NOTES
6704
Vol. 77
was obtained 2.3 g. of material boiling above room temperature, with n% 1.2880 (reported3 for n-C3F,COCl, n Z 5 ~ 1.288). Infrared analysis showed this material to be fairly pure n-CsF?COCl. The amide, prepared by direct reaction with ammonia gas and recrystallized from CCl,, did not depress the melting point of an authentic sample of n-C3F7CONH2.6 Anal. Calcd. for CIHZFTNO:N, 6.57. Found: N, 6.64. Distillation of the filtered reaction product gave several fractions, b.p. 87-98’, which contained both unreacted (nCaFg)20 and n-CaF,CCI,. By refractive indices and infrared spectra the amounts were determined to be 6.0 and 2.8 A1C13 g., respectively. From this the ether conversion is calcuRfCFzOCFZRr (RfCCIzOCClzRr) ---+ lated as 70%, and the yields of n-CaF7COCl and n-C3F7RrCOCl RrCC1, cc13 are each 30% (based upon ether not recovered). In an otherwise identical reaction, run a t 150’ for 13 hours, the The absence of partially-reacted ethers or their conversion was 45% but the yield of n-C3F7CC13 was only cleavage products lends support to the postulation 8%; n-C3F7COCl was not recovered. Some hexachloroof the tetrachloro ether as a reaction intermediate. ethane was found in the distillation residues. Reaction of (n-C6F13)z0.-The ether (%-C6F13)20, 110 g. Although under these conditions the cyclic cy,a,aluminum chloride, 35 g. (120% of theory) were heated a‘-trichloroperfluoro ethers were stable,lb deca- and a t 230’ for 15 hours. Upon distillation of the filtered reacchlorodiethyl ether on heating splits spontaneously tion mixture there was obtained n-CaF11COC1, b.p. 87-93’, n% 1.2992 (reported6 for n-CsFIICOC1, b.p. go”, n z 5 ~ to CCl3C0Cl and C2C16.2 The reaction of ( T Z - C ~ with F ~ )aluminum ~~ chlo- 1.310) yield 20.6 g. (43%). Infrared analysis showed t o be pure n-CsFllCOCl. Anal. Calcd. for C S C ~ F ~ ~ O : ride gave known products. The acyl chloride, this C1, 10.7. Found: C1, 10.8. n-C3F7COC1,3-5 and the trichloride, n-C3F,CC13,6 The amide, recrvstallized from CCL. did not deDress the were identified by boiling points and infrared melting point of an authentic sample of ~ - C S F I ~ C O S H Z . ~ spectra. The acyl chloride n-CbFllCOC1,S from Anal. Calcd. for CGH2F11PLIO:N, 4.47; F, 66.8. Found: .4.40: --, F. 66.7. -cleavage of the ether (n-C6F1&0, was similarly h-,There was also obtained a fraction which proved t o be identified. The trichloride, 7z-CgFllCC13,is a new fairly pure n-CsF1,CCIJ: it had b.p. 140-144”, n 2 5 1.3403 ~ compound. and weighed 28.5 g. (51%). It was, however, contamiI t is now possible to characterize perfluoro ethers nated by some hexachloroethane; this was removed by means of the purification procedure previously described for the and determine their structure by cleavage followed cyclic rr,oc,or’-trichloroperfluoroethers.lb Twenty grams of by preparation of a suitable derivative of the pure n-CbFIlCC13, b.p. 143”, % Z ~ D 1.3383, was obtained. resulting perfluoroacyl chloride or chlorides. In Anal. Calcd. for C6C13Fll: C, 18.6; C1, 27.5. Found: the examples described here the amides were C, 18.6; C1, 27.5. Sixteen grams (15%) of unreacted (n-C6F13)20was reprepared. By mixed melting point determination covered. Yields are based on ether not recovered. with authentic samples of I Z - C ~ F ; C O N H ~and ~ - ~ The reaction of 25 g. of ( n - C ~ F l ~ )and ~ O 8 g. of AlC13 a t n-CgF11CONH25 these amides were shown to have 185” for 14 hours went to the extent of 77%. The yields of n-CiFl1COCI and n-C6F11CC13 were 51 and 63%, respecthe normal structure. This work verifies that the perfluoro ethers have tively.
ported. Members of a series of perfluoro (a-alkyl) cyclic ethers were treated with aluminum chloride to give simple replacement of the three alpha fluorine atoms by chlorine. A quite different result was obtained when perfluoro di-(n-alkyl) ethers were similarly treated. NO ci,ci,a’,a’-tetrachlorofhoro ethers were found, but instead the cleavage products corresponding to them were isolated, as illustrated by the equation
+
~
retained the normal structure of the hydrogencontaining ethers from which they were obtained by electrochemical f l u o r i n a t i ~ n . ~ ~ ~ Experimental Perfluoro Ethers .-The perfluoro ethers (n-C*Fs)zO and (n-CsFl3)sO used in this work had been prepared in these laboratories by electrochemical fluorination* of the corresponding ethers, ( z - C ~ H Qand ) ~ ~(n-C6H1s)s0.g Their boiling points, refractive indices and infrared spectra indicated the absence of impurities. Reactions with Aluminum Chloride .-Reactions and processing were very similar to those previously described.Ib Minor differences are noted below. Reaction of (n-C4Fs)20.-Twenty grams of the perfluoro ether ( ~ - C ~ F Q )and ? O the theoretical amount of aluminum chloride (8.0 9 . ) were heated together a t 175’ for 16 hours in a racking autoclave of 43-ml. volume. The autoclave was chilled and opened: volatile materials were caught in a liquid air trap on warming and partial evacuation. There (2) Malaguti, A n n . Chim., [31 16, 13 (1846); E. € I . Huntress, “Organic Chlorine Compounds,” John Wiley and Sons, Inc.,New York, N. Y . , 1948, p. 224. (3) Technical Bulletin, “Hepta5uorobutyric Acid,” Minnesota Mining a n d Manufacturing Company, Kew Products Division, S t . Paul 6, Minnesota, 1949. ( 4 ) D . R . Husted and A. H. Ahlbrecht, Abstracts, 116th A.C.S. Meeting, September, 1948, p. 10K. ( 5 ) A . H. Ahlbrecht, D. R . Husted, T. S. Reid a n d G. H . Smith, Jr., THISJoWnN.4L, 78, in press (1956). (6) G. V. D. Tiers, H. A. Brown a n d T. S. Reid, ibid., 76, 5978 (1953). (7) J . H. Simons, “Fluorine Chemistry,” Academic Press, Inc., New York, S . V.,1950, p . 414 ( 8 ) J . H . SirngJna, U . S. Patent 2,509,383 (March 4, 1950).
- >
Acknowledgment.-I am indebted to Dr. H. E. Freier for the analytical data, to Dr. W. E. Keiser and Dr. J. McBrady for the infrared spectra, and to Mr. J. D. Keating for assistance with the autoclave reactions.
COXTRIBUTION S o . 100 FROM THE CENTRAL RESEARCH DEPARTMENT MINNESOTA MINING& MANUFACTURING COMPANY ST.PAUL6, MINNESOTA
The Chemistry of Perfluoro Ethers. 111. Synthesis of wTrichloromethylperfhoroacy1 Chlorides by Cleavage of Cyclic Perfluoro Ethers’ BY GEORGEVANDYKETIERS RECEIVED SEPTEMBER 8, 1955
PerAuoro (0-alkyl) cyclic ethers have been found to react with aluminum chloride to give simple substitution of the three a-fluorine atoms by chlorine.2a Alpha substitution also occurred with perfluor0 di-(n-alkyl) ethers, but was then followed by cleavage of the ether.2b This study has now been extended to two cyclic perfluoroethers which do not bear an a-perfluoroalkyl group. These ethers split very much as had (1) Presented a t t h e 126th hIeeting of t h e American Chemical Society, New York, N. Y.,1954, Abstracts, p. 27-M. ( 2 ) (a) P a r t I. G. V. D. Tiers, THISJOURNAL 77, 4837 (1955), (b) P a r t 11, t b z d , 77, 6703 (1955).
Dec. 20, 1955
NOTES
the non-cyclic ethers, but the cleavage product contains both the trichloromethyl and the acyl chloride groups, as illustrated by the equations
properties with the cyclo-CaFsO used in this study. An extremely low yield was obtained in the case of cyclo-CbFloO, but this was attributed to increased difficulty in fluorinating the tetrachloro ether cyclo-C~C14F~O.Henne and Richter commented that extensive decomposition occurred by way of breaking the ring, although decomposition products were not described: certainly the work reported in the present paper supports this view.
YFZ-CF:)~
2[YF2-CCL>
CFz-CFp
]
o +
CF2-CClp CCl,(CF2)2COCl CFz-CClz
CF20 CS-CFz
6705
Experimental Perfluoro Ethers.-Perfluorotetrahydrofuran and perFrom the absence of symmetrical products such fluorotetrahydropyran were prepared by electrochemical as CC13(CF&X13 or ClCO(CF2)ZCOCl i t is rea- fluorination of the corresponding cyclic ethers.@-8 The soned that the chlorine shift (which accompanies boiling point of cyclo-C4F8O was in agreement with prereported values.oj8 In the case of perfluorotetrasplitting) is intramolecular. The ease of splitting viously hydropyran, physical constants have been reported for a of these ethers is in marked contrast to the stability highly purified samples: the boiling point, 34.0”, is someof the CY, CY, a‘-trichloroperfluorocyclic ethers,2a and what higher than earlier values.6p8 The ether sample used demonstrates that the mere presence of a five- or in this work was almost pure, as judged by its refractive index and its infrared spectrum: the isomeric cyclic ether, persix-membered ring is not responsible for resistance fluoro-2-methyl-tetrahydrofuran@ was not present. to cleavage. 4,4,4-TrichlorotetrafluorobutyryI Chloride.-PerfluoroThe w-trichloromethylperfluoroacyl chlorides are tetrahydrofuran, 21 .6 g. (0.10 mole) and aluminum chloride, new compounds. They have been characterized 13.3 g. (0.10 mole), were heated together a t 170’ for 13 hours in a 43-ml. rocking autoclave.2 Volatile products by their physical properties and infrared spectra, were caught in a liquid air trap. After crude fractionation and by conversion to their amides. They may be in a vacuum system, 9.1 g. of unreacted cyclo-C4F80 was reused to identify or to prove the structure of the covered and identified by molecular weight and infrared parent ether, or alternatively as synthetic inter- spectrum. The conversion was 57% based on cyclo-C,F80. (Note: insufficient AlCll was used.) A filtered solution of the mediates. An example of the latter is the prepara- reaction product in C ~ F I ~was O distilled to give 4,4,4-trition of the new trichlorotrifluoropropene CCI8CF= chlorotetrafluorobutyryl chloride, b.p. 148’, n Z 5 1.4105, ~ CF2, according to the scheme yield 8.6 g. (55%) based on unrecovered perfluoro ether. Pot residues contained hexachloroethane. Anal. Calcd. HzO for C4CldF40: C1,50.3; C1 (hydr.), 12.6. Found: Cl, 50.1; CC13CF&2FzCOCl C1 (hydr.), 12.3. The amide was prepared by reaction with NaOH ammonia gas, and was recrystallized from CClr, m.p., 280” 126.5-127.0’ (cor.), not altered by further recrystallization. CC18CF2CF2C02Na -+CC13CF=CF2 Anal. Calcd. for C I H Z C ~ ~ F ~ KC1, O : 40.5; F, 29.0; N, Found: C1, 40.5; F, 25.7; N, 5.33. The thermal decomposition of sodium salts of 5.34. 5,5,5-TrichlorohexafluoropentanoylChloride .-Perfluoroperfluorocarboxylic acids to give perfluoroolefins tetrahydropyran, 19.8 g. (0.075 mole), and aluminum chlohas been reported previously. ride, 13.3 g. (0.10 mole), reacted a t 180’ for 13 hours. The proposed structure, CF2=CF-CC13, is A filtered solution of the reaction product in CeF120was confirmed by examination of the infrared spectrum. fractionated and 6.0 g. of unreacted cyclo-CsFloO was reand identified by infrared analysis. From this the The carbon-carbon double bond absorption is found covered conversion is estimated as 70%. As in the preceding exa t 5.63 p, as compared with 5.56 p for C3FB.4a ample, there was little or no intermediate fraction, but pot A shift to longer wave length has been found on residues contained hexachloroethane. There was obtained chloride, b.p. 165’, @D passing from CCl,=CCI-CF, (6.30 p ) to CC12= 5,5,5-trichlorohexafluoropentanoyl 1.3966, yield 9.3 g. (54%), based on unrecovered perfluoro CC1-CC1, (6.46 ,IL).~ These compounds also serve to ether. Anal. Calcd. for CJCIIFGO:C1, 42.7; C1 (hydr.), illustrate the pronounced shift to longer wave 10.7; F,34.4. Found: C1, 43.0; Cl(hydr.), 10.4; F, 33.9. lengthiwhich accompanies replacement of vinylic The amide, recrystallized from CC14, had m.p. 138.3fluorine by chlorine in perhaloolefins. This shift is 138.4’ (cor.) not altered by further recrystallization, Calcd. for C ~ H Z C ~ ~ F &c, O :19.2; N, 4.49. Found: 0.20-0.25 p per substitution, and has been found to Anal. C, 19.2; N, 4.61. apply also to the substitution of vinylic fluorine by a 4,4,4-TrichlorotetrafluorobutyricAcid.-The acyl chloride CClaCF2CF~COC1,30.0g. (0.106 mole), was mixed with perfluoroalkyl group. 4b The a,a,a’,a’-tetrachloro ethers, postulated as water, 1.91 g. (0.106 mole), in a narrow-necked flask and to stand until about the theoretical weight loss (as intermediates in the cleavage reaction, are them- allowed HC1 gas) had occurred. The reaction mixture was distilled selves known compounds. They have been syn- to give pure 4,4,4-trichlorotetrafluorobutyric acid, b.p. thesized by chlorination of the corresponding 212’, n Z 6 D1.4132, neutral equivalent 262,264 (calcd. 263.5), fluorohydro ethers, and have been treated with yield 23.0 g. (81%). The sodium salt was prepared and was a t 110”. Anal. Calcd. for C4C13F4Ka02:F , 26.6; antimony trifluorodichloride to achieve the syn- dried Na, 8.08. Found: F, 26.2; Na, 8.26. thesis of the perfluoro cyclic ethers6 In the case 3,3,3-Trichlorotrifluoropropene.-Following the general of cyclo-C~FsO,H e m e and Richter obtained in fair procedure of LaZerte, et U Z . , ~ the sodium salt, CC1,CFIyield a product which compared closely in physical CF2C02Na, 13 9 g. (0.049 mole), was found to decompose a t 250-280’ in mcuo. Volatile products were fractionally (3) J. D.LaZerte, L. J. Hals, T . S. Reid and G . H . Smith, TEIS distilled to give 3,3,3-trichlorotrifluoropropene,b.p. 89JOURNAL, 711, 4525 (1953). 91”, @D 1.4027, yield 4.8 g. (50%). Anal. Calcd. for (4) D. C.Smith, el al., “Spectroscopic Properties of Fluorocarbons CsC13Fa: F, 28.6. Found: F, 28.7. CC&(CR)3COCl
___f
and Fluorinated Hydrocarbons,” NRL Report 3567, Naval Research Laboratory, Washington, D. C.,1949; (a) p. 63 and p. 126, (b) p. 142. (5) D. C. Smith, et ai., “Infrared Spectra of Fluorinated Hydrocarbons,” NRL Report 3924, Naval Research Laboratory, Washington, D. C.,1952, p. 63 and p. 68. (6) A. L. Henne and S . B. Richter, THISJOURNAL, 74, 5420 (19.52).
(7) J. H. Simons, “Fluorine Chemistry,” Academic Press, Inc., New York, N . Y.,1950, p. 414. (8) E. A. Kauck and J. H . Simons, U.S. Patent 2,594,572 (April 29, 1952). (9) T. J. Brice and R. I. COOP, THISJOURNAL, 7 5 , 2921 (1953).
6706
NOTEs
Vol. 77
Acknowledgment.-Mr. H. L. Anderson carried out the preparation of the 3,3,3-trichlorotrifluoropropene. I thank Dr. P. Itr. Trott for the pure isomer of cyclo-C6F100 used in this study. I am indebted to Dr. H . E. Freier for the analytical data, to Dr. Ti’. E. Keiser for the infrared spectra, and to Mr. J. D. Keating for assistance with the autoclave reactions.
thate of czs-2-phenylcyclohexanol.8 This sample of 3-phenylcyclohexene gave an infrared spectrum almost identical with that from the reaction of 3bromocyclohexene with phenylmagnesium bromide.9 Reduction of 3-phenylcyclohexene gave phenylcyclohexane identical with a sample obtained by reduction of 1-phenylcyclohexene. I t is evident from the ultraviolet absorption spectrum of the sample of 3-phenylcyclohexene obtained by CONTRIBUTION KO.101 FROM THE the xanthate method that no more than a small CENTRALRESEARCH DEPARTMEXT MINSESOTA MINIXGAXD MANUFACTURING COMPAYY percentage of 1-phenylcyclohexene can be present. ST.PAUL6, MINNESOTA These results support the structure assignments and analyses of Alexander and iNudrak.8 The extent of rearrangement of 3-phenyl- to 1Elimination Reactions in Cyclic Systems. 111. Mechanism of the Formation of 1-Phenylcyclo- phenylcyclohexene on treatment with bases was hexene from Trimethyl-trans-2-phenylcyclohexyl- followed by means of changes in refractive index. Under the conditions used by Arnold and Richardammonium Hydroxide son’ for the Hofmann reaction no rearrangement BY JOSEPH WEINSTOCK AND F. G. BORDWELL was observed. RECEIVED SEPTEMBER 1, 1965 I n order to be sure that the non-homogeneous nature of our reaction mixture was not responsible for In a recent publication‘ it was shown that the the failure of rearrangement, 3-phenylcyclohexene product from a Hofmann degradation of trimethyl- was heated with potassium hydroxide in alcohol trans-2-phenylcyclohexylammonium hydroxide was solution a t 100’. After one-fourth hour the extent practically pure 1-phenylcyclohexene. Rather than of rearrangement was 0%; after 24 hr., 4.5%; assume that this product resulted from a direct cis after 96 hr., 9.8%; after 192 hr., 14.2%. elimination and that the reaction constituted an A slow rate of isomerization for a similar system exception to the general rule that trans eliminations was observed by Quelet lo for the base-catalyzed are favored,?the authors postulated that 3-phenyl- transformation of allyl-9-bromobenzene to propenyl-p-bromobenzene. Refluxing for 24 hours ’/ \ with a solution of potassium ethoxide in ethanol caused 40y0 isomerization; seven hours reflux with potassium amyloxide in amyl alcohol gave 75% isomerization. The slower rate of isomerization observed by us for 3-phenylcyclohexene is consistent cyclohexene was first formed in the reaction, by a with the fact the ionization of a hydrogen in an trans elimination, and that this initial product was open-chain system, exemplified by ethyl acetoacerearranged under the reaction conditions to 1- tate, is about one hundred times as fast as in a phenylcyclohexene. comparable cyclic system, exemplified by 2-carbOur investigations of elimination reactions in ethoxycyc1ohexanone.l’ 1,4-Pentadiene also apcyclic system^^-^ suggested that the increased acid- pears to be stable to alkali since it is formed from ity of the hydrogen attached to the carbon holding 1,5-diiodo- or 1,5-dichloropentane by treatment the phenyl group might be sufficiently great to with alcoholic potassium hydroxide.’? allow a cis elimination involving this hydrogen to The possibility that traces of silver oxide might occur more rapidly (giving 1-phenylcyclohexene) in some way cause rearrangement of 3-phenylcyclothan the trans elimination involving a relatively hexene to 1-phenylcyclohexene also was eliminated, non-acidic hydrogen (giving 3-phenylcyclohexene), since 3-phenylcyclohexene was found to be unwhich was suggested.’ To test this hypothesis affected by heating with a suspension of silver oxwe have prepared 3-phenylcyclohexene and investi- ide under the reaction conditions. gated its tendency to rearrange in the presence of I t is highly probable that 1-phenylcyclohexene is base. formed in the Hofmann degradation‘ by a cis elimiAn attempt was made a t first to obtain 3-phen- nation, as indicated in the equation. This is then ylcyclohexene from trans-2-phenylcyclohexanol by another example wherein increased acidity of the the method of Price and Karabinos.6 Since this hydrogen being eliminated promotes a cis eliminamethod was found to give a mixture of product^,^ tion in preference to a possible trans elimination. 3--6 it was abandoned in favor of pyrolysis of the xanExperimental (1) R. T. Arnold and P. N. Richardson, THISJOURNAL, 76, 3649 (1954). (2) See t h e discussion in “Structure and Mechanism of Organic Chemistry,” b y C. K. Ingold, Cornel1 University Press, Ithaca, N. Y . , 1953, Chapter V I I I . (3) J . Weinstock, R. G. Pearson and F. G. Bordwell. THISJOURNAL, 76, 4748 (1954). (4) F. G. Bordwell and R. J . Kern, ibid., 7 7 , 1141 (1955). ( 5 ) F. G. Bordwell and M . L. Peterson, i b t d . , 7 7 , 1145 (1955). (6) C. C. Price and J . V. Karabinos, i b i d . , 62, 1159 (1940). (7) See C. J. Collins a n d H . J. Schaeffer, Abstracts of Minneapolis Meeting of t h e American Chemical Society, September, 1955, p. 5 6 - 0 ; C . C . Price, J . A. XlcCov and E. Eliel, ibid., p. 5 5 - 0
Materials.-cis-2-Phenylcyclohexanol was prepared by the method of Price and Karabinose as modified by Berti.l3 (8) E. R. Alexander and A. Mudrak, THISJOURNAL, 72, 1810 (1950). (9) A. Berlande, Bull. sac. chim., [ 5 ] 9, 644 (1942). W e wish t o thank Dr. C. J. Collins for providing us with samples of 3-phenylcyclohexene made by this method. (IO) R. Quelet, Bull. SOC. chim., [ 4 ] 46, 80 (1929). (11) R. G. Pearson and R. L. Dillon, THISJOURNAL, 76, 2439 (1953). (12) N. Demjanow a n d M . Dojarenko, Be?., 40, 2589 (1907); H. B. Dykstra, J. F. Lewis and C. E. Boord, THISJOURNAL, 62, 3396 (1930). (13) G. Berti, T H I SJOURNAL, 76, 1213 (1954).
