Methyl Alpha-Cyanoacrylate I. Free-Radical Homopolymerization

VOL. IV, ISSUE NO. 11, PAGES 231-236 (1960) JOURNAL OF APPLIED POLYMER SCIENCE Methyl a-Cyanoacrylate. I. Free-Radical Homopolymerization* A. J. CANALE, E. GOODE, €3. KlNSINGER,t J. R. PANCHAK, KELSO, R. K. GRAHAM$ W. J. R. L. and Research Division, Rohm & Haas Company, Bristol, Pennsylvania The preparation of high polymers of methyl When the monomer is suitably inhibited with a a-cyanoacrylate by free-radical initiation does not Lewis acid against anionic polymerization and when seem to have been adequately described in the precautions are taken against extraneous moisture literature. Procedures for the preparation of the (such as using acid-washed and carefully dried monomer,'S2 stabilization against anionic polymerequipment and conducting all handling operations ization,2 and some properties of the bulk polymer' in a dry box), reproducible results can be obtained have appeared in the patent literature. A discusin free-radical polymerizations. sion of the chemistry of cyanoacrylate monomers as adhesives has a ~ p e a r e d . ~ With the exception of Bulk Free-Radical Polymerization a recent study by Mihail, Lupu, and D a ~ c a l u , ~ Establishment of the free-radical nature of the most of the properties appear to have been measpolymerization was based on the observations that ured on low molecular weight materials initiated by polymerization could be initiated with such water, alcohols, and other weak bases. Such becommon radical initiators as azobisisobutyronitrile havior in the presence of weak bases is similar to (AIBN) and could be inhibited with diphenylthat of vinylidene cyanide5 and nitroethylene.6 picrylhydrazyl. The rate data also appear to Unlike these monomers, however, methyl a-cyansupport a free-radical polymerization. oacrylate (hereafter abbreviated as MCyA) polyPrecise rate data are very difficult to obtain, merizes readily by free-radical techniques to form even when anionic polymerization is suppressed, high molecular weight polymer. In the present as an almost immediate acceleration of rate occurs communication we report some observations on the free-radical homopolymerization of MCyA and some properties of the polymer in bulk and in solu.04 tion. A subsequent paper will discuss the copolymerization behavior of this interesting monomer. TlAL RATE POLYMERIZATION BEHAVIOR = .I398 / h .03 Anionic Polymerization As has been disclosed1v2 as might be expected and from its similarity to vinylidene cyanide, MCyA is readily polymerized by traces of base, even by materials as weakly basic as alcohols and water. [Coover et al.3 have shown that such facile anionic polymerization occurs with monomers possessing two groups exhibiting strong electromeric (-E) effect,~.' Similar behavior was noted8 for purified dimethyl met hylenemalonate , CH,=C (COzC *.] H3) * Presented in part a t the 136th Meeting of the American Chemical Society, Atlantic City, New Jersey, September 1959. t Present address: Department of Chemistry, Michigan State University, East Lansing, Michigan. $ To whom correspondence should be addressed. .02 .01 0 0 250 TIME 500 750 (SECONDS) Fig. 1. Rate of bulk polymerization of methyl a-cyanoacrylate at 6OoC., initiated by azobisisobutyronitrile. 231 232 CANALE, GOODE, KINSINGER, PANCHAK, KELSO, AND GRAHAM at low conversion. This effect is similar to that found for methyl acrylate,*O although the autoacceleration is detected here at even lower conversions. Some turbidity is noticed throughout the polymerization, but no settling out of polymer was found; the polymer formed at high conversions is optically clear. This behavior is somewhat similar to that of methacrylonitrile in bulk.” TABLE I Comparison of kps/kl Values of Various Vinyl Monomers at 6OOC. kp2/kt x 104, 1./mole/see. Vinylidene cyanide Styrene Acrylonitrile Methacrylonitrile Methyl methacrylate Methyl a-cyanoacrylate Vinyl acetate Methyl acrylate .20 - (Table I) of this value with those for severaI other common monomers. .I 5 . r 9 0.116 0.491b 0. 697b 1.44 2.10 18.5 106.0 Reference 5 14 15 15 16 This work 16 16 a Does not appear to form homopolymer with free-radical initiation. Measured in dimethylformamide solution. e Assumes 100% efficiency of initiation. .05 0 0 02 04 06 (AIBN, oe m l l 1 0 1 2 Io5 Fig. 2. Rate of bulk polymerization of methyl a-cyanoacrylate at 60°C. as a function of initiator concentration. The data from a typical dilatometric polymerization are shown in Figure 1; extrapolation to initial rate was made from times of less than 200 sec. A duplicate polymerization was performed at the same time with no AIBN present, and the rate of this control was subtracted from the overall rate to obtain the AIBN-initiated rate to compensate for any spurious anionic or free-radical reactions. It is realized that error could be introduced because of expansion of the confining liquid under nonsteady-state conditions12 and also because the rate of the “control” is competitive with the AIBNinitiated reaction, so that the corrected rate is a relatively small difference between relatively large numbers. A series of these measurements was applied to the determination of the rate-initiator concentration relationship, which is shown in Figure 2 to be: rate (yo/hour) = 175 [AIBNIOJ; rate (fraction/sec.) = 4.86 X 10-4[AIBN]o.5 = R , / [ M ] . From this relationship, from the known valuesI3 for the decomposition of AIBN at 60°C. (1.1 X sec.-l> and arbitrarily assuming 1 0 0 ~ o efficiency of initiation, k P 2 / k ,is calculated as 0.021 l./mole/sec. Although this calculation requires several approximations, it allows a comparison It can be seen that MCyA is quite unlike vinylidene cyanide in its free-radical homopolymerization behavior. Examination of models suggests a lower degree of electrostatic hindrance for addition of MCyA to its own polymeric radical relative to vinylidene cyanide, but the striking difference in behavior is not obviously explained. The relatively high value of k P 2 / k ,may be related to a low value of k,. Mutual electrostatic and steric repulsion would make mutual combination of radicals relatively unlikely. Termination by disproportioiiation would involve abstraction of a hydrogen atom from the ester methyl group (unlikely) or from a p-carbon; examination of models suggests that abstraction of hydrogen from the pcarbon is much more hindered than that, from a methyl group such as that in methyl methacrylate. Inhibition, Branching, and Chain Transfer No detailed study of the inhibition of free-radical polymerization of MCyA has been made. Many of the common inhibitors, especially those which function as antioxidants, contain basic functional groups which initiate anionic polymerization of MCyA. Diphenylpicryhydrazyl and benzoquinone are relatively effective inhibitors. Furfurylidene malononitrile, a very effective inhibitor for styrene” and a relatively effective inhibitor for methyl methacrylateX8 and methyl acrylate, l 9 acts only as a ret,arder of polymerization in this system. An effective inhibitor of MCyA would be expected to act as an electron donor, so t.hat a METHYL or-CYANOACRYLATE. I contribution to the transition state could be made by resonance forms of the type 233 strongly suggests that these polymerizations were not free-radical, but anionic in nature. CN -CH2- L :a-- SOLUTION PROPERTIES - -(Inhibitor)a' AOOCHJ with a consequent lowering of the activation energy. 2o Furfurylidene malononitrile would be expected to act as an electron acceptor in such systemsls and so would be much less effective. Branching by chain transfer with polymer is unlikely with MCyA, since there are no labile hydrogen atoms in the polymer. No crosslinked polymer has ever been detected, even in high conversion polymerizations, but no attempt has been made to detect low degrees of branching. Synthesis of pure model compounds for chain transfer studies has proved difficult. Attempts to regulate the molecular weight of bulk or solution polymerizations with dodecyl mercaptan indicated that little change was effected. Also, conversions seemed to be somewhat higher in the presence of mercaptan. It is conceivable that the mercaptan is capable of acting as a weak anionic initiator. Mercaptans would not be expected to be very effective transfer agents in the radical polymerization of MCyA, since polar contributions to the transition state are less favored than in the case of styrene;21 however, it was unexpected that essentially no control could be achieved. Solution Polymerization An attempt to study polymerization in solution to minimize the autoacceleration effect found in bulk was unsuccessful when benzene was used as a solvent, since the polymer precipitated as a highly swollen gel. Many of the solvents in which homogeneous polymerizations could be conducted are extremely difficult t,o obtain completely free of water or other ionic initiators. Near the completion of this study, it was found that AIBNinitiated solution polymerizations could be conducted in carefully purified isobutyronitrile or nitromethane, but no kinetic data were obtained in these solvents. Mihail et al.4 have described the emulsion polymerization of cyanoacrylate esters initiated with hydrogen peroxide in the absence of emulsifying agent. The molecular weight was reported to decrease with increasing temperature. The extreme sensitivity of MCyA to even traces of water Solvents for the homopolymer include propionitrile, pyridine, nitroethane, nitromethane, acetonitrile, dimethylformamide, ethylene carbonate, succinonitrile, and butyrolactone. Swelling ratios in these solvents were measured on polymer crosslinked with a small amount of ethylene glycol dimethacrylate; a rough estimate of the square root of the cohesive energy density from these data gives a value of ca. 14.5, relatively close to that for polyacrylonitrile. Many of the above solvents are also solvents for polyacrylonitrile. Polymers of MCyA are insoluble in alcohols, which indicates the absence of strong hydrogen bonding. The homopolymer is also insoluble5 in common aromatic solvents (benzene, toluene), ketones (acetone, methyl ethyl ketone), and chlorinated solvents (chloroform, ethylene dichloride). Polymers were purified for analysis by dilution with purified nitromethane and precipitation in the nonsolvent, diethyl ether. Anhydrous ether was used to prevent inadvertant polymerization of remaining monomer. After thorough washing with ether to remove monomer, the polymer was redissolved in nitromethane and reprecipitated into ether or methanol containing a little hydkochloric acid. Preliminary data have been obtained on the intrinsic viscosity-molecular weight relationships for MCyA homopolymer in acetonitrile (a poor solvent) and nitromethane (a good solvent). Measurements were carried out a t 30°, which is below the theta temperature for this polymer in acetonitrile. The data for four whole polymers give the relationships: [q] = 6.43 X 10-5M,0-77 (nitromethane, 30°, ~ w - r a n g 106 to 5 e [TI = x 106) 8.45 X 10-4M,0-45 (acetonitrile, 30°, MW range 5 X lo3 to 3 X lo5) It is hoped to publish more details on the dilute solution behavior at a later date. The present relationships are useful in showing that high molecular weight polymer is formed with radical initiators. Solutions of MCyA polymer in dimethylformamide or dimethyl sulfoxide exhibited a decrease in intrinsic viscosity on long standing at room tem- 234 CANALE, GOODE, KINSINGER, PANCHAK, KELSO, AND GRAHAM perahre or on heating. This behavior is reminiscent of that observed for poly(vinylidene cyanide) in the presence of water and bases.6 BULK PROPERTIES The homopolymer must be prepared from very pure monomer to give clear, transparent sheet castings; traces of impurities can lead to yellowing of the material. Molding of samples for physical testing is difficult because of the thermal degradation discussed below. Dilatometric measurement of the glass temperature could not be accomplished because of degradation at temperatures only slightly above the glass temperature. From the Vicat softening point and A.S.T.M. heat distortion measurements, the glass temperature was estimated to be 165OC. < T o < 17OoC., which agrees well with earlier results.22 TABLE I1 Physical Properties of High Molecular Weight Poly(methy1 a-Cyanoacrylate) No. 1 No. 2 No. 3 ~~ Plex 1 8 1 ~~~ Heat distortion temp., "C., at 264 psi (ASTMD-648-56) 163b 157b 160b 96bo Impact strength, Izod, unnotched, X 1 in. section (ASTM-D256-56) 5.8 6.3 5.6 7.0d Flexural strength, psi (ASTM-D790-49-T) 16,500 15,100 17,700 16,000 Modulus in flex554,000 557,000 530,Ooo 450,000 we, psi Deflection at break, in. 0.46 0.45 0.49 0.6 Reduced specific viscosity, dl./g." 1.8 2.5 2.0 Average values not intended for specification purposes. b Conditioned 48 hr. at 50°C. Not annealed; value increased ca. 15" on annealing. in. secd Estimated from Charpy unnotched on 1/p X tion. 0 In nitromethane at 25°C. at 2 gJ1. a The cast sheet exhibits physical properties comparable with other high molecular weight thermoplastic materials such as poly(methy1 methacrylate). Some typical properties of our best preparations are shown in Table 11, where they are compared with typical data for Plexiglas 11. THERMAL AND IRRADIATION DEGRADATION On heating in air above 18OoC., the homopolymer of MCyA yellows and bubbles. Some monomer is liberated, and the intrinsic viscosity of the residue drops sharply. Under these conditions, most of the relatively nonvolatile monomer cannot escape and is repolymerized to low molecular weight material. The actual weight loss is low. A few i vacuo experiments indicated the decrease in n molecular weight to be less pronounced, but the amount of volatile material increased sharply. The relative ease of thermal depolymerization has been utilized in the synthesis of MCyA monomer from low polymer.'S2 Attempts to add stabilizers or to vary the nature of the end groups have been unsuccessful in preventing this degradation. Whether initiation of depolymerization occurs randomly or preferentially a t weak links has not been determined; possible sources of weak links might be tail-to-tail placement caused by termination of polymerization by combination or ketimine links from polymerization t,hrough the nitrile group. Such ketimine links, previously reported for polymethacrylonitrile,28 could not be detected by ultraviolet spectroscopy. Gamma irradiation in vacua of MCyA homopolymer causes a decrease in molecular weight which appears to be proportional to dose. In Figure 3 are plotted reciprocal viscosity-average molecular weights for MCyA and MMA homopolymers as a function of irradiation time. From the relative slopes of these lines and the known energy requirements for irradiation degradation 12 POLY-(METHYL S - C Y A N O A C R Y L A T E ) 9 _ I (D I 2 . _ . x > I I - 3 POLY-(METHYL METHACRYLATEI - 0 0 4 S 1 2 T I M E OF I R R A D I A T I O N 1 6 20 (HOURS) Fig. 3. Reciprocal viscosity-average molecular weight, l/MwJplotted against time of irradiation for poly(methy1 a-cyanaoacrylate)and poly(methy1 methacrylate). METHYL a-CYANOACRYLATE. I of poly(methy1 methacrylate) (60 e.v. per mainchain scission)124 poly-MCyA appears to require only ca. 18 e.v. per main-chain scission. Not enough data are available to show that some concurrent crosslinking is not occurring; as Shultz has pointed it is possible that such a decrease in apparent viscosity-average molecular weight can be found with some crosslinking also occurring. This behavior would be unexpected, however, in view of the rather general rule that polymers of backbone structure -CHz-CXYonly degrade.25 EXPERIMENTAL Monomer of high purity was prepared by the method of McKeeverlZ8 which the condensate of in formaldehyde and methyl cyanoacetate is treated with acetic anhydride to remove water liberated in the cracking of the condensate. The monomer had the following properties: m.p. l°C., b.p. 46OC. (1.5 mm.), ng 1.4431, dy 1.100 g./ml., dm 1.067 g./ml. The monomer was inhibited against anionic polymerization with boron trifluoride-acet.ic acid complex and stored in a refrigerator in polyethylene bottles . Rates of polymerization were followed at 6OoC. dilatometrically, mercury being used as the confining liquid. The density of the polymer was 1.304 g./ml. at 25O and 1.289 g./ml. at 60°, and the contraction on polymerization was 18.7 ml./mole. In many of the polymerizations, a second dilatometer containing no initiator was filled and degassed at the same time. Rates of both polymeritions were measured and the rate of polymerization of the control subtracted from the overall rate. Reagent grade nitromethane and anhydrous diethyl et.her were used as diluents and precipitants. Dodecyl mercaptan was redistilled before use. Diphenylpicrylhydrazyl and furfurylidene malononitrile were supplied by Dr. J. L. Kice. The intrinsic viscosity measurements were determined with a Ubbelohde-type viscometer at 3OOC. Light-scattering measurements were determined with a Brice-Phoenix light scattering photometer at 3OOC. Many members of the Research Division have contributed t phases of this work. We wish especially to thank Dr. o C. H. McKeever, Dr. H. R. Raterink and Mr. S. Wise for monomer preparations and W. H. Snyder, J. Dzomba and L. Foell for polymerieatlon studies. Irradiation studies were performed using the cobalt-60 source of the University of Pennsylvania; we are grateful to Dr. R. E. Hughes for his assistance. 235 References 1. Ardis, A. E., U. S. Pat. 2,467,926-7 (1949). 2. Joyner, F. B., and G. F. Hawkins, U. S. Pat. 2,721,858 (1955). 3. Coover, H. W., Jr., F. B. Joyner, N. H. Shearer, Jr., and T. H. Wicker, Jr., SPE Journal, 15,413 (1959). 4. Mihail, R., A. Lupu, and L. Dascalu, Rev. chim. (Bucharest), 9, 606 (1958). 5. Gilbert, H., F. F. Miller, S. J. Averill, R. F. Schmidt, F. D. Stewart, and H. L. Trumbull, J . Am. Chem. Soc., 76, 1074 (1954). 6. Weiland, H., and E. Sakellarois, Ber., 52, 898 (1919). 7. Ingold, C. K., Structure and Mechanism in Organic Chemistry, Cornell Univ. Press, Ithaca, N. Y., 1953, pp. 64 ff. 8. McKeever, C. E., and H. R. Raterink, private com- munication. 9. Schildknecht, C. E., Vinyl and Related Polymers, Wiley, New York, 1952, p. 307. 10. Walling, C., J. Am. Chem. Soc., 70, 2561 (1948). 11. Schildknecht, C. E., op. cit., pp. 294-95. 12. Burnett, G. M., Trans. Furaday Soc., 46, 772 (1950). 13. Results compiled by C. Walling, J . Polymer Sci., 14, 214 (1954). 14. Tobolsky, A. V., and J. Offenbach, J . Polymer Sci., 16, 312 (1955). 15. Bamford, C. H., A. D. Jenkins, and R. Johnston, Proc. Royal SOC.(London), 2398, 214 (1957). 16. Matheson, M. S., E. E. Auer, E. B. Bevilacqua, and E. J. Hart, J . Am. Chem. Soc., 71, 497, 2610 (1949); ibid., 73, 5395 (1951). 17. Haward, R. N., and E. T. Borrows, U. S. Pat. 2,650,899 (1953). , 18. Kice, J. L., J . Am. Chem. ~ o c .76,6274 (1954). 19. Kice, J. L., J. Polymer Sci., 19, 123 (1956). 20. Mayo, F. R., and C. Walling, Chem. Revs., 46, 191 (1950). 21. Walling, C., J. Am. Chem. Sac., 70, 2561 (1948). 22. Deutsch, K., E. A. W. Hoff, and W. Reddish, J Polymer Sci., 13, 565 (1954). 23. Talat-Erben, M., and S. Bywater, J . Am. Chem. Soc., 77,3710, 3712 (1955). 24. Shultz, A. R., P. I. Roth, and G. B. Rathmann, J. Polymer Sci., 22, 495 (1956). 25. Miller, A. A., E. J. Lawton, and J. S. Balwit, J . Polymer Sci., 14, 503 (1954). 26. McKeever, C . H., U. S. Pat. 2,912,454 (1959). synopsis Methyl a-cyanoacrylate, when suitably inhibited by Lewis acids against anionic polymerization, can be polymerized readily with free-radical initiators to form hard, clear, high molecular weight polymers. The ratio of k,a/k, is approximately 0.021 a t 6OOC. if 100% efficiency of initiation by azobisisobutyronitrile is assumed. Acceleration occurs a t very low conversion to polymer during bulk polymerizations. The homopolymer is thermally unstable a t temperatures only slightly above the glass temperature (estimated to be 165-170°C.). Degradation of the polymer under gamma irradiation is pronounced, only 18 e.v. being required per main-chain scission. 236 CANALE, GOODE, KINSINGER, PANCHAK, KELSO, AND GRAHAM R6sum6 Zusammenfassung L’a-cyanoacrylate de mkthyle, lorsqu’il est convenablement inhibe par les acides de Lewis vis-his de la polymerisation anionique, peut &re rapidement polym6ris6 par des initiateurs radicalaires pour donner un polymbre dur, clair et de haut poids mol6culaire. Le rapport kp2/k,est approximativement de 0,021 B 60°C si I’on admet 100% d’efficacit6 d’initiation par l’azobis(isobutyronitri1e). L’acc616ration a lieu a des tres faibles degres de conversion durant la polymerisation en bloc. L’homopolymbre n’est thermiquement instable qu’B partir d’une temperature Iegbrement superieure B la temperature de transition vitreuse (estim6e 165-170°C). La dhgradation du polymbre par irradiation gamma est prononc6e e t elle ne nCcessite que 18 6lectron-volts par scission de chaSne. Methyl-a-cyanoacrylat, das durch die Anwesenheit von Lewis-Sauren gegen anionische Polymerisation entsprechend geschutet ist, kann mit radikalischen Startern leicht zu harten, klaren, hochmolekularen Polymeren polymerisiert werden. Das Verhaltnis kp2/k, betragt unter Annahme einer 100% Starterausbeute mit Aeo-bis-isobutyronitril bei 60°C etwa 0,021. Bei der Polymerisation in Substane tritt bei sehr niedrigem Umsatz eine Reaktionsbeschleunigung ein. Das Homopolymere ist bei Temperaturen, die nur schwach oberhalb der Glasumwandlungstemperatur (zu 165-170°C bestimmt) liegen, thermisch instabil. Der Abbau des Polymeren unter Gammabestrahlung ist erheblich; es sind nur 18 Elektronvolt zur Spaltung der Hauptkette erforderlich. Received March 17, 1960 Revised June 14, 1960