2434
VOL.28
Notes Inhibitors i n the Palladium-Catalyzed Hydrogenation of Aryl Nitro Groups HAROLD GREENFIELD Naugatuck Chemical Division of U . S . Rubber Company, IVaugatuck, Connecticut Received March
4, 1963
I n liquid-phase hydrogenations of aromatic nitro compounds to aromatic amines, the slurry catalyst of choice usually is powdered palladium on carbon. This is so both for preparative laboratory organic syntheses and for industrial processes, such as the hydrogenation of dinitrotoluene to toluenediamine. Thus, the inhibitory effects of possible impurities in such reductions are of practical importance as well as of academic interest. There is very little literature on the inhibition of palladium-catalyzed hydrogenations of nitro groups. Extrapolation of results from the many excellent published studies of poisoning is difficult since inhibitory effects depend on the nature of the substrate being hydrogenated. For example, sodium hydroxide and sodium methoxide do not inhibit the reduction of nitrobenzene to aniline under conditions where reduction of a less strongly adsorbed substrate, such as an olefin, is severely inhibited. A number of inorganic anions, inorganic cations, and organic compounds were tested as inhibitors in the palladium-catalyzed hydrogenation of p-nitrotoluene to p-toluidine. The study was made under mild conditions of temperature and pressure, and the potential inhibitors were a t a level of about 5 mole yo based on p-nitrotoluene, much higher than trace contamination. Thus, lack of inhibition indicates that the contaminant is quite innocuous, and slight inhibition suggests that the contaminant might not be troublesome at a lower concentration and /or under more severe reaction conditions. S o inhibition was shown by sodium nitrate, acetate, sulfate, carbonate, phosphate, hydroxide, fluoride, bromide, and chloride. Increasing the amount of sodium chloride to 115.4 mole To based on p-nitrotoluene did not change the reaction rate. Inhibition was caused by both sodium and potassium nitrites, with an approximately threefold increase in reaction time. This is of interest because of possible contamination with nitrite ions in some products from nitrosation and nitration reactions. Sodium sulfite caused complete poisoning after about one-third of the reaction had been slowly completed. Complete poisoning was produced by sodium iodide, cyanide, sulfide, and bisulfite. Decreasing the amount of sodium iodide to 2..>6 mole To based on p-nitrotoluene still resulted in complete poisoning. S o inhibition was shown by ferrous chloride. The reaction time was increased about 1..>-fold by nickel(TI) nitrate, and about twofold by ferric chloride, fer(1) It Baltzlv J
4.m Chem S o c
7 4 , 4586 (1952)
ric nitrate, cobalt(I1) nitrate, and chromium(II1) nitrate. More severe inhibition was caused by zinc nitrate, with an approximately tenfold increase in halflife. Complete poisoning was produced by copper(1) chloride, copper(I1) chloride, copper(I1) nitrate, silver nitrate, aluminum nitrate, and lead nitrate. Poisoning by copper compounds is of practical importance because their use in many preparative reactions sometimes results in the contamination of aromatic nitro compounds. For example, copper compounds are used in the replacement of the halogen of aryl halides by nucleophilic reagents as aqueous alkali, ammonia, amines, and cyanides; in replacement of the diazonium group of aryl diazonium salts by a variety of reagents; and in the synthesis of diaryls by condensation of aryl halides or diazonium salts. KO inhibition was caused by n-octyl chloride, chlorobenzene, bromobenzene, aniline, piperidine, sodium methoxide, phenol and p-nitrophenol, using 5.13 mole yo based on p-nitrotoluene. Titrophenols have been reported to be severe poisons in the palladium-catalyzed hydrogenation of dinitrotoluenes in the absence of solvents or diluents.2 There was no indication of such behavior under our experimental conditions. The reaction time was increased about twofold by p-nitroso-K-methylaniline and about sixfold by p-nitrosodiphenylamine. The addition of S-nitrosodiphenylamine produced an induction period of about one hour followed by hydrogenation a t about one-third the rate of the uninhibited reaction. The inhibition caused by both C-nitroso- and S-nitrosoarylamines is noteworthy because of the presence of nitroso groups in certain aromatic nitro compounds, either as part of the nitro compound or a contaminant. The effect of phenyl disulfide was investigated at very low concentrations. KOinhibition was caused by 0.0026 mole yo based on p-nitrotoluene, and an approximately fivefold increase in reaction time was produced by 0.00,51 mole Yo based on p-nitrotoluene. Thus, phenyl disulfide produces considerable inhibition a t a level of about 50 moles of inhibitor per million moles of nitro compound. Experimental Each experiment was run in a Paar shaker type pressure reaction apparatus (Paar Instrument Co., series 3910) a t room temperature and a hydrogen pressure of about 1 a t m . A 50-ml. charge was used, containing 2.68 g. (19.5 mmoles) of p-nitrotoluene in a mixture of 2-propanol and water ( t : 1 volume ratio) and 0.05 g. of a 5Vc palladium-on-carbon catalyst containing 50% water, z.e., 0.025 g . of jr& palladium on carbon (Engelhard Industries, I n c . ) . The concentration of p-nitrotoluene was 5 3 . 5 g./l. or 0.39 mole/l. T h e concentration of 5% palladium on carbon was 0 . 5 g./l., 0.93 wt. c 5 based on p-nitrotoluene, or 1.3 g. /mole p-nitrotoluene. One millimole of the compound to be tested as a n inhibitor (0.02 mole/l., 5.13 mole % based on p-nitrotoluene) usually was added. Reayentgrade chemicals were used throughout. (2) I.. D. Winstrom and R D. Samdahl.
U.S.
Patent 2,976,320 (March
21. 1961); British Patent 832,153 (April 6 , 1960).
SEPTEMBER, 1963 For the most part, reaction rates were compared on the basis of half-life, the time required for completion of approximately one-half of the reaction as determined by hydrogen absorption. The total gas absorption was about 5 p s i . , and the half-life for an uninhibited reduction of p-nitrotoluene to p-toluidine was about 9 i. 1 min., with the rate roughly zero order with respect to p-nitrotoluene. Since there was no attempt to control temperature, the results are semiquantitative. Severtheless, gross diff erences in rate were reproducible and meaningful.
Controlled Potential Electrolytic Oxidation of Anthracene in Acetonitrile'
SOTES
2435
recovered by adding water). The electrolysis product was shown to be bianthrone (I) (lO,lO'-bianthroiiyl). Based on the observed formation of the dipyridiiiium perchlorate salt and on the amount of electricity consumed, Lund stated that the electrolytic oxidation of anthracene resulted in the loss of two a-electrons.6 The production of bianthrone, however, would indicate that the anode reaction occurring in the present system involved the loss of only one a-electron from each anthracene molecule to form the free radical ion species (11) which then dimerized. The additional oxidation of any 9,9'-bianthryl formed was in fact predicted by
KARLE. FRIEXD A N D W. E. OHXESORGE
0
Department of Chemistry, University of Rhode Island, Kingston, Rhode Island Receiced January 25, 1963
H Controlled potential electrolysis has been established I1 as a useful technique in synt'hetic organic chemistry, 0 especially when t,he desired product is itself easily I of this approach oxidized or r e d u ~ e d . ~ aThe ~ b ~efficacy ~ The following considerations would account for the to synt,hesis problems is well illustrated by the work of observed peaks in the mass spectra and for the failure Lingane, Swain, and Fields who successfully reduced to observe a peak a t m ' e 386. Cleavage of the labile 9-(o-iodophenyl)acridine to the dihydro derivative.4 10-10' bond in bianthrone would likely occur upoii Theory and techniques of controlled potential electrolysis have been adequately discussed by L i ~ ~ g a n e . ~electron impact and could produce the benzylically stabilized 10-anthrone radical and the highly stahle Most controlled potential electrolytic syntheses have aromatic fragment (111) with m e 193. The pcak at utilized direct electrolysis of the starting material a t the m / e 165 would appear upon loss of CO (twenty-eight working electrode to produce the desired product, mass units) from 111 in a nianner analogous to that alt'hough so-called substitution techniques also have reported for anthraquinone.* been used.',2 Recent voltammetric studies of aromatic hydrocarbons is acetonitrile solution have suggested that the anodic waves of these compounds are due to the removal of a-electrons. The ion radicals formed are very reactive species and in the case of naphthalene polymerize rapidly to coat the anode and cause a I11 rapid decrease of the electrolysis current.6 Lund also showed that pyridine in a n electrolysis solut,ion of ant'hracene in acetonitrile reacted with the 9,10dihydroanthracinum radical ion to form the dipyridiiiium salt. The present work was undertaken in an Experimental attempt to extend the technique to the use of other nucleophiles which might produce useful derivatives Reagents a n d Solutions.---I.:astnlan (Practical) acetonitrile was distilled from phosphorus perltoside and used to prepare :dl of anthracene. Ethanol was selected with the expecsolutions. Cliemic:tlly pure sodiurll perchlorate ( h i e n d Drug tation that the 9,lO-diethyl ether derivative of anthraand Chemical C o . ) , dried a t 150" for 21 h r . and stored over cene would be obtaind. The electrolytic addition of phosphorus pentoside in a vacuum desicc:itor, was used to ethoxy groups to furan using the constant current prepare the supporting electrolyte. .inyhydrous silver perchlorate was used on special order from Fisher Scientific C o . Other technique has been report'ed.' chemicals were reagent grade quality. Electrolysis of a saturated acetonitrile solution of Apparatus.--=\ Sargent 3Iodel XI. polwograph was used to anthracene which contained supporting electrolyte and record current-voltage curve?. A modifiedg Iielley, JoneP, arid was 0.45 )I/ in absolute ethanol produced a white comFisher controlled potenti:al coulometric titrator10 was used as a potentiostat for all elertrol\-eea and also to deterniine the esact pound which was partially soluble (the remainder was (1) Abstracted f r o m t h e thesis submitted b y K. E . Friend f o r t h e M.S. degree in chemistry, University of Rhode Island, .July, 1962.
( 2 ) (a) 31. ,J. Allen. "Organic Electrode Processes," Reinhold Publishing Co.. New York. N. Y . , 1938: (h) A . P. Toniilov, l ' s p . K h i m . , 1462 (1901): Rum. Chem. Rea., G:39 il9F1). ( 3 ) F . D. Popp, H. P. Pcliiiltz, Chem. Reo., 62, 19 (1962). (4) .J. .J. Linaane, C . G . Swain. and 11,Fields, J . A m . Chem. S o c . , 6 6 , 1348 (1943). (3) .J. .I. Linaane. "Electroan;ilyticnl Chemistry." 2nd E d . , Interscience Publishing Co.. New York. S . Y . , 19.58. ( 0 ) H. I,nnd. . 4 d ~Chem. .Ccand., 11, 1323 (1957). ( 7 ) N. Claiison-riaas. ibid., 6 . 569 (19.52).
working electrode potenti:als for the electrol>-ser :LS described later. Utrnviolet ahsorption spectr:t were recorded on a Beckman AIodel DR-2 spectrophotometer, infrared spectrlt on a Baird-.itonlie LIodel R l I - I spectropliotonleter, :ind 1n:iss spectra on a Eendiv Time of Flight and on :i Conso1id:tted lIodel 21-103C mass spectrometer. (8) J. H. Beynon, ">lass Spectrometry and I t s -1pplications to Organlc Chemistry," Van Nostrand Publishing Co.. Princeton, N. .J., 19fi0, p p . 2 7 1 , 272. (9) W.,J. C u r r e n , 31,s. thesis, Unix-ersity of Rhorle Island. 1900. ( I O ) 31. T . Iielley, H. C . Jones. and D. .J. Fisher. d n d . Chem., 31, 488 (19.59).
2436
NOTES
A platinum electrode was the anode; for voltammetric studies a wire with a n exposed surface area of about 0.15 cm.2 was used and a foil electrode (surface area about 6 cm.2) was used for electrolyses. A silver wire coiled in a helix of ca. 1.3-cm. diameter and immersed in a solution of silver perchlorate and sodium perchlorate in acetonitrile served as the reference electrode and as the cathode. I t was necessary to isolate the silver electrode from the anthracene solution during electrolysis b y means of a cylinder fitted with a fine porosity fritted disk to avoid reduction of the electrolysis products to the starting material. All solutions were deaerated for a t least 30 min. by bubbling with nitrogen which had been passed through vanadous chloride solution, concentrated sulfuric acid, and finally acetonitrile. .In atmosphere of nitrogen was maintained above the solution during electrolysis. Current-Voltage Curves.--An approximate current-voltage curve (uncorrected for the iR-drop in the cell) a-as recorded on each electrolysis solution over the voltage range: 0 to f 3 v. (m. .L\g/-Ig-). The exact position of the anodic wave w s then determined point b y point using the potentiostat by setting the platinum electrode potential manually and observing the current. Electrolyses were performed with the anode potential set to a value about 0.1 v . more positive than the halfwave potential determined from the latter current-voltage curve, which was corrected for t h e iR-drop in the cell. The electrolysis was allowed to proceed until the current decreased t o a constant value. lIagnet,ic stirring continued throughout the electrolysis. -In :xetonitrile solution 0.5 AYin sodium perchlorate and 0.1 Y in silver perchlorate and a solution of ethanol in acetonitrile plus supporting electrolyte showed no appreciable current a t potentials less anodic than +1.2 v . (es. Ag/Ag+). *Inthracene in this solvent is oxidized a t potentials more anodic than about +0.88 v . (us. Ag/Ag+). Properties of the Electrolysis Product .-The compound was not appreciably soluble in absolute ethanol, ether, water, benzene, or chloroform, hut could be recrystallized from acetone. On heating it turned brown slowly near 250' and melted to a dark brown liquid a t about 275". .Iddition of concentrated sulfuric acid to tlie product caused the formation of a green color; upon standing for less than 1 min. the color became a ruby red. S o color change was observed upon heating a solution of the electrolysis product in benzonitrile, anisole, or xylene. The results of direct elemental analyses for C, H, and 0 were 87.24, 4.80, and 8.22' L,, respectively, and are in excellent agreee calculated for bianthrone: C, 87.02; H, 4.70; . Duplicate molecular weight determinations lenzene) of 319 and 329 were reported; the low results are attrihuted t o experimental error. The ultraviolet absorption spectrum of the electrolysis product in ahsolute ethanol solution showed a peak a t 267 mp (log E 4.:3W, a shoulder a t 300 mp (log E 3.9), and a minimurn a t 242 mp (log e 4.07). T h e infrared absorption spectrum (potassium broniide pellet) showed strong peaks a t 1320 c m . 2 , 1660 cm.? (conjugated carbonyl), and peaks of moderate intensity a t 1580 cm-1, 600 c m - 1 , $130 cm.-l, and 790 c m . 3 . l I a s s spectra were obtained by introducing the sample directly into the electron beam of the mass spectrometer" and showed, petiks a t w e 193 and 165; a t the sample pressures used no peaks were evident a t higher mle v a h e s . Chemical Synthesis of Bianthrone.-Bianthrone was prepared cllemically by the x t i o n of concentrated nitric acid on a suspension of mthracene in glacial acetic acid.l2 T h e melting point and infrared and ultraviolet spectra of this preparation were identic:rl to those of the electrol product. T h e reactivities with the reagents previously mentioned were also the same for the two preparations. T h e mass spectrum of this compound when injected directly into the electron beam was identical with that ot)t:iined from the electrolysis product.
Acknowledgment.--The
authors are grateful to Mr.
&\. I,. Burliiigame and to M r . James A. McCloskey. Department of Chemistry, Massachusetts Institute of Tcchiiology, for their interest and for their assistance
in obtaining and interpreting the mass spectral data. (11) l i , Riemann, " l l a s s Sigertrometry," I I c G r a w H i l I Book Co.. Inc. S e i \ - Y o r k , S . 8 . . 19(i2, pp, 33, 34. (12) F;. .Iosephg a n d I'. Ratit. "Elsevier's Encyclopedia of Organic Chemist r y . " Val, 1:I. Elsevier Publishing C o . . S e w York, N . Y.,1946, p. 766.
YOL. 28
Beckmann Rearrangement of Arylglyoxylonitrile Oxime p-Toluenesulfonatesl TRAVIS E. STEVEKS Rohm and Haas Company, Redstone Arsenal Research Division, Huntsville, Alabama Received February
4, 1963
Oxime esters readily undergo the Beckmann reand references to literature on this subject are available in recent papers.4,5 A brief study of the rearrangement of the p-toluenesulfonates (I) of five pheiiylglyoxylonitrile oximes is reported here. Characterization of these oxime tosylates, prepared from the sodium salt of the corresponding phenylglyoxylonitrile oxime6 and tosyl chloride, is summarized in Table I.' TABLE I OXIMETOSYLATES ( I ) Ar-C-CS
-
--
Anal.-----
I _ 7 _
I. A r Phenyl p-Chlorophenyl p-Nltrophenyl p-Methoxyphenyl Mesityl
M.P., Yield, OC. 70 135 154 146 141 152
64
66 34
55 86
--
--
59.98 53.81 52.17
4.03 3.31 3.21
9 33 8.37 12.27
60.12 53.74 52.11
4.09 3.40 3.11
9 01 8.18 11 98
58.18 63.14
4.27 5 30
8 48 8 18
58.07 62.99
4 43 5 47
8.32 8 14
X
C
Found H
---
C
Calcd. H
N
The a-cyano tosyl esters (I) appeared to be much more stable than ordinary oxime tosylates.2s4 They did not undergo solvolysis in refluxing ethanol, and could be recovered unchanged either after a three-hour period of refluxing in benzene containing an equiralrnt of aniline or after a pyridine solution mas warmed on the steam bath for two hours.8 When treated with excess sodium ethoxide in ethanol or with ethanolic potassium hydroxide, the first four tosylates listed in Table I were converted to diethyl (1) This research was supported b y t h e Advanced Research Projects gency under Army Ordnance contract no. DA-01-021-ORD-11909. (2) P. Oxley a n d W. F. Short, J . Chem. Soc.. 1,514 (1948); TV. Z,Heldt, J . Am. Chem. Soc., 80, 5880, 5972 (19.58). (3) J. C. Craig a n d A . R. S a i k , z b i d . , 84, 3410 (1902). (1) W. Z. Heldt, J . O r g . Chem., 26, 109.7 (1901). ( 5 ) J . P. Freeman, zbid.. 26, 3507 ( I Y f i l ) . (6) X I . R. Zimmerniann. J . prakt. Chem.. (2) 66, 3.53 (19021. (7) .ilthough both syn a n d anti forms of all the oximes apparently exist.6 isomeric tosylates were isolated only in the mesityl case (see Experimental). T h e anti configuration of the oxime tosylates indicated in Table I a n d in the discussion following cannot be considered to have been established; a s y n configuration also could explain the results obtained here. Since t h e type of migration observed here corresponds t o t h a t usually associated with a n anti configuration. these oximes tosylates are n-ritten in the anta configuration. (8) With I , h r = p-chlorophenyl, some rearrangement probably occurred during chromatography o n a n alumina colu1nn3 for only 75% recovery of tosylate was possible. characterization of a n y rearrangement product \vas not attenipted. Rearrangement of phenylglyoxylonitrile oxime to Kphenyloxaniide b y phosphorus pentachloride in ether. followed b y hydrolysis. has been reported.6 b u t t h e p-nitro-, p-chloro-, a n d o - c h l o r o l ~ l ~ ~ n y l g l y o x y l o nitrile oximes n e r e reported to be unaffected by this treatnwnt. l y e have observed t h a t both t h e phenyl and p-chlorophenyl ouiines a n d Iihosr~ilorlls pentachloride in ether o r nrethylene chloride yield not only some oxamide. b u t also significant amounts (15-30Ct) of the ol-cliloroirtiinoarylacrtonitril~, T h e results of this s t u d y ivill be reported later
SEPTEMBER, 1963
NOTES OEt
N-OTS
Ar-
-CY
OEte
--+
/
Ar-N=C
ETOH
1
I
+ ArCN
\
I1
OEt
0
I11
-1-arylimidocarbonates II.9 A small amount of aryl cyanide also was f0rmed.I To determine the amount of nitrile and imidocarbonate produced in these ethoxide-induced rearrangements, the reaction mixtures were chromatographed on silica gel. It has been shown that diethyl K-arylimidocarbonates are converted cleanly to N-arylurethans (111) during passage through a silica gel column.'O The results of those experiments are summarized in Table 11. TABLE I1 BECKMANN REARRANGEMENT OF POXIMINO TOSYLATES (I) -Yield, Nitrile
I, Ar
%"---Urethan
.. 62 Phenyl, run lb run 2c 8.3 60 p-Chlorophenyl, run 1 5.6 64.0 run 2 4.5 63 p-Methoxyphenyl, run 1 6.4 76.7 run 2 6.8 64 p-Nitrophenyl, run 1 8.3 24.9 Mesityl, run ld Trace Yone a Yield of product isolated from t h e chromatographic column. Run 1 was conducted with about 2.5 equivalents of potassium hydroxide in ethanol at about 60". R u n 2 refers t o experiments conducted with about 2.3 equivalents of sodium ethoxide in ethanol a t reflux. Cleavage of t h e tosylate ester and recovery of oxime accounted for about 50% of t h e reaction products (see Experimental).
The stability of these tosylates indicates that nucleophilic attack of ethoxide ion on the carbon atom of the C=X function, rather than a prior f r a g m e n t a t i ~ n , ~ initiates these rearrangements. Although it appears that the rearrangement proceeds either through a concerted process such as that illustrated by A or
nwt " Y Y
B. Y=CN or OEt
A
,OEt
Ar-N=C
\
Y
-
(9) Rearrangement of t h e tosylate I , A r p-nitrophenyl, gave a much more complex product mixture t h a n indicated here, probably due t o condensations involving the nitro group, However, Some of t h e expected products were obtained (Table 11). ( 1 0 ) T . E. Stevens, J . O r g . Chem., 26, 3 4 5 1 (19131). Characterization of several diethyl imidocarbonates a n d their conversion to urethans are reported here. T h e infrared spectra of these imidocarbonates have strong C=N absorption a t 5.9-6.0 p , a n d their presence in t h e reaction mixtures converted directly t o urethans was confirmed b y infrared spectral examination of the crude products.
2437
through an intermediate with an electron deficient nitrogen (B),7811no definitive mechanistic choice can be made at this time. The timing of the cyanide ion displacement is unknown also. The small amount of nitrile formed may arise from nucleophilic attack at the nitrile function of the cyano tosylatel*; thus it was of interest to see if the rearrangement of the mesitylglyoxylonitrile oxime tosylate, in which attack a t the oxime C=K function mas hindered, would lead to mesitonitrile. However, cleavage of the tosylate with regeneration of the oxime appeared to be the main reaction path. This would indicate that attack at the cyano function is not a favorable process with these tosylates, and might be considered additional evidence tor their anti c~nfiguration.~ Experimental13 T h e procedure used for t h e preparation of the first four cyanooxime tosylates listed in Table I is illustrated by the following description. Preparation of Phenylglyoxylonitrile Oxime p-Toluene Sulfonate.--A solution of sodium ethoxide was prepared by dissolving 2.3 g. (0.10 g.-atom) of sodium in 50 ml. of absolute ethanol. This solution was cooled in an ice bath while 11.7 g. (0.10 mole) of phenylacetonitrile and 10.3 g. (0.10 mole) of butyl nitrite was added dropwise. T h e solution then was stirred at ambient temperature for 45 min. T h e volume of ethanol was concentrated to 30 rnl. a t reduced pressure, 30 ml. of ether was added, and the sodium salt of phenylglyoxyloilitrile oxime was removed by filtration (9.0 g . ) . This wss suspended in 100 rnl. of benzene and 10.2 g . of p-toluenesulfonyl chloride was added portionwise. T h e mixture was refluxed for 2 hr., cooled, and filtered. T h e filter cake was washed with benzene. T h e filtrate was washed with water, dried over magnesium sulfate, and evaporated a t reduced pressure. T h e residue 13.9 g., was recrystallized from 95% ethanol. Thus, the tosylate of phenylglyoxylonitrile oxime, 10.2 g. (64GjC based on sodium salt), m.p. i34-135", was obtained. Mesitylglyoxylonitrile Oxime.-The reaction between butyl nitrite and mesitylacetonitrile was conducted as described before except t h a t the ethanol solution was refluxed for 2 hr. T h e oxime was isolated after chromatography on silica gel and recrystallized from chloroform-hexane, m.p. 113--114'. Anal. Calcd. for CllH12N20:C, 70.19; H, 6.43; N , 14.89. Found: C, 70.14; H, 6.83; S , 14.95. I n subsequent runs, the oxime isolated had a n1.p. of 70-77" and was apparently (infrared and analysis) a mixture of syn and anti forms. No isomerization was observed during recrystallization or chromatography on silica gel, b u t the isomer of m.p. 113114' could be separated by chromatography. Mesitylglyoxylonitrile Oxime p-Toluenesulfonate .--A solution of 1.10 g . of the oxime (m.p. 113-114°) and 1.2 g . of p-toluenesulfonyl chloride in 20 ml. of acetone was treated dropwise at 0 " with 2.5 ml. of 2.6 S aqueous sodium hydroxide. After 15 niin., 5 ml. of water was added and stirring was continued for 3 0 min. without the cooling bath. Water (25 m l . ) was added, and the precipitate was filtered and washed with water. T h e tosylate weighed 1.80 g., m.p. 150-152". It was recrystallized from hexane-chloroforni, ni . p , 152-153 O . When the tosylate was prepared from the oxime mixture of m.p. 70-77", the crude product melted a t 132-136". Anal. Calcd. for C18HI&.'03S: C, 63.14; H, 5.30; X , 8.18. Found: C, 63.12; H, 5.34; N , 7.99. T h e tosylate characterized earlier ( m . p . 152-153") could be obtained by recrystallization of the mixture. T h e isomeric tosylate
( 1 1 ) Since mechanism B would not require anti migration of the aryl g r o u p , the sun configuration of t h e oxime tosylates cannot he ruled out. (12) Tosylates of arylaldehrde oximes d ~ c o m p o s ereaddy to a iniytrirp of nitrile and isonitrile; see E. Muller a n d B. Narr, Z . S a l u r / o r s r h . , 16b, 84.5 (1961). (13) Melting points are uncorrected.
2438
XOTES
VOL. 28
was obtained as the first fraction elected from a silica gel column, Synthesis of Di(methylthiomethy1) Polysulfides m.p. 149-150". I t s infrared spectrum was quite different from t h a t of the 152' isomer. MARIOR. ALTAMURA, TORSTEN HASSEISTROM, AND :ZnuZ. Calcd. for Cl~Hi8N2O3S: C, 63.14; H , 5.30; N, 8.18. LOUISLONG,JR. Found: C , 53.07; H , 5.40; S , 7.62. Reaction of Sodium Ethoxide and Phenylglyoxylonitrile Oxime Tosy1ate.--4 mixture of 5.00 g . (16.7 mmoles) of the tosylate and Pioneering Research Division, U . S . Army Natick Lahoratorips, 100 rnl. of absolute ethanol was refluxed while 13.9 nil. of 2.60 N .Vat&, Massachusetts sodium ettioxide in ethanol was added dropwise. T h e mixture was refluxed an additional 2 hr. and then the volume of ethanol Received January 18, 1863 was reduced to about 25 nil. a t reduced pressure. T h e residue was diluted with water and the product was partitioned between sym-Dialkyl polysulfides have been known for a water and methylene chloride. Evaporation of the methylene chloride left 2.49 g. of residue. Distillation of this residue gave long time.I2 Some of these have been detected in diethyl T-phenylimidocarbonate, 1.73 g., b.p. 100" ( 5 m m . ) , natural p r ~ d u c t s-5, ~while others have found industrial n z o D 1.5110, infrared spectrum identical with t h a t of an authentic applications.6 Recently, a series of methylthioalkyl sample .lo isothiocyanates, of which methylthiomethyl isothioDiethyl N-(p-Chloropheny1)imidocarbonate.-A solution of 2.0 cyanate' is the lowest member, has been isolated from g. (6 rnmoles) of the tosylate of p-chlorophenylglyoxyloiiitrile oxime in 7 5 ml. of absolute ethanol was stirred at 60" while 13 natural products. Symmetrical dialkyl polysulfides, ninioles of potassium hydroxide in I O ml. of ethanol was added wherein the alkyl moiety contains a sulfur atom, as dropwise. T h e solution then was stirred overnight a t ambient exemplified by the methylthiomethyl group, have not temperature. After t h e volume o f ethanol in the reaction mixture been described. had heen concentrated to about 20 nil., water was added. and the mixture was extracted with niethylene chloride. T h e residue CH3--S-CH*--S.-CH*--S--CH3 obtained on evaporation of the metliylene chloride was distilled to give diethyl S-(p-chlorophen~-l)imidocarb(~nate, 0 . 8 g., n Z o ~ Methylthiomethyl mercaptan, a key compound in the 1.5295, b.p. 100" ( 1 m m . ) , infrared spectrum identical with t h a t present investigation, was prepared in good yields by a of a n authentic saniple.l0 modification of the method described by Feh& and Isolation of Urethans and Nitriles.-The procedures used to 170gelbruch.g This modification obviated the danger of obtain the data given in Tahle I1 are illustrated. a n explosion. Due to its instability and permeating, ( A ) Ethanolic Potassium Hydroxide a n d p-Chlorophenylgarliclike odor, much difficulty was encountered in glvoxylonitrile Oxime Tosy1ate.---A solution 3.35 g . (10 mmoles) of the tosylate was treated with 22 nimoles of pot handling and characterizing this compound. It was in etlranol as described earlier. The 3.0 g . of crude residue obfound practical not to isolate the mercaptan for pretained was c1ironi:itogr:tphed on a silica gel column. Pentane in parative work, but to use it in absolute ether solution. methylene chloride, methylene chloride, and ethyl acetate in Di(methylthiomethy1) sulfide, included for comparanietliylenr chloride were iised n.s surcessive eluents. The first tive purposes, mas prepared by a new method from fraction char:icterized was p-chlorohenzonitrile, 0.077 g . (5.677 ), rn .p, 92-93". infrared spectrum identical with t h a t of an authentic chloromethyl methyl sulfide and the hydrate of sodium sample. The major fraction, eluted t)y 5% ethyl acetate in hydrosulfide (from commercial source) in benzene. methylene chloride, was recrystallized from hexane and gave pThe di- and trisulfides were obtained in whstantial chl(ircinhen?-liiretti~n, 1.28 g . (61( i,), m . p . 68-69" (reportedlo amounts as by-products from the reaction used to make m . p . 68-60'). the tetrasulfide, narnely by treating methylthiomethyl (B) Ethanolic Sodium Ethoxide plus p-Methoxyphenylglyoxylonitrile Oxime Tosy1ate.--A solution of 4.95 g. (15 mmoles) mercaptan and sulfur monochloride in dry benzrnc in cif the tosyl:tte in 100 nil. of hot absolute ethanol was stirred the presence of pyridine a t a low temperature --a modiwhile ?dl n i l . of 1 . I 9 S sodium ethoxide in ethanol (36 mmoles) fication of the general procedure for the synthesis of was added. The mixture was refluxed for 2 hr. T h e organic polysulfides of Patel, P t a/.,'" who did not obtain the prodiict was isolated hy estmction with methylene chloride and tetrasulfide The di- and trisulfides were obtained in W:LS rhroniatographed on :t silica gel column. T h e first fraction characterized W:IR an isonitrile, 0.127 g . (6.4?{), m.p. 61-62", the pure state by fractional dktillation i n high vacuum, infrared spwtruni identical with that of authentic material. T h e while the pure tetrasulfide was isolated from the immajor fr:wtiori, p - n i e t t i o x y p h e n ~ l ~ ~ r e t l i ~weighed ~n, 2.27 g. purities in the residue by extraction nith solvents. ( 7 6 . 7 < ; , ) . One recrystallization from hexane gave 1.88 g. of I n another experiment the same reaction wa5 carried ureth:in, ni.11. 65-66' (reported14m . p . 66-67"). out but in the absence of pyridine. The vacuum Ethanolic Potassium Hydroxide and Mesitylglyoxyloi C) nitrile Oxime Tosy1ate.--When 1.7 g. ( 5 ninioles) of the tosylate fractionation of the reaction product yielded the diof m.p. 152" (presum:ibly o f cnnfipuration 1) was allowed t o react and trisulfides, and an appreciable residue, which c o d d with 11' nimciles of piit:iesirim hydroxide in 40 nil. of ethanol for 80 not be distilled. This polysulfidic substance iindcrwcnt niin,, the only products characterized after cliromatography on autodesulfuration a t room temperatiire, a bchavior silica gel were ~ i i e ~ i t ~ l g l y o s y l o n i toxime. rile 0.38 g . ( d o ? ) , m.p. 114-114". :ind 0.10 g . of an oil that appeared (infrared and analysis) to he mesitylglyoq Ionitrile oxime ethyl ether. (1) P. Klason. J . p r n k l . Chem., 1 6 , 210 (1877). .Inn/. Calcd. for CliHi6XL'90: C , 72.19; €I, 7.45; N , 12.96. ( 2 ) R . Connor, "Organic Chemistry," Vol. 1. H. Gilman, E d . , .John Wilpy Found: C , 72.06; TI,7.69; N , 14.3. and Sons, I n c . , K e w York, N . Y . , 1943. p. 864. ( 3 ) E. E . R r i d , "Organic Chemistry of Bivalent Sulfur," T'ol. i3, Chernical I n another experiment, 2.60 g. (7.6 nimnles), of the mixed Publishing '20..Inc.. New York. N . Y . , 1960, 1,. 3 6 2 . tosyhtes cif 1ii.p. 132-136" was allowed t o react with 15.1 mmoles (4) S. D. Bailey, M , I,. Basinrt, .I. L. I l r i a c o l l . a n d A . I . h l c r a r t h y . .I. of potassium hydroxide in 80 ml. of ethanol a t 60" for 3 hr. T h e F o o d S c i . . 26, lR3 (1961). p r o d l i c t mixture R:IS rhroniatographed nn silica gel to give, in ( 5 ) L. Jiroriaek, Chem. List?,. 5 0 , 1840 fIS.50): Collprfion C z w h . Chprn. iirdtlr ( I f elution froin the c o l i i n i n , 0.21 g . of the supposed 0-ethyl Commun., 2 2 , 1491 (1457). oxinip. 0 . 1 1 g . of mesitonitrile, m . p . 49-51' (reported15 m.p. (13) Rrf. 3 , p. 392. 50-52"). and 0.79 g . ( 5 5 ( ' : ) of the recovered oxime, m . p . 72(7) T. Hanselstrom. R . C . Clapp, I.. T . Alann, .lr., and I.. T,ona. .Ir.. 76'. .I. O r g . C h r r n . . 26, 3020 (1801). ( 8 ) .4 ICjaer. "Fortrchritte der ChPmie oraanischpr SatilrntoRP." X'ol, 18. 1,. Zrctimriater. Ed., Sprinnrr-Verlaa, W'irn, 1 9 G O . 11. 1k4. (9) F. Frh6r a n d K.V o w l b r u c h , Chem. Rcr., 9 1 , 88fl (19.58). (10) P. P. Patpl, I. Srnniipta, a n d C.. C . Clrakravarti. .f. lndinr! I n n l . S r i . , 1.38, 73 ( l 9 3 0 ) ,
NOTES
SEPTEMBER, 1963 previously observed to occur in the higher polysulfides." l 2 I n about three weeks, the separation of sulfur ceased and the compound reached the stable state corresponding to an octasulfide. This transformation is shown in Table I. The use of pyridine in this reaction has favored the formation of lower polysulfides. TABLE I ACTODESULFURATION OF A DI(METHYLTHIOMETHYL) POLYSULFIDE PRODCCT T i m e , ----Found---days C
0a 3 24 60
...
H
S
. .
----CaIcd.(CHISCH~IS, C
..
12
> 10
...
10.89 2 . 3 2 86.79$ 10 10.85 13.17 2 . 9 7 8 4 . 2 6 8 12.68 13.20 3 . 1 2 84.84 8 12.68 a -4 sample could not be taken for analysis a t transformation. Obtained by difference.
H
S
...
...
2 . 2 8 86.88 2.66 84.66 2.66 84.66 this stage of
LIethylthiomethyl pentasulfide was prepared from chloromethyl methyl sulfide and anhydrous potassium pentasulhde substantially by the same procedure used by Riding arid Thornasl3 for making di(ally1) pentasulfide. By analogy to the structure of the dialkyl polysulfides,'d--'8 it is assumed that the polysulfidic bridge of the di(methylthiomethy1)polysulfides has an unbranched chain structure. The physical properties of the di(methylthiomethy1) polysulfides are given in Table 11.
2439
atmosphere at a bath temperature of 45-50'. T h e residue was subjected t o a vacuum of 100 m m . for 5 min. A vapor phase chromatogram of this concentrate showed two peaks, one for the thiol and one for ether. A rough estimate of the amount of mercaptan present was made from the ratio of the areas of the mercaptan and ether peaks. The retention times for the thiol and ether were 2.30 and 0.74 min., respectively, when a 0.1-pl. sample was used with a Perkin-Elmer "Q" column (Apiezon "L" grease) a t 165' and 15-p.s.i. helium gas pressure. The best yield of pure mercaptan was 7.49 g. (69.0%, lit.9 58%) from 20.0 g. of starting material. Instead of the pure mercaptan, a thiol-ether concentrate, freshly prepared, was preferred for the reaction purposes. The concentrate could be conveniently stored in the presence of nitrogen a t about 3". The pure product was obtained b y vacuum distillation under nitrogen a t 47 mm. It was a clear, water-white, mobile liquid with a very sharp, penetrating, garliclike odor; b.p. 54-55.3" [lit.Ob.p. 60" (47 mm.)] . An elemental analysis and infrared spectral d a t a , now reported for the first time, are given subsequently. Anal. Calcd. for CZH&: C, 25.50; H , 6.42; S, 68.08. Found: C, 26.14; H , 6.83; S, 67.87. T h e infrared spectrum showed, in addition t o the characteristic weak band for the thiol group at 2500 em.-', strong absorption bands a t 2830, 1430-1420 (doublet), 1205, 986, and 699 em.-'. Methylthiomethyl mercaptan, in the pure state or in ether solution, slowly decomposed and deposited a white solid on standing. From related work (unpublished results) on the previoua thiol, a solid of this nature was found t o be a mixture of thioformaldehyde polymers. Trithiane was not identified as a component of this mixture. Di(methylthiomethy1) Sulfide.-In a n attempt to prepare methylthiomethyl mercaptan b y the addition of chloromethyl methyl sulfide to a dry benzene suspension of the solid (previously formed in situ by the distillation of the water of hydration as an azeotrope with benzene froni commercial sodium hydrosulfide)
TABLE 11 DI(METHYLTHIOMETHYL) POLYSULFIDES CHzSCH2 S n CH?SCHx Di ( m e t hylthiomethyl) n Monosulfide Disulfide Trisulfide Tetrasulfide Pentasulfide
1 2
3 4 5
B.p., 'C. (rnm.)
Formula
110.8-111.8(13.5)n 8 3 . 2 - 8 5 . 8 (0.021) 118-119.8 (0.021)
CiHioSa CaHioSa GHioSs CiHloks
... ...
CnHioS,
Feh4r and Vogelbruchg reported:
-1101. wt.Calc,l. Foun3
nioD
1 . 8 8 1 1 (20)' 1.6280(20) 1.6628(20) 1 . 7 0 7 2 (30)* 1.7077 ( 7 8 . 6 ) h
b.p. 113' (13 m m . ) ;
164 186 218 251 283
155 212 250 268
nZ0D1.5818;
Experimental l g Methylthiomethyl Mercaptan.--S-(Methylthiomethy1)isothiuronium chloride' (19.6 g . , 0.114 mole) was hydrolyzed by stirring with 24.6 g . of 5 K sodium hydroxide solution at room temperature for 4 h r . T h e oil t h a t separated as an upper layer was extracted with several portions of peroxide-free ether. The combined ether solutions were washed with water and dried over anhydrous sodium sulfate. After filtration, the solvent was removed b y distillation through a T'igreux column in a nitrogen (11) R . C. Fuson. C . C. Price, D. >I. Burness, R. E. Foster, W. R . Hatcha r d , a n d R. D. Lipscomh, J . O r g . Chem.. 11, 487 (1920). (12) S . Bezri and P . Lanea, Onsz. chim. ital.. 80, 180 (1950). (19) R. W. Ridinw a n d ,J. P. Thomas, .I. ChPm. S o c . , 123, 3271 (1923). (14) .J. I h n o h u e and V.Schomaker, .I. Cheni. P h p . . 16, 192 (1948). (1.5) I. M.Dawson. A . R I . Mathieson, and .J. 11.Robertson. J . Chem. SOC.. 322,1250 (1918). (10) .J. E. Baer and A I . Carrnack, J . A m . C h ~ m Soc.. . 71, 1215 (1949). (17) L. AI. Kushner, G . Gorin, and C. P. P m y t h , i h i d . . 7 2 , 477 (1950). (18) F. FehPr, G . Kraiisr. and K. Vouelbruch, Chem. Ber.. 90, 1590 (19,571. (19) S o d i u m sulfhydrate crystals ( h y d r a t e of s o d i u m hydrosulfide) was obtained frorn Fisher Scirntific Co. Sulfiiv rnonocliloride mas purchased froin IXstillation Products Indiistrirs. Potassium pentasulfide was prrpared by t h e method rlpscribetl by .I.Rule ani1 .J. S. Thoinas. J . Chem. Soc., 106, 2819 (1914). and chloromethyl methyl stilfide by t h a t of H. RichteenChem. , Be?.. 8 6 , 142 ( 1 9 5 3 ) . Vapor phase ctironiahain and R . A l f r ~ ~ l s s o n torrraphy rnrasureirients were taken on a Perkin-F:liner vapor fractorneter, AIorlel 1 X C . I n f r a r d ahsorption s p e r t r a were obtained froin a PerkinElmer Infracord. L I o d ~ l 137. AIolcculsr w i g h t s were deternrined hy a Rlrchrolah vapor prcssure osmonieter. Model 301. T h e boiling points are iinrorrerted.
157*
---Calcd.--
C 31.13 25.77 21.99 19.18 17 00
Found---------
r
H
S
6 . 5 3 62.34 5 . 4 1 68.82 4 61 7 3 . 3 9 4.02 76.80 3 . 5 7 79.43
mol. wt., 140;
Y,61.58c/;,
c 3 0 . 7 1 , 3 0 77 2.5 63, 25.51 22.36,22.29 19 09, 19 10 16.56, 16.64
H 6 64,6.57 5 .58, 5 . 4 9 4 . 8 3 , 4 69 4 . 0 9 , 3 96 3 . 4 6 , B 59
S
62 1 3 , 6 1 84" 68 97. 69 24 73 6 6 , 7 3 7.5 77.03,76 17 80 1 5 , 8 0 . 3 9
The value a t 20" was>1.71.
and by t,he reflux of the resultant mixture for 10 h r . , the sulfide was obtained instead as the major product in the form of a light yellow, niohile oil with a peppery odor. Ten grams of chloromethyl niethgl sulfide gave approximately 2.0 g. of di(methy1thiomethyl)sulfide, the physical constants (Table 11) of which agreed with those of FehCr and Vogelbruchg for the same compound prepared by a different method. A gas-liquid chromatogram showed a retention time of 15.2 min. for the sulfide when a 0.25-pl. sample was used on a PerkinElmer "Q" column a t a column temperature of 166" and a helium gas pressure of 20 p.s.i. Di(methylthiomethy1) Disulfide.-One of the fractions from the fractional distillation of the crude tetrasulfide, vidr infra, was pure di(methy1thiomethyl) disulfide. It) was a yellow, mobile liquid with a soft, garliclike odor. From 5.2 g . of methylthiornetkiyl mercaptan, 1.54 g. of the pure disulfide was otitained. The physical constants of this compound m d the other polysulfides are given in Tahle 11. Di(methylthiomethy1) Trisulfide.-This trisulfide was ohtained from the same source as the disulfide b y vacuum fractional redistillation of a higher boiling fraction of the crude prodnct. It was a liqnid, resemhling the disulfide closely in appenrnnce and odor. The same amount of mercaptan gave 2.0 g . of this s u hstance . Di(methylthiomethy1) Tetrasulfide.zO-To a solution of 5.2 g . (0.056 mole, ;i.iC/;excess) of methylthiomethyl mercaptan, as an
(20) This reaction should he performed under a ne11 ventilated hood
2440
NOTES
VOL. 28
ether concentrate,21 and 5 ml. of dry pyridine in 10 ml. of dry attempt to distil i t under vacuum (0.08 mm.) in a nitrogen a t benzene was added dropwise a t 5-10" with agitation a solution of mosphere gave evidence of decomposition, accompanied by a n in3.5 g. (0.026 mole) of sulfur monochloride in 10 ml. of dry benzene crease in pressure to 0.1 mm. Two small quantities of oil, 1.4 g. over a period of 1.25 hr. White, crystalline pyridine hydrochloand 1.3 g., presumably impure di- and trisulfide, were collected. ride soon precipitated. The addition of more benzene (10 ml.) When the bath temperature had reached 185" and the vapor was necessary to loosen the crystalline mass. After the reaction 123", distillation ceased and the pressure dropped to 0.08 mm. mixture had come to room temperature, i t was stoppered and The residue was a clear, yellow-orange, moderately viscous oil allowed to stand overnight. Water (50 ml.) was added and the (3.5 g.) with a mild odor not resembling the previous polysulfides. mixture was shaken for 15 min. The benzene layer was sepaThe oil soon became opalescent and, on standing a t room temrated and the aqueous layer extracted with 10 ml. of benzene. perature, slowly underwent autodesulfuration with the separation The benzene layers were combined and washed successively with of .sell defined crystals of rhombic sulfur, identified by melting 40 ml. of 1 .V sulfuric acid, 40 ml. of 1 S sodium bicarbonate, point (109.1-111.5°) and molecular weight (found, 256, 257; and thrice with 40-ml. portions of water. After the benzene solucalcd. for Sa, 256). Shaking the polysulfidic material with acetion was filtered and dried over anhydrous sodium sulfate, the tone produced the same transformation in much less time. The solvent \vas removed in uacuo a t 55-60" under nitrogen. When autodesulfuration which took place on standing at room temperathe solvent ceased to distil, the pressure slowly was reduced to ture was observed over a period of time. At intervals, the clear, 34 mm. The crude product (7.9 g . ) was fractionated a t 0.021 supernatant oil was carefully removed from the crystals of sulfur, mm. under nitrogen. The portion t h a t distilled a t 83.2-88.5" and its composition determined by elemental analysis. The was essentially pure di(methylthiomethy1) disulfide and that results are given in Table I. At the end of 3 days, the composiwhich distilled a t 90.2-li 1O was impure trisulfide. Redistillation of the polysulfide corresponded to t h a t of di(methy1thiotion of the latter fraction under the same conditions gave pure methyl) decasulfide. It corresponded to that of an octasulfide di(methylthiomethy1) trisulfide. Distiilation was discontinued to a t the end of 24 days, and thereafter remained constant. apoid decomposition when the residue did not distil upon raising Acknowledgment.-The authors are grateful to Dr. the bath temperature to 165'. The residue (2.3 g., 35.4%) was impure tetrasulfide. The method of solvent e x t r a ~ t i o n ~ ' ~D. ~ ~Stanley ~~ Tarbell of the University of Rochester for previously employed for the isolation and purification of higher his helpful suggestions; to Dr. W. Davidson for the polysulfides was used for this compound. Extraction with molecular weight determinations; to Mr. F. Bissett absolute alcohol, filtration of the alcohol solution, and removal of of our laboratory for the gas chromatography and the solvent a t room temperature in vacuo and under nitrogen spectrophotometric determinations; to Mr. C. Digave the alcohol-soluble tetrasulfide. From t>his substance, after extraction with absolute chloroform and removal of the Pietro of the Analytical Laboratory for the microsolvent under the same conditions, the di(methylthiomethy1) analyses; and to Mr. A. Deacutis for the preparation tetrasulfide was obtained as a pale yellow, moderately viscous of potassium pentasulfide. oil with the usual garliclike odor. The yield was 1.39 g. (21.40/,). Di(methylthiomethy1) Pentasu1fide.-To a suspension of 12.1 g. (0.051 mole) of potassium pentasulfide in 100 ml. of anhydrous ether, protected from moisture and previously flushed with The Formation of Thiachroman as a Major dry nitrogen, was added a solution of 9.8 g. (0.101 mole) of chloromethyl methyl sulfide in KO ml. of anhydrous ether over a period Product in the Claisen Rearrangement of Allyl of 0.5 hr. a t room temperature, with magnetic stirring. The Phenyl Sulfide' reaction mixture was flwhed again with dry nitrogen, stoppered, and allowed to react a t rnom temperature for 13 days. During this time, the mixture was stirred 8 hr. daily for 9 days. The C A L Y. hfEYERS," COST.4NTINO RINALDI, AND LUCIANO BONOLI solid component changed from a coarse, orange-red powder to a pale yellow crystalline material. The reaction mixture was filIstituto d i Chimica Industriale, Universitd d i Bologna, tered under slight vacuum, the solid residue was washed with Bologna, Italy several 10-ml. portions of ether, and the washings were combined wit,h the filtrate. The resultant ether solution of the pentasulfide Received March 12, 1963 was washed with three 50-nil. portions of water, dried over anhydrous sodium sulfate, filtered, and the ether removed by distillation a t 34-40' under a slight vacuum and in a nitrogen atmosUnder conditions effecting the Claisen rearrangement phere. The residue was taken to, and kept a t , 50" and about of allyl phenyl ethers, thia analogs afford mainly the 35 mm. for 10 min. The yield of crude pentasulfide was 10 g. corresponding propenyl phenyl sulfides (prototropic ('70%). By extracting the crude product with absolute ethanol (disti!led over magnesium) to remove the alcohol-soluble impuriisomerization). Only recently, however, the obserties, and by suhjecting the residue to a high vacuum a t room vations of Kwart and Hackett3 suggested that, through temperature to eliminate the last traces of solvent, the pentaamine catalysis, the thia analogs also may be induced sulfide was obtained as a pale yellow, slightly viscous oil with the to follow the path of the Claisen rearrangement; uiz., usual garliclike odor. The yield was 3.9 g. (27.3%). The allyl phenyl sulfide (I), dissolved in a high-boiling pentasulfide was stable with respect to autodesulfuration on standing a t room temperature. amine, was converted ( 15-2070) into 2-methyl-2,3Di(methylthiomethy1) Polysulfide Products Higher than Pentadihydrobenzothiophene (TI) by a single distillation sulfide.*O-To a solution of 7.4 g. (0.079 mole) of methylthio(atmospheric pressure). methyl mercaptan in 2 j ml. of dry benzene was added a solution During the course of our a convenient of 6.3 g . (0.039 mole) of sulfur monochloride in 15 ml. of dry benzene according to the procedure described earlier for the preparation of I1 was sought and we followed, essentetrasulfide. No pyridine was used. Copious evolution of tially, the method of Kwart and Hackett. In addition hydrogen chloride gas took place at the beginning of the reaction. to isolating the desired compound we found for the When the reaction mixture had reached room temperature, however, only a little hydrogen chloride gas was evolved which (1) This s t u d y is p a r t of a series dealing with the nature of organic sulfur groups a n d is supported b y grants froni the Petroleum Research Fund. The gave a weak test with hlue litmus paper and F i t h ammonia. preceding paper is b y C. Y. hleyers. G. Lombardini. and L. Bonoli, J . Am. With the exception that the benzene solution had to be washed C h e m . Soc., 84, 4603 (1902). ( a ) Inquiriesshould be directed to C. Y. Afeyers, six times with KO-nil. portions of ice-water to ohtain a negative Department of Chemistry, University of Southern California, Los Angeles 7, test for chloride ion, the procedure for working up the reaction Calif. mixture and removing the solvent was the same as that used for ( 2 ) E. S . Karaulova. D. Sh. hIeilanova, a n d G. D. Gal'prrn, Zh. Ohshch. the tetrasu!fide. The weight of the crude oil was 7.6 g. An K h i m . , 27, 3034 (1957). _____ ( 3 ) H. KtL-art and C. bf. Hackett, J . Am. Chem. Soc., 84, 1754 (1962). (21) T h e concentrate was a n approximately 61% solution of the merc a p t a n in d r y ether. (22) J. R. Thomas and R . W , Riding. J . C h e m . Soc., 126, 2460 (1924).
(4) C . Y. hIeyers, C. Rinaldi, and L. Bonoli, 144th N a t i o n d Meeting of the American Chemical Society, Los Angeles, Calif., April, 1963, Abstracts of meeting, p. 4M.
KOTES
SEPTEMBER, 1963
2441
than or very close to that of the amine solvent.10 In first time another major Claisen product, the isomeric our studies the initial mixture was allowed to rej?ux thiachroman (111).5 From these observations a new for several hours, the quinoline was then converted into and simple preparation of thiachromans is offered, its hydrochloride, and the sulfur-containing products further evidence of the greater stability of six- than were removed by ether extraction and separated by five-membered cyclic sulfides is suggested, and more fractional distillation. In this way both I1 [b.p information regarding the Claisen path of this reaction 113-114' (15 mm.); lit." b.p. 118-120' (21 mm.); 1,lis provided. dioxide, m.p. 116-117'; lit." m.p. 115-115.5'1 and I11 In several experiments involving two- to four-hour were detected and isolated. Only a very small amount of refluxing of equimolar quantities of I and quinoline, lower-boiling material was isolated, the balance of I conversions of about 25% into I1 and about 30% into being represented as high-boiling fractions. l 2 I11 were realized. The latter was identified by its Significantly, the observed per cent of conversion of I b.p. 125-127' (14-15mm.) [liL6"b.p.128-130' (l5mm.)] into I1 was more or less the same froin the single disand its alniost quantitative transformation (one-hour tillation (when much unchanged I but no 111 mas obreflux with excess hydrogen peroxide in acetic acid) t'ained) and from the several-hour reflux (when virt'ually to the 1,l-dioxide (IT), m.p. 89.5-90' (from water, no I but much I11 was isolated). This suggests that the then ligroin; lit.'m.p. 88-89'). Moreover, 111was prepared directly from thiachromaiione (T')* by r e d u ~ t i o n , ~ formation of I1 is rapid but involves reversible reactions (or a reaction cycle), while the reaction leading to I11 is then converted to the 1,l-dioxide, melting point the slow but unidirectional. The formation of I11 experisame as that of IT' either alone or in mixtures with it. nientallyI2b from I1 offers additional evidence for this mechanism. This reasoning is justified by related observations. Keither Ia nor Ib has been isolated from t'liese reactions. If Ia is formed it may be tautoi~ierized'~ to Ib and, as s u g g e ~ t e d ,the ~ latter rapidly cyclized t'o I1 which is detected early. Moreover, I1 mould be the expected quinoline, (Markownikoff) product from direct cyclization of Ia, polymers although its formation by this route may be less iinportant because of the lower reactivity of the iioncoiijugated double bond and because the amount of Ia is depleted by its isomerization to Ib. While 111 cannot be formed directly from I b it may be formed by t'he Ib cyclization of Ia. The formation of 111, then, would Ia (slow) (rapid)lt be slow not only because of reasons just mentioned, but because 111 is not the preferred (Llarkownikoff) product. Finally, it should be noted that cyclic six-niembered sulfides exhibit much less strain than the corresponding five-membered s y ~ t e n i s . ' ~While I1 is reasonI11 ably cleaved under these conditions, 1 2 h , re-forming Ia aiid Ib, 111 is stable and through its continuous formation becomes the major product when the mixture is allowed to reflux over an extended period. A very recent report,,l6of the formation of six-niembered sultones rather than the expected five-membered IV ones and explained on the basis of t'he greater stability of the former (the lat'ter being strained), virtually The isolation of 111 in this study, but not in that parallels the observations and conclusions arising from reported by Iiwart and Hackett, may reasonably be explained by comparing their respective reaction pro(IO) Compound 111 has a n appreciably hietier hoi!ine iioint. ancl, Pven if i t were present in small aniorints, it probably ~ v o u l dnot have been detertrrl cedure and results with our own. Their single distillain the distillate. K e found t h a t xr-iien a mixture uf qiiinolinp containing 20;G lions of solutions of I aiid quinoline (or K,S-diethylof I11 was distilled in coinnion equipnient al,ureciable amounts of 111 allpeared in t h e distillate only after a large p a r t of t h e quinoline had Iwen aniline) yielded I1 along with mainly unchanged I and distilled. small or t'race amounts of propenyl phenyl sulfide and (11) E. N. Karaulova, D. S h . h f e i l a n o r a , and G . D. Gal'pern, Dokl. A k n d . t'hiophenol. A11 of these have boiling points lower S o u k S S S R , 123, 99 (1958).
1
\
(a) T h e o x y analogs of I1 (2-niethyl-2,3-dihydrohenzofurans) and I11 fclironians) a r e iiii1)ortant side-products observed follouing t h e Claisen rearrangement of allyl plrenyl ethers. See I,. I. Sniith, Chem. R e c . , 27. 287 (1940): and I). 5. Tarbrll. O r g . Reortzons, 2 , 1 5 - 1 9 (1944). (fi) ( a ) .J. von Braiin, Ber., 43B, 3220 (1910); ( b ) C. Angelini a n d G. Grandolini. A n n . chim. (Ronie), 46, 235 (10.50), report b.p. 124-126° (IO rnm.) . ( 7 ) 1;. 0. Bordrvell and \V. H . lIcICellin, J . A m . Chem. S O C . , 7 3 , 2251 [l9.-)1). ( 8 ) 1.: lirollpfeiffer and H. Schultze. Ber., 66B, 1819 (1923). ( 9 ) Hatistartory y i e l d s n-ere obtained by a simple reflux \r-ith amalgatirated zinc in liyrirocliloric acid (until tests n ith 2.-l-(linitrophenylhy(lra~ine R'ere negatii-e) and t h e g r o d u r t isolated in the subsequent steam-distillatr. Tile \Volff-ICishner method (ref. 7 ) and lithium aluminum hydride ( r e f , h h ) have also been rised.
(12) (a) The propenyl phenyl sulfide t h a t probably was fornied is Pnstly converted into polynieric material (ref. 2 , 3 ) . fb) From a n,ixturP of I1 .and quinoline t h a t n-as refluxed f o r several hours there was isolat?d unchanged 11, small amounts of 111, and polymeric niaterial representing a b u u t on?third of the original charge of 11. (13) T h e well known tautomerization of allylbenopnes t o proppnylbenzenes (e.g., eugenol to isoeugenol) occurs in a m a t t e r of minutes in alkaline medium. (1.1) C. 1 '. Meyers, S. Ghersetti. and .I. RIangini. 140th S a t i o n a l l I e e t ing of the American Chemical Society, Chicago, Ill.. Heytrrl,ber, 1901. Abstracts of meeting. p. RT. (1.5) H. D. Hartough and S. L. Meisel, "Compoiinds with Condensed Thiophene Rings," Interscienre Publishers, Inc., Ken. York. N. I'.. 1954. pp. 3 4 f f . (16) \V. E. T r u r e . D. K. Burdge, and R..J. Hteltenkaml,, . J . O r g . C h ~ m . . 27, 3913 (19fi2).
NOTES
2442
this study of the cyclic sulfides. Similarly, the facile cleavage of cyclic five-membered sulfones under conditions not affecting the six-membered systems mas attributed $0 ring strain in the former.4 The isolation of both I1 and I11 in this study provides greater evidence that the classical Claisen rearrangement product (Ia) is the initial intermediate formed directly from I (cj. ref. 5). However, since neither we nor Kwart have actually isolated this classical product (or Ib) under conditions leading to the formation of I1 and now also 111, this evidence may be circumstantial and an alternate initial path from I cannot yet be definitely precluded. The Claisen route, with the initial formation of Ia and I b as very reactive transients (i.e., nonisolable)" derives further support from related studies. There are indications that benzenesulfenyl anions (ArS-) are much more reactive nucleophiles than phenoxy anions; e . g . , 1 % - attacks benzotliiophene 1,l-dioxide forming quantitatively the P-PhS-adduct, while PhO- is completely unreactive with this substrate.L8 Much greater niesonieric delocalization of the negative charge in the phenoxy systems reasonably accounts for this and siniilar observations. l Y It follows rationally that while o-allylphenols can be isolated from usual Claisen rearrangements, the corresponding thiols may simply cyclize too rapidly to be detected per se especially under the high-temperature and alkaline conditions employed. It has been suggested*" that I initially might cyclize directly into 111-anion which then is reversibly transformed into the anions of la, Ib, and 11, all in equilibrium, but irreversibly into I11 by proton abstraction. The experiniental facts-sequence of formation of I1 and 111, cleavage of I1 but not 111 under these conditions, etc.-together with independent evidence of the greater stability of six- than five-membered cyclic systems containing a sulfur atom4 1 4 - 1 6 do not support this argument. h closer examination of these reactions, necessarily under less vigorous conditions, should prove interesting and fruitful. (17) While this was suggested by K w a r t to explain t h e absence of I a , n o supporting evidence was offered. (18) F . (;. Bordwell a n d W. H. AlcRellin, J . A m , Chem. Soc., 72, 1985 (1950). (19) See C . C . Price a n d R . Oae, "Sulfur Bonding," T h e Ronald Press c o . , New T o r k . s.I-,, 19fl2, p. 2 5 . (20) Tlrc s i i t h o r s grstefullj- acknowledge R referee's suggestion.
Azasteroids. IV. 3Iicrobiological Dehydrogenation of C-Ring Azasteroids' ROBERT H AIL?[
R A\D
ROBERT MLIR
Rzologzcal and C hernaCul Research Davaszons, G D Searle und Company, Skokze, Illznozs IZecczbetl February 26, 1563
;\.mstcroid- ha\ e attracted sonie attention a? a direci t liicli t o aearch for 3teroid hormone analogs. Introductioii oi nitrogen into ring C (nith concurrent tion in
(11 P a r t 111 R H l l a z ir J O r g Chem 2 8 , 348 (1963)
VOL.
28
homoannulation) has been reported by Mazur2 and Z d e r i ~ . A11 ~ ~ our attempts to convert suitable precursors to A-ring unsaturated compounds by chemical methods have failed-for example, bromination-dehydrobromination,6 selenium dioxide oxidation,' arLd heating with dichlorodicyanoquinones-and Zderic has mentioned having the same experience. We have now found that fermentation of 12a-aza-3Phydroxy-C-homo-5a-pregnane-12,20-dione (I) with Nocardia sp. A.T.C.C. 14558 gave 12a-aza-C-homo-1,4pregnadiene-3,12,20-trione(11). Compound I1 was obtained in two polymorphic forms having identical solution infrared spectra. The structure was confirmed by quantitative hydrogenation and by n.m.r. spectrum. The latter showed the typical complex pattern in the 350-4t50-c.p.s. region due to interaction of the C-4 proton with the C-2 proton (itself part of an AB system). Similarly, fermentation of 3P-acetoxy-12a-aza-17ahydroxy-C-homo-5 a-pregnane-12,20-dione (II I) with Nocardia sp. X.T.C.C. 14359 gave 12a-aza-17a-hydroxyC-homo-l,4-pregnadiene-3,12,20-trione (IT'), with typical n.m.r. spectrum in the 350--43O-c.p.s.region. A different result was obtained by fermenting 36-acetoxy12a-aaa-C-homo-5a-pregnane-12,20-dione (V) with Arthrobacter sp. ,Z.T.C c. 14360 which yielded 12a-azaC-homo-5a-pregn-l-ene-:3,12,20-trione (VI) ; the 1i.m.r. spe,trum showed the .1R pattern of a AWketone.
RO
w
o
m
H
I, R = R ' = H 111, R = Ac, R' = OH R = Ac, R' = H V,
0
IT, R = H I V , R' = OH
H
vI It is interesting that the ultraviolet spectra of all products showed a hypsochromic shift of about 4 mp usually associated with 11-keto steroids. To our knowledge, there is only one previous example of direct conversion of a ring X/B saturated steroid 3-alcohol derivative to a A1 4-3-ketone by ferrnentati~n.~ (2) R . H. Maeur, J . Am. Chem. Sac., 82, 3992 (1960). (3) R. H. hlaaur, ibid., 81, 1454 (1959). ( 4 ) J. A . Zderic, H. Carpio, a n d D. C. Limon, J . O r g . C h e m . , 27, 1125 (1962). (;7) .J. .1. Zderic a n d J. Iriarte, i b d . 27, 1750 (1902). (6) E. S. RothInan a n d 11, E. \Tall, J . A m . Chem. Sac.. 78, 1744 (1956). (7) C. l l e y s t r e , H. Frey, \\'. Voser, and A . WPttstein. Hela. Chim. Acto, 3 9 , 7 3 4 (1956). (8) D. Burn. D. N. Kirk, a n d 1'. Petrow. ['roc. Chem. Soc., 14 (1900). (9) V. E. Oriponi, Neil E. Ripler. a n d .J. J. Goodman, U. S. Patent, 3,047.468 (July 31, 1902).
SOTES
SEPTEMBER, 1963 Experimental
2443
'The n.m.r. spectrum was interpreted as follows: 19-CH, ( 5 7 ) , 18-CH3 ( i 5 ) , 21-CI13 (130), 2-H (doublet, 345 and 3 5 5 ) , NH (417), 1-H (doublet, 425 and 435). J1.2 had the value of 10 C.P.S.
We would like to thank R. T . Dillon and associates for analyses, rotations, and spectra. Melting points are uncorrected. The analytical samples were dried for 2 hr. a t 100' under moderate vacuum (about 10 m m . ) . N.m.r. spectra were obtained on a Varian A-60 spectrometer a t 10% consentration in deuteriochloroform using tetramethylsilane as an internal, standard. Steroid Epoxy Ketones. 11. 2,3-0xygenated Chemical shifts--figures in parentheses--are reported in cycles Steroids from la,2a-Oxidocholestan-3-one per second downfield from the standard. Colunin chromatography was carried out by S . Bilek and M . Blaumeiver (direction E . G . Ilaskalakis). Quantitative hydroWILLIAM REIWHA N D ROSALDLELIAHIEL genations were performed by i V . b l . Selby. 1Za-Aza-C-homo-l,4-pregnadiene-3,12,20-trione(II).--SoKedzae Chemacal Laboratory, Mzchzgan State C'ntuerszty, cardia s p . A.T.C.C. 14558 (Searle A20-16) was grown as a subEast Lanszng, Jlichzgan merged culture in a stainless steel fermentor in 35 1. of medium containing 200 g. of Ilifco Nutrient Broth and 10 g . of silicone emulsion (1)ow Corning .4ntifoam A F Emulsion). The culture Received 2?farch 13, 1863 was agitated by means of a paddle-type stirrer operating at 200 r.p.m. and was aerated with 10 1.p.m. of sterile air which entered In the preceding paper of this series,' base-catdyzed through a sparger located below the agitator. The incubation ring contract'ion of appropriate +epoxy ket'ones? After an initial growth period of 30 temperature was 25'. h r . , 10.0 g. of 12a-aza-3S-hydroxy-C-homo-5a-pregnane-l2,20- was suggested as a novel approach to norst'eroids. dione (1)s in 200 m l , of acewne and 50 ml. of methanol was added However, this mode of reactioii was not observed when and incubation continued for 14 h r . 4,j-oxidocholestan-3-one was treated with rnet'liaiiolic The culture was extracted with two 18-1. portions of methylene base, the major product being 4-metho~y-~~-cliolestenchloride and the combined extracts distilled t o dryness. The %one (VI).1'4 We now report the result's of a similar residue, 10.26 g., was crystallized from 200 nil. of 1 : 1 benzenecyclohexane and the desired product separated as irregular study of lcr,2cr-oxidocholestan-3-one (I). prisms, 6.10 g. (62Yc), m . p . 180-184". Recrystallization from Treatment of I with refluxiiig methanolic sodium benzene raised the m . p . to 183-185'; [ a ] % $48" ( c 1, methanol); hydroxide yielded 2-methoxy-A1-cho1esten-3-one (11)
?:X :
240 mp
(e
16,100).
as the major product. This st'ruct'ure assignment was Anal. Calcd. for C21H27?;03: C , 73.87; H , 7.97; E,4.10. Found: C, 74.03; H , X.lO; N , 4.06. based on the infrared spectrum,: : , :A 3.95 and 6.23 p , A sample (19.5 mg.) in 957, ethanol was hydrogenated over the ultraviolet spectrum, X ~ ~ " 263 f ~nip ( e B O O ) , the 5% palladium on carbon (4.0 mg.) in the apparatus of Clausonn.m.r. spectrum (vinyl hydrogen and methoxyl hydroI(aas.10 Hydrogen uptake ceased at 102% of two double bonds. gen appear as singlets, 4.18 and 6.57 T , respectively, Similar results were obtained with model ~ 1 , ~ - 3 - k e t o n e s . T h e n.1n.r. spectrum was interpreted as follows: 19-CH3 with an area ratio of 1 : 3 . 3 ) ,and the elenieiital analysis. (io),8-CH3 (75.5), 21-CH3 (129,5), 4-H (broad singlet, 361), I n addition, acid hydrolysis of I1 to the diosplieiiol IIIa 2-H (doublet of doublets, 367 and 377),XH (422), 1-H (doublet, and an independent synt,hesis of I1 from diosphenol 422 and 432). J 1 , 2had the usual value of 10 c.p.s. while J2.4 IIIb provide chemical evidence for this structure. was about 2 c.p.s. The reaction probably proceeds by opening of tlir oxiIn a second run under identical conditions, the product had m.p. 200-203". T h e infrared spectrum i n chloroform was identirane ring through methoxide ion att,ack at C-2 followed cal with t h a t of material with m . p . 183-186'. by @-eliminationof water. 12a-Aza- 17~-hydroxy-C-homo-l,4-pregnadiene-3,12,20-trione An isomeric derivative, 3-methoxy-A3-cholesten-2-one (IV).--The fermentation was conducted as described except (IV),was prepared by methylation of a mixture of dios t h a t the organism employed was .l:ocardia s p . A.T.C.C. 14559 (Searle A20-17). Following an initial growth period of 26 hr., phenols IIIa and I I I b with alkaline dimet,hyl sulfate. 6.0 g . of 3p-acetoxy-lZ'a-aza-lia-h~drosy-C-homo-5a-pregnane- The infrared and ultraviolet spectra of I\' were siniilar 12,20-dione (III)3 dissolved in 250 ml. of acetone was added and to those from 11; however, the n.ni.r. spectrum of thr incuhation continued for 21 hr. The crude material obtained former exhibited a doublet a t 4.83 7 , having au area by methylene chloride extraction was chromatographed on silica gel. Elution with ethyl acetate gave the desired product, 1.01 ratio to the methoxyl resonance a t 6.49 T of 1 : 2.9, in g . ( l i ( L ) , m.p. 252-256". Crystallization from methanol contrast to the singlet vinyl resonance observed for 11. yielded squnre prisms, m . p . 260-262"; [ a l Z 3 +45' ~ ( c 1, chloroThe exclusive formation of isomer IY in this reaction is form); "::A: 240 mp ( e 16,300). interesting in view of the fact that' there are no ob\ious .4na2. Calcd. for CzlH2,S04: C, 70.56; H , 7.61; E,3.92. steric factors favoring alkylation of one diosplieiiol over Found: C, 70.46; H , 7.93; X , 4.33. T h e n . m . r . spectrum was very similar to t h a t of compound 11. the other. Since equilibriuiii between t,hr two dios12a-Aza-C-homo-5a-pregn-1-ene-3,12,20-trione (VI).-The phenols and their conjugate bases is undouht'rdly fermentation was carried out as described for compound I1 established in t'he alkaline niediuin employed in this except t h a t the organism used was ilrthrobacter s p . A.T.C.C. react'ion, we suggest this select,ivity reflect,s a difCereiice 14.560 (Searle B22-46). After an initial growth period of 26 hr., 10 .0 g . of 36-acetoxy- 12a-am-C-homo-5a-pregnane- 12,20-dione in the stability of the diosphenol conjugate bases and or: ( V ) 3 in 200 ml. of acetone was added and incubation continued the ~net~hylat'ion derivatives.: =2n instructive coiitrast for 21 hr. The crude product was chromatographed on silica is provided by the preparat'ioii' of :~-niethoxy-S'-cllolesgel. Elution with ethvl acetate yielded compound 1.1, 2.56 g. ten-4-one (171) from diospheiiol T' through siniilar (26c/;), m.p. 194-196'. Crystallization from 1 : 1 benzenetreatment with dimethyl sulfate. In this case, steric cyclohexane gave clusters of needles, m . p . 206-20'7.5'; [ a ]% $ 3 2 " ( c 1, methanol);H: : :A 227.5 nip ( e 11,100). ilnal. Calcd. for C21H29x03: C, 73.43; H , 8.51; S , 4.08. Reusrh and R . LeXIahieii, J . A m . Clie77~.Soc.. 85, 1660 (190:3). (1) Found: C, 73.56; H,8.37; S , 3 . 9 9 . (2) I T , Treibs. Bpr., 64, 2178, 2513 (19.31). A sample (21.63 mg.) in Mfparsas a doublet n?th J 3 c.p.s. R'e have confirmed the essential f r ~ a t r i w sof thip stildy. ( $ 3 ) 111 nddition to tlir C-19:C-I interaction present in I I I b b u t not in I I I a . t i l ? frrriiicr i s o m p i h a s serious compressions brtiveen C-19 :C11 a n d C-llasters thesis of Rlelvin Vi. Schlechter presented to t h e Virginia Polytechnic I n s t i t u t e in 1988. ( 3 ) Supported b y a research g r a n t (AP-88) from t h e Division of Air Pollution, Bureau of S t a t e Services. Kational I n s t i t u t e s of Health, U. S. Public Health 8erT-ice. ( 4 ) F o r recent n-ork, see ( a ) L. K. Brice and R. D . K a t s t r a , J . A m . Chem. Soc.. 8 2 , 2609 ( I S G O ) , a n d ( b ) F. A . Vingiello. h l . 0 . L. Spangler. and J. E. Rondurant. J . O w . Chem., 26, 2091 (1900). t c ) T h e mechanism is shown onlj- t o r 2-henrylplienyl 4-pyridyl ketone: h o a e r e r . it presumably would be t~ssrntlallj-tlie same for t h e t w o Isomeric pyridyl ketones and it is very eirnilai t o a previouslx published rnrchar~ismfor t h e cyclization of 2-benzyl-
of the 4-pyridyl ketone, the pyrotonated species corresponding to I1 has its positive charge at a greater distance from the carbonyl group, and this may favor, relative to the other isomers, the formation of the diprotonated species and account for its relatively faster rate of reaction. It also should be noted that the 2pyridyl ketone, in forms corresponding to I11 and IV, may form a five-membered intramolecular hydrogen bonded ring structure with resulting stabilization of these structures. As the pyridyl ketones (I) presumably exist in acid environment primarily as pyridinium ions (11),acid in excess of that required to protonate the basic nitrogen should be required to effect cyclization. This was shown to be the case. 2-Benzylphenyl 3-pyridyl ketone under a five-hour refiux with an equimolar amount of 48y0hydrobromic acid in acetic acid gave no 9-pyridylanthracene, whereas an 88y0yield resulted when excess acid was used.
h e n / r i i l i e n o n e s . 41J C . Irovrd prel,aration, see P. Polss. P h . D . thesis, Virrinia Polytechnic I n s t i t u t e , Rlacksbrlrg, Va., December, 1962, p. 109.
(7) C . (1949).
The three isomeric pyridyl ketones needed for this study were prepared through the interaction of the Grignard reagent of 2-bromodipheiiylniethane6 and the appropriate cyanopyridine followed by hydrolysis to the final product. Only the 2-pyridyl ketone was obtained in pure crystalline form. The 3- and 4-pyridyl isomers were obt,ained as distillable oils which were characterized via their picra'tes and by essentially
K. Bradshe, and F. A . Vingiello, J .
Am. Chem. Soc.. 71, 1434
KOTES
SEPTEMBER, 1963
2449
TABLE I
Isomer
2-Pyridyl 2-Pyridyl 2-Pg ridyl 2-Pyridyl 3-Pyridy l 3-Pyrid yl 3-P) ridyl 3-Pyridyl
Method of cyclization
A B
C D A
B C D
4-Pyridyl A 4-Pyridyl B 4-Pyridyl C D 4-Pyridyl a Calcd. for CIgH,,X.
Yield,
%
NEW PYRIDYLAKTHRACEKES (1'1) ____-___----Analyses, M.p.. Carbon "C. Calcd.a Found
____
-
70------Hydrogen---Calcd. Found
r--Nitropen--Calcd. Found
92 Quant. 62 Quant.
163-165 163-1 65 163-165 163-165
89.38
89.25
5.13
5.31
5.49
5 61
93 Quant. 60 Quant.
197-1 98 197-198 197-198 197-198
89.38
89.52
5.13
5.26
5.49
5.40
97 Quant. 65 Quant.
199-200 199-200 199-200 109-200
89,38
89.16
5 13
5.18
5 49
5 54
Experimental8 -In 2-Benzylphenyl 2-Pyridyl Ketone.--A Grignard reagent was prepared from 21.1 g. (0.086 mole) of 2-bromodiphenylmethane and 2.1 g. (0.086 mole) of magnesium in 75 m l . of anhydrous ether. T h e Grignard reagent was stirred and a solution of 8.0 g. (0.084 mole) of 2-cyanopyridinell in 30 ml. of anhydrous ether was added dropwise over a period of 90 min.; a slight amount of heat was evolved. The resultant dark brown solution was heated under reflux for 8 hr. during which time small portions of solvent were added to prevent the stirrer from being stopped. Ttie mixture was then carefully decomposed with cold concentrated hydrocliloric acid and the two layers which formed were stirred a t room temperature for 8 hr. The organic la)-er was discarded and the aqueous portion was heated under reflux for 4 hr., cooled to room temperature, and made alkaline with dilute sodium hydroxide solution. The basic solution was extracted several times with an ethyl ether-acetone mixture, and the combined organic extracts were dried over anhydrous magnesium sulfate. The solution was concentrated and the residual oil distilled. Tlie fraction distilling between 208-211" (3 m m . ) was collected; 12 g . (595;). The distillate crystallized on cooling and on recrystallization from ethanol yielded white prisms, m.p. 61-62'. .4nal. Calcd. for CIgHljrV'O: C , 83.49; H , 5 . 5 3 . Found: C, 83.26; H , 5.65. Picrate 2-BenzvlDhenvl 3.Pvridyl Ketone.--A Grignard . . ~ of _. reagent was prepared from 37 g. (0:15 mole) of 2-bromodiphknylmethane in 100 ml. of anhydrous ether and 3.78 g . (0.16 mole) of niitgnesium. A solution of 15.5 g . (0.15 mole) of 3-cyanopyridine'? in 3 0 ml. of anhydrous ether was added dropwise over a period of 3 h r . T h e reaction was worked up as was described for the isomeric 2-pyridyl ketone. The fraction distilling between 204-207D (8 nim.) was collected; 14 g. (33%). This pale yellow oil resisted all attempts at crystallization and darkened after several days. A hot solution of 1 g. of the prior, freshly distilled ketone in 6 ml. of ethanol was added to a hot solution of 1 g. of picric acid in 3 inl. of ethanol. The solution was then cooled and theprecipit:tte which formed &-a? recrystallized from ethanol, yielding yelliiw flakes, m.p. 116-118". Anal. Calcd. for C2bHlsK\r'r0s:C , 59.76; H , 3.61; K, 11.15. Found: C , 59.83; H , 3.71; S , 11.26. Picrate of 2-Benzylphenyl 4-Pyridyl Ketone .-This compound was prepared in a manner similar t o t h a t described before for the isomeric 3-pyridyl compound. T h e Grignard reagent, prepared friini 24.7 g. (0.10 mole) of 2-hromodiphenylmethane and 2.5 g . (0.10 g.-atom) of magnesium, was allowed to react with 10 g. (0.10 mole) of 4-cyanopyridine13; the product was worked u p as ~
~
( 8 ) A l l inc-ltina points a n d boiling points are uncorrected. ( 4 ) All a n a l y s e s )yere c a r r ~ e do u t by Geller Microanalytical Laboratories. Barrlnnia. S . i l 0 ~T h c assistanw of 111. Thomas J . Delia in the preparation of both t h e analyticsl sarnl)lQsand the sanrples needed f o r carcinogenic activity testing
is rratrfiilly ackno~!I~dyerl. ( 1 1 ) L.Craig, .I. .4m.Chem. Soc., 66, 231 (1934). ( 1 2 ) f 1 , .idkins, P I a ! . , z h i d . . 66, 1291 (1914). (I:%) I ) . E 11 HYDROBROMIC ACID-CATALYZED CYCI.IZ.4TIO.V SEVERAL ICETOKES IS ACETICACIDA T 100"
THE
K Ketone
2-Benzylbenzophenone 2-Benzylphenyl 2-pyridyl ketone 2-Benq-lphenyl 3-pyridyl ketone 2-Benxylphenyl 4-pyridyl ketone
Product
9-Phenylanthracene B-(P-Pyridyl)anthracene 9-(3-Pyridyl)anthracene 9-(4-Pyridyl)anthracene
x
(lir.
OF
102 1)
0 727i0.01 2 i5
i0.02
2 G!) f 0 06
8 . 3 7 f 0 1 .i
24.50
KOTES
procedure given in ref. 4a. All reactions were performed in a stock solution similar to one previously u ~ e d . ~ , ' 4This solution was prepared by combining 700 ml. of redistilled glacial acetic acid, 166.4 ml. of redistilled hydrobromic acid, and 43.6 ml. of redistilled water. The ketone concentration was 500 X 10-5 31. The results are summarized in Table 11. ( 1 4 ) 1:. A . Vinqiello. SOC.,
J . G . Van O o t , and H. H. Hannabass, J . A m . Chem.
74, 4R4G (1Y.52).
Comments on N.m.r. Spectra of Some Optically Active Dimethylcyclohexenes
VOL. 28
contains 60% 2,4-, 38y01,3-, and 2% eso-olefin, was separated by repeated fractionations through a 90 X 0.5 cm. spinning hand column (Podbieliniakj. On the basis of rated plates of the column4 and compositions of several fractions we estimate the difference in boiling points between (+)-2,4-dimethylcyclohesene and ( +)1,3-dimethylcyclohexene to he less than l o and probably less than 0.5"; we detected no difference in boiling points during fractionation, hut gas chromatographic analyses of distillation fractions revealed the 2,4-olefin to be higher boiling. All analyses were done on a 73 ft X 0.25 in. LAC-446 on firebrick (20% by weight,) gas-liquid chromatographic column (herograph) from which the olefins emerge in the order: exo-; 1,3-; 2,4-. Table I compares sonie new and old physical constants. Proton magnetic resonance spectra of the olefins were obtained with an A-60 n.ni.r. spectrometer (Varian Associates); pertinent information is shown in Table 11.
GERARD V. SNITHASD PHILIPJ. TROTTER
Discussion Department o j Chemzstry, Illznozs Instztute of Technology, ('hzcago 16, Illznoas Kecezved January 81, 1963
We had occasion to prepare three isomeric optically active dimethylcyclohexenes: (+)-2,4-dimethylcyclohexene, (+)-1,3-dimethylcyclohexene, and (-)-3methylniethylenecyclohexane. Confusion surrounded their identities, but we offer absolute proof of structure, 1i.ni.r. proof which discloses methyl group positions on c yclohexene.
Optical rotations of our 2,4- and 1,3-olefins differ considerably from previous reports. We believe this arises from different methods of establishing purity and not from racemization; e.g., our 2,4-olefin has higher and our 1,3-olefin lower rotation than previously reported samples. We mere able to establish compositions of mixtures by gas-liquid chromatography, an aiialytical tool not available to former workers. Because of their close boiling points, mixtures of these two olefins appear never to have been separated. The proton magnetic resonance spectra offer absolute
TABLE I SOME PHYSICAL COXSTANTS OF THE CYCLOOLEF~XS Temp., QC.
Temp.,
B.P.,
Temp.,
11. x "C. Ref. 1.4442 25.5 0,802 25 130-1 750 +135.16 589 29 2,4-L)imethylcyclohexene 25 ,803 25 129 ... f112.90 579 25 h 1.4467 1.446 25 ,801 25 128 760 Racemic 25 ,806 27 124.5-125 719 91.4 589 27 d 1.4448 1.4518 20 , . . ., . 127-129 ... Racemic 25.5 ,799 25 130-131 750 34.20 589 29 1.4465 1,3-l)iniet hylcyclohexene 1.4480 25 ,807 25 127 .. . 65.36 579 25 b 1.443 25 ,798 25 137 760 Racemic f 1.4493 20 ,805 20 124.5 740 Racemic . . . . . . ... , . , 124-129 758 Racemic 0 3-?rIetIi~lniethylenec~clohexane 1 ,4429 2,5,5 ,783 25 121 750 - 4 5 . 9 3 589 29 1.4433 25 ,791 25 120.5 .. - 4x32 579 25 h This w o r k ; rotations of pure olefins were calculated from rot,ations and compositions of various binary mixtures assuming rotations are additive. For purit,ies of samples, see Table 11. * M. Mousseron, R. Richard, and R. Granger, Rd1. SOC. c h i m . France, 13, 222 "Pli!-sical Properties of Chemical Compounds," Vol. 1, R. R. Ilreishach. Ed., American Chemical Society, Washington, 11. C., (1$46), S.Siege1 and &I.Thnkel, "ildvanccs in Catalysis," Vol. IX, Academic Press, Inc., S e w York, S . Y.,1956, p . 15. 19.55, pp. 497, 499, J. E. Xickels and W. Heintzelman, J . Org. Chern., 15, 1142 J . Meinwild and R. F. Crossman, J . 9 7 n . Chem. Soc., 78, 992 (1956). Ionovan, a n d P. D. Koch, J . Ore. Chem.. 27, 4630 (19fi2). (9) ,J. A . Divon and P..I.Naro, i b i d . , 26, 2094 (1960). (10) P. A . Liebnian and B. J. Gudzinorvicz, Anal. Chem., 33, 931 (19G1). (11) P. D. Koch. R . AI. ICliss, D. V. Lopiekes. and R. 6. Wineman, J . Ore. Chem.. 2 6 , 3122 (1901).
(12) K. B. Wiberg a n d B. J. Nist, J . A m . Chem. Soc.. 8 3 , 1228 (1901). (13) E. A . LaLancette and R. E . Benson, ibid., 83, 4867 (1901). (14) L. W. Marrison, J . Chem. Soc.. 1614 (1981). ( 1 6 ) J. M . Derfer, E. E. Pickett, a n d C. E . Boord, J . .4m. Ciaem. Soc.. 71, 2482 (1949). (16) Cyclopentyl compounds a t 890 and 930 em.-'. ref. 17, 11. 31; cyclopentyl derivatives a t about 977 cni. -1, ref. 1 4 : bicyclopentyl. hicyclohexg-I,
seven dicyclopentyl-, and two dicyclohexylalkanes averaging near 895 and 936 cm. - 1 , rei. 18; tricyclopentyl- and l , ~ , . i - t r i c ~ c l o p e n t ) . l c y c l o i l r x aantr 890 and 930 cm. -1, unpublished n o r k froin this laboratory. (17) L. J . Bellamy. "The Infrared Spectra oi Complex Molecules." 2nd Ed., Methuen, London, 19.58, p.205. (18) S. V. Markova, P. A . Baahulin, A . F. Plate, and V. I. Rtanko, Opi. ond S p e c t r y . ( r S S R j (Eng. Trans].), 8 , 2GO (19GO).
SEPTEMBER, 1963 may be cyclobutyl ring deformation or additional peaks from the 1,l-disubstituted cyclohexyl rings. The latJter do not establish uneyuivocably the presence of the cyclobutyl ring without further substantiation. Since Wiberley and co-workers19 quoted variations in C-H stretching absorptions for three-, four-, five-, and six-membered alicyclics, closer examination of the 3300-2700-cm. -l region was made with lithium fluoride optics. Cyclobutyl C-H stretching bands have been assignedz0near 2985 and 2907 cm.-l for the asymmetric and symmetric vibrations, respectively. I n the spectrum of V, peaks were found at 2923 (s), 2855 (m), and 2842 (m) crn.-'. The 2923- and 2855-cm.-' bands correspond to the respective asymmetric and symmetric C-H stretching vibrations of the cyclohexyl methylene groups, but the 2842-cm.-' peak is well below the frequency expected for cyclobutyl methylene groups. In the lithium fluoride infrared spectrum of a homologous compound, dispiro [5.1.5.0]tridecane, the cyclopropyl methylene C-H asymmetric and symmetric stretching vibrations appeared 30 to 40 cm.-' lower in frequency than those assigned in the literature.I0 A similar shift may take place with the cyclobutyl C-H frequencies in dispiro [5.1.5.1Itetradecane, resulting in revised expected frequencies near 2950 and 2870 em.-'. However, heavy cyclohexyl bands partially overlap these frequency ranges. Thus the 2842-cm.-' band required further investigation and lithium fluoride infrared spectra were obtained for compounds I and 111. The dione exhibited cyclohexyl C-H bands a t 2923 and 2848 cm.-' and the monoketone a t 2928 and 2855 cm.-l with no noticeable shouldering seen near 2840 cm.-'. The 284S-cm.-' peak in the dione spectrum was intermediate between the 2855-, 2842cm.-I doublet seen in V. The appearance of roughly equal intensity peaks on either side of the unperturbed vibration indicates that a common Fermi resonance effect, may be present. It is concluded from the evidence in the 3300-2700cm.-l region that the cyclobutyl C--H stretching bands in I' are obscured by the strong cyclohexyl absorptions and that the cyclohexyl symmetric stretching absorption is complicated by Fermi resonance. X Raman spectrum of Ir was obtained for comparison with the infrared spectrum and for evidence of cyclobutyl ring-breathing absorptions. KO genuine coincidences in t'he Raman and infrared spectra were observed below 1700 cm. - I . This suggests, by the rule of mutual exclusion, that the molecule possesses a center of symmetry The over-all simplicity of the infrared spzct'rum also indicates a highly symmetrical form, considering that there are 108 vibrational degrees of freedom for the CI4Hz4hydrocarbon. Any projected rearranged products arising from this synthesis route moultl be expected to have less symmetry and to show a richer spectrum throughout the 1500-650-cm. region. If Batuev's21 ring-coupling t.reatment were followed for the cyclobutyl ring-breathing vibration, the observed peaks in the Raman spectrum of T' near 990 and 945 c m - I would give a mean frequency of 968 cm.-I. The unperturbed cyclobutane vibration is assigned a t (19) S. E. Wiberley. S. C. Bunce. a n d W. H. Bauer, Anal. Chem.. 31, 217 (1900). (20) R . P. Stein, B. S. thesis. Rensselaer Polytechnic Institute, 1958; cf. ref. 19. ( 2 1 ) 31. I. B a t u e v , Izu. Akad. N a u k S S S R , Old. K h i m . N a u k , 3 (1947).
KOTES
2453
970 cm. -l.' However, additional comparison spectra with model dispiranes are needed before such a treatment could be used for structural evidence. The nuclear magnetic resonance spectrum of V provided final conclusive evidence of the proposed dispirotetradecane structure. It has only two peaks: one a t r 8.62 corresponding to cyclohexyl protons, and one at r 8.52 showing a shift from the literature value of r 8.04 for cyclobutyl protons.l2 However, increased shielding of the latter protons by the cyclohexyl rings in this structure could reasonably cause a shift of the observed amount and direction. The homolog dispiro [5.1.5.OItridecane showed'' a chemical shift of r 10.01 for the cyclopropyl protons, compared to literature values of T 9.78.12,22The observed signal for these cyclopropyl protons was a t higher field than that for previously reported methylene protons in organic compounds, while the cyclohexyl proton signal was a t 7 8.58, close to the usual value. The sharp, symmetrical shape of the ring proton peaks in the n.m.r. of Jr follows Chamberlain's observationZ3that saturated alicyclics produce sharp bands if unsubstituted or gem-disubstituted, while lower degrees of substitution result in unresolved broad bands at 30 or 40 Rlc. The observed lack of multiplet structure from spin-spin coupling in the 40- and 60-Mc. spectra would be anticipated since the cyclobutyl protons are adjacent to quaternary carbon atoms. This absence of multiplicity also was noted in the n.m.r. spectra of 1,l'-dimethylbicyclohexy18 (60 Me.) and dispiro 15.1.5.0]tridecane1' (40 Mc.), both of which have two quaternary carbon atoms adjacent to their respective group protons. I n the spectrum of the dispirotetradecane (V), the two observed peaks were in a ratio of approximately 4.5: 1 (average of four determinations). compared to the theoretical value of 5:l. Experimental Infrared spectra were obtained with a Perkin-Elmer Model 21 double beam spectrophotometer with sodium chloride optics. Frequencies were checked with a polystyrene film reference and estimated t o be accurate and reproducible to 0.01 p . Other infrared spectra were obtained for the 2700-3300-cm.-l region with a Cary Model 14 recording spectrophotometer with lithium fluoride optics. The Raman spectra were determined on a Cary Model 81 Raman spectrophotometer on a concentrated solution of the solid in carbon tetrachloride. The n.m.r. spectra were determined on a Varian Model 1'4300B high resolution spectrometer (40 hfc.) or Model A-60 (60 Mc.) with tetramethylsilane as a n internal reference. The solid was dissolved in deuteriochloroform for examination. Vapor phase chromatograms were determined on a PerkinElmer Model 154-C vapor fractometer, equipped with a 6-ft. column of 16% m-bis(m-phenoxj7phenoxy)benzene on 35-80mesh Chromosorb ITr,operating a t 190", 15-p.s.i. inlet pressure, lOO-cc./min. helium outlet flow. Dispiro[S.l S.11tetradecan-7-one (III).2-The preparation was based on the work of Ralborsky and Buchman.2 The intermediates had the following properties: dispiro[5.1.5.1]tetradecan-7,14-dione ( I ) , m.p. 163-164', lit.2 m.p. 164-165", 58T0 yield; 14,14-bis(ethylmercapto)dispiro[5.1.5.1]tetradecan7-one (111, m.p. 58-59", lit.2 m.p. 57.5-58.3', 81%; yield; dispiro [5.1.5.1]tetradecan-7-one (111),2 m . p . 8 5 8 7 " , from acetoni(22) L. M. J a o k m a n , "Applications of S u c l e a r Magnetic Resonance in Organic C h e m i s t r y , ' ' Pergamon Press, Kew York, N. Y.,1969, p . 5 2 ; J. A . Pople. W. G . Schneider. and H. J. Bernstein, "High-resolution Xuclear Magnetic Resonance." McGraw-Hill Book Co., Inc., N e w Y o r k , N. Y., 1959, p. 236; S.Brownstein, Chem. Rev., 69, 472 (1959). (23) K.F. Chaml.trlain, A n a l . Chem.. 31, 56 (1959).
2454
NOTES
trile, lit.2 m.p. 89-90.5", 2,44initrophenylhydrazone, m.p. 183-184", lit.2 m.p. 181-182". Infrared absorption bands, in cm.?, for diketone I: 2920 (s), 2841 ( m ) , 1742 ( s ) , 1701 (w), 1445 ( m ) , 1357 (w), 1305 ( m ) , 1176 (s), 1136 (w), 1033 (w), 984 ( m ) , 915 ( w ) , 864 (w); for monoketone 111: 2920-2850 (s), 1767 ( s ) , 1724 (w), 1445 (s), 1430 ( w ) , 1346 ( w ) , 1220 (w), 1155 ( w ) , 1124 ( m ) , 1031 (w), 995-980 ( m ) , 926 ( m ) , 860 ( w ) , 840 ( w ) . The diketone I gave only a monosemicarbazone, isolated in 94% yield as tiny white needles, m.p. 210-212", from 507, ethanol. Its infrared spectrum was consistent with its assigned structure. A n a l . Calcd. for C I ~ H ~ ~ NC~, O 64.9; ~ : H , 8.3; N , 15.1. Found: C , 65.2; H, 8.6; S , 15.1. 7,7-Bis(methylmercapto)dispiro[5.1.5.1]tetradecane (IV).zA Pyrex bomb containing freshly fused zinc chloride (13.4 g., 0.098 mole), ketone I11 (24 g., 0.116 mole), anhydrous sodium sulfate (13.4 g., 0.094 mole), and methyl mercaptan (123.2 g., 2.56 moles) was allowed to stand a t room temperature for 48 hr. with occasional shaking. The excess mercaptan was allowed to boil off and the residue was extracted with ice-water and ether. The ether extract was washed with cold 57, aqueous sodium hydroxide solution, then with water. After drying over sodium sulfate, the ether was removed under reduced pressure. Crude solid, m.p. 111-114", was recrystallized from acetonitrile to give 25.1 g. (76%) of IT;, m.p. 117-119". Walborsky and BuchmanZ reported a 487, yield for this compound and a melting point of 83.5-84.5'. They did not report a sulfur analysis for their product. A n a l . Calcd. for C16H&: C , 67.6; H , 9 . 9 ; S, 22.5. Found: C , 67.4; H, 9.9; S, 22.4. Dispiro[5.1 S.l]tetradecane (V).2-The following were combined and refluxed for 48 hr.: dithioketal IT' (24 g., 0.084 mole), no. 28 Raney nickel repeatedly washed with ethanol prior to use (720 g.), and 95'; ethanol (2000 ml.). The Raney nickel was filtered off. The ethanolic filtrate was concent>rated to about half its initial volume and diluted with water. The resultant milky solution was extracted thoroughly with petroleum ether (b.p. 60-90"). The extract was dried over sodium sulfate and solvent removed by distillation, A pale yellow liquid residue resulted, which was distilled under reduced pressure. The product began to distil a t 154" (51 mm.), but, since it started to solidify in the condenser, the distillation was halted. The pot residue, which solidified on cooling, and the solid from the condenser weighed 14.0 g. (86'2 yield of V). After recrystallization from acet,onitrile the product melted from 56.5-58'. Walborsky and Burhmanz reported a 17%)yield of their hydrocarbon and different physical properties. It may be significant that their product was isolat,ed by distillation from sodium, while the use of sodium was avoided during t,he isolation of this product. A n a l . Calcd. for ClrHz4: C , 87.4; H , 12.6; mol. wt., 192; R,M(molar refractivity), 60.76. Found: C, 87.4, 87.6; H , 12.7, 12.6; mol. wt., 190, 191 (Rast method); R M ,59.50. Infrared absorption bands, in em.-': 2920-2850 (s), 1447 (s), 1430 (ni), 1340 ( w ) , 1285 ( m ) , 1245 ( w ) , 1145 (a),1067 ( w ) , !)4X (ni),!330 ( m ) , 848 ( m ) . 3-Cyclohexylspiro[cyclohexane-l,4-( 2'-pyrazolin-5'-one)] (VI). -The procedure of Barton, et d . , 5 was followed, using 10 g. of sodium (0.40 g.-atom), 500 ml. of diethylene glycol, 50 g. of anhydrous hydrazine (1.56 moles), and 30.8 g. of diketone I (0.14 mole). A total of 25.2 g. (iiC,b)of TT, glistening white plates having t: bluish luster, m.p. 174-175', was isolated from ethanol. A n a l . Calrd. for ClrHz2NYO:C , 71.8; H , 9.5; ?;, 12.0. Found: C , 72.1; H , 9 . 5 ; N , 12.0. The infrared suectrurn (KBr wllet) exhibited S-H stretrh a t 3185 mi.-' (strong intensity) and a t 3067 ern.-' (medium intensity), carbonyl stretrh a t 1695 rm.-', and broad absorption between 741 and SO0 rm.-' attributed to secondary amide N-H deformation , 1 7 To rule out the possihility that VI contained a primary amide group, i t was subjected to N-methylation according to the procedure of Loudon and OggZ4applicable to cyclic amides. The crude reaction product, a yellow oil, showed about 15 area of unchanged VI by v.p.c. A portion of the major reartion product was trapped out from the column and its infrared spectrum obtained. There was no ahsorption in the N-H stretch region, nor in the amide S-H deformation region, supporting the structure
(2.1) .I.
L).
Loudon and .J. Ogg. J . Chem. Soc., 739 (19553
VOL. 28
assignment of VI. The carbonyl absorption shifted very slightly toward higher frequency (1698 cm.-l). Dispiro[5.1S.11tetradecane-7,14-diol.-The procedure used 1)y Walborsky4 to prepare dispiro[4.1.4.1]dodec:tne-6,12-diol was adapted for the preparation, using a solution of lithium aluminuin hydride (4.7 g., 0.125 mole) in 700 ml. of anhydrous ether, and 55 g. of diketone I ( 0 . 2 4 9 mole). After recrystallization from henzene-ligroin, 37 g. (66%) of VI1 was obtained, m . p . 174-175". A n a l . Calcd. for C,,Hz,O,: C , 74.9; H , 10.8. Found: C , 75.0; H , 10.9. cis-trans isomers are possible for the diol but no effort was made to determine whether i t was the cis or trans glycol (cf. Walborsky,). However, after initial acidification of the hydrolysate, the aqueous phase had remained slightly turliid. I t was made strongly acid with 6 .V sulfuric acid, and again extr:icted with ether. This latter extract was worked up sepnrately and provided 0.2 g. of a solid melting a t 198-200", presumably a higher-melting isomer. Ditosylate of Dispiro[5.1S.11tetradecane-7,14-diol.-l'he ])r(icedure of Marvel and Sekerazj was modified and adapted for this preparation. From 20.3 g. of diol was obtained 4 7 . i g . of crude ditosylate (darkens a t 135", m.p. 210-220'). rllthough this yield was nearly quantitative, attempts to recrystallize led to considerable loss of product. From tetrahydrofuran two wops were obtained: 8.7 g. of small, white, felted needles, ni.1). 224225a (clear brown melt); and a second crop, 6.5 g. of small lustrous plates, m.p. 238-240' (clear melt, not darkening until several degrees above the melting point). The high- and low-nielting forms of the ditosylate can be accounted for on the hasis of cis-trans isomerism. No effort was made t,o identify the s1ierific stereoisomers. The yield of isolated, purified product was 15.2 g. (31%). ilnal. Calcd. for C ~ ~ H S 6 O 6 SC2,: 63.1; H , 6.8; S , 12.0. Found for low-melting form: C , 63.5; H, 6.8; S, 12.3. Found for high-nielt>ingform: C , 63.2; H , 6.8; S, 12.0.
Acknowledgment.-This work was supported by Materials Central, Aeronautical Systems Division, U. S. Air Force, under contract AF 33(600) 38448 We wish to thank Professors Richard C Lord of the Massachusetts Institute of Technology arid Wilbert H. Urry of the University of Chicago for valuable consultation and assistance, arid Messrs. J V Pustinger and T;liilliam R. Smith for assistance i n dctermination and interpretation of spectra. ( 2 5 ) C S M a r i e l and V C Sekera "Organic Synthpscs ' Coll Vol 111. John Wile1 a n d Sons Inc New York N Y 1955 p 300
Tautomerism of 5a,lla-Dehydro-7-chlorotetracycline. Preparation of 5-Alkoxy-7chloroanhydrotetracy-clines M. SCHACH vox WITTEKAC,F. A. HOCHSTEIN, A N D CHARLES R. STEPHENS Medical Research Laboratories, Chas PfLzer & Company, Inc., Groton, Connecticut Received March 4, 1963
I n 1958 McCormick and co-workers' reported the isolation of a microbial metabolit,e, which was shown to be a dehydro analog of the antibiotic chlorotetracycline2.Y( l a ) . This metabolite was produced by a blocked mutant of Streptomyces aureofaciens. On (1) J. R . D. LIcCormick, P. A . Miller, J . A . Gronicti, N. 0. Sjolander; a n d A . E'. Doerschuk, J . A m . Chem. Soc., 80, 5572 (1958). (2) l3. 11. Duggar, U . S.P a t e n t , 2 , 4 8 2 , 0 5 6 (1949). (3) C. R. Stephens, L. €1. Conover, R. Pasternack, F. A . Hoctistein, W . T. Rloreland, P . P. Regna. F. J . Pilgrim, K. J . Brunings, a n d R. B. n'oodward, J . A m . Chem. Sac., 76,3568 (1954).
SEPTEMBER, 1963
2455
NOTES
Treatment with dry hydrogen chloride in various alcohols resulted in a series of heretofore unknown tetracycline derivatives-the 5-alkoxyanhydrochlorotetracyclines* (4). These substances show ultraviolet absorption and the antimicrobial properties similar to la, R = C l b, R = H
further fermentation4 with a normal S. aureofaciens strain the metabolite was reduced to chlorotetracycline. It was thus suggested4 that the dehydro analog was a biogenetic intermediate. The structure 7-chloro-5a,1la-dehydrotetracycline (2) was assigned on the basis of the following observations. (i) The substance could be hydrogenated to a mixture of tetracycline (lb) and
OH OH 0
0 4
those of anhydrotetra~ycline.~This transformation is consistent with a starting intermediate of structure 5 which, by a simple allylic rearrangement, could generate the necessary C-5 carbonium ion.
2
its C-5a epimer. (ii) Ultraviolet absorption changes were noted in the chromophore attributable3 to the BCD ring system. (iii) An unspecified form of the metabolite showed infrared absorption a t 5.8 p. These early data define C-5a as a center of unsaturation. The present paper describes additional observations on this metabolite which demonstrate that C-5 may act as an unsaturated center. Of particular interest is a new synthetic route to 5-oxygenated tetracycline analogs. In our hands, the dehydrochlorotetracycline metabolite was initially isolated as a yellow crystalline hydrochloride that showed no indication of carbonyl absorption below 5.9 p (potassium bromide). Analytical, optical rotation, and ultraviolet absorption data on these crystals were similar to those reported earlier.l6 When the ultraviolet spectrum of the metabolite was determined in the presence of magnesium chloride, a strong shift was noted in the region above 350 mH-a phenomenon previously shown by extensive model studies6 to involve complex formation with an enolizable 11,12dicarbonyl in the tetracycline series. This, as well as the observed pK,'s of the dehydro compound (3.3, 5.0, 9.6) suggest that its principal chromophore is a readily enolizable 11,12-diketone. Strong acid treatment clearly implicated the 5position as a reactive center in the metabolite. Thus, boiling the compound in aqueous hydrochloric acid resulted in a mixture of substances with ultraviolet absorption strongly reminiscent of the apoterramycins' (3, acid degradation products of 5-hydroxytetracycline). Me W
NMel O
CONH2 "
0
OH O H 0 3
(4) J. R. D. McCormick, N. 0. Sjolander, P . A . Miller, V. Hirsch, N. A. Arnold. a n d .4. P . Doerschuk, J . A m . Chem. Soc., 80, 6460 (1958). ( 5 ) A comparison sample of the McCormick metabolite was kindly SUPplied b y D r . B. L. Hutchings of t h e American Cyanamid Co. (6) C!. L. H. Conover, Chem. SOC.(London) (Spec. Publ.), 48 (1956).
OH OH 0 0 5, R = C I 6, R = H
The tautomeric nature of dehydrochlorotetracycline also was demonstrated by the isolation of two different crystalline tautomers of the amphoteric metabolite. One form, from water, shows no resolvable carbonyl absorption below 6 (potassium bromide or dioxane) When this form was warmed in chloroform, it was converted to a ketonic tautomer, A,, 5.83 p (chloroform or potassium bromide). Because of the infrared spectrum it is attractive to assign the double bond in the ketonic tautomer of 5 to the 5,5a-position (cf. 7) rather than the previously proposed 5a,l la-position. l o Recently, Scott and Bed-
7
ford have drawn similar conclusions1] as to the position of the double bond in dehydroaureomycin. However, the n.m.r. spectrum of the ketonic tautomer in deuteriotetrahydrofuran shows no absorption in the region of 3.3-6.0 r , the position expected for nonconjugated olefinic protons, and provides no evidence for a 5,5aisomer in this solvent. The presence of a reactive center a t 5a in the metabolite 5 together with several other recent observation^^,^^ (7) F. A. Hochstein, C. R . Stephens, L. H . Conover, P. P. Regna, R . Pasternack, P. N. Gordon, F. J . Pilgrim, K. J. Brunings, and R. B. Woodward, J . A m . Chem. Soc., 7 6 . 5455 (1853). (8) T h e stereochemistry a t C-5 in these compounds has not been determined. (9) C. W. Waller, U. 5 . P a t e n t 2,744,932 (1956). (10) Cf. R . B. Turner a n d D . M. Voitle, J . Am.. Chem. Soc., 1 3 , 1403 (1951), for a detailed s t u d y of a,@-and @,-,-unsaturated forms of the model Bystem 1-acetyl-2-methylcyclohexenone. (11) A. I. Scott a n d C. T. Bedford. ibid., 84, 2271 (1962). (12) (a) D. Pereman, L. J. Heuser, J . D . Dutcher, J . WI. B a r r e t t , a n d J . A . Boska, J . Bacleriol., 80, ( 3 ) , 419 (1960); (h) J. R . D. McCormick, P . A. Miller, S.Johnson, N. .4rnold, and N. 0. Sjolander, J . d m . Chem. Soc., 84, 3023 (1962).
NOTES
2456
leads to the speculation that a common biogenetic intermediate such as 6 may explain the observed formation of both tetracycline and 5-hydroxytetracycline by various Streptomyces strains. Such an intermediate as 6 might be visualized as undergoing either fermentative reduction or hydration (peroxidation-reduction) in the fins1 step of biogenesis. We feel this route to be an attractive possibility, although the conversion of 5a,6-anhydro-5-hydroxytetracyclineto 5-hydroxytetracycline has been accomplished by an S . aureofaciens strain.lZb Experimental
hexane. T h e crystals so obtained were dried a t 60" (0.01 mm.) for 18 hr. They showed very strong infrared absorption et 5.83 p either in potassium bromide pellet or in chloroforni solution. T h e crystals decomposed indefinitely above 181 O. Anal. Calcd. for C22H21N208C1: C , 55.41; H, 4.44; N , 5.88. Found: C, 54.93; H , 4.65; X , 5.71;, , ,A (CHC1,) 366 mp (log c 3.58). T h e n.m.r. spectrum of the ketonic tautomer was measured in octadeuteriotetrahydrofuran. X o signal was observed in the region of 3.3-6 7. T h e aromatic protons werp indicated as doublets a t 2.45 and 3.1 7. T h e infrared spectrum of the solution used for n.m.r. measurements showed a strong peak a t 5.82 p . Acid Degradation of Dehydrochlorotetracyc1ine.-Ten milligrams of dehydrochlorotetracycline hydrochloride was dissolved in 5 ml. 1 hydrochloric acid and heated a t 95' for 24 min. An amorphous precipitate formed which showed an ultraviolet absorption spectrum of, , ,A 248, 320, 373 mp, reminiscent of apoterramycin. 5-Methoxy-7-chloroanhydrotetracyclines.-~40.5% solution of 7-chlorodehydrotetracycline hydrochloride was heated under reflux in 0.1 .V methanolic hydrochloric acid for 17 hr. T h e solution was concentrated under reduced pressure, filtered, and the crude product precipitated with ethyl acetate. Further purification was achieved by recrystallization from ethyl acetate. Paper chromatography showed the product t o be homogeneous and diff went from anhydrochlorotetracycline or anhydrotetracycline; , , ,A (MeOH.HC1) 229, 272, 337, 433 mp; log E 4.37, 4.58, 3.45, D ( c 0 . 2 , MeOH). 3.86; [ c Y ] ~ ~-229' Anal. Calcd. for C23H2aS208C12'/zHzO: C , 51.50; H , 4.70; X , 5.23; CI, 13.18; OCH3, 5.79; C-CH,, 2.80. Found: C , 51.29; H , 4.81; N , 5.28; C1, 15.10; OC&, 6.58; C-CHj, 2.97. 5-Ethoxy-7-chloroanhydrotetracycline.-This compound was prepared by the procedure described before, substituting et,hanol for methanol; A, (MeOH HC1) 229, 274, 335, 438 mp, log e 4.39, 4.55, 3.55, 3.83; [CX]"D -181" ( c 0 . 2 , MeOH). Anal. Calcd. for C21H2BN20&12: C , 53.24; H , 4.84; N , 5.18; OC2H5, 8.32. Found: C, 53.71; H , 5.25; N , 4.85, OC2H5, 9.1. Other 5-alkoxy-7-chloroanhydrotetracyclines, including the isopropoxy and benzyloxy derivatives, were prepared in a similar manner by heating the reaction mixture under reflux or to !OOo for several hours and were characterized by paper chromatography and by their ultraviolet absorption spectra.
Isolation of Dehydrochlorotetracycline .-A Streptomyces aureofaciens mutant was grown on a medium similar to t h a t used for the production of chlorotetracycline.2 This medium included 75 g. of cornstarch, 25 g. of corn steep liquor, and 10 ml. of soybean oil per liter, plus the usual organic salts and calcium carbonate. After a 2-day inoculum incubation, a 250-gal. tank was run for 5 days a t 26" with 12 cu. f t . of sterile air/hr./gal.; termina.1 p H , I .3. T h e broth, 157 gal., was adjusted t o p H 2 with sulfuric acid; 75 lb. of Supercel was added, filtered, and filtrate adjusted to p H 8.5 with sodium hydroxide. T h e precipitate which formed was filtered on a press and washed with water to yield 16 kg. of wet cake. This cake contained small amounts of chlortetracycline as well as dehydroa,ureomycin, as a metal complex. T h e wet broth precipitate (16 kg.) was slurried in isopropyl alcohol (17 1.). After the slurry was acidified to p H 1.9 with concentrated hydrochloric acid, sodium chloride ( 8 k g . ) , butanol (34 l . ) , and Supercel (650 g . ) were added. T h e mixture was filtered and the phases were separated. T o the aqueous phase was added the filtration residue, isopropyl alcohol (8.5 l . ) , butanol (17 I.), and sufficient concentrated hydrochloric acid t o adjust the p H to 1 . 5 . After filtration the phases were sepa.rated. The combined organic phases were concentrated under reduced pressure to a volume of 15 I . and filtered from precipitated solids. T o the filtrate were added 0.01 .V hydrochloric acid (3 1.) and hexane (34 1.). After separation of the phases the organic phase was extracted twice with 0.01 S hydrochloric acid (1.5 1.). T h e combined acid extracts were freeze dried; residue, 416 g., was dissolved in methanol (2.5 1.) and filtered from insolubles. On addition of ethylacetate ( 1 . 1 1.) crystals formed on cooling and standing for 2 days. The crystals were collected and washed with ethanol; yield, 107 g. A small sample was recrystallized from methanol-ethanol. I t s infrared spectrum wits ident,ical with t h a t of a n authentic ample.^ Anal. Calcd. for C22H?2S2C1208: C, 51.48; H, 4.32; K , 5.46; C1, 13.81. Found: C, 51.28; H, 4.58; N , 4.99; C1, 13.05. T h e ultraviolet absorption spectrum in methanol-hydrochloric acid showed A,, 254, 383 mp (log E 4.3, 3.6), while in methanolsodium hydroxide the ahsorption was shift,ed to 247, 260 (sh), 424 nip (log e 4 . 3 , 4.3, 4.6), and in methanol-magnesium chloride to 238, 268, 410 mp (log e 4.47, 4.4, 4.04). The optical rotation was determined in 0.6770 solution in 0.03 .V hydrochloric acid. The value was found t o change with time as follows: [ C X ] ~ ~ D +6.8" (15 min.); +12.5" (105 min.); + l o (19 hr.). The pK,'s of another sample of dehydrochlorotetracycline hydrochloride were determined in 0.1 ?; potassium chloride s01ution.l~ T h e values found were p a . 3.3; 4.98 & 0.10; 9.64 f 0.10; neut. equiv., 528 (calcd. 512 ) . Qualitat,ive observations of ultraviolet absorption z's. p H taken in aqueous solution (e 4.10-5 'If) indic a t e that an anion forms with increasing p H in the range p H 3.1 (A,,, 375 nip) to pH 6 . 0 (A,,, 405 m p ) ; the p H a t the inflection point is ca. 4 . 6 . This furnished a qualitat,ive corroboration of the pk', 2 value. Amphoteric Dehydrochlorotetracycline .-One gram of dehydrochlorotetracycline hydrochloride was slurried in 50 ml. of water and adjusted to p H 7.0. T h e solution so formed was filtered, tlien adjusted to pH 3.0. il light yellow crystalline solid separated (0.7 g , ) , which showed onl>- a trace of 5.8-p infrared absorption in either potassium bromide pellet or dioxane solution. This mnterinl (0.659.) was hoiled in chloroform for 4.5 hr., filtered from n trace of insoluble residue, the filtrate reduced to dryness and recrystallized from 25 ml. of chloroform containing 4 ml. of
-.
(13) P. P. Regna, I. A . Solornons, K. hIurai, A . E. Timreck, K. J. BrunInas. a n d IT. -4, Lazier. .I. A m . C h e m . SOC.,1 3 , 4211 ( 1 9 5 1 ) .
VOL. 28
Acknowledgment.-We are indebted to Rlr. E, Tynan and Dr. F. W. Tanner, Jr., for the fermentation of the dehydroaureomycin. Drs. R. L. Wagner, Jr., Kotaro Murai, and their associates provided the physical measurement data.
Quinazolines and 1,4-Benzodiazepinese XI. Synthesis and Transformations of 7-Chloro-2,3dihydro(and 2,3,4,5-tetrahydro)-5-phenyl-lH-l,4benzodiazepine' LEO H. STERNBACH, E. REEDER, A N D G. A .
ARCHER
Department of Chemzcal Research, Research Dzvzsaon, Hoffmann-La Roche, I n c , .Vutley, 'Yew Jerseu Received March do, 1968
Our interest in benzodiazepine derivatives prompted us to study methods for the synthesis of 2,3-dihydro-5(1) Paper X , L. H. Sternbach, R . I a n Fryer, 0 . Keller, \T. Metlesirs, G . Sach. and N. Steiger, J . .Wed. Chem.. 6, (3) 261 (1963). ( 2 ) The material contained in this paper a n d the synthesis of analogs 1jf I V , bearing in position 7 , a hydrogen or bromine atoni. a methyl. rarboxy. or carboniethoxy group, are described in the Hoffmann-1.a Roche Bi,lrian P a t e n t 020773 (Derwent Abstracts of F e h . 8, 1903). This aujilication also contains derivatives of I V bearing a n additional substituent in t h e phenyl ring (Z'-F, C1, O C H d a n d analogs of XIc bearing a n amino. dimerllylaniino. or cyano group in position 7 . P a r t of this material a i l 1 be described in a further communication.
KOTES
SEPTEMBER, 1963 phenyl-IH-1 ,4-benzodiazepines. Using 2-amino-5-chlorobenzophenone as starting material, we developed three methods leading to the desired 7-chlor0-2~3-dihydro-5-phenyl-1H-l,4-benzodiazepine(IV). 3 a The most useful procedure was the reduction of the corresponding benzodiazepin-2-one ( V ) 3 b with lithium alumiiium hydride4 in tetrahydrofuran, which gave a good yield of the desired product (IV). Another approach was the conversion of the aminobenzophenone I into a benzamidoethylamino or phthalimidoethylamino
A
H+ OH'
V
H,C.--CHi, -
CI
NH-CHz
'N' H -t A1CI3
I
1
I
I
C6H5
I
c1
C=N
I
N-CO \
I
I
\ c1
derivative 11, followed by hydrolysis to remove the protecting acyl group. This method was less advantageous since the intermediates of type I1 were obtained in low yield. The hydrolysis of I I a and b did not offer any particular difficulties; the initially formed aminoethylaminobenzophenone I11 cyclized spontaneously during the isolation procedure. The third method, in which the aminobenzophenone I was condensed with ethylenimine, gave only a 3.5% yield of the desired product. Compound I V could be acylated or methylated to yield the corresponding 1-substituted derivatives X I . Acid hydrolysis of IV led to the aminoethylaminobenzophenone 111, which was isolated in the form of its hydrochloride. Treatment of this compound with alkali resulted in its cyclization to IV. The lithium aluminum hydride reduction method was also applied to the 1,3,4,5-tetrahydrobenzodiazepiiione VJII,3 which was readily converted into the corresponding tetrahydrobenzodiazepine VIb; this compound was isolated as the hydrochloride, since the free base could not be obtained in crystalline form. The same compound also was obtained by lithium aluminum hydride reduction of IV. The reduction of the 1-methyl derivative VIJ3 with lithium aluminum hydride yielded mixtures16from which was obtained a crystalline product to which structure I X was assigned. This structure was proved by the preparation of IX by lithium aluminum hydride reduction of X, which was prepared by catalytic hydrogenation of VII. Experimental
H-C?
CHs
c1
2457
a N - ' > C H z YH-NH
c1
I CsH5
IX
R3 I
, XIa, CH~ RB
N-CHz
b,
c1
C,
R3=CHO = COCHs R3 = CH3
( 3 ) ( a ) T h e synthesis of the 7-nitro analog of I V b y another method was first descrihed by J . A . Hill, A . W. Johnson, a n d T. J . King, J . Chem. Soc., 1430 ( 1 9 6 1 ) ; ( b ) L . H . Sternbach a n d E. Reeder, J . Org. Chem., 26, 4936 (1961). ( 4 ) M.Uskokovik, .J. Iacobelli, and W. Wenner, ibid., 27, 3606 (1962),
ilsed the saiiie method for the reduction of 3H-1,4-benzodiazepine-2,6( I H,4H)-dione.
All melting points are corrected. T h e infrared and ultraviolet absorption spectra of starting materials and reaction products were compared in order to establish structural changes. T h e infrared spectra were determined in 3% chloroform solutions or in potassium bromide pellets using a Perkin-Elmer Model 21 spectrophotometer, and the ultraviolet absorption spectra in isopropyl alcohol or in 0.1 ll' hydrochloric acid. 7-Chloro-2 ,3-dihydro-5-phenyl-lFZ-l,4-benzodiazepine (IV) . A. From 7-Chloro-1 ,3-dihydro-5-phenyl-2FZ-l,4-benzodiazepin-2-one (V).-To a stirred suspension of 24 g. (0.63 mole) of lithium aluminum hydride in 400 ml. of tetrahydrofuran (freshly distilled over potassium hydroxide) was added during 1 hr. a solution of 86.4 g . (0.32 mole) of L' in 1200 ml. of tetrahydrofuran. The mixture was refluxed until it turned deep brown (ea. 5 min.). It was then cooled and the excess reducing agent was destroyed by the addition of wet ether. The grey suspension which formed was filtered through Hy-flo and gave a clear yellow Polution, which was dried over sodium sulfate and concentrated in vacuo. The residue was crystallized from a mixture of methylene chloride and petroleum ether7 and gave 58.6 g- . (71YG)of yellow needles or prisms melting at 172-1740, A n d . Calcd. for CI6Hl1C1X2: C. iO.17: H , 5.10. Found: ~. .. . C, 69.80; H, 5.01. B. From 2-Amino-5-chlorobenzophenone (I) and Ethylenimine.-To a cooled, stirred suspension of 8.9 g. (0.067 mole) of anhydrous aluminum chloride in 100 ml. of dry benzene was added 20 g . (0.09 mole) of I. The reaction mixture was heated to reflux temperature, then the heating was stopped, and a S O ~ U tion of 1.9 g. (0.045 mole) of ethylenimine in 25 nil. of henaene was added in small portions. The reaction mixture was stirred for about 30 min. and poured onto ice. .4fter the addition of 30 g. of potassium hydroxide, the organic layer was separated, extracted with 2 LV hydrochloric acid, and discarded. The aqueous acid ( 5 ) T h e inirared spectra of tetrahydro and dihydro derivatives showed characteristic differences in the 1600-cm. - 1 region. T h e dihydro d e r i v a t i ~ e s had a strong band a t 1612-1615 cni.-', which can be ascribed to rhr 4 , 5 . double bond, since i t is not present in the tetrahydro derivative-. ( 6 ) The infrared spectra of noncrystalline fractions obtainrd from t h ~ reaction indicated the presence of XIc. ( 7 ) T h e petroleum ether had a boiling range of 30-F0°.
2458
NOTES
layer was made alkaline and extracted with ether, and this extract was dried and concentrated in vacuo. The residue (1.7 g.) was crystallized from ether and yielded 0.8 g. (3.5%) of IV.8 C . Via N-[2-(4-chloro-2-benzoylanilino)ethyl] benzamide (IIa). -A solution of 2.3 g. of I(O.01 mole) and 2.3 g. of p-bromoethylbenzamide (0.01 mole) in 25 ml. of dimethylformamide was heated on a steam bath for 16 hr., diluted with water, and extracted with methylene chloride. The organic layer was dried and evaporated in vacuo. The residue was crystallized from a mixture of ether and petroleum ether, yielding 0.9 g. (23%) of crude react,ion product. The analysis sample was crystallized from acetone-petroleum ether and formed yellow needles of 5-[2-(4-chloro-2-benzoylanilino)ethyl]benzamide( I I a ) melting a t 143-144". Anal. Calcd. for C ~ Z H I ~ C ~ NC,~ O69.75; ~: H , 5.06; N , 7.39. Found: C, 69.46; H , 5.16; N , 7.61. A solution of 1.1 g. of I I a in a mixture of 15 ml. of concentrated hydrochloric acid and 10 ml. of ethanol was refluxed for 56 hr. The reaction mixture was diluted with water and extracted with methylene chloride to remove starting material. The aqueous layer was made alkaline with 3 ,V potassium hydroxide and extracted with methylene chloride; the extract was dried and evaporated in, vacuo. The residue was crystallized from ether and yielded 0.34 g. (45%) of crude IV.8 D . Via N- [ 2-(4-Chloro-2-benzoylanilino)ethyl] phthalimide (IIb).-A solution of 2.3 g. (0.01 mole) of I and 2.5 g. (0.01 mole) of p-bromoethylphthalimide in 30 ml. of dimethylformamide was refluxed for 16 hr., diluted with water, and extracted with methylene chloride. The organic layer was dried and concentrat,ed in Vacuo to dryness. The residue was crystallized from ether to give 0.6 g. (14%) of crude reaction product. The analysis sample was crystallized from ether and formed yellow prisms or needles of N- [2-(4-chloro-2-benzoylanilino)ethyl] phthalimide ( I I b ) melting at 171-173". Anal. Calcd. for C23H17Cl?u'203: C , 68.23; H, 4.23. Found: C, 68.03; H , 4.26. T o 30 ml. of 707, sulfuric acid, heated to 135", was added 1 g. (2.4 mmoles) of I I b , and the temperature was raised to 179". After 30 min., the solution was poured onto ice and extracted with methylene chloride to remove phthalic acid and unchanged starting material. The aqueous layer was made alkaline with 40% potassium hydroxide and extracted with methylene chloride. The organic layer was dried and concentrated in vacuo to dryness (0.5 g , ) . The residue was crystallized from ether and yielded 0.2 g. (327,) of crude IV.* Hydrochloride.-The base I V was treated with an excess of methanolic hydrogen chloride, and the salt was crystallized by the addition of ether and petroleum ether. It formed yellow prisms melting a t 245-247". Anal. Calcd. for CljHl4Cl2N2: C, 61.45; H , 4.81. Found: C, 61.78; H, 4.65. Formyl Derivative (XIa).-The mixed anhydride of formic acid and acetic acid was prepared by the addition of 6.8 ml. of 9870 formic acid to 16.4 ml. of acetic anhydride cooled in an ice bath. This mixture was heated a t 50" for 2 hr., then cooled, and added to a solution of 40 g. (0.156 mole) of 1 V in 300 ml. of methylene chloride. This solution was kept a t 25' for 17 hr. and concentrated in z'acuo. The residue was treated with aqueous ammonia and ether. The ether phase was dried and yielded crystals which, after recrystallization from a mixture of methylene chloride and petroleum ether, .gave 15 g. (34Y0) of white prisms melting a t 116-119'. ilnal. Calcd. for C16H~&1?;20: C , 67.49; H , 4.60; N , 9.84. Found: C, 67.51; H, 4.54; N , 9 . 7 2 . Acetyl Derivative X1b.-A solution of 1 g. of IV in a mixture of 15 ml. of pyridine and 10 ml. of acetic anhydride was left a t room temperature for 5 hr., and then concentrated in vacuo to dryness. The residue was crystallized from ether and then from a mixture of methylene chloride, ether, and petroleum ether, to form colorless prisms melting a t 165166' (92%). Anal. Calcd. for C 1 7 H l ~ C l S 2 0C, : 68.34; H , 5.06. Found: C, 68.41; H, 4.78. 2-(2-Aminoethylamino)-5-chlorobenzophenoneHydrochloride (HI).--A solution of 2 g . of compound IV in a mixture of 20 ml. of ethanol and 20 ml. of 2 S hydrochloric acid was refluxed for 19 hr. and then Concentrated in vacuo to dryness. The residue was crystallized from a mixture of methanol and ether to yield 0.4 g . (8) The product was identified x i t h a n authentic saniple by mixture melting point determination and coniparison of the infrared spectra.
VOL. 28
of the crude hydrochloride. After recrystallization from the same solvent mixture, the product formed yellow needles which softened a t 170" and melted a t 172-174" dec. Anal. Calcd. for CI5H1&12?rT~0:C, 57.89; H , 5.18. Found: C, 58.16; H, 5.23. Attempts to liberate the free base by treatment of the hydrochloride with alkali resulted in cyclization to IV.
7-Chloro-2,3-dihydro-l-methyl-5-phenyl-lH-l,4-benzodiazepine (XIc).-To a solution of 10.2 g. (0.04 mole) of IV in 100 ml. of dimethylformamide was added a t 50°, with stirring, 2 g. of a 53% mineral oil dispersion of sodium hydride. The mixture was cooled, and 3.6 ml. of methyl iodide added. ilfter stirring for 30 min. a t room temperature, the reaction mixture was poured into i c e w a t e r and extracted with methylene chloride. The organic layer was separated, dried, and concentrated in vacuo. The residue (8.3 g.) was dissolved in a small amount of a mixture of ether and petroleum ether (1: 1j, and adsorhed on a chromatographic column (3.5-cm. diameter) prepared with 300 g. of Woelm grade I alumina. The column was eluted first with 2.8 1. of a 50% ether-petroleum ether mixture, then with 500 ml. of a 75% ether-petroleum ether mixture, followed by 500 ml. of absolute ether. The eluates were combined and concentrated in vacuo. The residue was crystallized from a small amount of a mixture of ether and petroleum ether to yield 2.8 g. ( 2 5 9 ; ) of crude reaction product. llfter recrystallization from the same solvent mixture, or from pentane, the pure product was obtained as colorless prisms melting a t 95-97'. Anal. Calcd. for C16H15C1N2:C, 70.97; H , 5.58; N , 10.35. Found: C, 70.75; H , 5.35; K, 10.12. Further elution of the column with U.S.P. ether yielded unchanged starting material. 7-Chloro-2,3,4,5-tetrahydro-5-phenyl-lH1,4-benzodiazepine Hydrochloride (VI). A.-To 6.9 g. (0.18 mole) of lithium aluminum hydride in 150 ml. of dry tet,rahydrofuran was added a solution of 21 g. (0.077 mole) of i-chloro-1,3,4,5-tetrahydro-5phenyl-2H-1,4-benzodiazepin-2-one(VIII)3 in 300 ml. of tetrahydrofuran. The addition was carried out with stirring, a8 rapidly as the foaming permitted. The mixture was heated to reflux, then cooled, and stirred a t room temperature until the reaction subsided. I t was then refluxed for 30 min., decomposed with ethyl acetate and wet ether, and filtered over Hy-flo. The organic layer was separated, dried, and concentrated in vacuo. The oily residue was dissolved in methanol and acidified with a slight excess of methanolic hydrogen chloride. On addition of acetone, the crude hydrochloride (16.4 g., 7 2 % ) crystallized and was separated by filtration. The analysis sample was recrystallized from a mixture of methanol and acetone; it formed slightly yellow plates melting a t 259-260" dec. Anal. Cafcd. for C&&~&Z: C, 61.03; H, 5.16. Found: C, 61.01; H , 5.24. B.-To a stirred suspension of 1.5 g. (0.04 mole) of lithium aluminum hydride in 100 ml. of dry tetrahydrofuran was added a solution of 5.1 g. (0.02 mole) of 1 V in 50 ml. of tetrahydrofuran. The reaction mixture was refluxed for 3 hr., cooled, decomposed with 200 ml. of wet, etlier, and filtered over Hy-Ho. The filtrate was dried and concentrated in vucuo. The residue (5.5 g.) was crystallized from a mixture of ether and petroleum ether to yield 1.9 g. of starting material. The mother liquors were conrentrated in vucuo to dryness. The residue was dissolved in methanol, and acidified with an excess of methanolic hydrogen chloride. Ether was added, and the precipitated hydrochloride (2.2 g., 60% yield) was filtered off
7-Chloro-2,3,4,5-tetrahydro-l-methyl-5-phenyl-lH-l,4-benzodiazepine (IX). A.-A solution of i-chloro-1,3,4,5-tetrah2ldro-lmethyl-5-phenyl-2H-1,4-benzodiazepin-2-one(X)9 (40 g . ) in anhydrous tetrahydrofuran (500 ml.) was added dropwise over 2 hr. to a refluxing solution of lithium aluminum hydride (15.2 g . ) in tetrahydrofuran (500 ml.). Refluxing was continued for a further 2.5 hr.; then the excess lithium aluminum hydride was decomposed by careful addition of saturated aqueous sodium sulfate (100 ml.). The solution was then dried with anhydrous sodium sulfate and filtered. FJvaporation of tlie filtrate gave the crude product, which was dissolved in ether and extracted with 1 ?i hydrochloric acid. The acid extract was made basic with sodium hydroxide solution and extracted with methylene chloride. The extract was evaporated, and the residue was converted to (9) Compound X was prepared i n the same manner as V I I L 2 I t crystaliiaed from ether as prisms, melting a t 141-145'. Anal. Calcd. for C X H M C I K ; ~C~,: 68.57; H , 5 . 5 0 . Found: C. 6.5.8G; H. 5.27.
SEPTEMBER, 1963
KOTES
2459 ?OR
the monohydrochloride by treatment with the calculated amount of niethanolic hydrochloric acid. The hydrochloride was crystallized by the addition of ether and separated (m.p. ea. 250"). It was then reconverted to the base I X , which on recrystallization from pentane formed pale yellow needles, m.p. 60-62" (527,). A n a l . Calcd. for Cl6HI7C1N2: C, 70.44; H, 6.28. Found: C2, i0.33; H , 6.47. B .-Reduction of VI1 with lithium aluminum hydride in tetrahydrofuran, by the same procedure as t h a t used for reduction of X and purification of the product by the same methods, gave I X in 20"4 yield.* The hydrochloride was prepared from the base and methanolic hydrochloric acid, as previously described, and was obtained as colorless needles (from methanol-ether), m.p. 258-259". -4nal. Calcd. for Cl6HI8C12Np: C, 62.14; H, 5.87; X , 9.06; C1, 22.93. Found: C , 62.44; H, 6.10; Pi, 9.09; C1, 22.56. The monopicrate was prepared from the base and picric acid in ether and was obtained as yellow prisms, m.p. 202-204" (from ethanol). A n a l . Calcd. for C22H2oClNsO7: C, 52.64; H, 4.03; E, 13.95. Found: C, 52.89; H , 4.02; N, 14.17.
?OR CHaCOsH
C6H5 Ia,R=H b, R = C H a
IIa, R = H b, R = CH3 At/ihL,
Acknowledgment.--We are indebted to Dr. A. 1Votchane, Mr. S. Traiman, and Dr. V. Toome for the the infrared and ultraviolet spectra, and to Dr. ill Steyermark and his staff for the microanalyses. Mr. L. A . Dolan was helpful in the preparation of larger amounts of starting materials and intermediates.
H I
?OR
H
COCH3
Quinazolines and 1,4-Benzodiazepines. XII.' Preparation and Reactions of 2,3-Dihydro-lH-l,4benzodiazepine 4-Oxides
VI
WERNERMETLESICS,GLADYS SILVERMAN, ASD LEO H. STERKBACH C6H5
Department of Chemical Research, Research Division, Hoffmann-La Roche, Inc., Xutley, New Jersey Received March 20, 1963
I n order to obtain nitrones of type V the oxidation of the acylated 2,3-dihydrobenzodiazepines (Ia and b) with peracetic acid2 was studied. The primary oxidation products, the oxaziridines IIa and b, were obtained in good yield. They isomerized on heating to the nitrones IIIa and b which, in turn, rearranged to IIa and b on exposure of dilute solutioiis to d a ~ l i g h t . ~I n contrast to the nitrones, the oxaziridines IIa and b liberated iodine from an acidic potassium iodide solution. According to e ~ p e c t a t i o n , ~the ultraviolet spectra of the oxaziridines showed only an inflection a t ca. 238 mp, whereas the nitrones IIIa and b had maxima a t ca. 234, 260, and 310 mp. These maxima, also shown by compound V, are characteristic of coinpounds containing a nitrone function in conjugation n i t h a phenyl g r o ~ i p . ~ The acyl derivatives IIJa and b were hydrolyzed with alkali to the nitrone V, which could be reacylated to the starting materials 111. This shows that under the chosen conditions the iiitrones did not undergo any structural changes. Further proof was obtained by ( I ) Paper XI, L. H. Sternbach. E. Reeder, and G . A . Archer, J . Org. Chem.. 28, 24.56 (19631. (2) W. D. Emnions, J . Am. Chem. Soc.. 79, 5739 (1957). (3) See I,. H. S t e r n b a r h , B. -1. Koechlin, and E. Reeder, J . 0,g. Chem., 17, 4071 (1962), for coinparison and for earlier references. (4) A I . .J. Karnlet a n d L. A . Kaplan, ibid., 22, 576 (1957).
VI1
treatment of with phosphorus trichloride which gave the known diazepiiie 191. The reduction of 111 with lithium aluminum hydride did not yield 1-alkyl derivatives. Depending 011 reaction conditions, either the hydroxylaniine IV or I'I was obtained from IIIb.5 Both products could be reoxidized with mercuric oxide to give the corresponding iiitrones V and IIIb, respectively. \J
Experimental
A11 melting points are corrected. T h e infrared and ultraviolet absorption spectra of the compounds described were determined to establish structural changes. Identity was proved hy mixture melting point and comparison of infrared spectra. The ultraviolet spectra were determined in isopropyl :tlcoliol u9ing a Cary Model 14 spectrophotometer. 7-Chloro-4,5-epoxy-l-formyl-2,3,4,S-tetrahydro-S-phenyl1H1,4-benzodiazepine (IIa).-Peracetic acid w v : prepared ~ by dropwise addition of 2.3 ml. of acetic anhydride to a mixture of 3 nil. of methylene chloride, 0.6 ml. of '30% hydrogen peroxide, and 1 drop of coricentrnted sulfuric acid a t 0". This mixture was kept in an ice hatli for 15 min., a t 25" for 30 niin., and then added to a solution of 5 . 3 g. (0.019 mole) of In1 in 10 nil. of acetic acid. T h e solution was left a t 25' for 17 hr. and then made alkaline by addition of ice and itqueoua ammonia. Crystals separated which, after recrystallization froni a mixture of nieth>.lene chloride and petroleum ether, formed 3.5 g . (63%) of white prisms melting a t 150-152'. ( 5 ) Compound IIIa. on reduction with lithium aluminum hydride, also I V , which was not isolated b u t oxidized directly to V (over-all yield, C Q . 25%). gave
2460
XOTES
VOL.
28
A n a l . Calcd. for C&3C1N202: C, 63.90; H , 4.36; K, 1-Acetyl-7-chloro-2,3,4,5-tetrahydro-5-phenyl1H- 1,4-benzo9.32. Found: C, 64.24; H,4.22; S , 9 . 2 2 . diazepin-4-01 (VI).-A solution of 12.6 g. (0.04 mole) of I I I b in 7-Chloro-l-formyl-2,3-dihydro-5-phenyl-lH-l,4-benzodiaze-250 ml. of tetrahydrofuran was added a t 15' to a solution of 0.76 pine 4-Oxide (IIIa).-A sample (1 9.) of I I a was melted in an oil g. (0.02 mole) of lithium aluminum hydride in 100 ml. of tetrabath kept a t 195". After 3 min. the dark melt was cooled and, on hydrofuran. The solution was kept a t 15-20' for 1 hr., diluted addition of methylene chloride and ether, crystallized to yield with ether, decomposed with 4 ml. of water, and filtered. The 0.5 g. of crystals melting a t 132-138". Recrystallization from a filtrate was concentrated and the residue (12 9.) was dissolved in mixture of methylene chloride and hexane yielded white prisms benzene and adsorbed on a column containing 350 g. of neutral melting at 136-139". From a mixture of methylene chloride alumina (Woelm, grade I ) . Elution with a mixture of methylene and ether a dimorphic form melting a t 150-153" was obtained; chloride and ethyl acetate (1:2) gave 0.6 g. of starting material I I I b in the first fractions. Later fractions gave oils which , , ,A 233 mp ( e 21,000),, , ,A 259 mp ( e 14,000); A,, 307 mp ( e 10,000). crystallized on standing. These were recrystallized from a A n a l . Calcd. for Cl6H&1N2O2: C, 63.90; H , 4.36; N, mixture of methylene chloride and ether to give 2.3 g. (18%) of 9.32. Found: C,63.72; H, 4.62; X,9.67. white prisms VI melting a t 161-163". This product gave a mixture melting point depression with A n a l . Calcd. for Cl7K1iC1N202: C, 64.45; H , 5.41; N, I I a and could be reconverted to I r a by exposure to daylight ( 2 8.84. Found: C, 64.52; H, 5.67; N, 9.02. days) in a 1 % isopropyl alcohol solution. This product was reoxidized to I I I b in the manner described l-Acetyl-7-chloro-4,5-epoxy-2,3,4,5-tetrahydro-5-phenyl-lH- for the conversion of IV to V. 1,4-benzodiazepine (IIb).-This compound was prepared in 83y0 yield from I b in the same manner as described for the preparation Acknowledgment.--We are indebted to Dr. T'. of IIa from I a . Crystallization from ether gave colorless prisms Toome and Mr. S. Traiman for the spectrophotometric melting a t 161-163". determinations and to Dr. A1 Steyermark and his staff A n a l . Calcd. for C17H,6ClN20z:C, 64.87; H , 4.80; X , for the microanalyses. 8.00. Found: C, 65.07; H, 5.03; N, 9.01.
l-Acety!-7-chloro-2,3-dihydro-5-phenyl-lH-1,4-benzodiazepine 4-Oxide (IIIb).-Compound IIb was rearranged under conditions used for the preparation of I I I a . The product I I I b was obtained in 76% yield and formed, after crystallization from a mixture of methylene chloride and petroleum ether, colorless prisms melting a t 218-220"; A,, 234 mu ( e 21,000), A,fifl 260 mp ( E 12,000), A,, 310 ( e 12,000). Anal. Calcd. for CliH16C1N202: C, 64.87; H , 4.80; N, 8.90. Found: C, 65.17; H , 4.79; N,8.83. This product I I I b was photoisomerized in the same manner as I I I a , and the oxaziridine I I b was identified in the customary way. 7-Chloro-2,3,4,5-tetrahydro-5-phenyl-lH-1,4-benzodiazepin4-01 (IV).--A solution of 12.6 g. (0.04 mole) of IIIb in 250 ml. of tetrahydrofuran was added to a solution of 1.52 g. (0.04 mole) of lithium aluminum hydride in 100 ml. of tetrahydrofuran. The temperature of the solution rose t o 28". After stirring for 1 hr. a t room temperature, ether and 7 ml. of water were added. Filtration and evaporation of the solution gave white prisms which, after recrystallization from a mixture of ether and hexane, melted a t 167-169". T h e yield was 6.9 g. (63%). ilnnl. Calcd. for C15HljC1N20: C , 65.57; H , 5.50; N, 10.20. Found: C , 65.51; H . 5.48; N, 10.28. This compound (0.3 g.) was oxidized with 0.7 g. of mercuric oxide (30 min., 25") in a mixture of 6 ml. of acetone and 1 ml. of water. This product (0.2 9.) was identical with V. 7-Chloro-2,3-dihydro-5-phenyl-lH-l,4-benzodiazepine 4-Oxide (V).-A solution containing 30 g. (0.095 mole) of I I I b in 300 ml. of methanol and 150 ml. of 1 S aqueous sodium hydroxide was refluxed for 6 hr. Concentration of the solution gave yellow needles which, after recrystallization from methanol, melted a t 240 mp ( e 20,000), 242-215". The yield was 22.8 g. (87%);, , ,A A, 262 mp ( e 15,000), , , ,A 303 m p ( e 7000). Anal. Calcd. for C16HlaClN20: C, 66.06; H , 4.80. Found: C, 66.15; H , 4.75. The formyl derivative I I I a was hydrolyzed in the same manner. T h e following reactions,were carried out with V. A . Conversion to VI1.--A solution of 0.15 g. of V in 7 ml. of chloroforni containing 0.25 ml. of phosphorus trichloride was refluxed for 30 min. T h e mixture was cooled, poured onto ice, made basic with aqueous sodium hydroxide, and extracted with methylene chloride. Evaporation gave yellow flakes which, after recrystallization from ether, melted a t 171-173" and were found to be identical with V1J.l B. Formylation to 1IIa.-To a cooled mixture of 13.6 ml. of 98s; formic acid and 32.8 ml. acetic anhydride which had been kept a t 50' for 2 hr. was added 12.6 g. (0.046 mole) of V. A red solution formed which, after standing for 20 min. a t 25', turned yellow and was made alka!ine by addition of ice and aqueous ammonia. Extraction with methylene chloride and dlization yielded 10.5 g. (i65b) of I I l a . C. Acetylation to 1IIb.--A solution of 0.5 g . of V in 5 mi. of acetic :inhydride and 7 ml. of pyridine was kept a t 25' for 20 hr. Concentration in C ~ C I L Ogave a viscous residue which crystallized on addition of ether. Recrystallization from a mixture of methylene chloride and petroleum ether gave 0.32 g . of crystals melting a t 213-216", which were identical with I I I b .
A Novel Deoxygenation Method for Pyridine N-Oxide E D W A RE. D SCHWEIZER A N D GEORGE J. O'KEILL
Department of Chemistry, University of Delaware, S e w a r k , Delaware Received March 8 , 1963
In recent years considerable interest has been manifested in the possible methods that may be employed to deoxygenate pyridine S-oxide.' We wish to report a novel deoxygenation reaction for pyridine S-oxide using dichlorocarbene. Dichlorocarbene was prepared employing the following carbene precursors : potassium t-butoxide and chloroform,2 sodium methoxide and chloroform, sodium methoxide and methyl trichl~roacetate,~ and phenyl(trichloromethyl) m e r ~ u r y . ~ Dimethyl carbonate was found t o be present following the reaction of the dichlorocarbene prepared from sodium methoxide and chloroform with pyridine Noxide, and its presence suggests the following mechanism.6
OCClz
+
2NaOCH3
-
OC(OCH3)z
+
2NaC1
(1) (a) T. R . Emerson a n d C . W. Rees, J . Chem. Soc., 1917 (1562); (h) D. I. Relyea, P. 0. Tawney, a n d A . R. Williams, J . Org. Chem., 2 1 , 477 (15621, a n d references cited therein. (2) W. E. Doering a n d A . K. Hoffmann, J . Am. Chem. Soc., '76, 6162 (1954). (3) H. E. Winherg, J . Org. Chem., 24, 264 (1955). (4) W.E. Parham a n d E. E. Schweizer, ibid., 2 4 , 1733 (19.5'3). ( 5 ) D . Seyferth, J. 11,Rurlitch, a n d J. K. Heeren, ibid., 2 1 , 1492 (1562j: (6) Similarly, 5-diazofluorene with pyridine X o x i d e in refluxing benzene gives over 50% yield of fluorenone a n d 35% fluorenone azine, suggesting a similar carbene mechanism, in contrast to the reaction of 9-diaiofluorene with benzene alone which gives only difluorene. E. E. Schaeizer and G. .J. 0'Neill, unpublished results.
SEPTEMBER. 1963
NOTES
The yields of pyridine, when phenyl(trichloromethy1)mercury was employed as the carbene precursor, were from 63y0 (based on the carbene precursor when its mole ratio to pyridine S-oxide was 1: 3 ) to 25YG(when the ratio ~ 7 a s1 : 1). When methyl trichloroacetate' was used as the carbene precursor the yields of pyridine were 40%, 28%, and 18% (when the haloacetate to pyridine ratios were 3:1, 2:1, and 1:1, respectively). The reaction of potassium t-butoxide and chloroform with pyridine S-oxide gave trace amounts of pyridine. The usefulness of pyridine S-oxide and dimethyl sulfoxide as oxygen-donating species with a number of nonhalocarbenes6 is being explored also.
246 1
was observed when an authentic sample of carbonate was introduced with t h e benzene concentrate. T h e infrared spectrum of the benzene layer contained al1,the characteristic carbonate peaks of particular interest are the peaks a t 790 cm.-' and 1265 cm.-l, which are not found in the spectrum of methyl orthoformate, indicating t h a t the carbonate was obtained.
The Formation of Substituted Pyridines from the Reaction of the Sodium Salt of Alalononitrile with Haloforms in the Alcohol-Alkoxide System A. PAULKRAPCHO AND P. S. HUYFFER'
Experimental Materials.-Powdered sodium methoxidea and potassium tbutoxide9 are always transferred in a drybox under an atmosphere of dry nitrogen. T h e methyl trichloroacetate'o was prepared according to Dumas, b.p. 151" (atmpspheric pressure), and pyridine X-oxide,l1b . p . 120" (3 m m . ) , was distilled prior t o use. Phenyl(trichloromethy1)mercury as Carbene Precursor.Phenyl(trich1orornethyl)merc~ry~~ (17.5 g . ) , pyridine S-oxide (12.T g.), and 100 ml. of anhydrous thiophene-free benzene were refluxed with stirring for 44 hr. T h e benzene was distilled, b.p. t o 110", and treated with a saturated solution of picric acid in benzene. T h e pyridine picrate (3.12 g., 2 3 7 , ) was recovered (melting point and mixture melting point were identical with an a u t hen tic sample ) . T h e residue remaining after the benzene was distilled was concentrated to dryness. This second distillate was extracted with three 5-ml. portions of 10% hydrochloric acid and the combined extracts were made strongly basic with potassium hydroxide pellets while cooling. T h e basic solution was then extracted with ether, dried (potassium hydroxide), and distilled. Pyridine (1.41 g., 407,) was collected, b.p. 114'. I t s infrared spectrum and picrate were identical with an authentic sample. Therefore, a total yield of 63% of pyridine was obtained. T h e yields of pyridine in all the other experiments mentioned were determined as shown in the following procedure. Methyltrichloroacetate a s Carbene Precursor.-Sodium methoxide (17.6 9 . ) was added slowly from an enclosed erlenmeyer flask through a flexible tubing to a cold (1&15') solution of rnethyl trichloroacetate (18.5 9 . ) and pyridine S-oxide (49.5 9.) in 2.50 ml. of anhydrous, thiophene-free benzene. T h e mixture was stirred a t ice-water bath temperature for 1 hr. after completion of the base addition. The mixture was transferred to a rotary evaporator and evaporated t o dryness at water aspirator pressure and a temperature not exceeding 65". T h e distillate was collected in two D r y Ice-acetone traps. An aliquot (3.0025 g . ) of the benzene distillate (166.1 g . recovered) was treated with a saturated solution of picric acid in benzene. T h e resulting dried pyridine picrate (melting point and mixture melting point checked with an authentic sample) weighed 0.229 g . and represented I .96'?& by weight of the aliquot. Therefore, the pyridine ohtained was calculated to be 3.25 g. (40% yield). A similar solution of the pyridine X-oxide used was treated in the same manner employed for the work-up described previously and the benzene concentrate was shown to give no picrate of either pyridine or pyridine S-oxide. Chloroform a s the Carbene Precursor.-Sodium methoxide (17.6 g . ) , chloroform (12.3 g . ) , and pyridine N-oxide (49.5 g . ) were dlowed to react in henzene (250 ml.) solution as described previously. The work-up procedure was identiral with t h a t used in the preceding experiment and the calculated yield of pyridine was 17%. The gas chromatogram of the benzene concentrate showed R peak for dimethyl carbonate; peak enhancement s l ~ o ( 7 ) The ratio of sodium methoxide t o methyl trichloroacetate was always 3:1 in order t o absorh t l i e phosgene proposed as a n intermediate in the reaction. ( 8 ) The hlatheson Co., Inc., East Rutherford, N. J. (9) LISA Research Corp., Callery. Pa. (10) J. Dumas, A n n . Chem.. 83, 111 (1839). (11) Reilly T a r and Chemical Corp., Indianapolis 4 , I n d . (12) Prepared according t o E. E. Schwiezer and G. J. O'Neill, J . O r g . Chem.. 38, 851 (1963).
Department of Chemzstry, I;niverszty of Vermont, Burlington, Vermont Received March 20. 1963
In continuation of a study of the reaction of carbanions with haloforms in the alcohol-alkoxide system, me have investigated the reaction of the sodium salt of malononitrile and wish to report a convenient one-step synthesis of pyridines of type I. The reaction of the sodium salt of malononitrile with chloroform and sodium ethoxide in ethanol has been reported by Kotz and Zornig3 as yielding a product formulated as either I1 or 111. The questionable structural assignment rested solely on an elemental analysis, and a degradation attempt was reported as being unsuccessful. X reexamination has shown that the structure of the product is I (R = CIH,) rather than I1 or 111. This was readily established by an infrared comparison and by a mixture melting point deterniination with the pyridine prepared according to the procedure reported by Little, et aL4
I
'
1/2 H,O
I1 H
To test the general applicability of this method for the synthesis of 2-amino-3,5-dicyano-6-alkoxypyridines, several alcohols and alkoxide ions were utilized. The results of this study are recorded i n Table I. It can be seen from Table I that fairly good yields of the pyridines were obtained in all cases except with potassium t-but'oxide in t-butyl alcohol where no pyridine could be isolated from t>hereaction mixture. (1) National Defense Education Act Fellow. (2) A . P. Krapcho, P. S. Huyffer, and I . Starer, J . Oru. C h o n . , 37, 3096 (1962). (3) A . Kotz a n d W.Zornig, J . p r a k t . C l e m . , 183, 425 (1906). ( 4 ) E. L. Little, J r . , W J. Middleton, D. I).Coffnlan, V. A . Enpelhardt a n d G. N. Sausen, J . A m . C h e m . S o c . . 8 0 , 2832 (19.58).
NOTES
2462
VOL.
28
TABLE I
cyclization rate, since the material not converted into the pyridine was identified as unchanged starting ALKOXYPYRIDIKES material. Pyridine I' Reactionb CrudeC M.P.,~ Mechanism of Pyridine Formation.--From the R time yield, % OC. aforementioned results it may be concluded that the CzHs 25 min. 50 219-2206 intermediate precursor of I is the salt IV. In one run CHs 30 min. 50 253-254f utilizing sodium isopropoxide, isopropyl alcohol, malonoCHI' 4 hr. 65 253-254 nitrile, and chloroform, when the reflux period was n-C3H7 16 hr.* 35 168-16gi shortened, it was possible to isolate the sodium salt of i-CSH7 2 hr. 50 2 15-2 16' IV in a 27% yield. Similarly, the reaction using t-C4Hgk 17 hr. 01 ... potassium t-butoxide yielded 20% of the potassium a All were prepared according t o the general procedure desalt of IV. scribed in the Elxperimental section using the alcohol and alkoxide ion corresponding to the R group. Chloroform was utilized in In a previous paper we have proposed that the sodium all runs escepb .where noted. At the boiling point of the alcohol salt of tetraethylpropene - 1,1,3,3- tetracarboxylate, except where noted. The yields are based on malononitrile formed by treating the sodium salt of diethyl malonate and no attempt was made to optimize reaction conditions. with chloroform in the presence of ethanolic sodium d Melting points of pure crystallized samples. e Lit.4 m.p. 223-224", m.m.p. 219-220". f L i t . s m.p. 258-259", m.m.p. ethoxide, is probably produced by an initial carbanion 253-255'. 0 With bromoform. li Performed a t room temperaattack on dichlorocarbene.2 The salt IV probably Anal. Calcd. for CloHloture under a nitrogen atmosphere. arises by the same mechanistic route. N 4 0 : C, 59.40; H , 4.98; N, 27.78. Found: C, 59.38; H, In order to obtain evidence for the possible inter5 . 2 5 , I '4naI. Calcd. as in i. Found: C, 59.65; H, 5.12; N, mediacy of dichlorocarbene, a few experiments were 27.62. With potassium t-butoxide in t-butyl alcohol. No water-insoluble material was isolated. performed using sodium trichloroacetate.6 The thermal decomposition of sodium trichloroacetate in the According to the results of previous investigations presence of the sodium salt of malononitrile using 1,2the pyridine system (I) undoubtedly arises from the dimethoxyethane as solvent yielded none of the sodium basic cyclization of the salt of 1,1,3,3-tetracyanopropene salt of IV. However, the use of tetrahydrofuran as (IV). It has been reported that, if the potassium salt solvent yielded small amounts of IV (about 1% as ascertained by ultraviolet spectroscopy). (NC),C=CH-C( CPu')z- M + The results of the sodium trichloroacetate experiIV ments can probably be attributed to the insolubility of of IV is refluxed with a solution of aqueous methanolic the sodium salt of malononitrile in the aprotic solvents potassium hydroxide, I (R = CH,) is formed; but with added sodium trichloroacetate. This was ascerudder similar conditions with aqueous ethanol as soltained by performing the r e a h o n in tetrahydrofuran vent, 2-aminod-cyano-6-ethoxy-3-pyridinecarboxarnide using a short reflux period. The sodium salt of malonr e ~ u l t s . ~The synthesis of I (R = C2H,) has been reonitrile precipitated in a 90% yield. ported from the reaction of the salts of IV in ethanol The failure to obtain any t-butoxypyridine in the with sulfuric acid and also from the reaction of the potassium t-butoxide-t-butyl alcohol-chloroform system sodium salt of malononitrile with ethoxymethylenemight be due, to a great extent, to the reaction of the malononitrile.4 I n the latter reaction, control of the t-butoxide ion with the initially formed dichlorocarberie , temperature is important in determining the nature of with the consequent destruction of this ion making it the product. If the reaction is allowed to proceed unavailable for the cyclization step. This also is given without cooling, the pyridine I (R = CZH,) is obtained some support by the fact that the reactions of dihaloin ethanol medium, but, if the reaction is run at O", carbenes with alkoxide ions to produce carbonium ion the sodium salt of IV is formed. intermediates fall in the following order: tertiary > We have investigated the cyclization of the potassium secondary > primary.' salt of IF' under the reaction conditions utilized for the halo form-malononitrile reaction and in several cases Experimental8 have found that high yields of the corresponding alkoxypyridine (I) can be obtained. The results of these A . General Procedure from the Ha1oforms.-To a sodium alkoxide solution prepared by treating sodium ( 0 . 2 g.-atom) with experiments are recorded in Table 11. PREPARATION
AiYD
PROPERTIES
OF
2-AMINO-3,5-DICYANO-6-
TABLEI1 Pyridine Io
R
Alkoxide
Crude yield, %
Reflux period. hr.
Sodium methoxide 70 1 CH, Sodium ethoxide 70 1 CzH, Sodium isoproooxide 66 1 t-CSH, t-CaHs Potassium t-butoxide 20 17 * All were prepared according to the general procedure described in the Experimental section.
It is apparent from Table I1 that the pyridine system (I) is produced in very good yields with short reaction periods except in the potassium t-butoxide run. The low yield in this case is significant and represents a slow (5)
S.G . Cottis and H. Tieokelmann, J . Ore. Chem., 26,
79 (1961).
about 150 ml. of the dry alcohol, Eastman practical-grade malononitrile (0.1 mole), dissolved in about 30 ml. of the alcohol, was added. After the haloform (0.05 mole) was added, the mixture was heated gently with magnetic stirring, and in a few minutes a vigorous exothermic reaction occurred with precipitation of the sodium hdide. The mixture was then reflused for the desired period, filtered, and the residue washed with hot alcohol. The filtrate was concentrated on a Rinco evaporator, and the mixture was poured into ice-water. The precipitated solid was recrystallized from the corresponding alcohol. The infrared spectra of the crude and crystallized material were identical. (6) W. M . Wagner, Proc. C h e m . Soc., 229 (1959). (7) P. S. Skelland I. Starer, J . .4m. Chem. S o c . , E l , 4117 (1959).
(8) All melting points were determined on a Fisher-Johns nieltinp: point block and are uncarrected. Infrared spectra. were recorded on a l'erkioElmer Model 21 recording siiectronleter. Ultraviolrt analyses were performed on a Reckman D K recarding spectrometer. The analyses were performed by the Schwarakopf Microanalytical Laboratory and the Galbraith Laboratory.
SEPTEMBER, 1963
SOTES
2463
The pyridines had identical ultraviolet spectra in acetonitrile exhibiting :thsorption maxima a t 271 mp (log E 4.2) and 315 m p (IfJg S 1.1). l':ihle I records the pertinent data for each pyridine prepared by this Iiroredure. B. General Procedure from the Potassium Salt of 1,1,3,3Tetracyanopr0pene.-To a solution of the sodium alkoxide preIiared l)y trexting sodium metal (0.002 g.-atom) in about 20 mi. of the dry :tlrohol, the pot:tssium salt of IT9 (0.0055 mole) was :iclded. T h e solution was refluxed for the desired period and the solvent ~ V ~ Lthen S rernoved with a Rinco evaporator. Water was :tdded, and the solid whirh separated was removed b y filtration. 'I%(, infrarrd spertra of these solids were identical td those of the Iiyridines previously prep:tred. The pertinent data for each run i.s presented in T:thle 11. In the cxse of the t-butoxypyridine the same procedure was followed excaept t h a t potassium metal was substituted for scdium. The infr:ired spectr:i of this product was similar to those of the pyridine s y s t e m previously prepared; the ultraviolet spectra in aretonitrile were identical. The compound commenced t o r h m g e color a t about 215' and melted a t about 320". C. Isolation of the Salt of IV. ( a ) From Malononitrilethe reacticn compoChloroform--Sodium 1sopropoxide.-After nents werc c.ombined as in procedure A , the mixture was stirred f o r 2 hr. a t room ternper:iture and then refluxed for 15 min. The i)rrc.iliit:tted solid was filtered from the hot solution. The filtrate w:is evaporated to drj-ness with a Rinco evaporator and yielded :I solid whose infr:ired spectrum indicated the presence ( J f the salt of IT7 along with the sodium salt of malononitrile. Thc n1tr:tviolet specstrum exhibited a molar extinction coefficient of 11,000, which corresponds t o 305; of the salt of IT' (molar e ~ t i n r t i o ncmoefficient of t h r potnssium salt of IV in methanol is
Acknowledgment.-This research was supported in part by U. S. Public Health Service grant no. GM-08241-02.
:iviolet slJectrum of the s;tmple exhibited no :it)s(ir1~tionat 344 nip, indicating the absence of the sodium s:dt of I\.. I n the ( x s c of tetrahydrofurm :IS solvent the ultraviolet absiir1)tion cwrrespcinded to :thout 1 % of the s d t of IT' ( e of 420 :tt :4).
(1) P a r t V I 1 of the series "Studies in Phosphinemethylene CIIeiiiiBtry": P a r t V I , I). Seyferth, J. Ic-o 5.68, X N O ~6.45, 7.36, and hc-0 8.17, 9.02 p i n potassium bromide. .3nal. Calcd. for C6H91Y05: C , 41.16; H, 5.18; N , 8.00. Found: C , 41.08; H , 5.57; N , 7.83. The cyclic carbonnte was unchanged by refluxing with excess acetyl chloride. FVhen heated with 370 aluminum chloride a t 120" it was converted to polycarbonate. The product was taken (1.5) If. Plaut.
U . S. Patent 2.978.484; C h e m . Abstr., 66, 15934 (1961)
VOL. 28
up in acetone and precipitated with water. The colorless solid had the following absorption bands in potassium bromide: Xc=o 5.65, A x 0 2 6.43, 7.40, Xc-o 8.02 M . Calcd. for ( C G H ~ N ~ ~C,) ~41.16; : H, 5.18; S , 8.00. Found: C , 41.09; H , 5.67; N, 7.62. Partial isomerization occurred also when the cyclic carbonate was heated with excess sulfuryl chloride 0.5 hr. a t 75". 2-Methyl-2-nitrotrimethylene Sulfite.-The known sulfite16 was obtained in B%yield. It melted, resolidified, and remelted a t 110", boiled without change a t 234' (cor.), (575 mm.), and had X N O ~6.41,7.41, AS-o 8.48, hother 9.76, 10.15, 10.85, and 11.83 1.1 (in potassium bromide). The crystalline sulfite (0.09 g.) could be dissolved in 4 ml. of boiling water (92O, 570 mm.) and crystallized unchanged on cooling. Anal. Calcd. for C4H,N06S: C, 26.52; H , 3.89; N, 7.73. Found: C, 26.85; H , 4.29; N, 7.07. It was similarly unaffected by refluxing with sodium iodide in acetone and was stable toward potassium permanganate in acetone. 2-Methyl-2-nitrotrimethylene sulfite (0.10 g.) was dissolved in 100% nitric acid (2 ml.), heated a t 93" for 0.5 hr., poured on ice, and extracted with dichloromethane. The dried extracts furdininished 0.11 g. (87%) of 2-methyl-2-nitro-1,3-propanediol ~ ~ 7.42, X N O ~6.02, 7.81, 11.95 p trate," @D 1.4710; X N 6.40, in a liquid film. Authentic dinitrate, prepared by direct nitration of the diol, had n Z 5 ~1.4711, d252s1.485, and an identical infrared spectrum. l,j-Dichloro-2-methyl-2-nitropropane .-The reaction of 2methyl-2-nitro-l,3-propanediolwith sulfuryl chloride and pyridine gave 687, of crude product which on fractionation yielded about 307, of colorless liquid, b.p. 50-52" (1.5 mm.), with n Z 6 ~ 1.4730 and h o 2 6.41, 7.41 in a liquid film. ilnal. Calcd. for C4H7C12N02: C, 27.93; H , 4.10; C1, 41.23. Found: C, 27.78; H, 4.11; C1, 40.90. This compound is presumably identical with the one described in the literature.I8
(16) S. P. Lingo, U. S. Patent 2,471,274;C h e m . Abstr., 43, 6222 (1949). (17) J. 4.Wyler, U. 9. Patent 2,195,551;C h e m . Abstr., 34, 5283 (1940); W. de C. Crater, U. S. Patent 2,112,749;C h e m . Abstr., 33, 3964 (1938). (18) R. Preussmann, ArzneimitteZ-Fo7sch., 8,638 (1958).
sym-Difluorotetrachloroacetone as a Source of Chlorofluorocarbene B. FARAH A N D S. HORENSKY General Chemical Research Laboratory, Allied Chemical Corporation, Morris Township, New Jersey Received April 92, 196.9
Hine, Ketley, and Tanabe' suggested that chlorofluorocarbene was an intermediate in the reaction of dichlorofluoromethane with strong nucleophiles. Attempts by Parham and Twelves2 a t preparing l-chloro1-fluor0 bicyclo [4.1.O]heptane using dichlorofluoromethane as the carbene precursor and cyclohexene as the acceptor were only moderately successful in that very low yields of an impure product were isolated. Consequently, an efficient chlorofluorocarbene precursor and therefore, a direct route to l-chloro-l-fluorocyclopropanes is not available. We wish to report a convenient method for the preparation of chlorofluorcarbene in good yields by the reaction of sym-difluorotetrachloroacetone3 with potas(1) J. Hine, A. D. Ketley, and K. Tanabe, J. Am. Chem. SOC.,88, 1398 (1960). ( 2 ) W.E. Parham and R. R. Twelves, J . Ore. Chem.. 33, 730 (1958). (3) Allied Chemical Corporation. General Chemical Division. Morris Township. N. J.
NOTES
SEPTEMBER, 1963 sium t-butoxide4 in nonprotonic media. Pure 7chloro-7-fluorobicyclo[4.1.O]heptane (I) and l-chlorol-fluoro-2-phenyl-2-methylcyclopropane (11) were readily isolated by the introduction of cyclohexene and methylstyrene as carbene acceptors.
(I-
2495
Triphenylphosphine Oxide-Hydrogen Per oxide Adduct'
R.D.
TEMPLE,^ Y. T s c ~ oAND , J. E. LEFFLER
Department of Chemistry, Florida State University, Tallahassee. Florida
I
I1
Received M a y 3, 1963
Our observations parallel those reported by Kadaba and Edwards6 and by Cassie and Grant6 for the alkaline decomposition of hexachloroacetone. The degradation of sym-difluorotetrachloroacetone may differ from that of hexachloroacetone in that chlorofluorocarbene could be produced through a concerted elimination of chloride and fluorodichloroacetate (path A) rather than through the occurrence of a fluorodichloromethyl anion7 (path B).
A recent publication by Oswald and Guertin3 reports the preparation and isolation of several trialkylamine oxide-hydrogen peroxide adducts. This prompts us to report a similar type of compound which has been prepared and investigated in this laboratory. When a solution of triphenylphosphine or triphenylphosphine oxide in dioxane is treated with an excess of 30% hydrogen peroxide at O', a crystalline complex [(C&)3PO]z.Hz02 is obtained in high yield. CFCl2COCFC12 + BLO- + Anal. Calcd. for C36H3204P2:C, 73.21; H, 5.47; 0 P, 10.49; equiv. wt., 304; mol. wt., 608. Found: C, 73.79; H, 5.62; P, 10.52; equiv. wt., 303 CFClzAOBU + : CFCl + C1- (path A) (iodometry), 282 (gas evolution); mol. wt., 523 + 30 CFClzCOCFClz + BuO- +CFClzCOOBu + CFClz- + (cryoscopic in benzene). :CFC1 + C1- (path B) The triphenylphosphine oxide-hydrogen peroxide The infrared spectra (A,, p ) of the chlorofluorocycloadduct is very soluble in dioxane, chloroform, acetone, propanes prepared were consistent with their structures : dimethylformamide, and methanol, sparingly soluble in I, 3.48 (cyclopropane CH),* 8.98, 9.15 (CF), 10.15 benzene, ether, and chlorobenzene, and insoluble in (cyclopropane ring), 6.90, 9.52, 10.24 (cyclohexane ring), water, cyclohexane, and pentane. It melts at 132-133' and 12.58 (CCl); 11, 3.42 (cyclopropane CH), 8.60, with vigorous gas evolution. The complex is stable for 8.75, 8.90 (CF), 10.48 (cyclopropane ring) and 11.98 several months at room temperature but decomposes (CCl). The observed shift of the CCI stretching band on heating with or without solvent to give oxygen and from the 13.5-14.5-p region to shorter wave lengths is triphenylphosphine oxide. The infrared spectrum in due to interaction with fluorine in I and with both flouchloroform or Nuiol is very similar to that of triphenylrine and the aromatic ring in II.9 phosphine oxide, but an additional band is present at 3.2 p . This suggests a hydrogen-bonded complex, as Experimental found for the amine oxide adduct^.^ Although the adduct appeared to be largely associated 7-Chloro-7-fluorobicyclo [4.1.O] heptane.-A suspension of 112 g. (1.0mole) potassium tert-butoxide in 500 ml. of cyclohexene in solution, the oxygen-evolving species is probably free was cooled to 0" and treated in 1 hr. with 116 g . (0.5 mole) of hydrogen peroxide. The rate of decomposition was not sum-difluorotetrachloroacetone. The reaction mixture was reproducible, apparently being subiect to catalysis by stirred for 3 hr. a t 0-5", filtered and the filtrate frartionated to the surface of the glass reaction vessels. The half-life give 27 g. (36T0) 7-chloro-7-fluorobicyclo[4.1.O]heptane, b.p. 72-73" (38 mm.). of the adduct in chlorobenzene at 79.5' and at concenAnal. Calcd. for C,H&lF: C1, 24.9; F , 12.8. Found: trations of 0.01 to 0.07 molar varied irregularly from C1, 25.1; F , 12.6. 2000 to 3000 seconds. Copper turnings, precipitated l-Chloro-l-fluoro-Z-methyl-2-phenylcyclopropane.-A suspensilver, and manganese dioxide catalyze the decomposision of 112 g. (1.0mole) of potassium tert-butoxide in 250 ml. of tion. a-methylstyrene and 250 ml. of hexane was treated with 116 g. (0.5 mole) of difluorotetrachloroacetone. There was obtained The adduct in benzene is capable of oxidizing benz41 g. (44%) of l-chloro-l-fluoro-2-methyl-2-phenylcyclopropane, hydro1 to benzophenone and anthracene to anthrab.p. 69-70" ( 5 mm.). quinone, even in the absence of light. Warming the Anal. Calcd. for C .I.O H : 19.2: Found: .. I ~ C I FC1, , F.. 10.3. adduct with styrene gave methanol-insoluble polymer C1, 19.0; F, 10.3. and oxygen. (4) M.S.A. Chemical Corporation, Callery, Pa. ( 5 ) P. K. Kadaba and J. 0. Edwards, J , Ore. Chem., 46, 1431 (1860). (6) F. W. Grant and W. B. Cassie, ibid., 26, 1433 (1860). (7) W. E. Parham and E. E. Schm.eiser, ibid., 44, 1733 (1858). (8) W. R. Moore and H. R. Ward, ibid., 26, 2073 (1860). (8) L. J. Bellamy, "The Infrared Spectra of Complex Molecules," John Wiley and Sons. Inc., New York, N. Y., p. 331 (1858).
(1) This investigation was supported i n part under a contract with the Office of Naval Research. (2) National Science Foundation Graduate Fellow. (3) A. A. Oswald and D. L. Guertin, J . Org. Chcm., 98, 651 (1863).
