NOTES
638
by its infrared spectrum which was identical with the spectrum of known sulfur dioxide." The sulfone was recrystallized from ether to give white crystals, m.p. 75-78'. Further recrystallizations gave material melting a t 82-84'. The substance crystallized with 0.5 mole of ether which was not removed completely even on pumping a t O.0Smm. pressure with gentle heating. The ether was detected by a proton n.m.r. spectrum of the episulfone. A typical ether spectrum was observed in deuterioacetone in addition to complex multiplets centered a t 425 and 445 c.p.8. (relative to tetramethylsilane at 60 Mc.). The relative areas of the absorption peaks for the aromatic protons vs. the ether protons were 20:5 corresponding to the empirical formula C28HZ~O&O.5C4H100.The areaa under the two absorptions for the methylene protons in the ether were in the ratio of 6:4. (14) "Sadtler Standard Spectra," Midget Edition, The Sadtler Research Laboratories, Philadelphia, Pa., 1962, No. 841.
VOL.30
Anal. Calcd. for C28H2001S~0.5C~H~00: C, 73.59; H , 5.15; S, 6.55; mol. wt. (assume dissociation), 326. Found: C, 73.45; H, 5.26; S, 6.66; mol. wt., 339, 348 (boiling point elevation in acetone). The molecular weight of a sample (m.p. 128-129') recrystallized from water-acetic acid, in which the etherate did not form, was determined from the elevation of the boiling point of ethylene dichloride. Anal. Calcd. for CzsH200&: C, 74.31; H, 4.48; S, 7.09; mol. wt., 453. Found: C, 74.83; H, 4.80; S, 6.79; mol. wt., 428, 426, 453. The ultraviolet spectrum had no defined maxima, X (dioxane) 260 and 235 mp.
Acknowledgment.-We wish to thank Dr. W. A. Sheppard and Dr. E. A. LaLancette for helpful discussion.
Notes Preparation of 1,2,3-Triazolesfrom 7-Azido-l,3,5-~ycloheptatriene.A Displacement from Nitrogen J. J. LOOKER Research Laboratories, Eaatman Kodak Company, Rochester, New York 14660 Received October IS, 1964
Tropylium azide (7-azido-l,3,5-~ycloheptatriene) has been prepared previously,' and its n.m.r. spectrum shows that it has the covalent structure I. The ionic form I1 may exist in ionizing solvents.2 The covalent
I
II
structure is also suggested by its physical state (liquid) ; ionic cycloheptatrienylium compounds are high-melting solids. The observation that no solid separated from a mixture of methyl tropyl ether and hydrazoic acid in ether further supports this assignment since separation of ionic tropylium bromide occurred when hydrogen bromide was passed into a solution of ditropyl ether.3 Phenyl and benzyl azide can be added to a variety of acetylenes and olefins to produce triazoles and triazolines, respectively. 4--6 Tropylium azide, with the covalent structure I , should undergo addition in a (1) C. E. Wulfman, C . J. Yarnell, and D . S. Wulfman, Chem. Ind. (London), 1440 (1960). (2) D. S. Wulfman, L. Durham, and C . E. Wulfman. %bid., 859 (1962). (3) W. von E. Doering and L. H. Knox. J . Am. Chem. Soc., 76, 3203 (1954). (4) A . Michael, F. Luehn, and H. H. Higbee, A m . Chem. J . . 90,377 (1898). (5) F. Moulin, Helu. Chim. Acta, 8 6 , 167 (1952). (6) W. Kirmse and L. Homer, Ann., 614, 1 (1958).
0
w I
similar manner to form adducts with the seven-niembered ring attached to nitrogen. Addition of the azide I to acetylenes did occur when an electron-attracting group was attached to the triple bond. Dimethyl acetylene dicarboxylate, dibenzoylacetylene, and propargyl aldehyde condensed smoothly and gave triazoles 111, IV, and V, respectively. Di-
0 m
Iv
acetylacetylene afforded only tar under the same conditions. Without a strong electron-withdrawing group present on the triple bond (phenylacetylene or diphenylacetylene), addition did not occur. Two attempts were made to prepare triazolines from tropylium azide. l14-Naphthoquinone was recovered unchanged when heated wit,h azide I. Maleic anhydride gave an unresolvable mixture which may have resulted from a competing Diels-Alder addition of the dienophile to the cycloheptatriene ring. The structures of adducts 111, IV, and V are supported by their n.1n.r. spectra. Each spectrum has a triplet, representing one proton, between r 5.2 and 5.4 which is due to the proton a t position 7 of the 1,3,5cycloheptatriene ring. The remaining bands in each spectrum also agree with the proposed struchres. The formyl group in structure V was assigned to the 4-
NOTES
FEBRUARY 1965
k position on the basis of the polarization of the reactants and the mechanism of dipolar addition.? The ability of the cycloheptatrienyl group to leave a carbon, nitrogen, or oxygen atom has been established.8 The adducts 111, IV, and V would be expected to undergo a similar displacement in acid. When compounds I11 and IV were treated with hydrogen chloride in ether, tropylium chloride separated and triazoles VI and VI1 were isolated from the respective solutions.
639
mole) of dibenzoylacetylene was added and the mixture was heated for 2 hr. on a steam bath. Removal of the solvent and recrystallization of the residue from ethanol gave 3.2 g. (58%), m.p. 75-77'. Anal. Calcd. for CnH1,NaOz: C, 75.2; H, 4.7; N, 11.4. Found: C, 74.9; H, 4.5; N, 11.4. 4,5-Dibenzoyl-l,2,3-triazole(VII).-Hydrogen chloride was passed through a solution of 1.0 g. (0.0027mole) of 1-(2,4,6cycloheptatrienyl)-4,5-dibenzoyl-l,2,3-triazole(IV) in 100 ml. of ether for 30 min. and the residue was collected (0.34g., 100% tropylium chloride). The a t r a t e was concentrated, washed with water, and dried t o give 0.75 g. (100%) of the triazole VII, m.p. 164-165'. This compound has been previously reported but with no mention of the preparation or the melting point.1° The triazole VI1 was also obtained in 13% yield by allowing 2.3 g. of dibenzoylacetylene to stand for 2 days a t room temperature with an excess of hydrazoic acid in benzene-ether. The product melted at 164-165' after recrystallization from ethanol and did not depress the melting point of material obtained from IV. Anal. Calcd. for CisHi1NaOz: C, 69.3; H , 4.0; N, 15.2. Found: C,69.0; H, 4.0; N, 15.3. 1- (2,4,6-Cycloheptatrienyl)-4-formyl-1 ,2,3-triazole .-Tropylium azide was prepared from 6.0 g. (0.092 mole) of sodium azide and 14.0 g.. (0.080mole) of tropylium tetrafluoroborate and extracted into methylene chloride. The organic extract was dried, concentrated, and taken up in a solution of 4.3 g. (0.089 mole) of propargyl aldehyde in 50 ml. of carbon tetrachloride. The solutionwas heatedon a steam bath for2hr. and concentrated, and the residue was distilled to give 10.0 g. (67%) of product, b.p. 92-94' (0.07mm.), 72% 1.5750. Anal. Calcd. for CloHBNaO:C, 64.1; H, 4.9; N, 22.4; mol. wt., 187. Found: C, 64.2; H, 4.9; N, 22.2; mol. wt., 187 (mass spectra). Attempted Preparation of 4-Formy1-1,2,3-triazole .-Hydrogen chloride was passed through a solution of 5.0 g. of 1-(2,4,6-cydoheptatrienyl)-4-formyl-l,2,3-triazole(V) in 100 ml. of ether for 30 min. The solid was collected and the filtrate was washed with water, dried, and concentrated. A small amount of unidentifiable oil remained. Attempted Condensation of Diphenylacetylene and Phenylacetylene with Tropylium Azide.-A solution of 6.7 g. (0.050 mole) of tropylium azide and 8 g. (0.045 mole) of diphenylacetylene in benzene was heated at reflux for 3 days. Removal of the solvent and distillation of the residue gave 7.5 g. (94%) of unchanged diphenylacetylene. When toluene ww used as a solvent and the reaction time was 24 hr., 70% of the starting material was recovered unchanged. Phenylacetylene and tropylium azide (0.02mole of each) were heated in benzene a t reflux for 2 hr. Upon distillation, 64% of the azide was recovered. No higher-boiling material remained in the flask.
e)
w
VI
Cleavage probably occurred by protonation of the substituted nitrogen atom, with subsequent loss of the organic cation. Triazole V did not give a cleavage product under these conditions. Experimental
7-Azido-l,3,5-~ycloheptatriene1(I .-A) solution of 7.1 g. (0.040 mole) of tropylium tetrafluoroborate in 60 ml. of water was stirred while a solution of 3.0 g. (0.046 mole) of sodium azide in 10 ml. of water was added dropwise. An oil separated which was extracted into benzene or methylene chloride. The extract was washed with water, dried, concentrated, and distilled immediately, b.p. 36-40' (0.1 mm.). Redistillation gave 3.0 g. (56%), b.p. 70-72' (7mm.), n% 1.5461 [lit.lb.p. 50' (1 mm.), nZ6D1.54661. The infrared spectrum showed the presence of the azide group a t 2120 cm.-l. On standing a t 25', a solid separated and the liquid became black. The azide could be stored for short periods (24hr.) at -10' under nitrogen. When methyl tropyl ethera was allowed t o stand overnight in an ether solution of excess hydrazoic acid, no solid separated. 1-(2,4,6-Cycloheptat~enyl)-4,5-dicarbomethoxy-l,2,3-triazole (III).-A solution of 1.7 g. (0.013 mole) of freshly distilled tropylium azide and 1.8 g. (0.013 mole) of dimethyl acetylenedicarboxylate in 10 ml. of carbon tetrachloride was heated carefully on a steam bath until the initial evolution of heat stopped. After an additional 30 min. of heating a t reflux, the solvent was removed and the residue was recrystalhed from limoin-ether ; yield 2.6 g., 73%, m.p. 63-64'. Anal. Calcd. for ClIH~nN~O4:C.. 56.7: H , 4.8: N. 15.3. Found: C, 56.5; H, 4.7;- Ni 15.2. The product was obtained in 64% yield (7.0 g.) when undistilled tropylium azide, from 7.1 g. (0.040 mole) of tropylium tetrafluoroborate and 3.0 g. (0.046 mole) of sodium azide, waa heated for 2 hr.with 5.6 g. (0.040mole) of dimethyl acetylenedicarboxylate in 60 ml. of carbon tetrachloride. 4,5-Dicarbomethoxy-l,2,3-triazoleg (VI).-A solution of 3.5 g. (0.013mole) of l-(2,4,6-cycloheptatrienyl)-4,5-dicarbomethoxy1,2,3-triazole (111) in 100 ml. of ether was treated with hydrogen chloride for 1 hr. and the tropylium chloride which separated was collected; yield, 1.7 g. (100%). The filtrate was washed with water, dried, concentrated, and recrystallized from benzene to give 0.8 g. (34%), m.p. 129-130' (lit.9m.p. 133'). Anal. Calcd. for CsH,N~O4: C, 39.0; H, 3.8; N, 22.7. Found: C, 39.1; H, 3.9; N, 23.0. l-(2,4,6-Cyclohepta~enyl)-4,5-dibenzoyl-l,2,3-triazole (IV) Tropylium azide was prepared from 2.0 g. (0.031mole) of sodium azide and 5.0 g. (0.028mole) of tropylium tetrafluoroborate and extracted into 50 ml. of benzene. After drying, 3.5 g. (0.015 ,
_
.
I
.-
(7) R. Huisgen, Proc. Chem. Soc., 357 (1961). (8) K. Conrow, J . Am. Chem. SOC.,81, 5461 (1959). (9) s. Yamad8, T. Mizoguchi, and A. Ayata, Yakwaku Zas8hi, 77, 452 (1957); Chem. Absfr., 81, 14697e (1957).
Acknowledgment.-The author wishes to express his thanks to Miss Thelma J. Davis, Mr. David P. Maier, and Dr. J. Kenneth O'Loane for discussions concerning infrared, mass, and n.m.r. spectra, respectively. (IO) D.G. Farnum and P. Yatea, J .
070.
Chem., a7, 2209 (1962).
17-Azasteroids. 1II.l The Synthesis of N-Hydroxy-l7a-aza-D- homosteroids SUMANAS RAKHIT AND MARCEL GUT wmcesler Foundation f o r Experimental Biology, Shrewsbury, Massachusetts Received September 3, 1964
Recent publication^^-^ from this laboratory described the synthesis of various 17-azasteroids. It was of (1) This work was supported, in part, by a National Institutes of Health Grant H-5266. (2) 9. Rakhit and M. Gut, Tetrahedron Letters, 223 (1964). (3) 9. Rakhit and M. Gut, J. Org. Chem., 29, 859 (1964). (4) 8. Rakhit and M. Gut, SfeTOidS, 4, 291 (1964).
640
NOTES
Figure 1.
J Figure 2.
0
L
J
Figure 3.
interest to prepare an azasteroid which had a negative group attached to the nitrogen in order to test the feasibility of such a synthesis. A steroidal heterocyclic hydroxylamine of the type A (Figure 1) has not been reported in the literature, except hydroxamic acids of the type B which were synthesized by the Schering group and reported by Robinson and co-workers.5 Extensive work has been done on the synthesis of the substituted hydroxylamines by thermal decomposition of the corresponding amine oxides.6 Although utih a t i o n of this reaction has been exhaustively investigated for the synthesis of olefins and substituted acyclic hydroxylamines, not much work seems to have been done in the area of heterocyclic amines. Pyrolysis of N-ethyl and N-methyl tetrahydroquinoline oxide are reported7 to yield tetrahydroquinoline together with acetaldehyde and formaldehyde, respectively. Cope8 (5) C. H. Robinson, 0. Gnoj, A. Mitchell, R. Wayne, E. Townley, P. Kabasakalian, E. P. Oliveto, and D. H. R. Barton, J. A m . Chem. Soe., 88,1771 (1961). (6) A. C. Cope, Org. Reactionr, 11, 361 (1960). (7) Y.Y. Dodonov, J. Qen. Chem. USSR, 14, 960 (1944); Chern. Ab&., 89,4612(1945). (8) A. C. Cope and N. A. LeBel, J. A m . Chem. Soc., 82,4656 (1960).
VOL.30 has reported formation of a ring-cleaved unsaturated hydroxylamine and a saturated bicyclic compound upon thermal decomposition of N-methyl-a-picoline oxide. However, Thesinge obtained the desired N-hydroxypyrrolidine by pyrolyzing N-ethylpyrrolidine oxide. Rogers'O described the synthesis of substituted acyclic hydroxylamines as well as some heterocyclic hydroxylamines which involved a reverse Michael-type reaction on oxides of compounds of the type R2NCH2CH2X (X = COOR', CN, or COPh). In this note we intend to describe the synthesis of two N-hydroxy-l7a-aza-D-homosteroids,which were prepared by the thermal decomposition of the oxide of the corresponding N-ethyl derivatives. These oxides in turn were prepared in good yield by treating the appropriate N-ethyl compounds with m-chloroperbenzoic acid. Decomposition of the N-oxides of the type C could, however, yield ring-cleaved unsaturated hydroxylamines of the types D and F along with the expected hydroxylamine E and ethylene. When the amine oxides 3 and 7 were pyrolyzed by refluxing them in xylene under nitrogen, only single products were obtaked in each case which were subsequently proved to be the desired N-hydroxy compounds 4 and 8, respectively. Proof of structure of these compounds as well as of the intermediates, was provided by their n.m.r. spectra. The n.m.r. spectra of the N-ethyl derivatives 2 and 6 showed a signal for three proton triplet centered at 60 C.P.S. for the C-21 methyl group and the C-18 methyl protons appeared a t 52.5 C.P.S. In the pyrolyzed product there was no signal for the C-21 methyl group and inspection of the low-field region of the spectra revealed only one vinylic proton around 345 C.P.S. Alternate structures D or F for the pyrolyzed product would have shown signals for C-21 methyl and also two additional vinylic protons in the low field region of the spectra. The C-18 methyl protons in the hydroxylamines appeared as expected at a low field of 62.5 C.P.S. Moreover, the elemental analysis of the products corresponded for the hydroxylamine structures 4 and 8. Thus, this sequence of reactions provides an entry into a new class of compounds and could possibly be extended to other azasteroids (Figures 2 and 3). Experimentalll
N-Ethy1-17a-aza-D-homoandrost-4-en-3-one (2) .-To a solution of 1.00 g. of 17a-aza-D-homoandrost-4-en-3-one ( 1)12 in 30 d.acetone waa added 0.9 g. of diethyl sulfate and a solution of 1 g. of potassium hydroxide in 3 ml. of water. The mixture was heated under reflux for 6 hr. The solution was then concentrated in VQCW), cooled, and poured into cold water. The semicrystalline precipitate was collected, washed, and dried to give 1.05 g . of 2, m.p. 148-151' dec. A portion was recrystallized from ether to yield pure N-ethyl-17a-aza-D-homoandrost-4-en&one (2): m.p. 159-161'; VE;1660 (3-keto), 1625 (A4) cm.-l; n.m.r. 52.5 (18-CHs), 60.0 (21-CH8), 68.0 (19-CHs), 3455. 237 mp (log E 4.09). ( C 4 H of A4) c.p.5.; (9) J. Thesing and W. Sirrenberg, Chem. Ber., 92,1748(1959). (10) M. A. T. Rogers, J. Chem. Soc., 769 (1955). (11) The microanalyses were performed by Schwarzkopf Microanalytioal Laboratories, Woodside, N. Y. Melting points were taken on a FisherJones hot stage and are corrected. The infrared spectra were recorded from a pressed potassium bromide pellet on a Perkin-Elmer Infracord spectrophotometer. The n.m.r. spectra were recorded on a 60-Mc. Varian V-4300B spectrometer using 20% solutions in deuteriochloroform and tetramethylsilane as an internal standard. The ultraviolet absorption spectrum was taken on a Cary 14 spectrophotometer. Woelm neutral alumina, activity 111, was used for chromatography. (12) B. M. Regan and F. N. Hayes, J. A m . Chem. Soc., 78, 639 (1956).
FEBRUARY 1965
NOTES
Anal. Calcd. for CzlHaaNO: C, 79.94; H, 10.54; N, 4.44. Found: C, 79.62; H, 10.64; N, 4.66. N-Hydroxy-l7~-aza-D-homoandrost-4-en-3-one (4) .-A sohtion of 900 mg. of N-ethy1-17a-aza-D-homoandrost-4en-3-0ne (2) in 40 ml. of methylene chloride w a cooled ~ to 0-5". TOthis, a solution of 700 mg. of 85% m-chloroperbenzoic acid in 30 ml. of methylene chloride was added slowly within a period of 10 min. The mixture was then allowed to stand for about 15 min. a t room temperature, after which it was washed successively with a cold sodium carbonate solution and water and was dried over sodium sulfate. Removal of solvent yielded 950 mg. of oily N-oxide 3, which was used without further purification. The N-oxide 3 was dissolved in 30 ml. of xylene and heated under reflux in a nitrogen atmosphere for 20 min. The solvent was removed under reduced pressure and 850 mg. of an oily product was obtained which showed the presence of a hydroxyl band in the infrared. This oil in benzene solution was put on an alumina column. Elution of the column with benzene containing 50-75y0 ethyl acetate yielded identical fractions, m.p. 155158'. These were combined to yield 600 mg. of N-hydroxy-l7aaza-D-homoandrost-4-en-3-one(4). A portion of this was recrystallized for analysis: m.p. 159-160'; v=: 3200 (N-OH), 1660 (3-keto), 1625 (A4) cm.-l; n.m.r. 62.5 (C-18 CHa), 70.5 ' mp (log o (C-19 CH3), 348.0 (C=CH of A4) c.p.5.; ~ ~ ? a n O237 4.08). Anal. Calcd. for ClsHzsNOa: C, 75.20; H, 9.63; N, 4.62. Found: C, 74.96; H , 9.70; N, 4.45. N-Ethyl-l7a-aza-D-homoestra-1,3,5( lO)-triene-3-methylEther (6).-A solution of 1.8 g. of 3-methoxy-17a-aza-D-homoestra1,3,5(10)-triene12( 5 ) in 40 ml. of acetone waa refluxed with 2 g. of diethyl sulfate and a solution of 2 g. of potassium hydroxide in 6 ml. of water for 4.5 hr. The solution was poured onto ice and the precipitated solids were filtered, washed, and dried to yield 1.6 g. of 6 , m.p. 100-103'. A portion of this waa recrystallized from ether-hexane to give pure N-ethyl-17a-aza-D-homoestra1,3,5(lO)-triene-3-methyl ether (6): m.p. 104-105O; 1600 (C=C aromatic), 1560 (C=C aromatic), 1030 (0-CHa) cm.-l. Anal. Calcd. for C21H31NO: C, 80.46; H, 9.97; N, 4.47. Found: C, 80.65; H, 9.96; N,4.66. N-Hydroxy-17a-aza-D-homoestra-1,3,5(1O)-triene-3-methyl Ether (8).-A solution of 1.1 g. of the N-ethyl derivative 6 in 30 ml. of methylene chloride was cooled to 0-5" and a solution of 770 mg. of 85% m-chloroperbenzoic acid in 10 ml. of methylene chloride was added within a period of 10 min. The mixture waa allowed to stand a t room temperature for another 15 min. then w a ~washed with cold solution of sodium carbonate and then water, and was dried over sodium sulfate. Removal of solvent yielded 1.15 g. of oily N-oxide 7. This waa dissolved in 50 ml. of toluene and heated under reflux in a nitrogen atmosphere for 15 min. Toluene was removed in vacuo, leaving 1.08 g. of a brown oil. The oil in benzene solution was put on an alumina column. Elution of the column with benzene containing 7 5 4 0 % methylene chloride provided identical fractions, m.p. 140-142'. These were combined to yield 700 mg. of N-hydroxy-l7a-aza-D-homoestra1,3,5(10)-triene-3-methylether (8). A portion was recrystallized from ether for analysis: m.p. 143-145'; v E 3200 ~ (N-OH), 1600 (C=C aromatic), 1560 (C=C, aromatic), 1030 (0-CH,) em.-'. Anal. Calcd. for Cl&N02: C, 75.71; H, 9.03; N, 4.65. Found: C, 75.58; H, 9.00; N, 4.31.
in these neopentyl-type systems.2 In the dehydration of spir0[4.5]decan-6-01 (1, R = OH; m = 2; n = 3) over alumina about 50y0 of spiro[4.5]dec-6-ene (2, m = 2; n = 3) has been found along with the octalins resulting from ring expansion. The dehydration of this spiranol with sulfuric acid leads exclusively to octalim2b The zinc chloride dehydrations of spiro[5.5]undecan-l-o1 (1, R = OH; m = 2; n = 4) and spiro[4.5]decan-l-ol (1, R = OH; m = 1; n = 4) yield mainly cyclohexyl-1-cyclopentenevia a ring contraction route2d and 1,2,3,4,5,6,7,8-octahydroazulene via a ring expansion The phosphoric acid dehydration of spiro[4.4]nonan-l-o1 (1, R = OH; m = 1; n = 3) leads to 4,5,6,7-tetrah~droindane.~~ The zinc chloride dehydration of the related system 2,2dimethylcyclohexanol has been reported to yield olefins of rearranged carbon skeleton. The products obtained from the acetolysis of a-tosyloxyspiranes (1, R = tosyl) are also predominantly rearranged 01efins.~
v::
The Decomposition of Spiranone p-Toluenesulfonylhydrazones. A Convenient Synthetic Route to Sphenes A. PAULKRAPCHO AND R. DON" Department of Chemistry, University of Vermont, Burlington, Vemnont Received Julv 8, 15'6.4
The dehydration of a-hydroxyspirans (1) to produce spirenes (2) is djlKcult to perform because of the tendency for carbon skeleton rearrangement to occur
641
R
NNHTos
1
2
3
The basic decomposition of the p-toluenesulfonylhydrazones (tosylhydrazones, 3) seemed to be potentially useful as a simple synthetic route to certain spirenes (2). The relative migratory aptitude of various ring sizes could also be determined and contrasted to the results of the products formed under ionic conditions. The formation of olefins from the decomposition of tosylhydrazones was first investigated by Bamford and Stevens.6 Friedman and Shechter have investigated the relationship between the products and the nature of the solvent in these decompositions (protonic or aprotic).6 The decomposition of alkyl tosylhydrazones in aprotic solvents yields olefins via hydrogen migration and cyclopropanes via intramolecular insertion.6b The observation has been made that hydrogen migration occurs more readily than carbon skeleton rearrangement in the carbenoid decompositions. For example, the decomposition of the tosylhydrazone of pinacolone leads to t-butyl ethylene (52%) and 1,1,2trimethylcyclopropane (47%). The decomposition of the tosylhydrazones of cyclopentanone and cyclohexanone yields cyclopentene and cyclohexene, respectively.' The tosylhydrazones of higher cycloalkanones produce intramolecular insertion products in addition to the cycloalkenes. In the decomposition of the tosylhydrazones of the C , through Clo ketones extensive 1,3-, 1,5- and l,&transannular insertion (1) National Science Foundation Undergraduate Researoh Participant, 1983-1964. (2) (a) P. A. Nsro and J. A. Dixon, J . Org. Chem., 26, 1021 (1961); (b) P. A. Naro and J. A. Dixon, J . A m . Chem. SOC.,81, 1881 (1959); (0) H. Christol, R. Jacquier, and M. Mousseron, Bull. 80c. chim. France, 348 (1957); (d) {bid., 1027 (1957); (e) H. Christol, M. Mousseron, and F. Plenat, ibid., 543 (1959); (f) R. Mayer and W. Tapelmann, Chem. Ber., 91, 1784 (1958). (3) H. Meerwein and J. Schafer, J . prakt. Chem., 104,289 (1922). (4) A. P. Krapcho and M. Benson, J. A m . Chem. SOC.,84, 1038 (1982). (5) W. R. Bamford and T. S. Stevens, J . Chsm. SOC.,4735 (1362). (8) (3 L. Friedman and H. Shechter, J. A m . Chen. SOC.,81,5512 (1959); (b) ibid., Sa, 1002 (1980). (7)L. Friedman and H. Shechter, ibid., 88, 3159 (1961).
642
NOTES
VOL.30
TABLEI TOSYLHYDRAZONE DECOMPOSITIONS Tosylhydrasone
2,2-Dimethylcyclohexanone Spiro [4.5]decan-6-one Spiro [5.5]undecan-l-one Spiro [4.5]decan-l-one Spiro [4.4]nonan-1-one
Products (system 2)
Crude yield, %
v.p.0. analysis, %a
81 75 82 90 80
99 95 99 99
3,3-Dimethylcyclohexene Spiro[4.5]dec-bene ( m = 2, n = 3) Spiro[5.5]undec-l-ene(m = 2,n = 4) Spiro[4.5]dec-l-ene( m = 1,n = 4)
...
On a Ucon Polar column.
products result.7 Powell and Whiting have investigated the decomposition of various decalone sulfonylhydrazones to ascertain the factors controlling the ultimate position of the double bond in the octalins produced.8
Results The routes to the synthesis of the spiranones and the corresponding tosylhydrazones are described in the Experimental section. The tosylhydrazones were decomposed in refluxing diglymeusing commercial sodium methoxide as the b a ~ e .The ~ ~diglyme ~ ~ ~ mixture was refluxed for about 15 min. and the products were isolated from the mixture by steam distillation. The crude products were analyzed by vapor phase chromatography and were tentatively identified by infrared analysis. The olefinic carbon-hydrogen stretching region (3000 cm.-’), the carbon-carbon double bond stretching region (1650 cm.-l), and the olefinic carbonhydrogen out-of-plane deformation frequencies (750 cm.-1) were particularly revealing where only a single reaction product was obtained.IO The structural assignments were established by nuclear magnetic resonance spectroscopy. The details of the infrared and n.m.r. analyses are listed in the Experimental section. The products from the decompositions were quantitatively hydrogenated to the corresponding spiranes (except the products from the decomposition of the tosylhydrazone of spiro [4.4]nonan-l-one which absorbed only 66% of the calculated amount of hydrogen for one double bond). The hydrogenated products were identzed through the comparison of their infrared spectra and vapor phase chromatographic retention times with those of authentic spiranes prepared by the WOE-Kishner reduction of the corresponding spiranones. The results of this study are tabulated in Table I. The decomposition of the tosylhydrazone of 2,2dimethylcyclohexanone is included for comparison. From the data in Table I it can be seen in the common-size rings studied here that excellent yields of spirenes are produced, except in the decomposition of the tosylhydrazone of spiro [4.4]nonan-l-one where a competing ring expansion and an insertion product (8) J. W. Powell and M. C. Whiting, Telrahedmn, 18, 168,278(1961). (9) For other examples: (a) S. J. Cristol and J. K. Harrington, J . Ow. Chem., 88, 1413 (1963); (b) G.L. Closs, L. E. Closs, and W.A. Ball, J . A m . Chem. SOC.,86, 3796 (1963); (c) C. A. Crob and J. Hostynek, Helr. Chin. Acta, 46, 1680 (1963). (10) (a) H. B. Henbast, G. D. Meakins, and G. W. Wood, J. Chem. SOC., 800 (1954); (b) I(.Nakanishi, “Infrared Absorption Spectroscopy,” HoldenDay, Inc., Ban Francisco, Calif., 1962,p. 24.
53 12
... ... ...
4,5,6,7-Tetrahydroindane(4) Spiro [bicycle(2.l.O)pentane-4,1’cyclopentane] (5) Unidentified 5
...
...
Spiro[4.4]non-l-ene(m = 1,n = 3)
32 3
result. The contrast between these results and the products obtained under ionic conditions where skeletal rearrangement predominates should be noted. A selective hydrogen migration predominates in each case studied. In the case of the tosylhydrazone of spiro [4.4]nonan-l-one the transition state free energy for the l13-insertionis close to that for olefin formation
5
4
and competes with hydrogen migration which leads to the introduction of a strained double bond into the five-membered ring.ll The formation of the ring expansion product in this reaction is energetically favorable owing to the release of ring strain in the fivemembered ring on expansion to a six-membered ring.12 However, the absence of any insertion product in the decomposition of the tosylhydrazone of spiro [4.5]decan-1-one is of interest although in this case no ringexpanded product is to be expected (ring strain increase in the six to seven transition). This study is currently being extended to larger and smaller adjacent rings and to larger rings bearing the tosylhydrazone group to assess migratory aptitudes and intramolecular routes. Experimental Elemental analyses were performed by the Galbraith Microanalytical Laboratories, Knoxville, Tenn. Infrared spectra were recorded on a Perkin-Elmer Model 21 spectrometer. The vapor phase chromatographic analyses were performed on the Aerograph A-90-P. All melting points are corrected; all boiling points are uncorrected. All 60-Mc. n.m.r. spectra reported here were determined neat or in carbon tetrachloride solutions with tetramethylsilane as the internal standard and chemical shifts are reported in .r-values. Synthesis of the Spiranones.Spiro[4.4]nonan-l-one was perpared in SO% yield from 2-(w-bromobutyl)cyclopentanone by the procedure described by Mayer, Wenshuh, and Topelmann.l* Spiro[4.5]decan-6-one was prepared in a 75% yield by the prooedure described by Dixon for the pinacol rearrangement of 1,l‘dihydr~xy-l,l’-dicyclopentyl.~~ Spiro [5.5]undecan-l-one was prepared in 40% yield by the reaction of 1,5dibromopentane with cyclohexanone in the presence of potassium &butoxide.” Spiro[4.5]decan-l-one was prepared in 65% yield from 2-(w-bromopentyl)cyclopentanone by the cyclization procedure described (11) J. G. Traynham and M. F. Sehnert, J . A m . Chem. SOC., 78, 4024 (1956). (12) I(. 9. Pitser and W. E. Donath, ibid., 81, 3213 (1959). (13) R. Mayer, G. Wenshuh, and W. Topelmann, Cham. Be?., B l , 1616 (1958).
FEBRUARY 1965
NOTES
by Mayer.'O The 2,2-dimethylcyclohexanonewas prepared in 40% yield by the alkylation procedure described by King.14 Spiro[5.5]undecan-l-one Tosylhydrazone (3, m = 2; n = 4). Typical Procedure.Spiro[5.5]undecan-l-one (10 g., 0.06 mole), p-toluenesulfonylhydrazine (11.2 g., 0.06 mole), and 35 ml. of absolute ethanol were refluxed for 2 hr. Upon brief refrigeration crystals separated which were collected and recrystallized from of the pure tosylhydrazone abolute ethanol t o vield 16 e. . - (80%) . of m.p. 112-113' d&. Anal. Calcd. for Cl&N202S: C, 64.65; H, 7.84. Found: C, 64.44; H, 7.80. Decomposition of the Tosylhydrazones. Spir0[5.5]undec-lene (2, m = 2; n = 4).-Spiro[5.5]undecan-l-one tosylhydrazone (15.0 g., 0.045 mole) was added to 60 ml. of diglyme (distilled from lithium aluminum hydride). Dry reagent grade sodium methoxide (2.5 g., 0.045 mole) was added in one portion. The apparatus was connected t o a mercury bubbler and the solution was heated with magnetic stirring. Upon brief warmhg, a solid white cake developed t o which 20 ml. more of diglyme was added. On further heating the white solid redissolved and a clear red solution was obtained. A vigorous evolution of nitrogen occurred and a white solid separated out. After heating a t reflux for 20 min. the gas evolution ceased and the red color disappeared. The mixture was cooled, 100 ml. of water was added, and the product was steam distilled from the mixture. The top layer in the distillate was separated and washed once with cold water to remove any diglyme. This layer was dried over anhydrous potassium carbonate and weighed 5.5 g. (82% yield). The vapor phase chromatographic analysis using a 22-ft. Ucon Polar column showed only a trace of contamination in this undistilled material (less than 0.5%). The spiro[5.5]-undec-l-ene was distilled at 205-207" (740 mm.): infrared (neat, NaCl plates) 3050 (=C-H stretch, m), 1660 (C=C stretch, vw), and 728 cm.-l (out-of-plane deformation of = C H , vs); n.m.r. (neat) 7 4.50 (singlet, vinyl protons), 8.10 (multiplet, allylic protons), and 8.45 (multiplet, all other protons), relative areas 1:1:7. Anal. Calcd. for CllH18: C, 87.92; H, 12.08. Found: C, 87.92; H, 12.12. Identification of the Spirenes.Spiro [5.5]undec-1-ene was hydrogenated in acetic acid using platinum as a catalyst and absorbed 101 2% of the calculated amount of hydrogen. The hydrogenated product was isolated by diluting the acetic acid with water. The top hydrocarbon layer was separated and washed with a 10% sodium bicarbonate solution. The layer was dried over potassium carbonate and distilled, b.p. 209-210' (740 mm.), lit.l6 b.p. 212' (723 mm.). The infrared spectrum and the retention time of this product were identical to spiro[5.5]undecane which was prepared by the Wolf-Kishner reduction of spiro[5.5]undecan-l-one.l6 Spiro[4.5]dec-6-ene (2, m = 2; n = 3).-The tosylhydrazone was obtained in a 88% yield and melted a t 153-154' dec. Anal. Calcd. for C1TH24N20eS: C, 63.73; H, 7.55. Found: C, 63.51; H, 7.61. The decomposition of the tosylhydrazone yielded 75% of the spirene, b.p. 34' (2 mm.), micro b.p. 181' (740 mm.). Anal. Calcd. for ClOHl6: C, 88.16; H, 11.84. Found: C, 87.93; H, 11.79. The vapor phase chromatographic analysis of the undistilled material on a Ucon Polar column indicated one major component contaminated by trace amounts of two other components (less than 5%) : infrared (neat, NaCl plates) 3040 (m), 1650 (vw),and 730 cm.-l (vs); n.m.r. (neat) 7 4.55 (singlet), 8.15 (multiplet), and 8.50 (multiplet), relative areas 1:1:6. The sample absorbed 98 2% of the calculated amount of hydrogen. The compound obtained upon hydrogenation waa identical in retention time with spiro[4.5]decane (Ucon Polar column). The infrared spectrum of the reduced product was also identical with that of spiro[4.5]decane. Spiro[4.5]dec-l-ene (2, m = 1; n = 4).-Spiro[4.5]decan-lone tosylhydrazone was obtained in 85% yield and had a melting point of 152-153" dec. Anal. Calcd. for Cl1HMN2O2S: C, 63.73; H, 7.55. Found: C, 63.64; H, 7.50. The decomposition of the tosylhydrazone yielded 90% of the spirene: micro b.p. 177' (740 mm.); infrared (neat, NaClplates) 3080 (m), 1620 (vw),and 740 em.-' (vs); n.m.r. (CCL solution)
4.45 (singlet), 7.75 (multiplet), 8.35 (multiplet), and 8.60 (broad singlet), relative areas 1:1:1:5. Anal. Calcd. for Cl&: C, 88.16; H, 11.84. Found: C, 88.01; H, 11.75. The analysis on the Uoon Polar column showed one major component with less than 1%contamination. The sphene absorbed 101 =k 2% of the calculated amount of hydrogen. The hydrogenated product was identical in retention time and infrared spectrum with spiro[4.5]decane.16 3,3-Dimethylcyclohexene.-The 3,3-dimethylcyclohexanone tosylhydrazone was obtained in 90% yield and melted at 129130' dec. Anal. Calcd. for CI6HnNzO~S:C, 61.20; H, 7.53. Found: C, 61.00; H, 7.34. The decomposition of the tosylhydrazone yielded 81% of 3 , s dimethvlcvclohexene: b.D. 112-113' (740 mm.), lit.16 b.p. 115116" (f45"mm.); infrared (neat, NaCl plates)'3070 (mj, 1660 (vw), and 730 cm.-l (vs). The V.P.C. analysis on the Ucon Polar column showed one component with a trace of contamination. The sample absorbed 98 f 2% of the calculated amount of hydrogen. The product isolated from the hydrogenation was identical in retention time and infrared spectrum with authentic 1,l-dimethylcyclohexane, b .p. 114-115' (740 mm.), lit." b.p. 117-118'. The latter compound was prepared by the Wolf-Kiahner reduction of the ketone. Product Identification from the Decomposition of the Tosylhydrazone of Spiro[4.4]nonan-l-one (3, m = 1; n = 3).-Spiro[4.4]nonan-l-one tosylhydrazone was prepared in a 90% yield and melted a t 146-147' dec. Anal. Calcd. for ClsH~zN~OzS: C, 62.72; H, 7.24. Found: C, 62.38; H, 7.17. The decomposition of the tosylhydrazone yielded 80% of the crude product. The analysis using the Ucon Polar column showed the presence of four components in the following percentages: 53,32,3, and 12, (arranged in order of increasing retention time). Upon catalytic hydrogenation the mixture absorbed 66% of the calculated amount of hydrogen for one double bond. The vapor phase chromatographic analysis of the hydrogenated product indicated three components in the percentages 56,32, and 12, arranged in the order of increaaing retention time. The component found to comprise 32% of the original mixture showed no change in retention time on hydrogenation and thus is inert to catalytic hydrogenation under the conditions utilized here (platinum and acetic acid a t 23"). The component of shortest retention time (53%) in the hydrogenated product was readily identified as spiro[4,4]nonane through the comparison of the retention time and infrared spectrum to those of an authentic sample of this compound prepared by the Wolf-Kishner reduction of the spiranone. Thus the component present in the original decomposition to the extent of 53% can be identified as spir0[4.4]non-l-ene.l* The component of longest retention time in the original decomposition mixture (12%) was identified as 4,5,6,7-tetrahydroindane,4, by the comparison of the retention time with that of authentic sample of this compound prepared by the lithium-ethylamine reduction of indane according to the procedure of Benkeser.18 The reduction of indane produces this compound aa the major product. The analysis of the infrared and n.m.r. data of the original decomposition products and the corresponding hydrogenated products allows one to assign a structure to the component produced in 32% yield: infrared (neat, original decomposition products) 3080 (m), 3050, 1620 (vw),1000 (m), and 740 cm.-l (vs); infrared (neat, catalytic hydrogenation products) 3050 (m) and 1000 cm.-1 (m). Upon catalytic hydrogenation the bands at 3050 and 1000 cm.-1 remain, clearly indicating the presence of a cyclopropane ring with a CH2 grouping in the original decomposition mixture"; n.m.r. (CCL solution, original decomposition products) 7 4.48 (singlet, vinyl protons), 7.70 (multiplet), 7.8 (multiplet), and 9.15 (multiplet due to cyclopropane CHZ); n.m.r. (CCL solution,
I".
*
(14) F. E. King, T. J. King,and J. P. Toplisa, J . Chem. Soc., 919 (1957). (15) J. A. Dixon and P. A. Naro, J . Org. Chem., 26, 2094 (1960).
643
7
(16) (a) H. C. Brown and G . Zweifel, J . A m . Chem. SOC., 8S, 2550 (1961); (b) N. D. Zelinsky, K. Packendorff, and E. B. Choohlowa, Chem. Ber., 88, 98 (1935). (17) H. L. Goering and F. H. McCarron, J. A m . Chsm. Soo., 78, 2270 (1956). (18) D. J. Cram and B. L. Van Duuren, ibid., 77, 3576 (1955). (19) R. A. Benkeser, R. E. Robinson, D. M. Sauve, and 0. H. Thomas, ibid., 77,3230 (1955). (20) See 9. A. Liebman and B. J. Gudzinowicz [Anal. Chem., 88, 931 (1961)I for an infrared study of oyolopropane systems.
NOTES
644
hydrogenated products) r 8.5 (complex broad multiplet with side peaks) and 9.15 (multiplet). The absorption a t r 9.15 is consistent with a cyclopropyl CH2 grouping in both the original and the hydrogenated reaction products.21 Thus in the original reaction mixture to the extent of 32% can be assigned structure 5 on the bask of the infrared and n.m.r. data. No attempt was made to identify the component originally present in a 3% yield.
Acknowledgment.-The partial financial assistance of National Science Foundation Grant 19490 is gratefully acknowledged. The authors wish to thank Dr. Lillya of the Chemistry Department of the University of Massachusetts for running the n.m.r. spectra. (21) (a) W. R. Moore and H. R. Ward, J . Om. Chem., 26, 2073 (1960); (b) J. P. Chesick [ J . Am. Chem. SOC.,84, 3250 (1962)l reported the c y c b propyl CHa grouping st T 9.5 in bicyclo[2.1.0]pentane. (c) K. B. Wiberg and W. J. Bartley, ibid., 82,6375 (1960).
9-Dicyanomethylene-2,4,7-trinitrofluorene, A New Electron Acceptor TAPANK. MUKHERJEE~ AND LEONARD A. LEVASSEUR
Energeties Branch, A i r Force Cambridge Research Laboratories, Bedford, Massachusetts, and Tyco Laboratories, Inc., Waltham 64, Massachusetts
VOL.30 TABLE I CHARQE-TRANSFER COMPLEXES OF DTF Decompn. range, OC. 258-261 400 212-215 365 330-340 195-198
Donor Anthracene Coronene Fluorene Perylene Pyrene Diaminodurene N,N-Dimethyl-2naphthylamine 215 CIodoaniline 203-204 2-Naphthylamine 210-212 Phenothiazine 235-238
We wish to report the preparation of 9-dicyanomethylene-2,4,7-trinitrofluorene(DTF), an acceptor which has superior complexing properties to its precursor, 2,4,7-trinitrofluorenone (TNF) . I n addition DTF forms stable anion-radical salts of lithium and triethylammonium cations. Although the ammonium acetateacetic acid catalyzed condensation of fluorenone and malononitrile proceeds smoothly in benzene, this procedure, when applied to TNF, led to the recovery of starting material. This failure can be attributed to the donor activity of the solvent, which evidently converts TNF completely to its benzene complex. The alternative procedure of piperidine-catalyzed condensation in methyl alcohol gave an excellent yield of DTF. NOz
A
CN CN
The structure of D T F has been assigned on the basis of elemental analysis and infrared spectrum. Hydrolysis in concentrated sulfuric acid led to 9-biscarboxamidomethylene-2,4,7-trinitrofluorene. Alkaline hydrolysis with sodium hydroxide in dilute tetrahydrofuran failed to regenerate 2,4,7-trinitroiluorenone. Solid charge-transfer complexes of DTF with aromatic hydrocarbons and amines are easily prepared by mixing chloroform or dichloromethane solutions of the components. In comparison with the TNF-hydrocarbon complexes, it was generally observed that the extent of dissociation into components on crystallization is markedly less with the corresponding DTF-hydro(1) To whom inquiries regarding this article should be aent at the Air Force Cambridge Research Laboratories, Bedford, Mass.
62.91 45.37 61.67 59.60
3.39 1.90 2.78 2.51
l ;l
N 12.93 7.00 13.22 11.38 12.38 18.59
-FoundC H 66.69 2.80 72.73 2 . 7 0 65.73 2.93 70.08 2.90 67.69 2.79 58.92 4.20
N 12.79 8.72 13.27 11.11 12.27 18.30
15.72 14.30 16.59 14.92
62.81 45.53 61.59 59.89
15.48 14.45 16.48 14.21
3.33 2.17 2.76 2.93
carbon complexes. In Table I some of the donor-acceptor complexes of D T F are listed. The advantage of D T F as a complexing agent is evident from spectral studies of the charge-transfer transition bands, the results of which are presented in Table 11. TABLE I1 CHARGE-TRANSFER SPECTRA FOR HYDROCRBON-DTF AND HYDROCARE~ON-TNF COMPLEXES AND ENERQIES OF HIQHEST OCCUPIED MO’s OF THE HYDROCARBONS Symbol
Received July 24, 1964
Calcd. for -complexC H 66.54 2.78 72.39 2.58 65.78 2.86 70.23 2 . 7 8 67.95 2.67 59.20 4.00
Donor
DTF5
TNFb
x ic
1 Acenaphthylene 550 415 i: 15 0.637 2 Anthracene 662 541 i 4 0.414 1,2-Benzanthracene 640 522 i 5 3 0.452 4 0.439 1,12-Benzoperylene 675 570 f 5 l12-Benzopyrene 600 510 i 5 5 0.497 0.371 3,4Benzopyrene 726 590 i: 5 6 3,PBeneoteraphene 686 561 5 7 0.405 Chrysene 576 482 f 10 0.520 8 0.539 9 Coronene 650 512 i 5 0.499 1,2,3,4Dibenzanthracene 610 505 i: 5 10 0,398 1,2,3,4Dibenzopyrene 647 548 f 7 11 0.442 12 1,2,4,5-Dibenzopyrene 689 562 f 2 Fluoranthene 500 430 i 15 0.618 13 0.347 745 620 f 5 Perylene 14 Phenanthrene 500 435 i 15 0.605 15 561 470 i: 10 0.501 Picene 16 0.445 650 520 i: 5 17 Pyrene Triphenylene 508 425 i:15 0.684 18 Benzo[ghi]fluoranthene 510 430 f 15 0. 63b 19 0. 505b 3,4Benzofluorene 615 495 i 5 20 0. 52b 2,3-Benzofluorene 570 485 f 5 21 0. 51b 1,2Benzofluorene 588 492 i: 5 22 625 518 i 10 0. 47b Decacyclene 23 Fluorene 523 425 i 10 0. 635b 24 C. A. CoulSee ref. 2. a Spectra takenin CHzClz solution. son and R. Daudel, “Dictionary of Values of Molecular Constants,” 2nd Ed., Mathematical Institute, Oxford, England, and the Centre de Mecanique Ondulatoire Applique, Paris, France, 1959.
*
The second column lists the donors, while the third and fourth columns show charge-transfer band maxima of D T F and TNF complexes, respectively. A few of the DTF-hydrocarbon spectra were measured in KBr disks and the absorption maxima were found to be within the same range with those obtained in solution. Since no sharp peaks could be obtained in the measurement of the spectra of TNF-hydrocarbon solutions, Lepley’s2data on the spectra of solid TNF-hydrocarbon complexes is used for comparison. The fifth column (2) A. R. Lepley, J . Am. Chem. Soc., 84, 3577 (1962).
FEBRUARY 1965 records the Huckel coefficients of the highest occupied molecular orbital (xi) of the donor. The charge-transfer bands of D T F complexes occur at considerably longer wave lengths, and are well separated from those of the components. In comparison, most of the observed charge-transfer regions of the T N F complexes are overlapped by donor absorption. The linear relationship3between the charge-transfer transition energy (E,) and the donor orbital coefficient according to eq. 1 has been found to apply to the DTF-hydrocarbon
complexes (see Figure 1). In this equation, 0 is the carbon-carbon resonance integral ; the constant is related to the energy of the lowest unfilled orbital of the acceptor, E'. The straight line of Figure 1 was obtained by the method of least squares. The slope of Figure 1 gives -3.13 e.v. as the value of the carbon-carbon resonance integral which is in good agreement with previous values. From the intercept of the straight line with the z-axis, the energy level of the lowest unoccupied molecular orbital of D T F is found to be E' = a -0.2lP. If we accept the generalization that for complexes of the same bond type (e.g., s-s,n-s,etc.) a decrease in the transition energy is associated with the increase in the formation constant14then D T F is intermediate in Tacid strength between tetracyan~ethylene~ (E' = a -0.110) and TNF2 (E' = a -0.300). The association constant (K) of the DTF-pyrene complex was determined by the Benesi-Hildebrand procedure.6 The observed value of K (23.0) is higher than that for the TNF-pyrene complex (20.9). However, this evidence must be viewed with caution since the Benesi-Hildebrand procedure is subject to considerable error.' Further, it was shown by Melby and coworkers* that, unless appropriate consideration of the geometry of the s-complexes is made, the K-value as a sole measure of relative s-acid strength may lead to erroneous conclusions. The lithium salt of the anion radical of D T F has been obtained as stable blue-black microcrystals from hot solutions of D T F and lithium iodide in acetonitrile. Similar experiments with sodium, potassium, and copper iodides gave highly colored but analytically impure paramagnetic ion-radical salts. Under similar conditions, T N F does not form any ion radicals. Electron spin resonance spectroscopy showed a single sharp line for the solid DTF- -Li+ anion-radical salt. Thus far, we have been unsuccessful in obtaining the fine structure of the anion radical in solution and, consequently, a detailed structural assignment cannot yet be made. The electrical and photoelectrical properties of the D T F charge-transfer complexes and ion-radical salts will be described elsewhere. (3) M. J. 9. Dewar and A. R. Lepley, J . Am. Cham. SOC.,88, 4560 (1961). (4) R. E. Merri6ld and W. D. Phillips, ibid., SO, 2778 (1958). (5) M. J 8.Dewar and H. Rogers,ibid., 54, 395 (1962). ( 6 ) H. A. Benesi and J. H. Hildebrand, ibid., 71, 2703 (1949). (7) For a discussion. see P. R. Hammond, J . Chem. SOC.,479 (1964). (8) L. R. Melby, R. J. Harder, W. R. Hertler, W. Mahler, R. E. Benons, and W. E. Mochel, J . Am. Chem. Soc., 84, 3374 (1962).
645
NOTES
0.6
-
0.5
.
0.4
-
?%
k
14
16 1/x X IO-' om.-'.
18
20
Figure 1.-Plot of the frequencies, I/& of the charge-transfer bands of D T F complexes against the molecular orbital coefficient, xi, for the highest filled orbital of the hydrocarbon. Experimental@
Materials.-2,4,7-Trinitrofluorenone (TNF) waa recrystallized from a 3:1 nitric acid-water mixture, washed with water, and dried under vacuum over silica gel. Anthracene, dibromoanthracene, pyrene, chryeene, and fluorene were purified by chromatography. Synthetic coronene waa purified by repeated crystallization which waa followed by sublimation and chromatography.10 All the other hydrocarbons were used in the form obtamed from Rutgerswerke-Aktiengesebchaft,Frankfurt-amMain, West Germany. The purity of these materials was checked by melting point, absorption spectra, and thin layer chromatography. In several cases, further purification by chromatography was necessary. Spectra.-Meamrements were made with a Cary-14 spectrophotometer. Solution spectra were measured from 300 to 800 mp with equal volume mixturea of 0.01 M acceptor and 0.02 M donor in Spectrograde dichloromethane. In cases of poor donor solubility, saturated solutions were used. Potassium bromide diem containing weighed amounts of pure comp!exes were prepared by standard techniques.ll The disks were dipped in Nujol to prevent light scatter and moisture absorption. The electron spin resonance spectra were measured with a Varian e.8.r. spectrometer using 100-kc. field modulation. Charge-Transfer Complexes.Solid r-complexes were prepared by the following general procedures. A 1.5 X 10-8 M solution of the aromatic hydrocarbon in chloroform or dichloromethane waa mixed with a 1.6-1.7 X lo-* M solutions of DTF in the same solvent. Most of the reported ?r-complexes precipitated immediately. These were collected, waahed with cold solvent, and dried under vacuum. In some instances, however, insolubility of the hydrocarbon in the nonpolar solvent necessitated mixing hot dilute solutions of the ingredients. On gradual cooling the womplexea separated out aa fine needles or microcrystals. Q-Dicyanomethylene-2,4,7-trinitrofluorene(DTF) .-To a SWpension of 3.15 g. (1 X 10-l mole) of powdered T N F in 250 ml. of boiling methyl alcohol, 2.0 g. (3 X lo-' mole) of malononitrile and 2 drops of piperidine were added. After 10 min. the mixture waa allowed to cool to room temperature. The precipitate was filtered, washed with methyl alcohol, and recrystallized from acetonitrile to furnish 3.4 g. (93%) of large (1-1.5 cm.a), yellow crystals of DTF, m.p. 266-268". (9) All melting points are uncorrected. Microanalysis were performed by the Galbraith Laboratoriea, Inc., Knoxville, Tenn. (10) The preparation and purification of coronene were carried out by Monsanto Research Corporation, Dayton, Ohio, under Contract AFl9(604)-8487. (11) G. M. Wyman, J . Opt. SOC.Am., 46, 965 (1955).
NOTES
646
Anal. Calcd. for Cl&aN,Oa: C, 52.89; H, 1.38; N, 19.25. Found: C, 52.49; H, 1.56; N, 19.33. Ultraviolet absorption spectra in dichloromethane show maxima a t 365 mr (log c 4.38). Infrared spectra in acetonitrile and in a KBr disk showed no absorption for the cyano group. Acid Hydrolysis of DTF.-DTF (300 mg., 0.8 x lo-* mole) waasuspended in 25 ml. of concentratedsulphuric acid and warmed for 20 min.; then the mixture waa poured on cruahed ice. The precipitate. waa washed free from acid and dried to give 330 mg. of white solid, m.p. 355-360' dec. A sample recrystallized from acetonitrile had m.p. 35S360" dec. and analyzed for biscarboxamidomethylene-2,4,7-trinitrofluorene. Anal. Calcd. for ClsH9N,0e: C 48.12; H, 2.27; N, 17.54. Found: C, 47.96; H, 2.28; N, 17.30. Lithium-DTF Anion-Radical Salt.-Lithium iodide (1.8 g., 4 X mole) and DTF (1.5g., 1 X lo-* mole) were each separately dissolved in 25 ml. of hot Spectrograde acetonitrile and the solutions were mixed under a nitrogen atmosphere. On cooling, the anion-radical salt separated aa blue-black microcrystals (1.24g., 7557,), m.p. 290' (explodes). Anal. Calcd. for ClsHsLiNaOs: C, 51.87; H, 1.36; Li, 1.88; N, 19.92. Found: C, 51.57; H, 1.60; Li, 1.62; N, 18.57. Electronic absorption spectra taken in acetonitrile show maxima a t 740 mp (log e 3.21); 562 mp (log c 3.29); 525 mp (log e 324); and 425 mp (log e 3.20). In the infrared, relatively strong absorption a t 2200 cm.+ for the CN group waa observed. The e.8.r. signal intensities of the microcrystals corresponded to about 5 X low spins/mole.l* Attempted Preparation of Et3NHC-(DTF)2- .-In a boiling solution of 0.716 g. (1.9 X mole) of DTF in 40 ml. of dichloromethane waa dissolved a solution of 0.10 g. (1.0 X lo-' mole) of TEA in 10 ml. of the same solvent. The mixture waa cooled; the solid (50 mg.) waa filtered and waahed with cold solvent, m.p. 185-190" dec. The analysis gave values close to 1: 1 composition. Anal. Calcd. for C B H ~ O N ~ O c,~ 56.89; : H, 4.34; N, 18.10. Found: C, 56.67; H, 4.37; N, 16.13. The above preparation was repeated using acetonitrile and THF aa solvent. The ion-radical salt was obtained after evaporation of the solvent to a small volume. The resulting paramagnetic solid did not analyze for either the 1 : 1 or the 1 :2 complex; in every case a low nitrogen value was obtained." (12) The e.6.r. measurements were carried out by Dr. Kedma Bar-Eli.
The absorption of weighed solid samples w w compared with t h a t of nolid diphenylpicryl-hydrazyl. The accuracy of the measurements is estimated to be &25%. (13) NOTEADDWIN PROOF.-We have now prepared a number of salts of D T F anion radical by the metathetical reaction between the onium iodides and lithium DTF. Typical examples of the cations are triethyl methyl ammonium, N-methyl quinolinium, triphenyl methyl phosphonium, and triphenyl methyl arsonium ions. These salts have the composition (cation)* (DTF)?.
a-Myrcene (2-Methyl-6- methylene- 1,7-0c tadiene) B. M. MITZNER, E. T. THEIMER, L. STEINBACH, AND J. WOLT Research and Development Deparlment, International Flavors and Fragrances, Inc., Union Beach, New Jersey
30
VOL.
i
'c=cI?, cl;, and
cy,-c=I
CH
.c-myrcene
I
/c\
-'
ond \
CH
CHl
-cy>-c-
I
-C!Y-
\
. . . . . . . . I . . . . . . . . i . . . . ,. . . . I . .
LI
I
I
70
bo
10
Figure 1.-N.m.r.
40 ??M
. . I
. . . . . . . . . . . . .... . . . .lJ t
J.0
2.0
1.0
spectrum of cr-myrcene.
I
I
I
1
I
7.0
. . . . . . . I.... . . . .I I
6.0
I
.... , . . . . t . . . . , . . . . i . . . . , , , , . t . . . . , . . . . r J
5.0
Figure 2.-N.m.r.
4.0 ?PM
1.0
3.0
1.0
spectrum of Pmyrcene.
since the corresponding abstracts do not mention ac myrcene.
I
I1
The formation of I would not be predicted viu a dehydration through carbonium ion 111, and has not been reported as a product of such dehydration by other workers.**' IA
Received February 6, 1964
a-Myrcene (2-methyl-6-methylenel,7-octadiene)(I) has neither been reported as found in nature nor unequivocally prepared synthetically. The only mention of I in the literature is that of G. V. Pigulevskii, et u Z . , ~ who deduced its formation in the iodine-catalyzed dehydration of linalool (11) based upon Raman spectral data. Other apparent referencesZ are not pertinent, (1) G. V. Pigulevakii and N. t. Prokudina, Dokl. Akad. Nauk SSSR, 67, 283 (1949). (2) F. K. Beilnbin, "Handbuch der Organischen Chemie," Vol. I, 4th
Ed., 3rd Supplement, 1958, p. 1063.
111
On the other hand, cis- and trans-ocimene, Fmyrcene (7-methyl-3-methylene1 ,&ctadiene), and dipentene can be predicted aa products and have, in fact, been so reported when using acid-type catslysts.6.6 (3) B. A. Arbuaov and W. 9. Ambramov, Cham. Bur., 67, 1942 (1934), and older references cited therein. (4) F. Aacoli and V. Creacenzi. Chim. I d . (Milan), 40, 724 (1968). (6) M. Ghd-Schumacher and U. Wicker, Cham. Be.,@a, 974 (1960). (6) International Flavora and Fragrances, Inc., unpublished work.
NOTES
FEBRUARY 1965
s--
~~~
647
_______
Figure 4.-Infrared
We have prepared a-myrcene by the vapor phase pyrolysis of myrcenyl acetate (IV) the mechanism for
Iv which has been reported previously.' This pyrolysis yields a- and #?-myrceneas substantially the only products and in the statistically predicted 3 :1ratio. We have also found a-myrcene in the pyrolysate of #?-pinene. The n.m.r. agrees precisely with the structure of amyrcene, Figure 1. For comparison #?-myrceneiS included, Figure 2. The infrared spectrum of a-myrcene,' Figure 3, has a very intense band a t 890 cm.-l, as expected for a compound with two vinylidene groups, and a relatively strong band a t 1650 cm.-l attributable to the isolated vinylidene group only.s The shoulder a t 1640 cm.-l and the relatively intense absorption a t 1600 cm.-' are assigned to the conjugated dienic system.* The 1650-crn.-l band is absent in the spectrum of #?myrcene, Figure 4, and the 1640- and 1600-cm.-l bands are present. (7) B. M. Mitzner, E. T. Theimer, and 8. Lemberg, Gun. J . Chem., 41, 2097 (1963). (8) L. J. Bellamy, "The Infrared Speotra of Complex Molecules," John Wiley and Sons, Inc., New York, N. Y., 1958,p. 34.
spectrum of 8-myrcene. Experimental Pyrolysis of Myrcenyl Acetate.-Myrcenyla cetateQ(99%) waa pyrolyzed at 400" in a 14 X 1 in. glass tube filled with stainless steel protruded packing at a rate of 1 g./min. using nitrogen as a carrier a t 10 ml./min. flow rate. The pyrolysate from 4540 g. of the ester was washed neutral with water and sodium carbonate. The washed pyrolysate (2872 9.) was dried and distilled through a 7-ft., 1-in.-diameter column containing stainless steel protruded packing a t a 20:l reflux ratio. This fractionation effectively separated the CY- and p-myrcene. A 500-g. fraction, b.p. 44" (10 mm.), was CY-myrcene. The rest of the distillate consieted of a mixture of CY- and p-myrcene and unchanged myrcenyl acetate. The a-myrcene cut was found t o be pure by g.1.c. analysis using a 10-ft., */le-in.-i.d., copper column filled with 10% Carbowax 20M on 60-80-mesh silane-treated Celite a t a column temperature of 90" and an inlet pressure of 25-p.s.i. helium with a flow rate of 60 ml./min. The relative retention times of CY- and p-myrcene on this column (@-pinene= 1.OO) are 1.06 and 1.15, respectively. The physical constants of a-myrcene are b.p. 44" (10 mm.), n% 1.4661, dS26 0.7959, 224.5 mM (e 18,600); of p-myr; cene are 224.5 mp ( E 18,600),8 225 mp ( e 20,130). The maea spectrum confirms the empirical structure with a strong molecular ion a t m/e = 136. Hydrogenation of 0.1 mole of a-myrcene, neat, a t room temperature and 30-p.8.i. HZ using 0.3% Adam PtOz catalyst, yielded a product whose infrared spectrum was identical with that of 2,6ðyloctane. Three moles of hydrogen/mole of terpene were absorbed. @-PinenePyrolysate.-A preparative gas chromatograph containing a 10-ft., I/le-in.-i.d. copper column packed with 20% Carbowax 20M on 60-80-mesh Chromosorb P waa used. p-Pinene pyrolysate (200 ml.) prepared by the procedure of Goldblatt and Palkin" from 99% p-pinene waa injected in 1-ml. por-
cz"""
XzzoH
(9) J. H. Blumenthal, U. S. Patent 3,075,003(Jan. 22, 1963). (10) H.Farnow, D~agocoRept., 7 , 124 (1958). (11) L. A. Goldblatt and 8.Palkin, J . A m . Chem. Soc., 68, 3517 (1941). '
NOTEB
648
tions a t 100’ and 30-p.s.i. helium inlet pressure a t a flow rate of 200 ml./min. Only the front half of the myrcene peak was trapped, and 5 ml. of material was thus collected which contained approximately 60% a-myrcene, and 40% p-myrcene. Retrapping of this material in the same manner yielded a-myrcene which waa 90% pure by analytical g.1.c.
Acknowledgment.--The authors are grateful to Dr. S. Lemberg and Mr. M. Jacobs for interpreting the n.m.r. data.
VOL.30
compound was treated further with phenylmagnesium bromide, the products isolated were triphenylmethanol, tetraphenylethanol, and biphenyl. These products correspond to a decomposition by the mechanism shown in eq. 4. The biphenyl is believed to be a by-product from the Grignard reagent and not a decomposition product of 2-benzyl-1,1,2,2-tetraphenylethanol.
?
-
HO-CPhpCPhzCPh
Attempted Synthesis of Hexaphenylcyclopropane B. S. GORTON
HOCPhs
Research Division, Electrochemicals Department, E. I . du Pont de Nemours and Company, Inc., Wilmington, Delaware
8‘
8
0-CPhzCPh- CPh
1 PhMgBr
2. HzO
0
PhCPh
1c 0-
+ Ph&=CPhI I
I
I. PhMgBr
2. Hz0
HCPhzCPhzOH (4) Received July 27, 1964 Experimental
Korshakl has reported the synthesis of polydiphenylmethylene. The polymer was derived by a polyrecombination reaction in which diphenylmethane was treated with a stoichiometric amount of t-butyl peroxide, and was reported to have molecular weights of from 10,000 to 900,000. The high degree of steric hindrance in the chain and the possibility of hydrogen abstraction from the benzene ring during synthesis suggest that structures other than a linear one may be formed, although infrared spectra did not indicate the presence of any disubstituted benzene rings. To study this question, an attempt was made to prepare the polymer by an alternate route. It has been reported that cyclopropanes can be polymerized either by heat and pressure2 or by titanium coordination catalysts. The results suggested that, if hexaphenylcyclopropane could be prepared, it might undergo ring opening to form polydiphenylmethylene. The synthesis of hexaphenylcyclopropane was attempted by two techniques. One was the addition of diphenylmethylene to tetraphenylethylene; the other, the synthesis of hexaphenyl-lJ3-propyleneglycol. In planning the addition reaction, the series of tri-, tetra-, and pentaphenylcyclopropanes leading up to hexaphenylcyclopropanes were tried as models. The tri- and tetrasubstituted compounds, 1,1,2,2-tetraphenylcyclopropane and 1,l,2-triphenylcyclopropane, were made, but both the penta- and hexasubstituted ones failed. Only hexaphenylpyrazoline was isolated in the latter case. The other planned route to hexaphenylcyclopropane is shown by eq. 1-3. However, our first step to hexaphenylpropylene glycol failed. Instead, 2-benzoyl1,1,2,2-tetraphenylethanolwas isolated. When this
-
~~
1. PhMgBr
PhzC(CO0CnHs)z
2.
HO[Phz]aOH
3310
PBra
HO [CPhz]aOH +Br [CPhn]aBr
(2)
Zn
Br[CPh,]aBr +CaPhs
1,1,2-Triphenylcyclopropane.-A solution of 0.033 mole of diphenyldia~omethane~ in 6 d.of ethyl ether was added to 25 d. of styrene. After the solution waa irradiated for 2 hr. with a G. E. ultraviolet sun lamp (RS275-R-40),the red color of the diphenyldiazomethane was discharged and a yellow color had formed. The solution waa diluted with ea. 500 ml. of petroleum ether. A precipitate of polystyrene waa filtered off and discarded. The petroleum ether filtrate waa evaporated under vacuum to an oil which was diluted with methanol. Some more polystyrene precipitated and waa filtered off. The filtrate waa once again evaporated to an oil and allowed to stand for several days. Crystals formed which were soluble in petroleum ether and moderately soluble in methanol. Recrystallization from methanol gave about 4 g. of white needles, m.p. 51-52’. Ana2. Calcd. for C21Hls: C, 93.37; H, 6.63. Found: C, 93.39; H, 6.75. The infrared spectrum of the compound (KBr pellet) shows peaks a t 2.90 (s), 3.30 (E), 6.23 (E), 6.68 (E), 6.90 (E), 9.27 (m), 9.67 (m), 10.3 (m), 12.8 (a), 13.05 (E), 13.3 (m), 13.55 (m), and 14.2-14.4 (broad, strong) p. The n.m.r. spectrum of the sample in CCL (Varian S-60A, 60 Mc., with tetramethylsilane aa standard) waa more complex than expected. The aromatic frequency waa split unequally to -420 and -429 c.p.8. in ratio of 2: 1, respectively. The Ha line was split twice equally to -175, -168, -166, and -159 c.p.8. as expected. However, Hp1 and Hpz showed seven lines instead of theexpectedeight: -120, -114, -115, -109, -104, -100, and -95 c.p.8. The ratio of the sum of areas of Ha, Hpl, and Hgz to the aromatic hydrogen waa 1:5 aa expected. No propene structure was present. 1,1,2,2-Tetraphenylcyclopropane .-Approximately 0.05 mole of solid diphenyldiazomethane was prepared by evaporation of a petroleum ether solution under vacuum (see above). The solid waa dissolved in 100 g. of l,l-diphenylethylene6 and the mixture waa irradiated under a G.E. ultraviolet sun lamp for 2 days. The white crystals which formed were filtered off and washed with small amounts of petroleum ether to give about 8 g. of a product with m.p. 167-170”. Skelle has reported the reaction of diphenyldiazomethane with 1,l-diphenylethylene, but no details or properties of the product were given. Anal. Calcd. for CzrHx: C, 93.65; H, 6.35. Found: C, 93.65; H, 6.48. The infrared suectrum (in KBr) was 2.95 (a), 3.30 (81, 6.25 (a), 6.70 (E), 6.9b (E), 9.25 (m), 9.65 (m), 9.85 (m), 10.95 (m), 12.52 (E), 13.2 (E), 13.9 (a), and 14.3 ( 8 ) p. The n.m.r. spectrum of the sample (as above) had one peak a t -417 and one at -148 c.p.6.; the ratio of their areas was
(3)
(1) (a) V. V. Korshak, S. L. Sosin, Proc. Acad. Sci. USSR, Cfiem. Seo., Eng. TFansZ., 184, 517 (1960); (b) V. V. Korshak, S. L. Sosin, Vysokomole k d . Soedin., 8, 1332 (1961). (2) A. A. Shohetinin, I. N. Topchieva, ibid., 4,499 (1962). (3) M.Yamada, M. Yanagita, 6. Makoto, Rept. Inst. Phys. Chem. Res. (Tokyo), 87, 429 (1961).
( 4 ) L. I. Smith and I(. L. Howard “Organic Syntheses,” Coll. Vol. 111, John Wiley and Sons, Inc., New York, N. Y., 1955, p. 351. (5) C. F. H. Allen and 8. Converse “Organic Syntheses,” Coll. Vol. I, John Wiley and Sons, Inc., New York, N. Y., 1956,p. 226. (6) R. M. Etter, H.S. Skovronek, and P. 8.Skell, J. Am. Cfiem. SOC.,81, 1008 (1959).
FEBRUARY 1965 10:1, respectively. No peaks were present which could be msigned to a propene struckture. Attempted Preparation of Pentaphenylcyclopropane. A. Ultraviolet Catalyzed.-A saturated solution of triphenylethylene' in petroleum ether wm prepared. The solution formed by adding 0.05 mole of solid diphenyldiazomethane to the solution was irradiated for 4 days using the G.E. sun lamp, by which time the color was discharged. Fractional crystallization of the mixture from a large excess of petroleum ether yielded only crude diphenylketazine, m.p. 141-152' (identified by infrared spectra), unchanged triphenylethylene, and slight amount of unidentified oil. B. Thermal Decomposition.-To a melt of 100 g. of triphenylethylene was slowly added a saturated solution of 0.1 mole of diphenyldiazomethane in petroleum ether. The mixture was heated for 2 days a t 100'. Only unchanged triphenylethylene could be identified in the reaction mixture. 3,3,4,4,5,5-Hexaphenylpyrazoline.-To a solution of 0.2mole of diphenyldiazomethane in 250 ml. of benzene was added 28.5 g . of tetraphenylethylene (Columbia Organic Chemicals, Columbia, s. C.). The mixture was refluxed on a steam bath for 3 days to give an orange-colored solution. The solution on cooling deposited crystalline tetraphenylethylene. The solution was concentrated further and crystals were removed at intervals. Diphenylketazine, identsed by infrared spectrum, came out next, and finally several fractions of crystals melting from 903 30' were obtained. The crude crystals were purified by vacuum sublimation to give about 12 g. of a yellow, waxy material which was recrystallized from ethanol to give crystals, m.p. 123-131". Anal. Calcd. for C39H80Nz:C, 89.0; H, 5.70; N, 5.32. Found: C, 88.44; H , 5.94; N, 5.44. The infrared spectrum of the compound (KBr pellet) shows peaks a t 2.90 (m), 3.30 (m), 6.20 (a), 6.30 (m,, 6.40 (m), 6.72 (a), 6.95(s),7.60(s), 7.82(m),8.50(m),9.30(m),9.72(s),10.45 (m), 10.75 (w, shoulder), 10.95 (m), 12.83 (s), 13.05 (s), 13.42 (s), 13.75 (m), and 14.2-14.4 p (broad, strong). Decomposition of Hexaphenylpyrazoline. A.-A 1-g. sample of hexaphenylpyrazoline was heated in an oil bath a t 240-250" a t atmospheric pressure for 48 hr. in a vessel attached to an ice trap for collecting products and a nitrogen stream to remove volatiles. The batch was cooled slowly for 2 days to yield a black maas of crystals. When the mass was washed with ethanol, the color was removed leaving a cluster of cream-colored fine needles, m.p. 216-222". The infrared spectrum of a sample of the needles confirmed that it was tetraphenylethylene. B .-A second 1-g. sample (as a fine, powdery film)of the hexaphenylpyraeoline was irradiated with the G.E. ultraviolet sun lamp for 4 days. No change was detected in the infrared spectrum of the material. 2-Benzoyl-l,1,2,2-tetraphenylethanol.- Asolution of 0.8 mole of phenylmagnesium bromide (Arapahoe Chemical Company, Boulder, Colo.) in 1000 ml. of absolute ethyl ether was refluxed for 5 days with 60 g. of ethyl diphenylmalonate.8 The solution was poured into an ice-ammonium chloride mixture and the mixture was extracted with ether. The ether layer was dried and evaporated to yield 79 g. of yellow-orange crystals, m.p. 84". The crude crystals were crystallized from ethanol to give light orange crystals, m.p. 118'. Anal. Calcd. for C3sHza0z: C, 87.20; H, 5.72; 0,7.05. Found: C,87.07; H,5.86; 0,6.53. The infrared spectrum showed a carbonyl absorption a t 5.95 p. The complete spectrum (KBr pellet) was 2.90 (e), 3.30 (m), 5.95 (s), 6.28 (m), 6.70 (m), 6.92 (a), 7.55 (m), 7.65 (m), 7.85 (m), 8.29 (m), 8.48 (w), 8.62 (m), 9.68 (m), 9.87 (m), 13.15 (s), 13.50 (m), and 14.2-14.4 (8, broad) p , Reaction of 2-Benzoyl-l,1,2,2-tetraphenylethanolwith Phenylmagnesium Bromide.-A solution of 10 g. of 2-benzoyl-1,1,2,2tetraphenylethanol together with 1.O mole of phenylmagnesium bromide in 250 ml. of ethyl ether was refluxed for 4 days. The solution was then poured into an ice-ammonium chloride mixture and extracted with ether. The ether layer was separated,. dried over anhydrous sodium sulfate, and then concentrated to yield two fractions of crystals: A, m.p. 155-240°, and B, 5562". Fraction A was fractionally sublimed at 140' (0.2 mm.) and 210' (0.5mm.) to yield two seta of crystals. (7) H. Adkins and W. Zartman "Organic Syntheses," Con. Vol. 11, John Wiley and Sons, Ino., New York, N. Y., 1955, p. 606. (8) A. C. Cope and S. M. MoElvain, J. Am. Chem. Soc., 64, 4322 (1932).
NOTES
649
The infrared spectrum of an authentic sample of triphenylmethanol, m.p. 161-163', was identical with that of the lower subliming material, m.p. 155-164'. Anal. Calcd.: C, 87.7; H, 6.16; 0,6.16; mol. wt., 260. Found: C, 87.33; H, 6.39; 0,6.30; mol. wt., 258, 254 (in ethylene chloride). The higher subliming material, m.p. 213-236', analyzed for tetraphenylethanol (likem.p. 232-236). Anal. Calcd.: C, 89.0; H, 6.33; mol. wt., 350.3. Found: C,88.33; H, 6.50; mol. wt., 344,352. "he fraction B waa sublimed at 100' (0.25 mm.) to give crystals, m.p. 68-71', which analyzed for biphenyl (m.p. 69-71'). The infrared spectrum of an authentic sample of biphenyl waa found to be identical with that of the sublimed material. Anal. Calcd.: C, 93.50; H, 6.50. Found: C, 92.52; H, 6.54.
Acknowledgment.-The author wishes to acknowledge the helpful suggestions by Dr. H. K. Sinclair on some of the mechanism interpretations. (9) A. MoKennie and J. 9. W. Boyle, J . Chem. Soc., llQ, 1139 (1921).
Tetraphenylbenzyne DIETMAR SEYFBRTH AND HORST H. A. MENZEL Department of chemistry, Massachusetls Institute of Technology, Cambridge, Massachusetts 02139 Received September 11, 1964
The availability of 1,2-dibromotetraphenylbenzene during the course of a recent study in these laboratories1 has prompted this brief investigation of the possibility of generating tetraphenylbenzyne, previously unknown, via this precursor. The procedure chosen was that used by Wittig and PohmerZato generate benzyne itself from o-bromofluorobenzene in the presence of furan as a benzyne-trapping agent. This reaction, carried out
under nitrogen a t room temperature during 6 days, gave 1,Pdihydr0-5,6,7,8-tetraphenylnaphthalene 1,4endoxide in 66% recrystallized yield. Attempted column chromatographicpurification of the endoxide on neutral alumina led to decomposition to 1,2,3,4-tetraphenylnaphthalene. Thin layer chromatographicanalysis showed that the latter was not present in the original crude reaction product. This brief study has shown that the four phenyl substituents in 1,2-dibromotetraphenylbenzene do not hinder benzyne formation, nor do they appear to make the benzyne intermediate significantly less stable than (1) D. Seyferth, C. Sarafidis, and A. B. Evnin, J . Orpanometal. Chem., a, 417 (1964). (2) (a) G. Wittig and L. Pohmer, Chem. Ber., 89, 1334 (1956). (b) See related experiments with benzyne, l,a-naphthyne, and 9,lChphenanthryne: R. Huisgeu and J. Sauer, Angew Chem., 73,91 (1960).
NOTES
650
benzyne itself. Stabilization through conjugation of the tetraphenylbenzyne is not deemed likely in view of probability that the adjacent phenyl rings are not coplanar with the central ring. A comparison of the relative reactivities of tetraphenylbenzyne and the structurally related 9,lO-phenanthryne toward a diene or of phenyllithium and lithium piperidide toward tetraphenylbenzyne2b should shed more light on this question.
VOL.30 SCHEME I -0 M+ CH3----CHz(CHz)n--CHz-X I 2b
CH3-CO-CH*-(CHz)n-CH2-X la, 12 = 2+x = c1 b,n=3;X=Cl c,
+
4
x-G'J I
n=2;
Experimental
M+
1-
CH3
Preparation of 1,4-Dihydro-5,6,7,8-tetraphenylnaphthalene 1,4-Endoxide.-Lithium amalgam was prepared2 from 45 g. (0.225 g.-atom) of mercury and 0.12 g. (0.018 g.-atom) of lithium in a Schlenk tube under argon. To the amalgam then waa added 3.2 g. (5.3 mmoles) of 1,2-dibromotetraphenylbenaene, 35 ml. of dry diethyl ether, and 35 ml. of dry furan. The reaulting suspension (maintained under an argon atmosphere) was shaken using a mechanical shaker for 6 days at room temperature. During this time the liquid phase became light yellow in color. The Schlenk tube was cooled and opened, and benzene and methylene chloride were added to effect solution of all organic components. The filtered, yellow organic solution was evaporated a t reduced pressure, leaving 2.8 g. of crystalline residue, m.p. 78-86'. The latter was recrystallized from carbon tetrachloride-petroleum ether to give a solid with m.p. 182-184'. Further recrystallization gave 1.9 g. (66%) of the desired product, m.p. 214-217' (decomposition with gas evolution). AnaZ.3 Calcd. for Ca&40: C , 91.04; H, 5.39; 0, 3.57. Found: C,90.44; H, 5.41; 0,3.52. The n.m.r. spectrum4 showed a triplet a t 5.55 and a singlet at 6.75 followed by a multiplet from 6.9-7.2 p The was grated area ratio of the triplet to the singlet p us multiplet 1:11. The triplet is assigned to the bridgehead protons a t the 1,4-positions, and the olefinic protons appear to overlap with the aromatic protons. A similar situation has been observed in the case of benzonorbornadieneP Attempted purification of the product by column chromatography (Woelm alumina, neutral) waa unsuccessful. Green-violet coloration was observed, and elution with 1; 1 benzene-cyclohexane gave a product which did not contain oxygen. Its m.p. 203-205' and its analysis suggest that it is 1,2,3,4-tetraphenylnaphthalene (lit.e m.p. 201-203°,68 19&200°6b). Anal. Calcd. for C34H24: C, 94.41; H, 5.59. Found: C , 94.28; H, 5.90. Analysis of the crude product of the tetraphenylbenzynefuran reaction by thin layer chromatography showed that tetraphenylnaphthalene was not present.
CH3
X=N(CH& I-
The Chemistry of Carbanions. VIII. The Intramolecular Alkylation of Ketyl Radical Anions1 HERBERT 0. HOUSE,JEAN-JACQUES RIEHL,AND COLING. PITY Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 09i39 Received September 18, 1964
Several alkylation reactions have been described2 which presumably involve either the intramolecular or the intermolecular alkylation of an intermediate radical
5
l'.
icH*jn CHa--CO-CHz(CH2)n-CH*-
1. 1
-0
M
CH3-CO-CHz
l'.
\
\
from
solvent
M'
3
'
i
CH~-CC=CH-,(CHZ)~-CH~
9'"'
(3) Analyses were performed by the Galbraith Laboratories, Knoxville, Tenn. (4) Measured in carbon tetrachloride solution using a Varian Associates A80 n,m.r, spectrometer. Chemical shifts are given in parts per million downfield from tetramethylsilane. (5) H. Menzel, Dissertation, Universitat Heidelberg, 1963. ( 6 ) (a) W. Herwig, W. Metlesics, and H. Zeiss, J . Am. Chem. SOC.,81, 8203 (1959); (b) H. Gilman, S. G. Cottis, and W. H. Atwell, ibid., 86, 1598 (1964).
2a
f
s,n=2;
0-
CH3,
/c\ CHZ,
/
M'
,CH2 (CHz))z
(CHz)n-CH3 5
4
anion. These intermediate radical anions have been derived from reduction of either aromatic nuclei or a,&unsaturated ketones with metals. We wished to learn whether a comparable intramolecular alkylation of a simple ketyl anion radical might be effected since such a procedure might offer a useful synthetic route to bridgehead tertiary alcohols. The accompanying equations illustrate the reaction of interest, namely 1 -+ 2 -+ 3 4 (see Scheme I). To examine this possibility the reactions of the ochloro ketones l a and l b with lithium in liquid ammonia, with the sodium-phenanthrene and sodiumnaphthalene radical anions, and with chromium(I1) ion [Cr(II) were examined. Although treatment of the chloro ketone l a with chromium(I1) acetate led to recovery of the starting material, the other two reducing systems did convert the chloro ketones la and l b to mixtures of the desired cyclic carbinols 4 -+
(1) This research has been supported by Grant No. AF-AFOSR-573-84 from the U. 8 . Air Force O 5 c e of Scientific Research. (2) (a) G. Stork and J. Tsuji, J . A m . Chem. SOC.,8S, 2783 (1981); (b) 8. Rakhit and M. Gut, ibid., 86, 1432 (1984); ( 0 ) J. Wiemann and F. Weisback, Compt. rend., 467, 1488 (1983); (d) D. R. Weyenberg and L. H. Torporcer,J . A m . Chem. Soc., 84, 2843 (1982); (e) D. Lipkin, F. R. Galiano, and R. W. Jordan, Chem. Ind. (London), 1867 (1983). (3) The acylation of anion radioala derived from saturated carbonyl functions ie exemplified by the acyloin reaction and by the cyclization of certain &keto esters. See C. D. Guteche and I. Y. C. Tao, J . Org. Chem., 48, 883 (1963): (4) The uae of the chromium(I1) ion was prompted by the observation of an intramolecular alkylation reaction when a suitably substituted a& uneaturated &bromo ketone was treated with chromium(I1) chloride. D. H. R. Barton and C. H. Robinson, U. S, Patent 3,028,337 (March 20, 1982); Chem. Abetr., 67, 3532 (1982).
FEBRUARY 1965
NOTES
and the corresponding saturated, acyclic ketones 5. The secondary alcohols corresponding to ketones 5 were also formed in reductions with lithium in liquid ammonia, presumably by further reduction of the ketones 5 . Unfortunately, the desired cyclization products 4 composed only 7% (lithium-liquid ammonia) to 28-33% (sodium-phenanthrene or naphthalene) of the volatile, neutral products, the major products being the ketones 5. I n an effort to suppress formation of that portion of the acylic product 5 which might be derived by direct cleavage of the carbon-chlorine bond (ie., 1 --t 6 + S), we examined the reductions of the quaternary ammonium salts IC, Id, and le. These salts were examined as displaceable functions because half-wave potential measurements5 of alkyl halides and the corresponding tetraalkylammonium salts suggested that the ammonium salts would be less easily reduced. However, the only products isolated from these reductions were the corresponding amines 8, 9, and 10 resulting from cleavage of a methyl group from the quaternary ammonium salts.*
their purity was established by gas chromatography.9 7-Chloro2-heptanone (lb) was isolated (55-7775 yield) as a colorless liquid: b.p. 101-102' (13 mm.); 12% 1.4459 [lit." b.p. 107108' (17 mm.), n% 1.44351 ; 2,4-dinitrophenylhydrazone m.p. 94.5-95' (lit." m.p. 95-96'); infrared absorptionloat 1720 cm.-' (C=O); ultraviolet maximum11 a t 274 mp (e 23.5); andn.m.r. peaks10 a t S 3.65 (2H triplet, J = 7 c.P.s., -CHTCl), 2.49 (2H, triplet, J = 7 c.P.s., -CHAD-), and 2.15 (3H singlet, C H r CO-). 6-Chloro-2-hexanone (la), isolated (77-79% yield) as a colorless liquid, b.p. 83-84' (10 mm.), n% 1.4424 [lit.@+lB b.p. 52.5-54' (2.3 mm.), 85.5-86.5' (16 mm.), n% 1.44141, has infrared absorptionlo a t 1720 cm.-l (C=O), an ultraviolet maximum a t 264 mp (e 39.3), and n.m.r. peaks10 a t 6 3.52 (2H triplet, J = 6 c.P.s., -CHTCl), 2.43 (2H triplet, J = 7 c.P.s., -CHrCO-), and 2.07 (3H singlet, CHS-CO-). After a solution of 13.43 g. (0.10 mole) of 6-chloro-2-hexanone (la) and 12.3 g. (0.15 mole) of sodium acetate in 60 ml. of acetic acid had been refluxed for 48 hr., the reaction mixture was poured into saturated, aqueous sodium chloride and extracted with ether. The ethereal extract was dried, concentrated, and distilled to separate 10.232 g. (65%) of purela 6-acetoxy-2-hexanone as a colorless liquid: b.p. 82-83' (2.9 mm.); 72% 1.4266; infrared absorptionlo a t 1725 (C=O) and 1740 cm.-l (ester ' 2 4 ) ; an ultraviolet maximumllat 275 mp ( 6 21.3); and n.m.r. peaks10 a t S 3.98 (2H triplet, J N 6 c.P.s., -CHrOAc) and 2.43 (2H triplet, J 6 c.P.s., -CH&O-) as well as two singlets (each 3H) a t 2.07 and 1.97 (CHrCO- and CHZ-CO-0-) . Anal. Calcd. for C8H1403: C, 60.74; H, 8.92. Found: C, 60.81; H, 8.83. After reaction of 13.43 g. (0.10 mole) of 6-chloro-2-hexanone with 18 g. (0.12 mole) of sodium iodide in 50 ml. of refluxing acetone for 18 hr., the reaction mixture was poured into saturated, aqueous sodium chloride and extracted with ether. The ethereal extract was washed with aqueous sodium thiosulfate, dried, and distilled t o separate 17.18 g. (76%) of 6-iodo-2-hexanone as a colorless liquid, b.p. 51-53' (0.1 mm.), @ D 1.5118 [lit." b.p. 117' (14 mm.)], which turned brown on standing. The product has infrared absorption10 at 1715 cm.-l (C=O) with n.m.r. peak@ a t S 3.16 (2H triplet, J = 7 c.P.s., -CHrI), 2.40 (2H triplet, J = 6 c.P.s., -CH&O-), and 2.07 (3H singlet, CHaCO-). Application of the same reaction procedure to 14.86 g. (0.10 mole) of 7-chloro-2-heptanone (lb) yielded 16.12 g. (67%) of 7-iodo-2-heptanone as a colorless liquid: b.p. 61-64' (0.1 mm.); n% 1.5064; infrared absorption10 a t 1720 cm.-1 (C=O); and n.m.r. peak@ at 6 3.16 (2H triplet, J = 7 c.P.s., -C&I), 2.37 (2H triplet, J = 7 c.P.s., -CHrCO-), and 2.05 (3H singlet, CHrCO-). Since this compound decomposed on standing to liberate iodine, no further characterization was attempted. Reparation of the Amino Ketones and Their Derivatives.After a solution of 5.62 g. (0.025 mole) of 6-iodo-2-hexanone and 3 g. (0.03 mole) of N-methylpiperidme in 40 ml. of absolute ethanol had been refluxed for 48 hr., the solution was concentrated under reduced pressure and diluted with ethyl acetate t o precipitate 5.949 g. of crude quaternary ammonium salt. Repeated recrystallization from mixtures of absolute ethanol and ethyl acetate afforded 5.606 g. (69%) of the pure methiodide of 6-(l-piperidino)-2-hexanone(IC)as white prisms, m.p. 82-82.5', which exhibited a single spot on thin layer chromatography.16 The product has infrared absorption18 a t 1715 cm.-l (C=O) with n.m.r. peaks1' at S 3.33 (3H singlet, CH3-N++), 2.67 (2H triplet, J = 6 c.P.s., -CHrCO), and 2.18 (3H singlet, CH3-CO-) as well as broad absorption in the regions 3.5-4.0 and 1.6-2.1. Anal. Calcd. for C12HJNO: C, 44.31; H, 7.44; I, 39.01; N, 4.31. Found: C, 44.39; H, 7.70; I, 39.25; N, 4.15. Application of the same reaction and isolation procedure t o 16 g. (0.066 mole) of 7-iodo-2-heptanone and 10 g. (0.1 mole) of Nmethylpiperidine in 40 ml. of absolute ethanol yielded 14.452 g. (64.5%) of the pure1&methiodide of 7-(l-piperidino)-2-heptanone
CHS- CO -CHz (CHs)%- CH2-N 8a, n -2 b , n -3 CHs -CH-CHi[CHz)n -CHz-N I OH 9a,n = 2 b,n = 3
3 3
CH~-CO-CHZ(CHZ)~CHZ-N(CH~)Z 10
Consequently, we have been unable to find reaction conditions which permit the desired cyclization reaction ( i e . , 1 + 4) to predominate over bond cleavage to form acyclic products (e.g., 1 -t 5 ) . Since available half-wave potential measurements6 suggest that formation of the ketyl 2 should be energetically favored over direct carbon-chlorine bond cleavage ( i e . , la or l b 3 6 ) , our results imply that the acyclic products 5 isolated have been derived, a t least in part, by opening of the intermediate alkoxy radical 3 to form 6 followed by further reduction. However, the symmetry of the reactants we have employed does not provide any direct evidence for this implication since cleavage of either of two carbon-carbon bonds in the alkoxy radical 3 would lead to the same carbon radical 6 . Experimental' Preparation of the w-Halo Ketones.-1-Methylcyclopentanol and 1-methylcyclohexanol were each converted t o their hypochlorites and decomposed by previously described procedures.* The resulting chloro ketones were isolatedby distillation and (5) P. Zuman, CoEZection Czech. Chem. Commun., 16, 1107 (1950). (6) H. Smith, "Chemistry in Nonaqueous Ionizing Solvents," Vol. 1, Part 2, "Organio Reactions in Liquid," Interscience Publishers, Inc., New York, N. Y.,1963, pp. 187-190. (7) All melting pointa are corrected and all boiling points are uncorrected. Unless otherwise stated, magnesium sulfate was employed a8 a drying agent, The infrared spectra were determined with a Perkin-Elmer Model 237 infrared recording spectrophotometer fitted with a grating. The ultraviolet spectra were determined with a Cary recording spectrophotometer, Model 14. The n.m.r. spectra were determined at 60 Mc. with a Varian Model A-60 n.m.r. spectrometer. The mass spectra were obtained with a CEC Model 21-130 mass spectrometer. The microanalyses were performed by the Scandinavian Microanalytical Laboratory. (8) (a) J. W. Wilt and J. W. Hill, J . 070.Chem., 26, 3523 (1961); (b) F. D. Greene, M. L. Savitz, F. D. Osterholtz, H. H. Lau, W. N. Smith, and P. M. Zanet, ibid., 28, 55 (1963).
65 1
-
(9) A column packed with Carbowax 20M on Chromosorb W was employed. (10) Determined as a solution in carbon tetrachloride. (11) Determined as a solution in 95% ethanol. (12) T. L. Cairns and B. E. Englund, J . Org. Chem., 21, 140 (1956). . (13) A gas chromatography column packed with Versamid 900 on Chromosorb was employed for this analysis. (14) N. Zelinsky and A. Moser, Ber., 85, 2684 (1902). (15) A silica gel thin layer adsorbent was used with an eluent composed of chloroform-methanol-ooncentrated hydrochloric acid (6:6 : 1 by volume). (16) Determined as a solution in chloroform. (17) Determined as a solution in deuteriochloroform.
652
NOTES
as white prkms: m.p. 69-69.5'; infrared absorption18 at 1710 cm.-' (C=O); and n.m.r. peaks" a t 6 3.37 (3H singlet, CH,Nt+), 2.54 (2H triplet, J = 6 c.P.s., - C H A O ) , and 2.18 (3H singlet, CHrCO-) as well as broad absorption in the regions 1.62.1 and 3.5-3.9. Anal. Calcd. for C13H2JNO: C, 46.02; H, 7.73; I, 37.40; N, 4.13. Found: C, 46.05; H, 7.90; I, 37.60; N, 3.85. After a solution of 0.743 g. (5.0 mmoles) of 7-chloro-2-heptanone and 0.70 g. (7 mmoles) of N-methylpiperidine in 10 ml. of n-butyl alcohol had been refluxed for 48 hr., the mixture was concentrated under reduced pressure. Recrystallization from an absolute ethanol-ethyl acetate mixture separated 0.661 g. (53%) of the pure15 methochloride of 7-(l-piperidino)-2-heptanone (Id) as white needles: m.p. 130.5-131'; infrared &bsorption16a t 1720 cm.-l (C=O); and n.m.r. peaks" a t 6 3.33 (3H singlet, C H r N t + ) , 2.47 (2H triplet, J = 6 c.p.s.,-CHrCO-), and2.13 (3H singlet, CH3-CO-) . Anal. Calcd. for ClSH&lNO: C, 62.80; H, 10.48; N, 5.64. Found: C, 62.64; H, 10.71; N, 5.84. A solution of 13.43 g. (0.10 mole) of 6-chloro-2 hexanone (la) and 21.3 g. (0.25 mole) of piperidine in 50 ml. of absolute ethanol was refluxed for 12 hr. and then cooled and diluted with ether. After the precipitate of piperidine hydrochloride had been separated and washed with ether, the combined ethereal solutions were concentrated, and the residue was distilled. Purela B-(l-piperidino)-2-hexanone (8a, yield, 13.738 g. or 75%) was collected as a colorless liquid: b.p. 58-61' (0.03 mm.); n% 1.4639; infrared absorptionlo at 1725 cm. -1 (C=O); ultraviolet maximum11 275.5 mp (e 24.4); and an n.m.r. peak10 a t 6 2.03 (3H singlet, CHAD-) as well as complex absorption in the regions 2.1-2.5 and 1.3-1.7. Anal. Calcd. for Cl1Hz1NO: C, 72.08; H, 11.55; N, 7.64; mol. wt., 183. Found: C, 72.03; H, 11.63; N, 7.58; mol. wt., 183 (mass spectrum). Application of the same reaction and isolation procedure to 14.86 g. (0.10 mole) of 7-chloro-2-heptanone (Ib) yielded 14.872 g. (75.5%) of pure13 7-(l-piperidino)-2-heptanone (8b) as a colorless liquid: b.p. 81-84' (0.05 mm.); 122% 1.4640; infrared absorptionlo a t 1720 cm.-1 (C=O); an ultraviolet maximumll a t 277 mp ( E 24.4); and an n.m.r. peak10 at 6 2.04 (3H singlet, CH3-CO) as well as complex absorption in the regions 2.1-2.5 and 1.2-1.7. Anal. Calcd. for C12H23NO: C, 73.04; H, 11.75; N, 7.10; mol. wt., 197. Found: 73.03; H, 11.81; N, 7.03; mol. wt., 197 (mass spectrum). A solution of 4.58 g. (25 mmoles) of the amino ketone 8a in 20 ml. of ether was added to 260 mg. (6.8 mmoles) of lithium aluminum hydride in 30 ml. of ether. The resulting mixture was stirred for 15 min., treated with 0.11 ml. of water to precipitate the aluminum salts, and filtered. After the ethereal filtrate had been dried, distillation separated 3.607 g. (78%) of pure13 6(l-piperidino)-2-hexanol (sa) as a colorless liquid, b.p. 81-84' (0.04 mm.), 72% 1.4751. The amino alcohol has infrared absorption10 a t 3630 and 3400 (broad) cm.-1 (unassocd. and assocd. 0-H) with an n.m.r. peak10 a t 6 3.93 (1H singlet, 0-H) as well as a broad multiplet centered at 3.68 (lH, CH-0-), a doublet ( J = 6 c.P.s.) centered at 1.11 (3H, CHI-), and complex absorption in the regions 2.0-2.5 and 1.2-1.7. Anal. Calcd. for CllH23NO: C, 71.30; H, 12.51; N, 7.56; mol. wt., 185. Found: C, 71.04; H , 12.57; N, 7.58; mol.wt., 185 (mass spectrum). Reduction of 4.75 g. (25 mmoles) of the amino ketone l b by the same procedure yielded 3.835 g. (77%) of purela 7-(l-piperidino)-2-heptanol (9b) as a colorless liquid: b.p. 87-91' (0.1 mm.); 72% 1.4749; infrared absorptionlo at 3610 and 3380 (broad) cm.-l (unassocd. and assocd. 0-H); and an n.m.r.1° singlet at 6 3.55 superimposed on a multiplet (2H, CH-OH), a doublet (J = 6 c.P.s.) centered at 1.08 (3H, CHs-), and complex absorptionin the regions 2.0-2.5 and 1.1-1.7. Anal. Calcd. for ClzH26NO: C, 72.30; H, 12.64; N, 7.03; mol. wt., 199. Found: C, 72.36; H, 12.71; N, 6.93; mol. wt., 199 (mass spectrum). The sodium enolate of ethyl acetoacetate was alkylated with 3(N,N-dimethy1amino)propylchloridela and the crude product was hydrolyzed and decarboxylated following the general procedure described previously.19 Distillation separated pure20 6-(N,N(18) R. Marechal and J. Bagot, Ann. phurm. j ~ u n c .4, , 172 (1946); Chem. Abstr., 41, 5099 (1947).
VOL.30
dimethylamino)-2-hexanone (IO, yield 41.5%) as a colorless Iiquid: b.p. 80-81' (10 mm.); n% 1.4295 [lit.Zl b.p. 194-195O (720 mm.)]; infrared absorptionlo at 1720 cm.-1 (C=O); an ultraviolet maximumll a t 274 mp (e 22.7); and n.m.r.10 singlets a t 6 2.14 (6H, CHrN-) and 2.07 (3H, CHa-CO-) as well aa complex absorption in the regions 2.0-2.5 and 1.2-1.8. A solution of 3.635 g. (0.026 mole) of this amino ketone 10 and 4.34 g. (0.031 mole) of methyl iodide in 50 ml. of acetone was allowed to stand overnight and then concentrated under reduced pressure. Recrystallization of the residual salt from an absolute ethanol-ethyl acetate mixture separated 4.434 g. (60%) of the pure16 methiodide of 6-(N,N-dimethylamino)-2-hexanone(le) as white needles: m.p. 88.689'; infrared absorption18 at 1710 cm.-l (C=O); and n.m.r. peakszz a t 6 3.16 (9H singlet, CH3-Nf +), 2.69 (2H triplet, J = 7 c.P.s., -CHrCO-), and 2.24 (3H singlet, CHrCO) as weg as unresolved absorption of the regions 3.1-3.5 and 1.4-1.9. Anal. Calcd. for CaHzoINO: C, 37.91; H, 7.08; I, 44.51; N, 4.91. Found: C, 38.57; H, 7.58; I, 44.90; N, 4.65. Reduction of 7-Chloro-2-heptanone (Ib). A. With Lithium in Liquid Ammonia.-A solution of 3.55 g. (24 mmoles) of the chloro ketone l b in 300 ml. of ether was added, dropwise and with stirring over a 3-hr. period, to a solution of 378 mg. (54 mg.-atoms) of lithium in 400 ml. of liquid ammonia and 200 ml. of ether. The addition was interrupted at the end of 2 hr. to add an additional 90 mg. (13 mg.-atom) of lithium. After the addition was complete, the reaction mixture was stirred for 1hr. and then the excess lithium was destroyed by the addition of 1 ml. of methanol. The ammonia was allowed to evaporate through a 90-cm. Vigreux column and the remaining ethereal solution was washed successively with aqueous hydrochloric acid, aqueous sodium bicarbonate, and water. After the ether solution had been dried and concentrated, the volatile portion of the residual neutral fraction was analyzed by gas chromatography9 and found to contain, apart from ether and ethanol, 2-heptanone (5, n = 3, 67%, first eluted), 2-heptanol (26%, second eluted), and 1methylcyclohexanol (4, n = 3, 7%, third eluted). The products from a comparable reduction were identified both by comparison of retention times and by comparison of the infrared and mass spectra of collected samples with spectra of authentic samples. B. With the Sodium-Naphthalene Radical Anion.-The chloro ketone l b (743 mg. or 5 mmoles) was added to a solution prepared from 460 mg. (20 mg.-atoms) of sodium and 1.921 g. (15 mmoles) of naphthalene in 80 ml. of 1,2-dimethoxyethane. An additional small portion of sodium was added to the solution t o maintain the green color of the naphthalene radical anion. The resulting mixture was concentrated under reduced pressure and then mixed with saturated, aqueous ammonium chloride and extracted with ether. The ethereal extract was washed successively with aqueous hydrochloric acid, aqueous sodium bicarbonate, and water and then dried and concentrated. Analysisa of the volatile portion of the crude, neutral product indicated the presence of 2-heptanone (5, n = 3, 72%) and l-methylcyclohexanol (4, n = 3, 280/0). Attempts to reduce the chloro ketone l b with chromium(I1) chloride in aqueous acetone resulted in recovery of the starting chloride, and no 1-methylcyclohexanol (4, n = 3) was detected in the crude neutral product obtained by reduction of the chloro ketone l b with sodium and isopropyl alcohol. Reduction of 6-Chloro-2-hexanone (la) .-The chloro ketone l a (581 mg. or 4.3 mmoles) was added, dropwise and with stirring, to a solution prepared from 345 mg. (15 mg.-atoms) of sodium, 1.335 g. (7.5 mmoles) of phenanthrene, and 40 ml. of 1,2-dimethoxyethane until the dark green color of the radical anion waa just discharged. The crude, neutral product was isolated as in the previous experiment. The volatile portion of this product contained9 four components, A (2%), B (60%), C (33%), and D (Sa/,), listed in order of increasing retention time. Components B and C were identified as 2-hexanone (5, n = 2) and l-methylcyclopentanol (4, n = 2), respectively. For a comparable reduction employing 6-iodo-2-hexanone the volatile portion of the crude, neutral product containedg, in order of elution, an uniden(19) D. S. Breslow, R. S. Yost, H. G. Walker, and C. R. Hauser, J . Am. Chem. Soc., 66, 1921 (1944). (20) A gas chromatography column packec! with silicone gum, No. SE-30, suspended on Chromosorb W was employed for this analyeis. (21) A. Lipp, Ann., 289, 249 (1896). (22) Determined as a solution in deuterium oxide.
NOTES
FEBRUARY 1965
tp -
653
tified component (1%), 2-hexanone (31%), 2-hexanol (5%, GLC 1-methylcyclopentanol (22%), and cyclohexanone (41%) Reduction of the Methochloride of 7-( l-Piperidino)-2-heptanone (Id).-Lithium (158 mg. or 23 mg.-atoms) was added, portionwise with stirring, to a solution of 2.275 g. (9.2 mmoles) T of the salt Id in 600 ml. of liquid ammonia until the solution I ll maintained a blue color (indicative of excess lithium) for 15 6. Then the excess lithium was destroyed by the addition of methanol and the ammonia was allowed to evaporate through a 90cm. Vigreux column. The residue was mixed with saturated, aqueous sodium chloride and extracted with ether. After the ethereal extract had been washed with water, dried, and concentrated, the residual oil (1.132 g.) was found to contain1*three CH, OH volatile components: an unidentified component (4% first eluted), 7-(l-piperidino)-2-heptmone(8b, 34%, second eluted), and 7-(l-piperidino)-2-heptanol(9b, 62%, third eluted). The latter two samples were identified both by retention times and by comparison of the infrared and mass spectra of collected samples with the spectra of authentic samples. Reduction of the Methiodide of 6-(l-Piperidino)-2-hexanone (IC).-The preceding experimental procedure was applied to 1.882 g. (5.6 mmoles) of the salt IC,134 mg. (19 mg.-atoms) of lithium and 400 ml. of liquid ammonia being employed. The crude basic fraction, 875 mg. of liquid, contained13 the following volatile components: an unidentified product (7%, first eluted), m C H 0 thiokrtol 6-(l-piperidino)-2-hexanone(sa, 34%, second eluted), and 6-(1piperidino)-2-hexanol (sa, 59%, third eluted). As in the previous cme, the products were identified by comparison of retention times, infrared spectra, and mass spectra. Reduction of the Methiodide of 6-(N,N-Dimethylamino)-2- Figure 1.-Rearrangements and transformations of caryophylhexanone (le) .-The previously described reaction and isolation lene oxide and dihydrocaryophyllene oxide. procedure was followed with 2.121 g. (7.5 mmoles) of the salt le, 181 mg. (26 mg.-atoms) of lithium, and 500 ml. of liquid describe an interesting rearrangement taking place ammonia. The crude, basic product, 784 mg. of liquid, conduring gas chromatography of caryophyllene oxide, a tained's three volatile components: A (5%, &st eluted), B (17%, second eluted), and C (78y0, third eluted). Product B terpenoid occurring in a number of commercially imwas identified as 6-(N,N-dimethylamino)-2-hexanone(IO) by portant essential oils. comparison of the infrared spectrum of a collected sample with The transformation of caryophyllene oxide to comthe spectrum of an authentic sample and from its mass spectrum pounds possessing novel carbon skeletons has been the with peaks a t m / e 143 (Mf), 58 [abundant, (CH&N+=CH2 subject of a number of investigation^.^*^ The authors and CHs(OH)C=CH2+], and 43 (CH3C=O+). Component C was tentatively identified as 6-(N,N-dimethylamino)-2-hexanol found that the epoxide I decomposed completely when from the following spectral characteristics of a collected sample. subjected to gas-liquid chromatography through a The material has infrared absorption10 at 3620 and 3400 (broad) column of Reoplex 400 (10% on acid-washed Chromocm.-l (unassocd. and assocd. 0-H) with n.m.r.lo singlets at sorb W). Infrared and n.m.r. examination proved 6 2.62 (lH, OH) and 2.15 (6H, CHI-N) as well as a doublet (J = 6 c.P.s.) centered a t 1.10 (3H, CHa-) and broad absorption in the the major degradation product (62%) to be an unsaturegions 3.5-3.9, 2.1-2.4, and 1.3-1.6. The mass spectrum has rated aldehyde I1 with infrared absorption bands a t peaks at m / e 145 (M+), 130'(M+-CH3), 58 [abundant, (CH& 3080, 1632, and 885 cm.-l (>C=CH2) and 2688 and N+=CHt], 45 (CH&H=O +H), and 44 (CH$N+H==CH2).
.
p,,
Essential Oils and Their Constituents.
XXV1.l Rearrangement of Caryophyllene Oxide during Gas Chromatography ISHWAR C. NIGAMAND LEO LSVI Pharmaceutical Chemistry Division, Food and Drug Directorate, Department of National Health and Welfare, Ottawa, Canada Received September 2, 196'4
The application of chemical and instrumental methods of analysis to essential oils and their constituents requires discriminating interpretation of experimental results. I n a previous paper the authors reported the transformation of humulene monoxide, a constituent of oil of wild ginger, to the corresponding alcohol during column chromatography over grade I aluminaS2 It is the purpose of this communication to (1) Paper XXV: J. Soc. Cosmetic Chemists, in press. (2) I. C. Nigam and L. Levi, J . Org. Chem., 29, 2803 (1964).
+P
1726 cm.-l (-CHO) and n.m.r. peaks at 9.30 (-CHO), 4.40, and 4.61 p.p.m. (>C=CH2) (tetramethylsilane = 0). Catalytic hydrogenation of the aldehyde ( C I ~ H ~ ~showed O) the presence of one double bond. Chromic acid oxidation of the saturated alcohol (111) obtained yielded the corresponding saturated aldehyde V having infrared absorption bands a t 2698 and 1726 cm.-l and n.m.r. peak a t 9.26 p.p.m. This aldehyde was also formed when subjecting dihydrocaryophyllene oxide (IV) to gas chromatography under the experimental conditions. The saturated aldehyde V was, in turn, converted via desulfurization of its thioketal with Raney nickel to the corresponding saturated hydrocarbon VI. Gas chromatographic and infrared characteristics of this hydrocarbon were different from those displayed by caryophyllane. These observations suggested that a rearrangement of the caryophyllane carbon framework had taken place. Quantitative analysis of the n.m.r. spectra of aldehydes I1 and V established the presence of three and four methyl groups, respectively. Since during the rearrangement reaction the number of (3) D. H. R. Barton and P. de Mayo, J. Chem. Soc., 150 (1957); Quart. Rev., 11, 189 (1957). (4) E. W.Warnhoff, C5n. J . Chem., 42, 1664 (1964).
654
NOTES
methyl groups of each epoxide molecule remained the same and an aldehyde group was generated, it was concluded that during gas chromatography the n i n e membered ring of the caryophyllene skeleton undergoes contraction to an eight-membered ring. Figure 1 illustrates the rearrangement of the epoxides to 4,10,10- trimethyl -7 -methylenebicyclo [6.2.0]decane-4carboxaldehyde (11) and 4,7,10,10-tetramethylbicyclo[6.2.0]decane-4-carboxaldehyde (V) , and also shows the conversion of either aldehyde to the saturated hydrocarbon, 2,5,5,9,9-pentamethylbicyclo [6.2.0]decane (VI). Formulas shown in Figure 1 are also in agreement with Warnhoff's recent work.4 Rearrangement of the epoxides failed to take place when the support (acid-washed Chromosorb W) wa;s treated with methanolic KOH15and used under the experimental conditions described. In this case unchanged starting materials only were recovered. Evidently the transformation is triggered by active surfaces of the acidic column support. The reactions observed are of considerable value for the examination of essential oils and related products. They were applied successfully by the authors for detecting the presence of caryophyllene oxide in oil of marihuana for the first time.6 Experimental
Gas chromatographic apparatus and procedures were p r e V~OUSIY described.' Infrared spectra were charted with aPerkinElmer Model 221 instrument. N.m.r. spectra were recorded in carbon tetrachloride solution, with a Varian A-60 spectrometer. Caryophyllene Oxide (I).-The terpenoid was obtained by epoxidation (monoperphthalic acid) of pure caryophyllene (1.0 9.) isolated by preparative gas chromatography from oil of copaiba balsam. Recrystallization of crude product from methanol yielded pure substance (620 mg., m.p. 64'). Rearrangement of Caryophyllene Oxide to 4,10,10-Trimethyl- 7 -methylenebicyclo[6.2.0]decane-4-carboxaldehyde.Caryophyllene oxide (500 mg., in a few drops of benzene) was injected repeatedly into a gas chromatographic column packed with Reoplex 400 (10%) deposited on acid-washed Chromosorb W. (column temperature 200°, helium flow 75 ml./min.). Effluents corresponding to the major peak observed were collected in small tubes packed with glass wool and moistened with carbon tetrachloride. The aldehyde obtained (280 mg.) possessed the foldZ02o 0.993, retention time ~ lowing characteristics: n Z 61.499, 8.4 min. Anal. Calcd. for C16H240: C, 81.75; H, 10.97. Found: C, 81.4; H, 11.2. Dihydrocaryophyllene Oxide (IV) .-Caryophyllene oxide (I) (101 mg.) waa dissolved in 5 d.of ethanol and hydrogenated in the presence of Adam catalyst (8 mg.) at room temperature and atmospheric pressure (hydrogen uptake 13.6 ml., yield of crude dihydrocaryophyllene oxide 95 mg.). 4,7,10,10-Tetramethylbicyclo[6.2.O] decane-harboxaldehyde. A. From Aldehyde (II).-The conversion was carried out in two steps. Hydrogenation to Alcohol 111. A sample (200 mg.) waa hydrogenated in ethanolic solution in the presence of Adam catalyst (25 mg.) (volume of hydrogen absorbed 40 d.).The product of hydrogenation (192 mg.), recovered after usual processing, displayed strong hydroxyl absorption at 3390 cm.-'. Retention time on Reoplex 400 (lOo/o) was 10.9 min. (temperature 200"; helium flow 75 ml./min.). Oxidation of Alcohol III.-The alcohol (180 mg.) waa dissolved in pyridine (2 ml.) and allowed to react overnight with a solution of chromium trioxide (180 mg.) in pyridine (4 ml.). ( 5 ) R. B. Bates, D. M. Gale, and B. J. Gruner, J . Ore. Chem., 38, 1086 (1963). (8) M. C. Nigam, I. C. Nigam, K. L. Hands, and L. Levi, unpublished data. (7) I. C. Nigam and L. Levi, Con. J . Chem., 40,2083 (1962).
VOL. 30 The reaction mixture waa diluted with water and extracted with ether. The ether extract was treated with dilute sulfuric acid and, after washing with water, dried over sodium sulfate. The crude aldehyde (130 mg.) obtained following evaporation of solvent was chromatographed over 6.5 g. grade I1 alumina, using petroleum ether and benzene aa eluents. The sample recovered from the benzene fraction and purified by gas chromatography exhibited a retention time of 7.8 min. on Reoplex 400 (10%) (column temperature 200', helium flow 75 ml./min.). B. From Dihydrocaryophyllene Oxide.Smal1 samples (15 A ) of dihydrocaryophyllene oxide (IV) were passed through the Reoplex column at 200" and the major component was collected for analysis. Gas chromatographic and infrared characteristics of the product obtained were identical with those of the sample prepared in accordance with method A. Hydrogenation of the aldehyde yielded the saturated alcohol 111, its identity being established by infrared and gas chromatographic analysis. 2,5,5,9,9-Pentamethylbicyclo[6.2.0]decane (VI).-Boron trifluoride etherate (50 J.) was added slowly to an ice-cooled mixture containing the saturated aldehyde V (10 mg.) and ethanedithiol (45 pl.). Glacial acetic acid was added drop by drop until a homogeneous solution was obtained. The reaction mixture was kept at room temperature for 1 hr., poured into an aqueous solution of potassium carbonate, and extracted with ether. The ether extract waa washed repeatedly with 10% sodium chloride solution, dried over anhydrous sodium sulfate, and evaporated. The residue waa refluxed with 2 ml. of a BUSpension of Raney nickel overnight, the catalyst was filtered off, and the a t r a t e was diluted with 10 ml. of water and extracted with petroleum ether. The extract waa dried over anhydrous sodium sulfate and evaporated. The residue (9 mg.) was analyzed by gas chromatography. Retention time of hydrocarbon relative to caryophyllane was 0.69 (column 10% silicone nitrile XE-60, temperature 130', helium flow 75 ml./min.).
Acknowledgment.-The authors are indebted to Dr. 0. E. Edwards, Pure Chemistry Division, National Research Council of Canada, for helpful discussions, n.m.r. determinations, and microanalyses, to Dr. E. Guenther, Fritzsche Brothers, Inc., New York, for courteously providing a fraction of oil of copaiba balsam rich in caryophyllene, and to Dr. Y. R. Naves, L. Givaudan and Cie., Vernier-Genhve, Switzerland, for kindly supplying a reference sample of caryophyllene oxide for this study.
T h e Isolation of PaIustric Acid from Gum Rosin N. MASONJOYE,JR., VIRGINIA M. LOEBLICH, AND RAYV. LAWRENCE
Naval Stores Labaratory,l Olustee, Florida Received October 19, 1964
Palustric acid, one of the major components of rosin, was f i s t isolated by Loeblich, et a1.,2and its structure was established by Schuller, et da This acid represents about 10% of the acidic portion of pine gum and up to 18% of the acidic portion of gum rosin. It is a homoannular, conjugated-diene resin acid, 265-266 mp (E 3300), [ a ]+71.8', ~ and m.p. 162-167'. (1) One of the laboratories of the Southern Utilization Research and Development Division, Agricultural Research Service, U. S. Department of Agriculture. (2) V. M. Loeblich, D. E. Baldwin, and R. V. Lawrence, J . Am. Chem. Soc., 77, 2823 (1955). (3) W. H. Schuller, R. N. Moore, and R. V. Lawrence, {bid., BP, 1743 (1980).
NOTES
FEBRUARY 1965 CHART 1 Starting material: 200 g. of gum rosin 600 ml. of acetone 68 g. of 2,6dimethylpiperidine
+
+
Discarded
62 g., [a]D 18', 80 ml. MeOH, 80 ml. acetone C
+
+
for acid,
3
1 41 discarded
41'2, 11 g., [a]D
0
+ 73") C
+ 15'
&c,19 g., [.ID
7.5 g., [a]D 46", 8 ml. MeOH, 8ml. acetone C
+5
4
The composition of the starting rosin, the crude salt, and the salt at each step in the purification was deter&ed by gas chromatography of the methyl esters. Table I shows the progress of the purification. Experimental
211evaporated to I/* vol.
16 g., [a]D 35', 20 ml. MeOH, 20 ml. acetone C
([a]D 5.8 g.1calcd. Ia1D
+
655
M
1 511 discarded 51c, 6.2 g., [a]D
&c,3.8 g., [a]D
+ 32'
+ 46'
+ 52'
Palustric acid is an intermediate and a product of the thermal and acid isomerization of l e v o p i m a r i ~ ~ and *~ neoabieticss7acids. On thermal isomerization palustric acid yields an equilibrium mixture of abietic, neoabietic, and palustric acids.* The only method available for the isohtion of palustric acid has been by partition chromatography.2 The method described in this paper is baaed on the precipitation of the 2,6ðylpiperide salt of the resin acids from an acetone solution of pine oleoresin or rosin followed by selective crystallization of the salt from a methanol-acetone (1:l) solution. The yield of pure palustric acid from slash gum rosin was 4%. Gum rosin is preferable as a starting material since, aa shown in Table I, abietic acid is coprecipitated along with palustric acid and gum rosin has a more favorable palustric acid to abietic acid ratio than either S.D. wood rosin or tall oil rosin.
Isolation of the 2,6-Dimethylpiperidine Salt of Palnstric Acid from Gum Rosin.-WW gum roein (200 g., acid no. 168) waa dissolved in 600 ml. of acetone. To the roein solution waa added with stirring 68 g. of 2,6-dimethylpiperidine. The solution WBB allowed to stand at room temperature overnight and the precipitated salt waa removed by filtration. The salt waa removed from the filter and waahed twice by slurrying with Wml. portions of warm acetone. The resulting white, crystalline amine salt waa diesolved in an equal weight of hot methanol, filtered, and an equal volume of acetone waa added. A modified scheme of triangular recrystsllization diagrammed in Chart I waa used. Combination of fractions &C and 6& gave 9.6 g. of salt with [a]=D +52" (c 1.0, alcohol), A$, 265 mp ( e BOO), and m.p. 156-162' (sealed evacuated tube). Thia weight of amine salt represents a 4% yield of palustric acid baaed on the acid number of the rosin. Conversion of Amine Salt to the Acid.-A 10-g. portion of the amine salt waa converted to the free acid by diesolving it in 350 ml. of 95% ethanol and adding, with stirring, 100 ml. of cold 3 N HiPo,. Ice-water was added to the cold acidified solution until no further cloud appeared. The precipitated acid was washed with water to remove excess mineral acid and recyatallized once from a minimum amount of hot 95% ethanol. Since paluetric acid is isomerized to abietic acid by strong acids it should be separated from the acidified solution aa rapidly aa poasible. The specific rotation of the final product waa +69',O while a at 265266 mp waa 28.2, and m.p. 162-167'. Further recrystallizations showed no improvement in purity. (9) The diEerence in the specSc rotation of the free add and that dculated for the w i d from the salt is assumed to be caused by some resolution of the amine during recrystallization.
New Syntheses of Pyrrolo[2,3-d]-and Pyrrolo[3,2-d]pyrimidines1 EDWARD C. TAYLOR AND EDWARD E. GARCIA Department of Chemistry, Prtnceta University, Princeton, New Jersey Received October 7,1964
TABLEI COMPOSITION OF METHYL ESTERS W L E S %" of each peak off Peak*
Gum rosin
Crude nalt
Salt 21C
..
g.1.0.
Salt 31C
Salt 41C
.. .. ..
Me pimarate 6 2.0 .. Unidentified 2.8 .. .. .. .. Me elliotinoate 3.2 . .. Mepalustrate 20.4 47.3 76.9 83.7 96.7 .. .. Me isopimarate 21.2 5.4 Me abietate 33.6 37.2 22.0 16.3 3.3 Me neoabietate 12.8 8.1 1.4 .. Per cent of the material that comes off the column a t 225', 5% Craig Polyester. Each peak except 2 was identified by comparison with an authentic sample of the resin acid methyl ester.
.
..
Interest in pyrrolopyrimidines hss been greatly heightened as a result of the recent discovery that both the [2,34- and [3,2d-] systems occur aa heterocyclic bases in a number of antibiotics. Tubercidin (la) and Toyocamycin (lb) have been shown to be derivatives of Paminopyrrolo [2,M]pyrimidine,2 and Viomycin (2) has recently been shown to possess a dihydropyrrolopeptide
..
(4) V. M. Loeblich,D. E. Bddwin, R. T. O'Connor, and R. V. Lawrence J . Am. Chem. SOC.,77, 2823 (186.5). ( 6 ) D. E. Baldwin, V. M. Loeblich, and R. V. Lawrence, Wd., 78, 2016 (1966). (6) V. M. Loeblich and R. V. Lawrence, ibid., 79, 1487 (1967). (7) D. E. Baldwin, V. M. Loeblich, and R. V. Lawrence, J . Cham. Em. Data, 8, 342 (1968). (8) N. M. Joye, Jr., and R. V. Lawrence, J . Org. Chum.. 96, 1024 (1861).
& R-$
QQ
N A N
I
R' la, R = H ; R'=p-D-ribqse b, R=CN; R=D-ribose
H
2
(1) Thin work was sup~ortedin part by a research grant (CY-02661) b Princeton University from the National Cancer Institute, National Institutes of Health, Public Health Service. (2) For a recent discussion and literature referenccrs, E. C. Taylor and R. W. Hendeed, J . Am. Chem. Soc., 86, 861 (1864).
NOTES
656
[3,2d]pyrimidine moiety.8 We wish to describe several new synthetic methods for the preparation of derivatives of both ring systems. Synthetic approaches to pyrrolo [2,3d]pyrimidines (7-deazapurines) have recently been reviewed, and a new approach to this system via pyrrole intermediates has been discussed.2 Not included in this survey was a recently described approach to derivatives of this system by reaction of l13-dimethyl-6-aminouracilwith chloroacetaldehyde or chloroacetyl chloride to give 3 and 4, re~pectively.~We have found an additional synthetic approach to this ring system from 6-aminopyrimidine derivatives based upon a reaction first 0
VOL. 30
appeared to be available by adaptation of the nitreneinsertion reaction described by Cadogan and CameronW00d,13 who showed that treatment of o-nitrostilbene with triethylphosphite gave 2-phenylindole. Requisite starting materials for the preparation of pyrrolo [3,2d]pyrimidines by this route would be 5-nitro-6-styryl derivatives. Two representatives of this class of compounds have been prepared by the condensation of 1,3,6-trimethyl-5-nitrouracilwith pmethoxybenzaldehyde and pdimethylaminobenzaldehyde to give 7a and 7b, respectively. When these compounds were heated
R
8
0 I
CH3 8
CH, H 4
described by Hirose, Takagi, and UnoGand by Osterheld,6in which the reaction of phenacylpyridinium bromide with aromatic primary amines was shown to give substituted indoles. We have found that the condenssr tion of l13dimethyl-6-anilinouracil(5) , prepared from the corresponding 6-amino compound via the exchange amination procedure of Whitehead and Traverso17with phenacylpyridinium bromide yields 1,3-dimethyl-6,7diphenyl-7H-2,4(1H,3H)pyrrolo [2,3dIpyrimidinedione (6). This condensation would appear to be of potential interest for the preparation of further 6,7-disubstituted derivatives.
a t reflux under nitrogen with triethylphosphite, they were converted in moderate yield to the pyrrolo 13,241pyrimidines 8a and 8b. Since the reaction of trialkylphosphites with oxygen (to give trialkylphosphates) hm been shown to be photoinitiated,14 it appeared reasonable that a photoinitiated oxidation of triethylphosphite by a nitro group might take place, provided that atmospheric oxygen were excluded. Irradiation of a dilute solution of 7a in triethylphosphite under nitrogen for 114 hr. did indeed result in the formation of the pyrrolo [3,2d]pyrimidine 8a in small yield. The very low solubility of the starting pyrimidine in triethylphosphite may well have been the limiting factor in this conversion. Although the nitrene insertion reactions do not proceed so satisfactorily as in previously reported, nonheterocyclic cases, they may offer some synthetic advantage for the preparation of condensed pyrrole derivatives which might otherwise be difficultly accessible. Experimental"
6% C& 6
Previously available synthetic routes to isomeric pyrrolo [3,2d]pyrimidines involved the cyclization of a malonic ester, pyruvate ester, acetic ester, or methyl 6-substituent with an amino or acetylated amino group or a Schiff base in the 5-position of the pyrimidine ring.8-12 A potentidy attractive synthetic alternative (3) J. H. Bowie, A. W. Johnson, and G. Thomas, Tstrohsdron Lefteu, No. 16, 863 (1964): J. H. Bowie, D. A. Cox, A. W. Johnson, and G. Thomas, ibid., No. 46, 3305 (1064). (4) C. W. Noell and R. K. Robins, J. Hetmocyclic Chem., 1, 34 (1964). (5) 8. Hirose, 9. Taksgi, and T. Uno. Yakugoku Zosshi, 81, 1363 (1961); Chem. Abutr., 66, 7255s (1982). (6) K. Oeterheld, Diploma Thesis, University of Giessen, 1961; cited by F. KrBhnke, Anusto. C k m . , I n t a . Ed. Enol., 1, 226 (1983). (7) C. W. Whitehead and J. J. Traverso, J. Am. Chum. SOC.,81, 3973 (1960).
1,3-Dimethyld-anilinouracil(5).-A mixture of 1.55 g. (0.01 mole) of 1,3-dimethyl-6-aminouracil, 0.9 g. of aniline hydrochloride, and 1.5 ml. of aniline was heated a t 150' for 3 hr. Water and ea. 50 ml. of chloroform were added to the cooled sirup and, after thorough mixing, the chloroform was separated and waahed twice with water. Evaporation of the chloroform gave a white solid which waa recrystallized from chloroform-ethyl acetate to give 1.15 g. (50y0)of white crystals, m.p. 185' (lit.10 m.p. 181182'). (8) K. Tanaka, T. Sugawa. R. Nakamori, Y. Sanno, and Y. Ando, Yokuqoku Zosahi, 76, 770 (1956); Chem. Abutr., 49, 14001 (1955). (9) F. L. Row, J . Chem. Soc., 4116 (1954). (IO) K. Tanaka, et d., Japan Patent 223 (Jan. 21,1965): Chem. Abefr., 60, P16879 (1956). (11) K. Tanaka, T. Sugawa, Y. Sanno, and Y. Ando, Japan Patent 1375 (Feb. 27, 1958); Chem. Abetr., 69, P1389b (1959). (12) W. Pfleiderer and H. Mosthaf, Be.,90, 738 (1957). (13) J. I. G.Cadogsn and M. Cameron-Wood, Proc. Chem. Soo. (London), 361 (1962). (14) J. B. Plumb and C. E. G r i 5 n , J. Ore. Chem., 18, 2908 (1963). (15) Microanalyses were performed by the Spang Microanalytical Laboratory, Ann Arbor, Mich. Melting pointa were determined on a ThomasHoover silicon bath apparatus. (16) W. Pfleiderer and H. Ferch, Ann., 616, 52 (1958).
FEBRUARY 1965 Anal. Calcd. for ClzHlsNaOa: C, 62.32; H, 5.671 N, 18.17. Found: C, 62.00; H, 5.64; N,18.25. 1,J-Dimethyl-6,7-diphenyl-7H-2,4( lH,3H)pyrrolo 12,341p m - . idinedione (6).-Anbtimate mkture of 2.3 g. (0.01mole) of 13dimethyl-6-anilinouracil and 3.1 g. (0.011 mole) of phemcylpyridinium bromide was heated in an oil bath a t 215', under a strong stream of nitrogen, for 3 hr. The resultant brown &S was dissolved by boiling with cu. 15 ml. of ethanol. After overnight refrigeration, filtration gave 0.5 g. (15'%) of a cream-colored solid, m.p. 219-221 '. Recrystallization from ethanol yielded white needles, m.p. 223-224'. Anal. C a l c d M z o H l , ~ 72.49; C ~ H, 5.17; - N, 12.68. Found: C, 72.82; H, 5.36; N,12.71. 1,3-Dimethyl-S-nitro-6-( p-methoxy)styryhuacil (7a).-A mixture of 4.0 g. (0.02 mole) of 1,3,6-trimethyl-5-nitrouracfi,2.8 g. (ca. 0.02 mole) of anisaldehyde, 2 ml. of piperidine, and 90 ml. of ethanol was heated to reflux with magnetic stirring. Solution occurred shortly after the reflux temperature was reached. After 1.5 hr . a yellow solid began t o separate, and after 20 hr. of heating the suspension was filtered hot t o give a canary yellow solid, m.p. 186-187'. Concentration of the filtrate to a small volume yielded a yellow solid which was washed thoroughly with boiling ethanol. The combined yield was 5.2 g. (82%). Recrystalbation from 1-butanol gave yellow needles, m.p. 186-187". Anal. Calcd. for ClsH15NsOl: C, 56.78; H, 4.77; N, 13.24. Found: C, 56.72; H, 4.59; N, 13.20. 1,3-Dimethyl-S-nitro-6-( p-dimethylamino)styry1uracil (7b).A suspension of 2.0g. (0.01mole) of 1,3,6-trimethyl-5-nitrouracil, 1.5 g. (0.01 mole) of p-dimethylaminobenzaldehyde, 1 ml. of piperidine, and 45 ml. of ethanol was heated, with stirring, to reflux. The resulting deep red solution was heated under reflux for 20 hr. and then filtered hot to give 2.5 g. of orange-red crystah, m.p. 205-207" dec. Concentration of the filtrate t o a small volume gave a red solid which after recrystallization from ethanolbenzene yielded 0.3 g. of additional product; total yield 2.8 g. (85%). Recrystallization from aqueous dimethylformamide gave bright red plates, m.p. 207-208" dec. Anal. Calcd. for CI&N~O~: C, 58.17; H, 5.49; N, 16.96. Found: C, 58.42; H, 5.71; N, 16.80. 1,3-Dimethyld-( p-dimethylaminophenyl)-SH-2,4( 1H,3H)pyrrolo[3,2-d]pyrimidinedione (Sb).-A mixture of 1.65 g. (0.005 mole) of 1,3-dimethyl-5-nitro-6-(p-dimethylamino)styryluracil and 5 ml. (large excess) of freshly distilled triethylphosphite waa refluxed under nitrogen for 5.5 hr. After standing overnight a t room temperature, the mixture was filtered to give an ambercolored solid, m.p. >300". Distillation of the filtrate under reduced pressure (ca. 0.1 mm.) gave a red sticky mass which when triturated with ethanol yielded a few milligrams of additional product. Recrystallization of the combined solids from dimethylformamide followed by vacuum sublimation at 240-250' (0.05 mm.) gave 0.9 g. (60%) of pale yellow solid, m.p. 310-318' dec., Y : 3225 cm.-'. Anal. Calcd. for C I ~ H ~ ~ N ~C,O64.41; Z: H, 6.08; N, 18.78. Found: C,64.33; H, 5.99; N,18.79. 1,3-Dimethyld-( p-methoxyphenyl)-SH-2,4(lH,3H)pyrrolo[3,2-d]pyrimidinedione (Sa). Method A.-A mixture of 1.4 g. (0.0044mole) of 1,3-dimethyl-5-nitro-6-(p-methoxy )styryluracil and 4 ml. of triethylphosphite was refluxed for 7 hr. under nitrogen. After standing overnight at room temperature, the mixture was filtered and the solid was waahed with ether. Distillation of the filtrate gave a red viscous oil which when treated with ethanol gave a few additional milligrams. Recrystallization of the combined solids from aqueous dimethylformamide followed by vacuum sublimation at 250' (0.05mm.) gave 0.5 g. (40%) of white crystals, m.p. 304-306' dec., YZ 3200 om.'. Anal. Calcd. for C I S H ~ ~ N ~C, O ~63.15; : H, 5.30; N, 14.73. Found: C,63.03; H, 4.84; N,14.74. Method B .-A solution of 0.2 g. of 1,3-dimethyl-5-nitro-6-(pmeth0xy)styryluracil in 30 ml. of warm triethylphosphite in a quartz tube waa degassed by bubbling in nitrogen for a few minutes, and the tube was stoppered and irradiated for 114 hr. at 3500 A. in a Rayonet photochemical chamber. During this time a very small amount of solid separated. Filtration and washing with ether gave a few milligrams of a yellow solid, m.p. 300' dec., which was shown to be identical with the product obtained by method A above by a comparison of infrared spectra and by a mixture melting point determination.
NOTES
657
Organio Disulfides and Related Substances.
XIII. Oxidation of Some Aryl DisuLGdesla-c L~LMAR FIELDAND TIMOTHY F.PAR60N61d Department of Chemistry, Vanderbilt University, Nashville, Tennessee Received August 18, 196.4
A conversion of aliphatic disulfides to aliphatic thiolsulfonates (RS02SR)has been developed by Douglass and Farah which affords an elegant and valuable synthesis of aliphatic thiolsulfonates. The procedure involves chlorinolysis in the presence of acetic acid and then treatment with water. The sequence of events is shown by these equations.2 RSSR
+ CHsCOzH + 2C12 + RSOCl
+ RSCl + CHsCOCl + HC1 J2Hz0
RSOzSR
+ CHsCOzH + 3HC1
To some extent, we extended this method earlier to the aromatic series, obtaining good results with ptolyl, a mcarboxyphenyl, and o-carboxyphenyl disulfide (and its ester).l0 Further assessment of steric and electronic factors which might affect results with aryl disulfides remained of interest, however. The present paper reports extension of the DouglassFarah method to phenyl, p-chlorophenyl, and pnitrophenyl disulfide. The total picture now makes it evident that the method provides a rather general and quite useful synthesis of aromatic thiolsulfonates, as well as of aliphatic ones. For preparation of symmetrical aryl arenethiolsulfonates, we consider the Douglass-Farah procedure to be much the method of choice. Table I summarizes the results. It shows that pure products were obtained in high yields with or without use of solvent, that a thiol can be used (since it is easily OXIDATION
TO
TABLE I THIOLSULFONATE6 U6INQ CHLORINE
ReChloactant, rine, Reautant moles moles (CsHsS)z 0 . 4 5 0.90' ( ~ C H ~ C ~ H B0.32 )I 0.64' pC1CsHSH 0 . 3 0 0.45d (pN0zCeHB)z 0 . 0 5 0.12'1' a After recrystallization
-M.p., "C.Solvent Yield, Prod(ml.) %" uota Lit. None 84 42-44 4H6' None 89 74-76 78.5-79.5c CHzClz (35) 84 134-136 137-13V CHzClz (20) 75 180-182 180-180.5'
from methanol. Crude yields were 2C12( CH~COZH) 2Hz0+ nearly quantitative. (RS)z RSOzSR (+ CHsCOzH) 4HC1. L. Field, J . Am. Chem. Soc., 74, 394 (1952). 2RSH 3c1z (+ CHsC02H) 2Hz0 + RSOzSR (+ CHaCOzH) 6HC1. e G. Bulmer and F. G. M m , J . Chem. Soc., 680 (1945). *This amount of chlorine was necessary to convert the suspension to a solution (RSSR + 2RSC1).
+ + ++
+
+ +
(1) (8) Reported in part at the Southeastern Regional Meeting of the American Chemical Sooiety, Gatlinburg, Tern., Nov. 1-3, 1962. This inveatigation W8U supported by the u. s. Army Medical Researoh and D b velopment Command, Department of the Army, under Research Contraot No. DA-49-193-MD-2030. Abstrsotad from a portion of the Ph.D. Diasertation of T. F. P.lb (b) T. F. Parsons, Ph.D. Diauertation, Vanderbilt University, May 1964. ( 0 ) Paper XII: R. R. CreUh8W and L. Field, J . Ow. Chem., 80, 175 (1965). (d) Du Pont Postgraduate Teaching Asaisb ant, 1962-1963. (2) 1. B. Douelass and B. 8. Farah, J . Ow. Chem., 44,973 (1959). (3) LField. T. C. Owen, R. R. Crenahaw, and A. W. Bryan, J . Am. Cham. aoc., 88, 4414 (1961).
658
NOTES
oxidized to the disulfide), and that a nitro group (unlike 2-pyridinh~n)~ does not deplete electron density on the sulfur atom s a c i e n t l y to be damaging (a slightly lower yield of 75% can be ascribed to excess chlorine used to ensure destruction of all of the difficultly separable nitrodisulfide). It is worth adding that commercial p-nitrophenyl disulfide was purified by recrystallizing it from chloroform-methanol and acetone-methanol (the melting point rose from 140-155' to 179-181'), a method which is simpler than earlier ones.5 Heretofore our only poor results with the DouglassFarah route as applied to ring compounds have been with 2-pyridyl disulfide dihydr~chloride~ and p-carboxyphenyl disulfide.lC We now report failure of the method with 2,4,6-triisopropylphenyldisulfide (1). After attempts to oxidize disulfide 1 at 0 or 30' without solvent, or at 0 ' in methylene chloride with a stoichiometric amount or excess of chlorine, 1 was recovered in good yield. Steric hindrance associated with the four o-isopropyl groups seems the most probable cause of the failure even though &butyl disulfide, which might also be thought to be hindered, can be cleaved to tbutylsulfenyl chloride under mild conditions.6 Efforts then were made to convert the disulfide 1 to 2,4,6-triisopropylphenyl 2,4,6-triisopropylbenzenethiolsulfonate (2) with a peroxide, Hydrogen peroxide in glacial acetic acid at about 40' converted 1 only to 2,4,6-triisopropylphenyl2,4,6-triisopropylbenzenethiolsulfinate (3, 78% yield). Excess hydrogen peroxide in acetic acid at about 70' evidently oxidizes 3 rather quickly to the sulfonic acid, because when the disulfide 1 was submitted to these conditions the only material which could be isolated was the disulfide (1, 49%) and the product was strongly acidic. Interestingly, the thiolsulfinate 3 showed no change after two years at room temperature, a degree of stability which seems unusual for the class' and probably has its origin in steric factors. Attempted oxidation of the thiolsulfinate 3 to the thiolsulfonate 2 with hydrogen peroxide in acetic acid at 80' permitted only recovery of 3 (42%), although there were indicatjons of the presence of a little 2 (and also of the disulfide 1). Schultz, Freyermuth, and Buc recently reported use of molybdenum, vanadium, and tungsten catalysts for oxidation with hydrogen peroxide of sulfides to sulfones.8 Even though either the sulfide or the intermediate sulfoxide generally was soluble in water in this earlier work, one of these catalysts deserved trial with the refractory hindered disulfide 1. No oxidation of 1 occurred, however, with excess hydrogen peroxide and tungsten trioxide in dioxane, even at elevated temperatures. Addition of acetic acid likewise was without effect. However, in the presence also of hydrochloric acid, which is known to catalyze peroxide oxidations,g the thiolsulfonate 2 could be obtained in 83% yield. Both the tungsten trioxide and the mineral acid are necessary. Use of hydrochloric or sulfuric (4) L. Field, A. Ferretti, R. R. Crenshaw, and T. C. Owen, J . Med. Chem., 7 , 39 (1964). ( 5 ) C. C. Price and G. W. Staoy, J . Am. Chem. SOC.,68, 498 (1946). (6) W. A. Sohulze, G. H. Short, and W. W. Crouch, Ind. Eng. Chem.,
49, 916 (1950). (7) Cf. D.Barnard, J . Chem. Soc., 4675 (1957). (8) H.8. Schulta, H. B. Freyermuth, and S. R. Buc, J . O w . Chem., 98, 1140 (1963). (9) Cf. D. Barnard, L. Bateman, and J. I. Cunneen, "Organio Sulfur Compounds," Vol. 1, N. Kharaaah, Ed., Pergamon Press, Inc., New York, N. Y., 1961,p. 229-247 (and references aited therein).
VOL. 30
acid in the absence of tungsten trioxide resulted in recovery of the disulfide 1 (75%), and use of the trioxide without the mineral acid gave impure product from which 40% of the 1 used was recovered. Adjustment of pH in formation of the catalysts was unnecessary, tungsten trioxide itself being satisfactory. The structure of the thiolsulfonate 2 was established by its analysis and infrared spectrum, as well as by its conversion using 2-mercaptoethylamine to 2-aminoethyl 2,4,6-triisopropylphenyl disulfide hydrochloride (65% yield).l b Seemingly, therefore, a powerful but selective oxidant which should prove useful in other instances is available in the combination of hydrogen peroxide with tungsten trioxide in an acidic solution. Experimental10 Preparation of Thiolsulfonates by Chlorino1ysis.-The general procedure, based on that of Douglass and Farah,2 is exemplified by the oxidation of phenyl disulfide to phenyl benzenethiolsulfonate. Any variations are obvious from Table I. Chlorine (64 g., 0.90 mole) waa added, from a chilled container warmed occasionally by the hand, to a stirred solution of 98.6 g. (0.45 mole) of phenyl disulfide in 27.12 g. (0.45 mole) of glacial acetic acid a t about -5' during about 1hr. The mass changed from a white solid to a clear orange-brown liquid. Water (16.2 g., 0.90 mole) then was added slowly within 5 min.; hydrogen chloride evolved in large amounts. The cooling bath waa removed and the mixture was stirred for 1.5 hr., at which time it was pale yellow. The mixture was washed with water by decantation to remove the acetic acid, and the thiolsulfonate remaining then was dried in a rotating-flask evaporator under reduced pressure to a constant weight of 111.7 g. (99%). This material was taken up in hot methanol. Chilling (Dry Iceacetone) gave 94.7 g. (84%) of phenyl benzenethiolsulfonate as white prisms, m.p. 42-44'. The infrared spectrum contained characteristic strong -SOT bands a t 1340 and 1150 cm.-1; the other thiolsulfonates of Table I showed these same bands with little variation and all gave a positive acidic thiolsulfonate test upon treatment with a thiol." p-Tolyl p-toluenethiolsulfonate was prepared earlier by this general procedure, but no details were mentioned.* 2,4,6-Triisopropylphenyl 2,4,6-Triisopropylbenzenethiolsulfinate (3) .-2,4,6-Triisopropylphenyl disulfide (1, m.p. 87-89", 3.00 g.)12 was suspended in 120 ml. of glacial acetic acid and was heated a t 70' until a clear solution resulted. The solution was cooled to room temperature and hydrogen peroxide (1.44 ml. of 30% solution, plus 1.5 ml. of water) was added during 1 hr. The temperature had to be increased to 40' and 50 ml. more of acetic acid had to be added to keep the mixture homogeneous. The solution was stirred a t room temperature for 24 hr., during which time 0.7 g. of solid crystallized and w&s removed by filtration, m.p. 12&131', slight dec. The filtrate was poured into water (200 ml.); chilliig overnight resulted in 1.70 g. more of solid, m.p. 127-131', slight dec. Both solids had identical infrared spectra consistent with the thiolsulbate 3 (78% yield, characteristic -SO- absorption a t 1090 cm.-1) but not the thiolsulfonate 2. Combination of the two solids and recrystallization from methanol gave 3 as pale yellow needles having a constant m.p. 135-136" dec. (m.m.p. 8P115' with 1). Anal. Calcd. for CaoHrsOS2: C, 74.00; H, 9.50; S, 13.17. Found: C, 73.65; H, 9.41; S,13.22. 2,4,6-Triisopropylphenyl2,4,6-Triisopropylbenzenethiolsulfonate (2).-The disulfide 1 (1.00 9.) was dissolved and tungsten trioxide (0.02 g.) was suspended in a mixture of 15 ml. of glacial acetic acid, 15 ml. of dioxane, and 5 ml. of 12 N hydrochloric acid at 65'. Hydrogen peroxide (0.7 ml. of 30% aqueous solution) was added during 1hr. The mixture then was stirred for 24 hr. at 65' and let cool to ca. 25'. A trace of sodium sulfite was (10) Melting points are corrected. Analyses were by Galbraith Microanalytical Laboratories, Knoxville, Tenn. Infrared spectra were obtained using E Perkin-Elmer Model 137 B spectrophotometer with Nujol mulls. (11) Cf.L. Field, T. F. Parsons, and R. R. Crenshaw, J . Ow. Chem., 98, 918 (1964). (12) D. E. Pearson, D. Caine, and L. Field, {bid., 96, 867 (1960).
659
NOTES
FEBRUARY 1965 added to destroy any residual peroxide, and the solution W&B evaporated under reduced pressure to a yellow solid. Recrystallization (methanol) gave 0.88 g. (83%) of the thiolsulfonate 2, m.p. 105-107', which after further recrystallization had constant m.p. 108.5110°. The infrared spectrum contained strong bands (-SOT) at 1150 and 1340 cm.-*. Anal. Calcd. for Cs&aO&: C, 71.63; H, 9.22; S, 12.76. Found: C, 71.59; H, 9.02; S, 12.91.
5.8 6'
.
..'
Preparation and Structure of Dimethyl a-Conidendrin-8-methyl Sulfonate G. M. BARTON' AND J. F. MANVILLE Contribution No. P-88 from the Department of Forestry of Canada, Forest Products Research Branch, Ottawa, Ontario, Canada Received October 9, 1964
On the basis of Holmberg's preparation of a-conidendric acid in which concentrated sulfuric acid was used to open the lactone ring, a similar preparation of dimethyl a-conidendric acid from dimethyl a-conidendrin was attempted. Methylation of the product with diazomethane, however, failed to give the expected methyl dimethyl a-conidendrate and resulted instead in the 8-methyl sulfonate derivative of dimethyl a-conidendrin [2-naphthoic acid, l12,3,4-tetrahydro-3-hydroxymethyl-6,7-dimethoxy-4-( 3,4-dimethoxyphenyl)-8-methyl sulfonate, y-lactone or 3,4,3',4'tetramethoxy-2'-sulfonyl methyl-cyclolignan-olid (9.9') by the proposed lignan nomenclature of Freudenberg and Weingesa],as shown by I. In view of these results Holmberg's reported a-conidendric acid2 was probably the 8-sulfonic acid derivative. Rased on elemental analyses and a molecular weight determination an empirical formula of CzsHzaO~Swas obtained for this compound. A ferric ferricyanide4 color test for phenolic groups was negative. Since the infrared spectrum indicated that the lactone ring was intact and methoxyl analyses showed an additional methoxyl besides the four aromatic ring methoxyls expected, this fifth methoxyl group was apparently the methyl ester of a sulfonic acid group introduced into the compound. Although soluble in chloroform the new derivative was only moderately soluble in methanol, ethanol, and acetone, and WM insoluble in water. Comparison of aromatic ring proton magnetic resonance spectra (p.m.r.) of dimethyl a-conidendrin (11)
R
OCHs I,R =S08CH3. -~
II,R=H
(1) Research O5cer, Vancouver Laboratory, FPRB. (2) B. Holmberg, Suenek. Kem. Tidakr., 83, 56 (1920). (3) K. Freudenberg and K. Weinges, Tetrahedron, lS, 115 (1961). (4) G. M. Barton, R. 8. Evans, and J. A. F. Cardner, Nature, 170, 249 (1952).
6 in P P M Figure 1.-Aromatic proton magnetic resonance spectra of dimethyl a-conidendrin-&methyl sulfonate (I) and dimethyl a-conidendrin (11)in deuteriochloroform.
and this sulfur-containing derivative (Figure 1) together with stereochemical considerations strongly indicated that the sulfonic acid methyl ester group should be assigned to carbon 8. Substitution of any one 'of the protons 2', 5', or 6' by such a strong electron-withdrawing group as a sulfonic acid methyl ester would result in the displacement of the remaining two protons to higher &values. Since only one proton was so displaced, substitution must have involved protons 5 or 8. A choice between proton 8 and proton 5 was made on the basis of stereochemical models of the two possible structures (Courtauld atomic models). These models clearly indicated that substitution of a bulky sulfonic acid methyl ester group on carbon 5 would be restricted by the large aryl group on carbon 4. Since the infrared spectrum of the monosulfonic acid intermediate showed it to be a mixture of the monosulfonic acids of both dimethyl and a-conidendric acids as well as their lactones, no rigorous purification of the intermediate was attempted. In any case, reaction of the intermediate with diazomethane produced only one crystalline end product I which was easily isolated. Undoubtedly, the 15-fold excess of diazomethane used in the reaction was involved in remethylation of free phenolic groups as well as combination with residual traces of sulfuric acids. Experimental Carbon, hydrogen, sulfur, and methoxyl analyses were performed by the Clark Microanalytical Laboratories, Urbana, Ill. Molecular weights were determined on the Mechrolab vapor pressure osmometer, Model 301A, using acetone m a solvent. P.m.r. spectra were recorded on a Varian, Model A-60, spectrometer in deuteriochloroform with tetramethylsilane aa an internal standard. Infrared spectra were run in KBr d i s h on a Baird-Atomic infrared recording spectrophotometer, Model KM-1. Preparation of the Monosulfonic Acid Intermediate.-Dimethyl a-conidendrii (4 g.) prepared by the method of Holmberg6 wm thoroughly mixed with concentrated sulfuric acid (12.6 ml.) and left a t room temperature for 6 br. Water (45 ml.) wm then slowly added with stirring. The resultant precipitation was allowed t o continue overnight a t 4'. The excess acid solution waa decanted and the product was recrystallized from water (35 ml.) by gentle warming. The slightly colored crystals (4.7 g.) melted with decomposition at 220-230'. Dimethyl (~-Conidendrin-8-methyl Sulfonate.-Dimethyl (Yconidendrin-8-sulfonic acid (2 g.) in methanol (40 ml.) wm added (5) B. Holmberg, Chsm. Ber., S4, 2389 (1921).
NOTES
660
to a 15-fold excess of diazomethane in ether (110 m l . ) . a The slightly yellow solution was allowed to crystallize at 4' for several hours, filtered, and washed with cold methanol; yield 1.2 g., m.p. 245-247'. Recrystallization from boiling methanol afforded an analytical sample melting sharply at 250": ~2: 2980 (m), 1780 (vs), 1520 ( s ) , 1475 (s), 1365 (m),1270 (e), 1220 (m), 1050 (m), 990 (s), 870 (m), 770 (s), and 660 cm.+ (m). Anal. Calcd. for CaHleOpS: C, 57.9; H, 5.49; S, 6.70; OCH3, 32.4; mol. wt., 478. Found: C, 57.62, 57.45; H, 5.48, 5.52; S,6.72; OC&, 31.41; mol. wt., 476.
VOL.30 a
O
-
0
f PhPCIz
N?BF,-
CuBr
Acknowledgment.-The authors wish to thank Crown Zellerbach, Chemical Products Division, Camas, Washington, for their generous sample of a-conidendrin. (6) A. I. Vogel, "A Text Book of Practical Organic Chemistry," Longmans, Green and Co., London, 1948, p. 844.
Cyclodehydrohalogenation of Diarylphosphinous Chlorides' JACKB. LEVY,G. 0. DOG,
AND
LEOND. FREEDMAN
v
J
VI
benzophosphole 5-oxide (VIII) from o-biphenylylphenylphosphinic acid.l Since the structure suggested for IV is similar to that given for VII, we have investigated the possibility that cyclodehydrohalogenation occurs before the aluminum reduction in the prepara-
Department of Chemistry, North Carolina State of the University of Nmth Carolina at Raleigh, Raleigh, North Carolina Received September 83, 1964
PCl,
Cyclodehydrohalogenation of arylphosphonous dichlorides has previously been shown to be a useful method for the preparation of certain heterocyclic organophosphorus compounds.2 By means of this method we have prepared derivatives of dibenzophosphole (I), phenoxaphosphine (11), and 5,lOdihydrodibenz [b,e]phosphorin (111).
%NO2 170-180-
.
3
VI1
VI11
H I
I1
I11
We have now found that suitably substituted diarylphosphinous chlorides3 also undergo cyclodehydrohalogenation. Thus 10-phenylphenoxaphosphine4(VI) was prepared in 26% yield by the following series of reactions. Neither IV nor V was isolated from these reactions. (o-Phenoxypheny1)phenylphosphinouschloride (V) presumedly underwent cyclodehydrohalogenation during the preparation and yielded the desired cyclic compound. The intermediate formed from the reaction of a diazonium tetrafluoroborate with an arylphosphonous dichloride has not been isolated, but structures similar to IV have been proposed.a-6*6 Campbell and Way have suggested that intermediate VI1 was formed in their preparation of 5-phenyldi(1) Supported by Research Grant GM-09479 from the National Institutes of Heklth, U. 9. Public Health Service. (2) G. 0. Doak, L. D. Freedman, and J. B. Levy, J . Org. Chem., 29, 2382 (1964). (3) L. D. Quin and R. E. Montgomery, ibid., 28, 3315 (1963). (4) 10-Phenylphenoxaphosphine and ita oxide have been prepared by F. G. Mann and I. T. Millar [ J . Chem. Soc., 3746 (1953)l. (5) L. D. Quin and J. S. Humphrey, Jr., J . Am. Chem. SOC.,88, 4124 (1961). (6) P. C. Crofta, Qucrrt. Rev. (London), 19, 341 (1958).
tion of 10-phenylphenoxaphosphine. Accordingly, o-phenoxybenzenediazonium tetrafluoroborate was allowed to react with phenylphosphonous dichloride, and the resulting mixture waa hydrolyzed. The uncyclized compound, (o-phenoxypheny1)phenylphosphinic acid, waa isolated in 52% yield. This result strongly suggests that cyclodehydrohalogenationoccurred after the aluminum reduction in our preparation of 10-phenylphenoxaphosphine. Some cyclization did occur, however, when the reaction mixture was distilled rather than hydrolyzed. Thus, when the highest boiling fraction of the distillate (b.p. 175-177' at about 5 p ) was treated with an alkaline hydrogen peroxide solution (to hydrolyze the mixture and to oxidize any trivalent phosphorus compound which may have been formed from thermal decomposition of a pentavalent phosphorus compound) , the resulting mixture gave a 3% yield of 10-phenylphenoxaphosphine lo-oxide and a 32% yield of (o-phenoxypheny1)phenylphosphinic acid. It is of interest to note that we have been unable to cyclize this acid by the method used by Campbell and Way for the cyclization of o-biphenylylphenylphosphinicacid. (7) I. G. M. Campbell and J. K. Way, J . Chem. Soc., 2133 (1961).
FEBRUARY 1965
NOTES
The reaction of p-tolylphosphonous dichloride with o-phenoxybenzenediazonium tetrafluoroborate, followed by reduction with powdered aluminum, gave a mixture from which we were able to isolate a 19% yield of p-tolylphenoxaphosphe and, after oxidation and hydrolysis, an 11% yield of (o-phenoxypheny1)-ptolylphosphinic acid. We have prepared two additional cyclic tertiary phosphines from o-biphenyldiazonium tetrafluoroborate. 5-Phenyldibenzophosphole718 was prepared in 21% yield, and 5-p-tolyldibenzophosphole was obtained in 7y0 yield. In the latter reaction a small amount of o-biphenylyl-p-tolylphosphinic acid was also isolated. Oxidation of all four cyclic phosphines with hydrogen peroxide yielded the corresponding phosphine oxides, 10-phenylphenoxaphosphine l o - o ~ i d e ,10-p-tolylphenoxaphosphine ~ 10-oxide, 5-phenyldibenzophosphole 5-oxide,?J and 5-p-tolyldibenzophosphole 5-oxide.
5-Phenyldibenzophosphole.-The reaction of o-biphenyldiazonium tetrafluoroborate (26.8 g., 0.1 mole) with phenylphosphonous dichloride (0.1 mole), followed by reduction with powdered aluminum, was carried out as described for the preparation of 10-phenylphenoxaphosphhe. The usual isolation procedure led to 8.1 g. of 5-phenyldibenzophosphole(b.p. 180" a t about 5 p). The yield after recrystallization from aqueous alcohol was 5.5 g. (21%), m.p. 92-94' (lit. m.p. 90-92',' 93-9408). The infrared absorption spectrum was in agreement with that previously reported.8 Anal. Calcd. for C18HlSP: P, 11.90. Found: P, 12.01. 5-Phenyldibenzophosphole 5-Oxide.-5-Phenyldibenzophosphole was oxidized in an alkaline solution of aqueous ethanol and hydrogen peroxide. After two recrystallizations from 6 N hydrochloric acid which contained about 5% ethanol, the melting point of 5-phenyldibenzophosphole 5-oxide was 165-167' (lit. m.p. 167-168',' 163.5-166.5' 8); 94% qeld. The infrared absorption spectrum was in agreement with that previously reported.' Anal. Calcd. for ClsHlaOP: P, 11.21. Found: P, 11.07. p-Tolylphosphonous Dichloride.-This compound (b.p. 90" a t about 5 p) was prepared by the method of Quin and Humphrey6 except that powdered aluminum wm used as the reducing agent rather than magnesium. The yield was 28.5%. 10-p-Tolylphenoxaphosphme and (o-Phenoxypheny1)-p-tolylphosphinic Acid.-The reaction of o-phenoxybenzenediazonium tetrafluoroborate (28.4 g., 0.1 mole) with p-tolylphosphonous dichloride and reduction with powdered aluminum were carried out as described for the preparation of 10-phenylphenoxaphosphine. A liquid (10.3 g., b.p. 180" a t about 5 p) was collected and dissolved in ethanol. Deaerated water was added to the solution to precipitate 5.5 g. (19% yield) of 10-p-tolylphenoxaphosphine, m.p. 53-57". The melting point of the phosphine after recrystallization from aqueous ethanol was 59-62'. Anal. Calcd. for C19H150P: P, 10.67. Found: P, 10.41. The original mother liquor from 10-p-tolylphenoxaphosphine was oxidized with hydrogen peroxide. The solution WEIS made alkaline, stirred with Norit, and filtered. Upon acidification, (o-phenoxypheny1)-p-tolylphosphinic acid (3.6 g., 11% yield) precipitated. After two recrystallizations from aqueous ethanol, the melting point was 176177'. Anal. Calcd. for Cl&o& P, 9.55; neut. equiv., 324.3. Found: P, 9.36; neut. equiv., 334.0. lo-p-Tolylphenoxaphosphine 10-0xide.-lO-p-Tolylphenoxaphosphine was oxidized in an alkaline solution of aqueous ethanol and hydrogen peroxide. After recrystallization from 6 N hydrochloric acid (which contained about 5% ethanol) the melting point of 10-p-tolylphenoxaphosphine10-oxide was 195-197'; 95% yield. An intense absorption a t 1200 cm.-l was assigned to the phosphoryl (P+O) bond. Anal. Calcd. for ClsH15O2P: P, 10.11. Found: P, 9.97. 5-p-Tolyldibenzophospholeand o-Biphenylyl-p-tolylphosphhic Acid.-The reaction of o-biphenyldiazonium tetrafluoroborate (26.8 g., 0.1 mole) with p-tolylphosphonous dichloride and reduction with powdered aluminum were carried out as described for the preparation of 10-phenylphenoxaphosphine. A liquid (3.2 g., b.p. 180' at about 5 p) was collected and dissolved in ethanol. Deaerated water was added to the solution to precipitate 1.9 g. (7% yield) of 5-p-tolyldibenzophosphole. The melting point of the phosphine after recrystallization from aqueous ethanol was 11C-113'. Anal. Calcd. for CI&P: P, 11.29. Found: P, 11.18. The original mother liquor from 5-p-tolyldibenzophos~hol~ was oxidized with hydrogen peroxide. The solution was made alkaline, stirred with Norit, and filtered. Upon acidification, o-biphenylyl-p-tolylphosphinicacid (0.2 g.) precipitated. The melting point after recrystallization from aqueous ethanol was 179-180'. Anal. Calcd. for C ~ Q H ~ O ZP, P : 10.05; neut. equiv., 308.3. Found: P, 9.91; neut. equiv., 311.6. 5-p-Tolyldibenzophosphole5-Oxide.-5-p-Tolyldibenzoph0~phole was oxidized in an alkaline solution of aqueous ethanol and hydrogen peroxide. The melting point of 5-p-tolyldibenzophosphole 5-0xide, after recrystallization from aqueous ethanol, was 190-192'; 94% yield. An intense absorption at 1200 cm.-1 was assigned to the phosphoryl (P-0) bond. Anal. Calcd. for C&&OP: P, 10.67. Found: P, 10.66.
ExperimentalQ-ll
10-Phenylphenoxaphosphine .-o-Phenoxybenzenediazonium tetrafluoroborate (56.8 g., 0.2 mole), 200 ml. of anhydrous ethyl acetate, 35.8 g. of phenylphosphonous dichloride (0.2 mole, Eastman P 6544), and 4 g. of cuprous bromide were allowed to react according to the procedure described for the preparation of 5-chlorodibenzophosphole.~After 45 min. of stirring the mixture was heated to initiate the vigorous reaction. Stirring was continued for 1 hr., and 3.6 g. of powdered aluminum (Mallinckrodt) was added to the flask. The mixture was refluxed for 1hr. and then stirred overnight. The usual isolation procedure was carried out2; the product (19.0 g., b.p. 150' at about 5 p) was collected as a liquid which later solidified. The yield after recrystallization from aqueous alcohol was 14.3 g. (26%), m.p. 97.5-98.0' (lit.4 m.p. 94.5-95.0"). Anal. Calcd. for Cl8H1@P: P, 11.21. Found: P, 11.38. 10-Phenylphenoxaphosphine 10-Oxide.-10-Phenylphenoxaphosphine was oxidized in an alkaline solution of aqueous ethanol and hydrogen peroxide. The precipitate which appeared on evaporation of the ethanol was removed by filtration, washed with water, and recrystallized from 6 N hydrochloric acid which contained about 575 ethanol; m.p. 176-179" (1it.h m.p. 173174"). The yield was 94%. An intense absorption a t 1205 cm.-l was assigned to the phosphoryl (P+O) bond. Anal. Calcd. for C18Hl&hP: P, 10.60. Found: P, 10.36. (0-Phenoxypheny1)phenylphosphinic Acid.--Phenoxybenzenediazonium tetrafluoroborate (14.2 g., 0.05 mole), 60 ml. of anhydrous ethyl acetate, and 8.5 g. of phenylphosphonous dichloride (0.05 mole) were allowed to react as in the preparation of 10-phenylphenoxaphosphine. After the diazonium salt was decomposed, stirring was continued for 45 min. The mixture was refluxed for 1 hr. and then filtered through a sintered-glass funnel. Water (50 ml.) was added to the filtrate, and the ethyl acetate was evaporated. After the addition of 150 ml. of 10% sodium hydroxide, the mixture was stirred for 1 hr., acidified with concentrated hydrochloric acid, and extracted with ether. When the ether layer was shaken with 20% sodium hydroxide, a sodium salt precipitated, which was removed by filtration, washed with ether, and then dissolved in water. The material which precipitated upon acidification of the aqueous solution was removed by filtration and recrystallized from aqueous ethanol; m.p. 164166'. The yield was 8.2 g. (52%). Anal. Calcd. for C18H1608P: P, 9.98; neut. equiv., 310.3. Found: P, 9.96; neut. equiv., 311.5. (8) G. Wittig and G. Geisaler, Ann., 680, 44 (1863).
(9) The infrared spectra were taken on a Perkin-Elmer Model 521 spectrophotometer. The samples were run in potassium bromide pellets. (10) Melting points were determined on a Fisher-Johns apparatus with a thermometer calibrated against U.S.P. reference standards aa previously described; cf. G. 0. Doak and L. D. Freedman, J . Am. Chem. rSoc., 18, 5658 (1951). (11) Analyses were performed by Galbraith Laboratories, Inc., Knoxville, Tenn.
661
NOTES
662
Heterocyclic Amines. I. 3-Thienyl Tertiary Amines’ JEREMIAEIB. SULLIVAN~ AND WALTERC. MCCARTHY CoUege of Pharmacy, University of Washington, Seattle, Wmhington 98106
Received September 18, is64
Acyl derivatives of aminothiophenes can be readily isolated, as can substituted aminothiophenes containing a stabilizing group, such as acyl, carboxyl, or sulfonyl, attached to the ring. Simple aminothiophenes, however, as the free base, have long been noted for their in~tability.~ Although 2-amin~thiophene~ could be isolated by vacuum distillation, the free base polymerized rapidly to gum. 3-Aminothiophene has never been isolated although its presence could be demonstrated by the preparation of acyl derivatives616 and by observation of its n.m.r. spectnun.7 The instability of amino, hydroxy, and mercapto derivatives of thiophene and furan has frequently been ascribed to the existence of tautomeric forms.8a-d If this be part of the reason for the instability of these compounds, the tertiary amines should be more stable than the primary or secondary amines, much as alkoxy and acyloxy derivatives of furan and thiophene are more stable than the parent hydroxy compounds. Because of the potential importance, if sufficiently stable, of thienyl tertiary amine derivatives as medicinal agents and as dyestuffs, a suitable preparative route has been sought. Since this article was submitted for publication the synthesis of several closely related 3thienyl tertiary amines by a different method has been reported.Ee 3-Thenoyl chloride (I) was converted through the azide 11, by rearrangement in methanol, to methyl N-(3-thienyl)carbamate (111). This urethan was alkylated in xylene with sodium hydride dispersion and
I
I11 IN.”.
Me1
Me I
V
rv
(1) (a) This work hse been supported in part by U. 8. Public Health Service Research Grant No. GM 10264. (b) Presented before the Division of Medicinal Chemistry at the 144th National Meeting of the American Chemical Society, Loo Angeles, Calif., April 1963. (2) Fellow of the American Foundation for Pharmaceutical Education, 1959-1960. (3) H. D. Hartough, “Thiophene and Ita Derivatives,” Interecience Publishers, Inc., New York, N. Y., 1952, pp. 228-235. (4) W. Steinkopf, Ann., 40% 17 (1914). (5) W. Steinkopf and T. H6pner, ibid., 601, 174 (1933). (6) E. Campaigne and P. A. Monroe, J . Am. Cham. Soc., 76, 2447 (1954). (7) R. A. Hoffman and 8. Gronowitz, Arkiu Kemi, 16, 515 (1961). (8) (a) Ref. 3, pp. 229, 287, 428; (b) A. P. Dunlop and F. N. Peters, “The Furans,” Reinhold Publishing Corp., New York, N. Y.. 1953, p. 170; (0) 8. Crronowits, “Advanaes in Heterocyclic Chemistry,” Vol. 1, Academic Pr-6 Inc., New York, N. Y., 1963, pp. 82-87; (d) A. R. Eatritzky and J. M. Lagowski, ibid. Vol. 2 1963, pp. 1-22; (e) F. A. Buiter, J. 3. Sperna Weiland, and H. Wynberg, Rsc. @or. chim., SS, 1160 (1964).
VOL. 30 methyl iodide to give methyl N-methyl-N-(3-thieny1)carbamate (IV). The latter compound was reduced with lithium aluminum hydride in tetrahydrofuran to 3-dimethylaminothiophene(V) , which was isolated by preparative gas chromatography on a QF-1 (fluorocarbon silicone) column. This tertiary amine is stable when stored under vacuum or under nitrogen atmosphere for several weeks, or for several months when similarly stored in the freezing compartment of the refrigerator, but the compound darkens rapidly and polymerizes to a tar within hours in contact with air. Attempts to isolate salts of this amine derived from proton acids have not been successful, but the quaternary salts derived from reactions with alkyl halides are white crystalline compounds, hygroscopic, but stable in the presence of air and moisture. Methyl N-(3-thienyl)carbamate was also alkylated with benzyl chloride and sodium hydride, and the resulting urethan waa reduced with lithium aluminum hydride to prepare another tertiary amine, N-methylN-benzy1-3-aminothiophenelwhich was isolated as the methiodide. Attempted ethylation of methyl N-(3-thienyl)carbamate by the usual method gave, instead of the expected derivative, a compound presumed to be a self-condensation product of the urethan, tris(3thieny1)isocyanurate (VI). Modification of the alkyla-
Q1p N
N
VI tion procedure by addition of the ethyl iodide at the start of the reaction then gave the desired ethylation product. Many 3-substituted thiophenes9 show a characteristic infrared absorption band in the range 760-780 cm.-’. The 3-thienyl urethans reported herein show absorption within this range, but 3dimethylaminothiophene and the 3-thienylammonium salts show the correspondingabsorption peaks outside this range.lo Methyl N-(3-thienyl)carbamate shows absorption peaks at 760 and 770 cm.-’. Methyl N-methyl-N-(3thienyl)carbamate, methy1 N-e thyl-N- (3-thieny1)carbamate, and ethyl N-ethyl-N-(3-thienyl)carbamateall show an absorption peak at 770 cm.-’. 3-Thienyltrimethylammonium iodide and 3-thienyldimethylethylammonium iodide both show an absorption peak at 788 cm.-l and a shoulder at 794 cm.-’, while 3thienyldimethylbenzylammonium iodide shows peaks at 791 and 797 cm.-’. 3-Dimethylaminothiophene shows an absorption peak at 748 cm.-l. 2-Thenoyl chloride was also converted to methyl N(2-thienyl)carbamate1 which was alkylated to methyl N-methyl-N-(2-thienyl)carbamateand methyl N-benzyl-N-(2-thienyl)carbamate. Reduction of the latter two compounds failed to yield any pure amine. (9) S. Gronowita, ref. Bc, p. 13. (10) Several of these infrared spectra are reproduced in the Ph.D. Thesis of J. B. Sullivan, University of Wsehington, 1963.
NOTES
FEBRUARY1965
Anal. Calcd. for C&pNS:
Experimental 3-Thenyl Bromide.-The preparation of this intermediate in the synthesis of 3-thenoyl chloride was modified by replacement of the benzoyl peroxide catalyst" with an equal weight of azobisiisobutyronitrile. There was no change in yield, but a cleaner reaction mixture was produced. Methyl N-(J-Thienyl)carbamate @I).-A solution of 25 g. (0.38 mole) of sodium azide in 94 ml. of water was added a t 0' to 13 g. (0.089 mole) of 3-thenoyl chloride11 in 100 ml. of acetone. The mixture was shaken for 30 min. at 0'. The solvent was removed in a rotary evaporator and the oily residue was extracted with ether. After drying over anhydrous magnwium sulfate and removal of the solvent, the crude, oily bthenoyl azide was dissolved in 100 ml. of methanol. The mixture was refluxed for 8 hr., the excess methanol was removed in a rotary evaporator, and the oily product was crystallized from ligroin and recrystallized from water, m.p. 77-78", yield, 12 g. (86%). The n.m.r. spectrum12 in deuterated chloroform was 6 = 3.74 (s, CHa, 3H), 6.94 (m, ring H, lH), and 7.20 (m, 2 ring-H and NH, 3H) p.p.m. Ethyl N-(J-Thienyl)carbamate.-This wm prepared sirnilarb to the above, except that the azide was rearranged in anhydrous ethanol. The product was recrystallized from water, m.p. 71-72'. Identity of thk product was confirmed by its n.m.r. spectrum in CCL: 6 = 1.25 (t, CHa, 3H), 4.14 (q, CH2, 2H), 7.05 (m, ring H, 3H), and 7.85 (broad band, NH, 1H) p.p.m. Methyl N-Methyl-N-(3-thienyl)carbamate (IV) .-A solution of 3.2 g. (0.02 mole) of I11 in 100 ml. of anhydrous xylene was added to a stirred suspension of 6.0 g. (0.125 mole) of 50% sodium hydride dispersion in 150 d.of anhydrous xylene, refluxed for 8 hr. under nitrogen, and cooled; 90 g. (0.6 mole) of methyl iodide was added. After refluxing an additional 8 hr. ,the mixture was filtered, the residue was washed with benzene, and the combined filtrate and washings were evaporated. The oily residue was vacuum distilled, and 3.0 g. (85%) of crude product was collected from 126 to 134' at 10 mm. Although contaminated by a small amount of mineral oil from the sodium hydride dispersion, it was suitable for subsequent reduction. The analytical sample was extracted with cold hexane, and again vacuum distilled, b.p. 136-137' (12 mm.). The n.m.r. spectrum in CC4 was 6 = 3.25 (9, N-CHa, 3H), 3.67 (9, O-CH3, 3H), 6.89 (t, probably 2-H of ring, 1H) (J = 2.3 c.P.s.), and 7.12 (d, probably 4,5-H of ring, 2H) (J = 2.3 0.p.s.) p.p.m.I3 Anal. Calcd. for C~H~NOLS!: C. 49.10: H. 5.30: N. 8.18: 0, 18.69; S, 18.73. F o - u d k C,'49.83;' H,'5.475 N; 8.22; 0, 18.56; S, 18.33. 3-Dimethylaminothiophene (V).-A solution of 3.0 g. (0.0175 mole) of IV in 25 ml. of anhydrous tetrahydrofuran was added dropwise, with stirring, under nitrogen to 2.0 g. (0.0525 mole) of lithium aluminum hydride in 50 ml. of anhydrous tetrahydrofuran. After refluxing for 24 hr., the reduction complex was decomposed with water. This mixture was filtered under nitrogen, the residue was washed with ether, and the solvent was partly removed in a rotary evaporator. The amine was purified by gas chromatography in a Beckman GC-2 instrument a t 100" through a 10-ft. QF-1 (fluorocarbon silicone) column with helium as the carrier gas. The n.m.r. spectrum in CCL was 6 = 2.79 (s, 2CHa, 6H), 5.86 (q, 2-H of ring, l H ) , 6.70 (q, 4-H of ring, lH), and 7.14 (q, 5-Hof ring, 1H) p.p.m.,Ju = 1.7, JU = 3.4, and J46 = 5.4c.p.s. (all f 0.2 c.P.s.). (11) E. E. Campaigne and W. M. IaSuer, J. Am. Chem. Soc., 10, 1555 (1948). (12) We wish to acknowledge our indebtedness to MI. B. J. Nist of the
Department of Chemintry, University of Washington, for the determination of Bll n.m.r. spectra reported herein. Theee spectra were meaaured at 60 Mc., and the chemical shifts are reported as &values (p.p.m.) from tetramethylnilane (internal standard). The multiplicity is shown b y e singlet, d doublet, t triplet, q quadruplet, and m incompletely resolved multiplet. Assignments are shown i n parentheses. Coupling constants are reported in 0.p.s. (13) This is an unexpectedly simple spectrum for the ring protons of a monosubstituted thiophene. It appears that the protons a t positions 4 and 6 muat experience a similar magnetic environment, and t h a t the coupling C O M t a n t s , Ju and Ju, muat be almost identical. Hoffman and Gronowits, ref. 7, report a similar n.m.r. spectrum for 34odothiophene in cyclohexane. (14) Elemental analyaea were by A. Bernhardt, Mfllheim (Ruhr), Germany. (15) T. Curtius and H. Thymen, J. prokt. Chcm., 66,5 (1902).
-
-
-
-
-
663 C, 56.65; H, 7.13; N, 11.01
S,25.21. Found: C, 56.79, 56.88; H, 7.26, 7.17; N, 10.90, 11.18; S, 25.14, 24.98. The amine was further characterized by conversion to ita methiodide, m.p. 211-212' after recrystallization from ethanol. Anal. Calcd. for C7H1JNS: C, 31.23; H, 4.46; I, 47.21; N, 5.20; S, 11.90. Found: C, 31.82, 31.70; H, 4.62, 4.71; I, 47.02,46.92; N, 4.92,5.05; S, 11.70, 11.58. N-Methyl-N-benzyl-3-aminothiophene Methiodide.-A solution of 5 g. (0.038 mole) of I11 in 150 ml. of anhydrous xylene was added to a suspension of 4 g. (0.083 mole) of 50% sodium hydride dispersion in 50 ml. of anhydrous xylene, refluxed for 12 hr., cooled, and 22 g. (0.17 mole) of benzyl chloride was added. After refluxing an additional 8 hr., the mixture was filtered, the residue was washed with benzene, and the combined filtrate and washings were evaporated. The product was isolated by vacuum distillation, with 3.32 g. (45%) collected at 140-160' at 0.3 mm. A solution of 2.5 g. (0.011 mole) of this benzylated urethan in 50 ml. of anhydrous tetrahydrofuran was added d r o p wise, with stirring, under nitrogen, to 2 g. (0.053 mole) of lithium aluminum hydride in 50 ml. of anhydrous tetrahydrofuran. After refluxing 24 hr. and hydrolyzing the reduction complex with water, the mixture was filtered and the residue was washed with ether. The combined filtrate and washings were dried with anhydrous magnesium sulfate and evaporated, an equal volume of benzene was added to the oily residue, and methyl iodide w t ~ then added. The solid product was recrystallized from methanol, m.p. 152.5-153.5'. Anal. Calcd. for ClaHiJNS: C, 45.23; H, 4.67; I , 36.76; N, 4.06: S, 9.29. Found: C, 45.27; H, 4.90; I, 36.81; N, 3.96; S, 9.26. Tris(3-thieny1)isocyanurate(VI) .-A solution of 10 g. (0.064 mole) of I11 in 150 ml. of anhydrous xylene was added to a suspension of 6 g. (0.125 mole) 0500 O gave a product identified as N-methallylpropionuntil the reflux temperature increased t o the level desired. The amide (11). This identification was made by analysis, combined overhead and residue mixture was distilled, and the products were suitably identified. The reaction temperatures infrared spectroscopy, and by hydrolysis to methallylreported are approximately the maximum attained in the tube, amine. The N-methallylpropionarnide was further this maximum occurring about 1 in. below the top of the heated characterized by conversion to 2-ethyl-5,5-dimethyl-2zone. oxazoline (111). N-Acylated allylarnines were also Pyrolysis of 2-Ethyl-4,4-dimethyl-2-oxazoline. N-Methallylpropionamide (11). Method A.-From 344.7 g. of the oxazoline obtained by pyrolysis of 2-phenyl-4,4-dimethyl-2(neut. equiv.8 130) passing through the reactor in 6.4 hr. a t 540' oxazoline and 2,4-dimethyl-2-oxazoline. A temperathere was obtained 339 g. of material with neut. equiv. 161. ture of 500-600' is required. Yields of 44-90% were Then 334 g. of this material was again passed through the reactor obtained in 6.2 hr. a t 560' to yield 323 g. of material with neut. equiv. 263. Distillation yielded 146 g. of recovered oxazoline I and 135 N-C (CH3)2 0 CH3 g. of amide 11, b.p. 135-142' (20 mm.), for an oxazoline converII It I sion of 56% and an amide yield of 71%. CH3CHzC ,0,hH2 CH3CH2C -NHCHzC =CH2 Anal. Calcd. for N-methallylpropionamide, C~HUNO: N, 11.01; iodineno., 199.6. Found: N, 11.20; iodine no., 193.8, I I1 193.7. N-CHz Method B.--An 85% yield of amide a t a 78% conversion of II I oxazoline was obtained at a reactor temperature of 565'. The CHsCH2C\O/C(CH3)2 reflux temperature increased from 133 to 180' in 6.8 hr.
-+
R-&-oR'~
R-~--T;S-R'~
(1)
.
I11
This rearrangement of oxazolines possibly proceeds via the N-acylaziridine as follows. CH3
I
N-C-CHs
''\&H~
1 1
b
H CH2
61
f$,c-cH3 R-C-N,
__t
LHz
(1) R. Royer and D. G. Neilson, Chem. Reu., 61, 190 (1961).
(2) 9. Gabriel and R. Stelzner, Ber., 28, 2929 (1895); A. A. Goldberg and W. Kelley, J. Chem. Soc., 1919 (1948); F. Winternits, M. Mousseron, and R. Dennilauler, B d l . SOC. chim. France, 382 (1956). (3) P. E. Fanta, U. S. Patent 2,766,232(1956); J . Org. Chem., 28, 72 (1958). (4) P. B. Talukdar and P. E. Fanta, ibid., 24,526 (1959); H.W. Heine, M. E. Felter, and E. M. Nicholaon, J . A m . Chem. Soc., 81, 2202 (1959); D. V. Kashelikar and P. E. Fanta, ibid., 82, 4930 (1960); H. W. Heine, Angew. Chem., Intern. Ed. Engl., 1, 528 (1962); P. E. Fanta, L. J. Pandya, W. R. Groakopf, and H. Juang Su, J. 01.9. Chem., 28, 413 (1963). (5) Boiling and melting points are uncorrected. Melting points were determined with a Fiaher-Johns apparatus. (6) H. L. Wehrmeister, J. Org. Chem., 28, 2587 (1963). (7) R. Oda, M. Okano, S. Tokeura, and F. Misumi, Bull. Chem. SOC. Japan, 86, 1219 (1962); Chem. Abstr., 57, 12454 (1962). (8) The neutralization equivalents of the oxazolines were determined by potentiometric titration with perchloric acid using acetic acid as a solvent.
FEBRUARY 1965
NOTES
Methally1amine.-Hydrolysis of the amide obtained by both methods A and B yielded methallylamine which was identified by the preparation of a picrate, m.p. 204206' (lit. m.p. 202.5204.5°,a 202-206°s), and aphenylthiourea derivative, m.p. 76-77' (1it.a m.p. 78-79"), and by benzoylation to N-methallylbenzamide, m.p. 66-67" (lit.8 m.p. 69.5-70.5"). 2-Ethyl-5,5-dimethyl-2-oxazolhe @I) .-An 18.5-g. portion of amide product obtained by method B, b.p. 130-135" (20 mm.), was added to 25 ml. of concentrated HaO4 with st%g and cooling (temperature kept a t 30-35") in 15 min. The mixture was then poured onto 200 g. of crushed ice. Sodium hydroxide (40 g.) and water (50 ml.) were added (ice bath cooling), and the mixture was extracted with three 50-d. portions of ether. Dktillation of the dried extract gave 9.0 g. of 2-ethyl-5,5-dimethyl-2oxazoline, b.p. 13S144' (lit.* b.p. 141"). Anal. Calcd. for C7H13NO: N, 11.01; neut. equiv., 127.2. Found: N, 11.05; neut. equiv., 129.0. A picrate of this material, m.p. 144-146' (lit.* m.p. 147-149'), was shown to differ from the picrate, m.p. 151-154", of Z-ethyl4,4-dmethyl-2-oxazoline by mixture melting point (123-128'). Pyrolysis of 2-Phenyl4,4-dimethyl-2-oxazoline. N-Methallylbenzamide.-N-Methallylbenzamide, b.p. 113-117" (0.05 mm.), was obtained in a 21% yield (based on oxazoline charged) by pyrolysis of 2-phenyl-4,Pdimethyl-2-0xazoline at 597' using method A. Pyrolysis of the oxazoline at 559' by method B but at reduced pressure (50 mm.) gave a 79% yield of amide (28% oxazoline conversion), b.p. 136-142' (1 mm.), in 26.5 hr. The reflux temperature rose from 152 t o 159.5". Analysis of the product by gas chromatography indicated a purity of 93.5%. Anal. Calcd. for N-methallylbemamide, CllHlsNO: N, 8.00; iodine no., 144.8. Found: N, 8.09; iodine no., 143. Similarly prepared material, b.p. 141-143' (1 mm.), WM recrystallized from petroleum ether to m.p. 68-69' (lit.8 m.p. 69.5-70.5'). A mixture melting point with an authentic sample was not depressed. Pyrolysis of 2,4-Dimethyl-2-oxazoliae. N-Allylacetamide.N-Allylacetamide, b.p. 102.105' (10 mm.), lit.10 b.p. 113-116' (15 mm.), was obtained in a 44% yield (46% oxazoline conversion by pyrolysis of 2,4dimethyl-2-oxazoline at 586" using method A. Anat. Calcd. for N-allylacetamide, C6H8NO: N, 14.13; iodine no., 256. Found: N, 14.80; iodine no., 236. The structure of this impure product was also supported by an infrared spectrum [bands a t 3.06 (NH) and at 6.08 p ( C d ) ] and by a n.m.r. spectrum [ 6 1.97 (CHaCO), 3.75 (=GCHrN-), 5 (=CH2), 5.7 (=CH-), and 7.5 (-"-)I. A slight amount of impurity was also indicated by the presence of small extraneous peaks throughout the n.m.r. spectrum."
would afford such a route. As an illustration, a-L-aspartyl-L-leucine has been prepared from p-nitrocarbobenzoxy-L-allylglycyl-L-leucineby oxidation with periodate in the presence of catalytic amounts of permanganate.2 The compound thus prepared is chromatographically pure and is indistinguishable from the product obtained by the condensation of &benzyl carbobenzoxy-L-aspartate and cleucine benzyl ester followed by hydrogenolysis, Coupling reactions with carbobenzoxy- or p-nitrocarbobenzoxy-L-allylglycine proceed smoothly either with dicyclohexylcarbodide or ethyl chloroformate. The original procedure of Lemieux and Rudloff2 for the oxidation of unsaturated fatty acids was modified to permit work a t higher concentrations. The course of the oxidation can be followed quantitatively by iodimetric titration or qualitatively by the pH rise. The most favorable pH range is between 7 and 9. The rate of oxidation is rapid a t 20' in solutions that are 0.1 M , 0.025 M , and 0.001 M with respect to periodate, substrate, and permanganate, respectively. The aspartyl derivatives were isolated either by extraction from the acidified solutions with ethyl acetate or by precipitation from the concentrated aqueous solution after reduction of the excess oxidants with sodium metabisulfite.4 Yields between 60 and 95% of the theoretical were obtained. Although the applicability of this approach to other a-aspartyl derivatives has not been studied, certain limitations are inherent in this procedure. Thus, appropriate protective groups are necessary for tyrosine, serine, and threonine. For example, it was found that N-carbobenzoxy-L-tyrosine reacted with permanganate but uptake of oxidant was drastically reduced with 0-acetyl-N-carbobenzoxy-L-tyrosineunder the conditions used for the oxidation of allylglycine derivatives. However, the carbobenzoxy derivatives of tryptophan, methionine, and cysteine compete with the olefin for oxidant, and peptides containing these amino acids cannot be synthesized directly by the procedure described above.
(9) R. Adams and T. L. Cairns, J . A m . Chem. Soc., 61, 2464 (1939). (10) 9. L. Gertler and A. P. Yerington, U. 5. Dep. ART,,ABS-88-34 (1955);Chem. Abetr., SO, 7111 (1956). (11) The 60-Mc. n.m.r. analysis and interpretation were supplied by J. L. Holcomb of Varian Associates, Palo Alto, Calif. Deuteriochloroform was used a8 solvent.
A New Synthesis of a-L-Aspartyl-L-leucine' ERNESTSONDHEIMER Department of Chemistry, State University Colkge of Forestry, Syracuse University, Syracuse, New York Received May 19, 1964
The major obstacle to the synthesis of a-aspartyl peptides is the presence of two nonequivalent carboxyl groups. One way of circumventing this dficulty is to utilize a precursor which can be converted under mild conditions to a p-carboxylic acid after formation of the a-peptide bond. The conversion of L-allylglycine (L-2-amino4-pentenoic acid) to a-aspartyl peptides (1) Supported in part by a research grant GM 05492 from the National Institutes of Health, Public Health Service.
665
Experimental5 L-Allylglycine was prepared from N-acetyl-DL-allylglycines by the use of hog kidney acylase.' p-Nitrocarbobenzoxy-L-allylglycine.-To 3.8 g. of L-allylglycme in 7.7 ml. of 4 N sodium hydroxide at 0' was added with stirring in four portions at 20-min. intervals 8.1 g. of p-nitrocaxbobenzoxy chloride in 21 ml. of 1,Pdioxane and 10.3 ml. of 4 N sodium hydroxide. Stirring was continued for 4 hr. at 0'. The mixture waa filtered, the precipitate was discarded, the filtrate was acidified with concentrated hydrochloric acid, and the oil was extracted with ethyl acetate. After washing with water, drying with magne&m sulfate, and evaporating the solvent, the extracts yielded an oil that crystallized from 25 ml. of benzene. After two recryti tallizations from benzene, the yield was 6.0 g. (68%), m.p. 81D (c 3.5, dimethyl formamide). 83", [ C X ] ~ ~+4.5' Anal. Calcd. for Cl&hN2Oe: C, 53.06; H, 4.80; N, 9.52. Found: C, 53.00; H. 4.39; N. 9.82. (2) R. U. Lemieux and E. Rudloff, Can. J . Chem., 8 8 , 1701 (1955). (3) P. M. Bryant, R. H. Moore, P. J. Pimlott, and G. T. Young, J . Chem. Soc., 3868 (1959). (4) M. Jacobson, M. Beroza, and W. A. Jones, J . A m . Chem. Soe., 88, 4819 (1961). (5) All melting points were determined on a microscope hot stage and are uncorrected. Analyses were by George Robertson. Florham Park, N. J. (6)A. Neuberger and G. H. Tait, J . Chem. Soc., 3963 (1962). (7) 8. Bbck and N. G. Wright, J . B i d . Chem., 818, 39 (1955). (8) H.I(. Miller and H. Waelsch, J . Am. Chsm. Soc., 74, 1092 (1952).
666
NOTES
p-Nitrocarbobenzoxy-I;-allylglycyl-L-leucineBenzyl Ester.To a solution of 1.176 g. (4 m o l e s ) of p-nitrocarbobenzoxy-Lallylglycine, 1.892 g. of L-leucine benzyl ester p-tosylate,s and 0.35 ml. of triethylamine in 40 ml. methylene chloride was added 0.940 g. of d i c y c l o h e x y l c a r b o d . The mixture was shaken overnight at room temperature; 0.5 ml. of acetic acid was added. The solvent was evaporated from the atered solutions and the residue taken up in 100 ml. of ethyl acetate. The extract was washed with 0.5 N hydrochloric acid, water, saturated sodium hydrogen carbonate, and water. Concentration to 20 ml. yielded 1.57 g. (79y0) of crude product, m.p. 140-142'. Recrystallization from methylene chloride and ligroin yielded 1.34 g., m.p. 146-148'. Anal. Calcd. for C&lNsO,: C, 62.76; H, 6.28; N, 8.45. Found: C, 62.84; H, 6.43; N, 8.98. p-Nitrocarbobenzoxy-L-allylglycyl-L-leucine .-This compound could be prepared either by saponification of the bemyl ester (74y0yield, m.p. 70-73') or by direct coupling. To a solution of 1.176 g. of p-nitrocarbobenzoxy-L-allylglycineand 0.55 ml. of triethylamine in 12 d. of tetrahydrofuran at -5' was added 0.38 ml.of ethylchloroformate. After 25 min., 576 mg. of L-leucine in 2.2 ml. of 2 N sodium hydroxide was added; the mixture was held a t room temperature overnight. After evaporation of the solvent and addition of 15 ml. of water, the solution was extracted with ethyl acetate and the extract was discarded. The aqueous portion was acidified and extracted with ethyl acetate, the extract was washed with water and dried over magnesium sulfate, and the solvent was evaporated. After two crystallizations from toluene, 0.96 g. (59y0) of product was obtained, m.p. 64-73'. Anal. Calcd. for C1oH26Na07: C, 56.01; H, 6.19; N, 10.31; neut.equiv.,407. Found: C, 56.28; H, 6.21; N, 10.55; neut. equiv., 410. Oxidation Procedure.-To 18 ml. of 0.115 M sodium metaperiodate at 20' was added 0.5 mmole of olefin; the pH was adjusted to 7.5 with 0.5 M sodium carbonate. The addition of 2 ml. of 0.01 M potassium permanganate initiates the oxidation. Usually, the reaction was stopped after 1 hr. by acidification t o pH 2. In some experiments, the pH of the mixture was readjusted to pH 7.5 after the first hour and a second increment of 0.01 M potassium permanganate was added. However, the effect of this treatment on the yields of oxidation products was slight. The aspartic acid derivatives have been isolated either by extraction with ethyl acetate or by crystallization from the concentrated reaction mixture. For the latter procedure, the cooled reaction mixture was titrated with 1 M sodium metabisulfite until a colorless solution was obtained, excess acid was destroyed with solid sodium hydrogen carbonate, the solution was concentrated under reduced pressure, and the pH was readjusted to pH 2. Storage of the rnixture overnight at 4' yielded crystalline product. Oxidation of 0.5 mmole of carbobenzoxy-L-allylglycinee gave carbobenzoxy-L-aspartic acid in 78% yield when extracted into ethyl acetate while p-nitrocarbobenzoxy-ballylglycine was converted t o p-nitrocarbobenzoxy-baspartic acid in 95y0 yield when isolated from the aqueous phase after sodium metabisulfite treatment. Conversion of p-Nitrocarbobenzoxy-L-allylglycyl-L-leucie to a-cAspartyl-L-leucine.-The a c i d ~ e doxidation mixture from 407 mg . of p-nitrocarbobenzoxy-L-allylglycyl-bleuche was extracted twice with 30-ml. portions of ethyl acetate. The combined extracts were washed with water and dried over magnesium sulfate; the solvent was evaporated a t reduced pressure. The residue, which was noncrystalline, was dissolved in 5 ml. of acetic acid and hydrogenated at room temperature and 1 atm. in the presence of 50 mg. of palladium black for 2 hr. Filtration, evawration of most of the acetic acid, and addition of ether yielded 158 mg. (64%) of crystalline a-L-aspartyl-cleucine, recrystabed from aqueous acetone for analysis. Anal. Calcd. for C10H18N206: N, 11.38; neut. equiv., 246. Found: N, 11.44; neut. equiv. (titrahd to pH 6.5), 251. The material prepared by the above procedure and a-kaspartyl-cleucine prepared by the method of Bryant, et a2.,*were indistinguishable; their infrared spectra as potassium bromide pellets were completely superimposable; each gave only one ninhydrin-positive spot after chromatography on Whatman No. I paper in n-butanol-acetic acid-water, 4: 1 :5 (v./v.), Ri 0.59 (Bryant, et al.,8 reported Rf 0.60); [ a I a K-9.8' ~ (c 3.3, 0.1 N hydrochloric acid) [Bryant, et al.,* reported [a]"D -9.7" (c 3.42, 0.1 N hydrochloric acid)].
VOL.30 Quaternization of Aziridines. Evidence for the Monomeric State of Products' GEORQE K. HELMKAMP, RONALD D. CLARK,^ AND JAMES R. KOSKINEN University of California, Riverside, California Received September 4, 1964
I n the synthesis of a series of ,&substituted amines, quaternary aziridinium salts were considered as potentially convenient alkylating agents. Protonated aziridines of varying stability have been isolated and described,a-Kbut there is no comparable description of exhaustively alkylated species prepared directly from aziridines. The exceptional reactivity of the quaternary salts toward nucleophiles has virtually restricted salt preparation to those exceptional instances in which the anion and solvent have low nucleophilicity,8 the carbon skeleton has special steric features,'~~ or (in one instance) a stable complex forms.s Quite recently, however, Bottini and VanEtten investigated the quaternization of cis- and truns-1,2-dimethyl-3-isopropylaziridine and demonstrated the formation of stable monomeric iodides.'O When cis- or truns-l12,3-trimethylaziridine (la) waa treated with methyl iodide, a white crystalline solid formed rapidly. The product was found to deCHa I
7%
I
R la, R=CH3
b, RxCHsCHz
compose readily on attempted recrystallization. Similar observations were made with isomers of l-ethyl2,3-dimethylaziridine (lb). If an analogy were made with the hydrochlorides of aziridines, it could not be assumed that the materials were monomeric species, for theae protonated compounds polymerized vigorously at room temperature.ar6 The instability of the quaternary salts to procedures used for crystallization, however, indicated a structural feature other than an unstrained, quaternized nitrogen atom of the piperazine type 4. Such piperazinium salt formation has been observed4v1l in the spontaneous dimerization of nitrogen mustards. (1) Supported in part by Cancer Research Funds of the University of California and Grant GM 8185 from the National Institutea of Health, U. 9. Public Health Service. (2) Taken in part from the Ph.D. Thesis of R. D. Clark. (3) 0.D. Jones, A.Langsjoen, Sr. M. M. C. Neumann, and J. Zomlefer, J. Ow. Chem., 9, 126 (1944). (4) C. Golumbio, J. 8. Fruton, and M. Bergmann, ibid., 11, 618 (1946). (6) R. D. Clark and G. K. Helmkamp, ibid., 49, 1316 (1964). (6) N. J. Leonard and K. Jann, J. Am. Chem. Soc., 84, 6418 (1960). (7) P. E. pants, L. J. Pandya, W. R. Groskopf, and H.-J. Su, ibid., 48, 413 (1963). (8) N. J. Leonard, K. Jann, J. V, Paukstelis, and C. I(. Steinhsrdt, ibid., 98, 1499 (1963). (9) N. B. Chapman, J. W. James, J. D. P. Graham, and G. P. Lewis, Chem. Znd. (London), 805 (1962). (10) A. T. Bottini and R. L. VanEtten, J. Ow. Chcm., SO, 675 (1966). We are grateful to Professor Bottiii for informing us of these results prior to publioation and for extensive disousaion on the monomeric us. dimeric nature of the produota. (11) P. D. Bartlett, S. D. Ross, and C. G. Swain,J. Am. Cham. Soc., 69, 2971 (1947).
FEBRUARY 1965
NOTES
When the methiodides of 1 were treated with silver 2,4,6-trinitrobenzenesulfonate,stable salts were obtained. In an attempt to ascertain the state of the compounds, molecular weights were determined by vapor pressure osmometry. With acetonitrile as solvent and benzil as a standard, the compounds appeared to be dimeric. In ethanol, with biphenyl as a standard, the results were equivocal. Table I presents a summary of data observed for species as related t o the two considered states of association.
When the optically active aziridine wa8 used, it was found to form an iodide with a moderately high specific rotation: [ C ~ ] ~+31.55' ~D (0.445 F in ethanol). The assigned structure then must be that of the monomer 3. I n order to provide further c o h a t i o n , the reaction of 3 with excess ammonia or dimethylamine was followed polarimetrically. Each reaction was first order in aziridinium salt (Table 11)through a t least three half-lives of material.
TABLEI EQUIVALENT WEIQHTSOF QUATERNARY AZIRIDINIUM SALTS DETERMINED BY OSMOMETRY'
Aziridinium salt; anion
Equiv. wt., calcd. Epuiv. wt., found -fo--inMono- AcetoDimer mer nitrile Ethanol
truns-l,1,2,3-Tetramethyl; iodide
141
106
157b 145'
t~uns-1,1,2,3-Tetramethyl; TNBS~
252
189
251' 233"
trans-1-Ethyl-l , 2,3-tnmethyl; iodide
153
114
153* 142'
trans-1-Ethyl-l , 2,3-tnmethyl;
138"
. .. 137'
.
264b .. 245' a Determined with a Mechrolab Model 301A osmometer, values based on rtssumption of complete ion dissociation. Benail aa reference standard. Biphenyl aa reference standard. 2,4,6-Trinitrobenzenesulfonateanion.
TNBS~
261
196
In the absence of definitive data concerning the monomeric or dimeric states of the quaternized aziridines, the use of optically active 1,2,3-trimethylaziridine (2) was considered. The alkylation of the active compound, eq. 1, would lead to an optically active monomer 3 in the absence of racemization. If the reaction proceeded beyond the monomeric state in the presence of excess alkylating agent, eq. 2, the dimeric material should have the m s o structure shown by structure 4. If the configuration were other than that of 4, it would have to have resulted from racemization or &-opening of the immonium ring. Only in the latter instance could it be possible to obtain an optically active dimeric product. By analogy with other ring-opening reactions such as acid-catalyzed reactions of epoxides, there is evidence for retention of configuration as a specific process only in those instances in which substituents are arranged for neighboring group participation. l 2
667
TABLEI1 FOR THE REACTION PSEUDO-FIRSFORDER RATECONSTANTS OF 1,1,2,3-TETRAnaEITWL~I~INmM IODIDE WITH AMINES"
R08Ct-t
8DIh
Amine concn., F
Rate constant, min. -1
0.0237 0.0232 (CHs)zNH 3.32 0.0679 0.0668 a Followed polarimetrically at 400 nlrc and 25". a"
4.74
Half-life, min.
29.2 29.8 10.1 10.3
The reaction with dimethylamine, eq. 3, yielded an optically inactive product, the bkdimethylaminobutane (5) with a specXc rotation of [ C Y ] ~ 0.00 W 0.01'. The ammonia reaction, on the other hand, yielded an optically active product (6,eq. 4) with a specific rotation of [ c x ] 16 ~ ~f~2'.~ ~
*
(3)
The mere fact that the alkylation product of an aziridine reacted readily with a nucleophile verifies the conclusion that the three-membered ring remained intact, for the dimeric product, a derivative of piperazine, should not have been as easily susceptible to attack by a nucleophilic species. The retention of optical activity in the formation of 6 excluded racemization as a principal feature of the process, and the formation of the inactive compound would characterize it as predominantly meso. Experimentalla
N-Alkyl-Z,3-Dimethylaziridines.--The aziridines were prepared from the corresponding epoxides by the Wenker method" as applied by Ghirardelli and Lucaa in the butene series." Yields of products vaned from 30 to 50%. t~am-l,2,3-Trimethylaziridine.-The compound waa prepared from 25.8 E. (0.36 mole) of tram-2.3-eaoxvbutane: vield. 10.0 e., 33%; byp.'71.Ck71.20 (735 j;n% i.4045. Anal. Calcd. for CbHllN: C, 70.53; H, 13.02. Found: C, 70.26; H, 13.20.
-.
(13) Microanalyses were by C. F. Geiger, Ontario, Calif. (14) H. Wenker, J. Am. Chem. SOC.,67,2329 (1936).
(12) R. E. Parker and N. 9. IsaaO8, Chem Rm., 69, 737 (1969).
(16) R. Ghirardelli and H. J. Lucas, ibid., 79, 734 (1967).
I
-
668
NOTES
truns-l-Isopropyl-2,3-dimethylaziridine.-The compound WBB prepared from 12 g. (0.17 mole) of truna-2,3-epoxybutane: yield, 6 g., 31%; b.p. 102-102.5' (735 mm.); n2% 1.4019. A d . Calcd. for C~HISN: C, 74.27; H, 13.36; N, 12.37. Found: C, 74.23; H, 13.22; N, 12.12. truns-l-Isopropyl-2,3-di1nethylaziridinium 2 ,C,d-Trinitrobenzenesulfonate.-The salt waa prepared as a derivative of the aziridine by adding a 10% solution of the aziridine in acetonitrile to a solution of 2,4,6-trinitrobenzenesulfonic acid in the same solvent until a distinct red coloration appeared. The red color was discharged by the addition of a small amount of the acid. After the product waa precipitated by the addition of ether, it was recrystallized twice from absolute ethanol; m.p. 211.5-212.5'. Anal. Calcd. for C18H18N400S: C, 38.42; H, 4.46; N, 13.79. Found: C, 38.41; H, 4.26; N, 14.09. truns-1,1,2,3-Tetramethylaziridinium Iodide.-The salt waa prepared by adding 2.1 g. (0.025 mole) of trans-1,2,3-trimethylaziridine t o 25 ml. of methyl iodide which had been cooled to about -50'. White crystals began to form immediately. After about 10 min., when the mixture had reached room temperature, the crystalline solid waa removed by filtration and air dried; yield, 5.4 g., 96%. The compound melted with decomposition in the region of 135-150'. Since the methiodide could not be recrystallized without excessive decomposition, it was converted to the 2,4,6-trinitrobenzenesulfonate salt. To 0.050 g. (0.00125 mole) of silver trinitrobenzenesulfonate in 15 ml. of acetonitrile waa added 0.301 g. (0.00125 mole) of the aziridinium iodide. Silver iodide (0.27 9.) precipitated immediately. The product waa precipitated by the addition of about 500 ml. of anhydrous ether and recrystallized from acetonitrile-ether. The product melted at 205-206' dec. Anal. Calcd. for C12H1eNdO&3: C, 36.73; H, 4.11; N, 14.28. Found: C, 36.88; H,4.44; N, 13.98. truns-l,1,2 ,3-Tetramethylaziridinium 2,4,6-Trinitrobenzenesulfonate.-In an alternate procedure for the preparation of the trinitrobenzenesulfonate, 1.O g. (0.0025 mole) of trimethyloxonium trinitrobenzenesulfonate16in 15 ml. of acetonitrile waa added to 0.21 g. (0.0025 mole) of truns-l,2,3-trimethylaziridinein 15 ml. of acetonitrile. The solution was allowed t o stand for 15 min. before precipitation of the product by the addition of about 500 ml. of anhydrous ether. The yields in this and subsequent analogous reactions were nearly quantitative. The infrared spectrum and melting point of the salt were identical with those of the product formed from the quaternary iodide. A mixture melting point showed no depression. cis-l-Ethyl-l,2,3-trimethylaziridiniumIodide .-The compound was prepared by a method analogous to that described for the trans-tetramethylaziridine. Methyl iodide waa used aa the alkylating agent for the N-ethylaziridine. The iodide salt again was unstable and could not be recrystallized; m.p. 116-128' dec . cis-l-Ethyl-l,2,3-trimethylazirid~um Bromide .-The bromide wm prepared from the reaction between cis-l-ethyl-2,3-dimethylaziridine and methyl bromide in which the latter waa the solvent. The highly hygroscopic product waa purified by sublimation; m.p. 134-138' dec. Anal. Calcd. for C7HlsBrN: C, 43.31; H, 8.33; Br, 41.17. Found: C. 43.31: H, 8.68; Br, 40.73. truns-l-Ethyl-l,2,3-trimethylaziridiniumIodide.-The compound was prepared directly from the aziridine and methyl iodide. The product decomposed on attempted recrystallization. Nitrogen snalysis (calcd., 5.81%; found, 5.74%) satisfactory, but the compound waa converted t o the t m t r o benzenesulfonate for further characterization. truns-l-Ethyl-l,2,3-trimethylaziridinium 2,4,6-Trinitrobenzenesulfonate .-The compound waa prepared by the alternative routes described for the corresponding N,N-dimethyl derivative. Both routes yielded identical producta; m.p. 212-214' dec. Anal. Calcd. for ClaH18N4OgS: C, 38.42; H, 4.46; N, 13.79. Found: C, 38.71; H, 4.78; N, 14.08. Equivalent Weight Deterrnination.-Meaaurements were carried out with a Mechrolab Model 301A vapor pressure osmometer. The solutions were between 0.04 and 0.06 F,ttssUming the molecular weight of the monomer. Standard clwes were determined with benzil and biphenyl aa the solute in acetonitrile or ethanol. D( +)-1,1,2,3-TetramethylaMddum Iodide.--The optically active iodide waa prepared in the manner described for the in-
VOL.30
active salt. It waa used directly for kinetic measurements because of its instability on attempted recrystallization. The corresponding trinitrobenzenesulfonate could not be utilized because it also reacted with the ammonia or dimethylamine to be used for ring opening. The specific rotation of the iodide, [a]*% +31.55', waa determined in ethanol. There waa no change of rotation with time in solution, so no reaction with solvent waa apparent. Even so, it waa impossible t o reisolate the compound in purer form. maso-Bis-2,3-dimethylaminobutae.-The compound waa isolated from the reaction mixture used for kinetic measurements. The reaction solution consisted of 0.050 g. of D( +)-1,1,2,3tetramethylaziridinium iodide dissolved in 2.00 ml. of 3.32 M aqueous dimethylamine. After the solution had stood for 2 hr. at 25', it waa saturated with potassium hydroxide and extracted with ether. The ether solution waa dried over potassium hydroxide pellets. The introduction of hydrogen chloride to the supernatant ether solution precipitated a white solid which was crystallized successively from ethanol and 1-propanol and sublimed. Because of inconsistent melting point characteristics and long melting point range, the hydrochloride waa converted to the trinitrobenzenesulfonate by the addition of 2,4,6-trinitrobenzenesulfonic acid to an ethanolic solution and precipitation with ether. The product waa recrystallized from acetonitrile containing a small amount of water; m.p. 263-264' dec. Anal. Calcd. for CZ&NNSOISSZ: C, 32.88; H, 3.59; N, 15.34. Found: C, 32.85; H, 3.62; N, 15.09. (2S:3R)-2-Dimethylamino-3-aminobutane.-The compound waa isolated as the hydrochloride from the solution used for kinetic meaaurements. The technique waa that described for the bisdimethylamino product. The hydrochloride was very hygroscopic, but it waa converted to a stable, easily handled compound by the addition of 2,4,6-trinitrobenzenesulfonic acid to an ethanolic solution. The trinitrobenzenesulfonate waa crystallized from nitromethane-ether; m.p. 255-258' dec. Anal. Calcd. for ClsHnNaOl&: C, 30.77; H, 3.16; N, 15.95. Found: C, 30.50; H, 3.39; N, 15.70. Kinetic Measurements.-The reactions of ammonia and dimethylamine with the optically active aziridinium iodide were followed polarimetrically at 400 mp, using a J. C. Rudolph and Sons manual spectropolarimeter with an oscilloscope for readout. The temperature of all runs waa kept a t 25.0'. The reaction mixtures were prepared a t the operating temperature and placed into a thermostated, 10-cm. polarimeter tube of all glass construction except for quartz windows. Measurements were made t o f0.003' rotation through at least three half-lives of the optically active reactant. The optical activity of the reactant and its concentration in solution were used to estimate the optical activity of products. I n the caae of the bisdimethylaminobutane, the lack of optical activity waa verified by noting the lack of optical rotation of its dihydrochloride.
Preparation of Oxamide from Hydrogen Cyanide and Hydrogen Peroxide R. C. SHERIDAN AND E. H. BROWN
~
.was
(16) D. J. Pettitt and G. K. Helmkamp, J. Ow. Chsm., 28, 3833 (lC68).
Division of Chemical Development, Tennessee Vulley Authority, Wilson Dum, Alabama Received September 1, 1964
Attfieldl in 1863 and Radziszewski2in 1885 reported that hydrogen cyanide and hydrogen peroxide react to form oxamide, (CONH&, but Rupp and Pfenniga found that carbon dioxide and ammonia also are produced. The reaction to produce oxamide may be written 2HCN
+ H202 +(CONH2)2
(1) Attfield, J . Chem. Soc., 16, 94 (1863). ( 2 ) B. Radziszewski, Be?., 18, 355 (1885). ( 3 ) E. Rupp and F. Pfennig, Chem. ZQ., 84, 322 (lQl0).
FEBRUARY 1965
NOTES
but the reaction apparently proceeds in a t least two stages such as
hydroxide decreased the yield of oxamide but had no effect on either the induction period or the rate of the reaction, Masson' reported that oxamide is not formed by the reaction of hydrogen peroxide with potassium cyanide. When the reaction was carried out at 0' in nonaqueous solvents (ethyl acetate, amyl acetate, or isobutyl alcohol) the results were about, the same as those obtained in aqueous solution: from 43.5 to 47.2% of the hydrogen cyanide was converted to oxamide, and 4.8 to 5.4 moles of hydrogen peroxide mole/of oxamide were consumed. The oxamide, however, appeared in a different crystal form which on recrystallization from water yielded the usual triclinic modification?
2HCN (CN)2
or, over-all ZHCN
+ HZ02
(CN)2
+ 2H20
+ 2H202 +(CONH2)z +
0 2
+ 3&02 -3 (CONHZ)~+ 2H20 +
0 2
McMaster and Langreck4 and Radziszewski2 prepared several amides by the reaction RCN
+ 2H202 ----t RCONHz + H2O +
0 2
and Wiberg6 studied the preparation of benzamide from benzonitrile by this reaction. A brief study was made of the reaction of hydrogen cyanide with hydrogen peroxide in an exploration of methods for the preparation of oxamide. Mixtures of aqueous solutions of hydrogen cyanide and hydrogen peroxide stand quietly with no apparent reaction for an induction period that decreases as the temperature of the cooling bath is raised, The mixtures then begin to effervesce, oxamide crystallizes, and the exothermic reaction becomes quite vigorous unless the mixture is cooled. The effect of the temperature of the cooling bath on the reaction is shown in Table I. TABLEI
REACTION O F HYDROQEN CYANIDE WITH HYDROQEN PEROXIDE IN
-Temp., Bsth
O C -
Max.
AQUEOUSSOLUTION' -Time, Induction
m i n . 7 Reaction
% of HCN converted to (C0"n)n
140 48.5 205 68 24 42.9 10 35.3 40 2 7.5 15 'Charge, moles: HCN 0.1253, HzO2 0.4944, H2O 4.44. 0 15 23 35
3 27 43 75
In the runs reported in Table I, the mole ratio H202:HCN in the charge was 3.95. In tests a t ,'O variation in this ratio from 1.65 to 4.70 made no significant difference in the fraction of hydrogen cyanide converted to oxamide (45 to 47%) or in the consumption of hydrogen peroxide, which ranged from 4.0 to 4.9 moles/mole of oxamide produced. Dilution of the reaction mixture with water increased the induction period and decreased the yield slightly. The spent reaction mixtures contained hydrogen peroxide, but no hydrogen cyanide. The gases evolved during the reaction contained ammonia and carbon dioxide, indicating this side reaction. HCN
+ HzOz = C02 + NHa
The oxamide was a homogeneous, white, crystalline product with the optical properties and powder X-ray pattern of the usual triclinic modification.6 Addition of the miscible organic solvents ethanol, methanol, dioxane, tetrahydrofuran, and acetone to the aqueous reaction mixture had no significant effect on the reaction, but N,N-dimethylacetamide both increased the induction period and decreased the yield of oxamide. Addition to the aqueous reaction mixture of small amounts of either sulfuric acid or potassium (4) L. McMaster and F. B. Langreck, J . A m . Chem. SOC.,39, 103 (1917).
(5) IC. B. Wiberg, ibid., 76, 3961 (1953). (6) E. M. Ayerst and J. R. C. Duke, Acta Cruet., 7, 588 (1954).
669
Experimental Stock solutions (usually about 10% HCN) were prepared from liquid hydrogen cyanide and chilled solvents. Aqueous hydrogen peroxide solutions were prepared by diluting reagent grade (30%) solutions; solutions of hydrogen peroxide in water-immiscible organic solvents were prepared by shaking mixtures of the solvents with 30% aqueous solutions of hydrogen peroxide and separating and ducarding the aqueous' phaaes. These solutions were then dried with anhydrous sodium sulfate and filtered. Hydrogen cyanide in the stock solutions and spent reaction mixtures was determined by titration with silver nitrate, and hydrogen peroxide was determined by titration with potassium permanganate. In each test of the reaction, the hydrogen cyanide and hydrogen peroxide solutions were mixed at the bath temperature in 250-ml. wide-necked volumetric flaaks. The mixtures were allowed to stand in the cooling bath, with occaaional swirling, until the reaction appeared to be complete; they were then filtered and the filtratetes were analyzed for hydrogen cyanide and hydrogen peroxide. The precipitated oxamide was washed with small portions of icewater, rinsed with acetone, dried at 105', and examined by X-ray analysis.e (7) 0.Masson, J. Chem. Soo., 91, 1449 (1907). (8) I. R. Lehr and E. H. Brown, N o t w e , 108,1168 (1964).
The Structure of the Dehydration Product, C1oH140, of 2-Hydroxyis~pinocarnphone~ TAKAYUIU SUQA,KENICHIRO MORI, AND TAMON MAWUURA
Department of Chemistry, Faculty of Science, Hiroshima University, Hiroshima, Japan Received M a y 81,196'4
On dehydrating 2-hydroxyisopinocamphone2(1) with anhydrous oxalic acid, Kuwataa isolated a terpene ~ dZ640.9450) along ketone (b.p. 205-210°, n 2 61.5034, with the major product, carvacrol, and proposed it to be p-mentha-4(8) ,6dien-2-one (2). The assignment of structure 2 was based on elementary analyses and nonidentity of the oxime, m.p. 103-104.5", and semicarbazone, map. 183-184') derivatives with derivatives in the literature. "Isocarvoxime" (3))m.p. 143-144', the oxime derivative of 2, has been synthesized from (1) Supported in part by a grant from the Ministry of Education in Japan. (2) The spatial arrangement of the hydroxyl group at C-2 has been clarified reoently by H. Schmidt [Chem. Ber., 98, 2485 (I960)]. (3) T. Kuwata, J . SOC.Chem. I d . Japan, 40,24 (1937); Chem. Abstr., 81, 3899' (1937).
NOTES
670
1
2
3
4
carv0ne4~and clearly is not identical with the oxime derivative of the terpene ketone considered to be 2 by Kuwata. We have now repeated the dehydration of 1 with anhydrous oxalic acid in accordance with Kuwata's report* and have obtained, in 9.8% yield, a terpene ketone, CloH1hO, whose properties are identical with those reported by Kuwata. An examination of the infrared, ultraviolet, and n.m.r. spectra of the CI0Hl40 ketone demonstrated its identity with eucarvone (4). This identity was further confirmed by the coincidence in melting points of the oxime, semicarbazone, and 2,44initrophenylhydrazone derivatives with those reported for the derivatives of eucarvone (4). The conversion of a pinane derivative to eucarvone by an organic acid is unique. Experimental
All melting points reported are uncorrected. (- )-2-Hydroxyisopinocamphone (1) .-This compound was prepared from (+)-.-pinene (b.p. 156.7', [ a I U ~32.72') by oxidation with potassium permanganate a t @-5" following the method reported.816 The crude compound thus obtained was purified by distillation: b.p. 96.5-96.7' (5 mm.), m.p. 34-35', [a]% -23.26" (c 1, ethanol); lit.at8 m.p. 35.5-36.5', DI.[ - 18.65' (c 1.4, ethanol). Anal. Calcd. for CioHlsOz: C, 69.73; H, 9.37. Found: C, 70.39; H, 9.45. Preparation of Dehydration Product, C1&O.-A mixture of 1 (200 g.), acetone (266 ml.), and anhydrous oxalic acid (333 g., m.p. 187-189') was heated under reflux for 6 hr. After acetone wm distilled off in vucuo, the residue was poured into a large amount of a saturated sodium chloride solution and then extracted with petroleum ether. The petroleum ether solution, after removing carvacrol (82.8 g., b.p. 111.5-112' at 9 mm., (4) 0. Wdaoh, Ann., 846, 266 (1906). (5) H. Gddsahmidt and E. Iiisser, Ba.,10, 2073 (1887). (6) T. Kuwata, J . SOC.Chem. I d . Japan, 89, 814 (1936).
VOL.30
identified by infrar6d spectrum comparison) by washing with 5% potassium hydroxide solution, gave a neutral reaction mixture (98.9 9.). The distillation of the neutral reaction mixture yielded 19.6 g. (9.8% based on 1 taken) of a fraction boiling at 82-84' (10 mm.) and showing the following physical constants: b.p. 205-208', n% 1.5055, 0.9445. Structural Confinnation of Dehydration Product.-The further rectification of the fraction boiling at 82-84' (10 mm.) gave a main fraction showing b.p. 84.2' (10.5 mm.), n% 1.5050, d*4 0.9455, and 303 mp (log e 3.81); lit. for eucarvone (4): b.p. 85-87' (12 mm.),' n34.8~1.5051,* daO, 0.948,O 303 mp (log e 3.83) .a Gas-phase and chromatostriplo chromatography indicated this oil to be homogeneous. Anal. Calcd. for Cl&fl4O: C, 79.95; H, 9.39. Found: C, 79.53; H, 9.22. The infrared spectrum (neat film) exhibited bands at 1665 (conjugated C=O), 1643, 1601 (C=C), 1425 ( - C H r C O - ) , 1390 and 1368 (+Me), 826 (>C=CH-), and 727 (Cis-CH= CH-) em.-'. The n.m.r. spectrum (neat, 40 Mc.) showed peaks at 7 8.94 (-CMez, relative intensity 6), 8.11 (CHa-C= C ,3), and 7.36 (-CH&O-, 2). The oil gave the following crystalline derivatives, the melting points of which coincided well with those reported for eucarvone (4)
-
e:"
Oxime, m.p. 105-106.5', 291 mp (log e 4.03), lit.11 m.p. 106'. Anal. Calcd. for CIOHI~NO:C, 72.69; H, 9.15; N, 8.48. Found: C, 72.77; H, 8.98; N, 8.38. Semicarbazone, m.p. 185-186', 310 mp (log E 4.29); l i t . 1 2 m.p. 186-186.5", 309 mp (log E 4.22). Anal. Calcd. for C1~HI~N~O: C, 63.74; H, 8.27; Found: C, 63.42; H, 8.01. 2,4Dinitrophenylhydrazone, m. 151-152', ' : :A 384 mp (log e 4.68); lit.'* m.p. 151-152', 2$388 mp (log e 4.44). Anal. Calcd. for C16HlsN404: N, 16.87. Found: N, 17.12.
e:E
Acknowledgment.-The authors are indebted to Arakawa Rinsan Company, Ltd., and Tokuch6 Kagaku Company, Ltd., for a gift of a-pinene, and to Dr. Shim Tonomura for microanalyses. (7) 0. Wallaoh and € Kohler, I. Ann., 889, 104 (1905). (8) R. A. Barnes and W. J. Houlihan, J . O w . Chem., 16, 1609 (1961). (9) A. Baeyer, Be?., 17, 812 (1894). (10) J. G. Kirohner, J. M. Miller, and G. J. Keller, Anal. Chem., 18, 420 (1951); J. M. Miller and J. G. Kirohner, ibid., 16, 1107 (1953). Lohr, I. Ann., 806, 239 (1899). (11) 0. Wallach and € (12) A. E. Gillam and T. F. West, J . Chem. SOC.,483 (1942). (13) J. R. B. Campbell, A. M. Islam, and R. A. Raphael, J . Chem. SOC., 4086 (1966).
