3648 NOTES

The Journal of Organic Chemistry

TABLE I DECOMPOSITION 1-PHENYGL-CYCLOHEXYL OF PERACETATE (I)

Solventa

Reaction time, hr

Phenol

Cyclohexanone

Product composition, I-Phenoxyeyclohexene

Caprophenone

Miscellaneous

Ethanol 62 53 34 C Acetic acidd 52 20 30 44 e 72 11 12 Acetic acidf 72 9 58 h Heptane 25 34 Pyridine 22 i i cu. 75 Ca. 25 Reaction was carried out in refluxing solvent, unless otherwise noted. Percentages of peak areas obtained in gc analysis with a Carbowax 20M column a t ca. 200". These data do not represent accurate mole per cent composition, but do indicate approximate relative amounts of the indicated products. e TWO peaks (13%) were found at very short retention times, one of them probably being solvent. d Reaction temperature, 80". e 5% phenyl acetate was detected, also unidentified high boiling material, ca. 15% of total observed area. f Reaction carried out a t room temperature (ca. 25'). 0 There was also 5% unidentified material. h 8% of what appears to be I-phenylcyclohexanol was detected. Some unidentified higher boiling material was present. i Traces of phenol and cyclohexanone were detected.
5

t-butylperoxy chloroformate, the mechanism of which is solvent dependent.* Furthermore, the nature of the ionic products in the decomposition of I depends on the particular solvent employed, as discussed above.
Experimental Section

G s chromatographic analyses were carried out with a Perkina

Elmer Model 154 instrument using a Carbowax 20M column (10% on 70-80 mesh Anakrom ABS) a t 200, and with an F & M Model 500 instrument using a 2-ft 10% silicone rubber (SE 30) on 60-70 mesh Anakrom ABS column. The F & M instrument was programmed for 70-340" a t 21/min. Peak areas were calculated by the peak height-width at half-height method. 1-Phenyl-1-cyclohexyl Peracetate.-1-Phenylcyclohexyl hydroperoxide was prepared from the corresponding alcohol with 90% (or 70%) hydrogen peroxide using the procedure of Hey, Stirling, and william^.^ Reaction of the pure hydroperoxide with acetyl chloride in pyridine gave the liquid peracetate in 87% yield. The identity of the peracetate was confirmed by its infrared spectrum (neat); strong bands a t 1754 (C=O), 1183 (C-0), and 754 and 694 cm-l (monosubstituted aromatic). The nmr spectrum (cc14) confirmed the structure of the peracetate, showing a multiplet (5 H) in the aromatic region near 6 7.3 ppm and a broad signal for the ring protons from ca. 1.4 to 2.3 ppm, which included a sharp singlet for -0COCHs at 6 1.72 ppm (total 13H). Decomposition of the Peracetate in Ethanol.-1-Phenyl-1cyclohexyl peracetate (6 g) was heated under reflux for 62 hr in absolute ethanol. The cooled reaction mixture was diluted with salt solution and extracted with ether. The ether solution waa washed with sodium bicarbonate and salt solution and dried over MgS04. An infrared spectrum of the recovered product, 3.7 g, showed no evidence of starting peracetate. Strong hydroxyl and carbonyl (1689 cm-l) bands were present. The product was analyzed by gas chromatography on a Carbowax 20M column a t 198" (Table I). Further analysis on the silicone rubber column programmed to 340" confirmed phenol and cyclohexanone as the only significant components in the mixture. Decomposition of I in Acetic Acid.-Compound I (3 g) was decomposed in 30 ml of glacial acetic acid a t 80" for 52 hr. The product (2 g) was worked up as described above, except that methylene chloride was used for extraction. Analysis of the product was carried out by gas chromatography on the Carbowax 20M and silicone rubber columns (Table I). In a similar decomposition, the crude product was washed with 10% sodium hydroxide and distilled, giving l-phenoxycyclohexene, bp 66-68' (0.15 mm). Anal. Calcd for C__~ H U O :C, 82.71; H, 8.09. Found: C, I . . . . 82.04; H, 8.19. The infrared spectrum (10% cc14) showed bands a t 1220 (s) (aromatic ether). 1669 (C-C str), 862 em-' (=CH out of plane deformation). The structure of the vinyl ether was confirmed by its nmr spectrum, which showed a multiplet (5 H) in the aromatic region near 7 ppm, an olefin proton (multiplet, 1 H ) at about 6 4.94 ppm, and the ring methylene protons (8 H ) in the range 1.4-2.4 ppm composed of two major overlapped multiplets corresponding to the ring allylic and nonallylic protons. Strong
(8) P. D. Bartlett and H. Minato, J. Amer. Chem. Soc., 86, 1858 (1963).

irradiation in the allylic proton region (175.8 cps upfield from the olefin signal) collapsed the olefin proton signal to a broad singlet. Decomposition o I in Heptane and midine.-Compound I f (2 g) was decomposed in 20 ml of refluxing heptane under nitrogen for 25 hr. The product (1.7 g) had an infrared spectrum which showed very little hydroxyl absorption and no cyclohexanone. There was a strong band at 1667 [CsH&(==O)-] and at 1220 cm-l. Analysis of the product was carried out on the Carbowax 20M and silicone rubber columns (Table I). The peracetate (1.3 g) was decomposed in 20 ml of refluxing pyridine under nitrogen for 22 hr. The infrared spectrum of the product showed virtually no hydroxyl absorption but showed strong bands a t 1667 and 1215 cm-l. Product analysis was carried out on a Carbowax 20M column (Table I).

Registry No.-I, 17012-35-0.

17012-34-9; 1-phenoxycyclohexene,

Acknowledgments-Thanks are extended to Dr. P. E. Butler of the Analytical Research Division for performing the nmr decoupling experiment and also to Mr. Jack Oliver for his technical assistance.

Reaction of 2-Acetylcyclohexanone with Malononitrile and Cyanoacetamide

FILLMORE FREEMAN, DONALD FARQUHAR, K. A N D ROBERT WALKER L.
Department of Chemistry, California State College at Lung Beach, Long Beach, California 90801 Received February 13, 1968

It has been reportedl1V2 without structure proof, that the pyridine-catalyzed reaction of cyanoacetamide (I) and 2-acetylcyclohexanone (11) gave a mixture of 3-cyano-2-hydroxy-4-methyl-5,6,7,8-tetrahydroquinoline (111) and 4-cyano-3-hydroxy-l-methyl-5,6,7,8-tetrahydroisoquinoline (IV). During studies of the reaction of malonic acid derivatives and dicarbonyl compounds, we found that the diethylamine-catalyzed reaction of malononitrile (V) or I with I1 gave a solid with the empirical formula Cl1Hl2N20. We have shown this product to have structure IV via elemental analysis, spectral analyses, and chemical reactivity. Also, we have prepared several previously unknown derivatives of IV. The nmr spectrum of IV in trifluoroacetic acid, which shows resonances at 7 6.98 (m), 7.35 (m), 7.48 (s, C-CH,), and 8.17 (m), does not distinguish
(1) H. K. Sen-Gupta, J. Chem. Soc., 107, 1347 (1915). (2) H. I. Sen and U. Bose, J . Indian Chem. Soc., 4, 61 (1929). (

Vol. 33, N o . 9,September 1968

CH3 CH3

NOTES 3649
which cyclizes via a nucleophilic attack of the unshared pair of nitrogen electrons at the acyl carbonyl carbon, followed by elimination of water.

IIIa
CN

IIIb

CN

I+II-

qrCN
XI

4 -

kH3 IVa

IVb

between I11 and IV nor does the ultraviolet spectrum, which is similar to 3-isoquinolinol, distinguish between I11 and IV. The infrared spectrum showed absorptions a t 4.45 p (conjugated CN), 6.05 and 6.21 (>C=O tf >C-OH), and 8.7 p (is~quinoline).~-+ Phosphorous pentachloride converts IV into a single chloride VI. The isoquinoline structure is evident from the infrared band a t 8.75 p and the absence of absorption in the 9.05-p r e g i ~ n . The ~ ~ ~ nmr spectrum of VI in CDC13 showed methylene hydrogens a t T 7.10 (m), 7.38 (m), and 8.16 (m) and methyl hydrogens a t 7.75 (s).
CN
COOH

@
Experimental Section

CN

HO CH, * I V

This procedure provides a simple and high yield route to otherwise difficulty accessible substituted 5,6,7,8-tetrahydroisoquinolines. The scope and mechanism of this reaction with divers dicarbonyl compounds and different malonic acid derivatives are being investigated.

CH3

CH,
WI

VI

VI1

Judicious hydrolysis of IV with 25% potassium hydroxide solution gave a single acid VII, mp 259-260" dec, and its decarboxylation product VIII, mp 236237", in the same experiment. Dornow and Neuse4 have reported that the corresponding quinoline isomers, IX and X, melt a t 242' and a t 217-218", respectively.

IX

Both VI1 and VIII showed the characteristic isoquinoline band in the infrared spectrum at 8.75 and 8.72 p, respectively. Finally, thermal decarboxylation' of VI1 gave VI11 in 73.6% yield. These data are consistent with the formulation of the condensation product as IV. The mechanism presumably involves attack of the anion derived from I or V a t the cyclic carbonyl carbona to give a Knoevenagel condensation product (XI)
(3) H.E.Baumgarten, W. F. Murdock, and J. E. Dirks, J . Orp. Chem., 46, 803 (1961). (4) A. Dornow and E. Neuse, Arch. Pharm., 487, 361 (1954). (5) H.Luther, ibid., 487, 383 (1954). ( 6 ) H.Junek, Montosh. Chem., 96, 1473 (1964). (7) H.Hart and F. Freeman, J . Amer. Chem. Soo., 86, 1161 (1963). (8) A. Lapworth and R . H. F. Manske, J . Chem. Soc., 2533 (1928); 1976 (1930). Cyclic ketonas generally react faster than acyclic ketones in carbonyl addition reactions.

General.-All melting points are uncorrected and were determined on a Thomas-Hoover apparatus. Nmr spectra were taken on a Varian A-60 instrument; infrared spectra were taken on a Baird-Atomic spectrophotometer; and ultraviolet spectra were taken on a Beckman DK-SA. Elemental analyses were performed by G. I. Robertson, Jr., Florsham Park, N. J., and Spang Microanalytical Laboratory, Ann Arbor, Mich. 4-Cyano-3-hydroxy-l-methyl-5,6,7,8-tetrahydroisoquinoline (IV). A. From Cyanoacetamide.-A mixture of 4.6 g (0.05 mole) of 2-acetylcyclohexanone,8 1.5 g (0.017 mol) of cyanoacetamide (Aldrich), 50 ml of 95% ethanol, and 0.5 ml of diethylamine (Eastman) was stirred for 18 hr a t room temperature. The reaction mixture was filtered and the solid was washed with dilute alcohol and dried to give 2.8 g of IV. The filtrate and the washings gave an additional 0.2 g after standing overnight (crude yield of 90%). Recrystallization from absolute ethyl alcohol gave 2.81 g (85%) of pale yellow needles: mp 357-359" dec; ir (KBr), (isoquinoline); uv 4.45 (CN), 6.05,6.21 (C=O c)C-OH), 8 . 7 4 ~ max (95% EtOH), 218 mp (e 18,824), 223 (19,049), 238 (6777), 343 (11,140); nmr (CF8COOH), 7 , 6.98, (m, 2, CH), 7.35 (m, 2, CH), 7.48 (s, 3, CHs), 8.16 (m, 4, CHI). Anal. Calcd for CllHlaNzO: C, 70.19; H, 6.43; N, 14.88. Found: C,70.15; H, 6.36; N, 14.70. B. From Malononitrile.-The procedure was the same as above except that the reaction time was 30 min and the yield was 27%. A mixture melting point determination showed t h s product to be identical with the product obtained with cyanoacetamide. 3-Chloro-4-cyano-l-methyl-5,6,7,8-tetrahydroisoq~oline (VI) was prepared by the procedure of Mariella and Leech.10 Recrystallization of the solid from dilute ethanol gave white needles: mp 100-100.5"; ir (KBr), 4.42 (CN), 8.75 (isoquinoline), 13.59 (CC1); nmr (CDCb), 7 7.10 (m, 2, CH), 7.38 (m, 2, CH), 7.55 (9, 3, CH,), 8.16 (m, 4, CHI). Anal. Calcd for ClIHI1NIC1: C, 63.93; H, 5.36; N, 13.55. Found: C, 63.91; H, 5.24; N, 13.60. 4-Carboxy-3-hydroxy-1-methyl-5,6,7,8-tetrahydroisoquinoline (VII) and 3-Hydroxy-1-methyl-5,6,7,8-tetrahydroisoquinoline (VIII).-A solution of 4.0 g (0.021 mol) of IV in 150 ml of 25% potassium hydroxide solution was refluxed for 5 days, cooled, and extracted continuously with ether for 48 hr to remove VIII. The ether was removed through a 14in. Vigreaux column, and the residue was dissolved in 95% ethanol, treated with Norit, filtered, and cooled to give 0.7 g (21%) of VIII: mp 236-237'; ir (KBr), 8.74 p (isoquinoline).
(9) R. M. Manyik, F. C. Frostick, Jr., J. J. Sandstron, and C. R . Hauser, J . Amer. Chem. Soc., 7 5 , 5030 (1953). (10) R. P. Mariella and J. L. Leech, ibid., 71, 331 (1949).

3650 NOTES
Anal. Calcd for CioHisNO: C, 73.59; H, 8.03; N, 8.53. Found: C, 73.66; H, 8.04; N, 8.55. The residue from the continuous ether extraction was made acid to congo red with dilute hydrochloric acid and the volume was doubled by the addition of water. The precipitated solid was filtered, dried, transferred to a Soxhlet extractor, and extracted with chloroform for 7 days. Evaporation of the chloroform gave a white solid which was dissolved in absolute ethyl alcohol, treated with Norit, filtered, and cooled to give 2.1 g (50%) of VII: mp 259-260" dec, ir (KBr), 8.75 p (isoquinoline). Anal. Calcd for ClIH13N03: C, 63.76; H, 6.32; N, 6.76. Found: C, 63.52; H, 6.49; N, 6.66. VI11 from VI1.-Compound VI1 was decarboxylated a t 264" via the procedure of Hart and Freeman.' Recrystallization from ethanol gave 0.6 g (73.6%) of white needles of VIII, mp 236237". The melting point was undepressed on admixture with the hydrolytic product (vide supra).

The Journal of Organic Chemistry

for preparing 8-methyl-2,7-cyclodecadien-l-one (4). One possibility appeared to be pyrolysis of the readily available tricyclic ketone 3.6

FQ?'@
0
3
4

When compound 3 was heated at 360" in a sealed Pyrex vessel, two isomeric products were formed in varying amounts (Table I). The major product was
PYROLYSIS
Run

Registry No.-I, 107-91-5; 11,874-23-7; IVb, 1701230-5; V, 109-77-3; VI, 17012-31-6; VII, 17012-32-7; VIII, 17012-33-8. Acknowledgment.-We are indebted to Dr. E. C. Friedrich" for some of the nmr spectra and to Mr. R. H. DuBois for certain of the infrared spectra.
(11) Department of Chemistry, University of California at Davis, Davis, Calif.

OF

TABLE I ~-METHYLTRICYCLO DECAN-&NE [4.4.o.Ozr']

-Reaction 3 mixture analyeis, %6 Other product

Amount of 3, mg

An Unusual Cyclobutane Pyrolysis

CLAYTON HEA'PHCOCK' AND BRUCE RATCLIFFE H. E.

1 2 3 4 5 6 7 8 5 Analysis by products.

527.5 70.0 24.2 523.4 69.2 24.4 538.9 70.0 23.1 65.9 517.1 27.2 527.9 66.2 25.7 534.0 66.6 25.6 68.0 547.6 24.3 553.2 67.9 24.4 internal standard showed 24.6% of

5.6 6.3 6.8 6.8 8.0 7.8 8.1 7.7a nonvolatile

Department of Chemistry, University of California at Berkeley, Berkeley, California 94720 Received April 23, 1968

The thermal decomposition of cyclobutane to ethylene2 is believed to involve the intermediacy of the tetramethylene diradical, which undergoes /3 scission to yield products (eq l).3 Numerous subCHz-CHz CHz-CHz

identified as the octalone 5 , by comparison with an authentic ample.^ The minor product was shown to be a cyclohexanone or cycloheptanone (Y,,, 1700 cm-') containing a terminal methylene group (vu= 890 cm-l). Due to the extreme difficulty in isolating this minor product in a pure state, its structure was not further investigated. The probable mode of formation of 5 is outlined in eq 2. The intramolecular diversion of the postulated

-I

CH2-CHz CHz-CHz
--f

CHz=CHz CHz=CHz
(1)

stituted cyclobutanes have been pyrolyzed and almost all undergo the reaction outlined in eq 1, the few exceptions being vinyl and 1,2-divinylcyclobutanes, which rearrange to cyclohexenes and cyclooctadienes, respectively. Srinivasan and Levi4 found that tricyclo [3.3.0.0296 1octane (1) undergoes thermal decomposition in the expected manner to yield 1,5-cyclooctadine (2). In connection with another study, we desired a method

3a

Q-0
U
2
(1) Fellow of the Alfred P. Sloan Foundation. (2) (a) G . T. Genaux and W. D. Walters, J . Amer. Chem. SOC., 78, 4497 (1951); (b) F. Kern and W. D. Walters, Proc. Nat. Acad. Sci. U.S., 89, 937 (1952); (0) F. Kern and W. D. Walters, J . Amer. Chem. Soc., 7 6 , 6196 (1953). (3) H. M. Frey, "Gas Phase Pyrolyses of Some Small Ring Hydrocarbons," in Advances in Physical Organic Chemistry, Vol. 4, V. Gold, Ed., Academic Press, New York, N. Y . , 1966. (4) R. Elrinivasan and A. A. Levi, J . Amer. Chem. SOC.. 3750 (1964). 86,

diradical 3a lends additional credence to the probable intermediacy of tetramethylene diradicals in cyclobutane pyrolyses.a The reason for the different modes of reaction of tricyclic compounds 1 and 3 is not clear.
Experimental Section Pyrolysis of 1-Methyltricyclo [4.4.0 .02q decan-&one (3).-Approximately 500 mg of ketone 36 was sealed at 1 mm in a 20-ml Pyrex vessel. The tube WBS heated a t 360" for 2 hr, cooled, and opened. Analysis of the reaction mixture by capillary vpc (150
(5) C. H. Heathcook, R. A. Badger, and J. W. Patterson, Jr., i b i d . , 89, 4135 (1967).