrared spec- ing models show that the conformationally rigid trum as previously reported material,' and 1.15 g. (97J of diene system of I deviates from planarity by an lp-acetylthi0-4-androstene-3,17-dione, m.p . 177.5179 angle of about 17.5', the skewness2corresponding to dec., [ a ]1 ~f lo,AMD -564', 240.5 mp, e 14,900. that of a right-handed helix. The structural eleAnal. Found: C , 69.76; H, 7.94. ment I1 present in compound I is quite unusual; la-Acetylthiotestosterone.-17p-Hydroxy-l,4-androsta- we are aware of only one other compound embodydien-3-one, 2.2 g., was heated with 2 ml. of ethanethiolic acid for 20 min. on the steam bath. Ether waa added and ing this feature, viz., the l,&cholestadiene of Tamm after 2 days crystals formed. They were recrystallized three and Albrecht.s In particular, the Ass-arrangement times from acetoneether-petroleum ether to yield 1.65 g. of does not seem to be represented among the numerous la-acetylthiotestosterone, m.p. 138.5-139' dec., e::""241 natural sesqui-, di-, and triterpenes. This fact mp, e 15,400, [ a ] D +128.5", AMD f126'. suggested that structure I1 might be markedly unAnal. Calcd. for CJ&O0S: C, 69.57; H, 8.34. Found: stable, and that compounds containing it might be C,69.90; H,8.40. Another 0.25 g. of material, m.p. 136.5-138' dec. waa ob- quite difficult to prepare. The desired substance O

tained from the mother liquors. Chromatography of the residue gave 0.15 g. of impure solid well aa noncrystalline material which may have included the p isomer.

(8) Rotations taken iu chloroform at 24 i 2 O .

(1) A. Moscowits, E. Charney, U.Weias. and H.Ziffer, J . Am. Chsm. Soo., 88,4661 (1961). (2) Theory (1) predict8 that the skewness (4.0.. s e ~ of e helicity) goverm the sign of the Cotton effect, as ita oontribution to the rotation far outweighs that of aaymmetrio substituents; a right-handed hrlicity produces a positive Cotton effect. (3) Ch. Tarnm and R. Albrecht. Hels. Chim. Acta, 42, 2177 (1959).

JULY, 1962


was obtained, however, by a pyrolysis reaction without undue difficulty. Compound I was prepared from the optically pure d-camphor-10-sulfonate of (+)-trans-1-hydroxy-4-keto-2-methoxy - loa -methyl- Azls- hexahydronaphthalene (111); a generous sample of this material, which is antipodal to the natural steroids, was obtained from Dr. Q. E. Thompson (Monsanto Chemical Company, St. Louis, Missouri). The conversion of I11 to (+)-trans-2-keto-lO-methylAa*6-hexahydronaphthalene (IV) has been described for the racemate by Woodward et aL14for the resolved material by Thompson et aLs

$J H

H I1





3.0: 5.2: 5.8); and mass spectroscopy (mol. wt. calcd., 146; found, 146). These findings prove that the desired compound I has actually been obtained, and that the possible shift of the dienic system t o the 2,4-position (calcd., A, 26Smp) has not taken place during pyrolysis. Furthermore, cleavage in the mass spectrometer into mass numbers 92 and 54 (toluene and butadiene, respectively) is the expected one for structure 1.' Several samples of I proved homogeneous by gas chromatography (the column, under the conditioiis used, readily separated cis- and trans-decalin). A slight contamination, presumably by an autoxidation product, was noted by mass spectroscopy (mass 160) and by two weak absorption bands (at 218 mp, arid 225 mp in cyclohexane) having variable intensity ratios with respect to the principal absorption band. The effect of acid on I was studied spectroscopically. The expected transformation into the 2,4diene seemed to take place only slowly after the addition of a few drops of alcoholic hydrochloric acid to a solution of I in ethanol, as indicated by broadening of the principal absorption band.

H Experimental


( +)-tr~m-2cu-Hydroxy-1Ocu-methyl-A~~~-hexahydronaphthalene (Va).-A solution of 4.3 g. of IV in 25 ml. of dry

ether was added to a slurry of 1.5 g. of lithium aluminum hydride in 25 ml. of dry ether. The solution was stirred a t room temperature for 3-4 hr., cooled to 0", and decomposed RO' with 10 ml. of a 10% aqueous solution of ammonium chloride. After several extractions of the aqueous layer with ether, the combined organic layers were dried, and the solVa. R = H 9 , I vent was removed in vacuo, yielding- 4.0 g. vb R=-CO - of an oil (Va) that did not crystallize. NQZ 3.5-Dinitrobenzoate (Vb).-The crude oilv Va (3.9 e . ) was dissolved in a small quantity of pyridine and added to a Reduction of IV with lithium aluminum hydride solution of 7.0 g. (excess) of 3,5-dinitrobenzoyl chloride in 55 in ether gave an oily alcohol (Va) which was char- ml. of dry pyridine. The solution was heated for two hours acterized as the crystalline 3,5-dinitrobenzoate on a steam bath, cooled, and poured into 500 ml. of aque(Vb). The hydroxyl in Va is assumed to be a, ous 570 sodium bicarbonate. The mixture was extracted several times with ether, and the combined ether layers were since Tamm and Albrecht3 found that the hydride washed several times with water, dilute acetic acid, and reduction of AI-cholestene-3-one proceeds stereo- finally with dilute sodium bicarbonate. They were next specifically to give A1-cholestene-3p-ol. The de- dried with sodium sulfate and concentrated in vacuo; the sired diene I was prepared from the ester (Vb) by residue of solid Vb was purified by crystallization from 60 ml. pyrolysis at 150-160' in vacuo. The dinitrobenzo- of methanol. Yield, 1.8g.; m.p. 114-115'. An additional amount, of less pure Vb was obtained by further concentraate has the advantage that the 3,5dinitrobenzoic tion of the mother liqiiors. For analysis, the compound was acid formed is not sufficiently volatile under our recrystallized several more times from methanol, without conditions to contaminate the diene. Compound I, change in m.p. Anal. Calcd. for C I ~ H I R N ~C, O ~60.33; : H, 5.06; N, a colorless oil even a t -60°, is volatile and readily autoxidized. For these reasons a satisfactory ele- 7.82. Found: C,60.13; H, 5.23; K,7.94. ( )-9a-Methyl-trans-l,4,9, lo-tetrahydronaphthalene (I). mentary analysis was not obtained; the compound Four hundred milligrams of Vb was heated in a Pyrex tube was characterized as the crystalline adduct (m.p. connected to a high-vacuum line (pressure mm.) at 150-155Ofor approximately 16 hr. The diene (I)formed was 109-111') with tetracyanoethylene. The identity of I was established by the following collected over a molecular sieve in a cooled trap ( -60") between the pyrolysis tube and the high-vacuum pieces of evidence: A, 261 mp, ( E 3450), calcd. by inserted line. Dinitrobenzoic acid and some undecomposed Vb Woodward's rules 263 mp6 (cf. 262 mp for 1,3- collected on the walls of the pyrolysis tube. The dry I was cholestadienea); integrated NMR (calcd. ratio for distilled, a t room temperature in 2racu0, into another remethyl-aliphatic-vinyl protons 3 :5 :6 ; found, ceiver. The NMR spectrum of I was determined in deuterw (4) R. B. Woodward, F. Sondheimer. D. Taub, K. Heusler, and chloroform: 60 kc. Optical rotatory dispersion (in cyclo-

-6 Li AIk1.a


W. H. McLamore, J . Am. Cham. SOC..74.4223


(5) A. J Speriale, J. A. Stephens. and Q. E. Thompson, ibid., 76, 5011 (1954).

(6) R. B. Woodward, ibid., 64, 72 (1942). (7) K. Biemann, Angew. CAem., 74, 102 (1962).





hexane) (see also Fig. 1 of ref. 1): [ c L ] ~ B ~ 268’; [ct]9sa 13,000°; [(Y]ZSS 0”; [a1234 -11,000’; [01]228 -6800’. Tetracyanoethylene Adduct of 1.-Approximately 80 mg. of I was treated at room temperature with a solution of 79.2 mg. of tetracyanoethylene in 1.0 ml. of freshly distilled tetrahydrofuran. The mixture turned red immediately; on standing for several days it became yellow. It was voncentrated under reduced pressure, and the residue was taken up in 8 ml. of hot hexane. On cooling, the adduct separated in long needles, which were recrystallized from hexanebenzene; m.p. 109-111’. Anal. Calcd. for C I , H ~ ~ NC,~ 74.45; : H, 5.11; N, 20.44. Found: C, 74.56; H, 5.11; N, 20.55.


Acknowledgment.-The authors are much indebted to Professor K. Biemann, M.I.T., for permission to quote the results of his mass-spectrographic investigation of I, and to Dr. E. D. Becker and hlr. R. B. Bradley of this Institute for the NMR data.



Specimens of Fomes n p p l a n a l u ~ collected ~~~ in northeastern Maine, were successively extracted with petroleum ether, diethyl ether, and 95% ethanol. A small quantity of crystalline material which separated from the petroleum ether extract yielded 3-0~0-24S-methy1-5cu-cholesta-7,22-diene (Ia) and 3P-hydroxy-24S-methy1-5~-cholesta-7,22diene (Ib) lo following chromatographic separation Ia. R = O b. R = ;-OH $I C.

R = -0COCH3


on activated alumina, The remaining hydrocarbon extract was saponified and small amounts of sterols Ia and I b were again isolated from the neutral comSteroids and Related Natural Products. XII. ponents. Tentative identification of these substances was based on elemental analyses, physical Fomes applanatus’t constants, and their infrared spectra. The structures were confirmed by comparison with authentic GEORGER. PETTIT A N D JOHN C. KNIGHT samples prepared from ergosterol. Both the disaccharide trehalose” and 5a-diDepartment of Chemastry, University of Maine, Orono, Maine hydroergosterol (Ib) were eventually isolated from the 95% ethanol extracts. However, attempts a t Received February 27, 196.2 isolating in a pure form reasonable quantities of The Polyporaceae family (Agaricales order) con- other neutral (or acidic) sterols from the petroleum tains tube-bearing basidiomycetous fungi of a ether, diethyl ether, or ethanol extracts were unfleshy-tough to woody consistency. Members of successful. For example, separation of the petrothis family are wood-inhabiting annuals or peren- leum ether extract based on isolating reactive carnials commonly known as bracket or shelf fungias bonyl compounds with Girard’s reagent T and Interest in locating new sources of lanostane-type sterols with digitonin held no particular advantage. Similar evaluation of other Fomes species is now tetracyclic triterpenes has led us to begin a phyto* ~progress. chemical survey of the Polyporaceae genus F o m e ~ . ~ in Although Fomes species have been recognized (6) Refer to: (a, F . applanatus), J . Zellner, Monatsh. Chem., 86, since 1878,a preliminary chemical examinations of 611 (1915) [Chem. Abstr., 9,3083 (191511; (b, F. fomentarkus), H. R. only six, apparently, have been described.6 One of Arthur, T. G. Halsall and R. D. Smith, J . Chem. Soc., 2603 (1958); (c, F. hartigii), L. Canonica and A. Fiecchi, Uarz. ehim. ital,, 89,818 these, Fomes applanatus, was reported to contain a (1959) [Chem. Abstr., 64, 22711 (196O)l: (d, F. juniperinus), M . pigment, mannitol and “ergosterol-like com- Anchel, A. Hervey, and W. J. Robbins, PTOC.Natl. Acad Sci. (U.S.), 655 (1952) [Chem. Abalr., 47, 4957 (1953)l; (e, F. oficinalis), pounds.”6a Our initial study has been concerned 38, R. M. Gascoigne, J. S. E. Holker, B. J. Ralph, and A. Robertson, with evaluating Fomes applanatus as a source of J . Chem. SOC.,2346 (1951) and J. Valentin and S. Knutter, PhUrm. Zentralhalle, 96, 478 (1957) [Chern. Abstr., 62, 13187 (1958)l; and ( f , lanostane derivatives. (1) Part X I , G. R. Pettitand D. hl. Piatak, J . O r g . Chern., 27, 2127 (1962). (2) This investigation was supported by P H S Research Grants CY-4074(C2) and CY-4074(C3) from the National Cancer Institute, Public Health Service and in part by National Science Foundation Research Grant G-19500. (3) North American Polyporaceae, for example, comprise 8 genera and approximately 235 species: L. 0.Overholts, “The Polyporaceae of bhe United States, Alaska and Canada,” University of Michigan Press, Ann Arbor, Michigan, 1953. (4) Isolation of the polyporenic acids from several Polyporus species (annuals) suggested that the closely related Fomes genera might also produce unusual lanostane derivatives. A summary of the elegant structural studies involving polyporenic acids has been prepared by E. R. H. Jones and T. G. Halsall, “Progress in the Chemistry of Organic Natural Products,” Vol. X I I , L. Zechmeister. ed., SpringerVerlag. Vienna, 1955,p. 44; cf. also, J. J. Beereboom, H. Fazakerley, and T. G. Halsall, J . Chem. Soc., 3437 (1957). ( 5 ) .4 distinguishing characteristic of this genus is the perennial nature of its species. Certain Fames have been known to survive yevrnty years.

8’. pinicola), T. Shibamoto, K . Minami, and T. Tajima, J. Japun Foreel Soc., 86, 56 (1953) [Chem. Abet/., 18, 6993 (1864)l. (7) This cosmopolitanspecies variously known as Boletus or Polyporua

applanatus and &s Qanoderma upplanalum, Polyporua meg5hnU, or EZ/uingia megaloma is most frequently encountered in hardwood regions. Samples measuring from ca. ten to forty cm., a t the widest point along their horizontal plane, were used in this study. (8) We are indebted to Dr. Martin A. Rosinski, Department of Botany and Plant Pathology, University of Maine, for identifying the plant material and for valuable counsel with several botanical problems. (9) The 24-methyl of Be-dihydroergosterol (Ib) has been assigned a BF configuration by K. Tsuda, Y. Kishida, and R. Hayatsu, J . Am. Chem. Soc., 82, 3396 (1960). Employing the valuable system proposed by R. S. Cahn, C. K. Ingold, and V. Prelog, Ezperientia, 12, 81 i1956), this position may now be designated S. Ketone Ia has also been isolated from a specimen of Fames fomentarius (See footnote 6b). (IO) Recently, sterol I b was found to be a constituent of Polyporus pinicola: T . G. Halsall and G. C. Sayer, J . Chem. Soc., 2031 (1959). (11) Isolation of trehalose from Fames pinicola has been reported by T. Shibamoto. K. hlinami. and T. Tajima, J . Japan For& Soc., 84, 390 (19.52).





treatment (8hr.)with maleic anhydride (0.35 g.) in refluxing zylene (20"l.). Following removal of solvent, in vacuo, the General Extraction Procedure .-In a typical experiment, yellow residue was heated (2 hr.) in refluxing methanol (60 specimens of Fomes upplanatus, collected (by Drs. B. Green, ml.) containing potassium hydroxide (2.5 g.). The mixture T. R. Kasturi, and the senior author, June, 1960) approxi- wae diluted with ether and washed with aqueous sodium mately 2-3 miles west of Duck Pond, Hancock County, bicarbonate and water. After removing the dry (magneMaine, were air-dried at room temperature,stripped of adher- sium sulfate) solvent, residual material was recrystallized ing bark, and ground to a coarsepowder. A 2.85-kg.sample from chloroform-methanol; yield 0.6 g. Four recrystallieaof the fungus sporophore was exhaustively extracted with hot tions from chloroform-methanol gave a pure specimen of petroleum ether (b.p. 65-70'). The total extract of yellow 5a-dihydroergosterol melting at 174-176" (lit.,lTm.p. 176'). oil weighed 26 g. A substantial portion (17.5 g.) of this An analytical sample was dried 24 hr. at 130' and 48 hr. a t material could be obtained using a room-temperature proce- loo", an vacuo. dure. Following treatment with petroleum ether the fungus Anal. Calcd. for C28H~0: C, 84.35; H , 11.63; active was dried, and the extraction procedure was repeated succw- H,0.25. Found: C, 83.81; H, 11.53; activeH, 0.36. sively with refluxing diethyl ether and 95% ethanol. The The acetate derivative (IC, prepared using 1:1 acetic ether extract led to a viscous brown oil weighing 5.0 g. and anhydride-pyridine, 2-hr . reflux period) recrystallized from the ethanol extract gave 92 g. of dark gum. chloroform-methanol as plates melting at 181-182' (lit.,18 Other specimens of Fomes applanatus collected by the m.p. 180-182' and 184-186.8'). authors south of Nicatous Lake (November, 1960) and near Anal. Calcd. for C30H4802: C, 81.76; H, 10.98. Found: the Pistol lakes (September, 1961) of Maine (Hancock C, 81.86; H, 10.66. County) gave lesser yields of the respective solvent extracts. The sterol I b and its acetate derivative were identical's Material from these sources was employed whenever larger with authentic specimens prepared6b116 from ergosterol. quantities of compounds described in the sequel were needed. The methanol mother liquor remaining after crystallizaIsolation of 3p-Hydroxy-24S-methyl-5a-cholesta-7,22-dienetion of sterol I b was concentrated to a viscous orange oil (Ib) and 3-0~0-24S-methyl-5~-cholesta-7,22-diene (Ia).(4.15 g.). A petroleum ether solution of this product (9.9 The crystalline material which slowly (several days) sepa- g., from several pooled experiments) was chromatographed rated from the crude petroleum ether extract (26 g.), de- on activated alumina.lg A fraction (0.27 g.) eluted by 2 : l scribed above, was collected and recrystallized from petro- petroleum ether-benzene partially crystallized (0.18 g., m.p. leum ether; yield 0.4 g. of colorless needles, m.p. 123-148'. 173-176') on standing. Repeated recrystallization from An 0.80-g. specimen corresponding to this fraction was chloroform-methanol yielded a pure sample of 3-oxo-24schromatographed in petroleum ether on activated a l ~ m i n a . 1 ~methyl-5a-cholesta-7,22-diene,m.p. 181-183 ' (lit. ,6b m .p, Elution with 1:2 petroleum ether-benzene gave 3-oxo-24s- 183-185'), [ a I 2 l 0.0" ~ (e, 0.81). R D in dioxane ( e , 1.69), methyl-5a-cholesta-7,22-diene(0.12 g.) as platelets melting 22-24'; [ a 1 2 9 6 -70', [a1317 +59" and [a1370+1.8'. a t 181-183". The fraction (Ib, 0.40 g.) eluted with 4:l Anal. Calcd. for C28H140: C, 84.78; H, 11.18. Found: benzenechloroform recrystallized from ethanol as plates, C, 84.57; H, 11.03. m.p. 170-171'. The latter specimen proved to be 5a-diKetone I a was identicaP with an authentic sample (m.p. hydroergosterol (Ib). Both ergosterol derivatives were 185-186') prepared6b from 5a-dihydroergosterol. A pure identicaP with authentic samples.6bn16 specimen of the 2,4-dinitrophenylhydrazone derivative crysFollowing removal of crystalline material from the original tallized from chloroform-ethanol as yellow needles, m.p. . extract it was heated 3 hr . in a refluxing solution composed of 212-215', [a]'*D 0.0" (c, 0.49). ethanol (125 ml.) and 5 N aqueous sodium hydroxide (125 Anal. Calcd. for Ca4Hd804N4: C, 70.80; H, 8.39; N, ml.). The solution was cooled and extracted with ether. 9.71. Found: G70.79; H, 8.40; N,9.80. The aqueous phase was acidified and acidic products were A combination of column and thin layer chromatographic methylated with diazomethane. A cursory examination of methods indicated that further separation of the residual the methyl esters, using gas chromatographic methods, in- neutral product would most probably be unprofitable. dicated that this mixture was composed of at least eleven Isolation of Trehalose.-When the 95% ethanol extract fatty acids. (vide supra) of Fomes applanatus was concentrated to ca. 350 Removal of solvent from the neutral ethereal solution (see ml., a yellow-brown solid (4.5 9.) separated (on standing). above) yielded an orange-red residue (4.59 g.). A portion of After collecting this material, the filtrate slowly deposited this mixture crystallized from methanol as pale yellow plates large colorless prisms (5.2 g.) of trehalose hydrate, m.p. 94(0.44 g.), m.p. 162-165'. A larger quantity (1.0 9.) of the 95'. Four recrystallizations from water-ethanol gave a crystalline material was prepared and acetylated (1:1 acetic pure sample melting a t 95-97' (lit.,20m.p. 97-98'). Thin anhydride-pyridine, 2 hr. a t reflux). The solid product layer chromatography of this substance on Kieselgel G using which separated upon cooling the reaction mixture was 4: 4: 2 isopropyl alcohol-ethyl acetate-water21 supported its collected, washed with ethanol, and recrystallized from the homogeneity. Mixture melting point determination and same solvent to yield plates (0.7 g.) melting a t 173-177'. both thin layer chromatographic and infrared spectral comFour recrystallizations from ethanol raised the melting point parison with a commercial (Nutritional Biochemicals Corp .) to 181-182'. The sterol (Ib, 0.7 g.) was further purified by sample of trehalose hydrate (m.p. 95-97') confirmed the structural assignment. (12) Melting points are uncorrected and were observed employing a Drying a t 100' for 24 hr., in vacuo, gave anhydrous treFisher-Johns melting point apparatus. Infrared and RD measufehalose, [ C Y . ] ~ ~+194' D (water; c, 1.14).22 ments were recorded b y Dr. R. A. Hill of this laboratory. The latter Anal. Calcd. for Cl2H22OI1:C, 42.10; H, 6.48; 0,51.42. were accomplished using 1.0-dm. annealed quartz cells and the optical Found: C, 42.04; H , 6.48; 0,50.92. rotatory dispersion accessory (0 = 79O44') for the Perkin-Elmer Model Experimentall28l a

4000A spectrophotometer. Optioal rotation (chloroform solution) measurements at the sodium D line were provided b y Drs. Weiler and Strauss, Oxford, England. Elemental analyses were performed in the laboratory of Dr. A. Bernhardt, Miilheim, Germany. (13) We wish to acknowledge the contributions of Dr. T . R. Kasturi during a preliminary investigation of Fomea applanatva solvent extracts. (14) Merck aluminum oxide "suitable for chromatography." (15) Established by mixture melting point determination and infrared comparison in potassium bromide. (16) G D. Laubach and K. J. Brunmgs, J . Bm. Chem. Soc., 74, 705 (lQ52).

(17) D. H. R . Barton and J. D . Cox, J . Chdm. Soc., 1354 (1948). Cf.also, ref. 6b imd 16. (18) R. C. Anderson, R. Stevenson, and F. S. Spring, J . Chem. Soc., 2901 (1952) and W. V. Ruyle, E. M. Chamberlin, J. M. Chemerda. G . E. Sita, L. M. Aliminosa, and R. L. Erickson, J. Am. Chem. Soc., 74, 5929 (1952). See also footnote 6b. (19) Aluminum Co. of America's grade F-20. (20) R. V. Lemieux and H. F. Bauer, Can. J . Chem., 32, 340 (1954). (21) E. Stahl and U. Kaltenbach. J . C h r ~ ? n ~ t ~6 ,g 351 . , (1961). (22) C. S. Hudson, J . A m . C h e m . Soc., 33, 1366 (1H16), reported [=ID +197".


2 698

Subjecting a 42-g. sample of the remaining 95% ethanol extract to basic hydrolysis (250 ml. of 575 aqueous sodium hydroxide) over a 4.5-hr. period, in refluxing ethanol (250 ml.), led to 1.9 g. of neutral material and a large quantity of dark polymer. Following acid hydrolysis of another 42-g. portion of ethanol extract, using 240 ml. of 6% hydrochloric acid and the same general procedure noted directly above, a 4.3-g. neutral fraction waa isolated. Thin layer chromatograms of both neutral products on silicagel G in 9: 1 benzene-ethyl acetate indicated that each waa a complex mixture. Partial resolution of the neutral material from acid hydrolysis using column chromatography on activated alumina1' substantiated this observation. However, a crude specimen (0.24 g., m.p. 162-168") of Sa-dihydroergosterol was eventually isolated from the 1.9-g. sample of neutral product by fractional recrystallization (methanol-chloroform aa solvent).

VOL. 27


siderably more drastic conditions are required for the isomerization of alkyl allyl sulfides' than are required for aryl allyl sulfides, and no condensation occurs with benzaldehyde when phenylthioglycolic acid is replaced by alkylthioglycolic acids.6 Failure of R'Li to react appreciably with the aliphatic vinyl sulfides could logically be attributed to prior anion formation as shown in the accompanying equations RCHzS-CH=CH*

+ R'Li + R'H

+ R-CH+CH=CHr Li

The Reaction of Vinyl Sulfides with Uyllithium. 11'


CsH-S-CH==CH-CHzLi or Q - S - C H = C H - C H ~

WILLIAME. PARHAM, MARTINA. KALNINS,~ AND DONALD R. TEE IS SEN*^' School of Chemistry of the University of Minnesota, Minneapolis 14, Minnesota

Received March 1, 1962

We have extended our study of the reactions of vinyl sulfides (I) with alkyllithium reagentss to in-

+ R'Li







clude cases in which R is vinyl and alkyl. The vinyl sulfides were allowed to react with an excess of alkyllithium in ether at 0'; the compounds studied and the results observed are summarized in Table I. It can be seen that the addition of RLi to RSCH =CH2 is quite sensitive to the structure of R, and proceeds to a reasonable extent, under the conditions studied, only when R is aromatic or vinyl. The facile addition of alkyllithium to phenyl vinyl sulfide, and the failure of phenyl vinyl ether to react,5 does indeed suggest a definite sulfur 3-dorbital effect; however, the reaction of the former may be promoted by the deshielding of the sulfur nucleus through the participation of forms such as III.6 There is other evidence to support such an activating effect by an aromatic substituent. Con(1) This work was supported in part by the O 5 c e of Ordnance Research, U.S. Army, Contract No. DA-11-002-ORD-2616 and DAORD-31-124-61-Gl3~ (2) In part from the M.S. thesis of M. A. Kalnins, the University of Minnesota, 1959. (3) In part from the Ph.D. thesis of D. R . Theiasen, the University of Minnesota, 1961. (4) Sinclair Research Fellow, 1959-1961. (5) W. E. Parham and R. F. Motter, J . Am. Chsm. SOC.,81, 2146

+ R'Li + \


However, this does not appear to be the case. When the reaction mixture obtained from n-butyl vinyl sulfide and ethyllithium was treated directly with carbon dioxide, the olefin was recovered (81%) and there was no evidence for the formation of an organic acid other than propionic. Similarly, and more unexpected, the carbonation of the reaction product obtained from phenyl propenyl sulfide and ethyllithium afforded only recovered sulfide (94.5 %) and propionic acid. The reaction of t-dodecyl vinyl sulfide (IV) with n-butyllithium is of interest since cleavage, rather than simple addition resulted. Failure t o isolate n-hexyl mercaptan from the reaction product suggests that the reaction proceeds as shown in (a), possibly through the complex (V), rather than by path b.





CHI I C H s ( C H z ) g C LiS-CH==CHz





4 ClHsLi COHir



(6) The sulfur 3-d-orbital may be made more reapectable by interaction with the antibonding benzene orbitals, interactions involving the 3-p-orbitals of sulfur with the benzene r system, and/or by inductive effecta.


~-CIZH~~S-CCH=CH~ + n-CrH9Li IV



+ n-CsH1$Li + C ~ H I I

JULY, 1962



TABLE I Alkyllithiuln

C4HoLi C2HsLi CIHoLi CIHeLi CzHsLi

Product isolated (yield, %)

CsHfi+CHzCHzC4Ho (51) C~HKSCH~CH~CZHI, (68) Recovered olefin (high ~ 8 0 ) ~ CzHsS(-CHz)fiCHa (low -5) Recovered olefin (-70) CdHgS-CHzCHzCiHg ( < l o ) Recovered olefin (88) C4HoS(CH&CH3 (not isolated)


Recovered olefin (81.3) Addition product ( 9.31 g., b.p. near 114"/1 mm., 72% 1.6000-1.6450. The recovered phenyl P-styryl sulfide amounted to about 54%. The olefin was identified, as described below, as l-phenylhexene-l(39% yield ) Ether was added to the water layer, obtained from the original reaction mixture, and the resulting mixture was made acidic by the addition of hydrochloric acid. Distillation of the dried, water-washed, ether extract gave phenyl mercaptan (4.3 g., 4oY0 yield, b.p. 164", n% 1.5837). Identification of I-Phenylhexene-1 .-The oil, described above, was redistilled (b.p. 74"/2 mm., n*% 1.5243). Anal. Calcd. for C12H16: C, 89.94; H , 10.06. Found: C, 89.22; H , 10.12; S, 0.00.


(11) W. E. Lawaon and E. M. Reid, J . A m . Chem. Soc., 41, 2821 (1925). (12) Obtained from Lithium Corp. of America. (13) B. T. Brooks and J. Humphrey, J . A m . Chrm. SOC.,40, 839 (1918). (14) H. R . Davis, Ph.D. thesis, University of Minnesota, p. 42 (1949).

VOL. 27

Ozonolysis of thia olefin gave benzaldehyde and n-valeraldehyde, identified by comparison of the corresponding 2,4dinitrophenylhydrazones with those prepwed from authentic samples. 1-Phenylhexene-1, prepared from n-butyllithium and Obromostyrene has n% 1.5377.16 A sample of l-phenylhexene-1, prepared (24% over-all yield) from benzaldehyde and 1-bromopentane, through the intermediate Grignard reagent with subsequent dehydration of the derived alcohol with phosphoric acid, boiled a t 97-100"/8 mm. and had n% 1.5528. The infrared and ultraviolet spectra of this sample of 1-phenylhexene-1 were identical to the corresponding spectra derived from the olefin described in the preceding section, and were consistent with the trans substituted ethylene structure. (15) C. S. Marvel, F. D. Hager, and D. D. Coffman, J . A m . Chem. Soc., 49, 2327 (1927).

Reactions of Triisopropyl Phosphite with Diphenyldiazomethane A.


Chemistry Division Research and Development Center, Armstrong &ork Company, Lancaster, Pa. Received March 1, 1964

St,audinger and his co-workers' prepared phosphazines from tert'iary phosphines and substituted diazomethanes. The phosphazines reported were basic, moderately stable, highly reactive comRIP

+ R'&=Ns


pounds; upon heat'ing they decomposed to give nitrogen and other products. An ylide was identified among the decomposition products from fluorenone triphenylphosphazine. c6H4 PhzP=N--N=C


/\Ci aH4


/ CsH4

+ Nz


Recently Bestmann and Buckschewski2 treated triphenylphosphine with a-ketodiazomethanes. The resulting a-oxotriphenylphosphazines in 80% ethanol were hydrolyzed to the corresponding aoxoaldehyde al-hydrazones and converted by methyl iodide to diazoketones and methylt,riphenylphosphonium iodide. Diazomethane and other diazoalkanes also have been treated with phosphorus trihalides at low temperatures.3 Phosphazines, however, were not isolated and, if formed, were unstable; nitrogen was evolved and moisture-sensitive products were obtained. Furthermore, no phosphazine was isoIat8ed when benzene phosphorus t'etrachloride ( l ) ( a ) H.Staudinger and J. Meyer, Helr. C h i n . Acta, I,619 (1911)). (b) H. Staudinger and W. Braunholta, ibid., 4, 897 ( l Q 2 l ) . (c) H. Stsudinger and G . LUsoher, ibid., 6 , 75 (1922). (2) H . Bestmann, H. Buckschewski, and H. Leube, Be?.. 92, 1345 (1959). (3) A. Ya. Yakubovich and V. .4. Ginsburg, Zh. Obahch. Khim., '22, 1534 (1952).

JULY, 1962


reacted with diazomethane at -40" .4 Trialkyl phosphites also react with diazomethane6 to yield distillable liquid trialkylphosphorazines6 in less than 60% yield. (ROa)P

+ CHzNz +(RO),P=N--N=CHz

These compounds, whose thermal stability is indicated by their distillation at low pressure, react vigorously with carbon disulfide to yield trialkyl phosphorothionates-a reaction analogous to that mentioned by Staudinger and Meyerla and examined in more detail by Braunholtz.' We found that triisopropyl phosphite reacts with diphenyldiazomethane to give an essentially quantitative yield of benzophenone triisopropylphosphorazine, a vhite crystalline solid identified by elemental analysis and its molecular weight. The reaction was carried out by adding a petroleum ether solution of diphenyldiazomethane [(CHs)rCHOl3-P PhzCNz +


[ ( CH~)Z-CBO]$-fi-N=CPhz

to an excess of triisopropyl phosphite in petroleum ether at room temperature. The relatively stable phosphorazine can be handled under ordinary conditions. After approximately one week, however, the material turned pink, an odor of t'riisopropyl phosphate developed, and diphenylketazine was isolated and identified by comparison with the authentic compound. These observations suggest either hydrolysis to benzophenone hydrazone and disproportionation of the latter or disproportionaH10

(RO)aP-N--N=CPh2 + (R0)sPO 1/2Ph&=N--N=CPhz


+ 1/2HzNNHz

tion of the phosphorazine. A third possible source 2(RO)aP=N--N=CPH* + (RO)aP=N-N=P( 0R)r

+ PbZC=N-N=CPhz

of diphenylketazine, suggested by Staudinger and Meyerla to arise by reaction of the carbene intermediate with a diazalkane as a dissociation product, is unlikely under mild conditions prevailing in our case. The evidence for dissociation and carbene (4) L. M. Yagupolskii and P. A. Yufa, Zh. Obshch. Khim., 28, 2853 (1958). (5) M. 1. Kabachnik and V. A. Gilyarov, Dokl. Akad. Nauk S.Y. S.R., 106,473 (1956). (6) Kabachnik a n d Gilyarov named their compounds phosphatoazines. Our nomenclature is preferable t o that proposed by the Russians because, aside from being more euphonious, i t is consistent with that tentatively approved by the general nomenclature committee of the Organic Division and of the NRC Subcommittee on Organic Nomenclature (see Chem. & Eng. News, S O , 4515 (1952), and "Handbook for Chemical Society Authors 1960," Spec. Publ. No. 14, The Chemical Society, London). Thus the adduct of diazoalkanes with phosphines are phosphazines (which Staudinger first used and seems t o be generally accepted) : with phosphinites (RoPOR') phosphinalines: with phosphonites RP(OR')P, phophonazines; and with phosphites (ROIsP, phosphorazines. The azines can be related in a simple and unambiguous way t o the parent structures: phosphinic, phosphonic, and phosphoric acids. Alternatively, all these azines can be related to phosphazines-w., dialkylalkoxyphosphazines, alkyldialkoxyphosphazines, and trialkoxyphosphazines. When a n assortment of substituents are present on the phosphorus stom, there is considerable merit in a nomenclature based on phosphine. (7) W. Braunholtz, J . Chem. Soc., 122, 300 (1922).


formation, however, is somewhat questionable. Thermolysis (170-180") of the triisopropylphosphorazine (I) failed to yield nitrogen as reported by Staudinger'" for the analogous compounds from tertiary phosphines ; instead, propylene and diisopropyl N-diphenylmethylenephosphorohydrazidate (111)were obtained quantitatively. The latter was identified by elemental analysis and molecular weight. The decomposition very likely proceeds through a transitory cyclic intermediate (11).

+ -

(i-PrO)sP- N-N=CPh,




CH3CH=CHa f (i-Pro)2 P (0)NH N =C Phz






The belief that the product is a phosphorohydrazidate was further confirmed by hydrolyzing the compound with concentrated hydrochloric acid : benzophenone and hydrazine mere isolated in 90% yield. Diisopropyl N-diphenylmethylenephosphorhydrazidate was obtained (75% yield) directly from an excess phosphite and diphenyldiazomethane by decomposing the phosphorazine at 178 f 2" without prior separation. Small amounts of triisopropyl phosphate and diphenylketazine were isolated also. A low yield of nitrogen gas seems typical of phosphorus azine reactions. Even under favorable conditions Staudinger and Meyerla never found nitrogen evolution from thermal decomposition of phosphazines greater than 90% of that available and in most cases it was about 50%. When various reactants are present nitrogen is not evolved,' but the nitrogenous products were not identified. Experimental Preparation of Benzophenone Triisopropylphosphorazine. -A purple solution of freshly prepared diphenyldiazomethane* (13.0 g., 0.07 mole) in 25 ml. of petroleum ether was added in 15 min. to a stirred colorless solution of triisopropyl phosphite8 (31.2 g., 0.15 mole) in 10 ml. of anhydrous low boiling petroleum ether at 25'. The reaction temperature rose to 41 O over a 25-min. period, but the purple color persisted. After 16 hr. at ambient temperatures the homogeneous reaction mixture became a clear light yellow. The solvent was distilled at 20 mm. (pot temperatures 25-30') and the unchanged phosphite (14.6 g., n% 1.4101) recovered at 0.3 mm. (pot temperature < 50'). When the yellow oily still residue (27.2 9 . ) was cooled, it solidified to a crystalline (8) Diphenyldiazomethane was prepared according to Org. Syntheses, Coll. Vol. 111, John Wiley & Sons, Inc., New York, N. Y., 1955, p. 351. The reaction product, a purple somewhat viscous oil, was used directly. (9) Triisopropyl phosphite (Virginia-Carolina Chem. Corp.) was redistilled and the fraction boiling a t 6243.5" (12 mm), n% 1.4113 was used in these experiments.



mass having a strong phosphite odor. A small portion of the solid, washed with cold petroleum ether, and pressed dry on a porous plate, melted at 62-65'. The product was soluble in petroleum ether, hexane, benzene, ether, carbon tetrachloride, chloroform, acetone, and ethanol. The main portion of reaction product, recrystallized once from petroleum ether, gave 17.4 g. of crystalline solid, m.p. 63-86'; several additional recrystallizations from the same solvent raised the melting point to 66-67". Anal. Calcd. for C~HslN203P: C, 65.65; H, 7.76; N, 6.96; P, 7.70. Found: C, 65.84; H, 7.67; N, 6.79; P, 7.56. Mol. wt. (determined cryoscopically in benzene): Calcd.: 402. Found: 387. Reaction of Diphenyldiazomethane with Triisopropyl Phosphite.-Triisopropyl phosphite (52 g., 0.25 mole) was added rapidly at ambient temperature to diphenyldiazomethane (13.0 g., 0.07 mole). After 45 min., the purple reaction mixture turned a clear yellow, and the reaction temperature rose to 37". At the end of 16 hr. at ambient temperatures no low-boiling condensate wm found in a DryIce trap connected to the system. The reaction mixture was heated slowly; at 150-170' gas was evolved and a low-boiling liquid condensed in the cold trap. After refluxing the reaction mixture for 20 hr. at 178 &3', the cold traD contained 3.0 g. (0.072 mole) of propylene identified by boiling point and infrared spectrum. The main reaction mixture (60.6 g.) was distilled in vacuo to give: (a) 33.6 g. (0.16 mole) of triisopropyl phosphite, b.p. 6767.5" (14.9 mm.), n'% 1.4104; (b) 2.1 g. of a clear colorlass liquid, b.p. 40-68' (2.4 mm.), n% 1.4082 (infraredindicated the presence of diisopropyl hydrogen phosphonatelo' and triisopropyl phosphatelOb; (0) 1.2 g. of triisopropyl phosphate, b.p. 68' (2.4 mm.), n20D 1.4068; and (d) 23 g. of a turbid residual oil. The residual oil solidified on cooling; it was triturated with cold hexane and filtered. The faintly yellow crystalline filter cake (17.4 g., m.p. 91.5-95'), after recrystallizing several times from hexane, melted a t 96.5-98'. Anal. Calcd. for ClsH2bN~03P:C, 63.32; H, 6.99; N, 7.77; P, 8.60. Found: C, 63.14; H, 6.60; N, 7.92; P, 8.52. Mol. wt. by cryoscopy in benzene: Calcd. 360. Found 360. The hexane filtrate was concentrated; the acidic yellow mushy solid (5.3 g.) was recrystallized from hot absolute alcohol to give 2.1 g. of diphenylketazine, m.p. 160-164'. Admixture with the authentic azine" gave no depression in melting point. The alcohol soluble material was not identified. Thermal Decomposition of Benzophenone Triisopropylphosphorazine.-A sample (1.0 g., 0.0025 mole) of the azine was heated a t 181' for 40 min. Propylene (52.8 ml. at STP, 0.0023 mole) was collected in a gas buret and identified by I.R. The light yellow residual oil (0.89 g.) solidified on cooling, m.p. 91.5-95'. The m.p. of mixture with authentic N-diphenylmethylenephosphorohydrazidate diisopropyl was not depressed. Hydrolysis of Diisopropyl N-Diphenylmethylenephosphor0hydrazidate.-The hydrazidate (5.0 g., 0.014 mole) was refluxed 24 hr. with 50 ml. of concentrated hydrochloric acid. Excess aqueous 20% sodium hydroxide was added and the immiscible oil extracted with ether. Distillation of the combined dried ether extracts left 2.3 g. (0.013 mole) of benzophenone, identified by m.p. and mixture m.p. The basic aqueous layer was acidified slightly and excess salicylaldehyde was added. The reaction mixture was shaken and the faint yellow solid present filtered, washed with low boiling petroleum ether, and air-dried to give 3.0 g. of (lO)(a) Reported b.p. 79O (14mm). H. McCombie, B. C. Saunders, and G. J. Stacey, J . Chem. Soc., 380 (1945). (b) Reported b.p. 83.5' (5 mm.), n% 1.40573,A. I. Vogel and D. M. Cowan, ibid., 16 (1943). (11) Diphenylketazine was prepared from benzophenone and diphenylmethyl hydrazone, m.p. 164.5-165.5', reported m.p. 164'. H. Szmant and C. McGinnis, J . A m . Chem. Soc., 7 2 , 2890 (1950).



material, m.p. 224.5-226'. A mixture m.p. with authentic salicylaldazine showed no depression.

Acknowledgment.-We wish to express our tippreciation to Miss Ellen Phillips for carrying out some of the experimental work.

Aconite Alkaloids. T h e Structure of Hetisine A.


The RockefeUer Institute, New York 81, N . Y . Received March 1, 1988

Data presented below confirm certain features of structure Il1,2but conflict with others, thus leading us to withdraw I as the tentative formulation for hetisine.




The presence of a methyl group on a fully substituted carbon atom is confirmed by NMR, the methyl group appearing as a sharp singlet a t 9.02 r in hetisine, 9.02 r in dihydroheti~ine,~.' and 9.06 r in hetisine diacetate.2 Similarly, the presence of the exocyclic methylene group is demonstrated by peaks which occur a t 5.16 r and 5.36 r (vinyl hydrogen) in the spectrum of hetisine and do not occur in the spectrum of dihydrohetisine. However, the spectrum of dihydrohetisine does contain a new peak (apparently a doublet, partially obscured by the original methyl) a t T = 9.00 attributable to a methyl group substituted on a carbon bearing a hydrogen. The NMR spectrum of Jacobs' des-N-methylhetisine4 has singlets at 9.06 r and 8.80 r and no peaks in the range 5.0 T to 5.4 r , which is consistent with the presence of the original methyl plus a new methyl group of the type 11,presumably arising from the exocyclic methylene of hetisine by alkyl The N-methyl group appears as a singlet a t 7.59 r. Absorption a t 4.37 r (s) and 4.84 7 (d., J = 5 c.P.s.) suggests the presence of a new double bond of type 111. The spectrum of hetisine diacetate has peaks a t 4.82 T , 4.98 r , 5.1 r, and 5.24 r corresponding in area to four protons. This indicates the presence, in addition to the exocyclic methylene group, of two protons on carbon atoms bearing acetoxyl groups, a feature not in accord with structure I. (1) A. J. Solo and 9. W. Pelletier, J . Am. Chem. Soc., 81, 4439 (1959). (2) S. W. Pelletior, Tetrahedron, 14, 98 (1961). (3) W.A. Jacobs and L. C. Craig, J . B i d . Chem., 143,605 (1942). (4) W.A. Jacobs and C. F. Huebner, ibid., 110, 189 (1947).


JULY, 1M2 In confirmation of this finding, the spectrum of the keto diacetate,2 obtained by chromium trioxideacetic acid oxidation of hetisine diacetate, has peaks a t 4.80 7,4.95T, 5.02 T, and 5.18 r corresponding in area to four protons while the spectrum of the hydrolysis product2 has peaks a t 5.13 r and 5.32 r corresponding in area to two protons. NOTEADDEDIN PROOF:After this note had been submitted for publication, a new structure,' based on X-ray studies was advanced for hetisine. Experimental The NMR spectra were determined at 60 Mc. on Model HR-60 and A-60 Varian spectrometers. Deuterochloroform was used as solvent and tetramethyhilane served as internal standard.

Acknowledgment.-The authors are indebted to Dr. ,4.Robertson, to the Department of Chemistry, Columbia University and to Varian Associates for aid in obtaining the NMR spectra. This work was supported in part by Grants RG 5807(C2) and RG 5807(C3) from the National Institutes of Health, U. S. Public Health Service. (5) M. Przybylaka, Can. J . Chsm., 40, 566 (1962).


= H) and 17~-hydroxyandrosta-1,4-dien-3-one (1.R' = OH, RW= H) they obtained 1,Cdimethyl19-norcholesta-1,3,5(10)-triene (1V.R' = CBHI,, R" = H) and 1,4-dimethylestra-l,3,5(lO)-trien17p-ol(IV.R' = OH, R" = H), respectively. When androsta-1,4-diene-3,17-dione (LR', R" = 0 ) was treated in the same manner, they were not able to isolate 1,1,17cu-trimethylestm-1,3,5(10)trien-17/3-01 (1V.R' = OH, R" = CH,) as a homogeneous product. In our investigation of the dehydration-rearrangement of the steroid^,^ which partly parallels theirs, (10)we found that 1,4,17a-trimethylestra-l,3,5 trien-17p-01 (IV. R' = OH, R" = CH,) can be obtained readily if the reaction mixture is acidified with dilute hydrochloric acid and the crude reaction product is chromatographed afterward on silica gel. The physical constants of our pure product support the structure formulated for it. Acetylation of the hydroxyl group at C-17 can be accomplished with isopropenyl acetate in the presence of p-toluenesulfonic acid to give i V (R' = OAC,R" = CHS).



1,4,17a-Trimethylestra-1,3,5(lO)-trien-17,9-01 (IV. R' = OH, R' = CHa).-To a solution of 8.00 g. (0.0281 mole) sndrosta-1,Pdiene-3,17-dione, m.p. 136.5-142', in 200 ml. of anhydrous ether, stirred, and heated under reflux, was added a solution of 40 ml. of a 3 M solution of methylmagnesium bromide. The reaction mixture was stirred and heated under reflux for 6 hr. and then stirred a t room temperature for 15 hr. The reaction mixture was decomposed with water and acidified with 1.7 N hydrochloric acid. The ethereal phase wm separated, washed successively with water and a saturated solution of sodium bicarbonate, dried over anhydrous sodium sulfate, and distilled to dryness under reduced pressure to afford a viscous orange oil. The oil was chromatographed on 800 g. of silica gel. Elution of the column with 2% ethyl acetate in benzene gave 3.00 g. (36%) of a solid, which melted at 143-145' after crystallization from ether-hexane. The analytical sample of IV (R' = OH, R" = CH3) was obtained as colorless laths after another crystallization from hexane, m.p. 143145.5'; XEo:H 269.5 mp ( e 287); XKBr 2.88, 2.98, 12.47 p ; [ c Y ] ~ D+127.5' (CHC4). Admixed with the starting material, it melted at 97.5-124'. Anal. Calcd. for CzlHsoO: C, 84.51; H, 10.13. Found: C, 84.35; H , 9.85. 1,4,17a-Trimethylestra-1,3,5(lO)trien-l7g-o1 acetate (IV. R' = OAc, R' = CH3.)-A solution of 0.55 g. of IV (R' = OH, R" = CH,), m.p. 138-142', 10 ml. of isopropenyl acetate, and ca. 0.1 g. of ptoluenesulfonic acid monohydrate wa8 heated under reflux for 1 hr. After m.0.1 g. of sodium acetate wa8 added, the reaction mixture was distilled nearly to dryness under reduced pressure. The residue was diluted with water and cooled in an ice bath. The solid was collected, waahed well with water, and dried. It waa crystallized first from acetone-hexane and then from hexane alone to yield 0.31 g. (47%) of IV (R' = OAc, R" = CHa) as color270 mp (e less dense crystals, m.p. 166.5-167'; 286); XKBr 5.82, 7.90, 7.98, 12.42 p; [ a ] = D +133' (CHCla). Anal. Calcd. for CaHlleOt: C, 81.13; H, 9.47. Found: C, 80.96; H, 9.49.

(1) H. Dannenberg and H.-G. Neumann, Ann., 646, 148 (1961). (2) M. J. Gentles, J. B. Moss. H. L. Hersog, and E. B. Herahberg, J . Am. Chsm. Soc., 80, 3702 (1958); H. Plieninger and G. Keilioh, Ber., 91, 1891 (1958).

(8) L. J. Chinn and R. M. Dodsony J . Ore. Chsm.. 24, 879 (1959); C. G. Bergatrom and R. M. Dodson, Cham. and Ind. (London), 1530 (1961). (4) Melting pointa were determined on a Fiaher-Johns melting block.

LELAND J. CHI" Division of Chemical Research, G.D . SearL and Go., Chicago 80, Illinois Received March 9, 1968

Recently Dannenberg and Neumannl reported the results of their study on the dienol-benzene rearrangement.2 They observed that the addition of methylmagnesium iodide to a compound possessing a A1t4-dienonesystem I gives a tertiary carbinol 11, which dehydrates to a ('semibenzene" 111, that rearranges to a substance whose ring A is in the form of a 1,2,3,4-tetrasubstitutedbenzene IV. From cholesta-1,4-dien-3-one (1.R' = CsH1,, R"






C Ha

- @I