Dec., 1946
PREPARATION AND POLYMERIZATION OF VINYL
TABLE I11 CONSTANTCOPOLYMERIZATION MIXTURE^ 60.6 wt.% dichlorostyrene, 39.4 wt.% butadiene Re-
Polymer
aC-
tion time, Yield,
1
@n
[li
% at 30° k‘ 15 7 . 2 0 . 6 4 2 0 . 3 6 4 25 1 1 . 6 ,666 ,347 47 1 7 . 6 ,725 ,364 57 3 1 . 4 ,890 ,351
hr.
p
Chlorine,
X 10: a t 27’ % 90 0.421 24.99 24.75 100 ,404 .. .. 130 .397 . 160 ,395 24.39 24.25
. . ..
. ..
Dichlorostyrene,
%
60.7
.. ..
59.4
I n presence of 0.1 weight per cent. of benzoyl peroxide a t 70’.
polymer composition is shown by the analytical results secured a t the 7.2 and 31.4y0extremes of conversion. Theyintrinsic viscosities were measured, and were found t o remain relatively constant up to about 12y0 conversion and then showed a marked rise, a 40y0increase being found a t 3lY0 conversion. A similar set of measurements with d.ichlorostyrene alone showed that the intrinsic viscosities were constant up to 21 .O% conversion, the intrinsic viscosity being 0.77 0.03. The number average osmotic molecular weights of the copolymers were determined and ranged from 90,000 to 160,000, increasing with
[CONTRIBUTION FROM THE
FLUORIDE
2467
degree of conversion. The viscosity, k’, and osmotic pressure, p, constants6 indicate, among other properties, the shapes and flexibilities of the polymer in the solvent, toluene; they were found to be constant within experimental error. Acknowledgment.-The authors wish t o express their appreciation t o the DOW,Monsanto and du Pont companies for the monomer samples; t o Dr. Carl Tiedcke for micro-analyses of the butadiene copolymers, and to Miss Doris Linberg for her assistance in much of the laboratory work. Summary Experimental results on two olefin-diene copolymer systems show that the simple copolymer composition equation is obeyed. The values of CY and p for the systems styrene-chloroprene, and dichlorostyrene-butadiene have been determined. An investigation was made of the change in viscosity and molecular weight with increasing degree of conversion of the constant copolymerizing mixture of the latter system. (6) Huggins, I n d . Eng. Chew., 36, 216 (1943); J . Phys. Chem., 42, 911 (1938); 43, 439 (1939). BROOKLYN,
N. Y.
RECEIVED MAY27, 1946
RESEARCH LABORATORY, GENERALELECTRIC COMPANY]
The Preparation and Polymerization of Vinyl Fluoride BY A. E. NEWKIRK The polymerization of vinyl fluoride has been reported to be difficult or i m p o s ~ i b l e . ’ *There ~~~ seemed to be no theoretical basis for this difficulty, and as the polymer should have useful properties an investigation was undertaken to prepare and polymerize vinyl fluoride.
cess. Vinyl fluoride has been prepared by the addition of hydrogen fluoride to acetylene both without a catalyst10t11J2,and with the aid of mercury (11) oxide on activated carbon.l3 The catalytic reaction in the vapor phase over catalysts of mercury (11) chloride, or mixtures of this and barium chloride, supported on active carPreparation of Vinyl Fluoride bon was used to prepare vinyl fluoride for .the Vinyl fluoride was prepared first by Swarts4 polymerization experiments. The latter catalyst by the reaction between lIl-difluoro-2-bromo- was superior in all respects. It gave higher conethane and zinc. This general reaction between versions, permitted higher space velocities, and an active metal and a halogenated ethane has had longer life. Typical results are shown in been used by other investigators for the prepara- Table 1. The mercury (11) chloride catalyst was tion of vinyl f l ~ o r i d e . ~ , Swart9 ~ . ~ , ~ has also used active for from eight to twenty-four hours, after phenylmagnesium bromide in ether, and alcoholic which its activity dropped to zero. Globules of potassium iodide, in place of the metal for remov- mercury were observed in the spent catalyst. ing halogen. The reaction of vinyl chloride or The crude gas from the reactor, after removal bromide with a metal fluoride is not described in of hydrogen fluoride, was mixed with ethane and the literature, and was tried with but slight suc- fractionally distilled a t atmospheric pressure in a low-temperature column. The addition of ethane (1) H. W. Starkweather, THISJ O U R N A L , 66, 1870 (1934). ( 2 ) F. Schloffer and 0. Scherer, German Patent 677,071 (1939). permitted the removal of acetylene as a n azeo(3) C . A. Thomas, U. S. Patent 2,362,960 (1944). trope.14 The vinyl fluoride fractions had a purity (4) F. Swarts, BUZZ. aced. m y . Belg., 383 0 9 0 1 ) .
(5) F. Swarts, ibid., 728 (1909). (6) A. L. Henne and T. Midgley, Jr., THISJ O U R N A L , 58, 882 (1936). (7) A. L. Henne, ,ibid., 60, 2275 (1938). (8) P. Torkington and H. W. Thompson, Trans. Faraday Soc., 41, 236 (1945). (9) F. Swarts, Bull. soc. chim., [41 26, 145 (1919).
(10) H . Plauson, U. S. Patent 1,425,130 (1922). (11) A. V. Grosse and C. B. Linn, THISJOURNAL, 64, 2289 (1942). (12) A. L. Henne, “Organic Reactions,” Vol. 2 , John W l e y and Sons, New York, S . Y . , 1944, p. 66. (13) J . Siill, German Patent 641,878 (1937). British Patent 469,421 (1937), U. S. Patent 2,118,901 (1938). (14) W. A. McMillan, THIS J O U R N A L , 68, 1345 (1936),
A. E. NEWKIRK
2468 TABLE I PREPARATIOX OF V I N Y L FLUORIDE
Catalyst HgClz-naC12 HgC12.BaC1:, Temp., " C . 97-103 97-10.1 83-10c3a Vapor input IHF 460-540 460 SGO C:2H? 440 440 881I cc.,'miii. Grams crude per hour 27.3 47.3 '33, W t . yo vinyl fluoride 51.5 82.0 x1.9 W t . % 1.1-diflunroethane 16.0 1.1 .. The reactSr heater was disconnected at the start of this run, T = 83 . The temperature in the middle of the reactor rost: gradually to 109 O during 5.7 hr.
i
of 94 mole Per cent. or better and were stored in steel N~ trouble was experienced for periods Of Over Of such purity six months. Experimental Substitution Reaction.-Vinyl chloride was passed over a mixture of antimony (111) fluoride and aiitimony (V) chloride heated in an iron tube at temperatures from 26 to 470' without t i e formation of vinyl fluoride. Mercury (11) fluoride reazted with vinyl chloride at 200" to produce a very small amount of vinyl fluoride. Vinyl bromide and mercury (11) fluoride reacted when sealed in a small bomb t o yield some vinyl fluoride, but the main reaction was the polymerization of thc vinyl bromide. Y,
Vol. 68
iron pipe three feet long which contained catalyst in the upper two-foot section. The reactor was fitted with a thermocouple well which permitted measurement of catalyst temperature a t any point on the axis of the bed. The outlet gas from the reactor was scrubbed with 10% sodium hydroxide solution and dried over calcium chloride. It was then collected in a trap cooled to - - 8 5 O , or by-passed to a gas density balance protected by a tube containiug sodium fluoride and sodium hydroxide. When starting a ruii, the apparatiii was disconnected at "A" and the reactor heated to 100" in a stream of hydrogen fluoride. Acetylene was then added, the scrubbers attached, and a< soon as the gas density balance indicated a n appreciable conversion (molecular weight = 40 or higherj the product trap was cooled. The effectiveness of the trap increased as the conversion of acetylene increased, and i t i one run iti which high conversions were obtained, less than 291, of the total product was caught in a liquid nitrogen trap placed after the Dry Ice trap. Catalysts: No. 1.-A solution of 145 g. of mercury (I1j chloride in 800 cc. of boiling water was poured oiito 535 g . of Columbia activated carbon pellets (grade 4sxlv, 6 to 8 mesh). The water was evaporated from this mixture with slow stirring and the mass dried at l0Cl" for two days. It was charged into the reactor and dried in a slow stream of hydrogen fluoride for eight hours starting a t 100" atid finishing a t 130'. The system was then closed off until used. No. 2.--.4 solution of 145 g. of mercury (11) chloride aiid 131 g. of barium chloride dihydrate in 350 cc. of boiling water was poured over 535 g. of carbon pellets, the water evaporated, the mass dried overnight a t Uso,and then in a stream of hydrogen fluoride as described above. Analysis.-Typical samples of vinyl fluoride purified by fractional distillation boiled in the range - 2 to -70" and had the following compositions in mole per w i l t . acetylene, ethane and vinyl fluoride: (1 I 2 , I , 97; (2) trace, 4, 96; (3) 0 , 2 , 98; a t i d i 4 ! 1, 5, 91. The compositioii of the vinyl fluoride samples and of some of the other fractions was determined as follows. The density of the gas was measured with ail Edwards gas density bala:icc, anti the sum of acetylene and vinyl fluoride tlctcrmined by absorption in h-oiriine mater. The composition was calculated assuining vinyl fluoride, ethaiie :iiid acctylciie ER to be the only substances present. This method had satisfactory accuracy. .hi analysis of an azeotrope fraction gave in mole per cent., C2HZ= 40.2, C2H6 = 59.0, ratio = 0.G3, CZHIF= 0.8. The published data are 40.75, 59.25, 0.688.14
Polymerization of Vinyl Fluoride Attempts to polymerize vinyl fluoride have been reported only A L L CONNECTING LINES I/P" IRON PIPE EXCEP? very briefly in the literature. AS riOTEn Starkweather' observed polymeriREACTOR SCRUR8FRS PROCJCT T R A P zation to "some extent" upon treating a toluene solution, s i t u rated a t -%', a t 67" and (iOOO Fig. 1 .-Apparatus Ior the prcparatioii of vinyl fluoride. atmospheres for sixteen hours. Addition Reaction: Apparatus.-This is showrl it1 Fig. 1. Scliloffer and Scherer? found vinyl fluoride very Corntnercial hytlrogcii fluoride from a cylinder was passed hard to polymerize, and Thomas3 has rc,portedthat tlirough ncctllc valvc, VI, at1 orifice flowinetcr, and was thcll iv: :tcctylclle \vhich had l,ccl, mctet.crl vinyl fluorirk ~ 1 heated ~ n in the prcsencc. of through a ,.otarnrter I,t'otcc(ecl b y a s a f r t y tra,,. The peroxide catalyst undergoes substantially n o "Saran" ~iianoiiieter tube of the hydrogen fluoride flow polymerization. Resinous materials have brcri meter beeatne (lark after about each three days of use and obtained in preparing villyl fluoride fronl acetylene was replaced. Liquid hydrogen fluoride was the maiioni- and hyilrogen fluoride, but have not beel, identieter fluid. The valve, VZ,was used to control the presfied urt"r.'0'"''2 sure in the gas inlet system as indicated by the mercury 1n the present investigation Vinyl fluoride was rnanoinekr placed a t the outlet to the rotameter. The gas mixture was passed into a reactor of two-inch diameter poltmt~rized by exposure to light of wave length
U
A.,
Calcium ste:trate and magnesium oxide were less than 2800 and by the use of benzoyl peroxide, lauroyl peroxide and acetyl peroxide. found t u impart increased color stability to polyConcentrated sulfuric acid caused charring of the \.,nylfluoride when used in an amount equal to 2y0 . monomer; dilute sulfuric acid had no effect. oi the lveight of the polynicr. li.hen heated i n a gas flame, polyviiiyl fluoride The conversion of monomer was increased froni 2y0 to 40y0, on the average, by the addition of a I!ipltt.d and tlier; burned with a yellow, very sooty solvent for the monomer and catalyst. Ace- fl;:nze. Combustion continued for several seconds aitcr the rei:!oval of the flame, then ceased. tone, ethanol and isopropanol were satisfactory. .I i j i - i t , f study wa.s made of the rate of deconipoConsiderable difficulty in obtaining reproducible 1 fluoride (prepared with benresults was wperienced with the acetone solvent, sitioii oi 1lol!-i-i~1~ benzoyl peroxide catalyst polymerizations. This n)?-1pvroxi(!e citalyst) in air and in nitrogen a t was due in part to the presence of both 1ower.boil- ! 7.: ' . 'I'lie etiiaiiol extracted polymer was much ing and higher boiling impurities i i i thc viiiyl mow st.:ttlc, t!iaii tlic crude polyiiicr, :tnd the polyfluoride which inhibited the poly~ncrizatioi~as mer w:ts iiiore stable in nitrogen than air, Fig. 2. shown by a ca.refu1 redistillatioii a n d polyi1ierization of the three fractions obtained. Teri per cent. of ethane had no effect on the yield of plyiiier. Variability of results was also influcnced hy the amount of water in the acetone a n d it WI.';LS found that the addition of 10y6water was beneficial. Reproducible yields with differen t prejmrations of vinyl fluoride were obtained after atlupting these two procedures. The polymeric vinyl fluoride, after drying, was obtained i n soft, friable white chunks which crushed 1;ery readily to a powder. The U.1'. polymerized mater.ia1 remained white indefinitely, but samples polymerized with the aid of benzoyl peroxide turned brown on standing. Polymers prepared in dry acetone darkened more readily than those prepared in acetone containing IOYl, water. It was found that extraction of the polymer with ethanol for four hours eliminated the darkening on storage of the main portion of the polymer which was ethanol insoluble. The ethanol soluble portion was purified and dried in a vacuum desiccator. I t became slowly dark brown in color, had a v e r y pungent odor, ;and the glass container was etched. I t seems likely that the extraction treatment fractionated the polymer and removed material of low molecular weight as well as other substances tizit catalyze the darkening, which is probably clue to the splitting out of hydrogen fluoride acid. This behavior is comparable to that reported for vinyl chloride1i which it is claimed can be made iiiore stable toward heat and light by the estraction of the lower molecular weight polymer by the solvent a.ction of aliphatic alcohols. Several samples of polyvinyl fluoride prepared with benzoyl peroxide catalyst were heated in a Fisher-Johns inclting point apparatus bc.twec.11 iiiicrosco!)ct covcr glasses. Softcniiig startctl :~rountl I 'Lo', ;I iitl a t I!)()* tlie saiiiplcs wvrc C Y M I I plc~tclyfluid. 'l'J.~(wMYLS i i o c v i t 1 r ~ i i c . rof ( I c ( ~ ) i i i l ~ sition. \17hcn worked on a stecl-sutfa~~e(l hotplate with a spatula, darkening sct: in ariuui(1 IO!)", the polymer powders could he worl;c(l i ! i L o :.i coherent mass a t I . 1 0 to ].ioo, and were ol)si~i-\.c.ti t o string out a t ] S O o . At the lattct- teiiilm-atiire they ranged in color from deep chocolate b r o w i to black clue to decomposition.
-
)-
(1.5) Deutwilie Celltiloid-Fabrik, German P a t e n t ti7Y,87(j (1!13Ll].
A. E.NEWKIRK
2-1'70
VOl. 68
3.0
2.5
-
s 2.0 r; .w
P
-:
1.5
4
s b
P.
1.0
0.5
0
0 Fig. I.-Thermal 2 ;: 1 3 ij Time, hours, I;ig. :3. -Thi:rmal decoinpoiitioii at 175' in nitrogen : €'\.I;, polyvinyl fluoride, ethaiiol iiimluble polymer; P\'Cl-l , polyviiiyl chIoritlc !jurifictl b y citraction with hot ~ y lchloride, P\.Cl-l, iwprop~-lalco 101: P\'CI 2 pol m t y l acctoacetatc a i plui 1 .:I', ; l ~ yIvvight of culciu iiihibitor, 0
I
ene fluoridc, 1,"-difluorocyclohesane, ancl 1,2tlifluorodiosnne decompose spontaneously t o olefin anti hyclmgen fluoride a t rootii temperature, and ra piclly form the dihydrosy derivatives on being passed through water. IJ-hile the head to tail arrangeinent iiiight be expected to be more stable tlictn thc others 111 the present instance, the st:tbility of the polymer is quite remarkable in .;ier; of tile instability and ease of hydrolysis of monofluorid cs. The tlctisi t y of polyvinyl fluoride determined pyknonietric,il!!. oil ;til ctlianol extracted sainp:e using ct1laiit:l ;LS thc fliiitl W;IS 1 .:jO : I t 25'. 'The sol~c,iitx:tiii~i oi :L iiuinber o f coiiilmiiiils on polyvinyl fluoric!c has eii cliialitatively tlctermirietl. I t is 1101. ;i]q)r M y solnblc :it rooiii ternpci-atui-ei i i :!ny solveut tried, but will dissolve in hot dioxane, cycli!hexanotie, isophorone, phorone, fenchonc, chlorohcnzenc, I ,1 ,2-trichloroethanc, 1 , 1 ,:!,2-t c t r x 1 i h - t icltliitnt', and ,ii--tricresyl phosphate. (.hi cooling the soliitii~ii,eitlicr precipitation of t h e polyiiier or gelation occurs. A 3:;; solution i n t i i x I , 1 ,?-trichloroethsne forms a soft gel a t 1*00!11 I c.li?jjc'r:lture. Thr iiiol(~c-~!l:ti-w i g h t of the etliaiiol insoluble, nietliyl et1iJ-L bct,)nc solii1,le ~wrtioiiot' the polymer was deteriiiirietl 1)ythe osiiiotic pressure method tlei aiid 1;ttoss" tu be L':i,000 * 5,000. ilS, L). 1. ~ 1 c . r d.iud iZ. 11. i ' u u a b , J . '1
2 3 4 5 Time, hours. decomposition a t 175' in air. same as Fig. 3.
1
G Notation
Polymeric vinyl fluoride appears to contain a crystalline arrangement for the most part. X film prepared by pressing a cover glass over a pool of molten polymer on a microscope slide, when examined after cooling, showed numerous small anisotropic areas. The X-ray powder diffraction pattern of the polymer powder had three sharp reflections. A fiber drawn from a pool of molten polymei- gave a pattern with nine reflections. These showed that the fiber was only very slightly oriented. Vinyl fluoride polymerized with vinyl acetate, vinyl chloride, and methyl methacrylate to form polymers containing fluorine. The copolymerization of vinyl chloride and vinyl fluoride has been disclosed by T h o ~ n a s . ~ Experimental Polymerization by Ultraviolet Light.-Three types of ultraviolet lanips were used for these experiments. (1) t h c General Electric Type H-4 lamp which transmits 110 ratlintioii a t wave lengths less than 2800 A , , (2) the same lami) without the outer glas: bulb, in which case consideraide radiatioii below 28OU A . is available, and (3) a four \v;ttt gci-iiiicictal laiiip with over 90r6 of the emitted radiatioii of wave length 25:37 A . Vinyl fluoride liquid in quartz capillary tubes at 35" polymerized on esposure to 2 arid 3, but not 1. The germicidal lamp caused polyiiierization of thirtp-