Notes The Formation of
a
Dimer of N-Vinylcarbazole
SUZANNE MCKINLEY,’ JEANV. CRAWFORD, AND# CHI-HUA WANC
Department of Chemistry, Wellesley College, Wellesley, Massachusetts 08181 Received December 89, 1966
There have been a number of recent studies on the polymerization of N-vinylcarbazole either by ionic or free-radical mechanisms. Scott2 and co-workers and Ellinger3 observed that T-complex electron acceptors such as p-chloranil and tetracyanoquinodimethane initiated the polymerization of N-vinylcarbazole. Scott indicated that the polymerization process was qualitatively unaffected by the presence of thiophene, a potent retarder of conventional cationic propagation, and acrylonitrile, a radical or anionically readily polymerized monomer. In the presence of water and certain amines, however, Scott reported the retard% tion of polymerization of this monomer. The atypical nature of N-vinylcarbazole polymerization with respect to either ionic or radical mechanisms is further complicated by the observahion that oxygen3 and hindered phenols4 have relatively little or no effect on this process. Scott, et al., suggested that this is a cationic polymerization initiated by a Wurster radical cation formed by oxidation of N-vinylcarbazole. Ellinger indicated that there is a “partial transfer” of one electron from N-vinylcarbazole to the electron acceptor as the initiation mechanism. However, where ferric ion is used as the oxidant instead of n-complex-forming electron acceptors, N-vinylcarbazole as well as 4vinylpyridine was polymerized in methanol solution.6 Since it is not easy to visualize a “partial electron transfer” with ferric ion, we believe that there is oxidation-reduction by complete electron transfer between the metal ion and N-vinylcarbazole or 4-vinylpyridine. The resulting radical cation from such oxidation-reduction initiated the polymerization of the monomer. Scott, et al., suggested that the initiation and propagation probably do not involve intermediates with independent radical and ionic functions which implies that the Wurster ion of N-vinylcarbazole does not exist with a localized carbon-radical function, e.g, the following. It may not be completely relevant to compare the
mechanism of ferric ion and a-complex electron acceptors in initiating polymerization of N-vinylcarbazole. However, in our more recent experiments with ferric ion as the oxidant, we obtained evidence for the existence of a localized carbon-radical intermediate by isolation of a dimer of N-vinylcarbazole wherein the formation of a new carbon-carbon bond is involved. The Structure of the Dimer.-With a high concentration of ferric ion (ferric nitrate, anhydrous or hydrate) in methanol-water medium (90:10), there was almost immediate dimerization of N-vinylcarbazole in good yield. The white crystalline compound melts at 191193”. It does not decolorize bromine in carbon tetrachloride or potassium permanganate in acetone-water solution. The infrared spectrum showed neither the N-H stretching frequency near 3600 cm-’ nor the C=C vibration frequency near 1670 cm-l characteristic of N-vinylcarbazole. Elementary, mass spectral, and nmr analyses correspond to the dimer, trans-l,Z-dicarbaeylcyclobutane, previously reported by Ellingee as a by-product from the reaction of N-vinylcarbazole with chloranil or tetranitromethane. Mechanism of the Formation of the Dimer.-The simplest possible mechanism for the formation of such a dimer with a cyclobutane skeleton should be a fourcenter-type reaction. It will give 1,3-dicarbazylcyclobutane as the chief product. However, ferric ion is essential for the rapid formation of the dimer. It is not reasonable that the function of ferric ion is only to facilitate the delocalization of the olefinic T bond in the N-vinylcarbazole. The nature of Fe+3-Fe+2 oxidation-reduction, as well as the chemistry of Nvinylcarbazole, is by no means simple either in the dark or under photolysis. Thus, it is quite feasible to account for the participation of ferric ion in this process as an oxidation-reduction reaction leading to a carbon-free radical which then dimerizes. \”
B
+
217e3‘+
+ (1) Participant in the Wellesley College Institute of Chemistry supported b y the Nations1 Science Foundation. (2) H. Scott, G . A. Miller, and M. M. Labes, Tetrahedron Lettera, No. 17, 1073 (1963). (3) L. P. Ellinger, Chem. I n d . (London), 1982 (1963). (4) C. H. Wang, unpublished result. (5) C. H.Wang, Chem. Ind. (London), 751 (1964).
-
2Fe2+
>”+.
k
4- 2Fe24
4- 2Fe3+
This mechanism is consistent with the observation that the rate of formation of the dimer increases with (6) L. P. Ellinger, J. Fenney, and A. Ledwith, Monatsh. Chem., 96, 131
(1965). 1963
NOTES
1964
VOL. 31
the concentration of ferric ion present, and can also account for the absence of any isomeric l,&dicarbazylcyclobutane. Experimental Section7 Hydrolysis of N-Vinylcarbazo1e.-To a solution of 1 g (5 x mole) of N-vinylcarbazole (Matheson Coleman and Bell, mole) mp 67")in 44 ml of 9:1 methanol-water, 0.01 g (2.5X of ferric nitrate (Mallinckrodt, Fe(N0&.9H20) was added. The mixture was stirred a t room temperature and a white precipitate gradually appeared. At the end of 4 days, 0.7 g of white solid was collected. This material was identified as carbazole by melting point and rnixture melting point determinations with known samples of carbazole, mp 238-241°, and by comparison of their infrared spectra. The mother liquor furnished acetaldehyde in about 50% yield based on the isolation of acetaldehyde 2,4-dinitrophenylhydrazone, mp 145-147', lit.8 mp 148". Comparable results were obtained when the samemolar concentrations of hydrochloric acid were used instead of ferric nitrate. Formation of the Dimer.-To a solution of 2 g (0.01mole) of N-vinylcarbazole in 88 ml of 9:1 methanol-water, 0.2 g (5 X mole) of ferric nitrate was added. The mixture was stirred and a white precipitate was observed within 10 min. At the end of 1 hr, the solid was collected by filtration. The yield was 0.4 g (20%), mp 189-193'. Recrystallization from an ethanol-acetone (1:1) solution raised the melting point to 191-193.5'. Anal. Calcd for CzsHzzNZ: C, 87.01; H, 5.74; N,7.25; mol wt, 386.5. Found: C, 86.74; H , 5.80; N, 7.31; mol wt, 373 (Rast method), 386 (mass spectrum).
Acknowledgment.-We are indebted to Dr. G. Dudek of Harvard University and Mrs. G. Dudek of this department for the mass spectral and nmr analyses. (7) Analysis m a s by Dr. M. S . Nagy, Massachusetts Institute of Technology, Cambridge, Mass. Melting points are not corrected. The nmr spectrum was recorded b y using a Varian 4-60 spectrometer, and mass spectral analysis was done by using a mass spectrophotometer, A.E.I. MS9. (8) G. R. Cleroo and TV. H. Perkin, Jr., J. Chem. Soc., 126, 1804 (1924).
3-Hydroxy-4-Substituted 1,2,5-Thiadiazoles. A New Synthesis STEVEA. MIZSAKAND MELPERELMAN Lill y Research. Laboratories, Indianapolis, Indiana Receieed July 29, 1966
Although 2,1,S-benzothiadiazole bicyclic systems have been known since the last century,' 1,2,5-thiadiazoles were not described until 1957. The monocyclic system has been obtained by oxidation of 2,1,3benxothiadia'zole Cleri~atives2-~ to 1,2,5-thiadiazole-3,4dicarboxylic acid and also by basic cleavage of 1,2,5thiadiazolo [3,4-d]pyrimidin-7(6H)-one (I) to 4-amino1,2,5-thiadia~ole-3-carboxamide.~ Carmack and associates have synthesized 3,4-dicyano-1,2,5-thiadiazole6by ring closure from thionyl (1) 0. Hinsberg, BeT., 42,2895 (1889). (2) (a) A. M. Khaletukii, V. G. Pesin, and T. Chou, Dokl. Akad. Nauk SSSR,114, 811 (1957); Chem. Abstr., 62, 46051 (1958); (b) V. G. Pesin, A. M. Khaletskil, and T. Chou, Zh. Obshch. Khim., 28, 2089 (1958); English translation, J . Ct'n. Chem. USSR, 88, 2126 (1958), Consultants Bureau, Inc., New York, N. Y. (3) (a) M. Carmack, L. M. Weinstock, and D. Shew, Abstracts of Papers, 136th National Meeting of the American Chemical Society, Atlantic City, N. J., Sept 1959, p 37P; (b) M. Carmack, D. Shew, and L. M. Weinstock, U. S. Patents 2,990,408 and 2,990,409 (June 27, 1961); Chem. Abstr., 66,4775 (1962). (4) I. Sekikawa, Bull. Chem. SOC.Japan, SS, 1229 (1960). (5) Y. F. Shealy and J. D. Clayton, J . Org. Chem., 2 8 , 1491 (1963). (6) D. Shew, Dissertat:ion Abstr., 20, 1593 (1959).
1
III
If
chloride and hydrogen cyanide tetramer (11) and have developed a general synthesis from substituted oxalimidates (III).' Two other interesting approaches have led to 3cyano-4-hydroxy-l,2,5-thiadia~ole,~ obtained from the reaction of potassium cyanide and sulfur dioxide in the absence of hydroxylic solvents, and to 3-phenyl-1,2,5thiadia~ole,~ formed by refluxing S4N4 with ethylbenzene. a-Amino acid amides undergo condensation with 1,2dicarbonyl compounds to give 2-hydroxypyra~ines.'~ We are indebted to Dr. R. G. Jones of these laboratories for calling to our attention the fact that this demonstrated similarity of a-amino acid amides (IV) to aromatic vicinal diamines suggests a route for the direct synthesis of substituted monocyclic hydroxythiadiazoles" (V). RCHC I 'NH2
+
SOCI,
-
HO \c+N, I
2"
R/C*N
/s
V
Iv
The initial experiments were carried out in chloroform with thionyl chloride and alaninamide hydrochloride. Very low yields of a product were isolated that had the characteristics expected of 3-hydroxy-4methyl-1,2,Bthiadiaxole. Improved results were obtained when the a-amino acid amides were allowed to react with thionylanilinelz in pyridine. A number of 3hydroxy-4-substituted 1,2,5-thiadiazoles were thus prepared in yields usually ranging from 20 to 60% (Table I). The stable aromatic character of the ring system is demonstrated by the marked phenolic properties of the hydroxyl function.8 The pKa values determined were generally in the range 6.4-7.3. A large-scale preparation of methioninamide (VI) ,'Oo1* without purification of the ester intermediate, allowed the isolation of a methylsulfonium chloride by-product (VII) in about 30% yield. CHzSCHzCHzCHCONH2
I
NH2 VI
.HCI
CHI
I
CH&3CHzCHzCHCONH2 i-
c1-
I
NHz VI1
.HC1
(7) R. Y. Wen, ibid., PS, 4121 (1963). (8)J. M. Ross and W. C. Smith, J. A m . Chem. SOC.,86, 2861 (1964). (9) V. Bertini and P. Pino, Angew. Chem., 77, 262 (1965). (10) R. G. Jones, J . A m . Chem. SOC.,71, 78 (1949). (11) Personal communications with Dr. Carmack, of Indiana University, indicate t h a t he has prepared the 1,2,5,-thiadiazolein the same manner. (12) A. Michaelis, Ann., 274, 173 (1893); Ber., 24, 745 (1891). (13) The a-amino acid esters were prepared according to the procedure of T. Curtius and F. Goebel, J. Prakt. Chem., [2] 87, 150 (1888). The a-amino acid amides were prepared as methioninamide described in ref 10.
JUNE 1966
NOTES
1965
TABLE I
Compd
R
Yield pure, %
MP, 'C
Formula
VI11 IX
CHz.=CH CHsSCHzCH2
54 57
92-94 65
CrH4NzOS CsHsNzOSz
Methioninamide methylsulfonium chloride hydrochloride (VII) was allowed to react with thionyl chloride in chloroform. The resulting product was 3hydroxy-4-vinyl-1,2,5-thiadiaxole, dimethyl sulfide having been eliminated during the reaction. Methioninamide hydrochloride yielded the expected methylmercaptoethylhydroxythiadiazole without complication. Experimental Section14
3-Hydroxy-4-methyl-l,2,5-thiadiazole(X). Method A.-To 10 g of alaninamide hydrochloride's suspended in 200 ml of chloroform was added 150 ml of thionyl chloride. The wellstirred reaction mixture was heated to reflux for 48 hr. The reaction was concentrated in vacuo to a brown solid, which was then taken up in 600 ml of chloroform. Thesolution was washed with two 150-ml portions of water and extracted with two 150-ml portions of 10% sodium hydroxide solution. The basic extracts were cooled to O " , made acidic with concentrated hydrochloric acid, and extracted with four 200-ml pogtions of chloroform. The latter solution was washed with water, followed by brine, dried over magnesium sulfate, and concentrated in vacuo. A yield of 150 mg (1.2%) of product was obtained as a yellow solid. It was sublimed a t 75" (0.3 mm), mp 143". Method B.-To 22 g of alaninamide, free base, suspended in 1500 ml of dry pyridine was added 121 g of thionylaniline.12 The weI1-stirred reaction mixture was kept under a nitrogen atmosphere while being heated a t 90" for 16 hr. The reaction was worked up as in the method A t o yield 8.6 g of a rust-colored solid. The compound was purified by chromatography on silicic acid and Supercel: 6.56 g (22% yield), mp 14&146", pK.' = 7.10, mol wt 120 (calcd 116.1). Methioninamide methylsulfonium chloride hydrochloride (VII) was obtained as EL by-product, in about 30% yield, during the preparation of me1,hioninamide (VIII), mp 168-172" from ethanol. Anal. Calcd for CeH1&LN20S: C, 30.63; H, 6.80; C1, 30.12; N, 11.91; S, 13.61. Found: C, 30.56; H, 7.00; C1, 29.88; N, 11.87; S, 13.80. 3-Hydroxy-4-Substituted 1,2,5-Thiadiazoles (Table I) .-Compounds IX-XI11 were prepared from thionylaniline and the appropriate a-amino acid amide. The 4-vinyl compound (VIII) was obtained as the only product isolated from the reaction starting with methioninamide methylsulfonium chloride hydrochloride (VII) and thionyl chloride. The molecular weight of each of the compounds in Table I was determined from the electrometric titration curve. The values found agreed with those calculated within the limits of experimental error.
%-
-Calcd, C
37.48 34.07 31.02 41.64 53.91 56.22
H
N
3.14 4.57 3.47 5.59 3.39 4.19
21.86 15.89 19.43 15.72
-Found, C
37.30 33.85 31.21 41.86 54.23 56.02
%----H
N
PK,'
3.24 4.60 3.23 5.79 3.78 4.35
20.98 15.94
6.80 7.0 7.10 7.25 6.98 7.1
19.67 15.68
measurements: Messrs. H. Howard and L. Huckstep. ultraviolet spectra; Rlr. L. Spangle, titration; Messrs, P. Landis and J. Klemm, nmr; and hlr. D. Woolf and Miss M. Hofmann, infrared spectra.
Dehydration of the Four Stereoisomers of 1-Decal01 over Thoria, Silica-Alumina, and Silico-Phosphoric Catalysts'
FREDERICK G. sCHAPPELL2 AND HERMaN
PINES
The Ipatieff High Pressure and Catalytic Laboratory, Department of Chemistry, Northwestern University, Evanston, Illinois, 60100 Received September 16, 1965
Acknowledgment.-We are indebted to Messrs. W. L. Brown, G. M. Maciak, H. L. Hunter, A. Brown, D . Cline, and C. u'.Ashbrook for elemental analyses and Mr. L. A. White for large-scale preparations of starting materials. In addition, we wish to thank Dr. H. Boaz and his group for the physicochemical
I n the preceding paper of this series it was shown that the dehydration of 1-decalols over aluminas takes place preferentially through a trans-elimination reaction. Thus cis,cis-1-decalol formed about 89% 1,9-octalin and only 5% cis-l,Boctalin, while cis,trans-1-decalol yielded 4% of 1,g-octalin and 92% cis-1,2-octalin. A similar trend was obtained with trans,& and trans,trans-1-decalol. The present paper reports the dehydration of the four stereoisomers of 1-decalol over thoria and over acidic type catalysts, namely, silica-alumina and silico-phosphoric acid. The experiments were made in a micropulse reactor.' Thoria.-Two sets of experiments were made, one using 30 mg and the other 500 mg of thoria (Table I). With the lesser amount of catalyst only ca. 3-13% of the decalols underwent dehydration to octalins, while with the 500 mg of thoria the dehydration amounted to from 26 to 61%, depending on the decalol used. The most resistant toward dehydration was the trans,trans-1-decalol, which is in agreement with a previous observation. Part of the decalols underwent epimerization and dehydrogenation to cis- and trans-decalones. The dehydration seems to proceed mainly via a trans-elimination reaction as evidenced by the formation of 1,g-octalin as the principal product from the dehydration of the cis,cis-1-decalol, and cis-1,2-octalin as the main olefin from the cis,trans-1-decalol. The stereospecificity of the dehydration of decalols in the presence of thoria was not so great, however, as in the
(14) All melting points are corrected. Ultraviolet, infrared, and nmr spectra were obtained in ethanol, chloroform, and deuterioohloroform, respectively. TitrEtions were carried out in a 66% aqueoua dimethylformamide system.
(1) Paper IX in the aeries of Dehydration of Alcohols. For the previous paper, see F. G. Schappell and H. Pines, J . 070. Chem.. 81, 1735 (1966). (2) Taken from a Ph.D. dissertation submitted to the Graduate School, June 1965. Monsanto Co. Fellow, 1963-1964.
NOTES
1966
VOL.31
TABLE I COMPOSITION OF THE PRODUCTS FROM DEHYDRATTON OF ~-DECALOLS OVER THORIA AT 350" Thoria, m 8 -t,c-l-OHa-
r
-c,c-l-OHa30
Products
500
-C,c,t-l-OHa30
500
30
--t,t-l-OHa-30
500
500
Octalins, mole % 8.4 51.4 13.3 61.0 4.6 39.7 3.4 26.3 t, c-1-OH 3.0 10.1 2.1 3.8 88.9 33.0 0.7 2.4 t,t-l-OH 0.7 10.1 1.2 2.6 11.5 95.9 53.8 c,t-l-OH 1.2 83.3 28.6 0.5 c,c-l-OH 84.7 10.7 0.5 t-1-onec 2.4 12.7 1.3 3.9 2.5 14.2 13.6 c-1-onec 0.8 3.8 1.5 0.4 1.5 3.9 Octalins, % trans-l,25.3 15.8 14.4 12.7 13.3 17.6 16.3 trans-2,34.1 5.3 5.8 8.6 9.8 1,962.0 45.1 19.5 14.5 56.7 44.8 53.9 36.8 ~is-1~217.8 18.7 40.6 34.5 8.5 12.2 9,lo20.2 26.8 24.1 18.7 30.6 27.6 19.9 24.9 a c,c-1-OH corresponds to cis,&-1-decalol; c,t-1-OH to cis,truns-1-decalol; t,c-1-OH to truns,c&-l-decalol; t,t-1-OH to trans,trunsI-decalol. For structures of these alcohols see ref 1. All of the decalols were 10% solutions in t-butyl alcohol. I n runs where 500 mg of catalyst was used only about 50% of material balanced was maintained. t-1-one corresponds to trans-1-decalone; c-1-one to cis-1-decalone.
TABLE I1 COMPOSITION OF PRODUCTS FROM THE DEHYDRATION OF 1-DECALOL OVER SILICA-ALUMINA AND SILICO-PHOSPHORIC ACID CATALYSTS Alcohola
c,c-l-OH c,c-l-OH C,tl-OH c,f-l-OH t,c-1-OH t,c-l-OH t,t- 1-OH t,t-1-OH
Temp, "C
250 350 250 350 250 350 250 350
c,c-l-OH 250 c,c-l-OH 350 c,t-1-OH 250 c,t-1-OH 350 t,c-1-OH 250 t,c-1-OH 350 t,t-1-OH 250 t,t-1-OH 350 a See footnote a in Table I.
-Products Low boiling
2.9 1.1 0.8 1.2 2.5 0.9 1.8 1.2
formed, mole % Octalins
50.6 51.7 52.5 47.4 48.4 51.1 45.5 51.7 44.3 68.6 30.2 51.2 36.3 57.7 26.2 40.6
Higher boiling
trane-l,2
Silica- Alumina 3.7 5.2 1.4 6.1 3.4 3.8
6.3 3.6 5.5 4.2 3.7 6.1 6.5 6.1 Silico-Phosphoric Acid Trace 0.5 5.0 3.8 3.6 7.8 7.3 12.2
case of the alumina catalysts; this may be due in part to the ease of dehydrogenation of decalols to decalones, with a consequent epimerization of the alcohols. Silica-Alumina and Silico-Phosphoric Acid.-The experiments were made with both catalysts at 250 and 350" (Table 11). The dehydration reaction was apparently accompanied by an extensive doublebond migration as evidenced by the large amounts of 9,lOoctalins in the product. The dehydration occurred only to a small extent through a trans elimination, as indicated by t,he presence in the olefins of 26-31% of cis-1,2-octalin from &,cis-1-decalol; cis,trans-1-decalol under similar conditions produced only 5-90/, of the cis-1,2-octalin. The major part of the dehydration reaction seems to proceed via a cationic mechanism. Conclusion.-The acidic type catalysts such as silica-alumina and silico-phosphoric acid dehydrate 1-decalols. The reaction is not very stereospecific and only part of the dehydration occurred via a transelimination reaction. Extensive isomerization of 1,9to 9,lO-octalin took place.
C-omposition of octalins, % trans-2,3 1,g ~i~-1,2
2.0 5.6 4.1 4.6 6.5 6.1 Trace 0.5 2.9 3.3 6.4 7.8
9.10
22.9 25.9 15.8 22.0 24.4 29.0 21.9 27.6
5.2 7.6 26.3 14.6
65.5 59.0 52.3 53.6 67.8 60.3 65.1 60.2
38.6 32.0 17.6 19.4 27.6 36.3 24.4 27.8
9.0 15.4 31.2 37.9
52.4 51.6 46.2 38.9 65.9 52.6 61.9 52.2
Thoria is a more stereospecific catalyst for the dehydration of 1-decalols than are the acid catalysts. Its activity and selectivity, however, is not so great as that of the alumina catalysts.' Experimental Section The apparatus, which consisted of a pulse microreactor, and the procedure used were the same as described previously.1 Thoria was prepared by heating 10 g of thorium oxalate in a vertical furnace a t 400" for 4 hr under a stream of nitrogen. The thoria thus produced wm compressed at 38,000 psi, crushed, and sieved to 2&40 mesh. Silica-alumina was the commercial cracking catalyst, type S-45, manufactured by Houdry Process Corp., Philadelphia, Pa. The catalyst was sieved to 20-40 mesh. Silico-phosphoric acid wm obtained from Universal Oil Products Co., Des Plaines, Ill. The catalyst which can be prepared by calcining Kieselguhr with phosphoric acid*,' was ground to 2040 mesh size. (3) V. N. Ipatieff, U. S. Patents 1,993,512,1.993,513, 2,018,005,2,018,066, 2,020,649 (1935). (4) F. G . Ciapetta and C. J. Plank, Catalysis, 1, 344 (1954).
JUNE 1966
NOTES
1967 TABLE I
The Base Strengths and Chemical Behavior of Nitriles in Sulfuric Acid and Oleum Systems
HALF-TIMESFOR FORMATION OF PROTONATED AMIDEFROM NITRILEAND/OR PROTONATED NITRILE -Half-time
N. C. DENO,ROBERTW. GAUQLER, AND MAXJ. WISOTSKY
70
Department of Chemistry, Whitmre Laboratory, The Pennsylvania State University, University .Park, Pennsylvania 16806 Receiced September 16, 1966
Protonated nitriles reside long enough in SOa-HzO systems to evaluate [the basicity of nitriles by nmr spectroscopy. Acetonitrile, propionitrile, and benzonitrile are half-protonated in 98-1009;b HzS04. Chloroacetonitrile is half-protonated in 30% oleum. The rates of formation of protonated amides from nitriles are reported as well as rates of conversion of protonated amides to acyl cations. The concentrations of benzoyl cation and protonated benzoic acid are equal in 20% oleum. Aliphatic Nitriles--Arnett in his review' stated that the basicity of nitriles were the least known among the common functional groups of organic chemistry. From cryoscopic studies, Hantzsch2 had reported acetonitrile to be half-protonated in 10070 HzS04. A more precise value of 99.6% HzS04 was found on the basis of conductimetric Using the Hammett HO value^,^ pK would be - 10.12. I n sharp contrast to the above value is the value of -4.2 reported for acetonitrile5 and -4.31 for propionitrile.6 These latter values were obtained by titrating the nitrile with sulfuric acid in a formic acid solvent. The "titrations" were necessarily carried out to only a small per cent conversion of B to BH+, and we believe that the small changes in acidity appearing with added nitrile were not due to protonation of the nitrile but were due Lo the presence of free nitrile much as t-butyl alcohol has, a small effect on acidity in &20% even though it does not protonate.' Although the judgement to discount the titration work can rest on the above argument, it is naturally reinforced by the present nmr studies which indicate acetonitrile to be half-protonated in 100% HzS04 and propionitrile to be half-protonated in 98% H2S04. The nmr work thus supports the cryoscopic and conductimetric studies and definitively settles the basicity of aliphatic nitriles. Chloroacetonitrile was found to be half-protonated in 30 f 5% oleum. Despite this weaker basicity, its gross rate of hydrolysis to the protonated amide is greater as shown in Table I. This means that protonated chloroacetonitrile hydrates more rapidly than the other protonated nitrile by at least a factor of 3-10. I n fact there is an added factor of about 100 because the value of H o in 100% H2S04(where acetonitrile is half-protonated) is -10.8 and the value in 30% oleum (where chloroacetonitrile is half-protonated) is -12.EL8 It is of course a general principle that the (1) E. M. Arnett, Progr. .Phys. Ore. Chem., 1, 223 (1963). (2) A. Hantzsch, Z. Physik. Chem., 66, 41 (1909). (3) M.Liler and D. Kosanovic, J . Chem. Soc., 1084 (1958). (4) h'f. A. Paul and F. A . Long, C h e n . Rev., 67, 1 (1957). (5) H.Le Maire and H. J. Lucas, J . A m . Chem. Soc., 7 8 , 5198 (1951). (6) L. P. Hammett and A. J. Deyrup, ibid., 64,4239 (1932). (7) N.Deno, T. Edwards, and C. Perizzolo, ibid., 79,2108 (1957). (8) M.A. Paul and F. A. Long, Chem. Rea., 67, 1 (1957).
a t 35 i Z0,min 7%SOa (in 96 15 34
-7% HzSOa (in Hz0)90
&Sor)65
Acetonitrile 125 45 25s 25 25b c Propionitrile 120 45 30d 25 25b , Benzonitrile .. 40 35 .. e .. Chloroaceto20 15 5 2 .. heRussian workers refer to nitromethane ( 5 ) N. Deno. R. (1966).
W. GaLigler, and M. J. Wisotaky, J. Org. Chem., 81,1967
(6) fl. L. Haldna, H. J. Kuura, H. E. Laaneste, and R. K. Puss, Zh. Fiz. Rhim., 87,863 (1964).
The Basicity of Alcohols and Ethers N. C.DENOAND JOHN0. TURNER Department of Chemistry, Whitmore Laboratory, The Pennsylvania State University, University Park, Pennsylvania 16802 Received September 16, 1965
Arnett and Anderson’ established the base strengths of aliphatic alcohols in aqueous mineral acids and Arnett and Wu2 established the base strengths of ethers. Their method was to observe the changes in distribution coefficient of the alcohol or ether between an inert organic solvent and various concentrations of aqueous sulfuric acid. The concentration of alcohol or ether in the nonaqueous phase was determined by gas chromatography. A representative selection of their estimates of base strengths appear in Table I. I n the present work, solubility studies were made in order to evaluate the basicities. Solubilities were used many years ago by Hammett3 and the method is virtually identical in theory with the distribution The novel feature was to include a phenyl group into the structure, but at a sufficient distance from the basic site to make its extinction coefficient independent of protonation. The observed extinction in the aqueous phase at 260-270 mp was thus a measure of the sum of the concentration of B and BH+. In all cases, the solubility decreased from 0 to 20% H2S04 due to salting out. From 20 to 35% H2S04 a shallow minimum existed. From 35 to 60y0 H2S04, the solubility increased. The data were treated on the assumption that the solubility of free alcohol or ether does not vary from 20 to 60y0H2S04and that the increase in solubility observed is due solely to an increasing proportion of protonation. The uncertainties and errors in this method have been d i s c u ~ s e d . ~ ~ ~ It was not felt that the data were of sufficient precision to test any acidity function equation so that we are content to simply list in Table I the best estimate of the per cent H&04 at which CB equalled CBH+. Also listed are estimates of the same value for comparable compounds from previous literature work. (1) E.M.Arnett and J. N. Anderson, J. Am. Chem. Soc., 86,1542 (1963). (2) E. M. Arnett and C . Y. Wu, {bid., 82, 4999 (1960); W, 1680, 1684 (1962). (3) L. P. Hammett, “Physical Organic Chemistry,” MoGraw-Hill Book Co., Inc., New York, N. Y..1940. (4) E.M. Arnett, Progr. Phys. Org. Chem., I, 223 (1963).
NOTES
1970
VOL. 31
TABLE I VALUESOF PERCENTSULFURICACIDAT WHICHTHE ALCOHOLOR ETHERIS HALF-PROTONATED % HISO(
.4loohols
Primary Methanol 1-Butanol 8-Phenyloctanol Secondary 2-Propanol 2-Butanol 11-Phenyl-2-undecanol Tertiary 2-Methyl-2-propanol (t-butyl)
pKa
Method and ref
41 39.5 41
-2.5 -2.3 -2.5
Raman spectrab Distributionc Solubility
(50) 36.5 39
-2.2 -2.3
Interpretation of kinetic& Distributionc Solubility
41.5 55 f 5 38.5 41
-2.6
Distributionc Interpretation of kinetics* Distribution" Solubility
2-Methyl-2-butanol -2.3 2-Methyl-1 l.-phenyl-2-undecanol -2.5 Ethers Dimethyl 54 -3.83 Distributionb Diethyl 52 -3.59 Distribution' Methyl butyl 52 -3.50 Distribution' Methyl 4-phenylbutyl 49 -3.2 Solubility Calculated from the equation, Ho = p K log (CB/CBH+), using the values of HOtabulated by M. A. Paul and F. A. Long, Chem. Rev., 57,1 (1957). b N. Den0 and M. J. Wisotsky, J . Am. Chem. SOC.,85, 1735 (1963). d P. D. Bartlett and J. 0. McCollum, ibid., 78, 1441 (1956). 6 N. Deno, T. Edwards, and C. Perizzolo, ibid., 79,2108 (1957). f See ref 2.
+
(1
4-Phenyl-1-butanol and its corresponding secondary and tertiary alcohols were studied. The results were erratic and although we suspect cyclization to tetralin derivatives, this was not verified.
Acknowledgment.-Support from the Petroleum Research Fund of the American Chemical Society is gratefully acknowledged.
Experimental Section
Bitriptycyl'
SolubilityDeterminations.-A mixture of 150 ml of the aqueous acid and 0.2 g of the alcohol or ether were swirled until an aliquot (260-270 mp). of the solution gave a constant extinction at A,, Preparation of Compounds.-The route to 8-phenyloctanol was as follows. Suberic acid was converted to the anhydride by refluxing with acetic anhydride. Excess acetic acid and anhydride were removed by distillation and the suberic anhydride condensed with benzene using aluminum chloride. This is the method of Hill,5 who reported a yield of 7-benzoylheptanoic acid of 78%. We obtained 65%. This acid was reduced to crude 8-phenyloctanoic acid in 80% yield by reduction with amalgamated zinc and hydrochloric acid using the Clemmenson conditions .e Reduction of the crude acid with LiAlHa in ether gave a 56% yield of 8-phenyl-1-octanol, bp 146-147' (2.5 mm). A n d . Calcd for C14H220: C, 81.5; H, 10.8. Found: C, 81.5; H, 10.9. 10-Phenyldecanoic acid, bp 224225' (14 mm), was prepared in a manner identical with the preparation of 8-phenyloctanoic acid, but using decanedioic acid in place of octanedioic (suberic) acid. The acid chloride, bp 180-184' (12 mm), was prepared by refluxing the acid with thionyl chloride and a trace of pyridine. The acid chloride was treated with dimethylcadmium by the method of Cason' to produce ll-phenyl-2-undecanone, bp 137139" (3 mm), in 40% yield. A LiAlHd reduction in ether converted the ketone quantitatively to 11-phenyl-2-undecanol, bp 133-134' (13 mm). A n d . Calcd for C1&0: C, 82.2; H, 11.4. Found: C, 82.0; H, 11.6. Treatment of 11-phenyl-2-undecanone with excess CHaMgI quantitatively produced 2-methyl-ll-phenyl-2-undecanol,bp 130-131" (7 mm). Anal. Calcd for C18H8~O:C, 82.4; H, 11.5. Found: C, 82.0; H, 11.6. Methvl 4-uhenvlbutvl ether was preuared by the reaction of Cphenyi-1-bbtanol wiih potassium m&al in eiher followed by addition of exceas iodomethane. The boiling point of the ether, 108-109' (11 mm), was in accord with the literature.8 (5) J. W. Hill, J . A m . Chem. Soc., 64, 4105 (1932). (6) E. L. Martin, O w . Reactions, 1, 155 (1942). (7) J. Cason, Chem. Rev.,40, 15 (1947). (8) J. von Braun, Rsr., 44, 2871 (1911).
CONSTANTINE KOUKOTAS, STANLEY P. MEHLMAN, AND LEONARD H. SCHWARTZ~ Department of Chemistry, The City College of the City University of New York, New York, New York 10081 Received November 19,1965
While numerous experiments conducted over the past fifteen years could have been expected to yield bitriptycyl (I), this hydrocarbon has yet to be unequivocally described. Thus, coupling fails to take place when 9-bromotriptycene is heated with copper, zinc, or sil~er.~14 Similarly, bitriptycyl has not been reported among the products resulting from the reaction of 9-triptycyllithium with cobalt chloride, copper chloride, or nickel chloride, or the silver-catalyzed thermal decomposition of 9,9'-ditriptycylmercury, or the copper- or silver-catalyzed thermal decomposition of 9,9'-ditriptycyldiselenide.4 Bartlett and Greene,s in the course of studying the thermal decomposition of ditriptoyl peroxide, isolated a small amount of a high-melting solid, "compound x," which they suggested might be bitriptycyl. We now wish to report a synthesis of bitriptycyl, to describe the properties of this hydrocarbon, and (1) This work waa supported in part by grants from the National Science Foundation (GP3511), the Research Corporation, and the General Faculty Research Committee of the City College of New York. The purchase of the Cary Model 14 spectrophotometer used in this work was made possible by a grant from the National Science Foundation (GP-3646) to the City College of the City University of New York. (2) To whom inquiries should be addrewed. (3) P. D. Bartlett and E. S. Lewis, J . A m . Chem. Soc., 74, 1005 (1950). (4) G. Wittig and W. Toohtermann, Ann. Chem., 660, 23 (1962). (5) P. D. Bartlett and F. D. Greene, J . A m . Chem. SOC.,76, 1088 (19.54).
NOTES
JUNE 1966 4’
6’
20
4
I further, to confirm the structure assignment made by Bartlett and Greene for their “compound x.” The present synthesis of bitriptycyl involves the addition of benzyne to 9,9’-bianthryl. For this purpose, benzyne was generated from anthranilic acid,s 1,2bromoflu~robenzene,~ and lJ2,3-benzothiadiazole 1,ldioxide.8 The yielda of bitriptycyl ranged from approximately 5-20%. The synthesis involving benzyne generated from anthranilic acid was the most convenient and generally gave the highest yields and the purest product. The isolation of bitriptycyl was considerably simplified by the extreme insolubility of this hydrocarbon in acetone. Bitriptycyl melts at 577” with decomposition, and in the absence of oxygen appears to be thermally stable up to its melting point. The proof of structure of bitryptycyl is based on the following facts. Carbon and hydrogen analysis is consistent with the formula C49Hs6. The mass spectrum shows a base peak a t m/e 506 (molecular ion), a strong peak at m/e 253 (triptycyl ion plus a double charged molecular ion) and a metastable ion at approximately m/e 127. These data serve to establish the molecular weight and also indicate that cleavage of the molecular ion to the triptycyl ion is an important process. The 280 mp (log E 3.69), ultraviolet spectrum, A:oal;&ne 272 (3.63), and 266 (sh) (3.38), compares favorably with the ultraviolet spectrum of triptycene, 279 mp (log e 3.67), 271 (3.55), and 265 (sh) (3.32). The infrared spectrum is relatively simple and agrees with the published spectrum of Bartlett and Greene’s “compoundx.” Our interest in the bitriptycyl system arises from the possibility that suitably substituted derivatives, e.g., a 2,2’-disubstituted bitriptycyl, may exhibit conformational stability. We are presently investigating this possibility. Experimental Section Infrared and ultraviolet spectra were determined on PerkinElmer Model 137 and Cary Model 14 spectrophotometers, respectively. Mass spectra were determined on an Associated Electrical Industries MS-9 spectrometer. Elemental analyses were performed by Schwarzkopf Microanalytical Laboratory, Woodside, N.Y. The melting point of bitriptycyl is uncorrected.9 Reaction of 9,9’-Bianthryland Anthranilic Acid and n-Butyl Nitrite.-Solutions of anthranilic acid (4.12 g, 30 mmoles) and n-butyl nitrite (3.49 g, 34 mmoles), each in 27 ml of 2-butanone, were simultaneously added during a 3.5-hr period to a stirred (6) L. Friedman and F. M. Logullo, J . Am. Chem. SOC.,86,1549 (1963). (7) G.Wittig and L. Pohmer, Chem. Ber., 89, 1334 (1956). (8) G.Wittig and R. W. Hoffmann, ibid., 96, 2718 (1962). (9) The melting point of bitriptycyl was determined in a sealed tube, under nitrogen, using a zinc chloride bath and a 10G-62O0 partial-immersion thermometer, supplied by the Chemical Rubber Company, Cleveland, Ohio.
1971
refluxing solution of 9,9’-bianthryl1° (0.50 g, 1.4 mmoles) in 25 ml of 2-butanone. The n-butyl nitrite solution was always in slight excess over the anthranilic acid solution. The reaction mixture was refluxed for an additional 1 hr and allowed to stand a t room temperature for a t least 12 hr in order to ensure complete precipitation of the product. The solid was filtered and repeatedly triturated with acetone. Two crystallizations from nitrobenzene gave a white solid: mp 577 f 5’ dec; yNujol 1284 (w),1151 (m), 1133 (w), 1036 (m), 915 (m) 808 (m), 771 (m), 753 (m), 745 (s), and 738 (s) cm-l; At:$ ( IO-cm cell) 280 mp (log E 3.69), 272 (3.63), and 266 (sh) (3.38); m/e 506 (base peak), mle 253 (56% of base peak), and m/e approximately 127 (metastable ion). For comparison, triptycene has ?:A: 279 mp (log E 3.67), 271 (3.55), and 265 (sh) (3.32). Anal. Calcd for C~OHZS: C, 94.83; H, 5.17. Found: C, 94.96; H, 5.15.
Acknowledgment.-The authors thank Dr. William Milne of the National Institutes of Health for the measurement and interpretation of the mass spectra. (10) F. Bell and D. H. Waring, J. Chem. Soc., 1579 (1949).
Magnetic Nonequivalence of a Thiophosphate Ester R. V. MOENAND W. H. MUELLER Analytical Research Division and Central Basic Research Laboratory, Esso Research and Engineering Company, Linden, New Jersey 7036 Received November 1.3, 1966
Magnetic nonequivalence due to molecular asymmetry has been reported for a variety of compounds.’-3 Roberts, et d J 4 studied nmr spectra of 1-phenylethyl benzyl ether and related compounds. I n particular, the effect of the proximity of an asymmetric center on the extent of magnetic nonequivalence was investigated. Sidall and Prohaska5 have observed nonequivalence of the alkoxy1 groups on a variety of phosphorus esters. This effect was explained by a preferred conformation of these esters. Bentrudeearecently- reported that
LCl8
shows magnetically nonequivalent methoxyl groups (A = 4 cps), ascribed to the asymmetric carbon in the molecule. (1) E.I. Snyder, J . Am. Chem. Soc., 86, 2624 (1963). (2) J. K. Randall, J. J. McLeskey, 111, P. Smith, and M. E. Hobbs, i b i d , , 86,3229 (1964). (3) R. M. Moriarty, J . Ow. Chem., 80,600 (1965). (4) G. M. Whiteaides, D. Holtz, and J. D. Roberts, J. Am. Chem. Soc., 86,2628 (1964). (5) T.H.Siddall and C. A. Prohaska, ibid., 84,3467 (1962). (6) (a) W.G. Bentrude, ibid., 81, 4026 (1965). (b) It was pointed o u t by a referee t h a t the nonequivalence in this compound presumably originates in a folded conformation I of higher energy than some open conformation 11, and increasing the temperature increases the population of I.
* 3-s
\
P=O
CHs
/
‘CH”
I
--+
*C=S
0-CH4Hs
P ‘’
‘0 I1
1972
NOTES
VOL.31
TABLE I SOLVENT EFFECT ON
Dipole moment
Solvent
THE
NONEQUIVALENCEO
Dielectric conet
Chemical shift of methyl triplet -P(OCHtCHa)z
Separation of triplets, CPS
Neat 1.17 0 Nonpolar Carbon tetrachloride 0 2.3 1.17 1.26 5 Hexachlorobutadiene 0 2.57 1.13 1.22 5 Benzene-ds 0 2.28 0.93 0.98 3 Carbon disulfide 0 2.64 1.13 1.21 4 Polar Chloroform-d 1.02 4.80 1.23 1.24 1 Acetonitrile 3.84 30 1.20 1.21 1 Acetone 2.89 20.7 0 1.20 Dimethyl sulfoxide 3.9 45 1.18 1.19 1 Pyridine 2.22 5.8 1.18 0 All the spectra (except the first) were run with -10 vol % of the compound in the above solvents a t ambient temperature. Chemical shift values are in parts per million downfield from tetramethylsilane (internal standard).
VOL. % COMP.
Figure 2.-Effect
IN CC14
of concentration on nonequivalence.
We now wish to report magnetical nonequivalence of protons separated from an asymmetric carbon center by a distance of six bonds. This effect was observed in 0,O'-diethyl S-(a-phenylethyl)thiophosphate,I (Figure l).' The nmr spectrum of I in CC1, shows two triplets
(7) The spectra were taken on 20% solutions in CClr. They were recorded with a Varisn Associates A-60 high-resolution spectrometer. Chemical shifts are in parts per million downfield from tetramethylsilane.
(A = 5 cps) for the methyls of the ethoxy groups. The separation of these triplets is too large to be a longrange splitting caused by coupling with the phosphorus nucleus (spin 0.5). It must be the result of the two methyls being magnetically nonequivalent. The longrange coupling with the phosphorus nucleus was observed on an expanded scale and found tobeJPoc-cHa = 0.6 cps. The separation of the triplets in CCh depends on the compound's concentration as shown in Figure 2. The greatest separation of the two triplets (A = 6.5 cps) was observed at 1 vol % concentration. When the compound was run without solvent a t ambient temperature only one triplet was observed. At loo", however, two triplets ( A = 2 cps) appeared.6b An 80 vol % solution in CC1, was observed a t different temperatures in order to establish that the chemical shift of the two methyl groups had not crossed over in going from a dilute solution a t ambient temperature to a neat solution at 100". The separation of the triplets was 1.4 cps a t 30" and 2.2 cps at 75" for the 80% solution. The fact that the separation increased with increasing temperature indicates that the triplet,s of the methyl groups had not crossed over. The values in Figure 2 are measured at ambient temperature. We have also studied the effect of other solvents on the nonequivalence of these ethoxy groups. The separation of the triplets was greatest (-5 cps) in nonpolar solvents and became 0 (0-1 cps) in polar solvents. The results are listed in Table I and the
JUNE 1966
1973
last column shows clearly the difference between the effectof polar and nonpolar solvents. These results are in qualitative agreement with the finding of Roberts and co-workers.s They reported that in the case of 1-phenylethyl benzyl ether the degree of magnetic nonequivalence bears an approximate inverse relation to the dielectric constant of the solvent. In view of the preceding, it was surprising to find that 0,O'-diethyl S-(1-indanyl) thiophosphate did not show a similar nonequivalence for the ethoxy groups (Figure 3). It should be noted in this connection that Robertss also found, that, in contrast to the methylene group in 1-phenylethyl benzyl ether, the corresponding protons in indanyl benzyl ether were equivalent in all solvents investigated. We have also studied a number of dmilar thiophosphates not containing an aromatic ring moiety and they all exhibit a single triplet at 1.33 ppm for all the methyl protons of the ethoxy groups. The indanylthiophosphate (Figure 3) also shows a triplet a t this position demonstrating that the T system of the benzene ring does not effect the methyl's chemical shift. In contrast, the two triplets from compound I are both shifted upfield. Thus, the two ethoxy groups apparently experience diamagnetic shielding, although to a different extent. More data will be necessary in order to draw valid conclusions as to the possible factors causing this nonequivalence. We feel, however, that the data supplied here are of interest to workers in this field and may stimulate such further research.
tail.' When an attempt was made to extend this to the preparation of 2,4-diazido-l,5-dicyano-l,3,5-triaza1,4-pentadiene (I), NC-X=C(N3)NHC(N3)=NCN, from the disodium salt of bis(5-tetrazo1yl)amine and 2 molar equiv of cyanogen bromide, there was obtained a base-insoluble product whose infrared spectrum revealed azido function a t 2140 cm-1 (Nujol mull or N,Ndimethylformamide solution), but unexpectedly no nitrile function (cyanoguanyl azides show both functions). I n addition the compound remained unchanged when treated with base under conditions which convert cyanoguanyl azide derivatives to tetrazolines. A better yield of the same compound was realized when only 1 mole of cyanogen bromide was employed per mole of bis(tetrazoly1)amine. The analytical data indicated that only one cyano group, rather than two, had been introduced per mole of starting bis(tetrazoly1)amine. Reduction of the compound with hydrogen sulfide in ammoniacal aqueous ethanol gave sulfur and melamine. A structure consistent with all of these facts is 2-amino-4,6-diazido1,3,5-triazine (IV), whose formation can be depicted by Scheme I. One of the tetrazole rings is opened by the SCHEME I
(8) G. M. Whitesides, J. J. Grooki, D. Holtz, H. Steinberg, and J. D. Roberts, J . Am. Chem. Soc., IST, 1058 (1965).
III
If 2-Amino-4,6-diazido-1,3,5-triazine RONALD A. HENRY
Chemistry Division, Research Department, U.8. Narial Ordnance Test Station, China Lake, California 93667 Receitled December 93, 1966
The opening of the ring in 5-(substituted-amino)tetrazoles by cyanogen bromide to give substituted cyanoguanyl azides was recently examined in some de-
I11
Iv
cyanogen bromide in the expected fashion to yield I1 which then undergoes an intramolecular cyclization (1) W. P. Norris and R. A. Henry, J . Ore. Chem., 49, 650 (1946).
NOTES
1974
between the nitrile group and the other tetrazole ring to give the substituted tetrazolotriaxine, 111. Because the triazine ring in the latter is electronegative, the second tetrazole ring is destabilized and is opened to yield the isomeric azidotriazine IV. Previous investigations2 have conclusively demonstrated with heterocyclic systems the existence of an azidoazomethine-tetrazole equilibrium, the position of which is dependent on solvent, temperature, and the electronegativity of the heterocyclic ring. With the present compound the evidence suggests that in solution at least the equilibrium strongly favors IV over III.3 For example, 2 moles of nitrogen/mole of IV were rapidly evolved when the latter was treated with either potassium iodide in trichloroacetic acid solution or potassium arsenite in potassium hydroxide solution. With either reducing agent, a plot of the volume of nitrogen evolved as a function of reaction time revealed no break which might indicate a rapid evolution of 1 mole of gas followed by the slower evolution of the second. Furthermore, with a fixed-thickness cell and with solutions (of the same molar concentration the intensity of the azido absorption in the infrared was the same in either N,N-dimethylformamide or trifluoroacetic acid ((probably because of protonation there was a shift of ca. 70 cm-1 to a higher frequency in the latter solvent). This behavior should be contrasted with that of Z-azido-4,6-dimethylpyrimidine which reveals azido absorption in trifluoroacetic acid a t 2180 cm-1, but none in N,N-dimethylformamide since the equilibrium now favors the isomer, 5,7-dimethyltetrazolo [l,ba]pyrimidine. Hart4 previously prepared 2-amino-4,6-diazido-l13,5triazine by ammonolyzing 2,4,6-triazido-1,3,5-triazine in diethyl ether; lie also reduced this compound to melamine with hydrogen sulfide. A base-soluble material of undetermined structure was also formed in the reaction of cyanogen bromide and disodium bis(tetrazolyl)amine, and was recovered by acidifying the mother liquors after the aminodiazidotriazine was removed. This compound, which contains an azido but apparently also no cyano group and which rapidly turns purple in the light, was difficult to purify because of its gelatinous nature in aqueous solutions and its insolubility or very low solubility in all common solvents. The analyses suggest a 2: 1 reaction product but its properties argue strongly against its being I. An attempt to extend this ring-opening, recyclization reaction to the preparation of 2-azid0-4~6-diamino1,3,&triazine, by reacting equimolar amounts of cyanogen bromide and the sodium salt of 5-guanylaminotetrazole, was unsuccessful. Experimental Section 2-Amino-4,6-diazido-I ,3,S-triazine.-A solution consisting of 8.55 g (0.05 mole) of bis(5-tetrazolyl)amine, 2.0 g (0.05 mole) of sodium hydroxide, arid 75 ml of water was cooled to 5 " . With
VOL.31
stirring, 5.3 g (0.05 mole) of cyanogen bromide was added in three portions during 5 min. Then 10 ml of acetone was added; the temperature rose to 8" and held for several minutes. When the temperature had dropped to 5', 20 ml more of acetone was added and the solution allowed to stand for 2.5 hr a t 0-5". The pH was then adjusted to the phenolphthalein end point and the product removed by filtration, washed well with cold water, and dried; yield, 5.75 g (64.7%). The gray product turned pink in sunlight. It was sparingly soluble in diethyl ether and in ethanol; a small portion (2.3 g) was recrystallized for analyses from 600 ml of 95% ethanol; the flat white needles did not melt up to 250' although they turned brown a t 210-220'. When dropped on a hot surface at 300', they exploded. Anal. Calcd for C3H2N10:C, 20.23; H, 1.13; N , 78.64; mol wt, 178; azide nitrogen, 47.2. Found: C, 19.90; H, 1.09; N, 78.8, 78.2; mol wt, 174; azide nitrogen,6 46.7; azide nitrogen,G45.8. When this reaction was performed using 0.1 mole of bis(tetrazolyl)amine, 0.2 mole of sodium hydroxide, and 0.2 mole of cyanogen bromide, the yield of aminodiazidotriazine was 2.3 g (12.9%). The aminodiazidotriazine (1.03 g) was dissolved in 150 ml of 95% ethanol, 50 ml of water, and 0.5 ml of concentrated aqueous ammonium hydroxide by warming on the steam bath; hydrogen sulfide was bubbled into the hot solution until no more sulfur precipitated. (No reduction occurred in the absence of the ammonium hydroxide.) The solution was filtered hot and the sulfur washed with hot water. The yield of dried sulfur was quantitative (0.37 9). The combined filtrate and washings were evaporated to dryness and the residue extracted with carbon disulfide to remove any traces of sulfur. The yield of crude product melting above 310" and considered to be melamine was quantitative (0.73 g) ; the infrared spectrum was essentially the same as that for authentic melamine. The X-ray powder pattern also confirmed melamine. The cold, aqueous mother liquors from the original reaction were acidified with an excess of concentrated hydrochloric acid. A voluminous, gelatinous, yellow solid separated; warming did not improve its physical properties. Some of the material was removed by filtration and washed well with water. The dried product decomposed explosively on a hot plate at 300'. The infrared spectrum revealed broad, hydrogen-bonded N H a t 3230-3030 cm-1, C=N at 1620, and azide a t 2140; the latter mas confirmed by the liberation of iodine when a solution of the compound in trifluoroacetic acid was treated with potassium iodide in acetonitrile. This material was purified by solution in aqueous sodium bicarbonate (rapid reaction) , filtering from a small amount of insoluble, and reprecipitation in an excess of dilute, aqueous hydrochloric acid. The gelatinous solid was filtered and washed repeatedly with water until free of salt and inorganic acid. The gray-purple, amorphous compound was dried under vacuum at 85' (25 mm) for 72 hr; the infrared spectrum mas identical with that for the crude product. This compound was insoluble in pyridine, acetone, ethanol, and tetrahydrofuran; it was slightly soluble in dimethylformamide and tetramethylurea; it dissolved in dimethyl sulfoxide with gas evolution and the formation of a dark-colored solution; it was sparingly soluble in hexafluoroacetone hydrate, but again the solution was very dark suggesting some decomposition. The nmr spectrum of a solution of the compound in deuterium oxide-sodium carbonate revealed only exchangeable protons; a saturated solution in tetramethylurea showed only a broad absorption centered at r 3.6 (proton on nitrogen). Anal. Calcd for (C2HaNo),: C, 21.62; H, 2.72; N, 75.66. Found: C, 21.42, 21.50; H, 2.60, 2.79; N, 75.74, 75.50. When some of this second compound in aqueous sodium bicarbonate was treated with hydrogen sulfide, sulfur was precipitated. The material recovered by acidifying the resulting solution showed neither azido nor cyano absorption in the infrared spectrum; the potassium iodide-trifluoroacetic acid test was also negative.
(2) C. Temple, Jr., and J. A. Montgomery, J . Or@. Chem., SO, 828 (1965), and references therein.
(3) Elucidation of the structure in the solid state is currently being attempted by Dr. G. J. Palenik from X-ray crystallographic data. (4) C. V. Hart, J . Am. Chem. Soc., 60, 1922 (1928).
(5) Determined gssometrically by Dr. W. R. Carpenter of this laboratory; W. R. Carpenter, Anal. Chem., 86, 2352 (1964). (6) I. Ugi, H. Perlinger, and L. Behringer, Ber., 91, 2330 (1958).
JUNE 1966 The Sodium-Ammonia Reduction of Bispropargyl- and Propargylallylamines. Steric and Conformational Effects1
1975
NOTES
R
= CMeS) was reduced in 68% yield to a product (bp 58-68' at 18 mm) which was composed of 64%
1-tbutyl-3-methylene-4-methylpyrrolidine(D, R = CMea), hydrochloride mp 183-185", and 36% diallylt-butylamine (E, R = CMea), bp 62-64' at 15 mm., G . F. HENNION AND R. H. ODES hydrochloride mp 124-126'. That ring closure followed the path A B D was established by the fact The Chemical Laboratories, University of Notre Dame, that N-allyl-N-propargyl-t-butylamine (B, R = C11e3), Notre Dame, Indiana 46666 prepared independently from t-butylallylamine and propargyl bromide, gave the same products in the same Received December S1, 1966 ratio (see Scheme I). The product composition data of Table I were obThe reduction of sterically crowded N,N-bis(1,ltained by glpc. When the amount of cyclized product dimethylpropargy1)methylaminewith sodium in liquid was small, as in Table I, entries 3, 4, 6 and 7, separation ammonia produced heptamethyl-3-pyrroline (a) in for purposes of characterization was achieved by 72% yield while the same reaction applied to N-(1,ldimethylpropargy1)-N- (1,l-dimethylallyl)methylamine preparative glpc using a 20-ft Carbowax 20 M column (cf. Table 11, V and XI). Compounds XIV and XV, c H 3 obtained in the ratio 36:64 as stated above, were separable as their hydrochloride salts by fractional crystallization from a mixture of ethyl acetate and absolute ethanol. That one of the reduction products was N,N-diallyla b alkylamine was easily established in all cases by indegave the isomeric hexamethyl-3-methylenepyrrolidine pendent synthesis from alkylamine and allyl bromide. That the other product was the isomeric 3-methylene(b) in 61% yield.3 The corresponding N-demethylpyrrolidine was shown by pmr and infrared examinaamines, conformationally less restrained with N-H tion. Thus the pmr spectrum of 1-tbutyl-3-methylenein place of N-CH3, also gave good yields of cyclized 4-methylpyrrolidine hydrochloride (XV HC1) in chloroproducts except that the product purities were not so form-d showed a doublet at 7 8.61 (3 H) assigned to high owing to some competitive reduction without cyclithe C-4 methyl group, a singlet at 7 8.55 (9 H) for the tzation. Since this new4 cyclization reaction seemed to butyl group, two doublets at 7 4.73 ( 2 H) assigned to hold promise for the facile synthesis of substituted pyrthe olefinic protons, and a multiplet from 7 5.5 to 7.5 rolines and pyrrolidines, we have studied the reduction (6H) assigned to the methylene protons (4), the prooutcome as related to steric and conformational effects ton at C-4 (I), and the nitrogen proton (1). Also the in the substrate. Acr clearly revealed by the data of infrared spectrum showed peaks near 6.0 and 11.0 p Table I, cyclization occurs as the major reaction only characteristic of R&=CH2. Furthermore, hydrogenawhen the unsaturated centers are conformationally tion of XV HCl gave a saturated product (XVI * HC1) restrained to close proximity. Thus the sodiumammonia reduction of N,N-dipropargylmethylamine whose pmr spectrum showed two sets of doublets and a singlet at 7 9.13, 8.85, and 9.0, respectively (15 H), (A, R = Me; not sterically crowded) gave a 72% yield assigned to the C-3 and C-4 methyl groups and the tof product which was found to be 95% diallylmethylbutyl, plus two multiplets centered at 7 7.75 (3H) and amine (E), independently synthesized from allyl bro7.0 (3 H) for the single protons at C-3 and C-4 and the mide and methylamine. The same result was had with two methylenes bonded to nitrogen. N,N-dipropargylethylamine (A, R = Et) which yielded The sodium-ammonia reduction of N-propargyl-Nmostly diallylethylaniine (92% of the product). benzyl-t-butylamine (Table I, entry 10) was included to On the other hand, N,N-dipropargyl-&butylamine (A, ascertain if ring closure of the ethynyl group with aromatic carbon could be achieved in this seemingly SCHEME I favorable instance. Cyclization did not occur, however, and the product proved to be entirely N-allyl-Nbenzyl-t-butylamine. Mechanistically, cyclizations induced by sodiumE B ammonia likely proceed essentially as suggested by Stork, et al. ,4a for cyclization of y-ethynyl ketones. As applied to our substrates, however, the reaction is believed to involve the union of a radical center (rather than a c arb anion^^) derived from one ethynyl group with the appropriate allylic or propargylic carbon atom in the other unsaturated group. Clearly, considerable L C D steric assistance is required for such reactions to occur ~in good yield and reactions B -.F D and A --t B --t D re(1) Paper no. 83 on subsitituted acetylenes. Previous paper: G. F Hennion and C. V. DiGiovanna, J. Org. Chem., $1, 970 (1966). quire less steric crowding than does the reaction3 (2) Eli Lilly Co,Fellow, 1964-1965. A 4 C. (3) G . F. Hennion and C. V. DiGiovanna, J. 078. Chem., SO, 2645 (1965). (4) For recently reported related cyclizations also achieved via reductions It is also worthy of note that two tertiary amine hywith sodium, see (a) G . Stork, S. Malhotra, H. Thompson, and M. Uchidrochlorides, those prepared from N,N-diallylisobayashi, J. A m . Chem. Soc.. 87, 1148 (1965); (b) M. Eakin, J . Martin, butylamine and from N-allyl-N-benzyl-t-butylamine, and W. Parker, Chemical Commun., NO. 11,206 (1965). --+
-
1
-
NOTES
1976
VOL.31
TABLE I SODIUM-AMMONIA REDUCTIONS OF BISPROPARQYLAND PROPARQYLALLYLAMINES R'R'NCHZCSZCH Entry
1 2 3 4 5 6
7 8 9 10
E =
a
bp 63-65'
R' F'ropargyl Propargyl Propargyl Allyl I'ropargyl F'ropargyl Allyl F'ropargyl Allyl Benzyl N,N-diallylalkylamine and (0.5mm).
R'
BP, OC (mm)
Reduction products' -Composition, Yield, % Acyclic E
%--
-
Cyclic D
Methyl 108-112 (atm) 72 95 5 Ethyl 71-74 (105) 70 92 8 n-Propyl 79-83 (75) 71 88 12 n-Propyl 76-82 (74) 79 84 16 Isopropyl 57-61 (18) 64 82 18 Isobutyl 74-81 (50) 79 86 14 Isobutyl 74-85 (47) 82 83 17 &Butyl 58-68 (18) 68 36 64 &Butyl 63-68 (18) 73 38 61 t-Butyl 40-51 (0.1) 82 loo* 0 D = l-alkyl-3-methylene-4-methylpyrrolidine. Product is N-allyl-N-benzyl-&butylamine,
TABLE I1 AMINESAND PYRROLIDINES RlR2R3N Compd
I I1 I11 IV
V VI VI1 VI11 IX
X XI
RZ
111 CzHs n-CaH7 7L-CmH7 n-CaH7 n-CaH7
R' HCSCCHz HCSZCCHz HCZCCHz HCECCHz HCZCCHZ CHz=CHCHz CHz=CHCHz CHsCHCHz -CHzCH(CHs)C(=CHz)CHr
CCIHT i-CsH? i-C4Ha I-CIHB X4H9 i-CdHp
HCEZCCHz HC=CCHz CHz=CHCHz CHFCH-CHz HCECCHz HCECCHz CHz=CHCHz HCECCHz CHGCHCHZ CHz=CHCHz -CHzCH(CHa)C(=CHz)CHr
XI1 XI11 XIV XV
HC3CCHz HC=CCHz t-CiHp CHsCHCHz ~-CIHP H C E C C H z C"z=CHCHs CH-CHCHz t-CdHs t-C4Ho -CHzCH(CHa)C(=CHz)CHr
XVI XVII
t-CtH8 t-C4H9
0
-CHzCH(CHs)CH(CHs)CHr HCEZCCHz CaHsCHz
See also Table I.
--Hydrochloride BP, MP, OC Method OC (mm) Yield, % Formula B 84-86(65) 55 CsHiaClN 142-144 A 90-92 (55) 32 CsHuClN 142-144 C 76-78 (47) 28 CoHiaClN 134-136 A" 78-79 (73) 15 CiiHi9NOr 119-121 See Table I, entries CiiHioNOr 191-192 3 and 4 A 72-74 (18) 32 CoHirClN 160-161 Aa 55-57 (23) 30 CsHiaClN 125-127 B 77-80(30) 57 CioHisClN 168-169 C 78-84 (50) 41 CioHisClN 135-137 B" 67-71 (50) 21 See Table I, entries CizHziNO4 168-169 6 and 7 C 74-76(18) 41 CioHiaClN 204-206 C 63-64 (20) 45 CioHisClN 186-1 87 A" 62-64 (15) 3 CioHzoClN 124-126 See Table I , entries CioHzaClN 183-185 8 and 9 89 b CioHzzClN 181-183 53 C 61-64 (0.3) CirHzoClN 201-203
* By hydrogenation of XV as previously described.*
or oxlate salt
c-
Calcd 60.95 62.97 62.23 57.62 57.62
Found 61.12 63.00 62.20 57.77 57.82
H-Calcd Found 7.68 7.90 8.22 8.45 9.29 9.45 8.35 8.57 8.35 8.44
62.96 61.52 64.68 63.98
62.95 61.48 64.89 63.95
8.22 8.30 10.32 10.53 8.69 8.79 9.66 9.65
-%
59.24 59.44
7-%
8.70
8.88
64.53 64.02 63.47 63.18
8.69 8.60 9.66 9.68 10.63 10.58 10.63 10.77
62.64 62.79 70.72 70.67
11.57 11.81 8.48 8.60
64.68 63.98 63.30 63.30
New compounds prepared as starting materials are included in Table I. No attempt was made to produce these in optimum yields. The methods used were as follows. Method A.-A mixture of 0.5 mole of alkylamine (RNHz), 1.2 moles of anhydrous powdered potassium carbonate, and 250 ml of ether was stirred mechanically, and 1.2 moles of allyl or propargyl bromide was added dropwise. The mixture ww boiled, with stirring, for 6 to 18 hr. Water was added to dissolve the salts. The ethereal layer was extracted with 100 ml of 6 N hydrochloric acid with cooling. The acidic aqueous solution was extracted with two 50-ml portions of ether (discarded) and then treated with 100 ml of 6.5 N sodium hydroxide solution. The layers were separated and the aqueous portion was extracted with three 50-ml portions of ether. The amine layer and ether extracts were combined, dried over anhydrous potassium carbonate, and then distilled.
Method B.-Propargyl or allyl bromide (0.75mole) was added dropwise with stirring to 1 mole of aqueous primary amine (e.g., 7001, ethylarnine) while maintaining the temperature a t 35-45'. After slow cooling to room temperature, 1.1 moles of sodium hydroxide solution was added dropwise with stirring over a period of 1 hr. The layers were separated, the aqueous portion was extracted with ether, and the extract was combined with the amine layer. The ethereal solution was dried over anhydrous potassium cai bonate and fractionally distilled. Method C.-The general procedure of method A was used in two steps with isolation of the intermediate secondary amine followed by realkylation to introduce the third group. For the first step 3-5 moles of primary amine was employed with 1.5 moles of potassium carbonate and 1.5 moles of allyl or propargyl bromide. The isolated product (distillation) was further alkylated using 0.5 mole with 0.55 mole of potassium carbonate and 0.5 mole of allyl or propargyl bromide. The yields listed in Table I1 are for the second step only and are corrected for recovered starting material. N-Benzyl-t-butylamine was prepared by a method previously described.6 Sodium-ammonia reductions were carried out by alternate additions of metallic sodium and ammonium chloride to substrates dissolved in liquid ammonia as reported earlier.3 Hydrochloride or oxalate salts were precipitated from ethereal solutions and usually crystallized from mixtures of ethyl acetate and absolute ethanol. Oxalate salts were prepared when the hydrochlorides were found to be very hygroscopic.
(5) B. L. Emling, R. Horvath, A. Saraceno, E. Ellermeyer, L. Haile, and L. Hudao, J . Ow. Chem., 94, 657 (1959).
(6) N. Bortnick, L. Luskin, M. Hurwitz, W. Craig, L. Exner, and J. Mirza, J . Am. Chem. Soc., 78,4039 (1956).
suffered rapid loss of an allyl group when crystallization was attempted from mixed solvents containing ethanol. The products were allylisobutylamine hydrochloride and benzyl-t-butylamine hydrochloride,6 respectively. The latter case is particularly noteworthy since N-dlyl-N-benzyl-t-butylamine hydrochloride has three different groups presumably liable to cleavage as carbonium ions. Experimental Section
JUNE 1966
NOTES
Acknowledgment.--The authors express their sincere thanks to Messers. H. L. Hunter, David Cline, .James Gilliam, and Charles Ashbrook of the Lilly Research Laboratories, Indianapolis, Indiana, for the analytical determinations and to Eli Lilly and Company for financial support.
Quaternary Salts from the Alkylation of Tertiary Amines with t-Propargylic Chlorides' G. F. HENNION AND C. V. DIGIOVANNA~
The Chemical Laboratories, University of Notre Dame, Notre Dame, Indiana 46556 Receiued December Si, 1966
Recently we reported3 that t-propargylic chlorides, RlR*C(Cl)-CrCH, alkylate trimethylamine to produce either propargylic or allenic quaternary salts depending on the size of R1 and R2, allenic products being produced when both are larger than CH3. Further study of this reaction, employing a wide variety of tertiary amines as well as assorted t-propargylic chlorides, has revealed that it may take three courses which are shown i n Scheme I. The particular SCHEME I R'R2C(C1)-CECH
+
'r
R3R4R5N.HC1
+ enyne
R3R4R6N - --+R1RzC(&ReR4R6)TC~CH(Cl-) L_t
R'R2C=C=CH-NR3R4RS(
C1-)
outcome in any one case depends not only on steric factors but also on the basicity and nucleophilicity of the particular tertiary amine eniployed. Quaternary salts, propargylic and/or allenic, appear to be the favored products frorri amines R3R4NCH3having PKb in the range ca. 3.5-7.5. While weaker bases react very slowly, if at all, and stronger ones seemingly favor elimination, other factors certainly are involved. Although it is not possible at this stage to disentangle and assess the solvation, basicity, nucleophilicity, and steric factors which favor quaternary salt formation over elimination, or determine whether the product will be propargylic or allenic, the following observations must be significant. The three reactions pictured above always proceed competitively; in most instances one, sometimes two, reactions are favored. Triethylamine, cyclohexyldirrtethylamine, dodecyldimethylamine, and 3-dimethylamino-3-methyl-1-butynereacted chiefly by elimination and thus produced their own hydrochlorides as the only isolable products. Methyldiethylamine gave viscous oils which could not be crystallized; infrared examination indicated that the oils were mixtures of amine hydrochloride and both quaternary salts. N-methylpyrrolidine gave the propargylic quaternaries by reaction with 3-chloro-3methyl-1-butyne and with 1-ethynylcyclohexyl chloride, but an oil, mostly the hydrochloride, by treat(1) Paper no. 84 on substituted acetylenes. Previous paper: G. F. Hennion and R. H. Ode, J . Ow.Chem., 31, 1976 (1966). (2) Eli Lilly Co. Fellow, 1962-1966. (3) G . F. Hennion and C. V DiGiovanna, J . Org. Chem., 80, 3696 (1965).
1977
ment with 3-chloro-3-ethyl-1-pentyne. N-methylpiperidine gave allenic quaternaries except again with 3-chloro-3-ethyl-1-pentynewhich produced an oily mixture of hydrochloride and both quaternaries. Dimethylpropyl-, dimethylallyl-, and dimethylpropargylamine, differing but little in steric features but significantly in basicity, gave good yields of quaternaries. Dimethylpropargylamine, the least basic of these three, uniquely produced the allenic and propargylic salts in about equal amounts, readily separated by selective extraction and recrystallization (see Experimental section). While no crystalline products were isolated from the treatment of pyridine with t-propargyl chlorides, the reaction of 2,4,6-collidine with 3-chloro-3-methyl1-butyne gave a 35% yield of a brown salt, mp 203204", for which C,H,N analysis and infrared examination indicated a propargylic structure. Propargylic and allenic quaternary ammonium salts [20 new] are described in Table I. It should be noted that all are hygroscopic and melt with decomposition; each one is a methochloride and hence was prepared from an amine CH3NR3R4. Structures were established by infrared and pmr examination as previously reported. The pmr spectrum of 3-(N-propargylniethylamino)-3methyl-1-butyne methochloride (Table I, compound 11) showed a singlet at T 5.4 assigned to methylene protons of the propargyl group, a singlet at T 6.5 for the N-methyl protons, and a singlet at T 7.9 for the Cmethyl protons, The allenic isomer (Table I, compound XIV) showed a multiplet at T 3.4 (allenic proton), a singlet at 7 5.6 assigned to the methylene protons of the propargyl group, a singlet at T 6.7 (Xmethyl protons), and a doublet with J = 2 cps at 78.1 assigned to the remaining methyl protons (both spectra in DzO with water as internal standard). Finally it should be stated that the various reaction products are believed to arise via zwitterion carbenes RIR2C=C=C :) as previously (R'RZ+C-C=Cdiscussed.
-
Experimental Section Infrared spectra were obtained for chloroform solutions; pmr spectra with chloroform-d solutions with TMS as the internal standard or with DzO solutions having water as the internal standard. The t-propargylic chlorides were prepared as previously described.4 Tertiary amines were purchased or prepared by methods given in the literature. Quaternary chlorides were prepared by the procedure described earlier3 except that the reactions were allowed to proceed for 24-72 hr depending on the rate of crystallization of product. I n most instances the product was recovered by filtration; in Lome experiments the solvent (usually acetone) was removed by evaporation or distillation in wucuo. Crystallization usually was achieved using acetone, acetonitrile, ethyl acetate, absolute ethanol, chloroform, or an appropriate binary mixture of these. Reaction of N,N-Dimethylpropargylamine with 3-Chloro-3methyl-1-butyne.-A 0.1-mole scale experiment gave 11 g (60y0 yield) of crude product after 7 2 hr. Approximately half of this material dissolved by boiling with a mixture of acetonitrile and chloroform. Several crystallizations of the material so extracted (same solvent mixture) gave 1-(N-propargylmethylamino)-3-methyl-l,2-butadienemethochloride (XIV), mp 119120". The residue from the original extraction was crystallized from a mixture of ethyl acetate and isopropyl alcohol to give 3-( N-propargylmethylamino)3-methyl-l-butyne methochloride (4) G . F. Hennion and A. P. Boiaselle, ibid., 86, 725 (1961).
1978
NOTES
VOL. 31
TABLE I QUATERNARY AMMONIUM SALTS Mol
Yield, Compd
R1
RZ
R3
%'
R4
formula
MP, "C
-%CCalcd
7 %
7 - 7 0 N--
Found
Calcd
Calcd
61.52 64.97 62.15 63.21 63.76 55.89 56.27 66.86 68.83 68.05 71.83 68.96 62.30
10.32 8.69 9.71 10.63 9.66 9.46 8.63 10.28 9.73 10.53 9.05 9.74 9.57
10.42 8.93 9.66 10.70 9.80 9.70 8.80 10.42 9.83 10.69 8.98 9.98 9.63
7.97 7.54 7.23 7.38 7.46 7.31 5.99 6.49 6.15 6.10 4.66 6.15 6.04
5.82 6.09
63.06 64.76 71.65 72.40 69.02 73.39 69.18
8.75 9.99 8.81 9.10 8.93 8.71 10.01
8.79 10.32 9.09 9.22 8.99 8.82 10.34
7.36 6.94 5.56 5.27 6.20 5.04 5.79
7.16 6.36 5.44 5.18 5.94 4.84 5.52
+
I I1 IIIC IV V VI VI1 VI11 IX X XId XI1 XI11
CHa CHa CHa CHa CHa CHa CHa
XIVC
CHa CHs CHI CHa CHI CiHa CzHs CiHa -(CHz)a-(CHz)s-(CHz)a-
CHa CHs CHa CHa CHa CHa CHa -( CH:!)a-(CH?)a-( CH,!)r -(CH:ds-( CH,!)6-( CH,!)r
Propargylic, RlRK!(NR'RCHa) e C H ((3-) CHa CzHs 53 CpNisNCl 181-183 61.52 CHs CHzCZCH 30' CioHdCl 177-178 64.68 CHa CHzCH=CHz 56 CioHisNCl 142-145 62.00 CHa CHzCHzCHa 84 CiaHzoNC1 199-201 63.30 -(CHZ)467 CioHisNCI 182-183 63.98 CHa CHzCHzOH 82 CpHisNOCl 168-170 66.39 CHa CHnCHzOCOCHa 65 CiIHzoNOzCl 166-167 56.52 CHs CzHa 48 CizHzzNCl 214-215 66.79 CHa CHzCH=CH% 43 CiaHzzNCl 194-195 68.55 CHs CHzCHzCHa 61 CiaHzrNC1 210-211 67.95 CHa CHaCsHs 32' C~HMNCL 175-176 71.85 -(CHz)d57 CiaHzzNC1 197-198 68.55 CHa CHZCH~OH 61 CizHnNOC1 181-182 62.18
H-Found
Found
7.77 7.48 7.21 7.35 7.24 7.19 6.00 5.94 6.03 6.27 4.50
+
xv
XVI XVII XVIII XIX
xx
Yield of crude material. CI?H&'Cl "/zCzHs013. Q
Allenic, R'RzCk=kCH-N(NR'R'CHa) (CI-) CHs CHzCSCH 3ob CinHieNCl 119-120 -(CHz)a90 CiiHzoNCI 151-153 CHI CHzCeHs 58 CiaHzzNC1 146-148 CHa CHzCeHa 52 CisHirNCl 112-114 CHa CHzCSCH 50 CisHmNCl 133-135 CHI CHzCsHa 32' Ci7HaNCl 149-151 -(CHz)s65 CirHzrNCl 144-147
50% of total product.
Analysis for CloHlsNCl.l/,HzO.
(11),mp 177-178'. The pmr spectrum of each product and of the original (rude material indicated that the latter contained the two products in nearly equal amounts. 3-( N-Methyl-p-acetoxyethylamino)-3-methyl-l-butyne methochloride (VII) was prepared by esterification of the hydroxy compound ( V I ) by heating with acetic anhydride (steam bath, 1 hr). The producxt, crystallized from acetonitrile, had mp 166-167" dec and showed infrared bands at 5.7 p (ester carbonyl) and at 3.0 and 4.7 p (ethynyl group).
d
63.14 65.49 71.55 72.29 69.16 73.49 69.54
Analysis for C10HIGNC1.1/4H?0.e ilnalysis for
SCHEME I
R
Acknowledgment.-The authors express their sincere gratitude to Air Reduction Company, New York, Xew York, for generous samples of t-acetylenic carbinols; to Alessrs. W. L. Brown, C. Ashbrook, D. L. Cline, J. Gilliam, and H. L. Hunter of the Lilly Research Laboratories, Indianapolis, Indiana, for the analyticalwork; and to Eli Lilly and Company for financial support.
Characterization of an Intermediate in the Dithionite Reduction of a Diphosphopyridine Nucleotide Model as a 1,4-Addition Product by Nuclear Magnetic Resonance Spectroscopy' WINSLOW S.CAVGHEY~ AND KARLA. SCHELLENBERG~
Department of Physiological Chemistry, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205 Received August 9, 1966
T h e yellow intermediate formed during the reduction of diphosphopyridine nucleotide (DPN) or its analogs was shown by Yarmolinsky and Colowick to decompose readily into the fully reduced compound (DPNH2) and ~ u l f i t e . ~The reduction could thus be represented (1) Supported in part by U. S. Public Health Service Grants HE-06079 and GM-11799. A preliminary report of this a o r k has been presented: W.S. Caughey and K. A . behellenberg, Federafion Proc., 48, 478 (1964). (2) Lederle Medical Faculty Award Scholar. (3) John and Alary R. Xarkle roundation Scholar in Medical Science. (4) M. B. Yarmolinski and S. P. Colowick, Blochim. Biophy8. Acta, 80, 177 (1956).
0
I R as shown in Scheme I. Here R could be adenosyldiphosphoribosyl, benzyl, or methyl. The yellow complex was studied only in solution where it was labile to air and neutralization, being stable only in strongly alkaline solution. Mauzerall and Westheimer5 subsequently demonstrated the absence of paramagnetism associated with the intermediate. The intermediate obtained from DPN exhibits a broad absorption band with Xma, at 357 mp (E 3200).4 This absorption band is broader than that of other addition products of DPN. Unlike the labile alkali, cyanide, or acetone complexes,6 the dithionite complex is not fluore~cent.~Kosower and Bauer suggested that (5) F. H. Westheimer, "The Mechanism of Enzyme Action," W. D. McElroy and B. Glass, Ed., Johns Hopkins Press, Baltimore, Ald.. 1954, p 356. (6) N. 0. Kapkn, Enzymes, 3, 105 (1960).
JUNE1966
NOTES
1979
TABLE I NMRDATAFOR PYRIDINIUM, lJ4-DIHYDR0,AND “YELLOW DITHIONITE INTERMEDIATE” SPECIES OF NI-BENZYLNICOTINAMIDE -Chemical Compd
H 4
N 1-Benzylnicotinamidechlorideb N1-Benzyl-1,4-dihydronicotinamide~ a
Dithionite intermediated (1,4-addition product) Parts per million from tetramethylsilane (see text).
I,
shiftsa
HI
Hs
Spin-spin splittings, opa 485 5,6
8.84 8.13 9.02 8.5 3.15 4.72 5.72 3.3 3.97 4.92 6.19 5.5 In D20. In CDCla. In D20 with 1 M NaOD.
the intermediate was not an addition product but a ?r complex (11) analogous to similar complexes of the pyridinium iodides (III).’ That changes in sub-
stituents on the pyridine ring affected the ultraviolet spectra of the dithionite products and of the iodides similarly was considered by these authors to constitute a basis for their suggestion. The intermediate obtained from dithionite and Nbenzylnicotinamide exhibits a similar broad absorption band although a t somewhat longer wavelength, Amax 372 mp ( E 3200). The benzyldihydropyridine derivative also exhibits an absorption maximum shifted about 15 mp to the red of the Amax for DPNHz (N-benzyldihydronicotinamide, Amax 355 mp;* DPNH2, Amax 340 mp). Earlier Wallenfels and Schuly had prepared and characterized by elemental analysis and ultraviolet spectra the addition products of a DPN model (R = 2,6dichlorobenzyl) wit!h sulfite, sulfide, mercaptobenzothiazole, and benzyl m e r ~ a p t a n . ~These adducts exhibited absorption bands with maxima a t about 350 mp, suggesting 1,4-addition products, and they dissociated into starting materials in dilute solution. Subsequently these authors isolated a crystalline addition product from a reaction with dithionite which they concluded was a 2-s~lfinate.’~ Proton nmr spect,ra of the alkali-stable yellow intermediate that is formed in the course of the dithionite reduction of N1-benzylnicotinamide chloride indicate that the intermediate is a 1,4-addition product as would be the case with a sulfinate or sulfoxylate group attached to the 4 position of the pyridine ring. All proton chemical shifts have been unambiguously identified with the aid of suitably deuterated reactants, and by comparison with the corresponding chemical shift for protons of the analogous pyridinium and 1,4-dihydropyridine compounds. The chemical shifts and spinspin splittings of the pyridine ring protons of the yellow intermediates resemble those of the fully reduced compounds rather than those of the fully oxidized compounds and are only compatible with a 1,4 adduct, and exclude 1,2 or 1,6 adducts, 7r complexes, or equilibrium mixtures of such postulated complexes (Figures 1 and 2 and Table I). (7) E. M. Kosower and S. W. Bauer, J . Am. Chem. Soc., 80, 2191 (1960). (8) D. Mauzerall and F. H. Westheimer, ibid., 17, 2261 (1955). (9) K. Wallenfels and H. SchUly, Angew. Chem., 69, 505 (1957). (10) K. Wallenfttls and €I. Schuly, Ann., 601, 178 (1959).
6.3 7.9 7.8
For the intermediate, Hz and H4 were shifted about 2 and 5 ppm, respectively, to high field compared with oxidized spectrum, whereas these protons were only about 0.1 and 0.8 ppm to low field from their position in the fully reduced species. Complexes of the type represented by I1 could be expected to exhibit nmr spectra which resemble the fully oxidized species far more closely than the fully reduced species” or, if a paramagnetic species resulted, broadening would be expected. Hanna and Ashbaugh have reported nmr spectra for a series of complexes between 7,7,8,8-tetracyanoquinodimethane and methyl-substituted benzenes.ll I n each case on molecular complex formation the acceptor protons were shifted to high field, but only to an extent far less than those shifts observed here in the intermediates produced with dithionite. The high-field shift which could result from charge neutralization in these intermediates (if regarded as charge-transfer complexes) would also be expected to be far too small to account for the shifts observed. On the other hand, the small shifts to low field in the spectrum of the intermediate compared with the fully reduced compound can be readily rationalized in terms of a 1,Paddition product. For example the low-field shift of 0.8 ppm for H4 appears entirely consistent with the replacement of one of the 4 protons with a deshielding sulfinate or sulfoxylate group. Less marked deshielding effects are expected for the other positions as observed. Solutions of the yellow intermediate and sodium hydroxide on cooling gave a crystalline, but unstable, yellow solid in good yield. The solid gave elemental analyses in accord with the presence of the sodium salt of either a sulfoxylate or a sulfinate group in the molecule, Solutions of the salt were yellow and exhibited an ultraviolet spectrum identical with the spectrum of the solution of yellow intermediate used in the nmr studies. A broad absorption band was found a t about 372 mp ( e -3000) for solutions which were prepared either by mixing dithionite and the oxidized species or by dissolving crystalline salt and also for solutions where concentrations of intermediate varied from 0.25 mM to 0.25 M or the concentrations of NaOH varied from 0.05 to 1 M . Furthermore, no intermediate other than the 1,4-addition product was detected in the nmr spectra. Thus the yellow color appears to be due to the 1,4-addition product. The nmr spectra of solutions of dithionite and N1benzylnicotinamide, the elemental analyses on the solid, and the correlation of ultraviolet spectra of solutions studied by nmr and of the solid all support the hypothesis that the yellow intermediate is the 1,4sulfinate or -sulfoxylate and thereby provide strong experimental support for the original suggestion of Yarmolinsky and Colowick that the yellow intermediate (11) M. W. Hanna and A. L. Ashbaugh, J . Phgs. Chem., 68, 811 (1964).
NOTES
1980
VOL. 31
L I
I
BENZYL
I
"DO
I
CHEMICAL SHIFT P.P.M. Figure 1.-Nmr
spectra of 1-benzylnicotinamide derivatives. Upper and lower spectra in deuterium oxide; middle spectrum in deuteriochloroform.
JUNE 1966
NOTES
1981
I
I BENZYL -CH2
HQI H
BENZYL ''6H5
'
BENZYL -CH2
I
H
H
H
Ha ti
H
CHEMICAL SHIFT P.P.M. Figure 2.-Nmr
spectra of Cdeuterated 1-benzylnicotinamide derivatives. Upper and lower spectra in deuterium oxide; middle spectrum in deuteriochloroform
VOL.31
NOTES
1982
formed during the dithionite reduction of DPN was an addition product. Experimental Section 1-Benzylnicotinamide chloride was prepared according to Karrer and Stare.I2 Found in the nmr spectrum in DzO were benzyl CH protons rm a singlet a t 5.81, phenyl protons as a singlet a t 7.41, H5 a.s a pair of doublets (J4,5 = 8.5 cps, J ~ , c = 6.3 cps) at 8.13, 134 as a pair of triplets (J4.5 = 8.5 cps, J4,g and J 2 , 4 = 1.6 cps) ai; 8.84, Hg as a pair of partially resolved trip~ J4,g = 1.6 cps) a t 9.02, HI as a lets (J6,g = 6.5 cps, J z , and broad “singlet” a t 9.27; HDO as a singlet a t 4.65. l-Benzyl-l,4-dihydronicotinamidewaa prepared by reduction of the nicotinamide chloride with dithionite according to Mauzerall and Westheimer.8 Nmr spectra in CDCls agreed closely with the data reported previously.13J4 Found were H4 as a pair of doublets (J4.5 = 3.3 cps, J4,g = 1.7 cps) a t 3.15, benzyl CHZas a singlet a t 4.28, H5 as a pair of triplets ( J ~ , = s 3.3 cps, J6,g = 7.9 cps) a t 4.72, He as a pair of quarters (J6,g = 7.9 cps, J4,6and J2.6 = 1.7 cps) a t 5.72 with amide protons on high-field side, HZas a doublet ( J z ,=~1.7 cps) a t 7.13, phenyl protons as a singlet a t 7.28. l-Benzyl-l,4-dideuterionicotinamideand 1-Benzyl-4-deuterionicotinamide Chloride.-To a solution of 1-benzyl-1,4-dihydronicotinamide (0.76 g, 3.5 mmoles) in dimethylformamide (5 ml) was added chloranil (0.91 g, 3.7 mmoles) in dimethylformamide (20 m1).16 After 10 E:ec while mixing thoroughly, 10 ml of 1 M HC1 was added. Th.e aqueous phase, which contained benzylnicotinamide chloride, was washed three times with ethyl acetate and evaporated to dryness. Crystals were obtained from ethanol in 70-90% yield, mp 229-232’. After three cycles of oxidation with chloranil in dimethylformamide and reduction with dithionite in deuterium oxide, the nmr spectra obtained for both the oxidized and reduced compounds were consistent with the presence of only deuterium a t the 4 position. In the nmr spectrum for the reduced compound in CDC13 were found benzyl C& protons as a singlet a t 4.28, H5 as a doublet (J5.g = 8.0 ops) a t 4.72, Hg as a pair of doublets (J5,o = 7.9 cps, J2.g = 1.’7 cps) a t 5.73, HZas a doublet (Jw, = 1.8 cps) a t 7.13, phenyl protons as a singlet a t 7.28. I n the nmr spectrum for the oxidiized compound in D20 were found benzyl CH, protons as a sin,glet a t 5.83, phenyl protons as a singlet a t 7.42, H5as a doublet (J6.g = 6.2 cps) a t 8.14, He as a pair of doublets (J6.6 = 6.3 cps, J2,g = 1.2 cps) a t 9.03, Hz as a doublet ( J Z ,=~ 1.1 cps) a t 9.28, HDO as a singlet a t 4.65. Dithionite Addition Products.-For the nmr studies the 4 hydrogen and 4-deuterio derivatives were both prepared in the same manner. Sodium dithionite (93% pure,lK 87 mg, 0.5 mmole) was dissolved in 0.5 ml of 2 M NaOD in D 2 0 under nitrogen. The benzylnicotinamide chloride (63 mg, 0.25 mmole) in 0.5 ml of D20 was added dropwise to the alkaline dithionite solution over a period of 1 min and the nmr spectrum was observed. I n the nmr spectrum for the solution from the 4-deuterio derivative were found benzyl CH2 protons as a singlet at 4.35, H5 as a doublet with the low-field peak a t 4.98 (the high-field peak was ~ 1.2 cps, hidden under the HDO peak), Hg as a doublet ( J z , = J6,6 = 7.8 cps) a t 6A9, phenyl protons as a singlet a t 7.23, HZ as a doublet ( J Z ,=~ 1.2 cps) a t 7.30. In the nmr spectrum for the solution from the 4-hydrogen derivative were found Ha as a doublet corresponding to one proton (J4,5 = 5.4 cps) a t 3.97, benzyl CH2 protons as a singlet a t 4.34 cps, H5 as a pair of doublets partially under HDO peak (J4,5 = 5.5 cps, J5,g = 7.5 cps) a t 4.92, Hg as a pair of doublets ( J 2 , 8about 1.0 cps, J6,g = 7.5 cps) at 6.19, phenyl protons as a singlet a t 7.23, HZas a doublet ( J 2 , g = 1.0 cps) a t 7.30. A solution of the 4-hydrogen derivative was prepared in the same manner except for .the use of ordinary water in place of Dmz0. This solution was transferred while minimizing exposure to oxygen via a syringe to an absorption cell with a 0.025-mm path length (Perkin-Elmer ultraviolet short path length cell no. 220-0070). With a 10% transmission neutral density screen in the reference beam, it was possible to (12) P.Karrer and F. J. Stare, H e h . Chin. Acta, 20, 418 (1937). (13) D.C. Dittmer and J. M. Kolyer, J . Org. Chem., 18,2288 (1963). (14) W. L. Meyer, H. :R. Mahler, and R. H. Baker, Jr., Biochim. Biophys. Acta, 64,353 (1962). (15) Facile oxidation of other dihydropyridines by chloranil has been reported: E. A. Hraude, ,J. Hannah, and R. Linstead, J. Chem. Soc., 3257 (1960). (16) W. Christiansen and A. Norton, Ind. Eng. Chem., 14, 1126 (1922).
observe a broad absorption band with E 2800 at 375 mp on a Perkin-Elmer Model 202 spectrophotometer (dithionite absorption in the reference cell amounted to an E of 110 at 375 mp). The spectra of more dilute solutions in cells of longer path lengths gave similar spectra. Thus when the original solution was diluted 1000-fold with either 1 or 0.05 M NaOH, a broad absorp372 mp (E 3000-3200). tion band was observed with ,A, To obtain a sample for elemental analyses a preparation was carried out in nondeuterated solvent and with tenfold larger amounts. The intermediate spontaneously crystallized from such solutions on standing at 3-4’ overnight. The crystals were isolated by centrifugation and pressed dry on unglazed porcelain followed by drying under vacuum a t room temperatue for 12 hr. The yellow crystals turned brown on standing exposed to air and the material underwent decomposition during attempts a t recrystallization. The yield was 53%. Anal. Calcd for ClaHlaNzNaOlaS: C, 51.99; H, 4.36; N, 9.33; Na, 7.66; S, 10.68. Found: C,47.75; H,4.67; N,8.93; Na, 7.54; S, 10.74. The low value for carbon could result from sodium carbonate formation during combustion (as is frequently encountered with sodium salts) or from lack of purity. The absorption spectra of a 0.25 m M solution of the crystalline material in 1 M NaOH exhibited a broad band with kmax 372 mp ( e 3100). The yellow crystals were insoluble in ethanol and freely soluble in water; a yellow oil, possibly the dihydropyridine, separated from the water solution a few minutes after mixing. When heated in a sealed evacuated capillary, the crystals shrank and turned brown a t 114-117’; further heating (to 200”) did not result in melting. Both solubility and melting characteristics were consistent with the crystals being a salt. I n contrast, the dihydropyridines were readily soluble in ethanol and sparingly soluble in water. Nmr Spectra.-All nmr spectra were obtained with a Varian Model A-60 spectrometer. Data in CDCla are reported as parts per million from tetramethylsilane as internal standard (6 values). Chemical shifts for the DzO solutions were estimated by the method of tube replacement; these values are thus only approximate as no internal or external reference was used. More precise chemical shifts were not considered necessary to support the conclusions of this paper.
Dihydropyran Derivatives of Secondary Aromatic Amines LAWRENCE G. VAUGHAN~ AND DAVIDN. KRAMER
Defensive Research Division,
U.8. A r m y Chemical Research and Development Laboratories, Edgewood Arsenal, Maryland
21010
Received January 10, 1966
Although dihydropyran has been used in numerous synthetic sequences as a blocking group for alcohols and acids, recent reviews indicate its use has not been extended to a r n i n e ~ . ~ -The ~ sole work in this area is that of Glacet, who prepared 2-tetrahydropyranylamines by the addition of aniline and N-methylaniline to dihy dropyran. We have now extended the scope of this reaction to five secondary aromatic amines. Since the tetrahydropyranyl group can be easily removed to regenerate the original amine, it offers promise as a base-stable blocking group for these amines. 596
(1) To whom inquiries should be sent: Central Research Department, Experimental Station, E. I. du Pout de Nemours and Co., Wdmington, Del. (2) J. F. W. MoOmie, Aduon. 070. Chem.,8, 191 (1963). (3) H.J. E. Loewenthal, Tetmhedron, 6, 269 (1959)(4) G. A. Swan, ”Technique of Organic Chemistry,” Vol. XI,A. Weisr berger, Ed., Interscience Publishers, Inc., New York, N. Y., 1963, p 457. (5) C. Glacet, Compt. Rend., 284,635 (1952). (6) C. Glacet and D. Veron, {bid., 248, 1347 (1959).
JUNE 1966
NOTES
1983
TABLE I N-( 2-TETRAHYDROPYR.4NYL)aMINES Starting amine
Yield, %
Diphenylamine Phenoxazine Phenyl-a-naphthylamine Phenothiazine 4-Methyl-4'-nitrodiphenylamine Recrystallized from methanol. Calcd: S,11.3. Found: S, 11.4.
MP, 'C
Formula
-Calcd, C
%-H
0
-Found, C
%H
0
60 89-90" Ci7Hi7NO 80.6 7.6 6.3 80.3 7.6 6.4 85 140-141b CnHi7N02 76.4 6.4 12.0 76.3 6.2 12.0 42 115-116" CziHaiNO 83.1 7.0 5.3 83.1 7.2 5.3 72 156-157b Ci7Hi7NOS 72.1 6.1 d 72.1 5.8 d 85 99-100' CislH2oN203 69.2 6.5 f 69.2 6.5 f b Recrystallized from acetonitrile. Recrystallized from petroleum ether (bp 30-60" ). d Anal. e Recrystallized from cyclohexane. f Anal. Calcd: N, 9.0, Found: N, 8.9.
The course of the reaction is illustrated by phenothiazine (l), which reacts with dihydropyran to give 10-(2-tetrahydropyranyl)phenothiaxine (2) in 72% yield. The proposed structure of this derivative is in
1 2
agreement with spectral data. The infrared spectrum showed no N-H stretching band in the 3400-cm.-l region (present in phenothiazine at 3390 cm.-l) but contained two strong bands attributed to C-0 stretching' at 1025 and 1070 cm-l. The ultraviolet spectrum (acetonitrile) of the product, Amax 251 mp (e 31,000) and 298 mp (e 3240), was similar to that of phenothiazine, Amax 253 mp (e 31,900) and 312 mp (e 5640). In the nmr spectrum, complex multiplets occurred at 7.1, 4.9, 4.0, and 1.7 1)pm. The signal at 7.1 ppm is attributed to the eight aromatic protons, that at 4.9 ppm to the methine proton, the signal at 4.0 ppm to the methylene protons adjacent to oxygen, and that at 1.7 ppm to the remaining six methylene protons. The integrated strengths of these peaks were in the expected ratio of 8 : 1:2: 6. All derivatives could be easily decomposed in alcohol to regenerate the starting amine in high yield. No attempt was made to identify the other expected product of this reaction, 2-ethoxytetrahydropyran (3).
Laboratories. Infrared spectra were recorded using a PerkinElmer 521 or a Beckman I R d A infrared spectrophotometer. All compounds were run &s 5% solutions in chloroform. Nmr spectra were recorded using a Varian Associates A-60 nmr spectrometer, and all chemical shifts are given in parts per million downfield from tetramethylsilane. Deuteriochloroform was used as the solvent in all cases. A Beckman DK-2 recording spectrophotometer was used to record ultraviolet spectra. For thin layer chromatography, silica gel G, Merck, was used. Melting points were taken on a Thomas-Hoover melting point apparatus and are uncorrected. 3,4Dihydropyran was distilled and stored over anhydrous potassium carbonate prior to use. All experimental proceduras were similar to that described for phenothiazine, with the exception of phenyl-a-naphthylamine. Tetrahydrofuran was used as the solvent with this amine, since no reaction occurred in diethyl ether. 10-(2-Tetrahydropyranyl)phenothiazine ( 2 j .-To a slurry of 15.0 g (0.0755 mole) of phenothiazine in 20 ml of anhydrous ether was added 13.8 g (0.164 mole) of dihydropyran and 1 drop of concentrated sulfuric acid (sp gr 1.84). The phenothiazine dissolved completely within 30 sec and a mildly exothermic reaction continued for 2 min. After 4 min, the adduct began to precipitate. It was then filtered and washed with 250 ml of 57, sodium carbonate solution. The adduct was dissolved in boiling acetonitrile and recrystallized as the solution cooled to room temperature. A total of 15.1 g (727, yield) was obtained. After two additional recrystallizations from acetonitrile, an analytical sample had mp 156-157". Anal. Calcd for C17H17NOS:C, 72.05; H, 6.05; S, 11.31. Found: C, 72.1; H , 5.8; S, 11.4. Regeneration of Phenothiazine.-A total of 15.1 g (0.0533 mole) of 10-(2-tetrahydropyranyl)phenothiazine was partially dissolved in 50 ml of boiling ethanol. When 1 drop of concentrated sulfuric acid (sp gr 1.84) was added, the solution turned dark brown and the compound dissolved completely. As the solution cooled, yellow crystals of phenothiazine precipitated. A total of 10.6 g (quantitative yield) m'as obtained. After drying, the crystals had mp 183-184.5", lit.8 mp 183-185". A mixture melting point with an authentic sample of phenothiazine showed no depression. (7) L. J. Bellamy, "The Infrared Spectra of Complex Molecules," 2nd ed, John Wiley and Sons,Inc.,New York, N. Y., 1958,p 119. (8) B. E. Baker and L. Briokman, J . A m . Ckem. Soc., 61, 1223 (1945).
2
n
$,Joc.~b 3 Data on derivatives of five secondary aromatic amines are summarized in Table I. The yields are based on single reactions and may be capable of significant improvement. The sole failure encountered to date has been with 2,4-dinitrodiphenyIamineJand is probably due to the combined resonance effects of the two nitro groups and steric hindrance by the ortho nitro group. Experimental Section Methods .-Analyses were performed by the Analytical R e search Branch of the Chemical Research and Development
Substituent Effects in Photochromic Nitrobenzylpyridines JULIUS WEINSTEIN,AARONL. BLUHM,AND JOHN A. SOUSA Pioneering Research Division, U . S . Army Natick Laboratories, Natick, Massachusetts Received December 3, 1966
Previous investigations showed that many derivatives and related compounds4 of 2-(2-nitroben~yl)pyridinel-~ (1) J. A. Sousa and J. Weinstein, J . 070.Ckem., 11, 3155 (1962). (2) A. L. Bluhm, J. Weinstein, and J. A. Sousa, ibid., 98, 1989 (1963). (3) J. Weinstein, J. A. Sousa, and A. L. Bluhrn, ibid., 29, 1586 (1964). (4) A. L. Bluhm, J. A. Sousa, and J. Weinstein, ibid., 99, 636 (1964); J. D. Marperurn et al., J . Phya. Chem., 66, 2434 (1962).
1984
NOTES
VOL. 31 TABLE I VALUESOF k AND t l / , FOR THE FADINQ REACTIONS OF 2-( 2-NITRO-4-SUBSTITUTED BENZYL)PYRIDINES I N ETHANOL AT 25"
4-Subatituent
k sec-1
t1/*,
msec
UP0
NHz
370 X 101 0.187 -0.660 OH 329 2.11 -0.357 H 56.3 12.3 0 c1 27.4 25.3 0.227 con 26.3 26.3 0.132 CONHz 5.94 117 COnEt 2.45 283 0.522 COzMe 2.42 286 0.463 CEN 1.17 592 0.628 NOz 0.122 568 X 10' 0.778 4 Values given by H. H. JaffB, Chem. Rev., 53, 222 (1953), except the value for the C02Me group, which was determined by H. van Bekkum, P. E. Verkade, and B. M. Wepster, Rec. Trav. Chim., 78,815 (1959).
c Oh
\
2.0
\ Q
CIN
ai; Figure 1.--Plot of log k values us. Hammett u p constants.
are photochromic in solution in ethanol and other solvents. Common to all of these aromatic compounds is the structural feature of a nitro and -C-H group ortho to each other.1-5 It is believed that the color change produced by ultraviolet irradiation is due to the formation of the tautorneric aci-nitro structure.
times greater when H rather than NOz is the para substituent. The greater the electron-repelling power of the substituent, the faster the reaction proceeds. I n Figure 1, log k values are shown to be linearly proportional to Hammett u value^.^ The value of the reaction constant ( p ) for the least-square line is -2.85, and the correlation coefficient for the line is 0.986. The correlation is considered excellent on the arbitrary scale proposed by Jaffe.lo A up value was not found in the literature for the CO-NHz group. This work suggests a value of 0.33. The observed substituent effect is probably due to differences in energy of activation, since it was found previously for the dark reactions of the p - C s N and p-NO2 derivatives that the energy of activation differs while the value of the entropy of activation is the same.' Experimental Section
N HO'
?I
Rate measurements of the fading reaction in organic solvents of several derivatives of 2-(2-nitrobenzyl)pyridine were carried out.'-3 The reaction was found to proceed by first-order kinetics with an activation energy in ethanol of about 5 kcal/mole and an entropy of activation of ca. -45 eu. Also, it was found that the rate of the reaction is affected by the nature of the substituent in the para; position of the benzene ring. This observation prompted us to extend the group of compounds previously reported and study the effect of para substituents on the rate of the dark reaction. Rate measurements were carried out in ethanol a t room temperature by flash photolysis techniques. I n every case studied, the fading process was found to follow first-order kinetics. The values of the rate constant, k , for the fading reaction are listed in Table I and can be seen to be markedly sensitive to the electronegativity of the para substituent. For example, in going from NO2 to NH2 the value of k is enhanced by a factor of lo4. Also, the rate constant is almost 500 ( 5 ) G . Wettermark, Nature, 194, 677 (1962). ( 6 ) R. Hardwick and H. S. Mosher, J . Chem. Phya., 86, 1402 (1962). (7) A. Ficalbi, Gazz. Chim. Ital., 99, 1530 (1963). (8) H. Morrison, and B. H. Migdalof. J . Ore. Cham., S O , 3998 (1965).
2-(2-Nitro-4-chlorobenzy1)pyridine.-A magnetically stirred solution of 3.0 g (0.013 mole) of 2-(2-nitro-4aminobenzyl)pyridinel and 3.4 ml (0.039 mole) of concentrated hydrochloric acid in 5.0 ml. of water was cooled to O " , and the amine diazotized by the gradual addition of 0.90 g (0.013 mole) of sodium nitrite in 5.0 ml of water. The diazonium solution was then added to a magnetically stirred, cooled solution of freshly prepared cuprous chloride in 5.7 ml of concentrated hydrochloric acid. The reaction mixture was allowed to warm gradually to room temperature, heated on a steam bath for 1 hr, cooled, and made basic with 5% sodium hydroxide. Solids were collected on a filter. They and the filtrate were extracted several times with benzene. The combined extracts were washed with 3% sodium hydroxide and water, and then dried over anhydrous magnesium sulfate. Evaporation of the benzene gave 1.5 g of a brown viscous liquid. After two distillations at reduced pressure, there was obtained 0.95 g of a light yellow oil. Anal. Calcd for C12H&1N202: C, 57.96; H, 3.65; C1, 14.26; N, 11.27. Found: C, 58.12; H , 3.80; C1, 13.99; N , 11.28. 2-(2-Nitro-4-hydroxybenzyl)pyridine.-A magnetically stirred solution of 3.0 g (0.013 mole) of 2-(2-nitr0-4aminobenzyl)pyridine1 and 2.94 ml (0.052 mole) of concentrated sulfuric acid in 4.0 ml of water was cooled to 0'. Diazotization was carried out at 0-5' by the gradual addition of 0.90 g (0.013 mole) of sodium nitrite in 4.0 ml of water. The resulting red paste was added to a continuously boiling solution of 4.0 ml of concentrated sulfuric acid in 20 ml of water. The solution was then cooled to room temperature; precipitated solids were removed by filtration. To the filtrate 3% sodium hydroxide was added gradually. A sizeable crop of unidentified material separated and was removed by filtration from the still slightly alkaline reaction mixture. On (9) L. P. Hammett, "Physical Organic Chemistry," McGraw-Hill Book Co., Inc., New York, N. Y., 1940, p 184.
(10) See Table I,ref a, p 230.
JUNE 1966
NOTES
continued careful neutralization a small amount of material precipitated and was collected on a filter. Several recrystallizations from ethanol gave 62 mg of pale yellow crystals, mp 139-140'. The infrared spectrum showed the presence of an OH group in the molecule. Anal. Calcd for C12H10N203: C, 62.60; H, 4.38. Found: C,61.91; H, 4.60. Sodium Salt of 2-( 2-Nitxo-4-carboxybenzyl)pyridine.-To 2-(2nitro-4carboxybenzyl)pyridine2 (0.1987 g) in 5 ml of water a stoichiometric amount of s,tandardized sodium hydroxide solution was added. The solution was then clarified by filtration, and the filtrate was evaporated to dryness with a current of nitrogen. The white residue was recrystallized three times from a mixture of ethanol and acetone. An infrared spectrum showed strong absorptions at 1410 and 1615 cm-l characteristic for the carboxylate anion group. (The precursor acid showed the carboxylic acid group at 1712 The preparation of all the other compounds was described previously.lJJ1 Anal. Calcd for ClrHJ\i2NaO~:C, 55.71; HI 3.24; N , 10.00. Found: C, 55.63; HI 3.13; N,9.74. Kinetic Measurements.-Samples for rate measurements were recrystallized or redistilled at reduced pressure before use. 2-(2-Nitro-4-aminobenzyl)pyridinewas purified by chromatographic adsorption on a silica gel (Davison Chemical Co., grade 950) column arid elution with 95% benzene5% ethyl acetate, folloffed by recrystallization from ethanol to give bright yellow crystals, mp 127-128". Spectro Grade absolute alcohol was used t o prepare 10-4 M solutions. The flash photolysis equipment and techniques used for the kinetic studies were described previously.1Ja Measurements were made at 25'. The decrease in absorbance of the visible absorption band produced hy ultraviolet irradiation was followed at 580 mfi with respect to time for all compounds except 2(2-nitrobenzy1)pyridine and 2-(2-nitro-4aminobenzyl)pyridine. For these compounds the dark reaction was followed at 400 and 500 mp, respectively. Values of the first-order rate constant, k , were calculated from the slope of the straight line in plots of log optical density us. time. The reactions were followed for at least three half-lives.
Acknowledgment-We wish to thank Mr. F. Bissett for the preparation of 2-(Znitrobenzy1)pyridine and Mr. C. DiPietro for the chemical analyses. (11) A .I.Nunn and K . Schofield, J . Chem. Soc., 583 (1952) (12) L. Lindqvist, R e v . Sei. Instr., 96, 993 (1964).
Aralkyl Hydrodisulfides. VI. The Reaction of Benzhydryl Hydrosulfide with Several Nucleophiles TAKESHIGE NAKABAYASHI, TEIJIRO KITAO,AND JITSUOTSURUGI
SHUNICHI KSWAMUR.4,
Department of Chemistry, Radiation Center of Osaka Prefecture, Sakai, Osaka, Japan Received September 3, 1966
I n a previous paper,' reactions of benzyl hydrodisulfide with several inorganic anions and benzenethiolate ion were studied. It was reported that hydroxide and sulfite ions, which have weak nucleophilicity toward sulfur, attack initially the sulfenyl sulfur atom of the hydrodisulfide, yielding about 0.5 mole each of hydrogen sulfide and diaralkyl disulfide and various amounts of the other products but no thiol, and also that cyanide and thiolate ions having strong nucleophicity at,tack both sulfenyl and sulfhydryl sulfur. Attack on sulfhydryl sulfur was postulated (1) Part IV: S. Kawamurs., Y. Otsuji, T. Nakabayaehi, T. Kitao, and J. Tsurugi, J . Ore. Chem., SO, 2711 (1965).
1985
to give the thiol besides other sulfur-containing compounds. I n the reaction mechanism, any steric factor of the substrate of the reactants was not taken into consideration. I n the present paper several inorganic anions and benzenethiolate ion were allowed to react with benzhydryl (diphenylmethyl) hydrodisulfide which may suffersteric hindrance as compared with the benzyl compound. The reactions were carried out under conditions similar to previous ones1 and indicated similar features. The products are summarized in Table I. The variety and amounts of the products from the reaction of benzhydryl hydrodisulfide with hydroxide ion are quite similar to those from the benzyl compound. Therefore, the reaction seems to proceed through the same mechanism as reported in previous paper, Le., through sulfenyl sulfur attack. I n the reactions of benzhydryl hydrodisulfide with cyanide and benzenethiolate ions, the distribution of the products was not greatly different from that from the benzyl hydrodisulfide reaction in the previous paper. However, a slight dissimilarity was observed about a little increased amount of the thiol in Table I compared with the corresponding one in the previous paper. This suggests that the attack on sulfenyl sulfur atom was sterically hindered, and that the attack on the alternative sulfhydryl sulfur contributed to the reaction more predominantly than in the case of the benzyl compound. The most prominent distinction between the results in Table I and those in the previous paper is that in the reaction of benzhydryl hyrodisulfide with sulfite ion about 0.1 mole of diphenylmethanethiol was obtained from 1 mole of the hydrodisulfide, while the thiol was reported not to be detected from benzyl hydrodisulfide. The formation of a detectable amount of the thiol may be interpreted by assuming that sulfhydryl sulfur attack contributes to the reaction a t the expense of the sulfenyl sulfur attack by analogy with the above discussion. However, this presumption is denied for the following reason. The distribution of the products with cyanide will be discussed as representative of the mechanism for both sulfenyl and sulfhydryl sulfur attacks. In the case of benzenethiolate the distribution is made complexing because of the participation of benzenethiolate in the products and therefore kept away. The reaction mechanism with cyanide is again cited (Scheme I) from the previous paper.' SCHEME I Attack on sulfenyl sulfur RSSH
+ CN- +RS-SH t
+RSCN + SH-
CN -
RSSH
+ RSCN +RSSR + HSCN
Attack on sulfhydryl sulfur RSSH
+ CN-+ RS-
RS-SH
+ RSSH
t
+RS-
CN ---t
RSSR
+ HSCN
+ SH-
A detailed examination of Scheme I leads to the relationships that the amount of the diaralkyl disulfide is equal to that of hydrogen sulfide, and that the amount of thiocyanate is equal to the sum of the amounts of thiol and hydrogen sulfide (or the disulfide), whatever
1986
NOTES THE PRODUCTS FROM THE
Nucleophile
cabs-
HIS,
Thiol, mole
mole
0. 46OC ((0.310)d 0.366 0.110
CN -
VOL.31
TABLE I REACTION OF (C&)#CHSSH
Free sulfur, g-atom
0.688
Combined sulfur, g-atom
Disulfide? mole
WITH
NUCLEOPHILES (1 :1 IN
SCN-, mole
SzOaI-, mole
MOLES)"
--Material
R
S
balance(CsHdaCH
0.202
0.270 0.185 0.272 0.639 91 100 91 0.390 0.327 0.256 0.140 89 74 62 OH0 0.430 0.407 0.477 86 91 95 Experiments were carried out in a small scale but the amounts of products were recalculated to a 1-mole scale for convenience of discussion. b The value containing the disulfide derived from polysulfide. 0 The sum of benzenethiol and diphenylmethanethiol. d The amount of benzenethiol only.
soa2-
proportion of sulfenyl or sulfhydryl attack contributes to the reaction. The result with cyanide in Table I satisfied the above relationships almost completely. A similar result was obtained also in previous paper' on the product distribution from benzyl hydrodisulfide with cyanide. As to the products with sulfite in the present paper, the amount of thiosulfate is far less than the amount of thiocyanate, and the distribution of the products is different from that with cyanide. The result excludes the mechanism by both sulfenyl and sulfhydryl sulfur attacks. Aloreover, the sulfenyl sulfur attack is eliminated also for the reason that Bunte salt RSS03-Na+ of diphenylmethanethiol could not be prepared in spite of all our efforts, while that of a-toluenethiol was prepared easily by the method in the literature.2 Probably the sulfenyl sulfur attack is prohibited by steric hindrance of both substrate and sulfite ion, the nucleophilic center of which is considered to be located a t the sulfur atom and hence surrounded by oxygen atoms. However, some extent of contribution of sulfhydryl sulfur attack to the reaction cannot be excluded. From the point of view of steric hindrance, triphenylmethyl hydrodisulfide, the sulfenyl sulfur of which is sterically most hindered in the present aralkyl series, was allowed to react with potassium hydroxide. Although all the products, unfortunately, could not be isolated, triphenylmethanethiol alone was detected qualitatively. Therefore, it seems clear that even the least bulky nucleophile such as hydroxide can hardly attack the su1fen:yl sulfur in triphenylmethyl hydrodisulfide. A previous paper3reported that triphenylphosphine attacks trrphenylmethyl hydrodisulfide exclusively on sulfhydryl sulfur atom. The remaining step of predominant reaction with sulfite may be hydrogen abstraction from the sulfhydryl group; this and the probable succeeding steps are indicated in Scheme 11. SCHEME I1 RSSH
+ SOs2- +RSS- + HSOa-
RSS- .f RSSH
H+
+2RSH + 25 or +RSSSR + SH-
Since sulfite ion converts the atomic sulfur formed in the reaction to thiosul.fate, the relationship that the amount of the thiol is equal to that of thiosulfate will hold. Furthermore, since trisulfide produced in the reaction is not desulfurated by sulfite at room temperature, the (2) R. A. Purgotti, Guzz. Chim. I t d , 90,25 (1690). (3) Part 111: J. Tsurugi, T. Nakabayshi, and T. Ishihara, J . Org. Chcm., 80,2707 (1965).
relationship that the amount of trisulfide is equal to that of hydrogen sulfide also will hold. The values indicated in Table I almost satisfy the above two requirements. Therefore, sulfite ion is considered to be forced to abstract the proton attached to the sulfhydryl group and/or attack the sulfhydryl sulfur atom. This leads t o the formation of a certain amount of various anions such as RS-, RSS-, and SH-, which in turn attack hydrodisulfide by various possible ways and make the distribution of products complex. I n view of the several reaction mechanisms with the nucleophiles, a brief summary will be made. As reported in the previous paper, in the absence of steric hindrance the weaker nucleophiles, such as hydroxide and sulfite ions, attack exclusively the sulfenyl sulfur of benzyl hydrodisulfide. I n the presence of steric hindrance, as in benzhydryl hydrodisulfide, only less bulky ions like hydroxide can react on the sulfenyl sulfur. More bulky ions, such as sulfite, cannot, however, react with the sulfenyl sulfur of benzhydryl hydrodisulfide but may induce the decomposition to yield a complicated distribution of products. Stronger nucleophiles, such as cyanide and thiolate ions, attack both sulfenyl and sulfhydryl sulfur atoms, but sulfhydryl sulfur attack in benzhydryl hydrodisulfide is more predominant than in the benzyl compound. Experimental Section Materials.-Benzhydryl' and triphenylmethyl6 hydrodisulfides were prepared by the methods reported elsewhere. The Reaction of Benzhydryl Hydrodisulfide with Potassium Hydroxide, Cyanide, and Benzenethio1ate.-Benzhydryl hydrodisulfide, 3-4.5 g, in dioxane was allowed to react with each of the above nucleophiles in a nitrogen stream a t room temperature. Crystals appeared in the solution after 2 or 3 hr. The reaction mixture was extracted with benzene. The benzene extract was examined and reaction products were identified by the previous method.' Other experimental conditions and procedures were similar to those of previous paper. Reaction of Benzhydryl Hydrodisulfide with Potassium Sulfite. -Benzhydryl hydrodisulfide, 3 g (0.0129mole) in 20 ml of dioxane was allowed to react with an equimolar amount of potassium sulfite dihydrate in 10 ml of water. The reaction mixture became brown, turned to yellow, and faded after a few minutes. After being kept overnight, it was neutralized (pH 6). Most of hydrogen sulfide evolved before the neutralization. Organic products were extracted with benzene after a large amount of water was added. The amount of thiosulfate ion in the water layer was determined by a previously described method.' For the determination of the amount of diphenylmethanethiol formed, see the following paragraph. An aliquot was used to determine the amount of polysulfidic sulfur by triphenylphosphine.' The remaining benzene extract was evaporated under reduced pressure and a white solid was obtained. This (4) J. Tsurugi and T. Naksbayashi, ibid., 14, 807 (1959).
(5) T. Nakabayashi, J. Tsurugi, and T. Yabuta, ibid., 19, 1236 (1964).
NOTES
JUNE 1966 solid showed mp 60-120' after desulfurization with sodium sulfite and was fractionated by treatment with n-hexane into two parts: a very viscous oil and white crystals. The latter was dibenzhydryl disulfide, mp 150-151 O ; the mixture melting point with an authentic sample was undepressed. The former, which could not be identified, was excluded from the material balance in Table I. Other procedures were the same as those given in a previous paper.' Identification and Estimation of the Reaction Products.-In the case of the reaction of benzhydryl hydrodisulfide with potassium hydroxide, sulfur was separated as follows. An appropriate amount of Idimethylformamide was added to the solid which was obtained from the benzene extract by removal of solvent. The solution immediately became red, and then turned to yellow. The yelloR crystals precipitated, were filtered, collected, and recrystallized from benzene: mp 118-119', mmp 11&119° with authentic sulfur. Diphenylmethanethiol was titrated with 0.1 h' sodium thiosulfate solution after the addition of an excess of 0.1 N alcoholic iodine solution. Quantitative analyses of other products-hydrogen sulfide, polysulfidic sulfur, inorganic ions, et?.-were carried out as described in a previous paper' and are listed in Table I. Reaction of Triphenylinethyl Hydrodisulfide with Potassium Hydroxide.-Triphenylmethyl hydrodisulfide, 1.99 g (0.0065 mole), in 25 ml of dioxane was allowed to react with an equimolar amount of potassium hydroxide in 10 ml of water under a nitrogen stream at room temperature. The solution became immediately red and then orange. Hydrogen sulfide, 0.0016 mole (0.249 mole/mole), was evolved when it was neutralized with 4 N hydrochloric acid after being kept overnight. In the reaction mixture a yellow precipitate was formed, which sintered a t 155' and melted a t 160' (unidentified with any known substances). Then organic material was extracted with benzene. The existence of triphenylmethanethiol in the benzene layer was recognized qualitatively by its characteristic odor and by using rubeanic acid as a detecting reagent for thiol. Other reagents, such as lead acetate, nitrosyl chloride, and sodium nitroprusside, for detection of thiols did not react with authentic triphenylmethanethiol.
Carbon-14 Isotope Effects in the Beckmarin Rearrangementla I. T. GLOVER^^
AND
V. F. RAAEN
Chemistry Division of Oak Ridge National Laboratory, Oak Ridge, Tennessee Received November 19, 1966
The transformation of ketoximes to amides (R2C= NOH + RCONHR) is catalyzed by such acidic reagents as phosphorus pentachloride, phosphorus pentoxide, and sulfuric acid. It is well established that the migrating group in the Beclrmann rearrangement approaches the nitrogen from the side opposite that of the departing oxygen atom. It is not possible to say whether the cleavage of the N-0 bond and the shift of the alkyl (or aryl) group are con.certed. Yukawa and Kawakami2 have reported relatively large reverse carbon-14 isotope effects in the rearrangement of methylene-labeled phenyl-2-propanone oxime and phenyl-1-labeled aceto= 1.052 and 1.121, respectively). phenone oxime These results were thought to favor a mechanism in which the migrating group participates in the fission of the N-0 bond. Small secondary reverse isotope effects of carbon-14 have been r e p ~ r t e d , ~ but , ~ heretofore no primary (1) (a) Research was sponsored by the U. 9. Atomic Energy Commission under contract with the Union Carbide Corp.; (b) U. 9. Atomic Energy Commission Postdoctoral Fellow under appointment from the Oak Ridge Institute of Nuclear Studies. (2) Y. Yukawa and M . Kuwakami, Chem. Ind. (London), 1401 (1961).
1987
reverse isotope effect of carbon-14 has been substantiated. We have repeated the work of Yukawa and Kawakami; the pertinent data are given in Table I. Our observations do not substantiate their findings. It is significant that considerable hydrolysis of the oximes to the parent ketones occurs during the reaction, Acetophenone oxime was treated with concentrated sulfuric acid at 62 i 1' for 40 min. The nmr spectrum of the chloroform extract of the acetophenone oxime reaction mixture shows methyl resonances characteristic of acetophenone a well as of the oxime and acetanilide (see Table 11). For this reason the determination of acetanilide was considered to be of little value. Hydrolysis of the phenyl-2-propanone oxime is likewise evident from the nmr spectrum of the chloroform extract of the reaction mixture. Furthermore, no resonances characteristic of nT-benzyl acetamide appear, and all attempts to isolate the amide failed. Since the oxime is partly destroyed by hydrolysis, little significance can be attached to the small normal isotope effect observed for both labeled species of acetophenone oxime. Within the limits of experimental error (
