482 COZEWITH, VER STRATE
Macromolecules
SCHEMEI
quinone
+
/-’
quinone incorporation (significant)
.___t__,quinone(b;;;:foration
vinyl monomer
1
T
inhibition
(charge-transfer complex) but inhibition with the other monomers. Two examples of significant incorporation and three examples of “slight” incorporation were observed with o-chloranil, whereas 2,3dichloro-5,6-dicyano-p-benzoquinone afforded four significant incorporations and two slight. I n general, however, only inhibition was observed in reactions with these quinones. We d o not know to what extent charge-transfer complexes influenced the observed polymerization reactions and, conversely, the inhibition to reaction. The facts that coloration almost always occurred upon admixture of quinone and vinyl
monomer and that quinones are known electron acceptors attest to the presence of CT complexes in the reaction mixtures, if not to their direct participation. We conclude that although some quinones can copolymerize with some vinyl monomers under free-radical conditions, the quinone copolymerizability and its redox potential are not generally related. Quinones of low, moderate, and high redox potentials have been treated with vinyl monomers over a wide range of double bond a ~ t i v i t y , but ~ ? ~the observed copolymerization reactions were spotty and inconsistent. However, some interesting copolymers were obtained from the present study, some of which arose by copolymerization via the quinone carbonyl groups to yield aliphatic-aromatic polyethers. The most significant example of a n aromaticaliphatic polyether obtained from the work presented in this paper was the product from 2,5,7,10-tetrachlorodiphenoquinone and styrene, wherein a high-molecular-weight copolymer was realized. We feel that other polyethers of high molecular weights can be prepared by use of quinones as comonomers, provided the proper reaction conditions are found.
Ethylene-Propylene Copolymers. Reactivity Ratios, Evaluation, and Significance C. Cozewith* and G . Ver Strate
Enjay Polymer Laboratories, Linden, New Jersey 07036. Received January 15, 1971
ABSTRACT: Reactivity ratios have been determined in ethylene-propylene copolymerization for a number of Ziegler-Natta catalyst systems. It was found that the reactivity ratio products varied with the composition of the catalyst components and were generally less than 0.5. A critical survey was also made of other published values for ethylene-propylene reactivity ratios. Considerable disagreement exists, which was primarily attributed to the method of copolymer analysis and the presence of multiple active species which are formed by many soluble Ziegler-Natta catalysts. It is shown that if multiple species are present, the reactivity ratios are a function of the mode of catalyst addition. Also, the reactivity ratios determined from the normal copolymerization equation do not correctly predict the sequence distribution in the copolymer.
T
he composition of the Ziegler catalyst used to produce ethylene-propylene rubber has a profound effect on both the polymerization process and the polymer properties. This effect arises in part from the variation in copolymerization reactivity ratios that occurs with a variation in catalyst system components. The value of the reactivity ratios (under otherwise constant copolymerization conditions) determines not only the relative monomer conversions but also the distribution of monomer sequences in the polymer chain. This distribution is of particular importance in ethylene-propylene rubber since the polymer crystallinity that results from long ethylene sequences affects the mechanical performance of the finished goods and the rheological behavior during processing. Reactivity ratios for ethylene-propylene copolymerization with soluble Ziegler catalysts have been reported in the literature (see Table 111), and some attempts have been made to relate these parameters to the monomer sequence distribution in the copolymer chain.’-* There is also information available o n how the values of the reactivity ratios actually affect (1) C. Tosi, A. Valvassori, and F. Ciampelli, Eur. Polym. J . , 5 , 575 ( 1969).
(2) Th. A. Veerkamp and A . Veermans, Makromol. Chem., 58, 147 (1961). (3) J. P. Jackson,J. Polym. Sci.,Parr A , 1, 21 19 (1962). (4) G . Ver Strate, and Z. W. Wilchinsky, ibid., Part A - 2 , 9, 127 (1971).
polymer crystallinity. This work was carried out to investigate that relationship further, and also to elucidate the dependence of the reactivity ratios o n the components of the Ziegler catalyst. We feel that this study makes at least two useful contributions. First, extensive efforts are made to use correct experimental and numerical techniques in the collection and evaluation of data. Second, the significance of the reactivity ratios and the reactivity ratio product as defined in the terminal copolymerization model5 is explored with regard to its correlation with the actual properties of the polymers obtained when they are produced by catalysts which commonly have multiple active polymerization centers. Experimental Section (A) Materials. The polymerization solvents, n-heptane (Phillips, 99 mol 2 minimum) and chlorobenzene (Matheson Coleman and Bell, bp 130-132”), were dried by passage through a column of 4A molecular sieves and were then stored under nitrogen until used. The n-heptane also obtained additional purification by treatment with silica gel. Oxygen and water were removed from the ethylene and propylene (99 mol purchased from J. T. Baker) by passage at 100 psig
z,
( 5 ) G. E. Ham, “Copolymerization,” Interscience, New York, N. Y., 1964, p 3.
ETHYLENE-PROPYLENE COPOLYMERS 483
Vol. 4, NO.4, July-August 1971
through a bed of powdered CuO held at 250" followed by a bed of 4A molecular sieves. The VCla, VOCl3, arid VO(OBU)~ employed in this study were purchased from the Stauffer Chemical Co. and used without further purification. The VCLI had a minimum purity of 9 5 % , with the remainder being primarily voc13, while the VOCL and VO(0Bu)s had purities of 9 7 z , with the contaminants being primarily CIZ and butanol, respectively. Recrystallization from benzene was used to purify the VO(acac),, which was also purchased from Stauffer. Elemental analysis indicated that the recrystallized material had 98% of the theoretical vanadium content. VO(acac)zC1and VO(acac)C12 were synthesized by the method given by Funk, et a/.,6 and elemental analysis for V and C1 gave the following results (i7, V calcd, found; % Cl calcd, found): VO(acac)2C117.0, 17.0; 11.8, 12.2; VO(acac)Clz 21.5; 22.5; 30.0, 28.7. Texas Alkyls supplied all of the aluminum alkyls, which were used as received. 'Typical purities were at least 96z. (B) Polymerization Procedure. A flow sheet of the polymerization unit is shown in Figure 1. The reactor was a 3-in. diameter cylindrical glass reaction flask containing a jacket through which water was circulated by a temperature control system to keep the polymerization temperature at 26.5 =k 0.5". A stirrer with two Fawcett mixed-flow impellers, 1.875 in. in diameter, located 2 and 5 in. from the bottom of the reactor, agitated the reactor contents at 3000 rpm. At the start of a run, 1 1. of solvent was pumped into the reactor and saturated with an ethylene-propylene mixture metered in the desired proportion through calibrated rotometers. Total gas flows of 3-5 I./min were used. When saturation was complete. the copolymerization was initiated by pumping catalyst and cocatalyst solutions frcm their reservoirs into the reactor at a rate of 0.5-1.0 cm3/min. The catalyst and cocatalyst were fed as dilute solutions ( M ) in heptane, ex1:ept for VO(acac)z, which was dissolved in dry benzene, Samples 'equal in volume to the amount of catalyst and cocatalyst solution fed were withdrawn at 15-min intervals and were analyzed gravirnetrically for copolymer content by evapoiation of the solvent. (C) Compositional Analysis. The composition of the polymer was determined by infmred spectroscopy (Beckman IR5A) using pressed films and measurement of the ratio of the 1155-cm-1 methyl band to the 720-cm-l methylene band. The method was calibrated by preparing 14C-tagged polymers. Two different catalyst systems were used to prepare the polymers to check for sequence distribution effects; in addition, both tagged ethylene and propylene were used. The accuracy of the calibration based on this study was *5 (absolute of the ethylene value at 95 % confidence. The precision of the ir measurement itself was i l absolute at 9 5 z confidence. Nmr analysis gave results in agreement with the lacdata, thus somewhat reducing the stated uncertainty. These results will be reported in detail elsewhere.' (D) Characterization. In addition to the determination of the average composition of the copolymers, certain of the samples were subjected to more detailed analysis. Column elution fractionation was performed using an ethanol-heptane solvent system. Fractions were analyzed for composition and molecular weight. Selected bulk samples were analyzed for crystalline fraction by techniques detailed elsewhere;4 annealing schedules were the same as reported therein. Orily calcrimetric results are reported here. These data were obtained using the du Pont differential scanning calorimeter (dsc). In some cases, the samples studied were prepared in a continuous-flow stirred tank reactor at monomer ratios corresponding to those present in the experiments described herein,
z
Figure 1. Polymerization unit as described in text. from monomer concentration-polymer composition data is usually carried out by keeping the dissolved monomer equilibrated with a gas phase of known composition so that monomer concentrations are constant throughout the polymerization. This primarily requires that the monomer gases be passed through the solvent at a rate much higher than their rate of reaction. However, this alone does not ensure saturation of the solvent. High stirring rates, low solution viscosities, and low reaction rates are also required t o prevent the monomer concentration from being affected by mass transfer. With most of the catalyst systems described in this paper, very high initial reaction rates result from batch catalyst addition and cause deviations from saturation. For this reason, a continuous catalyst feed method was used, in which separate solutions of the vanadium component and aluminum alkyl component are continually fed to a well-stirred reactor containing solvent previously saturated with a mixture of the monomers. The monomers are also continuously added in large excess. Once the copolymerization begins, the rate a t which active sites are initiated from the catalyst feed is quickly balanced by the rate at which active sites are destroyed by termination reactions. I n this steady state, the growing chain concentration is fixed and the polymerization proceeds at a constant rate which, by adjusting the catalyst feed rate, can be made low enough t o ensure the absence of significant mass-transfer effects. The typical copolymerization rate data given in Figure 2 show that after a n unsteady-state period during which polymerization poisons are consumed by the catalyst and the catalyst concentration increases from zero to the steady-state value, the relationship, between polymer concentration and I
2.5
I
I
I
I
I
-
Results (A) Reactivity Ratios. F o r solution copolymerization of ethylene and propylene, the determination of reactivity ratios (6) H . Funk, W . Weiss, and M . Ziesing, Z . Anorg. A l l g . Chem., 296, 36(1958). (7) I. J. Gardner, C. Cozewith, and G . Ver Strate, Rubber Chem. Technol., in press.
0
I
I
I
I
I
I
1
484 COZEWITH,VER STRATE
Macromolecules
TABLEI EFFECTOF REACTION TIMEON COPOLYMER COMPOSITION Catalyst system VO(acac)lC1voc11AlEtzCl Al(i-Bu)zCl Time, Polymer Time, Polymer min comp" min compa 75 85 95 105 115
Vanadium feed rate, mmol/hr Ethylene feed rate, l./min Propylene feed rate, l./min Polymerization rate, .M.hr) Ethylene conversion,
70 85 100 115 130 145 160
0.073
0.457 0.464 0.448 0.448 0.456 0.452 0.460 0.047
1.25
0.40
3.75
3.60
14.5
z z
Propylene conversion, a
0.629 0.637 0.620 0.615 0.629
3.07
9.0
4.7
1.5
0.62
that the Al/V ratio, above a certain minimum value, does not affect the copolymerization rate or copolymer composition. Under our steady-state polymerization conditions, the copolymerization equation has the form
where ml and in2 = the mole fractions of monomers 1 and 2 in the copolymer; M I and M Z = the concentrations of monom e n 1 and 2 in solution, and the reactivity ratios ri = k f i / k t j , etc., where the kgj's are rate constants for addition of monomer j t o a n i-ended chain. If the solution is in equilibrium with a gas containing mole fractions y l and y2 of monomers 1 and 2 , then
where Kl and Kt are the gas-liquid equilibrium constants for monomers 1 and 2. Making use of eq 2 and of the relationship
ml
+ mz = 1
(3)
eq 1 can be transformed t o
m,
Mole fraction of C2H4.
+
time is linear and the copolymerization rate is constant. The copolymer compositions corresponding t o the points in this figure are shown in Table I along with the polymerization conditions, As can be seen, the composition is not a function of polymerization time. Runs made at a constant monomer feed composition but varying catalyst feed rates indicate that under our experimental conditions mass-transfer effects were negligible if the polymerization rate was below 20 gj(1. hr) and a polymer concentration of 4-5 was not exceeded. A possible drawback of this continuous feed method is that more aluminum alkyl halide is fed than is necessary t o produce active catalyst. If either the reactivity ratios or polymerization rates change with AljV ratio, then steady-state polymerization may not be obtained. The results shown in Figure 1 and Table I indicate that this is not the case, in agreement with the results of other author^,*^^ who also found
1
= ___
+ 1]/[(3)(;) + 1 1 (4)
[(%)(E)
Because of the high degree of agitation in the reactor, the gas phase under the liquid surface was assumed t o be perfectly mixed and equal to the outlet gas in composition. This composition, which was the equilibrium value of Y C , H ~ / Y C ~ ~ ~ used in eq 4, was obtained by correcting the inlet monomer feed ratio for the small amount of the monomers reacted (less than 10% of the ethylene and less than 3% of the propylene). A value of 4.3 was used for K c ~ H ~ / K cin~ F heptane J~ solvent at 26.5", as given by Sherwood and Pigford,'O while in chlorobenzene at 30" a value of 5.2 was used, as given by Lukach and Spurlin.ll Reactivity ratios were determined by fitting the comonomer composition-copolymer composition data, which are shown in Figures 3-6, t o eq 4 by means of a nonlinear least-squares computer program that determines the values of the reactivity
8-
.7
7-
-
3: .*a
5e
.l-
E
Y
2
.4-
Y
-
C U R M CATALYST SYSTEM
i1
C U R M CATALYST SYSTEM
2 ETHYLENE IN FEED, Vc
,
2 4
/WCzH4
Vc
I
I
I
I
I
I
)
36
Figure 3. Feed ratio-polymer composition data with least-squares fits. (8) J. Obloj, M. Uhnat, and M. Nowakowska, Vysokomol. Soedin., 7,939 (1965). (9) E. Giachetti, P. Manaresi, R. Zucchini, and S. Bacciarelli, Chim. Ind. ( M i l a n ) , 48, 1037 (1966).
(10) T. I
