further chlorination A gave only CH&FKClrCH1 and CHICF&2HC1CH*Cll while B gave CHaCFsCHSCHCl,, then CH&F&HtCC& almost exchSively. These results show that the chlorination is directed away from an alpha CHa group, and tends to accumulate the chlorine atoms on the same carbon atoms. The failure of CH&1CF2CHzCHa to be chlorinated to CHC12CFtCHrCHa confirms the direction awav from a terminal gmup alpha to A CF2 group.
[CONTRIETJTION FROM
THE
Chlorine reactivity tests show that alpha chlorine atoms are immobilized and that beta chlorine atoms are repressed, while gamma chlorine atoms seem to be unaffected. COLUMBUS, OHIO
RECEIVED b h Y
7,lw"
(14) TW nvnlucript WM originaUy d v d on sep~llllkr14. 1943. aad after aumhtiion by the editodd board w u .eaptdfar publiatioa in Tma JOW~NA:.. It was, however, r e f a d to the NatiahPI Defenae R-h Committee and at their request waa withheld from pubtiation, in a m6dential file. until clarsncc was m t e d om May 6, 1946.
DBPARTMENT OF CWSMXSTRY AT THEOkio STATE UNIVERSITY 1
The Synthesis and Directed Chlorination of
1,1,1,-Trifluorobutane
BY ALBERTL. HENNEAND JAMES €3. HINKAMP
To find out how far a CFa group would extend its influence, trifluorobutane, CFsCHsCH2CHaI was synthesized and subjected to the action of chlorine. The synthesis probiem consisted in devising a practical method for obtaining an intermediate compound, CCldHCHSCH,, from which the desired CFaCHL!HtCHs was made by variations of the procedures previously used to synthesize CFICH~CHII.~ The chlorination as well as the isolation and identification of the chlorinated compounds were performed in a manner previously described. l l 2
Synthesis of CCh=CHCH2CHs The sequence CHOCHtCHPCH,, to CHCltCH2CHsCHa, to CHCl=CHCHzCHa, to CHC12CHClCHtCHa, to C C l d H C H & H a , was first tried. The transformation of butyraldehyde into its dichloride by means of phosphorus pentachloride gave only a 23% yield, which was finally brought up to 39% by a series of time-consuming improvements. Since these improvements called for hydrolysis of the reaction product at Oo, the risk of handling an incompletely hydrolyzed material prevented their application to larger quantities. The dehydrohaiogenation to CHCl= CHCH2CHsrequired a saturated solution of potassium hydroxide in boilin butanol, and gave 64'% of olefin together with 2 9 h o recovered dichloride. For large quantities it was deeA.iedsafer and more economical to start from technical CHsClCHC1CH2CHa. Dehydrohalogenation occurred easily with a 30% solution of sodium hydroxide in denatured alcohol and formed only two of the possible three olefins. The reaction was suaciently exothermicto maintain itself and it yielded 27% of CHt=CClCHtCHa, 46% of the desired CHC1=CHCHrCHs and 16% of recovered dichloride. Since CHFCClCH2CHa was easily transformed into CHSCFKH&HI used in the preceding study,2 this procedure proved to be very (1) Henne and Whalry, Txxs JOURNAL, U, 1157 (1942) (2) Hcnne and Hinkamp. ibid.. CT, 1194 (1946).
satisfactory. The addition of chlorine which synthesizes CHClsCHClCHpCHs was performed in the dark to prevent substitution. Pilot tests showed that water would greatly accelerate the rate of addition, but at the expense of increased substitution. When time matters little, it is best to operate without water, at about loo, and with a slight excess of olefin, for a net yield of 70-75%. In the presence of water, the reaction proceeds nearly four times faster, but the net yield drops to approximately 55%. The dehydrohalogenation of CHClrCHClCHsCHI proceeded very easily with either a ueous or alcoholic sodium hydroxide, and gave 98% of the desired CC12=CHCHnCHA.
Synthesis of CFsCH2CH2CHa At room temperature, C C l d H C H n C H , does not react with hydrogen fluoride. When the temperature is raised, addition takes place to form CFClsCHSCH&Ha, then substitution occurs to yield CFnClCH2CHsCHS. With longer reaction periods, higher temperatures, and a greater excess of hydrogen fluoride, the formation of CFtClCHzCH2CHs is progressively enhanced. A three mole quantity of CClr==CHCH&H8 was mixed with 24 moles of hydrogen fluoride in a steel container surmounted by a one meter length of half-inch steel pipe bearing a pressure gage and a releasing needle valve. The top twenty centimeters of the pipe were surrounded by a jacket which could be filled with Dry Ice. The reaction was started by puttiiig the cylinder in a hot water bath, its progress being shown by an increase in pressure due to the formation of hydrogen chloride. With Dry Ice in its container to ensure an efficient aephlegmation, the hydrogen chloride was slowly released into a water absorber by operating the needle valve so as to maintain the pressure between 17 and 20 atmospheres. When the pressure no longer increased, the reaction vessel was cooled in Dry Ice, and its contents were poured onto cracked ice, then worked up by
1198
ALBERTL. HENNEAND JAMES B. HINKAMP
steam distillation followed by fractional distillation. Comparative data are :
Vol. 67
hydrofluorinated product boil only 1.6' apart and their separation from an incompletely reacted mixture is exceedingly impractical. Reaction temperature, 'C. 65 100 In the various hydrofluorinations, it was asTime in hours 18 6 CF,ClCH&H&!H, formed, % ' 5.3 49.6 sumed that addition took place in strict accordCFC12CHpCH2CHsformed, % 33.0 10.3 ance with the rule of Markownikow. This asRecovered dichloride, % ' 38.0 5.3 was verified by subjecting both CFC12sumption 10.6 19 4 Tar, ? '$ CHzCH2CH3 and CF2C1CH2CH2CH3to halogen Material accounted for, yo 90 87 exchange with mercuric fluoride in the manner The dehydrohalogenation of CFZC~CHZCHZCH~ previously described. The resulting trifluoride to CF~=GHCH~CHS took place with ease but re- proved to be identical with the sample prepared quired precise experimental conditions to ensure by hydrogen fluoride addition to a double bond. good yields, because of the low boiling points of the Chlorination reagent and of the reaction product. In a three-liter, three-necked flask equipped In sunlight and in the presence of water, with a mercury-sealed stirrer, a dropping funnel C F S C H ~ C H ~ Cgave H ~ only two of the possible and a twelve-bulb reflux condenser, 180 g. (3.2 three monochlorides. CF3CH2CHC1CHa and moles) of potassium hydroxide was stirred with CFaCH2CH2CHzCl were formed in a ratio of 4 5 , 340 ml. of gently boiling absolute ethanol. This is while CF3CHClCH2CHs was completely absent, more than enough hydroxide to make a saturated As in previous cases1S2the group alpha to the solution. Over a period of three hours, 196 g. fluorinated cluster resisted chlorination; the beta (1.52 moles) of CFzClCHzCHzCHs was added and gamma groups were affected in a manner dropwise. The operating skill consists in adding which deviated but slightly from random chlorinathe halide a t exactly the rate at which CFF tion, in favor of the beta group. CHCH2CH3is produced and removed through the Illustrating again the previously observed tenreflux condenser to be liquefied in a Dry Ice cooled dency of the chlorine atoms to accumulate on the receiver. When the halide is added too rapidly, same carbon, further chlorination of CFsCH2it accumulates and automatically lowers the tem- CHClCHs yielded 8 parts of CF3CH2CCl2CHat o perature of the reaction to the boiling point of the 6 parts of CF~CHZCHC~CHZC~ while further chloazeotropic mixture of the halide with ethanol. rination of CF3CH2CH2CH2Clgave 2 parts of At this lower temperature dehydrohalogenation CF&!HzCH&HCl2 to 1 part of CF3CH2CHClis exceedingly slow. The yield obtained was CH2C1. When CFSCH2CH2CHC12 was chlorin83,39&, of pure difluoroolefin, and in addition 6% ated, the product far predominant was found to of unreacted reagent was recovered. be CF3CHZCH2CCl3. The hydrofluorination of C F d H C H 2 C H S Identifications to CF&!HsCH2CHs is a quantitative, highly exoWhile the experimental work was in progress, thermic reaction. One mole of the olefin was mixed with four moles of hydrogen fluoride a t the identities of the chlorides were postulated -80' in a steel container. The container was from the agreement between the observed boiling sealed, then allowed to warm up to room tempera- points and those predicted by analogies with ture. The reaction started spontaneously and propane derivatives previously studied. As boilproceded to completion within a few seconds, caus- ing points of isomers differ markedly (see table of ing the pressure to rise temporarily to 16 atmos- physical constants) and can be predicted within pheres. The reaction mixture was left to cool one or two degrees, the method is of considerable spontaneously to room temperature, after which help. The formulas were, however, verified as it was placed in a water-bath held a t 40'. The follows : Fluorination of the compound thought to be reaction products were then released through a water washer followed by an alkali washer, dried CF3CH2CC12CH3 yielded a pentduoride which over calcium chloride and condensed in an ice- proved identical with that obtained previously cooled receiver protected against losses by a trap from CClaCH2CF2CKa2and therefore had to be The yields were from 96 to 99% CFZCH2CFzCH3. The position of both chlorine held a t -80'. atoms in the starting material had thus been of pure CFaCH2CH2CH3. correctly postulated. Additional quantities of CFZCICHZCHZCH~ The monochloride from which CFsCHzCC12CH3 were obtained by working up the intermediate CFC12CH2CHzCH3. The latter was dehydrohalo- was obtained must perforce have been CF3CH2genated in 90% yield to CFCldHCHzCH3 CHClGH,. Its isomer must have been CFsCH2by means of a saturated solution of potassium CHzCHzCl because it boiled higher, and also behydroxide in boiling ethanol. To obtain CF2- cause it proved different from CF3CHClCH2CH3 ClCHZCHzCHs, the hydrofluorination of the olefin which was synthesized for the purpose of identiwas carried out substantially as above, except fication, as follows. Chlorine was added to ~ yield CF2ClCHC1CH2CH3. that heating a t 6 5 O for eight hours was needed to C F F C H ~ C H ~ C Hto complete the addition. This is a detail of much The latter was fluorinated with mercuric fluoride, practical importance because the olefin and its (3) Henne and Flanagan, T E I SJUURNAL, 66, 2302 (1843).
J ~ Y1945 ,
SYNTHESIS AND DIREXTED CHLORINATION OF 1,1, 1-TRIFLUOROBUTANE1199
TABLE I PEYSICAL CONSTANTS The analyses were performed by Dr. W. M. MacNevin and Mr. Joseph Varner, by the procedure described in Mr. Varner's Master's thesis, The Ohio State University, 1943. Freezing B. p. (760 Chlorine % Fluorine, CFClxCHaCHaCHi CFXC~CHXCHSCHI CFCI-CHCHxCHi CFPCHCHXCHI CFaCHrCHxCHi CFaClCHClCHtCHi CFiCHCl CHtCHi CFiCHaCHClCHi CFiCHaCHxCHA!l CFiCHxCClxCH: CFiCHxCFx CHx CFiCHiCHClCHxCl CFiCHsCHiCHClr
range 0.1 ,004 .1 .1
.02
.1 .02
.01
F. p., OC. -112.9 -119.39 -144.9 -164.95 -114.79 Below -100 Below -100 79.24 67.25 Ca. -48 35.01
-
-
mm.), OC. 96.3 56.03 54.4 3.71 16.74 96.98 58.3 65.63 86.60 89.90 40.14 115 110
d%
in hydrogen fluoride as a solvent, a process which is known to fluorinate a halogen cluster in preference to a single halogen atom.'qa The resulting CFaCHC1CH2CH3proved different from both the isomers formed in the chlorination of CFaCH2CH2CHa. The structure of CFsCH2CHClCH2Cl was derived from the fact that it was obtained from both CF3CH2CH2CH2C1 and CFaCH2CHClCHa. Purification and Physical Constants.-The Purification Procedure and the Physical measurements were carried out as indicated in the preceding paper,2 and with the degree of accuracy shown in the table. Reactivity Tests*-Tested with a boding solution of silver nitrate in alcohol,2 CFaCHC1CH2CHa and CFaCH2CHC1CH8 gave negative results while CFsCHzCHaCHzCl gave a faint turbidity. The interpretation that both the alpha and the beta chlorine atoms Were immobilized by the CFa &YOUPIwhile the gamma d o rine atom was Only repressed. The influence Of a CFa group extends, therefore, further than that of a -CFzgroup.2 Treated with alcoholic &&, CFsCHClCH2CHa did not react; CFICHICHCICH, gave CF,CH= CHCH~;and CF~CH~CH&H&~ gave an ether CF8CH2CH2CH2--*R* The first two results are similar to those recorded in the preceding
1.4007 1,3462 1.3851 1.3253 (0') 1.2921 1.3878 1.3369 1.3433 1.3505 1.355 1.2824
A R F Found chlcd. Found Cazd.
MR
Ux@D
1.1578 1.0754 1.0216 0 . 9 2 s (0') 1.0144 1.2716 1.1967 1.2133 1.2425 1.374 1.2666 1.390 1.382
30.41 25.47 24.90 19.97 20.51 30.24 25.4 25.54 25 41# 30 6 20.67
1.10 1.07 0.93 0.99 1.05 1.02 1.07 1.10 106 1.14 1.10
48.9
49.0
13.1
13.2
32.7
32.7
17.2
17.5
..
..
50.7
50.8
24.2 24.0 24.4 39.8
24.2 24.2 24.2 39.2
38.6 38.7 39.2 31.3 63.8
38.9 38.9 38.9 31.5 64.1
..
..
paper2for the monochloro derivatives of CHsCF2. CH2CHa. The formation of an ether in gamma position shows that the chlorine atom is too far from the CFa group to be d e c t e d by it and behaves like that of butyl chloride, which also gives an ether when so treated.
summary CFaCHnCH2CHa was synthesized by the action H S , of hydrogen fluoride on C C ~ ~ E H ~ Cwhich yields CF2C1CH2CH2CHa,followed by hydrogen removal to CF~:CHCH~CH~ then hydrogen addition to CFsCH2CH2CHs. ChlorinaGon did not affect the alpha CHI, but only the beta CH2 (4 parts) and the gamma C& (5 parts). ~~d~~ &lorination tended to on the carbon atoms already affected. The CFs group immobilized alpha and beta chlorine atoms and repressed a gamma chlorine atom, thus proving more effective in this respect than a CF2 group; it also accentuated the acid character of the alpha hydrogen atoms. COLUMBUS, OHIO
RECEIVED MAY7, 1945'
(4) This manuscript was originally received on October 11, 1943, and after examination by the editorial boud was accepted for publication in Tms JOURNAL. It was, however, referred to the National Defense Research Committee and at their request waa withheld from publication, in a confidential 61e, until clearance was granted on May 5,1945.
