Production of
POLYPERFLUOROVINY1 CHLORIDE E. A. Belmore, W. 31. Ewalt, aiid B. H. Wojcik HOOKER ELECTROCHETIIC4L COMPAYY, \ 1 4 G 4 R A
F4LLS,
\.
Y.
THE method developed by RIiller and associates a t Co1un1bia C'niFersity for t h e preparation of the polyperfluorovinyl chloride was successfully applied to pilot plant operation. Perhalovinyl pols mers \ a r j i n g in consistency from colorless, stable oils to solid resins were readil? obtained. Reaction conditions emplosed for t h e pol? merization of perfluorovinyl chloride were significant i n influencing t h e degree of polymerization and t h e t \ p e of pol?mer isolated. 4 molecular still was utilized for the segregation of t h e oil) constituents from the resinous products. Final treatment of the polyperfluorovin?l chloride oils \+ i t h cobalt trifluoride increased t h e thermal and chemical stabilitj of t h e oils. 4 procedure was dcbeloped for the contersion of solid resinous material t o oily products. 4 flow sheet shows the various steps in t h e s)nthesis, and data accumulated during pilot plant operation are included.
The selected fraction, distilled from the crude polymer of step 2, is treated :it 200" C. with a n equal weight of cobalt trifluoride. Thiq treatment saturates double bonds and replaces hydrogen and loosely attached chlorine atoms with fluorine.
-
PREPARATION OF PERFLUOROVINYL CHLORIDE
%gallon reactor, patterned after one used by Miller and co15-orliers, was set u p and a number of preliminary runs made t o :icquire firdthand experience with the reaction. From this study :I 110-gallon reactor ivas built, equipped with a packed column for removing the perfluorovinyl chloride. Figure 1 is a diagram of thi.; unit. .Inhydrous ethanol \vas mixed n-ith finely divided zinc dust in an :igitnted iron tank, and the slurry chnrged into the reactor. T h e circulnting trichloroethylene, used as a cooling medium, v-ns cooled to :i temperature between -50' and -60" C. with dry ice. Tlie triclilorotrifluoroethaiie~ obtained from Kinetic Chemicals, Iiic., n-n.5 fed directly into the reactor, which vias heated to T O " C. The rate of feed was controlled by means of a rotameter. Vapor tempei:iture, reactor temperature, and t h e pressure drop ttilougli A%
4
L.iBOR.iTORT method for the preparation of a fluoririecontaining oil having a high chemical inertness and certniii definite physical properties \vas norlied out b y \I7.T. IIiller and associates :it Columbia Vniversit~y. To make this oil. perfluorovinyl chloride v a s prepared aiid then polymerized, and thc. desired oil n-as separated from the crude polymer b y dist,ill:~tie):i under vacuum. The Hooker Electrochemical Company iv:i\ assigned tlie task of developing the laborztory process iiitci : I suitable plant operation. This paper describe; briefly tlie el?velopment work carried out in connection Tyitli thii problem : four steps w x e involved. The firpt is the preparation of perfluorovinyl chloride motiomei': A
CF2CI-CFCII CYLINDERS
POWDERED ZINC
EL SCALE
ClF?C-CCI?F
+ Zn *l+F2C=CFC1
7
ZnCI?
(l)
T h e 1,2,2-trichloro-l,l,2-trifluoroetlianeis dechlorinated by ziiic suspended in ethanol to give perfluorovinyl chloride. This conipound boils a t about -27" C. and is removed from the reactioii mixture through a fractionating column. The second step is polymerization of perfluorovinyl chloride:
F2C=CFC1
chloroform + (-F2C-CFCl-)= benzoyl peroxide
(2)
The pure monomer, dissolved in chloroform coritainiiiy the benzoyl peroxide catalyst, is heated in a pressure vessel whrw polymerization rapidly occurs. \Then the reaction is complete, the chloroform and other volatile materials are stripped off tci leave behind the polymer concentrate. The desired oil fractior is separated from this concentrate by distilling under liigli vacuum. T h e third step is the preparation of cobalt trifluoride:
+ F1+
2CoFa
,
REACTOR R E A c T ,.,o L.., i g r ~
2CoF2
(3)
Cobalt trifluoride is used in the following step and is prepared by treating anhydrous cobaltous fluoride with elemental fluorine, which rpndily oxidizes it t o the trivalent state. T h e fourth step is the treatment of dist,illed polymer:
1
MONOMER CYLINDERS
Figure 1. .ipparatus for RIonomer Preparation
338
March 1947
339
INDUSTRIAL AND ENGINEERING CHEMISTRY
the column ivere the primary indications of proper operating conditions. T h e fractionated prode ul uct \vas collected in steel cyliiiderb surrounded by a trichloroethylene-dry CHLOROFWH ice bath. .Ill lines WEIGH handling tlie prod?AH)( uct were either jacketed n i t h the t r i c h 1o r o e t h y 1en e c o o l a n t o r \ thoroughly insulated. In general the reaction seemed t o be rapid and straightf o r m r d and presented no unusual problems. Table I shows the d a t a from five runs carried out with this equipment. T h e yields of perfluorovinyl chloride varied from 79 t o 92Y0 of t h e o r y based on t h e rreight of trichlorotrifluoroethane used. This variation vas not due t o any inherent characteristic of the reaction b u t t o t h e skill and care of the operator. T h e use Figure 2. Flow Sheet for Polymer Preparation of adequate control instruments would perfluorovinyl chloride, and 8 pounds of benzoyl perosidc. Reprobably have produced consistently high yields. The amount action conditions recommended by Xiller Tvere 100" C . for 90 of zinc used was approximately 4OrC in excess of the theory reminutes. In order t o incrense the productivity of ench autoquired, b u t no attempt> were made to recover this exce-.: or the clave, pilot plant operations n-ere carried out :it 150" C. for 15 alcohol from the residue in the reactor. .liter these runs, niinutes. This change gave n polymer of a lover average molecuarrangements \\-ere made n-ith the D u Pont Company to supply lnr n-eight but did not markedly decrease t,he nniount of the fracthe monomer necessary for future n-orl;, so this p1ia.e of the tion suitable for the final product. Figure 2 i- tlie flow diagram problem was discontinued by the Hooker Conipnny. for the autoclave operation. The benzoyl peroxide, which contained abciut loc; n-:iter, was POLYRIERIZATION O F T H E PERFLUOROVISYL CHLORIDE \IONOJIER dissolved in chloroform, treated n-ith anhydrous sodium c x b o n ate, and charged into the :iutoclave. Since most of the perfluoroPreliminnry polymerization experiment< were c:ii~ried out in a vinyl chloride cont,ained a nonvolatile stabilizer, the monomer 22-gallon pressure veGsel, b u t for the pilot plant oper:ition :I 110~ v a iremoved from the cylinders in the pas p1in.e and passed gallon -tee1 autoclave n-as used. .Istandard charge for this unit through phosphorous pentoxide t o remove traces of moisture consisted of 000 pounds of alcohol-free chloroform, 100 pound. of and alcohol before being led into the autoclave. The reaction n-as c:irried out by starting the agitator :ind hcating the autoclave to 150' C. in nbout 6 or 7 minutes. The presT.4BI)E I. P R E P A R i T I O S O F P E R F L U O R O \ I " T L CHLORIDE IS 110GALLOSREACTOR s ~ i r equicldy rose to R ninxinium of nbout 300 pound. ~ C I square ' l l a t e r i a l s Charged, Lb Product inch and then rapidly decreased :ithe monomer v-:I> consumed. Run Abs CF?=CfCI, 7c of .ift,er 15 minutes a t 150' C. the autoclave was coolrd to 40To CFICI-CFCl, ethanol Zincn Ib theory 50 ' C. and slowly vented through a dry ice-cooled condenser to 195 150 151 88 8 273 200 150 157 84 0 300 recover t,he unrt?acted monomcr. This wa;;collected in cylinders 3 300 200 150 148 79 0 200 150 172 92.0 4 300 and recharged into t,he autoclave in sub-equent r ~ u i s . The re5 348 232 li4 200 92 0 covery of the monomer was complicated by the gradual accumulaTotal 1521 1027 774 828 87.3 tion of an appreciahle quantity of hydrogen fluoride which had a . 1 small aiiiount of spent sludge was left i n the reactor a t the end of each to be disposed of from time to time. The lnsae? from this crude run. system were probably high, b u t during 100 autoclave bntclies an
-
-
I
8
;
t
340
INDUSTRIAL AND ENGINEERING CHEMISTRY
FROM
CONDENSER
Vol. 39, No. 3
,
L11 F i g u r e 3.
DISTILLATION OF CRUDE POLYMER
Flow S h e e t for P o l y m e r C o n c e n t r a t i o n
average of 17% (17 pounds per batch) of the monomer charged was recovered. After venting, the aut,oclave was further cooled to about 25" C. and discharged into a 300-gallon holding tank (Figure 3). When five autoclave batches were accumulated, the material was treated with sodium carbonat,e and filter aid and pumped through a pressure filt,er. This step was added primarily to remore any iron t h a t might have been picked u p in the preceding operations. From preliminary work i t had been found that small amounts of iron in the crude concentrated polymer greatly :iccelerat,ed the rate of rracking and made distillation u d e r high vacuum practically impossible. The filtered chloroform solution was next transferred t o n simple 500-gallon enameled still, equipped with lead accessories, where the chloroform and polymers boiling below 100" C. were removed and returned to the storage tank. The filtered solution still contained some dissolved monomer; t,his as recovered by returning the uncondensed gases to the monomer recovery system. The residual polymer v a s transferred to a 100-gallon enameled vessel and further hented t o 150 C. while a vacuum T a s gradually built up to the masimum obtainable rvith water aspirator;.. At the same time a slow stream of air vas bubbled through tlie polymer. This treatment served to remove volatile materials and to decompose some of the more unstable components. After this, the remaining polymer was cooled t o 100-125° C. and discharged into clean steel cans. At this temperature the polymer was a tan oil which solidified into a sticky grease before room temperature was reached. I n order to isolate a fraction of the oil suitable for further processing, i t was necessary to distill the crude polymer under high vacuum. At distillation temperature the crude polymer was unstable and cracking readily occurred, accompanied by the evolution of much gas. Arrangements were made vvith Distillation Products, Inc., to carry out this distillation in commercial molecular stills. In their equipment the crude polymer was separated into five fractions of distillate, each containing polymer of successively higher average molecular weight, and a residue which was a brittle reainlike material melting a t approximately 90" C. All fractions were returned to the Hooker Company There they were blended i r i the following manner: Low boiling fractions containing no usable polymer rvere discarded, fractions containing
less than SO(> uvible polymer mere returned to Distillatiori Products, Inc., for redistillntion, fractions containing more thnn 80yo usable polytner mere processed through t,he nest step, and the rest \va> residue. Table I1 consisti of d:ita taken irom a report by Distillation Products, Inc., on tlie distillation of the first two samples of crude polymer suhmitted to them. These results xere obtained on a laboratorv still and are not strictly comparable to those from production units. The molecular weight determinations, however, give a good indicntion of the type of polymer that was being proccssed. PREPARATION OF COBALT TRIFLUORIDE
Figure 4 shows the flow sheet for this operation. The elemental fluorine, from a single cell of the type described by Murray, Osborne, and Icircherl, was led into the bottom of an insulated 1 I I u r r a y , R. 39, 249 11947).
I,., Oshorne. S. G..
and Kiroher, 11. S., IND.Exa. CHEM.,
TABLE11. DATAFROM REPORTO F DISTILLATION PRODGCTS, INC., ON FIRST T W O SAMPLES O F CRUDE COSCENTRATED POLYMER SUBMITTED TO THEY R u n 12, Polymerized a t 100' C. for R u n 15, Polymerized a t 160' C for 10 Minutes
".. *.-tion
No. Original 1 2 3 4 5 6 7 8 9 10
11
12 13 Cold t r a p Residue a
...
1.63 3.92 5.00 4.73 4.10 4.03 4.73 2.38 4.72 8.10 3.45 0.83 6.04 10.70 28.80
Total 100.0
,...
5.55 10.55 15.28 19.38 23.41 28 14 30,51 35.24 43.34 46.79 47.62 53,66 64.36 93.16
.. I., approx. 1040 640
-
~
No. Original 1
700
2 3 4
772
5
si2
6 7 8
1050. 1260 1580
2000
9 10 11
12 13 14 Cold trap Residue
tion ILOO.0 4.25 2 90 7.17 5.10 5.35 8.62 2.92 3.62 5.40 4.10 5.40 1.94 3.46 5.04 6.04 21.6
Totr
...
...
7.15 14.32 19.42 24.77 33.39 36 31 39.93 45.33 49 48 54.83 56 77 60.23 65.27 71.31 92.9
...
540 582 660 i60
960. 1080 lie0
..
2040
These and subsequent fractions were waxy solids a t room temperature.
March 1947
INDUSTRIAL AND ENGINEERING CHEMISTRY
341
lond Ira? t h w thr
