An Improvement in the Preparation of Acids via ... - ACS Publications

622

NOTES

VOL.

27

Notes An Improvement in the Preparation of Acids via Acetoacetic Ester JOHNJ. RITTER* AND THADDEUS J. KANIECKPJ Received August 31, 1961

The preparation of acids RCHzCOzH by the acetoacetic ester route is known to result usually in only fair or even poor yields because of the competing and usually preferred reaction to form methyl ketones. Dieckmann' reported the alcoholysis of &alkyl acetoacetic esters in presence of ethoxide to result in high yields of the products of acid cleavage:

+

CHICOCRZCO~C~H~ OCZHs----t 0CHaLCRzCO2CzHs +CHaCOzCzHs

-

Ac~H~

+ +

+

: C R ~ C O ~ C Z H ~CzHbOH ; ---t R~CHCOZCZHSOCzHs-

We have now found that the alcoholysis of monoalkyl acetoacetic esters, reported by Dieckmann and others to have failed, may be accomplished in (1) Present address: Evans Research and Development Corp., New York, N. Y. (2) Abstracted from a portion of the thesis submitted by Thaddeus J. Kaniecki to the Graduate Faculty of New York University in partial fulfillment of the requirements for the degree of Doctor of Philosophy. (3) Present address: Lever Brothers, Edgewater, N. J. (4) W. Dieckmann, Ber., 33, 2670 (1900); W. Dieckmann and A. Kron, Ber., 41, 1260 (1908). See also: W. M. Kutz and H. Adkins, J. A m . Chem. SOC.,5 2 , 4391 (1930); R. Connor and H. Adkins, J. A m . Chem. SOC.,54, 3420 (1932); L. J. Beckhad and H. Adkins, J. A m . Chem. SOC.,56, 1119 (1934); J. Finkelstein and R. C. Elderfield, J. Ot-g. Chem., 4, 365 (L939); W. B. Renfrow and G. B. Walker, J. Am. Chem. SOC.,70,245 (1956).

excellent yields by heating with catalytic amounts of alkoxide in excess of absolute ethanol. The complete absence of products of ketonic cleavage in either case is to be anticipated, as alkoxide attack on the ester carbonyl group would result only in ester exchange and would lead to no significant chemical alteration of this group. The success of the method reported here appesn to depend upon scrupulous care in the preparation of the antecedent materials, which must be dry and free of acetoacetic ester. Reaction is accomplished under reflux with continuous removal of ethyl acetate as the ethanol-ethyl acetate azeotrope. The reaction time varied in ten cases examined (Table I) from ten to twenty-four hours. Continuous removal of ethyl acetate was performed in early experiments under the assumption that this would aid in equilibrium shift and also possibly minimize obvious side reactions. However, a control experiment using a-allylacetoacetic ester without removal of ethyl acetate resulted in fair (70%) yield of ethyl 4-pentenoate in contrast to 88% obtained with continuous removal. It would appear that the principal value of ethyl acetate removal is simply that of avoiding temperature drop in the reaction vessel. This procedure was used throughout. It seems evident that the reaction begins with alkoxide attack a t the keto-carbonyl group, 5t9 shown above. The presence of water or acetoacetic ester would be expected to interfere by destruction of the alkoxide catalyst. The survival of the catalyst in the presence of monosubstituted acetoacetic esters might be due to the diminution of acidity of the alpha-hydrogen atom resulting from alphaalkyl substitution. It may be noted also that alkylideneacetoacetic esters, CH&OC(=CHR)C0&2Hs,

TABLE I

+

OQHs-

CH&OCHRCO~C~HI, C2HsOH

R

Ester0

GHS n-CsHr CaHs C'H, (methallyl) *C& i-CsHii ~-CIOHZI csHsc& C S E ~ C S(cinnamyl) H~ ClH702 (carbethoxymethyl)

1 1 1 1 1 1 1 1 1 1

__f

CHaC02CzHs

+ RCHzC02CzHs

C2HSOHo

Naa

Moles Used Ester

3.5 3.5 2.5 3.9 2.9 3.5 7.0 3.9 3.9 4.5

0.13 0.13 0.10 0.15 0.15 0.15 0.12 0.15 0.13 0.16

0.50 0.50 0.72 0.45 0.43 0.50 0.25 0.50 0.23 0.40

Yield, % ' 70b 82 88

83 81 88 84

84 92 84

B.P./P. 120-121/atm. 145-146/atm. 142-144/atm. 165-167/atm. 16&168/atm. 181-182/atm. 106106/1.5 mm. 93-95/2 IIUXI. 115-116/1 mm. 215-217/atm.

0 E ster:CaHbOH: Na is on 1mole ester basis. b Methyl and ethyl acetoacetic ester gave poorer yields than the higher homologs, apparently because of the difficu1t.yof their separation from acetoacetic eater by distillation.

FEBRUARY

1962

obtained by Knoevenagel condensation of crotonaldehyde and benzaldehyde with acetoacetic ester yielded only polymeric materials instead of the anticipated sorbic and cinnamic esters. It is suggested that allylic resonance may account for the failure of such esters to undergo the cleavage: 0-

C=CHR @

623

NOTES

I @ +C H s C b C C H R

~OIGHS

JOGH,

EXPERIMENTAL

General procedure. The same procedure was used for all compounds shown in Table I, except for variations in time required for completion of the reaction. Preparative details are given only for the alcoholysis of ethyl a-allylacetoacetate to ethyl 4-pentenoate: To a solution of 1.5 g. (0.065 g.-atom) of oxide-free sodium in 81 g. (1.76 moles) of absolute ethanol in a 500-ml. round bottomed flask waa added 122 g. (0.72 mole) of freshly distilled ethyl a-allylacetoacetate. The resulting light yellow solution waa refluxed under a 5-ft. distilling column packed with glaas helices and fitted with an automatic liquid-divider head. When the vapor temperature at the column head had dropped from 78' to 72' (the boiling point of the ethanol-ethyl acetate azeotrope), slow distillation was begun and the takeoff ratio waa adjusted to maintain the vapor temperature at 72-73'. After the calculated amount of azeotrope had been collected (22 hr.), the distillation waa stopped and the excess ethanol waa removed on the water bath. The f l a k contents were cooled, transferred to a separatory funnel, and washed with -100 ml. of cold 5% sulfuric acid. After separation of the ester layer the aqueous portion was washed twice with 25-ml. portions of ether, and the ether extracts were combined with the main portion of the ester. The resulting ether solution waa waahed successively with 50 ml. of water, 50 ml. of 27&sodium bicarbonate solution, and 50 ml. of water, and dried over anhydrous magnesium sulfate. The dry ethereal solution waa filtered from the drying agent, the ether waa removed and the product, ethyl Cpentenoate, waa distilled at ordinary pressure and collected a t 142-144'; yield, 81 g. (%yo).The still residue consisted mainly of unchanged ethyl a-allylacetoacetate.

Acknowledgment. The authors wish to express their thanks to the Food Machinery and Chemical Corp. for financial support of this investigation and also for providing generous quantities of necessary materials.

a t the allylic position. The yields in the N-bromot-butylamine reaction under certain conditions are comparable to those of the N-bromosuccinimide reaction. * I n the study of the allylic brominating ability of N-bromo-t-butylamine, cyclohexene was used as the olefin, the concentrations of the reagents and the type of solvent were varied, and the effect of azobisisobutyronitrile (AIBN) was studied. Light seemed to be inferior to ATBN as an initiator, and therefore was not investigated further. Our results appear in Table I. TABLE I THE EFFECTS OF V~RIABLES ON BROMO-&BUTYLAMINE AS -~

AN

THE REACTIONOF NALLYLICBROMINATING AGENT

~

Solvent Carbon tetrachloride Isohexane Isohexane Isohexme Isohexane Isohexane Isohexane Isohexane Benzene Benzene Benzene

Ratio of Moles N-Bromot-butjdamine/ Moles Olefin/ Catalyst or M1. Solvent Initiator

1/1.5/750 1/1.5/750 1/1.5/750 1/2/750 1/3/750 1/3/375 1/3/187.5 1/1.5/375 1/3/750 1/3/375 3/1/750

None Light" 0.2 g. AIBN 1 g.AIBN 1 g.AIBN 1 g.AIBN 1 g.AIBN 1 g.AIBN 1 g.AIBN 1 g.AIBN 1 g.AIBN

Time (Hr.)

6 6.5 3 5.5 5 3 1.5 5 5 5 5

Yield of 3-Bromocyclohexene

%" noneb

Md 23 38 37 59 31 40 30 42 17'

Yield of 3-bromocyclohexene baaed on N-bromot butylamine. b There appeared to be a competing reaction involving the carbon tetrachloride. 0 Light was supplied by sun lamp. Light seemed to favor the formation of tbutylamine hydrobromide. d All yields of 3-bromocyclohexene were corrected for the amount of product lost in working up the sample, using a technique similar to that of H. J. Dauben, Jr., and L. L. McCoy.* t. 22% 3,6-Dibromocyclohexene was produced; collected 95-100' (10 mm.). EXPERIMENTAL

Received August 14, 1961

Preparation of N-brmo-t-lmtylamine.a t-Butylamine, 0.2 mole, (freshly distilled) was mixed with 20 ml. of 10N sodium hydroxide and 100 ml. of water. Bromine, 0.2 mole, waa added dropwise over a period of 50-70 min. The reaction waa kept at G 5 " and was stirred constantly. After the bromine waa added, the sample waa extracted with ether, and the ether layer dried over magnesium sulfate. The ether then was removed by means of an aspirator. The N-bromdbutylamine remained in almost quantitative yield, about 30 g. It is a deep red-orange liquid with a strong unpleasant odor. It should be prepared immediately before use since it decomposes on standing with the formation of t-butylamine hydrobromide, a white crystalline solid.

Currently allylic bromination is most frequently accomplished through the use of N-bromosuccinimide.2 We have discovered that N-bromo-t-butylamine also reacts with olefins to give bromination

(1) This work waa supported by an undergraduate research grant (G-15672) from the National Science Foundation. (2) See H. J. Dauben, Jr., and L. L. McCoy, J. Am. C h .Soc., 81,4863 (1959). (3) See N. Kijner, J. prakt. Chem., 172, 64 (1901).

DEPARTMENTOF C H E ~ S T R Y NEW YORKUNIVERSIW NEW YORK,N. Y .

Allylic Bromination by N-Bromo- t butylamine' C. E. BOOZERAND J. W. MONCRIEF

624

NOTES

Bromination of cyclohexae. Cyclohexene, solvent, AIBN (when used), and N-bromo-t-butylamine were mixed in that order and in the ratios given in Table I. (About 0.16 mole of N-bromo-t-butylamine was used in most runs.) The mixture was distilled a t a reflux ratio of about 1:5 until a negative starch-iodide paper test for N-bromoamine was obtained. This was done to remove the t-butylamine as it was formed. Additional solvent was added at intervals to keep the volume of the reaction mixture approximately the same. Varying amounts of t-butylamine hydrobromide were formed in the reaction, but no quantitative measurenienta were made of it. After refluxing, the solid t-butylamine hydrobromide was filtered off, and the liquid was washed first with 10N sodium hydroxide and then with 6N hydrochloric acid. The nonaqueous layer then was dried and distilled, the 3-bromocyclohexene being collected a t 48-51’ at 10 mm. Further confirmation of the structure of the product was obtained through NMR and infrared analysis. The NMR spectrum had peaks a t 4.25 7 (2 vinyl protons), 5.32 T (allylic proton on brominated carbon), and a broad unresolved peak around 7.9 T (other protone). DEPARTMENT OF CHEMISTRY EMORY UNIVERSITY ATLANTA22, GA.

A Synthetic Procedure €or Secondary Bromides from Alcohols GLENNL. JENKINS A N D JAMES C. KELLETT, JR1 Received August 9 1961

In connection with other researches in these laboratories we desired to prepare certain substituted cyclopentyl bromides from the corresponding alcohols. Several workers have reported that halogenation of secondary alcohols with “conventional” reagents (PXa, HX) results in rearrangement of the p r o d ~ c t . ~As J the rearranging species is evidently a carbonium ion, an effort was made by Pines, Rudin, and Ipatieff to employ a bimolecular displacement of the p-toluenesulfonate ester of the alcohol with bromide The reaction conditions employed in their research resulted in nonrearranged bromides but were, by the admission of the authors, not synthetically feasible. By the use of a homogeneous system consisting of the p-toluenesulfonate ester of the requisite alcohol, aqueous dimethylformamide, and calcium bromide a t room temperature, we have been able to prepare secondary bromides in 5045% yields. Such conditions are expected to result in non-rearranged bromides. The course of the reaction with this system may be easily followed by quenching a 1-ml. aliquot of the (1) Present address: School of Pharmacy, University of North Carolina, Chapel Hill, N. C. (2) F. C. Whitmore and F. Johnston, J. Am. Chem. Soc., 60, 2265 (1938). (3) F. C. Whitmore and F. A. Kornatc, J. Am. Chem. SOC., 60, 2536 (1938). (4) H. Pines, A. Rudin, and V. N Ipatieff, J Am. Chem. SOC.,74, 4063 (1952).

VOL.

27

reaction mixture in 50 ml. of water, extracting the water with three 50-ml. portions of ether, removing the last traces of ether on a steam bath, and titrating the residual bromide ion with standard silver nitrate in the presence of fifteen drops of Dichlorofluorescein test solution and 1 ml. of 2% dextrin solution. Precautions must be taken to prevent the interaction of the alkyl bromide produced in the reaction and the dimethylformamide solvent. This interaction has been previously ~ t u d i e d The . ~ reaction between the alkyl bromide and dimethylformamide is sufficiently slow a t room temperature not to interfere with the synthesis of the halide; however, attempted distillation of a mixture of dimethylformamide with an alkyl bromide results in extensive decomposition. The work-up procedure results in a solution of the alkyl bromide with traces of dimethylformamide in petroleum ether; the dimethylformamide may be readily removed by passing this solution through alumina. As would be expected, the procedure is not applicable to alcohols sterically hindered to nucleophilic attack. Thus 2-methylcyclopentyl bromide was ohtained only in poor yields while 3-methyl- or 3ethylcyclopentyl bromide was obtained in good yields. EXPERIMENTAL^ Cyclopentyl bromide. Although rearrangement of this compound is undetectable, it was prepared in a study of solvents and reaction conditions before the application of these conditions to various substituted cyclopentyl bromides. Onefourth mole (21.5 8.) of cyclopentanol was esterified with ptoluenesulfonyl chloride according to the procedure of Streitweiser’ except that the product was extracted with ether, the combined extracts were dried, and the solvent was removed under reduced pressure. The crude ester wae then stirred a t room temperature for 12 hr. with a solution of 120 g. of calcium bromide (N.F.) in 600 ml. of dimethylformamide containing 1.5% water. The reaction mixture was then poured into ice water, the product which separated was removed, and the aqueous layer was extracted with 125 ml. of petroleum ether. The combined extract and product. after drying over potassium carbonate, was then pasped through a 50 X 2 cm. column containing ca. 100 g. of alumina which had been wetted with petroleum ether. The column was then e!uted with 100 ml. of petroleum ethw. The combined eluates were distilled to yield 19.0 g. (51.0%) of cyclopentyl bromide boiling a t 134.0-136.0’ and having n*: 1.4563; reported8 boiling at 135-136’ and having TI*: 1.4882. 3-Methylcyclopentyl bromide. In a similar procedure 0.7 mole (70.0 9.) of 3-methylcyclopentanol was converted into 73.0 g. (64.0%) of 3-methylcyclopentyl bromide boiling at 146.0-151.0° and having n’,”1.4762. (5) N. Kornblum and R. K. Blackwood, J. Am. Chem. SOC., 78, 4037 (1956). (6) All boiling points are uncorrected. Microanalyses were performed by Alfred Bernhardt, Mikroanalytisches Laboratorium in Max-Planck Institut, Mulheim (Ruhr), Germany. (7) A. Streitweiser et at., J. Am. Chem. SOC., 80, 2326 (1958). (8) C. R Noller and R Adams. J. Am. Chem. SOC.,48, 1080 (1926).

FEBRUARY 1562

Anal. Calcd. for CsHlzBr: C, 44.2; H, 6.8; Br, 49.0. Found:

C, 44.8; H, 6.9; Br, 48.2.

3-Ethylcyclopentyl bromide. In a similar procedure 0.62 mole (71.0 g.) of 3-ethylcyclopentanol was converted into 65.0 g. (59.3'34,) of 3-ethylcyclopentyl bromide boiling a t 172.0-177.0' and having na. 1.4780. Anal. Calcd. for CTHlrBr: C, 47.5; H, 7.4; Br, 45.1. Found: C, 48.6; H, 7.5;Br, 45.0. 2-Methylcyclopentyl bromide. I-Methylcyclopentanol, 1.19 moles (119.0 g.), was treated with equivalent quantities of the reagents described in the preceding experiment. The displacement was allowed to proceed for 24 hours. Workup vielded 36.0 g. ( 18.670) of product boiling at 45.5-49.5' a t 20 mm. and having n'," 1.4757; reportede boiling a t 150-151°. SCHOOL OF PHARMACY PURDTJE UNIVERSITY m E S T LIFAYETTE, I N D . (9) E. Buchta and S. Dauner, Ber., 81, 247 (1948).

Reaction of Ditolylethane with Gallium Bromide-Hydrogen Bromide in Benzene A. STREITWIESER, JR.,AND W. J. DOWNS Received July i7,1961

On the basis of stereochemical and isotope tracer experiments a new mechanism has recently been proposed for the Lewis acid catalyzed transalkylation of ethylbenzene in benzene.I In this mechanism a small amount of oxidation to a-phenethyl cation initiates a carbonium ion chain process: The aphenethyl cation alkylates benzene to form 1,ldiphenylethane, which is rapidly cleaved by acid under the experimental conditions to regenerate an a-phenethyl cation and benzene in which the two aromatic rings may have been interchanged. This mechanism has an obvious extension to related disproportionation reactions; it has the further necessary corollary that 1,l-diarylethanes react rapidly under these conditions, To test this corollary, a solution of 1,l-di-p-tolylethane in benzene was treated with gallium bromide and hydrogen bromide a t 50'. Aliquots of the mixture were examined at intervals by V.P.C. analysis after quenching with water. Even by the time the first aliquot was removed (15 seconds), the ditolylethane WBS converted completely to 1 l-diphenylethane. Toluene and a lesser amount of ethylbenzene were also identified as products. Hence, the reaction Ar&HCH1

625

NOTES

HBr GaBn

+ 2 CsHs e (CsH6)&HCHs + 2 ArH

phenethyl catioi with concommitant formation of other by-products. Small amounts of other compounds were found by V.P.C. but could not be identified. EXPERIMENTAL

The experimental technique was similar to that used in the earlier report.' A stock solution was prepared from 8.5 g. of sublimed gallium bromide and 101 g. of sodium-dried benzene and stored in a flask carrying a side arm closed with a serum cap. Fifty milliliters of the stock solution was transferred with a syringe to a one necked flask closed with a stopcock and a serum cap. Hydrogen bromide (0.49 g.) was admitted with a syringe needle and the flask wau brought to temperature in a 50" thermostat. 1,l-Di-p-tolylethane4 (0.50 ml.) was syringed in and 10 ml. aliquots were removed after. 15 sec., 5 min., and 1.5 hr. Each aliquot was quenched with, water and the organic layer was separated, dried, and examined by V.P.C. (7O0,tsilicone). In each aliquot, the toluene peak was 1%of the benzene peak; ethylbenzene increaseAfrom '/, during the run. A sample analyzed in the V.P.C. column a t 200' showed six additional peaks, the largest of which was 1,l-diphenylethane. The remaining peaks were small and could not be identified; however, 1,lditolylethane was found to be absent. DEPARTMENT OF CHEMISTRY, OF CALIFORNIA UNIVERSITY BERKELEY 4, CALIF. (2) J. S. Reichert and J. A. Nieuwland, J. Am, Chem. Soc., 45,3OYO (1923).

A Reappraisal Concerning the Variable Character of the Sulfone Group' CALY. MEYERS,GIANFRANCO MORETTI,AND LILLIAMAIOLI Received August 2, 1961

Szmant and Suld2 reported that, in benzoic acids of type I, a 4-N02 group decreased the acidity of the parent sulfone derivative but increased that of the corresponding sulfoxide and sulfide derivative, respectively. However,,similar substitution of the phenolic sulfone (11. X = SO,) increased its acidity. In both sulfone series a 4-NH2 group decreased the acidity, respectively. W I. R

+

CO,H; 11. R

R OH; X = SOz, SO, S

While it was noted that the decreased acidity of the 4-K0~-phenylsulfonylbenzoic acid was "entirely unexpected on the basis of additive inductive effects," the combined data were utilized in suggesting that the 4302- group varies in character depending upon the electronic nature of the

is orders of magnitude faster than the transalkylation studied earlier and is an allowable sequence as required in the proposed reaction mechanism. The (1) This study is part of a series dealing with the nature ethylbenzene also produced in the experiment un- of the sulfone group. The authors are grateful to the Pedoubtedly arises by hydride transfer to the a- troleum Research Fund of the American Chemical Society whose grants are making these studies possible. (1) A. Streitwieser, Jr., and L. Reif, J. Am. Chem. SOC., (2) H. H. Szmant and G. Suld, J. Am. Chem. SOC.,78, 82,5003 (1960).

3400 (1956).

626

NOTES

a? os -

\

0

\

\

0.C

a3

27

R=

0.8

0.4

VOL.

\ i

\ \ \

-

\ \ \ G

0.0

-1

0.1

0.4

-

\

0

\ \ \ \

-

\

0.1 a5 0.3

0.6

-

0.7

-

\ \

\ \ \

\

\

0 I

Fig. 1. A. Effect of u of R on pKa of p-(RCs&SOz)Cs&COzH, -in 80% aq. acetone; - - - - in 48% aq. in 48% aq. ethanol (see Table 11) ethanol (see Table I). B. Effect of u on R of pK, of RC~&SOZCHZCOZH

substituent--e.g., a NOz group decreases the posi-

the acidities of substituted phenylsulfonylacetic acids were also determined, but directly in 48%

tive charge at the S atom (04-0) by stabilizing

TABLE I SUBSTITUENTEFFECTS ON THE ACIDITY OF P-PIIENYLSULFONYLBENZOIC ACID

- ++I -

the double-bonded form

I I (0,5=O).2J I

During several attempts to confirm the anomalous acidity of this 4-NOzderivative, we also studied the 3-N02 isomer. Surprisingly, we found that neither compound was sufficiently soluble in 48% aqueous ethanol to allow pKa determinations under the conditions described by Szmant. It was found, however, that 80% aqueous acetone was very satisfactory and allowed relative acidities to be determined for the series. In addition to the 4 N H 2 derivative, we also studied the 3-NH2 isomer. The data (Table I and Fig. 1A) indicate that the group on the benzoic influence of the -SO2acid acidity is directly related t o the electronic nature of the substituent group measured in terms of Hainmett’s (r constants. The corresponding pK’s in 48% aqueous ethanol were then interpolated from the plots. The relative acidity of the 4 N 0 ~derivative is in sharp contrast to that reported by Szmant. To substantiate this observed relationship, (3) H. H. Szmant and J. M. Dixon, 78,4384 (1956).

J. Am. Chem. Soc.,

U PKa in 48% Aq. ethanol (of Substit- 80% Aq. Substituent uentIa Acetoneb Interpol’dc Reportedd

4NOZ 3-NOp H 3-NH2 4NHz

+0.778 $0.710 0.000 -0,161 -0.660

5.64 5.68 5.83 5.85 5.97

4.44 4.45 4.63 4.68 4.80

4.86

-

4.63

-

4.80

a J. Hine, Physical Organic Chemistry, McGraw-Hill, New York, 1956, p. 72. Determined (in triplicate) by titration of 0.002M solutions with 0.01M NaOH in 48% aqueous ethanol, using a Beckman Model G pH meter; values corrected where necessary by use of the Henderson Equation (S. Glasstone, Physical Chemistry, D. Van Nostrand Co., Inc., London, 1947, pp. 1003, ff.). Approximated values from Fig. 1A. Ref. 2.

aqueous ethanol. Again the data showed that the group is directly related influence of the -SO2t o the substituent’s u constant (Table I1 and Fig. 1B). As, in both types of acids, the degree of acidity increases with the corresponding electronegativity of the -SOTgroup, this electronegativity apparently is increased proportionally by 3- and 4 N 0 2 substitution, respectively, in both cases. This is contrary to the general conclusion of

FEBRUARY 1962

627

NOTES

methoxyphenols, isolated. Less reactive monocyclic aryl ethers such as a n i ~ o l e l -were ~ reported to remain unchanged or to give..water-soluble unidentified oils. A recent publication4on the oxidation of aromatic TABLE I1 hydrocarbons with trifluoroperoxyacetic acid SUBSTITUENT EFFECTS ON THE ACIDITYOF PHENYLSUL prompts us to report some results of our own. FONYLACETIC ACID Trifluoroperoxyacetic acid was found to oxidize anisole and diphenyl ether primarily by a process PKa of electrophilic monohydroxylation. Moderate U in 4870 Substituent (of Substituent)a Aq. Ethanolb yields of phenols were obtained when an equimolar amount of peroxy acid was added slowly over sev+O. 778 3.29 4N01 eral hours to a solution of the aryl ether in methyl3-NOz +O. 710 3.45 0.000 3.66 H ene chloride a t 15-25 '. Anisole (44% conversion) 4CHa -0.170 3.72 gave o-methoxyphenol (Ia) in 27% yield and p4OCHs -0.268 3.79 methoxyphenol (Ha) in 7% yield. From diphenyl 40H -0.357 3.84

Szmant. Moreover, a plot of sulfonylphenol acidities2 against u reveals a relationship parallel to those reported here for the corresponding benzoic and acetic acids.

Ref. a in Table I. Ref. b in Table I.

OR

Infrared studies of aryl sulfones4indicate that the S-0 force constant also varies regularly with the Q of the substituent, resembling analogous observations with ketone^.^ As both this relationship and that describing the relative acidities within the three series of sulfone derivatives are valid in each case even a t either extreme of substituent electronegativity, true variation in the character of the sulfone group, a t least to the extent proposed by Szmant, is not evident.

6"" Ia. R = CH, b. R = CeHs

OR

I

OH IIa.R=CHJ b. R=CeHs

ether (44% conversion) o-phenoxyphenol (Ib) and p-phenoxyphenol (IIb) were obtained in 35% and 12% yields, respectively. In neither case were we able to isolate any of the corresponding meta isoISTITUWJ DI CHIMICA INDUSTRULE mer. DI BOLOGNA UNIVERSIT~ BOLOGNA, ITALY It was of interest that the phenols isolated from the oxidations of both anisole and diphenyl ether were predominantly the ortho isomers. The modest (4) C. Y. Meyers, 140th Meeting, Am. Chem. SOC., total yields (34r47%) were explained by noting Chicago, Ill., Sept. 1961, paper No. 12, Div. of Phys. that the phenols underwent further rapid oxidaChem. tion, although no specific products were isolated. (5) N. Fuson, M.-L. Josien, and E. M. Shelton, J. A m . There was considerable doubt as to whether the Chem. Soc., 76,2526 (1954). observed ortho to para ratio (3:1) was due to a truly selective. process or was merely the consequence of a rate of subsequent oxidation greater The Oxidation of Anisole and Diphenyl Ether for the para isomer than for the ortho. In order to determine which of these alternatives was correct, with Trifluoroperoxyacetic Acid an equimolar mixture of o-methoxyphenol and pmethoxyphenol was treated with a one-half molar JAMESD. MCCLURE AND PAULH. WILLIAMS quantity of trifluoroperoxyacetic acid. Analysis of the unchanged methoxyphenol showed that approxReeeived August 7, 1961 imately twice as much para as ortho isomer had Friess' and co-workers studied the reactions of been consumed. The apparent selectivity of the peroxybenzoic acid with the methyl ethers of di- hydroxylation of anisole can thus be accounted for and trihydric monocyclic phenols in chloroform to a large extent by the more rapid oxidation of psolution. 1,4-&uinones were usually formed, in methoxyphenol. some cases with the loss of a methoxy group. More recently, the action of peroxyacetic acid2on a (1) S. L. Friess, A. H. Soloway, B. K. Morse, and W. C. number of substituted di- and trimethoxybenzenes Ingersoll, J . Am. Chem. Soc., 74, 1305 (1952). (2) H. Davidge, A. G. Davies, J. Kenyon, and R. F. has been examined. Once again l,Pquinones, formed by electrophilic 1,4-dihydroxylation and Mason, J . Chem. Soc., 4569 (1958). (3) H. Fernhole, Angew. Chem., 60A, 62 (1948); Ber., methoxy elimination, were usually the principal 84,110 (1951). products obtained. With neither peroxy acid were (4) R. D. Chambers, P. Goggin, and W. K. R. Musgrave, the presumed primary products of oxidation, J . Chem. Soc., 1804 (1959).

628

VOL. 27

NOTES

The Decarbonylation of a-Anilino-a,adiphenylacetic Acid by p-Toluenesulfonyl Chloride and Pyridine

EXPERIMENTAL Oxidation of anisole. A solution of trifluoroperoxyacetic acid6 (110 ml., 2.1 M ) in methylene chloride was added dropwise over a two-hour period to a stirred solution of anisole (21.6 g.; 0.2 mole) in 100 ml. of methylene chloride maintained a t 15-20'. Stirring was continued a t 25" for thirty minutes a t which time 95% of the peroxy acid had been consumed. The mixture was diluted with methylene chloride and washed with a 10% sodium bicarbonate solution. After drying the solvent was removed by distillation. The residual liquid was distilled through a semimicro Vigreux column to give anisole, 12.0 g., b.p. 51-53' (20 mm.) and fraction 2, 4.0g.,.b.p.55-90°(3mm.). Analysis of fraction 2 by gas-liquid chromatography at 190' on a column packed with DC 710 on firebrick indicated the presence of three components. Comparison of the chromatogram with those of authentic specimens showed that the constituents were anisole (4%), o-methoxyphenol (7377%), and p-methoxyphenol(18-23%). Fractional distillation of fraction 2 through a micro spinning band column at 4 mm. pressure yielded o-methoxyphenol, b.p. 53-55', and p-methoxyphenol, b.p. 85-87'. The latter isomer was recrystallized from methylene chloride-hexane to give white plates, m.p. and mixed m.p. 53-54'. Anal. Calcd. for CqHaO,: C, 67.7; H, 6.48; phenolic acidity, 0.805 eq./100 g. Found for o-methoxyphenol: C, 67.5; H, 6.51; phenolic acidity, 0.80 eq./100 g. Found for p-methoxyphenol: C, 67.6; H, 6.53; phenolic acidity, 0.81 eq./100 g. Oxidation of diphenyl ether. Trifluoroperoxyacetic acid in methylene chloride (110 ml. of 2.1 M solution) was added dropwise over a ninety-minute period to a stirred solution of diphenyl ether (34.4 g., 0.2 mole) in 120 ml. of methylene chloride a t 20-25'. Fifteen minutes after the addition was complete 95% of the peroxy acid had been consumed. The dark brown mixture was washed with 10% sodium bicarbonate, dried, and the methylene chloride removed under vacuum. Unreacted diphenyl ether, 19.0 g.; b.p. 102-104' (7 mm.), wae recovered by distillation through a Vigreux column. The residual solid was fractionally sublimed a t 60-80' and 0.5 mm. pressure. Each fraction was analyzed by gas-liquid chromatography a t 200" on a column packed with DC-11 on firebrick at a helium flow-rate at 60 ml./min. Comparison of each chromatogram with those of authentic samples showed that the total sublimate (8.1 g.) contained o-phenoxyphenol (5.7 g.), p-phenoxyphenol (1.9 g.), and diphenyl ether (0.5 9.). The first few fractiom of sublimate were triturated with cold n-hexane and filtered. The insoluble crystalline material w m recrystallized from n-hexane to give white needles melting a t 103-104'. The melting point was undepressed on admixture with an authentic sample of o-phenoxyphenol. AmE. Calcd. for ClnHloOn:C, 77.5; H, 5.42; phenolic acidity, 0.537 eq./100 g. Found: C, 77.8; H, 5.48; phenolic acidity, 0.525 eq./100 g. The last few fractions melted a t 77-81' and were recrystallized from n-hexane to give white plates melting at 828 3 O . The melting point was undepressed on admixture with an authentic sample of p-phenoxyphenol. Anal. Calcd. for C,aHloOz: C, 77.5; H?, 5.42; phenolic acidity, 0.537 eq./100 g. Found: C, 77.9; H, 5.35; phenolic acidity, 0.54 eq./100 g. SHELLDEVELOPMENT Co. DIVISIONOF SHELLOIL Co. EYERYVILLE,CALIF.

We have observed a novel decarbonylation of a-anilino-ala-diphenylacetic acid (I),by the action of p-toluenesulfonyl chloride and pyridine, leading to the formation of benzophenone anil (11). The byproducts (carbon monoxide and pyridinium ptoluenesulfonate) suggest that the reaction proceeds by formation of an unusual mixed anhydride, followed by a base-catalyzed elimination. The reaction was discovered during an attempt to prepare the N-ptoluenesulfonyl derivative of the amino acid I. The acid was recovered unchanged after treatment with p-toluenesulfonyl chloride and aqueous sodium hydroxide (even under forcing conditions), but reacted smoothly when anhydrous pyridine was employed as the base. However, the reaction products were those described above; none of the expected tosyl derivative was isolated. The base-catalyzed decomposition of a-(arylsulfonamido)carboxylic acids and acid chlorides has been reported previously by Wiley and his coworkers.'-8 I n the case of the free acids the decomposition products were aldehydes, disulfides, and carbon dioxide2; the acid chlorides gave aldehydes, sulfonamides, and carbon monoxide.8 A cyclic mechanism was suggested to account for these transformations.' The formal analogy between these observations and our results led us to postulate that the N-toluenesulfonyl derivative was formed first and subsequently underwent a rapid decomposition to the anil, carbon dioxide, and p-toluenesulfinic acid (as the pyridinium salt). However, the gaseous byproducts of the reaction gave no precipitate when passed through aqueous barium hydroxide and, in fact, proved to be exclusively carbon monoxide as determined by infrared methods. I n addition, the solid by-product was identified as pyridinium ptoluenesulfonate by comparison of the S-benzylthiuronium salt with an authentic sample. As sulfinic acids are relatively stable in the salt form,4 it is unlikely that pyridinium p-toluenesulfonate could have arisen from previously formed p-toluenesulfinate. These results suggest that the present reaction pursues a different course from that of the simpler

(5) W. I). Emmons and G. B. Lucas, J . Am. Chem. Soc., 77, 2287 (1955). Warning: as used in the preparation of tnfluoroperoxyacetic acid, a suspension of 90% hydrogen peroxide in methylene chloride could be detonated by a modification of the drop weight test method of F. Bellinger, et a!., Ind. Eng. Chem., 38, 310 (1946). (See Tech. Bull. SC:59-44, p. 17, Skiell Chemical Co., for description of apparatus.)

(1) R. H. Wiley and R. P. Davis, J . Am. Chem. Soc., 76, 3496 (1954). (2) R. H. Wiley and N. R. Smith, J . Am. Chem. Soc., 73,4719 (1951). (3) R. H. Wiley, H. L. Davis, D. H. Gensheimer, and N. R. Smith, J. Am. Chem. SOC.,74,936 (1952). (4) W. C. Truce and A. M. Murphy, Chem. Revs., 48, 69 (1951).

JOHN C. SHEEHAN AND JOHN W. FRANKENFELD Received August 28, 1961

FEBRUARY 1962

acids studied by Wiley. h reasonable alternative mechanism to explain our results would be attack of the acid anion on the p-toluenesulfonyl chloride to give the mixed anhydride, which then decomposes to the observed products. Mixed anhydrides (CsH5)zCC0zB I

629

NOTES

+

C7HiSOzCl

NHC~HS

CsHsN

0 AI

I1 C7H7S0,@C5H5'

+ NH

'C6H5N

of p-toluenesulfonic acid are known6 and have been proposed6as highly reactive intermediates in acylation reactions. Apparently the reduced basicity of the amino group and steric factors favor reaction a t the carboxyl group in this case. EXPERIMENTAL'

a-Anilino-a,a-diphenylaceticacid ( I).*aibTo a cooled (10') solution of 1.2 ml. (1.22g., 0.013 mole) of freshly distilled aniline in dry benzene was added, dropwise, a benzene solution of 1.4 g. (0.0057mole) of a-chloro-a,a-diphenylacetic acid.9 After the addition was complete (one-half hour) the reaction mixture was stirred at room temperature for an additi'onal one-half hour. Aniline hydrochloride was removed by filtration and the combined benzene fractions were washed with water, dried (magnesium sulfate), and concentrated to one-quarter volume. Upon cooling, 1.36 g. of solid, m.p. 165-170', was obtained. Addition of petroleum ether to the mother liquor and cooling afforded a second crop The total yield of crude acid was 1.6 g. (93%). After a single recrystallization from benzene-petroleum ether the m.p. was 178-180" (lit.ab m.p. 174-175'). The acid obtained in this way was identical to a sample prepared by the somewhat more laborious procedure of Klinger and Standke.84 Reaction of cr-anilino-a,a-d~pliphenylacetic acid ( I ) with ptoluenesulfonyl chloride and pyridine. To a suspension of 1 g. (0.0032mole) of a-anilino-a,a-diphenylacetic acid (I) and 0.8 g. (0.0047mole) of ptoluenesulfonyl chloride in 30 ml. of sodium-dried benzene was added 2.5 ml. of dry pyridine. The resulting light yellow solution was refluxed for seven hours. As the reaction progressed, the gas produced was swept from the reaction vessel in a slow stream of nitrogen, passed through a drying tube containing Drierite and Anhydrone, and collected in a gas cell. The infrared spectrum of the collected gases demonstrated that carbon monoxide (Characteristic doublet at 2120 cm.-l and 2160 cm.-1, identical to that of an authentic samplelo) was the only gas present (other than the nitrogen diluent).

After the removal of the gases, the reaction mixture was cooled to room temperature. The pyridinium salts were removed by filtration and washed with benzene. The combined benzene extracts were washed with dilute acid, dilute base, and water and dried. After evaporation of the solvent and crystallization from methanol 0.53 g. (62.4y0)of benzophenone ani1(11),m.p. 109-112"(lit." m.p. 112'), was obtained. The melting point was not depressed upon admixture with authentic benzophenone anil.I1 The infrared spectrum was identical with that of the authentic sample. Addition of phenylhydrazine to an alcoholic solution of the crude residue after the methanol crystallization afforded the phenylhydrazone of benzophenone, n1.p. 138-139' (lit.l* m.p. 137'). not depressed by admixture with an authentic sample. S-Benzylthiuronium salt. To a chilled aqueous solution of the crude salt mixture from the reaction described above was added an aqueous solution of S-benzylthiuronium ch!oride. The resulting heavy precipitate was collected by filtration and recrystallized from ethanol. The m.p. was 183-184' (lit.ls m.p. for the corresponding derivative of p-toluenesulfonic acid, 181-182') and the sample did not depress the n1.p. of an authentic sample prepared from sodium p-toluenesulfonate. The crude salt mixture failed to form a sulfone with 1chloro-2,4-dinitrophenol, in the characteristic test for sulfinic acids."

Acknowledgment. Financial support from a contract with the Office of Naval Research (Biochemistry Branch) is gratefully acknowledged. DEPARTMENT OF CHEMISTRY INSTITUTE OF TECHNOLOGY MASSACHUSETTS CAMJHZIDGE, MASS.

(11) E. Knoevenagel, J. prakt. Chem., (2)89, 37 (1914). (12) R. L. Shriner, R. C. Fuson, and D. Y. Curtin, The Systematic Identifiation of Organic Compounds, Wiley, New York, 1956,p. 318. (13) E. Chambers and G. W. Watt, J. Org. Chem., 6, 376 (1941). (14) J. P. London, J . Chem. SOC.,537 (1935).

The Reaction of Aldoximes with Alkali1 HENRYRAPOPORT AND WILLIAM NILSSON Received August U, 1961

During the course of another investigation2 it was found that phenylacetaldoxime, when subjected to the action of alkali a t 170°, was converted in 78% yield to phenylacetic acid. This reaction seemed to have sufficient potential as a synthetic tool to merit further investigation, and its application to a variety of aldoximes is the subject of the (5) See, for example, H. P. Kaufmann and L. S. Huang, present report. The only similar case of such a conversion of an Ber., 75B, 1214 (1942). (6) J. H.Brewster and C. J. Ciotti, J. Am. Chem. SOC., aldoxime to an acid by means of hot alkalia in77, 6214 (1955). volved the conversion of benzaldoxime and (7) Melting points are corrected. The authors are in- several substituted benzaldoximes, as well as furdebted to Mrs. N. Alvord for the infrared spectra. (8) (a) H. Klinger and 0. Standke, Ber., 22, 1212 (1889). (1) Sponsored in part by the United States Atomic (b) W. Schlenk, J. Appenrodt, A. Michael, and A. Thal, Energy Commission. Ber., 47, 484 (1014). (2) H. Rapoport and W. Nilsson, J . Am. Chem. Soc., (9) J. Klosa, Arch. Pharm., 288, 42 (1955). 83, 4262 (1961). (10) G. Jacini, Chimica e Industria, 29, 204 (1947); (3) E.Jordan and C. R. Hauser, J. Am. Chem. SOC.,58, Chem. Abstr., 44, 9844 (1950). 1304 (1936).

630

NOTES

VOL.

27

TABLE I PRODUCTS RESULTING FROM Aldoxime

THE

ACTIONOF POTASSIUM HYDROXIDE ON ALDOXIMES

M.P.

Phenylacetaldoxime Enanthaldoxime

Solventa

Time, Hr.

Temp.

Acid

A A C

3 3 16

190 190 80

80 63d 2 Id

A

3 8 3 6 3 3 6

170 120 170 120 190 170 120

88' 891 38h 4 68 95 15

101-1020 57c

Isobutyraldoxime

141-142e (b.p.)

Pivaldoxime

38-399

anti-Cinnamaldoxime syn-Benzaldoxime

131-132j 30-32k

B A A A A B

Yield, yo Other

-

I

50, oxime 20, amided

-

15, amide 42'

-

-

40, oxime 30, amide 4, nitrile

Solvents: A, diethylene glycol; B, 2-methoxyethanol; C, 80% aqueous ethanol. Reported6 map. 99-100'. Reported map.57' [E. Bamberger and F. Elger, Ann., 475,288 (192S)I. Enanthic acid was identified m ita amide, m.p. 93'; reported m.p. 93-94' [O. Aschan, Ber., 31, 2344 (1898)l. e Reported b.p. 141-142' [J. Petraczek, Ber., 15,2783 (1882)l. Identified as the amide, m.p. 128'; reported m.p. 128" [A. W. Hofmann, Ber., 15,977 (1882)l. 9 Reported m.p. 41' [A. Richard, Ann. chim. (Paris),[8], 21, 371 (1910)l. Identified m the anilide, m.p. 135"; reported m.p. 131-132' [G. Schroeter, Ber., 44, 1201 (1911)]. * This was separated into oxime (mostly) plus amide and nitrile. f Reported m.p. 135" [O. L. Brady and R. F. Goldstein, J. Chem. SOC.,1918 (1926)l. li Reported m.p. 35' [O. L. Brady and R . F. Goldstein, J. Chem. SOC.,1918 (1926)l.

furaldoxime, to the corresponding acids by heating in 2N aqueous sodium hydroxide a t 100'. The synoximes were transformed to acids in considerably lower yields than were the anti-isomers, and the unchanged oximes were recovered in considerable quantity. The anti-oximes were converted to acids in 38-62% yield, the remainder being converted to the more stable syn-isomer. In the present work the reaction conditions most frequently employed have been diethylene glycol as solvent a t a temperature of 170-190'. Under these conditions, all the aldoximes tested were converted by potassium hydroxide to carboxylic acids in good yields except for pivalic acid which was obtained in 38y0 yield. Here, however, high volatility may account for loss of material during the reaction and isolation. The reaction has been found to be quite general for aliphatic aldoximes as well as aromatic aldoximes. It may offer a useful alternative to the commonly employed conversion of aldehydes to acids through dehydration of aldoximes under acidic conditions followed by hydrolysis of the resulting nitrile. The results of the various reactions are presented in Table I. The reaction appears to be a base-catalyzed dehydration of the aldoxime to the niti'ile followed by hydrolysis of the nitrile through the amide to the acid. This sequence is borne out by the detection of H 0

R-

e

-NH2

+ R-COOH

AIkali is needed, since heating syn-benzaldoxime at. 120' for twelve hours in the absence of alkali in 2methoxyethanol yielded no nitrile, amide, or acid. The oxime was quantitatively recovered. The reaction is sensitive to steric factors. As seen from the table, lowering the reaction temperature from 190' to 120' does not affect the high yield of isobutyric acid from isobutyraldoxime, but it drastically reduces the yields of pivalic acid from the more highly hindered pivaldoxime, which remains largely unchanged a t this lower temperature. At 120', syn-benzaldoxime reacts slowly, only a 15% yield of benzoic acid being isolated after six hours. About 40% of the oxime is recovered, while the yields of nitrile and amide are about 4% and 30%, respectively. Hauser and Jordana similarly report only a 10% yield of acid a t 100' from synbtnzaldoxime. The reaction gives a virtually quantitative yield of benzoic acid a t 170'. Thus, it appears that a relatively high activation energy is associated with the dehydration of aldoxime to nitrile; a somewhat lower energy is required for conversion of amide to acid; and a much lower activation energy is associated with hydrolysis of the nitriles to the amides. The reaction is not merely a thermal dehydration followed by alkaline hydrolysis of the resulting nitrile, since thermal dehydration requires significantly higher temperatures and longer reaction times. It should be mentioned that the recently reported4 conversion of aldoximes to amides by means of nickel tetraacetate in xylene involves an entirely different mechanism, since the nitrile has been shown not to be intermediate in that case.

*OH-

both the nitrile and the amide as intermediates in those cases where the reaction was incomplete.

(4) L. Field, P. B. Hughmark, S. H. Shumaker, and W. S. Marshall, J. A m . Chem. SOC.,83, 1983 (1961).

FEBRUARY

1962

631

NOTES EXPERIMENTAL

Oximes were prepared by heating the aldehyde with hydroxylamine hydrochloride and sodium acetate in aqueous ethanol, except for phenylacetaldoxime which wm made by the method of Weerman.6 The aldehydes were all commercially available except for pivalaldehyde, prepared according to the method of Roberts and Teague.O The reaction of aldoximes with alkali. The general procedure used was as follows: The aldoxime (10 mmoles) was heated in 100 ml. of the solvent indicated at the temperature indicated in Table I with about 50 mmoles of potassium hydroxide, and the heating was carried out under a nitrogen atmosphere. At the end of the reaction time, the cooled mixture was diluted with water and extracted four times with half volumes of methylene chloride; these combined extracts being washed with 100 ml. of aqueous sodium chloride solution. Drying the organic extracts over sodium sulfate, filtering, and evaporating the solvent left as a residue the neutral fraction of the reaction mixture. The original aqueous alkaline solution then was acidified with hydrochloric acid to p H 7 , and again extracted with methylene chloride. A similar treatment of the organic phase yielded the weakly acidic products. Finally, acidification of the aqueous solution to pH 2 and similar extraction with methylene chloride yielded the strongly acidic products. The products were then identified by mixed m.p. and infrared comparison with authentic samples. The thermal dehydration of syn-benzaldoxime. A solution of 1.09 g. of syn-benzaldoxime in 80 ml. of 2-methoxyethanol was heated a t 120' for 12 hr. under a nitrogen atmosphere with no alkali present. The product was isolated m described above and found to consist entirely of unreacted oxime. A solution of 1.04 g. benzaldoxime in 50 ml. of diethylene glycol was heated 10 hr. at 200" under 3, nitrogen atmosphere. The product was entirely neutral and consisted of nitrile and amide. No unchanged oxime or acid wm present.

IV with lithium aluminum hydride afforded 1methylbenzocyclobutene (V), the ultraviolet spectrum of which was virtually superimposable upon that of benzocyclobutene itself.2 Reaction of IV with potassium t-butoxide in t-butyl alcohol gave methylenebenzocyclobutene (I), purified at 135' by gas chromatography. The new olefin, which absorbed bromine rapidly, was found to have an ultraviolet spectrum (Fig. 1) somewhat similar to

75

-

c

a

!ii

.0

p

f

r

EtoH

A,,

219 mp;

295 (3.62)

LAWRENCE RADIATION LABORATORY OF CHEMISTRY DEPARTMENT OF CALIFORNIA UNIVERSITY BERKELEY, CALIF.

225

1

I

I

250

275

300

Wavelength, mp

Fig. 1. Ultraviolet spectrum of methylenebenzocyclobutene (5) R. A. Weerman, Ann., 401, l(1913). (6) T. G . Roberta and P. C. Teague, J. Am. Chem. SOC., that of o-methylstyrene.3 The spectrum of I, 77, 6258 (1955).

Condensed Cyclobutane Aromatic Com-

however, showed more detail and higher resolution. As expected, mild catalytic reduction of I occurred readily with saturation of the olefinic double bond to give hydrocarbon V, as evidenced

pounds. XVI. Methylenebenzocyclobutene

WCH2

M. P. CAVAAND M. J. MITCHELL

Received September 6, 1961

This note describes a synthesis of methylenebenzocyclobutene (I), the simplest styrene analog in the benzocyclobutene series. Benzocyclobutene-1-carboxylic acid (11), prepared by an improved one-step hydrolysis of the corresponding nitrile, was reduced with lithium aluminum hydride to l-hydroxymethylbenzocyclobutene (111). Reaction of alcohol I11 with ptoluenesulfonyl chloride in pyridine gave the corresponding crystalline tosyl ester (IV). Reduction of tosylate ( 1 ) M. P. Cava, R. L. Litle, and D. R. Napier, J. Am.

C h .Soc., 80,2257 (1958).

I1 LiAlH,

I11

I

1

1

t-BuOK

IV

1

LiAlH,

OJ""" V

(2) M. P. Cava and D. R. Napier, J . Am. Chem. SOC., 80, 2255 (1958). (3) R. A. Friedel and M. Orchin, Ultraviolet Spectra of Aromatic Compounds, Wiley, New York (1951).

632

VOL.

NOTES

by change in the ultraviolet spectrum of the sample. Some further chemical transformations of I will be reported a t a later date. EXPERIMENTAL*

Benzocyclobutene-1-carboxylic acid (11). 1-Cyanobenzocyclobutenel (1.0 9.) was dissolved in saturated ethanolic potassium hydroxide (6 ml.). The solution was alIowed to stand for twenty-four hours a t room temperature, then diluted with water ( 2 ml.), refluxed for three hours, and poured into water (50 ml.). The resulting suspension was extracted with ether; then the aqueous layer was separated, acidified with 6.V hydrochloric acid, and re-extracted with ether. The second (acid-containing) ether extract wm dried over magnesium sulfate and evaporated to dryness. The residue was taken up in a minimum amount of warm petroleum ether (b.p. 3@-6O0), and the solution was decanted from an insoluble oily residue, seeded with a crystal of acid, and chilled to 10' t o give acid I1 (0.835 g., 73y0), m.p. 72-73'; reported1 76.5'. 1-Hydroxymethylbenzocyclobutene (111).A solution of acid I1 (1.505 g.) in ether (75 nil.) was added dropwise under nitrogen to a stirred solution of lithium aluminum hydride (0.830 g.) in ether. After being stirred overnight, the reaction mixture was treated with a saturated aqueous solution of sodium sulfate (ca. 4 ml., added dropwise). The resulting clear ethereal lager was drawn off, and the remaining aqueous sludge was wrashed with ether (25 ml.). The ether extracts were combined, dried over sodium sulfate, and evaporated to give alcohol I11 (1.250 g., 86y0) as a colorless oil, b.p. 93-95" (4 mm.); n y 1.5567, d s 1.071. The analytical sample was obtained by molecular distillation in .a semimicro apna.r,zt,i . -- - -i s-. Anal. Calcd. for C9HloO: C, 80.5G; H, 7.51. Found: C, 80.55: H, 7.82. 1-Hydroxymetkylbenzocyclobutene tosylate (IV). Finely powdered p-toluenesulfonyl chloride (1.24 g.) was added to a solution of alcohol I11 (0.888 g.) in pyridine (6 ml.). The reaction mixture was allowed to stand a t room temperature for twenty-four hours and was then poured into petroleum ether (100 ml.), b.p. 3GGO'. The resulting suspension was washed with cold 2N sulfuric acid (100 ml.), then with sodium bicarbonate (50 ml., 5%), and finally with water (100 ml.). The organic layer was dried over magnesium sulfate, filtered, and allowed to evaporate slowly at room temperature through a loose cotton plug. After standing for six dsys, the solution had deposited tosylate I V (1.153 g., 6OOj,) as clusters of pure white needles, m.p. 49-55', A second crop, obtained by chilling the mother liquor to 5", raised the total yield of crude product to 1.308 g. (73.6%). The sample, m.p. 73-74', was recrystallized from t-butyl alcohol. 1-Methylbenzocyclobutene (V). A solution of tosylate I V (1,000 g.) in ether (75 ml.) was added dropwise to a solution of lithium aluminum hydride (1.0 g.) in ether (75 ml.). The resulting suspension was stirred for four houra a t room temperature and was then treated with a saturated aqueous solution of sodium sulfate to decompose the excess hydride. The dried ether layer was concentrated to a small volume by carefully distilling off the ether through a fractionating column. The residual oil (1.2 9.) was passed through a gas chromatographic column (3Oy0 Apiezon M on firebrick, helium a t 105') to give pure 1-methylbenzocyclobutene (0.157 g., 380101, n y 1.5195, dS 0.924. Anal. Calcd. for CgHIO: C, 91.47; H, 8.53. Found: C, 91.14; H, 8.65. 1---

(4) Melting points are corrected. The analyses were carried out by Galbraith Microanalytical Laboratories, Knoxville, Tenn., and by Schwarzkopf Microanalytical Laboratory, Woodside, N. Y.

27

The ultraviolet spectrum (ethanol) showed the following maxima: Amax 260 (log e 3.08), 265.5 (3.24), and 271 (3.19) mp. Metitylenebenzocyclobutene (I), A solution of tosyIate I V (2.000 9.) in warm t-butyl alcohol (10 ml.) waa added to a solution of potassium t-butoxide (prepared from 0.300 g. of potassium and 6 ml. of t-butyl alcohol). The reaction mixture was stirred a t 75-80' for 15 minutes and was then poured into cold water (15 ml.). The resulting suspension was extracted four times with 5-ml. portions of petroleum ether (b.p. 3C-60°).'The aqueous layer was filtered to give an amorphous sulfur-free polymer (0.413 g.). The organic extracts were combined, dried over magnesium sulfate, and concentrated to 2 ml. by fractional distillation. Chilling the pot residue to -20' gave unchanged tosylate (0.275 g.), obtained as pure white needles. The mother liquor was vacuum distilled in an all-glass, hermetically sealed apparatus, and the distillate was subjected to gas chromatography (30% Apiezon M on [firebrick, helium a t 130') to give olefin I (0.165 g., 41%) aa a colorless liquid, n y 1.5679. Anal. Calcd. for CO&: C, 93.06; HI 6.94. Found: C, 93.87; HI 6.18. The infrared spectrum of methylenebenzocyclobutene showed absorption bands at 5.96, 6.01, and 11.6 p, characteristic of the exo methylene group. The ultraviolet spectrum is reproduced in Fig. 1. Hydrogenation of methylenebenzocyclobutene gave 1methylbenzocyclobutene. Thus, 0.01.62 g. of the olefin waa dissolved in ethanol (5 ml.) and wm hydrogenated at atmospheric pressure in the presence of palladium (0.001 g., 10%) on charcoal catalyst (reaction time, four minutes). The reaction mixture was filtered quantitatively through Celite, diluted to 10.0 ml. in a volumetric flask, and analyzed spectrophotometrically for 1-methylbenzocyclobutene. Anal. Calcd. for 1-methylbenzocyclobutene: 0.0165 g. Found: 0.0141 g.

Acknowledgment. This work was supported in part by a grant from the National Science Foundation. This aid is gratefully acknowIedged. EVANS CHEMICAL LABORATORY THEOHIOSTATEUNIVERSITY COLUMBUS 10, OHIO

'

Some Organofluorosilanes

L. W. BREEDAND MARYE. WHITEHEAD Received February 20, 1961

Organoalkoxysilanes with the general structure RSi (OEt)2R'Si (OEt)2R283are converted readily to the corresponding fluorosilanes when they are treated with boron-trifluoride etherate. The use of covalent halides in the preparation of halosilanes from alkoxysilanes is well (1) This research was supported by the United States Air Force, Air Research and Development Command, under Contract AF 33(616)-6916 and monitored by Materials Central, Wright Air Development Division, Wright-Patterson Air Force Base, Ohio. (2) L. W. Breed, J. Org. Chem., 25, 1198 (1960). (3) L. W. Breed, W. J. Haggerty, Jr., and F. Baiocrhi, J.Org. Chem., 25,1633 (1960). (4) C. Eaborn, Organosilicon Compounds, Academic Press, New York, 1960, pp. 170-171, 304-307.

FEBRUARY

1962

NOTES

Boron trifluoride etherate is effective in converting siloxanes to fluor~silanes,~-'but only meager details are reported on the use of this convenient reagent with alkoxysilanes. I n a patent, Sommer cites the preparation of difluoromethylphenylsilane, fluorotrimethylsilane, and silicon tetrafluoride from alkoxysilanes and boron trifluoride etherate without giving experimental details.' Only impure difluoromethylphenylsilane could be isolated from a mixture obtained by treating diethoxymethylphenylsilane with a 10% molar excess of boron trifluoride etherate. Several higher boiling fluorosilanes, however, were prepared in good yields by this procedure when a 50 mole per cent excess of boron trifluoride etherate was used. The properties of five new fluorosilanes prepared by this method are described in Table I. EXPERIMENTAL

p-Phenylenebis(difluoromethy1silane).A mixture of 43.1 g. (0.126 mole) of freshly distilled p-phenylenebis(dieth0xymethylsilane) and 35.5 g. (0.25 mole) of boron trifluoride etherate was heated for 3 hr. a t 58'. Fractional distillation gave 21.6 g. (827,) of the product boiling a t 99' a t 32 mm. The properties and analyses of this compound and other fluorosilanes that were similarly prepared from the corresponding alkoxysilanes are given in Table I. Dzjtuoromethylphenylsilane. After a mixture of 52.6 g. (0.25 mole) of diethoxymethylphenjlsilane and 26.0 g. (0.18mole) of boron trifluoride etherate was heated a t 6070' for 3 hr., fractional distillation of the product gave the following fractions boiling above 125': A, 125-135', 9.8 g.; B, 135-145', 16.4 g.; C, 145-155', 8.2g.; and D, residue, 10.6 g. Fraction B represented a 41.7% yield of impure difluoromethylphenylvilane. The boiling point reported for this compound is 141-142O.8 Hydrolysis of difluoromethylphmylsilane. Over a 15-min. period, a solution of 16.0 g. (0.102 mole) of difluoromethylphenylsilme (Fraction B, above) was added to 9.7 g. (0.24 mole) of sodium hydroxide in 100 ml. water. During the addition the mixture was maintained a t 0" by external cooling, and at the end of the addition the mixture was still alkaline. Insoluble salts were separated by filtration, and the filtrate was washed three times with 50 ml. portions of ether. Evaporation of the combined ether extracts gave 4.7 g. of crude product that was recrystallized from hexane to yield 1.9 g. (23%) of 1,3-dimethyl-1,3-diphenyldisiloxanediol melting 8C-82'. Reported melting points are 112-113OS and 82-84', 110-1 11 '.lo Anal. Calcd. for ClrHtsOgSiz: C, 57.89; H, 6.25;Si, 19.34. Found: C, 58.70;H, 6.80;Si,19.57.

RESEARCH INSTITUTE MIDWEST KANSAS CITY10,Mo.

( 5 ) F. P. MacKay, Doctoral Dissertation Series Publication No. 19311, Univ. Microfilms, Ann Arbor, Mich. (6)L. H. Sommer and G. R. Ansul, J . Am. Chem. SOC., 77,2482(1955). (7) L. H.Sommer, U. S. Patent 2,713,063 (1955). (8) Robert 0.Sauer, U. S. Patent 2,647,136 (1953); U. S. Patent 2,730,540 (1956). (9) M. F. ShostakovskiI, D. A. Kochkin, Kh. I. Kondrat'ev, and V . M. Robov, Zhur. ObshcheZ Khim., 26, 3344 (1956);Chem. Abstr., 51, 9514 (1957). ( I O ) W. H. Daudt and J. F. Hyde, J. Ana. Chem. SOC., 74,386 (1952).

633

634

NOTES

Chemistry of Certain Thiocyanates and Isothiocyanates Containing Silicon GEORGER. GLOWACKI AND HOWARD W. POST Received April 4, 1961

Bugorkova, Petrova, and Radionov' as well as have reported the addition of thiocyanogen to several olefinic compounds including some organosilanes. It may be assumed that the SCN groups attached themselves to form -SCN radicals. In reacting with silicon compounds however, there is always the possibility that rupture m y take place between silicon and carbon. I n fact, qualitative runs demonstrate this in the treatment of p-trimethylsilylphenol with thiocyanogen. On the basis of other work6 the isothiocyanate structure is assigned to the products of certain syntheses described below.

VOL.

27

silyl triisothiocyanate waa synthesized in 82% yield, b.p. 134-138' (3.0 mm.), ny 1.6350. Infrared absorption appeared a t 3070,2090,2000, 1600, 1410, and 1055 cm.-l. Anal. Calcd. for CsHaNsSaSi:C, 26.20; H, 1.31; N, 18.30; S, 41.90; Si, 12.25. Found: C, 26.04, 25.75; H, 1.33, 1.40; N, 17.84. p-Tolylsilyl triisothiocyanate. This compound was also prepared similarly to the above but using ammonium thiocyanate, in 57% yield, b.p. 182-185' (3.5 mm.), n y 1.6490. Infrared abeorption was recorded a t 3030, 2940, 2090,2020, 1605 and 1130 cm.-1. Anal., Calcd. for CloH7NaSaSi: C, 41.00; H, 2.40; N, 14.30; S, 32.80; Si, 9.59. Found: C, 39.50; H, 2.38; N, 14.12; S, 27.80; Si, 9.61. DEPARTMENT OF CHEMISTRY THEUNIVERSITY OF BUFFALO BUFFALO 14, N. Y.

The Preparation of Certain Carbon-functional Silathiols and Silathio Esters PAUL F. GAWRYS AND HOWARD W. POST

EXPERIMENTAL

%,3-Dithiocyanopropyltrimethylsilane.To 10.0 g. (0.087 mole) of trimethylallylsilane in 150 cc. of anhydrous ether, in a flask equipped with stirrer, etc., fresh thiocyanogen solution was added, containing 0.09 mole in 100 cc. of anhydrous ether. During the addition, the system was stirred vigorously and kept immersed in water a t 10". Three days later polymerized thiocyanogen was removed by filtration, the ether was distilled and the products fractionated giving 2,3-dithiocyanopropyltrimethylsilane, 6.5 g., 34% yield, b.p. 127131', na,7.' 1.4965. Infrared absorption was recorded a t 2970, 2180,2080, 1260, and 1415 cm.-l. Anal. Calcd. for C8HI4N&Si: C, 41.70; H, 6.06; N, 12.10; Si, 12.18. Found: C, 41.14; H, 5.76; N, 12.20; Si, 11.80. 1,%Dithiocyanoethyltrintethylsilane.In similar manner, this compound was prepared and isolated in 1.6% yield, m.p. 72-74", and showing infrared absorption a t 2970, 2180, 1260, and 1415 cm.-'. Allykilyl triisothiocyanate. To 90.0 g. (1.1 mole) of dry sodium thiocyanate in 500 cc. flask was added 176 g. (1.0 mole) of allyltrichlorosilane, freshly distilled. The flask, equipped now with calcium chloride tube and reflux condenser, was immersed in cold water at first, then the cooling was discontinued and refluxing was permitted. About 20 minutes later, refluxing was continued by heating, this time for 90 minutes. The cooled mixture was filtered through glass wool. Distillation gave allylsilyl triisothiocyanate, yield 70%, b.p. 126-128" (2.5 mm), n y 1.6140, with infrared absorption showing at 3080,2090,1995,1635, and 1055 em.-'. Anal. Calcd. for C6H5N3S3Si:C, 29.60; H, 2.06; N, 17.30; S, 39.50; Si, 11.50. Found: C, 29.34, 29.31; H, 1.98, 2.06; N, 17.20; S, 38.54; Si, 11.28. Vinylsilvl tm'isothiocyanate. Similarly to the above, vinyl(1) A. A. Bugorkova, L. M. Petrova, and V. M. Radionov, Zhur. Obshchei Khim., 23, 1808 (1953). (2) A. A. Bugorkova, V. F. Mironov, and A. D. Petrov, Zzvest. Akad. Nauk, S.S.S.R., OKhN, 474 (1960); English page 441. (3) V. F. Mironov, Yu. P. Egorov, and A. D. Petrov, Izvest. Akad. Nauk, S.S.S.R., OKhN, 1400 (1959); English page 1351. (4) A. D. Petrov, Yu. P. Egorov, V. F. Mironov, G. I. Nikishin, and A. A. Bugorkova, Izvest. Akad. Nauk, S.S.S.R., OKhN, 50 (1956) ; English page 49. (5) G. S. Forbes and H. H. Anderson, J . Am. Chem. Soc., 69, 3049 (1947).

Received April 19, 1961

Marvel and Kripps' have reported the addition of thiolacetic acid to dimethyldiallylsilane to form dimethylbis - (3 - mercaptopropy1)silane diacetate. Hydrolysis of this diester gave the corresponding dimercaptan. Cooper2 obtained trimethylsilylmethyl mercaptan by treating trimethylchlorome thylsilane with potassium hydrosulfide. Mironov and Pogonkinalat4prepared a series of carbon functional mercaptans of organosilicon compounds from the corresponding thiocyanates. Other methods already in the literature, were used herein.6-8 Saponification of these thiol esters, followed by their acidification yielded the corresponding monoor dimercaptans. Infrared data are included. The formulas of the addition products are written to indicate that addition had taken place in a manner known as "anti-Markownikoff ." The conclusion to this effect was reached on the basis of the results from analogous work by ~ t h e r s . l * ~ * ~ EXPERIMENTAL

Dimethyldiallylsilane, methylphenyldiallylsilane, diphenyldiallybilane, and d i m e t h y l b i s - $ - m e t h a l l y ~ l awere ~ prepared by the methods outlined by Nasiak and Post.' Trimethyl-pmethallylsilane was prepared by the procedure given by (1) C. S. Marvel and H. Kripps, J. Polymer Sci., 9, 53 ( 1952).

(2) G. D. Cooper, J. Am. Chem. SOC.,76,2500 (1954). (3) V. F. Mironov and N. A. Pogonkina, Zzuest. Akad. Nauk, U.S.S.R., Otdel. Khim. Nauk., 707 (1956). (4) V. F. Mironov and N. A. Pogonkina, Izvest. Akad. Nauk, U.S.S.R., Otdel. Khim. Nauk., 85 (1959). (5) D. C. Noller and H. W. Post, J . Ora. Chem.,. 17,. 1393 (1952). (6) L. D. Nasiak and H. W. Post, J. Org. Chem., 24, 489 (1959). (7) R. Nagel and H. W. Post, J . 01.0.Chem., 17, 1379 (1952).

FEBRUARY

1962

Petrov and Nikishin.8 Triethylvinylsilane was synthesized by the method outlined by Nagel and Post.' 4,4-DimethyL4-sib-l,7-heptylene dithioacetate and 4,4-diphenyl-4-sib-1,7-heptylene dithioacetate were prepared by the addition of thiolacetic acid t o the proper olefinic silane but were not isolated. They were not distilled but converted to their respective dimercaptans immediately upon synthesis. 4-Methyl-4-phenyl-4-sila-l,7-heptylenedithioacetate. This compound was prepared by treating 40.47 g. (0.20mole) of methylphenyl-diallylsilane in 75 cc. of cyclohexane, in a 200 cc. three necked flask with the usual stirrer, dropping funnel, and adapter, with 30.4 g. (0.40mole) of redistilled thiolacetic acid in 25 cc. of cyclohexane, slowly added, with the temperature kept a t about 65".After two hours of refluxing, excess thiolacetic acid and cyclohexane were removed by distillation a t reduced pressures. Above 200' a t 1 mm. decomposition occurred. The pot material was dissolved in an equal volume of methyl alcohol, and the colored impurities removed by Norite. The organic producta were then thrown out of solution by water, dried over anhydrous sodium sulfate, and analyzed, n y 1.5480, d:' 1.0819. Purification ww limited as distillation resulted in slight decomposition. Anal. Calcd. for Cl~HzsO~slSi:Si, 7.91, M.R., 104.1. Found: Si, 8.21,8.26,M.R., 102.9. d14,4-Trimethy1-4-silapent yl-1 thiolacetate. By a procedure similar to that used in the preceding preparation, this compound was prepared and isolated in 53% yield, b.p. 99"-100" (9 mm.), n g 1.4664,d:" 0.9158. Anal. Calcd. for CeHeOSSi: Si, 13.73, M.R., 61.28. Found: Si, 13.32,13.50,M.R., 60.51. 3,s-Diethpl-S-silapatyl-1 thiolacetate. This compound was prepared as indicated above, yield 42%, b.p. 117°-1190 (6.5 mm.) n g 1.4779,di' 0.9284.Infrared absorption occurred a t 3.3, 5.8,8.0, 11.8,and 14.3p. Anal. Calcd. for CloHnOSSi: Si, 12.85, M.R., 66.91. Found: Si, 13.05,13.11,13.09,M.R., 66.57. 4,4-Dimethy1-4-sibheptylene-l,7 dimercaptan. As described for the thiolacetates, 75 cc. of purified cyclohexane and 33.6 g. (0.24 mole) of dimethyldiallylsilane were treated with 36.5 g. (0.48 mole) of redistilled thiolacetic acid a t 55". After two hours of refluxing, the mixture was hydrolyzed by the slow addition of 40 g. (1.0mole) of sodium hydroxide in 100 cc. of distilled water and 40 cc. of 95% ethyl alcohol. The system was refluxed for 90 minutes, then neutralized with 6 N hydrochloric acid. After cooling, the upper layer was separated and combined with two 50 cc. benzene extracts of the water layer. After drying over 10 g. of anhydrous sodium sulfate, the material was fractionated yielding 4,4-dimethyl-4-silahepty1ene-ll7dimercaptan, 25% yield, b.p. 119.0-119.5° (6 mm.), ny 1.5034,d:" 0.9631. Infrared absorption was recorded at 3.3, 7.0, 8.0,8.5, 10, 10.8, and 12-13 p. Anal. Calcd. for C8HdZSi: Si, 13 46, M.R., 64.24.Found: Si, 13.06,13.51,M.R., 64.01. 4-Methy1-4-phenyl-4-silaheptykne-lJ7dimercaptan was prepared in the same manner, yield 25%, b.p. 161-162" (2.5 mm.), n'," 1.5580, dz5 1.0396. Infrared absorption ap13,and 13.6 p. peared a t 3.4,6.9,5.8,8.7,9.7,12.2, Anal. Calcd. for ClrH&Si: Si, 10.38, C, 57.7, H, 8.17, M.R., 84.07.Found: Si, 10.58,10.66,C, 56.5,H, 7.69, M.R., 83.87. d,4,4,6-Tetramthyl-4-sibheptylene-l,7 dimercaptan. This compound was also prepared as described above, 22% yield, b.p. 147-148' (8 mm.), ny 1.5031 4' 0.9534. Anal. Calrd. for CloH&Si: Si, 11.87, C, 50.79, H,10.2, M.R., 73.50.Found: Si, 11.65, 11.73, C, 50.53, H,9.6, M.R., 72.36. dJ4,~-Trimethyl-4-silapentyl-l-mercaptun. In a similar apparatus, 40.0 g. (0.19 mole) of 4,4-dimethyl-4-silapentyl-l thiolacetate in 80 cc. of purified cyclohexane was treated (8) A. D. Petrov and S. I. Nikishin, Zzvesf. Akad. Nauk, U.S.S.R., Otdel. Khim. Nauk, 1128 (1952);Chem. Abstr., 48, 1247 (1954).

635

NOTES

with 24.0 g. (0.60mole) of sodium hydroxide in 100 cc. of water and 50 cc. of ethyl alcohol a t 45-50'. The mixture was refluxed for five hours, then cooled and acidified with 6 N hydrochloric acid. The product was worked up as in the cases described above, giving a 9% yield, b.p. 69.5-70.5' (16 mm.), n y 1.4576, d:' 0.8597. Infrared absorption was recorded a t 3.3, 5.1, 6.1, 6.8, 7.0, 8.1, 9.4, 9.8, 11.4-12.7, 13.1,and 14.0 p. Anal. Calcd. for C?H&Si: C, 51.79, H, 11.17, M.R., 51.88. Found: C, 51.70,H, 11.41, M.R., 51.48. p-Triethylsilylethyl mercaptan. Similarly, this compound was prepared in 50% yield by the hydrolysis of the corresponding thiolacetate, b.p. 113' (23 mm.), n': 1.4730, d:' 0.8789. Infrared absorption occurred a t 3.3,6.8,7.0, 7.8, and 9.8I.(. Anal. Calcd. for C8H20SSi: Si, 15.91, M.R., 56.52. Found: C, 15.94,16.01,M.R., 56.30. DEPARTMENT OF CHEMISTRY THEUNIVERSITY OF BUFFALO BUFFALO14,N. Y.

The y-Ionones E. T. THEIMER, W. T. SOMERVILLE, B. MITZNER,AND S. LEMBERQ Received May I d , 1961

The base catalyzed condensation of citral with acetone, followed by acid cyclization, leads to a mixture of a- and p-ionones'; the ratio of the isomers formed is related to the acid used.* Similarly, methyl ethyl'ketone has been shown to condense with citral to form on cyclization the n-methylionones (a and p) and the isomethylionones (a and The 7-isomers (exocyclic methylene) have not previously been reported as products of these reactions, nor have they hitherto been characterized. We wish to report the isolation and characterization of pure y-ionone and y-n-methylionone from commercial products. The formtttion of the exocyclic methylene isomer is not unexpected, as the intermediate 11, prob(1) F. Tiemann and P. Kruger, Ber., 26, 2693 (1893); F. Tiemann, Ber., 31, 2318 (1898); 32, 827 (1899); 33, 883 (1900). (2) H.Hibbert and L. T. Cannon, J. Am. Chem. Soc., 119 (1924);Org. Syntheses, 23, 78 (1943). (3) M. G.J. Beets and H. van Essen, Rec. trav. chim., 77, 1138 (1958), and references therein, (4) Two reports of a derivative of y-ionone have appeared: (a) B. Willhalm, V. Steiner, and H. Schinz, Helv. Chim. Acta, 41, 1359 (1958)have reported the semicarbazone of this isomer (column chromatographed) , originally thought to be present as a mixture of 6- and 7-ionone. More recently, D. Szabo, Chem. Abstr., 54, 19748 (1960)has indicated the presence of y-ionone in the phosphoric acid cyclization of pseudo ionone by conducting the cyclization in the presence of 2,4dinitrophenylhydrazine. Thus far, our efforts to form derivatives of both pure new compounds reported here have led to viscous, intractable oil& with semicarbazide, 2,4dinitrophenylhydrazine and thiosemicarbazide, due to apparent product instability.

636

VOL. 27

NOTES

n

to both double bonds. Signals a t 5.32 and 5.53 p.p.m. are compatible with the hydrogen of the vinylidene methylene, g. The positions of the gem-dimethyl group (9.14, Y 9.10 p.p.m.) and of the methyl group attached to 1 (7.85 p.p.m.) as well as of the olefinic the carbonyl R1 = H, Rz = H; ionones, a,8, y proton alpha to the carbonyl, T = 4.05, (uncorR1 = CHa, RZ = H; n-methylionones, a, p, y R1 = H, RZ = C K ; isomethylionones, a,8, y rected, J = 15.2 c./s.), and the quartet of the oleably formed during cyclization, is common to the finic proton beta to the carbonyl (T = 3.41, 3.24, 3.14, and 2.98) conform to expectation. three isomers anticipated. The NMR spectrum of 7-n-methylionone (IV) suggests a doublet a t 7.51 p.p.m. ( J = 10.2 c./s.), again supporting the assignment of this signal to e. This doublet is partially obscured by the methylene protons, as shown below. The hydrogens of the I1 vinylidine methylene occur a t 5.50 and 5.27 p.p.m. We have found that the gamma components co- The decision between the two possible structures, ydistill with their respective isomers and, therefore, n-methylionone and y-isomethylionone, is resolved even under the most stringent distillation tech- readily by observation of the following facts: (1) niques only partial isomeric enrichments are ac- There are no methyl groups attached directly to a complished. The y-ionone was isolated by vapor- double bond and ( 2 ) the methylene group b appears phase chromatographic trapping from a mixture, as a quartet ( T =7.69,7.59,7.56,7.33).The adjacent methyl group, a, appears a t 8.81 and 8.93 r with a the composition of which is described By vapor phase chromatographic analyses anoma- suggestion of a shoulder a t 9.0 p.p.m., the latter siglous components have been observed, but not nal being overlapped by the gem-dimethyl grouping ; identified, in commercial methylionones. Enrich- the compound is, therefore, y-n-methylionone. The corresponding 7-isomethylionone has not ment of the component, which in vapor phase been isolated in pure form to date, but partial chromatography showed as a peak between a-nmethylionone and p-n-methylionone, was effected separation from the a-isomethyl isomer with which by fractionation of commercial cr-isomethylionone6 it is simultaneously eluted on vapor phase chromathrough a 4-foot protruded packed column (20:l tography has been achieved. It should be noted that the assignment of the reflux ratio). Vapor phase chromatographic examination of the distilled fractions showed that a- tertiary hydrogen, e, is most critical if the intei preisomethylionone had a retention time of thirty tation is to be unequivocal. A study of a-ionone minutes and the unknown isomer (tentatively desig- and a-n-methylionone shows the presence of this doublet., which is absent in the “beta” series. nated IV), thirty-four minutes. The structures under discussion are particularly We shall record the results of this KMR study in amenable to proton KC’NR ana lyse^,^ and the latter a future publication. firmly support the proposed structures. EXPERIMENTAL

?-Ionone. Vapor-phase chromatographic separation of the mixture (described in ref. 5) was accomplished by using a 10 ft. l/c-in. column of silane-treated Celite (60-80 mesh) impregnated (10%) with Craig polyester succinate; the aionone had a retention time of 27 minutes and the gamma iso?-Ionone I11

y-n-Methylionone

IV

The KMR spectrumlo of y-ionone (111) shows a clearly resolved doublet at, 7.49 p.p.m. ( J = 10.2 c./s.) attributed to the tertiary hydrogen e, allylic

(5) This sample was kindly supplied by Dr. J. Bain of The Glidden Co., Jackqonville, Fla. Its composition was approximately 357, of the presumed y-ionone and 65% a-ionone, as determined by vapor phase chromatography. The appearance of a terminal methylene absorption in the infrared (v$; 893 cm.-1) was strong evidence for the gamma content. (6) The commercial methylionones examined were rich (80%) in a-isomethylionone with smaller amounts of 0isomethylionone, a-n-methylionone, and 0-n-methylionone. (7) NMR spectra were measured on a Varian High Resolution Spectrometer, HR60. Samples were dissolved in carbon tetrachloride in 7-1070 concentration (by volume). The magnetic field strength waa 14,092 gauss and the oscillating frequency 60M c./s. Tetramethyhilane (1/4%) was used as the standard, a s described by G. V. I). Tiers, J. Phys. Chem., 62, 1151 (1958). (8) G. Buchi and N. C. Yang, Helv. Chim. Acta, 38, 1338 (1955)

FEBRUARY

1962

637

NOTES

PERTINENT NMR DATA a

I11

7.85

IV

-

b

C

-

4.18,3.91 4.14,3.89

7.69,7.59 7.46,7.33

Values d 3.41,3.24 3.14,2.98 3.37,3.20 3.11,2.93

mer, 30 minutes; column temperature, 180"; flow rate 50 ml./min. a t an inlet pressure of 46 p.s.i. Unequivocal evidence for the ionone skeleton of IV was obtained by the following: On atmospheric hydrogenation with platinum oxide in glacial acetic acid, three mole equivalents of hydrogen were absorbed; chromic acid oxidation (12 hr.) of the tetrahydroionol thus formed gave, on work-up, a ketone whose semicarbazone melted a t 172-174'. Admixture of the latter with the semicarbazone of authentic tetrahydroionone,* m.p. 175-177", prepared from a-ionone, showed no melting point depression. Anal.931o Calcd. for C13HZ00:C, 81.20; H, 10.48. Found: s omp, H C, 81.53; H, 10.47; mol. wt., 192; n'," 1.5003; h ~ ~ ~226 E 12,620. -,2sx 1682 cm.-l (conjugated carbonyl), 892 cm.-l (vinylidene methylene). yn-Methylionone. Vapor phase chromatography trapping of the distillate richest in 7-n-methylionone gave the pure compound; for this purpose, a Celite packing, impregnated with Dow Corning Silicone No. 710 (25% w./w.) in a */lein. ten-foot column, with a flow rate of 50 ml./min. (inlet pressure of 26 p.s.i.) a t B O " , was used. Support for the nmethylionone skeleton of IV was obtained as follows: On hydrogenation, as described above, three moles of hydrogen were absorbed; chromic acid oxidation of the n-methyltetrahydroionol gave a ketone whose semicarbazone melted a t 116-118" after repeated recrystallizations. The melting point of the latter was undepressed upon admixture with the semicarbazone of authentic n-methyltetrahydroionone, m.p. 116-118". Anal. Calcd. for CldH220: C, 81.50; H, 10.75. Found: C, 81.81; H, 10.65; mol. wt., 206, n y 1.4969; b.p. 76'/0.3 mm.; 227 mp, E 12,430. y:Zx 1682 cm.-l (conjugated carbonyl), 892 cm.-l (vinylidene methylene). RESEARCH DEPARTMENT INTERNATIONAL FLAVORS AND FRAGRANCES INC. UNIONBEACH,N. J. (9) Microanalyses by the Schwarzkopf Microanalytical Laboratory, Woodside 77, N. Y . (10) Molecular weights were determined by the lowionization, parent-ion technique on a Consolidated Electrodynamirs Model 21-103C mass spectrometer. See, for example, F. H. Field and S. H. Hastings, Anal. Chem., 28, 1248(1956).

Cyanocoumarins S. S. LELE,M. G. PATEL, A N D SURESIISETHNA Received M a y 15, 1.961

-

A few cyanocoumarins have been previously

but no attempt appears to have (1) B. B. Dey and A. Kutti, Proc. Indian Acad. Sci., India. 6,641 (1940). (2) C. H. Schroeder and K. P. Link, J . Am. Chem. Soc., 75,1886 (1953). (3) W. Baker and C. S. Howes, J.Chem. SOC.,119 (1953).

e

f,f'

B

7.58,7.41

9.14,9.10

5.50,5.27

7.61,7.42

9.12,9.08

5.49,5.27

been made so far to prepare them by the Rosenmund-von Braun reaction on the halogenated coumarins or to study their hydrolysis. The present work deals with these aspects. Methyl 7-methoxy-8-iodocoumarin-6-carboxylate, methyl 7-methoxy-8-iodo-4-methylcoumarinG-carboxylate, and methyl 5-methoxy-S-iodo-4methylcoumarin-G-carboxylate, when heated with anhydrous cuprous cyanide at temperatures mentioned in Table I, yielded the corresponding 8cyano derivatives (Ia and b) and (IIIb). With hot 10% alkali the cyano group of these coumarins remained intact, but the esters were hydrolyzed to the corresponding carboxylic acids which on decarboxylation gave the cyanocoumarins (ICand d) and (IIId), respectively, identical with those obtained by the,Rosenmund-von Braun reaction on 7 - methoxy - 8 - iodocoumarin, 7 - methoxy - 8iodo-4-methylcoumarin and 5-methoxy-8-iodo-4methylcoumarin. These cyanocoumarins al,$0 remained unaffected on boiling with 10% potassium hydroxide solution.

a R = H R ' = COOCHa

b CHI COOCHi

c H H

d CHa

H

e CHzCOOCHa H

7-Methoxy-3.-iodocoumarin, 7-methoxy-3-iodo4-methylcoumarin, and methyl 7-methoxy-3-iodo4-methylcoumarin-6-carboxylate on similar reaction with anhydrous cuprous cyanide yielded the corresponding 3-cyano derivatives (IIc, d, and b). The cyanocoumarjn (IIc) was previously prepared by Baker3 by the condensation of 2-hydroxy-4methoxybenzaldehyde with malononitrile in presence of piperidine. With hot 10% alkali these 3cyanocoumarins yielded the corresponding coumarin-3-carboxylic acids. Methyl 7-methoxy-8-iodocoumarin-4-acetate on a similar Rosenmund-von Braun reaction gave a mixture of 7-methoxy-8-cyano-4-methylcoumarin (Id) and methyl 7-methoxy-8-cyanocoumarin4acetate (Ie). On hydrolysis with sulfuric acid (goyo by volume) (4) M. Crawfqrd arid J. W. Rasburn, J . Chem. Soc., 2155(1956). (5)'C. Wiener, C. H. Schroeder, and K. P. Link, J . Am. Chem. Soc., 79, 5301 (1957).

638

NOTES

VOL.

27

TABLE I CYANOCOUMARINS~

No.

Coumarin

1 7-Methoxy-8-cyano-

2 7-Methoxy-3-cyano3 7-Methoxy-8-cyano-6carbomethoxy4 7-Methoxy-8-cyano-4 carbomethoxymethyl5 7-Methoxy-8-cyano-4 methyl6 7-Methoxy-3-cyano-4 methyl7 7-Methoxy-8-cyano-4methyl-6-carbomethoxy8 7-Methoxy-3-cyano-4 methyl-6-carbomethoxy9 5-Methoxy-8-cyano-4 methyl10 5-Methoxy-8-cyano-4methyl-6-carbomethoxy-

Rosenmundvon Braun Reaction Temperature

M.P.

%

Formula

180-190 180-190

269-270 225-226

60 40

C11H703N C11H703N

65.7 65.7

65.7 65.7

3.4 3.2

3.5 3.5

6.7 7.0

7.0 7.0

210-220

192

54

Ci3H006N

60.2

60.2

3.7

3.5

5.5

5.4

190-210 192

22

CuHii06N

61.8

61.5

4.0

4.0

5.3

5.1

220-230

289-201

55

C12HsOaN

67.0

67.0

4.3

4.2

6.5

6.5

180-185

223

75

CizHo03N

67.0

67.0

4.2

4.2

6.2

6.5

240-245

276

70

C14HilOsN

61.3

61.5

4.4

4.1

5.3

5.1

250-255 249

70

CirHiiOsN

61.8

61.5

4.4

4.1

4.8

5.1

260-270

227-230

46

CizHoOaN

67.3

67.0

4.2

4.2

6.8

6.5

225-230

236

44

CidHIiO5N

61.9

61.5

4.2

4.1

4.9

5.1

Yield,

c, % H, % N, % Found Calcd. Found Calcd. Found Calcd.

a Obtained on Rosenmund-von Braun reaction from the corresponding iodocoumarins described by S. S. Lele and S. Sethna [ J . Org. Chem., 23, 1731 (1958)] and S. S. Lele, M. G. Patel, and S. Sethna [J.Indian Chem. Soc., 37, 775 (1960)l.

TABLE I1 HYDROLYSIS PRODUCTS FROM THE CYANOCOUMARINS MENTIONED IN COLUMN 2, TABLE I Hydrolysis No. withu 1 3

A A

B 4

Ab

B 5

A

6

A, B

7

A

B 8

A B

9

A

10

A

B

Product Obtained Coumarin 7-Methoxy-Scarbamoyl7-Methoxy-Scarbamoyl-6carboxy7-Methoxy-8-cyano-6-carboxy7-Methoxy-8-carbamoyl4-acetic acidi-Methoxy-8-cyano-4acetic acid7-Methoxy-8-cyano-4methyl7-Methoxy-8-carbamoyl-4 methyl7-Methoxy-4methyl-3carboxy-b 7-Methoxy-8-carbamoyl4-methyl-6-carboxy- , 7-Methoxy-8-cyan0-4mcthyl6-carboxy7-Methoxy-4-methyl-3,6dicarboxy7-Methoxy-3-cyano-4-methyl6-~arboxy-~ 5-Methoxy-8-carbamoy1-4methyl5-Methoxy-8-carbamoyl-4 methyl-6-carboxy5-Methoxy-Scyano-4methyl6-carboxv-

c, % M.P.

Formula

Found Calcd.

H, % Found Calcd.

N, % Found Calcdy

277-279

60.7

60.3

3.9

4.1

6.8

6.4

267-268

54.8

54.8

3.3

3.5

5.7

5.3

228

59.0

58.8

2.6

2.9

5.9

5.7

235-236

56.3

56.3

3.8

4.0

5.1

5.0

274-275

60.3

60.2 3.5 (See Table I)

3.5

5.6

5.4

278

61.7

61.8

4.5

4.8

6.1

6.0

184-185

61.2

61.5

4.3

4.3

-

-

272

56.0

56.3

3.7

4.0

4.9

5.1

283

60.2

60.2

3.6

3.5

5.6

5.4 -

208-210

55.9

56.1

3.4

3.6

-

248

60.2

60.2

3.5

3.5

5.8

5.4

286

61.5

61.8

4.7

4.8

6.0

6.0

272

56.2

56.3

4.0

4.0

4.9

5.0

248

60.1

60.2

3.9

3.5

5.5

5.4

A = Sulfuric acid (goyo by volume); B = Potassium hydroxide (loyo). Baker et al. [ J . Chem. Soc., S 12 (1949)l. Hydrolysis was carried out by keeping with cold alkali (lo’%)for 48 hr.

FEBRUARY

1962

630

NOTES

TABLE I11 HYDROLYSIS O F %CARBAMOYL DERIVATIVES MENTIONED I N COLUMN 3, TABLE 11

S. No. 1 3

4 5 7 9 10

Method of Hydrolysis Hydriodic acida Alkali ( loyo)or Sulfuric acid (50y0) ii LL

'i

Sulfuric acid (50%) c

c, % Product Obtained Coumarin

M.P.

7-Hydroxy-%carboxy7-Methoxy-6-carbo~y-~

Formula

230 CioH805

-

7-Methoxy-4methyLc 7-Metho~y-4methyl-%carboxy-~ 7-Metho~y-4methyl-6-carboxy-~ 5-Methoxy-Pmethyl-8-carboxy5-Methoxy-4methyl-6,%dicarboxy-

-

Found Calcd.

H, % _____ Found Calcd.

58.2

58.3

3.1 -

2.9 -

-

-

-

-

-

-

-

-

-

-

-

-

-

286 284

C12H1005 ClaHloO7

61.2 56.2

61.5 56.1

4.6 3.6

4.3 3.6

-

-

-

-

Hydrolysis with hydriodic acid was carried out by heating the 8-carbamoyl derivative with hydriodic acid at 130" in an oil bath for 3 hr. S. Kumar, L. Ram, and J. N. Ray [J.Indian Chem. SOC.,23,365 (1946)l. Pechmann and Duisberg [Ber., 16, 2122 (1883)l. Limaye and Kullcarni [Rasayanam,1, 251 (1943)l; Chem. Abstr., 38, 4264 (1944). e Dalvi and Sethna [J.Indian Chem. SOC.,26,405 (1949)l.

the 8-cyanocoumarins gave the corresponding 8carbamoyl derivatives, while the 3-cyanocoumarins yielded the corresponding coumarin-3-carboxylic acids (Table 11).The 8-carbamoylcoumarin derivatives on further hydrolysis with sulfuric acid (50%) or alkali gave the 8-carboxylic acids. However, in the case of 7-methoxy-8-carbamoylcoumarin-4acetate only 7-methoxy-4-methylcoumarin was obtained and in the cases of methyl 7-methoxy-8carbamoylcoumarin-Bcarboxylate and its 4-methyl analog, the carbamoyl group was eliminated during the hydrolysis (Table 111). The cyanocoumarins (ICand d), (Ird), and (IIId) were demethylated to the corresponding hydroxy derivatives. EXPERIMENTAL

All melting points are uncorrected. Rosenmund-von Braun synthesis (see Table I). The iodo derivative (0.01 mole) was mixed with anhydrous cuprous cyanide (0.02 mole) and heated with stirring a t the specified temperature for 10 minutes. The mixture was powdered and extracted with either acetone or acetic acid. The residue from the extract recrystallized from acetic acid in needles. Addition of cyano derivative from a previous run aa recom-

mended by Koelsch-and WhitneyBto the reaction mixture improved the yield. A . Sulfuric acid hydrolysis (see Table 11). The cyano compound (0.5 g.) waa heated with sulfuric acid (90% by volume) on a steam bath for 2 to 3 hr. The solid obtained on pouring the reaction mixture on crushed ice waa extracted with sodium hydrogen carbonate to separate the acid wherever formed. The product obtained was crystallized from acetic acid. B. Alkaline hydrolysis (see Table 11).The cyanocoumarin (0.5 g.) was heated with aqueous alcoholic potassium hydroxide solution (lOoj,) on a steam bath for 2 to 3 hr. The product obtained on acidification waa purified through sodium hydrogen carbonate and crystallized from acetic acid. Hydrolysis with sulfuric acid (60%) (see Table 111). The carbamoyl derivative (0.5 g.) was heated with sulfuric acid (18 ml.; 50%) in an oil bath a t 120' for 3 hours. The product obtained on pouring the reaction mixture on crushed ice crystallized from acetic acid. Sulfuric acid (50%) waa prepared by adding concentrated sulfuric acid (6 ml.) to a mixture of acetic acid (8 ml.) and water ( 4 ml.). Demethylations (see Table IV). The methoxycyanocoumarin was dissolved in acetic anhydride and then heated with hydriodic acid a t 120' in an oil bath for 3 hours. The product obtained crystallized from acetic acid.

Acknowledgment. One of us (M.G.P.) thanks the Government of India for the award of a research scholarship.

TABLE IV HYDROXYCYANOCOUMARINS~ S. No.

Coumarin

M.P.

Formula

c, %

Found

Calcd.

H, % Found Calcd.

N,

?o

Found

Calcd.

7 8

7 5

~~

1 5 6 9 a

7-Hydroxy-8-cyano7-Hydroxy-€&cyano-Pmethyl7-Hydroxy-3-cyano-4-methyl5-Hydroxy-8-cyano-4-methyl-

305 CioH603N 272 CiiHiOaN 298 CiiH70aN 276 Cl1H?03N

64.2 65.4 65.4 65.9

64 2 65.7 65.7 65.7

2 6 3 2 3.8 3.6

2.7 3.5 3.5 3.5

Obtained by demethylation of the corresponding methoxycyano derivatives described in Table I. CHEMISTRY DEPARTMENT

(6) C. F. Koelsch and A. G. Whitney, J. Org. Chem., 6 , 795 (1941).

FACULTY OF SCIENCE OF BARODA M. S. UNIVERSITY BARODA-8, I N D I A

7 3

7 0

6 8 7.2

7 0 70

640

NOTES

The Preparation of Certain Trisubstituted 2,6-Bis(2’-pyridyl)pyridines1 FRANCIS H. CASE Received M a y 22, 1961

VOL.

27

catalytic reduction of 4,4‘-dinitrobipyridine dioxide in acetic acid, By a process similar to that described above, 2,6bis(4‘-hydroxy-2’-pyridyl)-4-hydroxypyridine was obtained from the corresponding triamino derivative. EXPERIMENTAL

In view of the sensitiveness of the color2 given by ferrous complexes of certain alkyl and aryl substituted 2,6-bis(2’-pyridyl)pyridines prepared in this l a b ~ r a t o r y ,it~ was considered desirable to introduce other substituents into this molecule in the positions pura to the ring-nitrogen. This was accomplished by a method analogous to that previously used4 in the preparation of similar derivatives of bipyridine.

d,B-Bis(d’-pyridyl)pyridinetrioxide. A solution of 8 g. of 2,6bis(2’-pyridyl)pyridine in 42 ml. of glacial acetic acid and 27 ml. of 30% hydrogen peroxide was heated for 2 hr. at 80”. After addition of a further 27-ml. portion of hydrogen peroxide the temperature was raised to 90’ and maintained there for 18 hr. The mixture was then poured into 400 ml. of acetone. After standing several hours, the precipitate was washed with acetone and removed by filtration. From two such runs 17 g. (88.170)of pure product was obtained melting at 320-321’ dec. It can be crystallized from methyl cellosolve-water. Anal. Calcd. for ClsHI1NsOa: C, 64.05; H, 3.94. Found: C, 63.80; H, 4.06. Z,Ci-Bis(4’-nitro-Z’-pyridyl)-$-nitropyridine trioxide. To a cooled mixture of 5.7 g. of 2,6bisl2’-pyridyl)pyridine trioxide, 20 ml. of concd. sulfuric acid, and 4.8 ml. of 20y0 fuming sulfuric acid was added 9 6 ml. of fuming nitric acid (sp. gr. 1.6). The mixture was heated und‘er a reflux condenser at 100’ for 1 hr. and then at 120’ for 4 hr. The contents of the flask were then poured into ice water and filtered. The tri-N-oxide of I was prepared in satisfactory The precipitate, after washing first with sodium bicarbonate yield by the action on I of hydrogen peroxide in solution and then with water, was dried, and crystallized acetic acid. Nitration of the oxide yielded the tri- from aqueous pyridine. The yield of nitro oxide, melting nitro oxide in relatively poor yield. This could be at 268-269’ was 1.9 g. (22.6%). An analyticalsample melted converted to 2,6-bis (4’-amino-2’-pyridyl)-4-amino- a t 272-273 Calcd. for CisHsNaOo: C, 43.27; H, 1.92. Found: pyridine directly by catalytic reduction in acetic C,Anal. 43.32; H, 2.06. acid. Removal of oxygen from I1 to form the tri2,6-Bis(4’-nitro-8’-py r i d yl) -4-nitrop yridine. A mixture of nitro compound was accomplished by refluxing 1.5 g. of 2,6bis( 4’-nitro-2’-pyridyl)-4-nitropyridinetrioxide with phosphorus trichloride. It was found that and 15 ml. of phosphorus trichloride was refluxed for 18 hr., poured on ice and made alkaline with ammonium with these trioxides removal of oxygen is more then hydroxide. The precipitate was removed, dried, and crystaldifficult than with the dioxides of bipyridine, lized from benzene. The yield of pure product, melting at so that phosphorus trichloride must be used either 225-226“ WM 0.7 g. (53.870). Anal. Calcd. for ClaHsNeOe: C, 48.91; H, 2.17. Found: undiluted or in more concentrated solutions in C, 48.86; H, 2.42. chloroform than previously. d,G-Bis(4’-amino-B‘-pyn’dyl)-4-aminopyridine. A mixture Treatment of I1 with acetyl chloride followed by of 2.2 g. of 2,6bis(4’-nitro-2‘-pyridyl)-4nitropyridinetriphosphorus trichloride in chloroform yielded 2,6- oxide, 5.0 ml. of glacial acetic acid, 1ml. of acetic anhydride, bis(4’-chloro-2’-pyridyl)-4-chloropyridine.Trimeth- and 2 g. of 10% palladium-on-carbon was shaken a t 30 lbs. oxy- and -ethoxy-2,6-bis(2’-pyridyl)pyridinesre- preeaure of hydrogen in a Parr reduction apparatus until more hydrogen was absorbed (5 hr.). After removal of sulted from the action of sodium methoxide and no catalyst by filtration the solution was evaporated to dryethoxide on I1 followed by removal of oxygen by ness using an aspirator. Enough water was added to dissolve phosphorus trichloride in chloroform. the resulting solid, followed by concd. ammonium hydroxide. A previous attempt4to hydrolyze 4,4’-dimethoxy- On standing in a refrigerator, crystals of the monohydrate 2,2’-bipyridine with hydriodic acid resulted in the separated (1.2 g. or 75%), which melted a t 288’. An analytical sample was prepared (m.p. 292-293”) by crystallization formation of the 4-hydroxy-4‘-methoxy derivative. from water. We have now obtained 4,4‘-dihydroxybipyridine Anal. Calcd. for C~IHIIN(.H~O: C, 60.81; H, 5.41; N, by the action of nitrous acid on diamino bipyridine. 28.38. Fbund: C, 60.25; H, 5.31; N, 28.03. Z,B-Bis(4’-methozy-d’-pyridyl)-4-methox~pyridine.To a The latter compound was conveniently prepared by cooled solution of 0.8 g. of sodium in 100 ml. of anhydrous methanol waa added 1.9 g. of 2,6-bis(4’-nitro-2’-pyridyl)Pnitropyridine. After stirring for 4 hr. a t 35-40’, solution wm nearly complete. The filtered methanolic solution was (1) This work was supported by Grant G 9645 from the then evaporated to dryness using an aspirator, and the residue was extracted twice with chloroform. The cold solution National Science Fonndation. (2) A. A. Schilt and G. F. Smith, Anal. Chim. A ~ t a , obtained by evaporating the chloroform to a volume of 50 ml. was treated with 10 ml. of phosphorus trichloride and 15; 567 (1956). (3) F. H. Case and T. J. Kasper, J. Am. Chem. SOC.,78, refluxed for 4 hr. After cooling and pouring on ice, the sohtion waa made alkaline with sodium hydroxide and filtered. 5842 (1956). (4)‘G. Maerker and F. H. Case, J. Am. Chem. SOC.,80, The orecbitate was dissolved in ether and the solution addeci to ihe chloroform layer in the filtrate. The residue 2745 (1958).

’.

FEBRUARY 1962 from the evaporation of the mixed solvents was then crystallized from methanol, yielding 0.9 g. (47.4%) of pure product, melting at 171-172". And. Calcd. for C18H17NaO::C. 66.87; H, 5.26; N, 13.00. Found: C, 66.70; H, 5.40; N, 13.11. $?,6-Bis(4'-ethoxy-2'-pyridyl)-4-ethoxypy~dine.To a cooled solution of 0.68 g. of sodium in 91 ml. of anhydrous ethanol was added 2.2 g. of 2,6bis(4'-nitro-2'-pyridyl)-Pnitropyridine. After stirring for 4 hr. at 60-65", solution waa nearly complete. The filtered alcoholio solution was evaporated to dryness using an aspirator and extracted with chloroform. After evaporation to a volume of 45 ml.; 9 ml. of phosphorus trichloride was added and the mixture refluxed for 3 hr. It was then poured on ice and made alkaline. The precipitate obtained, which was insoluble in ether and chloroform was dried and again treated with chloroform and phosphorus trichloride as before. After again pouring on ice and making alkaline, the product dissolved in ether and waa recovered by evaporation of the mixed solvents; yield, 0.8 g. (42.1%) of pure product melting a t 157-158'. Anal. Calcd. for C,1HpaNnOt: C, 69.04; H, 6.30. Found: C, 68.96; H, 6.49. 2,6-Bis(4'-chloro-b'-pyridyl)-4-chloropyridine. To a SUSpension of 2 g. of 2,6-bis( 4'-nitro-2'-pyridyl)-bnitropyridine in 20 ml. of glacial acetic acid a t 60' was added 12 ml. of acetyl chloride. After heating for 1 hr. on the steam bath, an additional 4 ml. of acetyl chloride was added and heating continued for 1 hr. The mixture was poured on ice and the solution neutralized with sodium bicarbonate. The crude oxide precipitating after separation by filtrat,ion and drying was suspended in 35 ml. of chloroform. To the cold mixture was added 4.5 ml. of phosphorus trichloride. After standing for 1 hr., it was refluxed for one hour on the steam bath, and then poured on ice and made alkaline. The resulting precipitate, after drying, was crystallized from benzene yielding 1.2 g. (75.0%) of product, m.p. 210-211°. An analytical sample melted a t 212-213'. Anal. Calcd. for C16HsNd24: C, 53.51; H, 2.38. Found: C, 53.49; H, 2.24. 4,4'-Dihydrozy-d,d'-bipyridine. To 7.5 ml. of concd. sulfuric acid a t 0' was added 0.95 g. of sodium nitrite. The mixture was allowed to warm to room temperature and then heated a t 65" until a clear solution resulted. To this solution was then added a solution of 1.2 g. of 4,4'-diamino2,2'-bipyridine in 5 ml. of concd. sulfuric acid a t 0-5'. After standing for 15 min., the reaction mixture was poured on 40 g. of ice, and the solution was allowed to stand overnight whereupon considerable evolution of nitrogen WBE observed. On adjusting to pH 6 with sodium hydroxide, a precipitate formed which was separated and crystallized from water. The yield of pure hemihydrate melting a t 342343" was 0.7 g. (58.3%). Anal. (sample dried a t 100'). Calcd. for CloHsNzOn: C, 63.83; H, 4.26. Found: C, 63.69; H, 4.30. Calcd. for Clcr H~NzOZ.~/ZHZO: HzO, 12.56. Found: HpO, 12.57. 2,6Bis( 4'-hydroxyb'-pyrid yl)-4-hydrox yp yn'dine. The procedure for this preparation waa the same as for that of 4,4'-dihydroxy-2,2'-bipyridine. From 1.2 g. of the triamino compound was obtained 1.2 g. of crude trihydroxy compound. The pure product was obtained as a dihydrate by crystallization from water, in which it is very sparingly soluble. It melts over 400". Anal. Calcd. for CjaH11NiOs.2H20:C, 56.78; H, 4.73; HzO, 11.36. Found: C, 56.92; H, 4.85; H20, 10.97.

Acknowledgment. The author is indebted to the G. F. Smith Chemical Co. for a generous supply of 2,6-bis(2'-pyridyl)pyridine. DEPARWENTOF CHEMISTRY

TEMPLE UNIVERSITY PHILADELPHIA 22, PA.

641

NOTES

Piperidine Ilerivatives with a Sulfur-Containing Function in the 4- Position' H. BARRERA~ AND R. E. LYLE Received June 6, 1961

A series of 1-methylpipendines having a functional group containing sulfur at the 4- position was desired for screening as potential antiradiation compounds. The obvious approach to the synthesis of these compounds, the nucleophilic substitution of 1-methyl-4-chloropiperidine, failed to give the desired product.3 An alternate route to l-methyl-4mercaptopiperidine (VIII) was suggested by the ease of formation of the hydrate of l-methyl-4piperidone hydrochloride,' which would suggest that the reaction of 1-methyl-4-piperidone with hydrogen sulfide should form a gem-dithiol with an ease similar to that observed with dibenzyl ketone.b The gem-dithiol thus formed could readily be converted to the corresponding mercaptan by reduction. The reaction of 1-methyl-4-piperidone (I) with hydrogen chloride and hydrogen sulfide in ether, the procedure used for the conversion of dibenzyl ketone to the gem-dithiol,bfailed to cause the introduction of sulfur, for the amine salt I1 precipitated before reaction with hydrogen sulfide occurred. By using a solvent, isopropyl alcohol, in which the amine salt would be soluble and precipitation of the product with ether the reaction led to a colorless, sulfur-containing product which released hydrogen sulfide on heating in aqueous solution. The color tests for sulfur-containing functional were inconclusive; however, the elemental analyses corresponded to the formula of l-methyl4-thiopiperidone hydrochloride (V) which was assumed t b be a trimer because of lack of color and analogy with polymerization of other thiones.9 Molecular weight determinations on V were incon(1) This research was supported by a contract, DA-49193-MD-2034, with the Office of the Surgeon General of the U. S. Army Medical Research and Development Command. A portion of this paper was presented before the Division of Medicinal Chemistry a t the 140th American Chemical Society Meeting, Chicago, Ill., September 3-8, 1961. (2) On leave from Extractos Curtientes y Productos Quimicos, S: A., Barcelona, Spain, 1959-1960, as Research Associate at the University of New Hampshire. (3) Unpublished results, H. Barrera and R. E. Lyle. With the exception of potassium xanthate, nucleophiles led to cleavage of the heterocyclic ring of 1-methyl-4chloropiperidine. (4) R. E. Lyle, R. E. Adel, and G. G. Lyle, J. Org. Chem., 24,342 (1959). (5) G. A. Berchtold, B. E. Edwards, E. Campaigne, and M. Carmack, J . Am. Chem. SOC.,81, 3148 (1959). (6) T. L. Cairns, G. L. Evans, A. W. Larcher, and B. C. McKusick, J . Am. Chem. SOC.,74, 3982 (1952). (7) H. Rheinboldt, Chem. Ber., 60, 184 (1927). (8) I. W. Grote, J . Biol. Chem.,93, 25 (1931). (9) E. Campaigne, Chem. Revs., 39, 1 (1946).

642

VOL.

NOTES

\

I11

VI11

HCI

I

F&-h'>Sl .HCl V

I1

-

tion and as a solid mull showed weak SH stretching bands a t 2525 cm.-' and no evidence of ammonium hydrogen. These data are consistent with the molecule of water being associated with l-methyl-4,4dimercaptopiperidine. The reaction of 111 with hydrogen chloride gave a salt identical with 1methyl - 4,4 dimercaptopiperidone hydrochloride (IV). The formation of a gem-dithiol from a ketone in the absence of an acid catalyst is most unusual. The 4-carbon atom of the gem-dithiol (I11 or IV) is in the same oxidation state as that of the thione. Thus it was assumed that in a reaction the unstable dithiol would be equivalent to the thione, and reactions anticipated for the thione could be obtained by using the gem-dithiol as starting material. The reaction of the gem-dithiol (111) with phenyllithium was investigated in an attempt to prepare l-methyl-4-phenyl-4-mercaptopiperidine, but no pure product could be isolated. The reaction of the gem-dithiol with sodium borohydride in isopropyl alcohol was more successful. If the reaction product was isolated immediately so that the 1methyl-4-mercaptopiperidine (VIII) which was formed was not allowed to stand in the alkaline medium, a good yield of VI11 was obtained. The mercaptan VIIZ, however, was rapidly oxidized to the corresponding disulfide IX by air on standing in an alkaline medium. In view of the anomalous cleavage reaction observed with nucleophilic reactions of l-methyl-4chloropiperidine, a the preparation of piperidines substituted in the 4- position with sulfur-containing functional groups requires an indirect route for synthesis. The most satisfactory preparation for 1methyl-4-mercaptopiperidine (VIII) has thus been shown to be through the gem-dithiol.

-

w H,S/I-ICI

27

f"

C&-N%'CN-CH3

s-s 92HC1 VI VII. base of VI

clusive due to the poor solubility characteristics and the salt character of V. The recrystallization of polymeric l-methyl-4thiopiperidone hydrochloride (V) caused a partial conversion to a new salk (VI) which still contained sulfur but which gave negative results with the above-mentioned color-test The elemental analyses and infrared spectrum of VI showed the salt to be hydrated; however, conversion of the salt VI to the base (VII) gave a solid which could be characterized as a dispiro-l,2,4trithiolane (VII), formed by the autoxidation of the thio ketone V. Such structures have been reported as the product of the reaction of ketones with hydrogen sulfide, sulfur, and a secondary amine'O and autoxidation of thio ketones. l1 The trithiolane (VII) was remarkably stable to heat EXPERIMENTAL and hydrolysis. The reaction of hydrogen sulfide with the gemI-Methyl-4,4-dimercaptop'pe-ridinehydrochloride (IV). A diol, 1-methyl-4-piperidone hydrochloride hydrate solution of 10.1 g. of freshly prepared 1-methyl4piperidone (11),in isopiopyl alcohol led to a good yield of a hy.drochloride hydrate" (11) in 400 ml. of isopropyl alcohol salt the properties of which showed it to be the was filtered and hydrogen sulfide was bubbled into the for 3 hr. After standing overnight, the solution desired l-methyl-4,4-dimercaptopiperidinehydro- solution deposited 9.2 g. (76%) of white l-methyl-4,4dimercaptochloride (IV). The gem-dithiol IV was very reac- piperidine hydrochloride (IV), m.p. 145' with resolidificative, for it was converted to the trithiolane VI on tion and remelting at 178-180". The reaction of I V with attempted recrystallization from alcohols and nitrous acids gave a green color, with lead acetate' an orange precipitate which darkened rapidly, and with Grote's eliminated hydrogen sulfide on dissolution in reagents a red color which turned green. These tests are water. consistent with the structure assigned. A remarkable reaction was obseyed on treatAnal. Calcd. for CsH14C1N&:C1, 17.75; N, 7.01; S, 32.10. ment of the base 1-methyl-4-piperidone (I) with Found: C1, 18.12; N, 7.32; S, 32.27. Concentration of the mother liquors from the isolation of hydrogen sulfide in isopropyl alcohol, for a solid gave the trithiolane ( V I ) vide infra. separated from solution with an analysis cor- I Vl-Methyl-4,4-dimercaptopipe-ridine (111). A solution of 14.5 responding to l-methyl-4,4-dimercaptopiperidine g. of 1-methyl-4-piperidone ( I ) in 100 ml. of isopropyl alcohol hydrate (111).The infrared spectra of I11 in chloro- waa cooled, and hydrogen sulfide waa added over a period of form solution gave evidence of the water of hydra- 4 hr. The solution waa allowed to stand overnight after (lO)(a) F. Asinger, M. Thiel, and G. Lipfert, Ann., 627, 195 (1959). (b) F. A. Kincl, Chem. Be-r., 93, 1043 (1960). (11) E. Campaigne and W. B. Reid, J . Org. Chem., 12, 807 (1959).

(12) Since the gem-dithiol could not be purified by recrystalliiation, an analytical sample of the gem-dithiol w m prepared by starting with carefully purified 1-methyl4 piperidone hydrochloride hydrate (11).

FEBRUARY

1962

which time 6.4 g. of solid had been deposited. The mother liquors were treated with hydrogen sulfide for an additional 6 hr. to yield an additional 2 g. (36% total) of 1-methyl-4,4 dimercaptopiperidine hydrate (111), m.p. 48-50'. Anal. Calcd. for C&aNO&: c, 39.74; H, 8.34; s, 35.37. Found: C, 40.30, 40.59; H, 8.43, 8.66; S, 34.48, 34.27. An isopropyl alcohol solution of l-methyl-4,4-dimercaptopiperidine hydrate (11)was treated with hydrogen chloride to form the hydrochloride (IV), m.p. 142-144", resolidified, m.p. 168-171", identical in infrared spectrum with I V isolated from 1-methyl-4-piperidone hydrochloride hydrate (11). Polymer of 1-methyl-4-thiopiperidone hydrochloride (V). A solution of 22 g. of 1-methyl-4-piperidone (I)in 180 ml. of isopropyl alcohol was saturated with anhydrous hydrogen chloride, and hydrogen sulfide was then passed into the solution for 5 hr. The addition of anhydrous ether caused the precipitation of an oil which partially crystallized on standing. The solid was removed by filtration and washed with isopropyl alcohol and ether. About 18 g. (55%) of V, m.p. 178-191°, wa8 obtained, but it could not be purified by recrystallization due to decomposition. Fractional precipitation of V from the reaction mixture obtained with pure 1-methyl-4piperidone (I) gave analytically pure VI m.p. 189-191". Anal. Calcd. for (C6HlZCINS),: C, 43.49; HI 7.30; SI 19.35; C1, 21.40. Found: C, 43.40, 43.11; HI 7.32, 7.15; S, 20.15, 20.12; C1, 21.43. The reaction of V with nitrous acid gave a green color and V gave an orange precipitate which darkened rapidly on reaction with alcoholic lead acetate.' Grote's reagent8 gave with V a red-purple color which changed to violet and finally blue. Dispiro-l,B,Q-trithiolanehydrochloride (VI). A solution of 1-methyl-4-piperidone hydrochloride hydrate (11) in isopropyl alcohol prepared from 28.3 g. of pure 1-methyl-4-piperidone (I) was saturated with hydrogen sulfide. The 1methyL4,4dimercaptopiperidine hydrochloride (IV) which precipitated was removed by filtration. The filtrate was concentrated and 5.9 g. (13%)of VI, m.p. 225-227", precip itated. Recrystallization of the solid from 95% ethanol gave VI, m.p. 233" dec. Anal. Calcd. for C1~Hz4C1~N2S~: C, 39.66; HI 6.66; C1, ~.H 37.78; Z O : H, 19.51; S, 26.47. Calcd. for C I ~ H Z ~ C ~ Z N Z SC, 6.87; C1, 18.59; SI 25.22. Found: C, 38.76, 38.67; HI 6.93, 7.07; C1, 17.83; S, 27.14. The base was prepared by neutralization of a solution of VI with potassium carbonate. Recrystallization of the base (VII) from ligroin gave colorless crystals, m.p. 78-80". Anal. Calcd. for CizHzzNzSa: C, 49.61; HI 7.63; SI 33.11. Found: C, 49.85; HI 7.87; S, 32.96. 1-Methyl-4-mercaptopiperidine (VIII). To a suspension of 6.4 g. of sodium borohydride in 50 ml. of isopropyl alcohol, 20 g. of l-methyl-4,4dimercaptopiperidine hydrate (111) was added in portions. An additional 40 ml. of isopropyl alcohol was added and stirring wa8 continued for 1 hr. The reaction mixture was heated at 58" on a water bath and stirred for 2 hr. Dilute hydrochloric acid was added until all of the solid dissolved, and the acidified solution was heated on the steam bath. The solution was neutralized with 20% potassium hydroxide until the addition of alkali caused no further clouding. The aqueous mixture was extracted several times with ether, and the ether extracts were dried. The ether was removed by distillation, and the residual oil was distilled under reduced pressure to give 10.1 g. (70%) of l-methyl-4mercaptopiperidine (VIII), b.p. 62' a t 0.8 mm., and 2.5 g. (17%) of the corresponding disulfide (IX), b.p. 180°, at 0.8 mm. The two bases were converted to their hydrochlorides in isopropyl alcohol to give 1-methyl-4-mercaptopiperidine hydrochloride, m.p. 172-173", and bis( l-methyl4-piperidyl) disulfide hydrochloride, m.p. 237-238". A d . Calcd. for CeH14ClNS:C, 42.97; H, 8.41; SI 19.12. Found: C, 43.21; H, 8.54; SI 18.20.

643

NOTES

Anal. Calcd. for C ~ ~ H Z & ~ Z NC,& :43.23; H, 7.86; 19.24. Found: C, 43.17; H, 8.04; SI 18.17.

s,

DEPARTMENT OF CHEMISTRY UNIVERSITY OF NEW HAMPSHIRE N. H. DURHAM,

Azasteroids. 1111~2 MILANUSKOKOV16,8 v. T c ~ O M E AND , ~ MARCEL GUT Received June 19, 1961

The preparation of 17,!3-acetoxy-19-nor-4-azaandrost-5-en-3-one (11)and its dehydrogenation to 4-azaestradiol 170-acetate (V) has already been mentioned in a preliminary comm~nication.~ Further experiments have now demonstrated that I1 can be prepared in superior yields from 17P-acetoxy-19-nor4-oxa-androst-5-en-3-one(I) by treatment of its. benzene solution with ammonia, whereby I1 precipitates from the reaction mixture. Alkaline hydrolysis of I1 gives 17/3-hydroxy-19-nor-4aza-androst-5-en-3-one (111), which is oxidized to 19-nor-4-aza-androst-5-en-3,17-dione (IV) with chromic acid. All the 3-keto-4-nza A5-steroids have a strong ultraviolet absorption with an absorption maximum in the region of 230-235 mp in neutral solution. TABLE I 4Azacholest-5-en-3-one4

4-Azapregn-5-en-3,20-dione4 17@-Hydroxy-4-azaandrost-5-en-3-one4 17~-Acetoxy-4azaandrost-5-en-3-one4 17~-Acetoxy-4-aza-19-norandrost-5-en-3-

h"

E

233 233 233 233

13,500 13,430 13,790 13,630

one4 234 4-Aza-4methyl-5-cholesten-3-one 234 4-Aza-4-methyl-5-pregnene-3,20-dione 234 4-Aza-4methyl-5-pregnen-2O~-ol-3-one 234 4-A~a-4,1ia-dimethyl-5-androsten-17~-01%one 234

9,630 13,4906 13,1806 13,4906 13,4906

With increasing hydrochloric acid addition to the methanolic solution the maximum gradually disappears. This can be explained by the fact that the absorption is due to the form IX6 and its disappearance to the formation of the protonated X. Upon repeating4 the oxidation of I1 with selenium dioxide in tert-butyl alcohol solution with a catalytic amount of either acetic acid or pyridine, (1) Azasteroids. 11, M. Gut and M. UskokoviC, J . 078. Chem., 26,1943 (1961). (2) This study.was supported by National Institute of Health Grant H-5266. (3) Hoffmann-LaRoche, Inc., Nutles 10, N. J. (4) M. UskokoviC and. M. Gut, Helv..Chim. Acta, 42, 2258 (1959). (5) Dr. N. J. 'Doorenbos, private communication. (6) NMR gives peaks for 1 H each, attached to N and Cs.

644

VOL. 27

NOTES OCOCH3

did not melt under 330'. Lack of material allowed only a qualitative determination of the ultraviolet a t 232 mp and 310spectrum, which showed ,A, 317 mp, which might be suggestive to formulation

I

~111.9

OCOCHB

QCOCH,

'

OCOCH:,

AN

O

VI11

i

OH-

The rotatory dispersion curve of ring A unsaturated amides lackdo the fine structure typical of A4-3-keto steroids. The completely aromatic characterll of 4-azaestradiol 3,174-diacetate is demonstrated by the fact that its ult,raviolet absorption maximum is not shifted by addition of acid or base. A shift of the absorption was, however, observed by subjecting the methanolic solution of either 4azaestradiol or 4-azaestradiol 17P-acetate to pH changes.

Figure 1

we found 4-azaestradiol 17p-acetate (V) to be a minor product (10%). Careful rechromatography of the fractions containing V allowed the separation I I N 9 5 % ETNANGL of a less polar product, identified as 4-azaestradiol 2 IN O.l& HCI(=IN UCI) 3,lTfi-diacetate (VI). The major product was polar 3 I N 0 . 1 KOU(= ~ IN KOH) and is formulated as log, 17p-dihydroxy-lY-nor-4azaandrost-5-en-3-one l7g-monoacetate (VII) on the basis of its properties and the reactions it undergoes. The elemental analysis indicates C19H2704N; infrared analysis shows the presence of a tertiary hydroxyl group (3520, 1200, and 1055 cm.-l). HH The hydroxyl resisted acetylation with acetic anhydride-pyridine and could not be oxidized with chromic acid. A solutbn of VI1 in acetic acid was saturated with hydrogen chloride a t 10' without undergoing aromatization, indicating that formation of a carbonium ion, followed by trans axial elimination does not occur as readily' with unsaturated lactams as with the conjugated ketone analog. A hypsochromic shift in the ultraviolet absorption, 200 250 300 350 compared to 11,from 234 mp to 230 mp is paralleled A m ? by the shift in the absorption* of 19-nor-lop-hy.Figure 2 droxytestosterone compared to 19-nortestosterone. The near infrared reveals the presence of a hydroxyl EXPERIMENTAL (1.417 p ) and of an NH-group (1.505 p ) . No color is produced with ferric chloride, thereby excluding a 1?~-Acetoxy-15-nor-~-azaandrost-5-en-~-one (11) from hydroxamic acid. A solution of VIJ in glacial acetic I. To the solution of 100 g. of 17p-acetoxy-19-nor-4-oxaacid was boiled for 11/2 hr., then evaporated, and the residue recrystallized from methanol. The prod(9) Compare M. Perelman, E. Fwkas, E. J. Fornefeld, uct underwent a transformation a t 310-315' and R. J. Kraay, and R. T. Rapala, J. Am. Chem. Soc., 82,

I,

( 7 ) Compare the aromatizations of the 17-acetate of 1Op-hydroxy-19-nor testosterone to estradiol 17-monoacetate under identical conditions. (8) J. P. Ruelas, J. Iriarte, F. Kincl, and C. Djerassi, J . Org. Chem., 23, 1744 (1958).

2402 (1960). (10) Proof of configuration for the 10-hydroxyl of VI1 can therefore not be obtained from rotatory dispersion measurements. ( 11) For 2-hydroxypyridine-pyridine equilibria compare S. F. Mason, J . Chem. SOC.,5010 (1947).

FEBRUARY

I

1962

20

47

I 2.

3 4

IN 95% ETMANOL IN 0.1 UCI IN 5.0N MCl IN 0.1n/ KOIi (= IN KOU)

&

HO N

200

645

NOTES

250

300

350

Figure 3 androst-5-en-3-one (1)12in lo00 ml. of dry benzene was introduced a stream of gaseous ammonia a t room temperature, whereby a copious precipitate deposited. The crystals were filtered off and recrystallized from methanol. The 55 g. of colorless prisms had a m.p. 301-310", dec. and the infrared and ultraviolet spectra were identical with those of a specimen obtained previ~uslg.~ 17'6-Hydroxy-lb-nor-4-azaandrwl-6-en-S-one (111) from 11. The solution of 15 g. of I1 in 150 ml. of 2% sodium hydroxide in methanol was left overnight a t room temperature, then a large amount of water was added, and the resulting suspension extracted with methylene chloride. The extract was washed with water, dried over anhydrous sodium sulfate, and evaporated to dryness. The residue was recrystallized from methanol to give 10.5 g. of 111, m.p. 248-258' (transformations at 160-180" and 244248'). [a]*: -5" ( c , 0.5 in , , ,A 234 mp ( e 8800); infrared absorption vmar 3600 CHCla); and 1020 (hydroxyl), 3200 (NH-group) and 1660 cm.-l (amide carbonyl). Anal. Calcd. for C&f,ozN: C, 74.14; H, 9.15; N, 5.09. Found: C , 74.24; H, 9.28; N, 5.32. Selenium dioxide oxidalion of 17p-acetoxy-19-nor-4-aazaandrost-6-en-S-one (11). To the solution of 5 g. of I1 in 900 ml. lert-butyl alcohol were added 9 ml. of acetic acid and 5.4 g. of selenium dioxide and the reaction mixture refluxed for 24 hr. After cooling, the liquid was decanted, the residue washed with ethyl acetate, and the combined solutions evaporated and chromatographed on silica gel. The fractions eluted u ith 25% ethyl acetate in benzene were pooled and rechromatographed (see below). The fractions eluted with 50% ethyl acetate-benzene gave, after recrystallization from ether, 650 mg. VII, m.p. >286", dec.; [CY]? -174' ( e , 0.8 in chloroform), infrared absorption maxima urnax 3520, 1200, and 1055 (tert. OH-group); 3180 and 3060 (XH-group); 1710 and 1250 (acetoxy group); 1664 cm.-l (amide carbonyl); ultraviolet 230 mp ( F 16,000). absorption maximum, , ,A Anal. Calcd. for C1~H27O4N: C, 68.44; H, 8.16; N, 3.20. Found: C, 68.17; H, 8.08; F,4.05. The mixture which wa8 eluted with 25% ethyl acetatebenzene (see above) was rechromatographed on silica gel, (12) I. A. Hartman, A. J. Tomasewski, and A. S. Dreiding, .I A n . Chem. SOC.,78,5662 (1956).

whereby the fractions with 15% ethyl acetate in benzene gave, after crystallization from methanol, 150 mg. VI, m.p. , ,A 212 >320°, dec.; [a]% -3" (c, 0.25 in chloroform); mp ( e 23,000), 254 mp ( e 16000); Amln 220 mp ( c 7600); unchanged in 0.1N HCl or 0.1.V KOH. Anal. Calcd. for CZIHZ~O~N: C, 70.56; H, 7.61; N, 3.92. Found: C, 70.60; H, 7.24; N, 4.10. The fractions eluted with 25% ethyl acetate-benzene gave 300 mg. V, [ C Y ] : +31' (c, 0.4 in chloroform); Amax 220 mp ( e 7800)18and 288 mp ( a 6600); ,,A, 259 mp ( e 3500), m.p. and infrared absorption as published previously; unchanged in 0.1 N HC1; 0.1 N KOH: Amax 221 mp ( e 16,100), 250 mp ( e 6000) (shoulder), and 334 mp ( e 5500); ,,A, 281 mp ( E 1200). 19-Aior-~-azaandrosl-5-en-S,l7-dione (IV) from 111. To the solution of 3 g. of I11 in 200 ml. methylene chloride was added 40 ml. of 2% chromic oxide solution in 80% acetic acid and shaken overnight a t room temperature. The methylene chloride layer was separated, washed with dilute sodium hydrogen sulfite solution, then with 2N sodium hydroxide solution and water, dried over anhydrous sodium sulfate, and evaporated. The crystalline residue was recrystallized from acetone to yield 2.6 g. of IV, m.p. 280-300", dec.; [ a ] y+18" (c, 0.7 in CHC13): ultraviolet absorption A,, 233 mp (C 9500) and 300-310 mp ( e 800); infrared absorption vmax 3200 (-HX), 1660 (amide carbonyl), 1740 cm.-1 (>CO). Anal. Calcd. for C17H2302N: C, 7 .69; H, 8.48; N, 5.12. Found: C, 74.59; H, 8.77; N, 4.92. Attempted dehydratzon of VII. A solution of 20 mg. of VI1 in 8 ml. of acetic acid a t 10" was saturated with hjdrogen chloride and left for 2 hours a t 5". The solvent wafi evaDorated in uacuo and the residue chromatographed on silica gel. The fractions eluted with 25% ethyl acetate in methylene chloride gave, after recrystallization from ether, 18 mg. prisms, m.p. >310°; Am,= 238 mp ( e 630). WORCESTER Fo UNDATION SHREWSBCRY, MASS.

FOR

EXPERIMENTAL BIOLOGY

(13) Prcviously reported erroneously as A,,

226 mp

( e 6460).

Spectral Properties of Some Aromatic Thiols' SIDNEY I. MILLERAND G. S. KRISHNAMURTHY Received June 39, 1961

As part of our broad concern with structurereactivity correlations, we reported on the kinetics of addition of arylthiols (nucleophiles) to a series of ethyl phenylpropiolates (electrophiles) , 2 Here we inquire whether there are independent measures of reactivity. Specifically, what relation, if any exists between the spectral properties of thiols and their nucleophilicity? There are a few reports on ultraviolet spectral (1) (a) Supported by the Office of Ordnance Research, U. S. Army. (b) Abstracted in part from the Ph.D. thesis of G. S. Krishnamurthy, Illinois Institute of Technology, June 1960. (2) G. S. Krishnamurthy and S. I. Miller, J. Am. Chem. SOL,83,3961 (1961).

646

NOTES

VOL.

27

TABLE I ULTRAVIOLET SPECTRA OF THIOPHENOLS A N D THIOPHENOLATES

p-t-CaHs P-CH~ m-CH3

H p-c1 m-COOCzHs P-NOZ

-0.197 -0.170 -0.07 0.00 0.23 0.36

-0.13 -0.12 -0.02 0.00 -0.24 0.04

1.27

0.64

data of aromatic thiol^.^ Our own data are summarized in Table I. All of the undissociated thiols had a structureless symmetrical strong absorption and a much weaker broad absorption in the region 295-265 mp.3 In going from the thiols to the corresponding thiolates, there is an expected shift of A,, to lower excitation energies, m-Carbethoxythiophenol also had a broad shoulder at ca. 244 mp a band presumably overlapped by the 222 mp band. The effect of any substituent in benzenethiol appears to displace the main absorption band to lower excitation energies. Typical irregular Vor J-shaped curves4 are obtained when AX,,, or vmax are plotted against u or UR values.5 Contrary to the suggestions of Rao, we take the view that these plots point to complex rather than simple correlations. It has been reported, however, that in a limited series molar refraction does correlate with v,ax.6 We believed that substituents of benzenethiol would influence the infrared S-H stretching frequency in a simple and systematic fashion. Certainly similar series provided examples of corrblations between Hammett or related substituent constants and frequency shifts.' Our infrared data for the S-H stretching frequency of benzenethiols are given in Table 11. The data of Josien et al. for a group of compounds not available to us are included.*" This is the most comprehensive table of such data and goes beyond that previously available.* Most strikingly these data point to little if any correlation between (3) (a) A. E. Gillam and E. S. Stern, An Introduction to Electronic Absorption Spectroscopy in Organic Chemistry, Edward Arnold, London, 1954; (b) M. 4. Murray, Anal. Chem , 21, 943 (19491; (c) American Petroleum Institute Research Project 44, Ultraviolet Spectral Data, Carnegie Institute of Technology, April 1959, Compound Serial Nos., 427, 435, 437; (d) R. N. Bapat, Proc. Ind. Acad. Sci., 50A, 183 (1959). (4) C. N. Rao, Chem. & 2nd. 1239 (1957); J . Sei. and Ind. Res., 17B, 56 (1958); Current Sci., 26,276 (1957). (5) R. W. Tnft and I. C. Lewis, J . Am. Chem. Soc., 81, 5343 (1059). (6) W. M. Schuhert, J. M. Craven, and H. Steadly, J . Am. Chem. SOC.,81, 2695 (1959), and related papers. (7) (a) H. H. Jaffe, Chem. Revs., 53, 191 (1953); (b) J. Bellamy, The Infra-red Spectra of Complex Molecules, Wiley, New York, 1958, Ch. 22.

238.6 237.0 239.5 237.0 247.5 222.6 244.0 318.0

11,000 9,500 5,800 7,500 11,000 25,000 8,300 15,000

270.2 270.2 271.0 270.0 279.0 242.0 280.5 424.5

17,700 16,000 17,500 16,000 19,500 18,350 11,500 13,150

TABLE I1 SPECTRAL SHIFTSOF THE S-H STRETCHING FREQUENCY OF ARYLTHIOLS RSH I N THE SOLVENT CARBON TETRACHLORIDE

0.0 -2.5 -2.5 -18.5 -2.5 3.5 -4.5 15.5 -10.5 14.5 4.0 1.0 16.0 15.0

0.0 -0.07 -2.39 -0.66 -0.170 0.23 -0.197 3.493 0,363 1.27 0.26 0.28 0.38 0.37

a Commercial products except as noted. Gift of PittConsul Chemical Co., "Thioxylenol" is made up of isomers of dimethylbenxenethiol. Gift of E. I. du Pont de Nemours and Go. u is taken from P. J. Bray and R. G. Barnes. J . Chem. Phys., 27, 551 (1957). Data of ref. 8a.

Av and Hammett u (or. v+), a result in sharp contrast to a recent report.sC Our conclusion from this work is that the spectral data of thiols are not very promising as measures of nucleophilic reactivity. EXPERIMENTAL

Materials. m- and p-Thiocresol were Eastman products. Benzenethiol and p-chlorothiophenol were gifts of Evans Chemetics Inc., while p-t-butylthiophenol was a gift of Pitt-Consul Chemical Co. These thiols were analyzed as 99.5 &o.5yOby amperometric titration with silver nitrate.$ p-Xitrothiophenol was prepared from p-chloronitrobenzene and sodium sulfide,'Oa m.p. 76-77' (lit.l0 77.5'). Amperometric analysis with silver nitrate indicated a purity of 97.5 f0.5%. m-Carbethoxythiophenol was prepared by esterification of the zinc reduction product of m-carboxybenzenesulfonyl~

(8) (a) M. L. Josien, C. Castinel, and P. Saumagne, Bull. SOC. chim. France, 423,648 (1957); (b) R. A. Spurr and H . F. Byers, J . Phys. Chem., 62, 425 (1958); (c) J. Jan, D. Hadzi, and G. Modena, Ricerca Sci., 30, 1065 (1960); ( d ) R. N . Bapat, 2nd. J . Phys., 33,295 (1959). (9) I. M. Kolthoff and W. E. Harris, Anal. Chem., 21. 963 (1949); 18,161 (1946). (10) (a) C. C. Price and G. W. Stacey, J . Am. Chem. Soc., 68. 499 (1946): (b) P. G. Bordwell and H. M. Andersen, J. Ah. Chek. Soc:, 75,6019 (1953).

FEBRUARY

1962

chloride.118 The reduction step gave ca. 3% yield of mmercaptobenzoic acid, m.p. 146147" (lit.11 146-147'). The thiol, 50 ml. of absolute alcohol, and 2 ml. of concentrated sulfuric acid were left for a week a t ca. 25". The crude ester wm distilled through a small column to yield mcarbethoxythiophenol, b.p. 95-100" (0.1 mm.), n2: 1.5612 [lit."b b.p. 147-149 (11 mm.) 1. Amperometric analysis with silver nitrate indicated a purity of 97.5 i 0.5%. In the above thiols, any impurities were probably the disulfides. These would not interfere with the infrared measurements of the S-H frequency. Some errors could and emax values. For example, be introduced in the ,A, Am,, 238-241 mp and log cmmax 4.2 have been reportedlg €or diphenyl disulfide as compared to Amax 237 mp and log ,,e, 3.88, which we have found for thiophenol. The ultraviolet spectra of seven undissociated thiols and their anions were taken on a Beckman DK2 recording spectrophotometer in the range 220-340 mp. Am,, and emax were determined and are summarized in Table I. Matched pairs of quartz cells of 1-cm. path length were used. Absolute alcohol was the solvent. Its impurities could not be detected spectroscopically in a 1-cm. cell when run against the reference water. Typical thiol spectra have been given by others.8 To obtain the spectrum of an undissociated thiol, one drop of concentrated hydrochloric acid was added to 100 ml. of both the sample solution and the reference solvent. This was enough to suppress the ionization of the thiophenols since the yellow color of the most acidic thiol, p-nitrothiophenol, disappears completely. To obtain the spectrum of a thiolate ion, 1 ml. of 0.1M sodium ethoxide was added to 100 ml. of both the sample solution and the reference solvent. The infrared spectra were taken on a Perkin-Elmer Model 21 spectrophotometer equipped with calcium fluoride optica. Sodium chloride or calcium fluoride cells of ca. 0.1 mm. thickness were used. The S-H region was scanned slowly a t least three times. The scale on the infrared records was 5 cm.-l per cm of chart. Spectrograde carbon tetrachloride was the usual solvent. A s the S-H stretching frequency of benzenethiol shifted from 2570.5 cm.-1 in the neat liquid to 2576.5 cm.-l when it was in dilute carbon tetrachloride solution, spectra for all thiols were taken on dilute solutions. Usually two such solutions of different dilutions were used. An internal standard spectrum of benzenethiol in carbon tetrachloride was repeatedly taken during these measurements. The precision of this standard was ca. 1.0 cm.-l while that of the other thiols was 1-2 cm.-l

647

NOTES

phenylcarbethoxymethylphosphoniumchloride (I). I n the same paper they also report that treatment of I with sodium hydroxide ( 2 ) gave a compound formulated as the betaine (111). Recently Isler and (CsHs)aP

+ ClCHzCOOCzHs ---+ ( CsHs)3~CHzCOOC~Hs,~l.nH~0 (1 )

1.n = 0 = 2

11. n

(C&)3P

/ \

C=O or ( C6H6)aP=CHCOOCzH6 (2)

\O/ I11

IV

co-workers2 showed that the bromide analogue of I gives triphenylcarbethoxymethylene phosphorane (IV) on caustic heatment, rather than 111. The discrepancy between these two reports led to a reinvestigation of the early work. Some of our results have been anticipated.a In following Michaelis's directions for the synthesis of the phosphonium chloride (I), difficulties were encountered. Attempts to recrystallize the crude product led invariably to decomposition. This problem was resolved when it became clear from interpretation of the infrared spectra, elementary analysis, and Karl Fisher determination that the product isolated was actually the dihydrate (11). This compound was best synthesized by simply mixing ethyl chloroacetate and a solution of triphenylphosphine in benzene. During standing a t room temperature, a crystalline product was formed. The reaction was arbitrarily considered to be complete after a week and the product was filtered and washed with absolute ether. The crude salt was allowed to dry in a i r 4 . e . in contact with DEPARTMENT OF CHEMISTRY water vapor-for one week. Under these conditions ILLINOIS INSTITUTE OF TECHNOLOGY the dihydrate (11), of analytical purity, was obCHICAGO 16, ILL. tained in high yield. The dihydrate holds its cyystalline water tena(11) (a) S. Smiles et al., J. Chem. SOC.,119, 1792 (1921); 121, 2022 (1922); (b) P. F. Wiley, J. Org. Chem., 16, 810 ciously. It survived recrystallization from methyl(1951). ene chloride-carbon tetrachloride. Long drying of (12) L. Bauer and J. Cymerman, J. Chem. SOC.,109 I1 over phosphorus pentoxide in vacuo raised the (1950); H. P. Koch, J. Chem. Soc., 394 (1949). melting point from 87' to 144'. This higher melting material is probably the anhydrous salt. On exposure to air, its melting point was soon lowered. Triphenylcarbethoxymethylphosphonium Since the dihydrate was accessible, stable, and convenient for further work, no attempts were Chloride Dihydrate made to characterize the anhydrous salt. An attempt to dehydrate the dihydrate (11) was WILLIAM J. CONSIDINE made by refluxing with benzene. The products were found to be triphenylmethylphosphonium Received July 6,1961 chloride (VI), carbon dioxide, ethanol and water. In a paper published in 1894 Michaelis and Gim- Formation of these products is represented by born reported' that the condensation of ethyl Equation 3. Reaction started as soon as the benzene chloroacetate and triphenylphosphinel gave tri(1) A . Michaelis and H. v. Gimborn, Rer., 27,272 (1894)

(2) 0. Isler et al., Helv. Chem. A., X L , 1242 (1957). (3) L. C. Smith, thesis, Rutgers University (1960).

648

NOTES

( C ~ H S ) ~ ~ C H ~ C O O2H20 C ~ H+ ~,C~

VOL.

27

was raised to 144' (dec.). When this dry material was stored in air for five days, its m.p. was 98" dec. [(C6H&PcH2COOH,cl]---f Dehydration of ti-iphaylcarbethoxymethylphosphonium chlo(C6Hs)3h&,c1 HzO 4- CzHbOH Cog (3) ride dihydrate (11).Triphenylcarbethoxymethylphosphonium chloride dihydrate (21.04 g.; 50 mmoles) was added to 100 V VI ml. of anhydrous benzene in a flask provided with a Deanhad reached its boiling point and was complete Stark trap and a nitrogen atmosphere. The flask was heated in a water bath and the solid melted to a yellowish oil. As within three hours. boiling initiated, an unusual amount of frothing w m obIn contrast to the above behavior, when the served. The gases (carbon dioxide) were passed through a hydrolysis of I1 was carried out a t room tempera- barium hydroxide trap to form a copious white precipitate. Boiling, under reflux, was continued for three hours, at the ture in dilute hydrochloric acid, the interesting analogue of betaine hydrochloride (V) was obtained. end of which time a lower layer with a constant volume of I .3 ml. had collected in the trap and the oily material in the The direct synthesis of V from chloroacetic acid flask had solidified. The bath was removed and the reaction and triphenylphosphine was recently reported by mixture was allowed to cool. The solid was isolated by filDenney.* Comparison of acids from both routes tration. The solid, its mother liquor, and the lower layer of the liquid in the Dean-Stark trap were worked up separately. showed them to be identical. Solid. The solid was washed with anhydrous benzene and As would be expected from the work on the dried over parafin shavings at 2 mm. An off-white solid was bromo analogue of I, treatment of I1 with aqueous obtained; yield 15.67 g. (50 mmoles; 100oJo).Recrystallizasodium hydroxide led to the phosphorane IV in a tion from methylene chloride-carbon tetrachloride gave quantitative yield; a phosphobetaine (111) was not white crystals of triphenylmethylphosphonium chloride which were stored over phosphorus pentoxide at 2 mm. for .obtained. 3 two weeks; m.p. 217-218.5' (lit. 212-21306 ; 222-224' a). The phosphorane I V is soluble in hydrochloric The infrared spectrum of this material was in excellent agreeacid and recoverable by immediate neutralization ment with that of an authentic specimen.6 with sodium hydroxide. If a solution of the phosMother liquor. The mother liquor (benzene) from the filphorane in hydrochloric acid is allowed to evapo- tration was analyzed by vapor phasd chromatography and to contain 0.73 g. of ethanol. rate slowly, at room temperature, crystals of V are found T r a p liquid. The liquid in the lower layer of the Deanobtained. Stark trap was found, by vapor phase chromatography, to consist of 0.78 g. of water and 0.53 g. of ethanol. The yield of water was 0.78 g. (43 mmoles; 87y0).The combined yield of ethanol was 1.3 g. (28 mmoles; 57%). - H+ Acid hydrolysis of triphenylcarbethoxymethylphosphonium ( C~H6)a~CH2COOC2Hs,C1 -+ chloride dihydrate (11). TriphenylcarbethoxymethylphosHzO phonium chloride dihydrate (4.21 g.; 10 mmoles) was dis( c ~ H ~ ) ~ ~ H ~ c o o H(4) , c ~ solved in 200 ml. of 1% hydrochloric acid and set aside to evaporate. After three weeks, the sticky solid left in the IV V casserole tyas dried over phosphorus pentoxide a t 2 mm. The yield was 3.58 g. (10 mmoles; 100~o).Two recrystalEXPERIMENTAL lizations from absolute ethanol-ether gave 1.15 g. (3.2 Materials. Commercial triphenylphosphine was purified mmoles; 32%) of white crystals; m.p. 219.5' dec. When by recrystallization from methanol and then hexane; mixed with authentic triphenylcarboxymethylphosphonium n1.p. 8C-81'. Ethyl chloroacetate (practical grade) was dis- chloride' (m.p. 218' dec.), the melting point was 218.5' tilled. Material used boiled a t 144-145'. Ethyl bromoacetate dec. The infrared spectrum was identical with authentic material. --Eastman White Label used as such. Triphenylcarbethoxymefhylene phosphorane (IV). A. All melting points were taken on the Mel-Temp apparatus. A Beckman Model DU Spectrophotometer was used for the Method of Mer* (from triphenylcarbethoxymethylphosultraviolet spectra and a Beckman Model IR4 for the infra- phonium bromide). Triphenylphosphine condensed with red. C, H analysis by Spang Microanalytical Laboratories, ethyl bromoacetate to give the phosphonium bromide. TiAnn Arbor, Mich. A Perkin-Elmer Model 154C Vapor tration with sodium hydroxide gave triphenylcarbethoxymethylene phosphorane (V). Recrystallizations from ethyl Fractometer was used for chromatographic analysie. Triphenylcarhethoxymethylphosphonium chloride dihydrate acetate-hexane, benzene-petroleum ether, and then cyclo(11).Ethyl chloroacetate (34.7 g.; 0.28 mole) was added to a hexane gave whiteprisms; m.p. 123.5-125.5' (lit. 116-117" s). Anal. Calrd. for CnHz,O2P: C, 75.84; HI 6.08; PI 8.89. solution of triphenylphosphine (65.6 g.; 0.25 mole) in 50 ml. of benzene. The solution was swirled to mix, stoppered and Found: C, 75.66; H, 6.12; PI 8.65. The ultraviolet spectrum had 223 mp, 267 m r set aside for one week. The crystals which formed were H mp, 268 mp 8G5, filtered, washed with ether and dried in air. Yield was 94.0 (E::". 839, 114); [ l i t . ' h ~ ~ , E t o222 1163) 1. The infrared spectrum is remarkable in that it g (0.22 mole, 89%) of a white crystalline powder; m.p. 87-88' dec (soften 86"). By recrystallization from a lacks any absorption in the carbonyl region, but rather has a methylene chloride-carbon tetrachloride mixture, large strong absorption a t 1672 cm.-I B. From triphenytcarbethoxymethylphosphonium chloride prisms of unchanged melting point were obtained. The 3521,3448 cm.-l (bonded-OH) dihydrate (11). Triphenylcarbethoxymethylphosphonium infrared spectrum had A:&, chloride dihydrate (2.10 g.; 5 mmoles) was dissolved in 20 as well as the expected bands. A n a l . Calcd. for C2zH&102P.2H20: C, 62.78; HI 6.23; P, ml. of water. Phenolphthalein was added and the solution 7.38; C1, 8.43; Hz0, 8.55. Found: C, 63.02; H, 6.30; P, 7.31; was made basic by the addition, during stirring, 0.5N sodium hydroxide to the end-point. The white crystalline C1, 8.76, 8.31; HzO (K. Fisher), 7.76, 7.75. When an aliquot of the dihydrate was crushed and stored (5) A. Michaelis and H. v. Soden, .4nn., 229, 295 (1885). over phosphorus pentoxide a t 2 mm. for one week, the m.p. (6) Kindly supplied by Prof. D. B. Denney. (7) D. B. Denney and L. C. Smith, Chem. & Ind., 290 (4) D. B. Denney and L. C. Smith, Chem. & Ind., 290 (1961). Sample supplied by Prof. D. B. Denney. (1961).

+

+

--

FEBRUARY

1962

649

NOTES

powder formed was isolated by filtration, washed thoroughly with water and dried in air. The yield was 1.74 g. (5 mmoles; 100%); m.p. 122.5-124.5-undepressed when admixed with authentic material as prepared under A . The infrared spectrum was the same as that of authentic. Acid stability of triphenylcarbethoxymethylene phosphorane (IV). Triphenylcarbethoxymethylene phosphorane (0.348 g.; 1 mmole) was dissolved in 10 ml. of 10% hydrochloric acid. Immediately upon attaining a clear solution, the pH was adjusted to the phenolphthalein end-point with a 10% sodium hydroxide solution. The white solid which formed was filtered, washed with water and dried in air. It weighed 0.330 g. (0.95 mmole; 95y07,)and was shown to be starting material by a mixed melting point determination and the congruency of infrared spectra. Acid hydrolysis of triphenylcarbethozymethylene phosphorane (IV). Triphenylcarbethoxymethylene phosphorane (3.48 g.; 10 mmoles) was dissolved in 100 ml of 10% hydrochloric acid and set aside t o evaporate. One month later, 3.55 g. (10 mmoles; 100%) of yellowish crystals were obtained. Two recrystallizations from absolute ethanol-ether gave 1.63 g. (4.6 mmoles; 46%) of white crystals; m.p. 218" dec. When mixed with authentic triphenylcarbethoxymethylphosphonium chloride (m.p. 219.5"), the melting point was 218.5' dec. The infrared spectrum waa identical with authentic material.

Acknowledgment. Grateful acknowledgment is made of the instrumental determinations of Mr. I. Simmons and of his discussions of the interpretations. Mr. H. Corbin and his associates performed the microanalysis; other than C and H. RAHWAY RESEARCH LABORATORY METALA N D THERMIT CORPORATION N. J. RAHWAY,

Bisulfite Adducts of Acrolein H. D. FINCH Received July 18, 1961

One molecular equivalent of sodium bisulfite reacts with acrolein to form a product which on hydrogenation and ring closure yields small amounts of 1,3-propane~ultone.~ As methacrolein and mesityl oxide by similar reactions yield 2methyl - 1,3 propanesultone' (3 - oxy - 2 - methylpropane - 1 - sulfonic acid sultone), and 1,1,3trimethyl-1,3-propane~ultone,~ both in good yield and as thc mesityl oxide adduct is known to be sodium 2-methyl-4-pentanone-2-sulfonate3; it might be assumed that the acrolein product is also the double bond adduct, sodium propanal-3-sulfonate. In contrast, the reaction of cinnamaldehyde with one equivalent of sodium bisulfite yields the carbonyl adduct, sodium 3-phenyl-2-propene-lhydroxy-l-~ulfonate.~ We have now obtained evidence that the initial product of the reaction between acrolein and

-

(1) C. W. Smith, D. G. Norton, and S. A. Ballard, J. Am. Chem. Soc., 75,748 (1953). ( 2 ) J. Willems, Bull. SOC. chim. Belg., 64,432 (1955). (3) A. Pinner, Ber., 15,589 (1882). (4)F. Heusler, Ber., 24, 1805 (1891).

sodium bisulfite at pH below 5 is the unsaturated carbonyl adduct, sodium l-hydroxy-2-propene-lsulfonate. The addition of sodium bisulfite t o the carbonyl group in this pH range is rapid and not strongly exothermic : NaHSOa

+ CH2=CH-CHO

+ CHz=CH-CH(

OH)S03Na

This adduct, at pH below 5 , slowly adds a second sodium bisulfite, yielding sodium 1-hydroxy-l,3propane disulfona te : CH2=CH-CH(

+

OH) S03Na NaHSOa + XaSO3-CH2-CH~-CH( OH) S03Xa

In the pH range above about 5.2 two reactions occur: The unsaturated carbonyl adduct disproportionates to the diadduct and a second mole of bisulfite adds either to the original unsaturated carbonyl adduct or to acrolein produced in the disproportionation. This reaction system is exothermic 2CH2=CH-CH(OH)S03Na + NaS03-CH~-CH2-CH( OH) S03Na

+ CH2=CH-CHO

+

CH2=CH--CH(OH)SOsNa NaHS03 + NaSOs-CH2-CH2-CH( OH) SOINa CHz=CH-CHO

+ 2 NaHSOa

--.)

NaSOa-CHz-CHz-CH(

OH) S03Na

As shown in the first equation above, the unsaturated carbonyl adduct alone disproportionates with evolution of heat at pH of about 5 . 2 or higher yielding the diadduct. In this case the expected amount of free acrolein is not obtained because acrolein reacts with the diadduct: NaS08-CH2CHzCH( OH)SO&a XaS03-CH2-CH2-CH-O-

+ xCH2=CH-CHO

--+

CH2-CH-

(

S03Na \ AH,>, NaS03-CH2-CH2-CH0

+

This last reaction has been carried out independently using a neutral, aqueous solution of recrystallized sodium l-hydroxy-l,3-propane disulfonate, and aqueous acrolein. The product probably contains a small amount of sodium propanal-3-sulfonate. So far it has not been possible to isolate sodium 1-hydroxy-2-propene-1-sulfonate from water solution. The identification of the compound is based on examination of the water solution by KMR spectroscopy, on the Raman spectrum of the solution, on ultraviolet determination of residual acrolein, and on chemical analysis. These data indicate the presence of the CHFCHR and R-CH(OH)S03Na groups. Sodium l-hydroxy-l,3-propane disulfonate when treated with aqueous hydrochloric or sulfuric acid splits out sulfur dioxide yielding sodium propanal-

6.50

NOTES

3-sulfonate, a white solid soluble in water, methanol, and 80% ethanol: NaSOJ-CH2-CH2-CH(OH)S03Na + H + +

VOL.

27

was added dropwise to the stirred solution and the p1-I and temperature were observed. At a pH of 5.2 an increase in temperature occurred and a strong acrolein odor developed in the solution. Analysis of the solution a t this point indiNaSOa-CH2CHz-CH0 H&03 N a + cated the presence of 0.002 mole/100 ml. of bisulfite ion and 0.086 mole/100 ml. of C=C. The solution was evaporated to The 2,4-dinitrophenylhydrazone and the oxime dryness at room temperature under vacuum yielding a white of this aldehyde are water-soluble, high-melting crystalline residue and distillate containing a total of 0.050 of acrolein. Recrystallization of the residue from aquesolids. Hydrogenation of the aldehyde over Raney mole ous ethanol gave white needles. nickel, conversion to free sulfonic acid, and cycliAnal. Calcd. for C ~ H B S Z O3H?O: ~N~~ C, 11.3; HI 3.8; S zation by vacupm distillation gives 1,3-propane- 20.1; C=C, 0.0 mole/100 g. Found: C, 11.6; HI 3.6; S, 21.8; C=C, 0.005 mole/100 g. sultont: in 79% over-all yield based on acrolein. The NMR spectrum of an aqueous solution of the crystals was identical with that of sodium l-hydroxy-1,3-propanedisulfonate. EXPERIMENTAL In a related experiment 280 ml. of a solution of sodium 1Preparation of sodium 1-hydroxy-1,S-propanedisulfonale. hydroxy-2-propene-1-sulfonate prepared from 58.6 g. of Acrolein, 27.0 g. (0.493 mole), was added over a period of acrolein and 95.0 g. of sodium metabisulfite at pH 3.4 was 20 min. to a stirred solution of sodium metabisulfite, 95 g. diluted with 500 ml. of ethanol containing sulfur dioxide (1 mole as sodium bisulfite), in 200 ml. of water. The temand left a t - 15'. Crystals (51 g., 32% yield on sodium biperature was maintained at 17-20' by cooling and the pH of sulfite) of sodium l-hydroxy-1,3-propanedisulfonatewere the reaction mixture was held a t 3.6-4.0 by addition of slowly deposited. small amounts of sulfur dioxide gas. Analysis of the reaction Reaction of sodium 1-hydroxy-1,s-propanedisulfonatewith mixture a t this point indicated the presence of 0.495 mole acrolein. A solution of 79.5 g. of sodium 1-hydroxy-1,3of C=C and 0.473 mole of bisulfite ion. The reaction mix- propsnedisulionate (0.25 mole) in 100 ml. of water (pH ture was left in an ice box over night (pH 3-4) after which 6.6) was added over 35 min. to a stirred solution of 84.0 g. analysis indicated 0.009 mole of C=C and 0.070 mole of (1.5 moles) of acrolein in 350 nil. of water at 20". After bisulfite ion. Ethanol was added until the mixture became stirring for 1hr. a t 20', the reaction mixture was evaporated cloudy. On standing, colorless needles were deposited. The to dryness a t room temperature, 1 mm. pressure, yielding catalysts were recrystallized from aqueous ethanol, yield 150.2 g. of water-soluble solid. Analysis of the distillate 120 g. from the evaporation indicated the presence of 0.105 mole Anal. Calcd. for C3H&07Na2.3HZO:C, 11.3; H, 3.8; 8, of carbonyl (5.9 g. as acrolein). Attempts to separate un20.1. Found: C, 11.6; H, 3.7; SI 20.7. changed sodium l-hydroxy-l,3-propanedisulfonate from The infrared spectrum of a Nujol mull of the crystals the solid by recrystallization from water and from aqueous showed no carbonyl bond (5.8 p). The high resolution NMR ethanol were unsuccessful. spectrum (40 mc./sec.) of a water solution of the crystals Anal. Found: C, 37.9; H, 5.3; S, 10.9; HzO, 5.56; hydroxyl showed higher field multiplets a t 127 cps (S-CH2-C) and value 0.166 eq./100 g., carbonyl value 0.291 eq./100 g. 163 cps (C-CH2-C) from a benzene external standard. Hydrolysis of sodium 1 -hydroxy-1,J-propanedzsulfonate. KOresonances were observed in the region from 20-60 cps A solution of 10.2 g. (0.032 mole) of sodium 1-hydroxy-1,3from benzene. propanedisulfonate in 100 ml. of water was treated with Preparation of sodium 1-hydroxy-2-propene-1 -sulfonate. 3.5 g. of 95% sulfuric acid in 25 ml. of water. The mixture Acrolein, 56.0 g. (1 mole), and a solution of sodium meta- was boiled to expel sulfur dioxide and was then neutralized to bisulfite, 95.1 g. (1 mole as sodium bisulfite), in 200 ml. of pH 7.1 nith sodium hydroxide. The solution was evaporated ,water, were added with stirring from separate burettes over to dryness under vacuum (50', 1 mm.) and the solid was a period of 45 min. to 50 ml. of water at 18-22'. The pH extracted with boiling 80% ethanol. The dry residue, 5.4 of the reaction mixture was held at 2.2 by addition of small g , contained 0.0013 mole of carbonyl by analysis. The amounts of sulfur dioxide gas. Feed rates were adjusted to extract was evaporated to dryness yielding 5.1 g. of white maintain st?jchiometric amounts of reactants in the reaction solid, soluble in water and methanol, carbonyl value 0.409 mixture. Analysis of the product indicated the presence of 'eq./IOO g. (calcd. for NaSOJCH2-CH2--CH0 .625 eq./100 0.046 mole of bisulfite ion and 0.975 mole of C=C. The 9). A solution of 0.6 g. of this solid in water reacted with 0.4 ultraviolet spectrum of the solution indicated less than 0.6 g. of 2,4dinitrophenylhydrazine yielding 0.4 g. of 2,4wt. yGacrolein in the product. The Raman spectrum indi- dinitrophenylhydrazone, a yellow, water-soluble solid cated the presence of C=C (1648 cm.-I) and the RCH- which was recrystallized from ethanol-water, m.p. 227-229' (0H)SOJa group (1046 cm.-l). No carbonyl band (1750 dec. Anal. Calcd. for CoH$N407SNa.H@: C, 30.2; H, 3.08; cm.-') was observed. The high resolution NMR spectrum (40 mc./sec.) showed two higher field multiplets a t 28 and 8, 8.94. Found: C, 29.5; H, 3.1; S, 9.0. 54 C.P.S.( C H e C and C=CHR) from a benzene external The infrared spectrum indicated the presence of water of standard. No resonances were found in the range from crystallization. 120-170 c.p.s. from benzene. The oxime of the aldehyde was prepared as follows. A soluDilution of the reaction product with ethanol containing tion of 5.5 g. of sodium l-hydroxy-l,3-propanedisulfonate sulfur dioxide gave a cloudy solution which slowly deposited in 50 ml. of water was treated with 2.5 g. of 95% sulfuric crystals of sodium l-hydroxy-l,3-propanedisulfonate.Evap- acid. The solution was boiled to eliminate sulfur dioxide oration of water from the crude reaction mixture at 0' and neutralized to xylene cyanol-methyl orange indicator. under vacuum gave a dry salt which sme!led strongly of Hydroxylamine hydrochloride, 1.2 g., was added and the acrolein and contained 0.158 mole of C=C/lOO g. (Calcd. solution was again neutralized (0.0163 eq. of sodium hyfor C3H5S04Na:0.625 mole/100 g.) droxide required, 94% of the sodium l-hydroxy-1,3-propane Disproportionation of sodium-l-hydroxy-d-propene-l-sulfo- disulfonate charged). The solution was evaporated to drynate. A solution (90 ml.) of sodium-l-hydroxy-2-propene-l- ness (50°, 1 mm.) and the dry solid was extracted with sulfonate was prepared as above from 31.2 g. (0.327 moles boiling ethanol. On cooling the extract crystals were deposas sodium bisulfite) of sodium metabisiilfate and 18.3 g. ited. These were recrystallized from 80% ethanol yielding (0.327 mole) of acrolein. Analysis of this solution indicated white crystals, m.p. 215-220" with decomposition. Anal. Calcd. for CaHe04NSNa.HzO: C, 18.6; H, 4.15; S, the presence of 0.0146 mole/100 ml. of bisulfite ion and 0.318 mole/100 ml. of C=C. Aqueous sodium hydroxide 16.6. Found: C, 19.0; HI 4.2; S,16.9; SO4--, 0.8.

+

+

FEBRUARY

1962

651

NOTES

The infrared spectrum indicated the presence of water of crystallization. Conversion of sodium l-hydroxy-l,3-propanedisulfonateto l,$propanesultone. A solution of 50.0 g. of crude sodium 1-hydroxy-1,3-propanedisulfonate (prepared from 8.9 g. of 95% acrolein and 29.8 g. of sodium metabisulfite) in 200 ml. of water was treated with 20.0 g. of 36% hydrochloric acid. The solution was boiled until the odor of sulfur dioxide was gone from the vapors (30 min). The solution was cooled, neutralized t o p H 7.05 with aqueous sodium hydroxide, and hydrogenated over Raney nickel a t 32-80' and 1300-770 p.s.i.g. Hydrogen adsorption amounted to 0.16 mole. The catalyst was removed by filtration. The filtrate was passed over Dowex 50 ( H + ) ion exchange resin to remove sodium ion. The solution was concentrated under vacuum and the bottoms were distilled, yielding 14.6 g. of propanesultone b.p. 96' (1 mm.), ester value, 0.819 eq./100 g., calcd. ester value 0.820 eq./100 g. (79% conversion on acrolein).

Acknowledgment. The author is indebted to J. L. Jungnickel and A. C. Jones for assistance with the NMR and Raman spectra. SHELLDEVELOPMENT Co. EMERYVILLE, CALIF.

Reactions of Hindered Phenols. 111. Reaction of Nitrous Acid with Hindered Phenols

0

0

I I1 X = C1, Br, CH2, i-C3H7

out a large amount of work on the nitrosation of substituted phenols.* We were interested in studying the reaction of nitrous acid on sterically hindered phenols. 2,6-Di-tert-butylphenol (111) gave on treatment with nitrous acid, an excellent yield of a compound melting at 221-222'. The ultraviolet spectrum showed A,, 302,418mw, emu 15,300 and 3700, respectively, which indicates predominantly a monoxime structure. The infrared spectrum (Nujol mull), 3330 cm.-l (OH), 1613 cm.-l (C=O), 1560 cm.-l (C=N), and 1042 cm.-l (Ny.OH stretching), supports an

xe I11

NOH IV

X Received July 1 1 , 1961

The tautomeric behavior of nitrosophenols (quinone oximes) is well known and on the basis of electronic spectra, Havinga and co-workers' have shown that p-nitrosophenol exists in solution as the phenol along with the quinone monoxime, whereas in the solid state it occurs as the oxime. Hadzi has recently shown on the basis of infrared studies that in the solid state it could be represented as the monoxime and in chloroform solution, the oxime structure predominates.6 X-ray determination of 3-chloroquinone-4-oxime and 3-methyl-6-chloroquinone-4-oxime has indicated that, the molecules exist in the quinone oxime form.6 If the quinone is sterically hindered by substituents in the ortho position as in I, then the product obtained by the action of hydroxylamine has the structure 11. Hodgson and co-workers have carried

'

(1) Part 11, Ref. 17.

(2) Deceased. (3) Present address: National Chemical Laboratory, Poona 8, India. (4) E. Havinga and A. Schors, Rec. trav. chim., 69, 457 (1950); 70, 59 (1951); A. Schors, A. Kraaijeveld, and E. Havinga, Rec. truv. chim., 74, 1243 (1955), see also L. C. Anderson and M. B. Geiger, J. Am. Chem. SOC.,54, 3064 (1932); L. C. Anderson and R. L. Yanke, J. Am. Chem. SOC.,56, 732 (1934). (5) D. Hadzi, J. Chem. Soc., 2725 (1956). (6) C. Romers, C. B. Shoemaker, and E. Fischmann, Rec. trav. chim., 16, 490 (1957). (7) F. Kehrmann, Ber., 21, 3315 (1888); 22, 3263 (1889); 23, 130 (1890). J. Prakt. Chem., [2] 40, 188, 257 (1889); [2] 42, 134 (1890).

=

tert-butyl

oxime structure (IV).6 Metro9 obtained a compound melting a t 219-220' by treating 2,6-di-tertbutylbenzoquinone with hydroxylamine hydrochloride. This obviously has the identical structure (IV). A number of methods are known for the preparation of 2,6-di-tert-butylbenzoquinone (V). Since the oxime (IV) was obtained in almost quantitative yield, the hydrolysis of the same appeared to be 0

0 V = tert-butyl

X a simple route for the preparation of V. Thus by the hydrolysis of the oxime (IV) with 20% hydrochloric acid in the presence of cuprous oxide, a 75% yield of V was obtained. Hart and Cassis16found that the action of nitric acid and acetic acid on 2,6-di-fert-butylphenol (8) H. H . Hodgson, J. Chem. Snc., 1401 (1931), and earlier papere. (9) S. J. Metro, J Am. Chem. SOC.,77, 2901 (1955). (10) A. F. Rickel and E. C. Koovman. " , J. Chem. SOP.. 3211 (1953). (111 C. F. H. Allen and D. M. Burness (to Kodak). U.'S. Patent 2,657,222. (12) E. Miiller and K. Ley, Ber., 89, 1402 (1956). (13) E. Miiller and K. Ley, Ber., 88, 601 (1955). (14) C. D. Cook, R. C. Woodworth, and P. Fianu, J . Am. Chem. Soc., 78, 4159 (1956). (15) H . Hart and F. A. Cassis, J . Am. Chem SOC.,73, 3179 (1951).

652

VOL. 27

NOTES

(111) gave 2,4-dinitro-6-tert-butylphenol or 3,5,3',5'tetra-tert-butyl-4,4'-diphenoquinone, depending upon the condition of the reaction. The nitrosophenols or the quinone oximes on oxidation with alkaline potassium ferricyanide are known t o give the corresponding nitrophenols.16 However, when a benzene solution of the quinone oxime (IV) was treated with alkaline potassium ferricyanide with a view to prepare 4-nitro-2,6di-tert-butylphenol, an excellent yield of a crystalline product, m.p. 141-142' dec., was obtained. The compound had an analysis corresponding to the molecular formula C42HOaOb?\Tz and had a molecular weight of 662 (depression of freezing point; Rast method could not be used because of the highly colored melt). The ultra-violet spectrum showed A,, 225, 293, and 310 mp; emax 12,400, 32,000, and 33,600 respectively. This indicates the presence of a dienone grouping and also the chromophoric group of the quinone oxime. The infra-red spectrum indicated the absence of hydroxyl and nitro groups but exhibited the characteristic twin band a t 1665 cm." and 1640 cm.-l due to the dienone grouping.'' When one part of the compound was heated with 50% sulfuric acid, two parts of the quinone oxime (IV), and approximately 0.5 parts of the benzoquinone (V) (calculated on the theoretical yields) were obtained. From this evidence, the structure VI has been proposed for the oxidation product.

-6 IX

c

NOH IV

+

II NO.

-

a+b

N

-

'0

x' 'x

1BOMtS"B

-

;3"-O 0 VI1

Ng0 b

a

i

)(x N

\

0 VI11

with lead dioxide in an ether solution. This showed the necessity of both base and an oxidizing agent for the reaction. EXPERIMENTAL

Preparation of 2,6-di-tert-butylbenzoquinoneoxime-4 (IV). To a solution of 2,6-di-tert-butylphenol (3 9 . ) in ethanol (25 ml.), concentrated hydrochloric acid (2 ml.) was added and the solution cooled to -5'. To the cooled mixture, a solution of sodium nitrite (1.1 g. in 5 ml. of water) was gradually added under vigorous agitation, maintaining the temperature at 0 to -5'. After complete addition (15 min.), the yellow product was agitated 30 min. longer and poured into OQ X\ ice water. The yellow precipitate (3.4 g.) gave on crystallization from benzene, shining yellow plates which melted a t 22 1-222 O , N. Anal. Calcd for C.,HLLO~N: C, 71.4; H, 9.0. Found: C, 71.3; H, 9.2. The light absorption spectrum showed Xs..' 302 and 418 mp; cmax 15,300 and 3700 respectively. Acetyl derivatave of the quinone oxime (IV). The quinone oxime (1.2 g.) was heated under reflux with acetic acid (5 ml.) and acetic anhydride (1 ml.) for about 10 min. The solution after cooling, was poured into water, collected on a filter and dried (1.1 g.). The acetyl derivative crystallized Although it is not possible to propose a definite from petroleum ether (b.p. 60-80') as pale yellow rhombic m.p. 102". mechanism, compound VI appears to be formed plates, Anal. Calcd. for CleHaOaPJ: C, 69 3; H, 8.3. Found: C, from IV by the following sequence of reactions. fi9.7; H, 8.4. The attack of base on the tertiary nitroso comPreparation of 8,f3-d~-tert-butylbenzoquinone (V). The pound (VII) possibly removes nitrous acid to give quinone oxime (0.45 g.) was dissolved in Methyl Cellosolve the phenol which forms the radical (VIII), and (10 ml.), acetone (1 ml.) and hydrochloric acid (34%, 5 ml. in 1.5 ml. of water), and cuprous oxide (1 g.) was added coupling with (a) gives the compound VI. to it. The mixture was refluxed for 1 hr. The solution was Attempts to condense two moles of the oxime carefully steam distilled and from the distillate a yellow (IV) with one mole of the benzoquinone (V) using crystalline product (0.33 g.) m.p. 68", identified as 2,6p-toluenesulfonic acid as the dehydrating agent, di-tert-butylbenzoquinone ( 75y0), was obtained. Reaction of the quinone oxime with alkaline potassium ferriwere not successful. It was also not possible to cyanide. To a vigorously stirred mixture of benzene (60 ml.), obtain VI by carrying out the reaction in alkali water (30 ml.), potassium ferricyanide (10 9.) and potassium only without potassium ferricyanide. Compound VI hydroxide (2 g.), a solution of the oxime (3.2 g.) in benzene could not also be obtained by oxidation of IV (200 ml.) was added during 1 hr. The reaction was carried out under nitrogen atmosphere. After stirring the solution for 4 more hr., the aqueous layer was separated, washed with (16) H. H. Hodgson and F. H. Moore, J. Chem. SOC., water, dried, and the benzene removed under reduced pres2260 (1925). H. H. Hodason and J. S. Wianall, . J. Chem. SOC,. sure (nitrogen atmosphere). This gave a colorless product (2.7 g.). From the filtrate which was colored dark purple, 320 ( i 9 2 8 i ' (17) M. S. Kharasch and B. S.Joshi, J . Org. Chem., 22, 50 mg. more of the colorless crystalline solid was obtained. It was soluble in ether, petroleum ether, and benzene. These 1439 (1957). I

FEBRUARY

1962

NOTES

653

solutions decomposed to a dark red solution in a few hours by paper chromatography. Connors and Ross' in a after contact with air. The solid (0.7g.) on quick CrYStaba- recent work have obtained by the opening of the tion from acetone (200ml.) under nitrogen atmosphere gave same piperazinedione only one peptide, VI. colorless long needles which melted a t 141-142'. Theoretically it became interesting to study Anal. Calcd. for Cd2H~06N2: C, 74.9; H, 9.0; N, 4.2. quantitatively the yield of the two peptides exFound: C, 75.0,75.2,74.3;H,9.1,9.3,9.0;N,4.3,4.0. Hydrolysis of compound VI- The crystalline compound pected after the hydrolysis. One could expect that (0.2 g.) Was dissolved in acetone (25 m1.1 by and assymetryof the molecule will favor the opening 50% sulfuric acid (5 ml.) was added to it. This was gently heated under reflux for about 5 min. The acetone was re- of one Peptide bond Over the other* However, moved under reduced pressure, the resulting solid dissolved there are practical difficulties in this study because in benzene and extracted with 5% sodium hydroxide. The of the similarity in the chemical properties of benzene layer on drying and concentratlon gave (0.0779.) the two peptides and their abnormal reaction with of the impure quinone. Steam distillation gave 34 mg. of ninhydrin.6 To overcome this, we incorporated in pure 2,6-di-tert-but,ylbenzoquinone. layer was acidified and the precipitate col- this substituted 2,5-piperazinedione a radioactive The lected (0.12 g.). Crystallization from benzene gave shining carbon. Starting from glycine-l-C14 we have preyellow flakes, m.p. 220', which were identical with the pared IV by the carbodiimide method. This peptide quinone oxime (IV), in their melting point, ultraviolet, and was then into the correspond~gpiperainfrared spectra.

POONA, INDIA

A Study with C" of the Hydrolysis of Unsymmetrical 2,SPiperazinedione into Dipeptides PATRICE TAILLEUR AND LOUIS BERLJNGUET Received July 18, 1961

zinedione (V), which after hydrolysis gave the two radioactive peptides. These were separated by paper chromatography and the radioactivity of each was determined. The relative proportion of the two peptides was 58% for IV and 42% for VI. In this case, the difference in the respective yields is not sufficient to allow for theoretical considerations. However, this new method using C14 incorporated into the piperaxinedione appears the ideal one for similar studies with other unsymmetrical piperazinediones.

The hydrolysis of 2,5-piperaxinedione has been extensively With unsymmetrical 2,5EXPERIMENTAL piperazinediones it should in theory give a mixture of ~ ~ ~ a r ~ ~ e n z o x y g z y ~(I). ~ eA ~ lsolution ~ c 1 4 containing two dipeptides. Previous work was done a t that 5.05 mg. of glycine-1-~14(total activity, 0.1 mc.) in IO ml. of time when the modern techniques of chromatog- water was made. An aliquot of 4 ml. was taken and 2 g. of raphy and electrophoresis %,ere unknown. The glycine were dissolved in it. Then sodium hydroxide and benzyl chloroformate were added and I waa isolated aa usual. analytical methods then used were based on the Yield: 4.4 g. (79%) m.p. 119-120°.6 Specific activity: 3.3 titration of free carboxyl or amino group liberated los c.p,m.~mg. during the hydrolysis. The emphasis was therefore Anal. Calcd. for CloHIINO,:N, 6.70.Found: N,6.75. more on the study of the ideal conditions and of the Bmzylamino-1-cyclopentanecarboxylate (11). This comrate of the hydrolysis rather than on the relative pound was prepared by a method recently descrihed.6 Benzyl N-carbobenzoxyglycgl-1-C'~-amino-l-cycZopentaneproportions and identities Of the products formed' carboxylate (111). To a solution of 4.0 g. of I, and 4.4 g. of I1 As no definite rules have been laid out for the in 40 ml. of tetrahydrofuran was added 4.2 g. of N,Niopening of unsymmetrical 2,5-piperazinediones1 dicyclohexylcarbodiimide, and I11 was isolated according to this method of synthesis of dipeptides, first pro- Sheehan and Hesss by recrystallizing with ethyl acetate and posed by Fischer and Shrauth,6 has been neglected. Petroleum ether. This yielded 6.8 g. (86%) of 111, m.p. 107-108°. Specific activity: 1.6 X 10' c.p.m./mg. In a recent publication' it has been shown that Anal. Calcd. for C28H26N205: N, 6.83. Found: N, 6.94. the partial hydrolysis of l14-diazaspiro[4.5]decaneGlycyl-l-C14-amino-l-cyclopentanecarboxylicacid (IV). 2,5-dione (V) by 1N hydrochloric acid gives two A solution of 4.1 g. of I11 in 30 ml. of ethanol containing a dipeptides: glycylamino-1 cyclopentanecarboxylic little acetic acid WLW hydrogenated for 6 hr. over 0.1 g. of palladium 10% on carbon. The catalyst was filtered, and the acid (IV) and amino-~-cyc~opentanecarboxylgly-solution evaporated to dryness, washed with acetone, and cine (VI),these two peptides having been identified filtered. The insoluble peptide ww recrystallized from water 1. s. Yaichnikov, J . Russ. Phys. Chem. sOc.9 58, 879 (1926). (2) E. Abderhalden and Mahn, z. Phy-qiol. them., 174,47(1928). w. Bas, J . ( 3 ) p. A* Levene, R. E. Steiger, and Chem.3 sl, 697 (lg29);.' A. Levene, R. E. steiger, A. Rothen, and M. Osaki, J . Biol. Chem.,86, 723 (1930). (4) M. Ludtke, Z . Physiol. Chem. Hoppe-Seyler's, 141, 100 (1924). (5) E.Fischer and W. Schrauth, Ann.,354,21(1907). ( 6 ) P.Tailleur and L. Berlinguet, Can. J . Chem.,39, 1309 (1961).

and acetone. Yield: 1.3 g. (70%), m.p. 277O.e Specific activity: 2.8 X 10' c.p.m./mg. Anal. C&d. for C,Hl4NN2Os: N,15.05. Found: N, 15.08. 1,4-Diazospir0[4.6] decane-2,6-dione-b-C14 (V). A mixture of 1 g. of IV and 6 g. of &naphthol9 was heated for 3 hr. a t 145' in an oil bath, with occasional stirring. After cooling, the yellowish Midue was thoroughly extracted two or three times with ether to remove the &naphthol. After djsolving

(7) T.A. Connors and W. C. J. Ross, J . Chem.SOC.,2124 (1960). ( 8 ) J. C.Sheehan and G. P. Hess, J . Am. Chem. Soc.,77, 1067 (1955). (9) N.Lichtenstein, J . Am. Chem. Soc., 60,560(1938).

654

NOTE8

VOL.

27

Diacetylorthnnilamide

JOHN G. TOPLISS Reeeived July SI, 1061

Ekbom' heated orthanilamide (I. RI = Rz = Ra H) with acetic anhydride and obtained a product m.p. 191.5-192.5° to which he assigned the structure I (R, = R2 = COCH3, R3 = 1%).Parke and =

Fig. 1.

l ~ ~ , , i , , , , ~ l , ~ , , , , , ~ , l , j ~ ~ ~ *:iiti.r ,,, T>:wti:il liydmlyri.:

N/R~

of V

aLk,

in a little hoiling water, the residue was purified with charI coal and crystallized out upon cwling. Yield: 0.63 g. (70Y'), m.p. 276". Sperifie activity: 2.2 X IO' c.p.m./mg. A d . Cdcd. for C8HsNr0.: N, 16.67. Found: 16.63. Williams2 stated that acetic anhydride and orthParlial hydrolysis of 1,4-di~aspiro14.6]&canc~,6diaeanilamide gave mainly diacetylorthanilamide and 6-C". The piperrrsinedione (0.2 g.) was hydrolyzed hy dissolving in 10 ml. of IN hydrochloric acid and hy hoiling that acetylation with pyridine and acetic anhydride during 5-6 minutes. The solution was cooled and the volume yielded the same compound almost quantitatively. was exactly completed to 10 ml. An aliquot of 0.1 ml. was The latter authors reported a melting point of 190" suhmitt,ed to paper ChromatoRraphy in 2,4,6trimethyl- and named this product in the Experimental secpyridine (1 part), 2,4lutidine (85%) (1 part), and water (2 parts), in order to separate the t w o peptides and the non- tion o-diacetylaminobenzenesulfonamide, thus assigning the eame structure as Ekbom. Recently, hydrolyzed pipernzinedione.

'NH-co

NH?

V

CH,--CH,

IV

Jr H2N

CO-KH-CHz-COOH

\-/

CH,--~~H> VI Measure oj the radiaaclim'ty (a) By e l u l i a from fhe paper. The paper hand was sectioned snd the radioactive spots of the two peptides and of the nonhydrolyzed piperazinedione were eluted with water. This water was collected on plnncheta which wme dried under an infrared light,. The mdicactivity was determined using a Nuclear Chicago detector D 47. T h e mean actual counts for fifteen determinations were 59 X 10'c.p.m. for IV, 43 X 10'c.p.m. for VI, and M) X 10' c.p.m. for the nonhydrolyzed V. This gave a proportion of 58% in favor of IV and 42% in favor of VI. ( b ) B y direci meamre on the paper. The paper hand was passed directly into a Nuclenr Chicago Aetigrsph. The areas of the two first curves ohtnined (Fig. 1) were determined and found to he in a proportion of 58% for IV and 42% for VI. The paper hand was nbo placed on a K;dak X-Ray Royal Blue Film during seven days, after which time the film was developed (Fig. 1). Three spots were found and their relative intensities were of the same order as the one3 found by the artigraph and the elution method.

the Acknmkdgment' The authors are National and Medical Research Councils of Canada and to Dr. L. M. Rabineau for the radioactivity determinations. DEPARTMENTOF BIOCAEMISTRY LIVAL UNIYER~IW QUEBEC.CANADA

Yale, Losee, and Bemstein3 in referring to the work of Parke and Williams showed the structure of the acetylation product as I (RI = R? = COCH3, R3 = HI. We have prepared diacetylorthanilamide, m.p. 196O, using pyridine and acetic anhydride according to the procedure of Parke and Williams.2However, a consideration of infrared spectral' and pK.' data shows that the compound is Z-acetylsulfamylacetanilide (I. RI = H, R? = & = COCHa). Two bands of medium to strong intensity attributable to -C=O vibrational modes are present in the infrared spectrum of diacetylorthanilamide a t 5.78 p and 6.00 p. The hand a t 6.00 p is also found in 2sulfamylacetanilide' (I. R1 = R3 = H, FL = COCH,) and clearly results from the presence of the acetamido substituent. The band a t 5.78 p c o r m sponds well with the expected lower wave length absorption of -C=O present in the group -SOr

0

II

N H C C H 3 and correlates with the absorption a t 5.80 p shown hy 2-acetylsulfamyl-N-methylacetanilide' (I. R1= CH3, Fb = R3 = COCH3). Further evidence is obtained by a consideration of absorp tions due to N-H stretchingvihrations. Diacetylorthanilamide has a sharp band in it.s infrared spectrum a t 2.95 p corresponding to the N-H absorption (1) Ekhom, &haw, K. Svmka Vel. Akad. Handl.. 27 (II), 3 (1942). (2) D. V. Parke and R. T. Williams, J . C h a . Soc., 176u

1950. (3)

H. L. Yale, K. Losee, and J. Bernstein, J . Am. C h . Soc., 82. 2042 (19W. (4) The infrared spectra of the compounds were determined as Nujol malls. ( 5 ) Determined in 66% dimethylformamide solution. (6) I.. RaRa, F a m o Ed. Sa'.,12, 483 (1957).

FEBRUARY

1962

655

NOTES

from an acetamido group. This band is absent in the spectrum of 2-acetylsulfamyl-N-methylacetanilide. Moreover no bands are present in diacetylorthanilamide in the 3-p region which are normally associated with the presence of an unsubstituted -SO2NH2 group. Confirmation of the fact that diacetylorthanilamide contains an acetylated sulfamyl group was obtained from a study of pK, data since the p K , value for this compound is 5.33, which may be compared with the figure 6.2 for 2acetylsulfamyl-N-methylacetanilide. In comparison, o-sulfamylacetanilide is so weakly acidic that its alkali titration curve does not show a break. These facts are only consistent with the formulation of diacetylorthanilamide as 2-acetylsulfamylacetanilide (I. R1 = H, Rz = Rs = COCHI). Acknowledgment. The author is indebted to Mr. R. Wayne for discussions in connection with the interpretation of the infrared spectra and Mr. J. McGlotten for the acidity studies. MEDICINAL CHEMICAL RESEARCH DEPARTMENT SCHERING CORPORATION N. J. BLOOMFIELD,

Facile Preparation of 178-Hydroxy-St9androstan-3-one and Its 17a-Methyl Derivatives R. BRUCE GABBARD AND ALBERTSEGALOFF Received JUIY 31, 1961

Investigation on correlation of steroid structure with androgenic activity required availability of 50-androstane derivatives. Accordingly, we investigated the stereochemistry of catalytic hydrogenation of the carbon-carbon double bond of A4-3-ketones under basic conditions. Addition of potassium hydroxide to the catalytic hydrogenation of A4-3-ketonesof the ergostane or spirostane series has been reported to lead to A/B cis (50) To our knowledge, however, use of potassium hydroxide in the catalytic hydrogenation of A4-3ketones of the androstane series has not previously been described. We observed that 17a-methylt,estosterone, or testosterone, in 2.5y0 ethanolic potassium hydroxide were easily hydrogenated, with palladium black as catalyst, to their 5p(1) F. Johnson, G. T. Newbold, and F. S. Spring, J . Chem. Soc., 1302 (1954). (2) A. F. Daglish, J. Green, and V. D. Poole, J . Chem. Soc., 2627 (1954). (3) M. Velasco, J. Rivera, G. Rosenkranz, F. Sondheimer, and C. Djerassi, J. Org. Chem., 18, 92 (1953). (4) R. Yashin, G. Rosenkranz, and C. Djerassi, J . Am. Chem. SOC.,73, 4654 (1951). (5) C. Djerassi, R. Yashin, and G. Rosenkranz, J . A m . Chem. SOC.,74,422 (1952).

dihydro derivativ6s. The product from the hydrogenation of 17a-methyltestosterone, 17P-hydroxy17a-methyl-5P-androstan-3-one, seems not to have been described previously. Infrared spectrophotometry was employed for the establishment of the absence of contaminating 5a-dihydro isomers. Both 17P-hydroxy-5a-androstan-3-one and its l7a-methyl derivative exhibit a characteristic peak a t 11.38 p , which is entirely absent for the 5p isomer. The presence of as little as 5% 5cr contamination could be detected in 17phydroxy- 17a-me thyl-5p-andros tan-3-one by this means, but in the case of 17p-hydroxy-5p-androstan-3-one 25%, but not 10% 5a contamination could be detected, perhaps because the 11.38 p band is comparatively more intense for the 17amethyl-5a-compound. Optical rotatory dispersion was found to be a better tool for analyzing quantitatively the presence of 5a impurities in the 5p compounds. It was found that the crude 170hydroxy-50-androstan-3-onefrom the reduction of testosterone had approximately 15% 5 a contamination.6 Consequently, partition chromatography on alumina was found necessary for the purification of the testosterone reduction product, whereby the less polar 5p product was easily separated from its 5 a isomer. When either 17~-hydroxy-17a-methyl-5~-androstan-3-one or 170-hydroxy-5p-androstan-3-one were assayed for androgenic activity, each possessed 1% of the activity of the testosterone reference standard. Both compounds were given subcutaneously, dissolved in oil. The assessment was based on stimulation of ventral prostate growth of castrated immature male rats. EXPERIMENTAL

17p-Hydroxy-lYa-methyl-6$-androstan3-one. Fifty grams of 17p-hydroxy-l7cu-methylandrost-Pene-3-onewas partially dissolved in 500 ml. of absolute ethanol containing 500 mg. of palladium black and 12.5 g. of potassium hydroxide previously dissolved in 25 ml. of distilled water, and hydrogenated a t an initial pressure of 45 lb. for 2 hr. Afterwards, the palladium catalyst ym removed by filtration through kaolin under reduced pressure, the filtrate neutralized with a sufficient amount of glacial acetic acid, diluted with 1500 ml. of distilled water, and placed in the cold (4') until the precipitated oil was completely crystallized, usually within 24 hr. The crude product was collected by filtration under reduced pressure, dried in vacuo over potaasium hydroxide, and added to 2000 ml. of boiling petroleum ether (b.p. 60-110°), wL,h facilitated separation of insoluble impurities from the soluble 58-dihydro reduction product. Decantation of the supernatant into another Erlenmeyer flask, boiling down to 500 ml., and setting the flask aside to cool in the cold (4') afforded 30 g. (60%) of pure 17j3-hydroxy-17a-methyl-5j3-androstan3-one,colorless glistening plates, [a]Y +3" (chloroform), double m.p. 74-76' and 119-121' (Kofler stage). Careful crystallization from dilute ethanol yielded the higher melting point form as

(6) We are greatly indebted to Dr. Fred Kincl of Syntex Laboratories for optical rotatory dispersion measurementa.

6%

NOTES

colorless needles, m.p. 120-121'. The compound gave a correct analysis for C20&02.7 17p-Hydroxy-6g-androstan4-one. Twenty-nine grams of testosterone was dissolved in 290 ml. of absolute alcohol containing 290 mg. of palladium black and 7.25 g. of potassium hydroxide previously dissolved in 15 ml. of distilled water, and hydrogenated a t an initial pressure of 45 lb. for 2 hr. The work-up of the hydrogenation mixture was the same as described above, except that purification of the dried crude product was accomplished by dissolving in 500 ml. of a solution of benzene and ether 2:1, passing the solution through a chromatographic column containing 150 g. of chromatographic grade alumina (Harshaw), and eluting with 3 1. of a solution of benzene-ether 2: 1. Evaporation of the eluate to dryness and crystallization from 500 ml. of hot petroleum ether (b.p. 60-llOo) afforded 17.3 g. (60%) of pure 17g-hydroxy-5g-androstan-3-one,colorless plates, m.p. 142-144' (Kofler stage), [a]: $32" (ethanol). Its infrared spectra was identical with that for an authentic sample of 17p-hydroxy-5p-androstan-3-one,and the mixed m.p. showed no depression.

Acknowledgment. We would like to thank Syntex, S. A., The Upjohn Co., and the Cancer Chemotherapy National Service Center of the National Institutes of Health, Public Health Service, for generous supplies of steroids for this work. This investigation was supported by a grant (CY3603) from the National Cancer Institute, Kational Institutes of Health, Public Health Service. DIVISIONOF ENDOCRIXOLOOY ALTONOCHSNER MEDICALFOUNDATION NEW ORLEANS, LA. (7) We are greatly indebted to the Cancer Preparations Laboratory of Merck Sharp and Dohme for determination of optical rotation and carbon-hydrogen analyses of the compound.

Reaction of Diosgenin Acetate with Hydrogen Chloride in Acetic Anhydride FREDERICK C. UHLE Received August I, 1961

VOL.

27

Treatment of I1 with three equivalents of potassium phthalimide in dimethylformamide a t 105' gave the phthalimido derivative 111. Hydrolysis of I11 with 5% ethanolic potassium hydroxide, followed by phthalamidic acid ring closure with N-cyclohexyl-N'- [2-(4- morpholinyl)ethyl]carbodiimide metho-p-toluenesulfonate, afforded the hemiketal IV, an intermediate in the synthesis of solasodine (VII) from kryptogenin.* Hemiketal dehydration of IV with glacial acetic acid a t 100' gave the phthalimido furostene V, an intermediate in the synthesis of solasodine (VII) from pseudodiosgenin.3 Hydrazinolysis of I11 afforded VI, a tetrahydropyridine derivative first prepared from solasodine (VII) with acetic anhydride in the presence of zinc c h l ~ r i d eSubjection .~ of the gross product from hydrazine treatment of 111 to acetate hydrolysis with aqueous ethanolic potassium hydroxide gave solasodine (VII), constituting a three-step synthesis of the alkaloid from the sapogenin. EXPERIMENTAL'

3~,16~-Diacetoxy-87-chloro-25a-6-choZesten-I~-me (11). The compound (CslH&lOs) (535.14) was prepared from diosgenin acetate with anhydrous hydrogen chloride and refluxing acetic anhydride according to the procedure of Miner and Wallis.] Sg, 16~-Diacetoxy-27-phthalimido-I6a-5-chobesten-~.8-one (111). A magnetically stirred solution of 267 mg. (0.0005 mole) of 3~,16~-diacetoxy-27-chloro-25a-5-cholesten-22-one (11)and 277 mg. (0.0015 mole) of potassium phthalimide in 2 ml. of anhydrous dimethylformamide was heated a t 105' for 24 hr.8 After the mixture had been diluted with 20 ml. of saturated aqueous potassium chloride, the precipitate was collected by filtrat,ion, washed with water and dried. Initial attempts to recrystallize the product from methanol were erratic, giving low melting material which did not improve and even appeared to deteriorate on repetition, suggesting an instability to prolonged heating in alcoholic solution. Recrystallization was accomplished by brief warming in isopropyl alcohol, followed by rapid chilling. Two such recrystallizations gave 200 mg. (62%) of very small plates, m.p. 160-180°. The analytical sample, from isopropyl alcohol, melted at 179-182'; [a]= 14" (chloroform); infrared spectrum (KBr): 5.80 (acetoxy), 5.65, 5.90, 13.8, 14.0 fi (phthalimido). Anal. Calcd. for CWH~INO, (645.81): C, 72.43; H, 7.96; N,2.17. Found: C,72.25; H, 7.82; N,2.43. 3p,2.8-Dihydroxy-27-phthalimido-~6a-6-furoatene (CssH,,NOs) (661.78) (IV). A solution of 129 mg. (0.0002 mole)

+

In 1956 Miner and Wallis described an interesting fission of the spiroketal ring juncture of diosgenin acetate (I) with anhydrous hydrogen chloride in refluxing acetic anhvdride to give an acetoxy chloro derivative.' The chloro compound, which (3) F. C.Uhle, J . Am. Chem. SOC.,83, 1460 (1961). crystallizes remarkably well from the resinous (4) Y. Sato, H. G. Latham, Jr., and E. Mosettig, J. reaction product in logo yield, was tentatively Orq. Chem., 22, 1496 (1957); Y. Sato and K. Ikekawa, J. considered by its discoverers to be 3P127-diacetoxy- Orq. Chem., 25, 786 (1960). 16~~-chloro-25a-5-cholesten-22-one. An attempt to (5) Melting points xere observed on a calibrated micro use the substance as starting material for synthesis hot stage. Microanalyses were performed by Dr. S. M. Nagp, of steroid pyrroline derivatives of a type recently Massachusetts Institute of Technology, Cambridge, Maps. were measured by Schwarzkopf Microanalytical prepared12however, has established that it must Rotations Laboratories, Woodside 77, K. Y. Infrared spectra were possess the structure 3P,lGP-diacetoxy-27-chloro- recorded with a Perkin-Elmer spectrophotometer, model 137. 25a-5-cholesten-22-one (11). Only those functional bands of significance in interpretation (1) R. S. Miner, Jr., and E. S. Wallis, J. Org. Chem., 21, 715 (1956). (2) F. C. Uhle and F. Sallmann, J. Am. Chem. Soc., 82, 1190(1960).

are mentioned. (6) The sparingly soluble 27-chloro compound was altogether unreactive with potassium phthalimide in dimethylformamide at 25' under conditions used with 27-iodo derivatives. a

FEBRUARY

1962

NOTES

C H , C O O U

VI

HO

657

JVw

of 3~,16&diacetoxy-27-phthalimido-25cr-5-cholesten-22-one longer exhibited the initial low melting point (70-72'). (111), 112 mg. (0.002 mole) of potassium hydroxide and 1 Fresh recrystallization from isopropyl alcohol, however, drop of water in 2 ml. of absolute ethanol was heated under gave material which melted a t 70-72', followed by solidireflux for 3 hr. After the mixture had been diluted with 10 ml. fication in characteristic fashion and remelting at 130of water, the clear solution was acidified with 6N aqueous 145'. This material, when mixed with the substance prehydrochloric acid. The precipitate was collected by filtra- pared from 111, again melted at 70-72'/130-145". Infrared tion, washed with water, and dried. To a solution of this spectra (IIBr) of the samples from the three sources were phthalamidic acid in 1 ml. of methanol was added 212 mg. identical: 5.92 (medium) (shoulder) (vinyl ether), 5.70, (0.0005 mole) of N-cyclohexyl-N'- [2-(4morpholinyl)-ethyl] 5.90, 13.8, 14.0 p (phthalimido). Sp,16j3-Daacetoxy-d8(87)-imino-25cu-5,22(hr)-cholestadiene carbodiimide metho-p-toluene~ulfonate.~After 3 hr. a t 25', followed by 20 hr. at O", the precipitate was collected by (ColH4&0,) (497.69)(VI). A solution of 129 mg. (0.0002 mole filtration and waa washed with methanol. The filtrate wm of 3~,16~-diacetoxy-27-phthalimido-25~-5-cholesten-22-one diluted with water t o give a precipitate which was collected (111), 64 mg. (0.002 mole) of hydrazine, and 1 ml. of methanol by filtration, washed with water, dried and combined with in 1 ml. of dichloromethane was kept at 25' for 3 hr.; a the material which had deposited directly from the metha- heavy deposit, presumably the hydrazine salt of phthalhynolic reaction mixture. Two recrystallizations from methanol draaide, had begun to separate. After 68 hours at 0", the gave 56 mg. (507,); m.p. 170-173" (varies somewhat with mixture was diluted with water and was extracted with rate of heating); rich infrared spectrum (KBr) identical with ether. The organic phase was washed with dilute aqueous that of the product prepared by sodium borohydride reduc- ammonia and with water and was concentrated under retion of 38-hydroxy-27-phthalimido-25a-5-cholesten-l6,22-duced pressure. The residue was triturated with 10 ml. of 10% aqueous acetic acid. After a very small amount of indione: 5.65, 5.90, 13.8, 14.0 M (phthalimido).' Although phthalamidic acid ring closure of the sodium soluble material had been removed by filtration through borohydride reduction product of 3p-hydroxy-2T-phthalicotton, the filtrate waa made basjc with dilute aqueous ammido-25a-5-cholesten-16,22-dione with iV-ethyl-N'-[2-(4 monia. An ether extract of the precipitate was washed with morpholiny1)ethyl)carbodiimide metho-ptoluenesulfonate water, dried over anhydrous magnesium sulfate, and filtered. at 25' had been accompanied by hemiketal dehydration,s The residue from vacuum evaporation of the filtrate was IV was stable to treatment with as much as 20 equivalents recrystallized from methanol to give 55 mg. (55%) of plates; of N-cyclohexyl-N'- [2-(4morpholinyl) ethyl]carbodiimide m.p. 184-192'; [ a ] D +43" (chloroform) ; infrared spectrum metho-p-toluenesulfonate in methanol for 20 hr. at 25". (KBr): 5.80 (acetoxy), 6.05 p (medium) (C=N).' 38-Hydroxy-27-phthalimido-25cr-5,20(22)-~urostadaene (CabGentle alkaline hydrolysis of VI gave a product still conH46NO4)(543.72)(V). A solution of 20 mg. (0.000035 mole) taining an acetoxyl reridue, presumably at C-16,' as attested of 3~,22-dihydrouy-27-phthalimid0-2,5c~-5-furostene (IV) in by infrared spectrum; more vigorous basic hydrolysis gave 12 drops of glacial acetic acid was heated at 100" for 10 solasodine (VII),* identified by melting point and infrared min.8 After the solution had been diluted with 10 ml. of spectrum. water, the precipitate wm collected by filtration, washed Solasodine [YP-Hydroxy-22(B7)-zmzno-25a-5-furostene] with water, and dried. Recrystallization from isopropyl alco- (C~H43N02) (413.62) (VII) A solution of 129 mg. (0.0002 hol gave 10 mg. (507,); m.p. 70-72", folloTved by solidifica- mole) of 3~,16~-diacetoxy-27-phthalimido-25a-5-cholestention and re-melting at 130-145'. 22-one (111),64 mg. (0.002 mole) of hydrazine and 1 nil. of Samples of V, prepared from kryptogenin3 and from methanol in 1 ml. of dichloromethane was kept a t 25" for 45 pseudodiosgenin,3 which had been stored for over a year no hr. After the mixture had been diluted with water, the organic solvents were distilled under diminished pressure. The (7) Aldrich Chemical Company, Milwaukee, Wis. precipitate was collected by filtration and was washed with (8) For acetic acid dehydration of 22-hemiketals, cf. H. dilute aqueous ammonia and with water. To a solution of the Hirschman and F. B, Hirschman, Tetrahedron, 3,234 (1958) ; precipitate in 4.5 ml. of ethanol was added a solution of 224 Y. Sato and N. Ikekawa, J . Org. Chem., 2 5 , 789 (1960). mg. (0.004 mole) of potassium hydroxide in 0.5 ml. of water.

658

VOL.

NOTES

After 20 hr. a t reflux temperature, the solution was concentrated under reduced pressure. The residue wm diluted with water and was extracted with ether. The organic phase waa washed with water, dried over anhydrous magnesium sulfate, filtered, and concentrated in vacuo. One recrystallization of the residue from methanol gave 25 mg. (30%) of white lustrous plates, m.p. 199-201" (plates in part characteristically rearrange to needles as the melting point is approached; melt solidifies to long needles) ; melting point of a mixture with a sample of naturally occurring solasodine (m.p. 199-201") 199-201"; infrared spectrum (KBr) identical with that given by a specimen of natural solasodine: 10.3, 10.4, 11.2, 11.5 (azaoxaspirane bands).

Acknowledgment. The author is indebted to Freda A. Doy for exploratory experiments; to S. B. Penick and Co., Inc., for a gift of diosgenin; and to the National Heart Institute of the National Institutes of Health, United States Public Health Service (H-2205) and the Eugene Higgins Trust for financial support. DEPARTMENT OF PHARMACOLOGY HARVARD MEDICALSCHOOL BOSTON15, MASS.

Synthetic Polysaccharides. VII. Preparation of Polyglucose Nitrate JOHN W. WOODAND PETERT. MORA Received August 3, 1961

In order to prepare water soluble cationic derivatives of the synthetic polysaccharides for macromolecular interaction studies and for biochemical applications, various reactions were recently investigated2 for the introduction of amine groups into the synthetic polygluco~e.~ One of these reactions was the attempted reduction of polyglucose nitrate with sodamide in liquid ammonia. Unfortunately, the reduction attempt gave only highly degraded unidentified products of low molecular weight. The preparation and the properties of the nitrate ester of the synthetic polyglucose are given, however, for the following reasons: The chemically synthesized polymer of glucose is a highly branched poly~accharide,~ having on the average three free hydroxyls available for esterification per anhydroglucose unit. A polymer with high nitrate content has a large number of nitrate groups in a very small space, held together by the covalently linked, branched, spherical carbohydrate ~ k e l e t o n In : ~ this regard it was interesting to note (1) Cf. for example P. T. Mora and B. G. Young, Arch. Biochem. Biophys., 82, 6 (1959); P. T. Mora, B. G. Young, and M. J. Shear, Makromol. Chemie, 38, 212 (1960); and B. G. Young and P. T. Mora, Virology, 12, 493 (1960). (2) J. W. Wood and P. T. Mora, in preparation. (3) P. T. Mora and J. W. Wood, J . Am. Chem. SOC., 80, 685 (1958), P. T. Mora, J. W. Wood, P. Maury, and B. G. Young, J . Am. Chem. Sac., 80,693 (1958). (4) P. T. Mora, J. Polymer Sci., 23,345 (1957).

27

that a polyglucose dinitrate detonated with one half of the impact force necessary for nitrocellulose of similar degree of substitution; also it ignited a t lower temperature (155'). These observations indicate that the close proximity of the nitrate groups attached to different monosaccharide residues also increases the instability of the polysaccharide nitrates as does the degree of substitution or the ratio of nitrogen to carbon and oxygen. Undoubtedly this method of nitration of polyglucose can be applied to the numerous different synthetic polysaccharides which were prepared by similar polycondensation of various other carbohydrates.6 EXPERIMENTAL

Conc,entrated nitric acid (70% reagent grade; sp. gr. 1.42) 400 ml. and 500 ml. of concentrated sulfuric acid (96% reagent grade; sp. gr. 1.84) were mixed in a round bottomed flask, 2 1. capacity, equipped with a glass stirrer and a thermometer. The mixture was then chilled (ice-salt bath) to -3" to -4" and 20 g. (0.123 anhydroglucose unit) of fineJy divided polyglucose (number average molecular weight M, = 6,600, intrinsic viscosity [ v ] = 0,05,.sample A, Ref. 2) was added as ra7idly as possible with stirring. The temperature of the reaction mixture rose to -2' for several minutes then dropped to -3" where it was maintained for 90 min. with continued stirring. The ice bath was replaced with a warm water bath and the mixture was heated gradually over a 20min. period to 34 i la,maintaining this temperature for 10 min. The mixture was then poured onto 2000 g. of crushed ice plus 300 ml. of water. After the ice had melted the white lumpy product was separated from the aqueous phase by centrifugation. In order to free the ester from acidic material the former was exhaustively washed with water, 5% sodium bicarbonate, and finally with water again until neutral to litmus, separating it from the washings each time by centrifugation. The product, after storing under water in the refrigerator ( +3") overnight, became slightly acidic again. Next, in order to preclude the possibility that acidic material was being entrapped in the slightly granular state of the ester, the latter was carefully ground to a fine powder under water with an all glass mortar. The slurry was transferred to a sheet of filter paper (Whatman #54) in a Buchner funnel and again washed with water until free of acidic material. After being dried for about a week in a vacuum (0.1 mm.) desiccator over calcium chloride (anhyd.) and sodium hydroxide pellets, the polyglucose nitrate was obtained as a fine white powder with a slight but still persistent odor of nitrogen oxides. The nitrate was quite soluble in acetone, abs. ethanol, 95% ethanol-abs. ether mixture (1:1); fairly soluble in 95% ethanol; very slightly soluble in abs. ether; and insoluble in water. On continued periods of storage in the dry state the nitrate was observed to give off increasing amounts of the yellow oxides of 'nitrogen when shaken or stirred, and it finally became lemon-yellow in color. It was then concluded that the persistence of the acidity in the nitrate was due to some auto-catalytic decomposition of the nitrate linkages in the ester structure, possibly initiated by presence of small amount of sulfuric acid esters6 and not to any mechanical entrainment of nitric acid in the granular or powdered material. From then on it was decided that for periods of storage longer than two t o three days the poly(5) P. T. Mora and J. W. Wood, J . Am. Chem. SOC.,8 2 , 3418 (1960). (6) R. W. Kerr in Chemistry and Industry of Starch, Academic Press, Inc., New York, N. Y., 2nd Ed. 1950, p. 303, mentions this possibility in the case of the instability of starch nitrates.

FEBRUARY 1962

659

NOTES

acetylcholine and related choline esters as defined by Dale many years ago.2 Further studies have shown that McN-A-343 is a ganglionic stimulant which acts at receptor sites or ganglionic cells distinct from those ganglionic sites which are activated by acetylcholine and which are blocked by conventional agents. Several related compounds which conform to structure I1 have been prepared due to the current interest in McN-A-343. These are, in general, very hygroscopic compositions which are obtained initially as gums or sirups. The trimethylammonium compounds are usually the highest melting materials in a given series and are more readily obtained in a pure state. The method of purification used for the compounds in Tables I-IV consisted of successive recrystallizations from absolute alcohol-ether mixtures. All the quaternary salts prepared were found to melt with decomposition. LABORATORY OF CHEMICAL PHARMACOLOGY No acetylenic carbamates of this type (11) NATIONAL CANCER INSTITUTE NATIONAL INSTITUTES OF HEALTH have been reported prior t o this work. However, BETHESDA 14, MD. the closely related acetylenic esters (111) have been prepared and tested as ganglionic blocking (7) We are indebted to Dr. W. C. Cagle and his co-workers agents by Bie14s5 and as fungistatic agents by at the G. S. Naval Propellant Laboratory, Indian Head, WatersS6The synthesis of 1-acetamido-2-butynylMd., for these analyses. trimethylammonium chloride also has been described by Marszak-Fleury.'

glucose nitrate could best be stablized by covering it with 10% sodium bicarbonate solution and storing in a refrigerator a t 3" to 5". The yield was 30.5 g. (97.9% of theor. for a di-nitrate substitution). Anal. Calcd. for [C6H806( N O P ) P ]C, ~ : 28.58; H, 3.20; N, 11.11.Found:C,26,72;H, 3.17;N, 11.15(Dumas), 12.14av. (Dupont Nitrometer'). On the basis of the nitrogen content (11.650j0,average) the ester contains 2.2 nitrate groups per anhydroglucose unit. The impact sensitivity test was run' according to a modified U. S.Bureau of Mines procedure. Our sample of polyglucose nitrate was detonated when placed in a combination metal cylinder and piston arrangement and a 5 kg. weight in turn was dropped from a height of 200 mm. onto the piston. A typical grade of nitrocellulose of 12.60% nitrogen usually detonates at a weight distance of 400 mm. under the same condition of testing. In the ignition test? small samples were placed in test tubes which were then heated in a Wood's metal bath at the rate of 5" per minute. Our sample ignited a t 155" whereas the typical nitrocellulose sample ignited at 189". The conclusion was that our sample of polyglucose nitrate is an extremely sensitive and unstable material and should be handled with due precautions.

Unusual Pressor Agents : Acetylenic Carbamates 'I?. R. HOPKINS,JAMES H. REA, P. D. STRICKLER, AND WILLIAMVANDERLINDE

Received August 3, 1961

R

@

'NCOOCH~C-CCH~NR~R~R~X@ / R1 0 I' RCOOCH~C=CCH~NR~R~R~X' I11

I n the course of a routine examination of compounds of diverse chemical nature for cardiovascular and autonomic activity, it was found that 4In addition to pharmacological studies, all [ N - (3 chlorophenyl)carbamoyloxy] - 2 - butynylcompounds prepared were screened as pesticides. trimethylammonium chloride (I) possesses unique It is of some interest that only 4-(3-chlorophenylpharmacological properties. I n the chloraloseanesthetized dog or cat this compound (also known carbamoyloxy)-2-butynyltributylammonium chloas McN-A-343) produces an initial depressor ride possesses post-emergent herbicidal activity response followed by a large pressor response a t 8 toward wild oats (Avenafatua) while being tolerant pg./kg., i.v. The pressor activity was partially toward wheat and barley a t identical rates. This blocked by Dibozane [1,4-(bis-1,4-benzodioxan-2- activity is of the same type, but of a lesser degree, yl methy1)piperazinel (1 mg./kg., i.v.) while hex- as that of the parent compound, 4-chloro-2amethonium [hexamethylenebis(trimethylammo- butynyl N-(3-chlorophenyl) carbamate (barban), nium bromide)] (1 mg./kg., i.v.) did not inhibit and now sold commercially as a selective wild oat actually potentiated this action. Unexpectedly, herbicide. atropine (1 mg./kg., i.v.) blocked both the pressor (2) H. H. Dale, J . Pharmacol. Exptl. Therup., 6, 714 and depressor activity. It was concluded by (1914). Roszkowskil that this material (I) did not fall (3)'A. P. Roszkowski, J. Pharmacol. Exptl. Therup., 132, into the classical pressor-depressor categories of 156 (1961).

-

'

(1) A paper entitled "McN-A-343: An Unusual Pressor Agent" was presented by A. P. Roszkowski a t the Fall Meeting of the American Society for Pharmacology and Experimental Therapeutics a t Seattle, Wash., 1960.

(4) J. H. Biel, U. S. Patent 2,867,619 (1959). (5) J. H. Biel, E. P. Sprengler, and H. L. Friedman, J . Am. Chem. Soc., 79, 6184 (1957). (6) J. A. Waters and G. A. Wiese, J . Am. Pharm. Assoc., 49,112(1960). (7) A. Marszak-Fleury, Compt. rend., 241, 752 (1955).

660

NOTES

VOL.

27

TABLE I 4N-CARBAMOYLOXY-2-BUTYNYLTRIMETHYLAMMONIUM CHLORIDESa

0

@

ArNHCOOCHzC=CCHzN( CH&C1 Crude Yield, AT

Formula

H 2-C1C8H4 3-ClCsH4 4-ClCsHr 2,5-C12CaH3 3-CFGH4 l-CloHT

M.P.

CaHibClNzOZ 226-228 dec. C14H18C12N20~ 117-1 19 dec. C14HlsC1~N~02 182-183 dec. C14H18C1zNz02 205-210 dec. ClaH17ClIN202 171-174 dec. C16HlsCIF3N20~ 195-197 dec. Cl~H21ClNZOZ 212-213 dec.

Calcd.

%

C

72 57 95 64 85 80 81

46.5 53.0 53.0 53.0 47 8 51.4 65.0

H

Found

N

7 . 3 13.6 5 . 7 8.8 5.7 8.8 5.7 8 8 4 . 9 8.0 5.2 8.0 6.4 8 4

Ionic C1

C

17.2 ll.2 11.2 11.2 10.1 10.1 10.7

46.5 53.1 52.8 52.8 47.7 51.2 64.7

H

N

7.3 6.0 5.9 5.5 4.7 5.4 6.3

C

13.3 8.5 8.7 8.6 7.7 7.9 8.0

Ionic 1 17.1 11.3 11.6 10.9 10.1 10.4 10.7

All compounds were prepared by Method A. TABLE I1 4N-CARBAMOYLOXY-2-BUTYNYLTRIETHYLAMMONIUM @

ArNHCOOCH&&CH2N( Crude Yield, Ar

Formula

M.P.

H CllHzlClNz02 196-197 dec. C1&C1NZo2 186-187 dec. CsHs 3-CICsH4 C1?H2&12NZ02 133-135 dec. 4-ClCeH4 CnHzrClzN~Oz 82-85 dec. 147-148 O Z dec. S - C F I C ~ H ~C I S H Z ~ C ~ F ~ N ~ a

C2H&C1 Calcd.

%

C

59 72 47 64 20

53.1 62.9 56.8 56.8 55.0

CHLORIDESO

e

H

N

8 . 5 11.3 7.8 8.6 6.7 7 . 8 6 . 7 7.8 6.2 7.1

-

Found

Ionic C1

C

14.3 10.9 9.9 9.9 9.0

52.9 62.6 56.8 56.6 54.9

H 8.5 7.5 6.8 6.7 6.0

N 10.9 8.3 7.6 7.6 6.7

C

Ionic I 14.2 10.8 9.8 9.7 9.4

All compounds were prepared by Method A.

The compounds in Tables I, 11, and I11 were prepared by two general methods:

and an ethyl homolog (VI) was synthesized from the reaction of 4-diethylamino-2-butynyl acetatesPl2 with ethyl bromide. The ethynylog of (A) RRiNCOOCHzCGCCHzX + RzRaR4N + choline iodide, 4-hydroxy-2-butynyltrimethylam@ RRlNCOOCHZC=CCHzNRzR3R~Xe monium iodide, has been described in a prior publication. l a (B) RRlNCOOCHzCSCHzCl $- 2RzR3NH +

+ RzR3XH.HX + R,X +

RRlNCOOCHzC=CCH*NRzRa RRiNCOOCHzCECCHzNRzRs

~ N H C O O C H z ( C H z ) 2 C H ,0R ( C H 3 ) J C a

@

RRINCOOCH~C=CCHZNRZR~R,X~

The methods employed for the preparation and the physical properties of the 4-halo-2-butynyl N-substituted carbamates used as starting materials have been described in a previous paper.8 In order to determine structure-activity relationships, the acetylenic bond was replaced by an olefinic bond in four examples (Table IV) and bv a single bond (IV). Acetylenic derivatives of choline chloride were prepared by the quaternization of 4chloro-2-butyn-1-olg with trimethylamine or triethylamine (Table V). In addition, the ethynylog (V) of acetylcholine chloride ivas prepared by the quaternization of 4chloro-2-butynyl acetat elO.ll with trimethylamine (8) T. R. Hopkins, R. P. Neighbors, P. D. Strickler, and L. V. Phillips, J. Org. Chem., 24, 2040 (1959). (9) W. J. Bailey and E. J. Fujiwara, J. Am. Chem. Sac., 77,165 (1955). (10) M. M. Fraser and R. A. Raphael, J. Chem. Sac., 4280 (1955). (11) J. Colonge and G. Poilane, Bull. soc. chim. France, 502 (1955).

0

CH&OOCH&GCCHZN(CH3)3Cl@ V 0 CH3COOCH2C-CCH2N(CzH5) &I0 VI

The pharmacology of the compounds in Tables I-V will be reported elsewhere. EXPERIMENTAL'^ Preparation of l+-chloro-B-butynyl N-(5-trifEuoromethylpheny1)carbamate. A mixture of 3-trifluoromethylaniline (8.1 g., 0.05 mole), pyridine (4.1 g., 0.05 mole) and 100 ml. of benzene was stirred and cooled to 5'. 4Chloro-2-butynyl chloroformate (8.4 g., 0.05 mole) was added dropwise a t such a rate so as t o maintain the temperature below 15'. After stirring a t ambient temperature for an additional 2 hr. the

(12) I. Marsznk and A. Marszak-Fleury, Compt. rend., 226,1289 (1948). (13) M. Olomucki. Ann. chim. (Paris), , 5 .. 845 (1960). ( i 4 j AH melting points are uncorrected.

FEBRUARY

1962

NOTES

66 1

reaction mixture was poured into 200 ml. of water, stirred well, the benzene lay& removed and dried over calcium chloride. Removal of the benzene under reduced pressure gave the product, 12.3 g. (85%), as a viscous oil. The product was purified by successive low temperature extractions with benzene-hexane mixtures. Anal. Calcd. for C12HSC1FJVOZ: C, 49.3; H, 3.3; C1, 12.2. Found: C, 49.8; H, 3.4; C1, 12.2. Preparation of 4-chloro-2-butynyl N-(9,5-dichlorophenyl)carbamate. A mixture of 4-chloro-2-butyn-l-o1 (10.4 g., 0.1 mole), 2,5-&chlorophenyl isocyanate (18.8 g., 0.1 mole), 100 nzl. of benzene and 5 drops of pyridine "as refluxed for 3 hr., cooled to room temperature, diluted with hexane and cooled to 0". The precipitated product, 9.7 g. (33%), was recrystallized from acetone-hexane, m.p. 70-71 '. Anal. Calcd. for CllH8ClaN02: C, 45.2; HI 2.8. Found: C, 45.1; HI 2.8. Preparation of 4-chlorobutyl N-(5-chlorophenyl)carbamate. A solution of 3-chlorophenyl isocyanate (15.4 g., 0.10 mole) and 4chlorobutyl alcohol (10.8 g., 0.10 mole) in 100 ml. of dry benzene -was heated for 3 hr., cooled and diluted with two volumes of petroleum ether. The resultant pink oil was separated and the residual solvent was removed under reduced pressure. There was obtained 16.0 g. (61%) of nearly colorless oil, n': I .5522. Anal. Calcd. for C11HlaC12N02:C, 50.4; HI 5.0; N, 5.3. Found: C, 50.3; H, 4.7; N, 5.4. Preparation of 4- [N-(S-ch1orophenyl)carba~noyloxy ]butyltrimethylammonium chloride. A solution of 4chloi-o-2-butyl N-(3-~hlorophenyl)carbamate (13.1 g., 0.05 mole) in 200 ml. of dry benzene was saturated with trimethylamine. After 18 hr. the resultant solid was recrystallized from an ethanolether mixture; yield 3.5 g. (220/,); m.p. 182-184'. Anal. Calcd. for C,aHzZC12N,02: C, 52.3; H, 6.9; N, 8.7. Found: C, 52.1; HI 7.1; N, 8.6. Preparation of .&chlOrO-&bUtenyl .V-(S-chloropheny1)carbamate. A solution of 3-chlorophenyl isocyanate (15.3 g., 0.10 mole) and 4chloro-2-buten-l-o1 (10.6 g., 0.10 mole) in 100 ml. of dry benzene was heated for 3 hr., cooled, and diluted with 600 ml. of hexane. The resultant heavy liquid was separated and the residual solvent was removed under reduced pressure. The product, 15.6 g. (60%), was a viscous, brown-colored oil, n y 1.5660. Anal. Calcd. for Cl,H11Cl2NOZ:C, 50.8; H, 4.2. Found: C, 51.0; H, 4.5. .$-Chloro-%butenyl N-phenylcarbamate was prepared in a similar manner; yield 75%; m.p. 40-42' after recrystallization from hexane. Anal. Calcd. for CllH12C1N02:C, 58.5; H, 5.3. Found: C, 58.25; H, 5.35. The following examples are representative of the methods of preparation of compounds in Tables I, 11, and 111. Method A. Preparation of 4-[N-(S-chloropheny1)carbamoyloxy 1-2-butynyltrimethylammnium chloride. A solution of 4chloro-2-butynyl N-( 3-chloropheny1)carbamate (18 g., 0.07 mole) in 400 ml. of dry benzene was saturated with anhydrous trimethylamine and was protected from moisture. After 24 hours the precipitated product (21.0 g., 95%) was recrystallized from ethanol-ether; yield 7.6 g. (3493); m.p. 182-183' dec. Method B. Preparation of 4-dimethylamino-2-butynylN-( 3chlorophay1)carbamate. A solution of dimethylamine (9.0 g., 0.20 mole) and 200 ml. of ether was cooled to 5'. A solution of 4-chloro-2-butynyl N-(3-~hlorophenyl)carbamate (12.9 g., 0.05 mole) in 150 ml. of ether was added dropwise while maintaining this temperature. After addition the mixture was refluxed for 4 hr. The precipitated amine hydrochloride was removed by filtration. The ether mother liquor was extracted with water, dried over magnesium sulfate, and concentrated by evaporation. The residue was treated with hexane to give 5.7 g. (43%) of crude product which was recrystallized from a benzene-hexane mixture; m.p. 108-109". Using a similar procedure, employing benzene as the solvent, a 66% yield of crude product was obtained.

662

VOL. 27

NOTES

TABLE IV 4N-ARYLCARBAMOYLOXY-2-BUTENYLTRIALKYLAMMONIUM

CHLORIDESa

e 0 A~NHCOOCHZCH=CHCHZNR~C~ Crude Yield,

Ar

R

C6H6 CaHs 3-C1C6H4 3-CICeHa

CH3 C2H5 CHI C2H5

Formula

M.P.

C14HnlC1N~02 186-188 dec. C17H2?C1N202 135-136 dec. dec. C I ~ H & ~ ~ N Z 181-183 OZ C17H26C12N202 dec.*

Calcd.

%

C

35 47 96 47

59.0 62.5 52.7 56.5

H

N

7.4 83 6.3 7.3

Found Ionic C 1

9.8 8.6 8.8 7.8

12.5 10.8 11.1

9.8

C 59.2 62.1 52.5 56.9

H

N

7.6 8.2 6.0 7.0

C 9.6 8.4 8.4 7.7

Ionic 1 12.7 10.8 11.2 9.8

A11 compounds prepared by Method A . * Indefinite, softened a t 110', decomposed a t 120". TABLE V COMPOUNDS RELATED TO CHOLINE CHLORIDE AND ACETYLCHOLINE HALIDES

e e

X RaNCHdkCCH20Z Crude Yield,

R

Z

CH? H C2H5 H CHa CH3C0 C ~ H S CHICO

X C1 C1 C1 Br

Formula

M.P.

C?Hl4C1NO 95-97dec. C1oH2oC1N2O2 86-93 dec. C&&lh'Oz 163-65dec. ClzHzZBrNO2 94-96dec.

Found

Calcd.

%

C

H

N

Ionic X

61 68 72 52

51.4 58.4 52.3 49.1

8.6 9.8 7.8 7.6

8.6 6.8 6.8 4.8

21.7 17.2 17.2 27.4

C

H

N

Ionic X

51.3 58.1 52.4 49.3

8.7 9.5 8.0 7.6

8.2 6.9 6.8 4.9

21.9 17.2 17.3 27.5

Anal. Calcd. for C I ~ H ~ S C ~ NC, ~ O58.5; Z : H, 5.6. Found: mole) in 100 ml. of dry benzene was saturated with gaseous C, 58.7; H, 5.5. trimethylamine at ambient temperature. The reaction mix4-Diethylamino-%butynyl N-( 3-chloropheny1)carbamate was ture was agitated for 2 hr., the solid removed by filtration prepared in a manner similar to that above, in benzene. and dried in a vacuum desiccator over phosphorus pentoxide. The product was a gummy solid which crystallized on ex- There was obtained 5.0 g. (72%) of off-white crystals; m.p. posure to air. There was obtained a 95% yield of a crude prod- 163-165'. Preparation of 4acetoxy-2-butynyltriethghmmonium brouct which was recrystallized from hexane; m.p. 77-78'. Anal. Calcd. for C16HlsC1N202: C, 50.9; H, 3.9. Found: C, mide. From a mixture of 4diethylamino-2-butynyl acetate (22.0 g., 0.12 mole), ethyl bromide (32.7 g., 0.03 mole) and 50.9; H, 3.8. Preparation of 4- [N-(3-chlorophenyl)carbamoyloxy]-3-buty- 250 ml. of benzene there deposited, after 2 weeks, 4.6 g. nyldimethylethylammonium iodzde. A solution of 4dimethyl- (15%) of product; m.p. 94-96' dec. After 2 months another amino-2-butynyl N-(3-~hlorophenyl)carbamate(8 g., 0.03 11.7 g. of product was collected bringing the total to 16.3 g. mole) and ethyl iodide (15.6 g., 0.10 mole) in 300 ml. of (52%). The combined solids were washed several times with benzene was allowed to stand at ambient temperature. After ether and dried in vacuum over phosphorus pentoxide, m.p. 24 hr., 9.2 g. (75%) of crude product was collected by filtra- 94-96" dec. tion and dried under vacuum. A small portion was recrysSPENCER CHEMICAL COMPANY tallized from an ethanol-ether mixture, m.p. 120-121' dec. KAN. Preparation of 4-[N-(3-chlorophenyl)carbamoyloxy]-d- MERRIAM, butenyltriethytammonium chloride. A mixture of 4chloro-2butenyl N-(3-~hlorophenyl)carbamate(20.8 g., 0.08 mole), triethylamine (15.2 g., 0.15 mole) and 150 ml. of dry benzene was stirred at ambient temperature for 24 hr. The benzene Electronic Effects in the Diels-Alder Reaction was decanted from the reaction mixture and the residue was between Methyl Phenylpropiolate and dissolved in 25 ml. of water. The water solution was added Tetraphenylcyclopentadienonesl Substituted to 150 ml. of benzene and the mixture distilled until the water was removed. The remaining benzene was decanted. The residue, 13.6 g. (47%), solidified after being subjected to MICHAELG . ROMANELLI~ A N D ERNESTI. BECKER3 vacuum for 24 hr. Recrystallization from ethanol-ether gave 11.5 g. (4070)of product, m.p. 110-120'. Received August 3, 1961 The following procedures were used for the preparation of the compounds in Table V. Preparation o f 4-h ydroxy-2-butynyltrimethylammonium chloPrevious studies in this laboratory have been conrzde. A solution of 4chloro-2-butyn-1-01 (10.5 g., 0.10 cerned with the Diels-Alder reaction between tetramole) in 200 ml. of dry benzene was saturated with trimethyl(Y = CH3, amine. The reaction mixture was stirred for 18 hr. and the cyclone and (1) Y-C=C-CeH6 bulk of the solvent was removed from the product by decantation. After removal of residual solvent under reduced pres(1) Taken from the thesis submitted to the Faculty of the sure there was obtained approximately 10 g. (61y0crude) of an amber colored gum. After cevcral recrystallizations from Polytechnic InAtitute of Brooklyn in partial fulfillment of an ethanol-ether mixture there was obtained 1.6 g. (10%) of the requirements for the degree of Bachelor of Science, 1961. (2) National Science Undergraduate Research Particia white amorphous solid, m.p. 95-97' (sealed tube). Preparatiomof ~-acetoxy-3-buti~nyltr~methyla"on~um chlo- pant, Summer 1960. (3) To whom inquiries should be scnt. rzde. A solution of 4-chloro-2-hutj nyl acetate (5.0 g., 0.034

FEBRUARY

1962

663

NOTES

"%

TABLE I METHYL2,3,4,5-TETRAKIS(ARYL)&PHENYLBENZOATES

R1

H3

R2

/

\

\

Rl

Rz

Carbon Calcd. Found

M.P.

~~~~~~~~~~~

~~

~~

~~

~

Hydrogen Calcd. Found

Chlorine Calcd. Found

~

c1

H

329-330

77.95

H CI CHIO H

c1 c1 H CHBO

252-253 288-289 267-268 256-257

77 I95 69.74 83.31 83.31

CHI0

CH30

222-224

79.22

77.39 77.89 77.46 69.96 82.98 83.48 83.45 79.30

CH20H, CHO, COOCHI, COOH, H)4 and (2) methyl substituted phenylpropiolate~.~~~ Their conclusion was that electron-withdrawing suhstituents accelerate the reaction. This investigation was concerned with electronic effects of substituents in the diene using methyl phenylpropiolate in all cases. EXPERIMENTAL

The starting tetracyclones were synthesized according to literature procedures as indicated. Tetracyclone, m.p. 220-221' (reported,' 219-220'); 2,5bis(4chlorophenyl)-3,4-diphenylcyclopentadienone, m.p. 236.5-237.5' (reported,g 239-240'); 3,4bis(Pchlorophenyl)2,5-diphenylcyclopentadienone, m.p. 249.5-250.5' (reported,8 253-254"); tetrakis(4chlorophenyl)cyclopentadienone, m.p. 286.0-287.0° (reported,S 297-299'); 2,5-bis(4methoxyphenyl) - 3,4 - diphenylcyclopentadienone, m.p. 195.2-196.0" (reported,g 195.0-195.4'); 3,4bis(Pmethoxyphenyl)-2,5-diphenylcyclopentadienone, m.p. 236.5-237.5' (reported,g 226.8-227.2'); tetrakis(4-methoxypheny1)cyclopentadienone, m.p. 247.2-248.2' (reported,$ 255-256'). Each of the cyclones synthesized here gave no depression with samples of the previously reported compounds. Phenylcyclohexane, purified according to Corson and Ipatieff,lo had a boiling point of 108-109' (12.5 mm.), n':

4.48

4.38 4.53 4.67 3.88 5.62 5.94 5.82 5.75

4.48 3.70 5.59 5.59 5.70

12.11

11.88

12.11 21.67

12.38 21.53

1.5265 (reported: n y 1.5190,11 ny 1.5254,12 ny 1.52215). Methyl phenylpropiolate boiled a t 77' (1 mm.), ny 1.5634, n': 1.5608 (reported,I3ng 1.5618). On standing, the ester developed a green color which interfered with the kinetic measurements; it was therefore distilled before use. The reaction was followed by measuring the rate of evolution of carbon monoxide. The kinetic runs and calculations were carried out as before.5 At the end of each experiment, the products were filtered from the reaction mixtures after cooling. They were recrystallized three times from either xylene or acetic acid (see Table I). RESULTS AND DISCUSSION

The over-all equation for the reaction is

-

+$&

COOCHB cI 0

Ar Ar

k)$(c00CH3+

co

+ Ill CI CbH,

Ar

CsHs Ar

The reaction displayed second-order kinetics even when carried to 90% completion. The reaction rate constants are given in Table 11. The reaction rate constant for tetracyclone itself agrees within 5% with previous values (1.48 X 1O-I molal-' set.-' and 1.56 X 10-3 molal-1 (4) J. J. Dudkowski and E. I. Beckcr, J . Org. Chem., 17, sec.-l 6 , after correcting for temperature using 201 (1952). E, of 19 kcal. (5) I. Benghiat and E. I. Becker, J . Org. Chem., 23, 885 The rate constants show that chlorine substitu(1958). tion in tetracyclone increases the reaction rate (6) D. N. Matthews, doctoral dissertation, Polytechnic constants while chlorine substitution in the ester Institute of Brooklyn, 1961. (7) J. R. Johnson and 0. Grummitt, Org. Syntheses, Coll. also increases the reaction rate constant: for methyl Vol. 111,806 (1955). (8) F. J. Thaller, D. E. Trucker, and E. I. Becker, J . Am. Chem. SOC., 73, 228 (1951). (9) S. B. Coan, D. E. Trucker, and E. I. Beckcr, J . i17n. Chem. SOC.,77,60 (1955). (10) B. B. Corson and V. K. Ipatieff, Org. Syntheses, Coll. Vol. 11, 151 (1943).

(11) J. F. hIcIienna 2nd F. J. Sowvn, J . A7n. Chem. SOC., 59,471 (1937). (12) B. B. Corson and V. N. Ipatieff, J . Ani. Chenz. Soc., 59,645(1937). (13) C. Moureu, P. T. Muller, and J. Varin, Ann. chim. [9] 2,269(1914).

664

NOTES

TABLE I1 REACTION RATECONSTANTS

$zR2 R

0

\ /

R* R2

Temp.

k X 103 (molal-' sec-1)

H H H H

176.4 176.4 176.3 176.3 Av.

1.58 1.57 1.58 1.62 1.59 i 0.02

c1 c1 C1 c1

H H H H €1

176.5 176.3 176.3 176.3 176.3 Av.

2.42 2.53 2.65 2.50 2.57 2.53 f 0.07

c1

17 18

H H H H H

c1 C1 c1 c1

176.3 176.3 176.3 176.3 176.3 Av.

2.37 1.93 2.17 2.38 2.07 2.18 & 0.15

28 30 31 32 33

c1 c1 c1 c1 c1

c1 c1

176.3 176.3 176.3 176.3 176.3 Av.

3.10 2.98 3.33 3.13 3.30 3.17&0.12

34 35 36 37

CHsO CH3O CH3O CHiO

H H H

176.3 176.3 176.3 176.3 Av.

2.13 2.30 2.13 2.15 2.18 i0.07

22 24

26 27

H H H H

OCHa OCHa OCHi OCHa

176.3 176.3 176.3 176.3 Av.

1.23 1.19 1.12 1.19 1.18&0.03

38 39

CH3O CHaO

CHI0 CHaO

176.3 176.3

40

CHjO

CHI0

Run No, 7 S 20 21

11 13 15

16 19 10 12 14

R1

H H H

H C1

c1 c1 c1

H

1.78 1.93 176.3 1.70 Av. 1.80 Z!Z 0.08

phenylpropiolate, k = 1.48 X molal-' set.-' and for methyl 4-~hlorophenylpropiolate,k = 2.25 X molal-' set.-'. Methoxy substitution in tetracyclone accelerates the reaction when in the para positions of the 2- and &phenyls, and in all four para positions, but decelerates the reaction when in the para positions of the 3- and 4-phenrls, whereas the methoxyl group slows the reaction when in the 4- position of methyl phenylpropiolate ( k = 1.10 X molal-' sec.-l). Thus, for the

VOL.

27

first time it is noted that a halogen may accelerate the Diels-Alder reaction whether in the dienophile or in the diene. The rate-substituent effects observed here are different from those in other Diels-Alder reactions in which no carbon monoxide is lost. For example, it has been shown that the p-chloro group has the expected effect of reducing the rate of reaction of 1-phenylbutadiene with maleic anhj-dride.14 On the other hand, in the reaction between tetracyclone and methyl phenylpropargylate, the absence of a wall effecW has been interpreted to mean that the evolution of carbon monoxide is not the rate-determining step in accordance with the concept of n'e~vman.'~ However, the absence of the wall effect has not been demonstrated for substituted tetracyclones and substituted methyl phenylpropargylates. It is also noteworthy that substitution in tetracvclone affects the reaction to a different extent depending upon the position of substitution in the diene; that is, substitution in the 2,5-phenyls is more activating than substitution in the 3,4phenyls. These results and those in the preceding paragraph are not easily understood in the present state of knowledge of the reaction. Further work is under way. DEPARTMENT OF CHEMISTRY POLYTECHNIC INSTITUTE OF BROOKLYN BROOKLYN 1, N. Y. (14) E. J. DeWitt, C. T. Lester, and G. A. Ropp, J . Am. Chem. Soc., 78, 2101 (1956). (15) M. S. Newman and E. Coflisch, Jr., J . Am. Chem. SOC.,80,830 (1958).

Preparation of Trimethylsilyl Ethers of Hindered 2,4,6-Trialkylphenols SIDNEYFRIEDMAN, MARVINL. KAUFMAN, AND IRVINQ WENDER

Received August 4, 1961

As part of the development of a procedure for preparing the trimethylsilyl ethers of coal and of phenols, we have recently reported the successful preparation of 2,Bdi-t-butylphenoxytrimethylsilane, using hexamethyldisilazane and trimethylchlorosilane as reagents and pyridine as solvent.' Pyridine was chosen because of its very good solvent action on coal, and the observation that it was necessary for derivative formation in the case of 2,6-di-t-hutylphenol and for complete reaction of coal. While most pheno!s will react readily with hexamethyldisilaxane, di-t-but,ylphenol requires the presence of pyridine and trimethylchlorosilane as well. At that time,' 2,6-di-t-butylphenol was the (1) S. Friedman, M. L. Kaufman, W. A. Steiner, and I. Wender, Fuel, 40,33 (1961).

FEBRUARY

1962

NOTES

665

chlorosilane (Dow Corning) in petroleum ether'; that used in later work was purchased from Peninsular Chemresearch co. Spectra. Infrared spectra were run on potassium bromide pellets. Ultraviolet spectra were run in cyclohexane. Preparation of ,9,4,6-tri-t-butylphenoxyLr~methylsilane. A solution of 5 g. of 2,4,6-tri-l-butylphenol,25 ml. of hexamethyldisilazame, and 25 d. of trimethylchlorosilane in 30 ml. of dimethylformamide was refluxed under nitrogen for 6 hr. The reaction mixture separated upon cooling into two layers. The upper layer was discarded. The lower layer was extracted with petroleum ether (b.p. 60-68'). The solvent was then evaporated from the extract, leaving 6 g. of crude product. Recrystallization from petroleum ether gave the 2,4,6-tri-t-butylphenoxytrimethylsilane,m.p. 85-87.5". Anal. Calcd. for C21H&Si: c, 75.45; H, 11.38; Si, 8.38. Found: C, 75.56; H, 11.54; Si, 8.27. The infrared spectrum of the product showed the absence of the hydroxyl absorption a t 2.75 p and the presence of the trimethylsiloxy absorption bands a t 7.9, 8.1, 9.75 (weakl, 11.90, 13.20, and 14.60 (weak) p.l Ultraviolet , ,A 2791 A ( e lolo), 2726 1 ( e 966). Low-voltage mass spectral analysis showed a peak at 334 mass units, with traces (< 1%) of higher and lower molecular weight homologs. The trimethylsilyl ether formation proceeded satisfactorily in the absence of trimethylchlorosilane but not in the absence of hexamethyldisilazane or dimethylformamide. A catalytic amount of dimethylformamide (0.1 ml.) in the presence of pyridine, hexamethyldisilazane, and trimethylchlorosilane failed to bring about trimethylsilyl ether formation. When oxygen was not excluded during refluxing, considerable oxidation took place. Preparation of Z?,,4,6-tri-tamy~phenoxyt~imeth ykihne. UPing a procedure similar to that used for the preceding experiment, 2,4,6-tri-t-amylphenoxytrimethylsilane was prepared. This was isolated as a liquid which was recrystallized from ethanol a t -20°, yielding a solid, m.p. 24.5-25.5', with a low-voltage mass spectrum showing a peak a t 376 mass units (as well as small amounts of masses 362 and 390), and with characteristic infrared absorption at 7.9, 8.0, 9.47 (weak), 11.90, 13.20, and 14.5 (weak) p. Hydroxyl 2797 absorption a t 2.75 p was absent. Ultraviolet, , ,A ( e 981), 2740 A ( e 930). Anal. Calcd. for Cn4HraOSi: C, 76.52; H, 11.77; Si, 7.46. Found: C, 76.46; H, 11.75; Si, 8.14. Preparation of 1,6-di-t-butyl-4-methyZphenoxytrimethylsilane. A solution of 5 g. of 2,6-di-t-butyl-4methylphenol (Eastman, recrystallized, m.p. 56') in 25 ml. of hexamethpldisilazane and 35 ml. of dimethylformamide was refluxed in a nitrogen atmosphere for 14 hr. The solvent and excess reagent were removed by distillation. The residue was chromatographed on alumina, and the main fraction recrystallized from petroleum ether (b.p. 60-68"). This gave 4.5 g. of product, m.p. 122.8-124', with the characteristic trimethylsiloxy absorption a t 8.0, 8.15, 9.75 (weak), 11.9, 13.15, and 14.45 (weak) p ; hydroxyl absorption at 2.75 p was completely absent. Ultraviolet, , ,A 2823 4 ( e 1127), 2754 ( e 1054). Low-voltage mass spectral analysis showed the expected peak a t 292 mass units. Anal. Calcd. for ClsHs2OSi: c, 73.90; H, 11.03; si, 9.60. Found: C, 74.07; H, 11.19; Si, 10.0. Hydrolysis of 8,4,6-tri-t-butylphenozytr~methglsilane. A solution of 1 g. of 2,4,6-tri-t-butylphenoxytrimethylsilane in a solution of 6 ml. of concd. hydrochloric acid in 25 ml. of EXPERIMENTAL methanol was refluxed for 4 hr. in a nitrogen atmosphere. Reagents. The 2,4,6-tri-t-butylphenol was generously The reaction mixture was extracted with petroleum ether, provided by the Ethyl Corp. The hexamethyldisilazane and the solvent evaporated. Infrared analysis showed the was prepared by bubbling ammonia through trimethyl- product to consist entirely of 2,4,6-tri-t-butylphenol.

most hindered phenoI readily available, leading to the conclusion that the system hexamethyldisilazane-trimethylchlorosilane-pyridine was capable of forming the trimethylsilyl ether of any phenol. A subsequent study of 2,4,6-tri-t-butyland 2,4,6-tri-t-amylphenol and of 2,6-di-t-butyl-4methylphenol revealed that substituents in the para position of hindered phenols prevented reaction from occurring with this reagent mixture. On the other hand, the trimethylsilyl ether of the unhindered 2,4,6-trimethylphenol is formed by simply refluxing the phenol with hexamethyldisilazane and a trace of trimethylchlorosilane as catalyst, although the catalyst is not necessary. It was thought that the presence of a base stronger than pyridine would be sufficient to cause reaction of the tri-t-alkylphenol. Piperidine was therefore substituted for the pyridine, but reaction still did not take place. A further attempt to form the trimethylsilyl ether of the tri-t-butylphenol was made using a recently published2 procedure which has been successful on carbohydrate derivatives. This method uses trimethylchlorosilane and pyridine as reagents and a two-phase solvent system of hexane and formamide. This procedure also was unsuccessful. When dimethylformamide, which is neither strongly basic nor strongly acidic, was used as a solvent for the reaction between hexamethyldisilaxane and tri-t-butylphenol, it proved to be singularly successful. One possible explanation is that in some reaction intermediate, ionic resonance involving charge distribution at the ortho and para positions of the benzene ring is involved. When the para position is vacant, as in 2,B-di-tbutylphenol, the charge is free to reside a t the para position and there be solvated (for example, by pyridine) relatively easily. Resonance stabilization becomes more difficult when no free ortho or para position is available, especially when the substituents are electronegative alkyl groups, and a solvent with dielectric constant higher than that of pyridine is needed to facilitate solvation of this charge. A similar explanation has recently been proposedJ to explain the difference in behavior of disubstituted benzenes in different solvents. Hydrolysis of the trimethylsilyl ethers required the use of alcoholic hydrochloric acid.l This is in contrast to the behavior of 2,4,6-trimethylphenoxytrimethylsilane , which is readily hydrolyzed by refluxing with aqueous ethanol.

(2) F. A. Henglein and B. Kosters, Chem. Ber., 92, 1638 (1959). (3) R. W. Taft, Jr., R. E. Glick, I. C. Lewis, I. Fox, and S. Ehrenson, J . Am. Chem. SOC.,82, 756 (1960).

a

(4) S. H. Langer, S. Connell, and I. Wender, J. Org. Chem., 23,50 (1958).

666

NOTES

VOL.

27

Attempts to hydrolyze the trimethylsilyl ether in B pyridine-ethanol-water mixture and in dimethylformamide were unsuccessful. U. S. DEPARTMENT OF THE INTERIOR BUREAU OF MINES AVE. 4800 FORBES

PITTSBURGH 13, Pa.

I1

I

AcO

Synthesis of Some Derivatives of Ursolic Acid’ ROBERT A. MICHELI~ I 17

I11 Received August 7, 1961

Mention of several new derivatives of ursolic acid (I. R’ = H, R = COOH) in a forthcoming publiati ion,^ which deals with the far ultraviolet absorption spectra of triterpenoids and steroids, makes a description of the synthesis and properties of these compounds timely. Ursolic acid was isolated from bearberry leaves (Arctostaphylos uva-ursi) by the method of Bilwith some modifications. Acetyl ham et methyl ursolate (I. R’ = CH&O--, R= COOCH,) was smoothly converted in 79y0 yield to the saturated, acid labile ketenesa,b I1 (R = COOCHJ by means of peroxytrifluoroacetic acid in the presence of sodium carbonate.6 Substitution of triethylammonium trifluoroacetate’ for the carbonate buffer afforded a product different from the desired dihydroketone I1 (R = COOCHI). This reaction products was only partially characterized and the reaction has not been investigated further. Conversion of the dihydroketone into the Aleenol diacetate I11 (R = COOCH3) (66% yield) was brought about by the well known sodium acetate-acetic anhydride methodngOn the other hand, the acid catalyzed reaction of the same unstable dihydroketone with isopropenyl acetatelo afforded the isomeric All-enol diacetate IV in 83Yo

V

VIa

HO

VIb

yield, probably without C-13 epimerization (see ref. 5b for a pertinent discussion of the a-amyrin analogues). Proton magnetic resonance spectra of the two enol diacetates, taken a t 60.0 mc. in deuterated chloroform, were provided by Dr. Robert L. Lundin. That for I V showed a broad singlet having a r value of 4.58 which is typical of olefinic protons, whereas I11 showed no such peak. Oxidation of the highly hindered 12,13-d0uble bond of 111 (R = COOCH,) with peroxytrifluoroacetic acid in the presence of disodium hydrogen phosphate6 was not successful. Under similar conditions the A“-enol diacetate IV could be readily oxidized to the a-hydroxyketone V. That V was indeed a keto1 was established from the infrared spectrum and by oxidation with bismutJh oxide in acetic acid to an a-diketone. A study of the ultraviolet and infrared spectra as well as

(1) This work was supported in part by Research Grant CF-4076 from the National Cancer Institute of the National Institutes of Health. (2) Present address: Western Regional Research Labora(8) It has been reported that simple olefins yield hyt o w , U. S. Department of Agriculture, Albany 10, Calif. (3) R.-4.Micheli and T. H. Applewhite, to be published. droxytrifluoroacetates and occasionally ditrifluoroacetates.’ (4) P. Bilham, G. A. R. Kon, and W. C. J. Ross, J . The crude reaction product from acetyl methyl ursolate gave a negative tetranitromethane color test after 1 or 48 Chem. SOC.,35 (1942). (5) ( a ) J. L. Simonsen and W. C. J. Ross, The Terpenes, hr. at room temperature; infrared (Ca), no hydroxyl, 5.56 Cambridge University Press, London, 1957; see Vol. V, p. (CF,CC)O-), 5.71 (broad), and 8.04 p. Hydrolysis was ac128. The .similarity of melting point and optical rotation complished with potassium bicarbonate in methanol [A. values make it difficult to distinguish the C-13 epimeric ke- Lardon and T. Reichstein, Helv. Chim. Acta, 37, 388 (1954)]; tones I1 ( R = COOCHI). There seems little doubt that the infrared (CSZ), 2.86 (hydroxyl), 5.72 (broad), and 8.05 p. product from this reaction corresponds to the known un- In some preliminary studies on As- and A’-sterols, however, fitable ketone since a-amyrin acetate (I. R = CH,, R , = the products appeared ,to be hydroxytrifluoroacetates; CHICO-) also affords the analogous acid labile ketone with typical infrared (C%), 2.9-3.0, 5.6C-5.65, 5.78-5.82, and pcroxytrifluoroacetic acid-sodium carbonate; ( b ) I. A. Kaye, 8.C-8.1 p. The hydroxyls are readily acetylated a t room M. Fieser, and L. F. Fieser, J. Am. Chem. Soc., 77, 5936 temperature with acetic anhydride and pyridine. (9) (a) L. Ruzicka and 0. Jeger, Helv. Chim. Acta, 24, ( 1955), discuss the epimeric a-amyrin compounds. (6) R. D. Emmons and A. S. Pagano, J. Am. Chem. 1178 (1941); (b) L. Ruzicka, 0. Jeger, J. Redel, and E. Volli, Helv. Chim. Acta, 28, 199 (1945). SOC., 77,89 (1955). (10) H. J. Hagemeyer and D. C . Hull, Znd. Eng. Chem., (7) W.D.Emmons, A . S. Pagano, and J. P. Freeman, J. 41,2920 (1949). A m . Chem. Scc., 76,3472 (1954).

FEBRUARY

1962

667

NOTES

positive ferric chloride and tetranitromethane tests proved that the latter exists in one of the monoenol forms VIa or VIb (R = COOCH,). reported the preparation of a Ruzicka et similar derivative in the a-amyrin series (I. R = CH3, R' = H) by chromic acid oxidation of the corresponding dihydroketone I1 (R = CHI) or the A12-enoldiacetate 111 (R = CH,). Although the physical properties such as molecular rotation and ultravioIet spectra are similar, the relationship of the monoenol reported here to the products from chromic acid oxidationgbhas not been firmly established. EXPERIMENTAL^^ Isolation of ursolic acid (I. R = COOH, R' = H). Four kg. of bearberry leaves12 was extracted with a methanol solution (12 1.) of potassium hydroxide (320 g.) a t room temperature for 24 hr. with occasional stirring. The filtrate from the above was acidified (hydrochloric acid), concentrated to 3 l., and the pale green precipitate was collected and dissolved in 5 1. of chloroform-ether (2: 1).Treatment of this solution with an equal volume of 10% aqueous sodium hydroxide deposited the crude sodium ursolate. Ursolic acid (25 g.) was obtained after acidification and three crystallizations from ethanol as colorless needles, m.p. 288-289.5°.13l 4 Acetyl methyl ursolate was prepared in the usual manner with acetic anhydride-pyridine and diazomethane; m.p. 246247.50,15 [a]': 62.5" ( c , 1.12). Methyl ursan-Sp-ol-l2-one-~8-oate acetate (11. R = COOCHa). To 13.5 g. of acetyl methyl ursolate in 200 ml. of dichloromethane and a suspension of 65 g. of anhydrous sodium carbonate was added a solution of peroxytrifluoroacetic acid (from 5 ml. of 90%]hydrogen peroxide and 28 ml. of trifluoroacetic anhydrideI6 in dichloromethane) with stirring over 0.75 hr. The mixture then was refluxed 0.5 hr., filtered, and the product crystallized from methanol as white needles; yield 11.0 g. (79yo). A sample was recrystallized twice from ethanol to constant rotation, [cy17 +28.6" (c, 1.94); infrared, 5.78, 5.91, and 8.07 p . The material melted with sintering 245-251' (open capillary) and 259-262' (evac. capillary).Ka Anal. Calcd. for C33Hj2O5: C, 74.96; H, 9.91. Found: C, 74.57; H, 9.92. Methyl A12-ursene-3p,ld-diol-28-oate diacetate (111. R = COOCH3). Five grams of I1 ( R = COOCH3) and 2.5 g. of freshly fused sodium acetate in 200 ml. of acetic anhydride were refluxed for 48 hr., diluted with ice water and extracted with ether. The residue, after evaporation of the solvent, crvstallized from methanol as straw colored needles; yield 3.5 g. (6670). One recrystallization from aqueous methanol afforded the analytical samnle as colorless needles; m.p. 175179' (with sintering), [el2: +50.6" ( c , l.Ol)j infrared 5.71, 5.78, 5.97 (weak), and 8.07 p, and a yellow color with tetranitromethane. Anal. Calcd. for CdjHjqOb: C, 73.64; H, 9.54. Found: C, 73.66; H, 9.61.

+

( 11) Melting points were taken in evacuated capillaries unless otherwise stated. Optical rotations were run in chloroform and infrared spectra were taken in carbon disulfide. Analyses by S. M. Nagy and associates. (12) This material, previously extracted with 50% alcohol, was generously donated by the S. B. Penick Co. ( 13) Elsevier's Encyclopedia of Organic Chemistry, Elsevier Publishing CO., Amsterdam, 1940-1952, vel. 14 and 148; pp. 565 and 10925. (14) Ref. 5, Vol. V, p. 134. (15) Ref. 13, p. 566. (16) Aged reagent gave poorer results.

Methyl A11-ursene-3p,12-diol-28-oate diacetate (IV). A solution of 6.0 g. of I1 ( R = COOCH3) and 0.9 g. of p-toluene sulfonic acid in 120 ml. of isopropenyl acetate was refluxed and slowly distilled for 24 hr. and t'hen concentrated to 50 ml. Ether and watcr were added arid the ethereal extract was washed, dried, arid evaporated in oucuo on the steam bath. The residue crystallized from methanol as pale tan needles qnd weighed 5.3 g. (83%); m.p. 219-221'. A sample was filtered through a plug of acid-washed alumina (Merck) and crystallized once fr:)m methanol (colorless needles), m.p. 220.5-221.5' (219-220' open capillary), [a]: -48.6' ( c , 1.37), infrared 5.71, 5.82, 6.00 (weak), and 8.08 p, and a yellow color with tetranitromethane. Anal. Calcd. for C3jHa06: C, 73.64; H, 9.54. Found: C, 73.42; H, 9.69. Methyl ursan-Sp,ll~,-diol-i2-one-28-oate Sacetate (V). A solution of peroxytrifluoroacetic acid (2 ml. of 90% hydrogen peroxide, 11 ml. of trifluoroacetic anhydride) in 10 ml. of dichloromethane was added over a 15-min. period to a mixture of anhydrous disodium hydrogen phosphate (25 g.), dichloromethane (50 ml.) and All-enol diacetate IV (0.90 g.). The mixture was refluxed for 0.5 hr. and then filtered. The filtrate,wss washed, dried, and evaporated and the residue crystallized from aqueous methanol. The yield was 0.38 g. (43%) of colorless cry$tals which gave no color with tetranitromethane but did give a positive bismuth oxide test (black precipitate after a few minutes refluxing in acetic acid). The analytical sample, which was obtained as stout crystals from a small amount of methanol, m.p. 237-242" (with sintering), [a]Y +1.3" (c, 5.04) and infrared 2.91, 5.77, 5.91, and 8.OT fi. Anal. Calcd. for C33HstOe: C, 72.75; H, 9.62. Found: C, 72.57; H, 9.65. Monoenol of methyl Sp-acetoxy-il,12-diketo-28-ursanic acid (VIa or VIb. R = COOCH,). Bismuth oxide (500 mg.) was added to a hot solution of V (380 mg.) in glacial acetic acid (15 ml.). After refluxing for 0.25 hr., the mixture vias partitioned between dichloromethane and water. The organic phase was worked up in the usual manner, and the product was crystallized from aqueous methanol; yield, 270 mg. (71%) of pale yellow needles which melted by 210' with sintering. The material was filtered through a plug of acidwashed alumina and recrystallized twice from a small volume of methanol (stout, colorless crystals); m.p. 229-231.5', [a]Y 134.4' ( c , 0.95),?A:' 290 mp ( e 9850) and infrared 2.93, 5.76, 5.97, 6.09, and 8.06 p. The product gave a typical enol test (blue-grey color) with ferric chloride in methanol and a yellow color with tetranitromethane. Anal. Calcd. for C3aHjo06: C, 73.03; H, 9.29. Found: C, 73.21; H, 9.44.

+

Acknowledgment. The author wishes to express his appreciation to Professor Louis F. Fieser for permission to publish this work. DEPARTMEST O F CHEMISTRY H.4RVARD UNIVERSITY CAMBRIDGE 38. MASS.

-

1 Formyl-2-methylisoquinoliniumIodide EDWARD J. POZIOMEK Received Au.gust 7 , 1961

It was obBemed recently that potentiometric titration of 2-formyl-1-methylpyridinium iodide (I) with aqueous alkali gave a smooth titration curve corresponding to that of a weak acid of

669

NOTES

VOL.

27

p K , 9.8-10.0.' It was concluded that, the titration ml. distilled water) with standard 0.1N sodium hydroxide, was due to a neutralization equilibrium (Equation either at 3-5" or room temperature, gave curves which indicated a p& value between 9 and 10 and a neutralization 1) and not a decomposition of the aldehydic equivalent of 299 f 5 (calcd. 299). Immediate back titrafunction. The ability of I to form a stable gem- tion with 0.1N hydrochloric acid indicated a function of p& 4.0 and a neutralization equivalent of 600. This neutrdization equivalent corresponda to one mole of acid from two moles of carboxaldehyde a would be expected from a Cannizzaro reaction. A Beckmann Model H2 pH meter was used in this work.

I

I1

Acknowledgment. Elemental analyses were performed by the Analytical Research Branch, U. S. Army Chemical Research and Development Laboratories, Army Chemical Center, Md. The author wishes to acknowledge the technical assistahce of Arthur Jones and Arthur Melvin glycol (11) as well as the acid character of I1 were and a criticism of the manuscript by Dr. David attributed to the strong electron-withdrawing N. Kramer. The sample of l-isoquinoline carboxcharacter of the pyridinium ring. Subsequently, aldehyde was kindly provided by R. M. Poirier of it was found that l-formyl-2-methylisoquinolinium Battelle Memorial Institute. iodide (111) (in which it would be expected that the HC=O I

PROTECTIVE DEVELOPMENT DIVISION U. S. ARMYCHEMICAL RESEARCHAND DEVELOPMENT LABORATORIES ARMYCHEMICAL CENTER,MD.

I11

a-electron density of the ring at the position of the carbonyl group is markedly lower than in the 2position of I)2was decomposed rapidly during a titration with 0.1N alkali a t 3-5'. Back titration indicated the presence of a conjugate base of a weak acid of p K , 4.0. It is reasonable to assume that either the Cannizzaro reaction or cleavage (loss of -CHO as formic acid) occurred. In any case, the discovery that I11 is attacked by dilute alkali at such a rapid rate is interesting and may stimulate a more thorough investigation. EXPERIMENTAL

Cortical Steroids as Acetal-Forming Compounds with Aldehydes and Ketones RINALDO GARDI,ROMANO VITALI,AND ALBERTOERCOLI Received August 7, 1861

The recent communication by Tanabe and Bigley' on the 17cr12l-isopropylenedioxysteroids has prompted us to report the results of our independent work on the cyclic acetals of steroids with a dihydroxy acetone side chain. We have prepared several 17a,21-cyclic acetals (Table I) by an acid catalyzed interchange reaction2 between cortical steroids and lower alkyl acetals of aliphatic, cycloaliphatic, or arylaliphatic aldehydes or ketonesea By our procedure (see Experimental) yields were often as high as 90%, particularly in the cases of the cyclopentanone and benzaldehyde derivative^.^ When a /3-oriented hydroxyl group is present at (3-11, this also undergoes the interchange reaction; in this instance, in addition to the expected acetal I,

1-Formyl-2-methylisoquinoliniumiodide. To 0.97 g. (6.2 X 10-3 mole)' of l-isoquinoline carboxaldehyde (m.p. 53"; reported3 55-55.5') dissolved in acetone and contained in a carbonated beverage bottle was added 8.8 g. (0.062 mole) of methyl iodide. The bottle was capped and heated in an oven a t 60' for six days. The reaction mixture was cooled and filtered to give 0.48 g. (26%) of a red crystalline solid (needles) m.p. 203-205' dec. An infrared absorption spectrum was determined in potassium bromide and the curve exhibited a strong absorption band at 5.86 p in carbonyl stretching region. Ultraviolet absorption maxima in water, 2.5 X 10-6 M , mM (log e ) : pH 6.5, 341(3.65), 28263.41); pH 12.0, 334 (1) M. Tanabe and B. Bigley, J . Am. Chem. SOC.83, 756 (3.68), 272(3.63). Anal. Calcd. far CilHIJNO: C, 44.1; H, 3.4; 0, 5.3. Found: (1961). (2) Nonsteroidal acetals have already been prepared in a C, 43.5; H, 3.4; 0, 5.6. Potentiometric titration of 1-jormyl-2-methylisoquinolinium similar way. Cf. M. DelBpine, Bull. SOC.chim. France (3) iodtde (111). Potentiometric titration of I11 (100 mg. in 10 25,574 (1901); Ann. Chim. (7) 23,378 (1901). (3) Steroids with dihydroxyacetone side chain do not (1) G. M. Steinberg, E. J. Poziomek, and B. E. Hackley, react directly with aldehydes and ketones, or they do so in a quite different manner as in the case of formaldehyde. See Jr., J.Org. Chem., 26,368 (1961). (2) This is based on a qualitative correlation of electron R. E. Beyler, R. M. Moriarty, F. Hoffman, and L. H. Sarett, densities of various heterocyclic rings. H. C. Longuet- J. Am. Chem. Soc. 80,1517 (1958). (4) Benzaldehyde, as well as many other carbonylic comHiggins and C. A. Coulson, Trans. Faraday Soc., 43, 87 pounds, should give two epimeric acetals due to formation (1947). (3) R. S.Barrows! and H. G. Lindwall, J. Am. Chem. SOC., of a new asymmetric carbon atom. Up to now, however, we 64,2430 (1942). have been able to obtain only one derivative.

FEBRUARY 1962

669

NOTES

TABLE I 1 7 4 DERIVATIVE Reaction with

Derivative of Cortexolone Cortexolone Cortexolone 3-ethyl enol ether Cortisone Cortisone Cortisone Cortisone Cortisone 3-ethyl enol ether 16a-Bromocortisone Prednisone Prednisone Prednisone Prednisone Prednisone Prednisone Prednisone Cortisol Cortisol Prednisolone Prednisolone PrednisoIone Sa-Fluoroprednisolone Cortisol Prednisolone Prednisolone Prednisolone

M.P.

[a]?

Benzaldehyde 162-163 +e5 Cyclopen tanone 178-180 93 Cy clopentanone 113-116 100 +loo Benzaldehyde 200-203 +152" Acetone 169-173 Cyclopentanone 242-243 +155.5 Cyclohexanone 237-239 +151 -45 Cyclopentanone 127-130 +183 Cyclopentanone 183-184 Acetaldehyde 196-1 98 +182 $144 Butyraldehyde 183-185 $144 (2-Methy1)butyraldehyde 220-224 $140 Caproaldehy de 105-108 +167b Acetone 200-202 f146.5 Cyclopentanone 201-203 +I40 Cyclohexanone 243-244 +142 Acetone 194-195 Cyclopentanpne 225-230 +120 Acetaldehyde 2 17-22 1 +loo Acetone 245-246 +104c Cyclopentanone 228-230 90 Acetone 2 10-212 +99 11@,17a,21-BIsDERIVATIVE (11. Y = C2Hs) Acetone Glass +lo8 Acetaldehyde 156-160 $140 Benzaldehyde 216-218 +110 Acetone Glass +loo

+ -

+

+

Carbon, % Calcd. Found

Hydrogen, % Calcd. Found

77.39 75.69 76.32 74.97 71.97 73.21 73.60 73.98 61.78 71.85 72.79 73.21 73.60 72.33 73.56 73.94 71.61 i2.86 71.48 71.97 73.21 68.87

77.35 75.90 76.45 75.09 71.88 72.91 73.42 74.05 62.20 71.96 72.39 73.18 73.57 72.32 73.53 73.59 71.58 72.64 71.30 71.79 73.17 68.57

7.89 8.80 9.15 7.19 8.05 8.04 8.24 8.43 6.58 7.34 7.82 8.04 8.24 7.59 7.60 7.82 8.51 8.47 7.82 8.05 8.04 7.46

7.69 8.83 9.12 7.17 8.05 8.00 8.13 8.41 6.43 7.36 7.89 8.17 8.12 7.58 7.51 7.82 8.37 8.48 8.00 8.07 8.01 7.23

71.28 70.71 76.26 71.57

71.24 70.58 76.17 71.39

9.08 8.35 7.27 8.70

8.80 8.43 7.08 8.45

+

a Reported' m.p. 180-185', [ a ] D 200' (chloroform). Rep0rted'm.p. 201-203", [a]= 214' (chloroform). 0 Reported' m.p. 243-247', [ a ] D 106' (chloroform).

+

a second product 11,s easily identifiable by infrared spectrum (lack of hydroxyl band)6 and by paper chromatography,' can be isolated. ( 5 ) While writing this paper, we have read that D. K. Fukushima and S. Daum, J. Org. Chem. 26,520 (1961), have obtained the 1I@-methoxymethyl ether of 17cu,20; 20,21bismethylenedioxyhydrocortisone as a by-product of hydrocortisone BMD preparation. (6) The infrared spectra of the 17c~,21-cyclicacetals in Nujol mull show a characteristic prominent band a t 11281125 cm.-l, as well as a significant shift (10-30 cm.-1) of the 20-carbonyl band towards higher frequencies. The compounds with an acetal group a t C-11 exhibit a strong and broad absorption between 1130 and 990 cm.-' (7) Using the modified Bush-type system E4 of W. R. Ebsrlein and A. M. Bongiovanni, Arch. Biochem. Rioph. 59, 90 (1958), prednisolone derivatives I exhibit an average R f corresponding to 0.75, whereas the derivatives I1 present a R f value of 0.9. 'iHZO\

co

w

0

I

Y = lower alkyl

CHzO

c=x x=c'

I

OY

\

w

0

I1

aryl

The ratio of these two derivatives is closely dependent on both the nature and the quantities of the reacting compounds. For example prednisolone reacts with cyclopentanone diethyl acetal, giving primarily the monoderivative I (yield 65%) ; on the contrary, on reaction with benzaldehyde diethyl acetal, the bisderivative I1 is obtained in 70% yield. In both instances, however, the minor component is detectable by paper chromatography. With the diethyl acetals of either acetone or acetaldehyde, nearly equal quantities of I and I1 are obtained; but the ratio can be shifted towards I or I1 by employing a lesser or greater amount, respectively, of the reagent acetal. Derivatives I and I1 can be easily separated from the mixture by chromatography on Florisil. The ability of the lip-hydroxyl group to give acetals is somewhat decreased by the absence of a free dihydroxyacetone side chain in the starting compound. However, by using the general procedure, we were still able to obtain from prednisolone 21-acetate the corresponding 11-(a-ethoxy)benzyl ether, although not in a fully satisfactory yield. Corticosterone 21-acetate gave the proposed substance in quantities detectable by papergram only. The parent steroid may be regenerated from all of the above described steroidal acetals, including the bisderivatives 11, by acid hydrolysis in a yield higher than 9501,; they are stable to base.

670

NOTES

We have also prepared some 3-enol ethers of 17a,21-acetals,* either by enoletherification of the preformed 17a,21-acetal or by acetalization of a 3-enol ether. The usefulness of all these compounds as intermediates for several reactions is quite evident; a further advantage is the possibility of obtaining, in 1lp,17aI2l-trihydroxy steroids, an adequate protection either of the three hydroxy groups or of those of the side chain only, by employing the benzaldehyde and the cyclopentanone acetals, respectively. Moreover, some of these acetals are of peculiar biological interest. For instance, prednisolone acetonide (administered orally in oily solution) and prednisolone cyclopentylidenedioxy derivative (applied locally) exhibit an enhanced anti-inflammatory activity in comparison with that of the parent alcoh01.~

VOL.

27

The melting point rose to 213-215' after one crystallization from methanol. Hydrolysis of the acetal (300 mg.) with methanolic hydrochloric acid as above described gave 175 mg. of prednisolone, m.p. 234-236". Prednisolone l7a18l-acetonide (I. X =