Water Effects on the Zwitterionic Polymerization of Cyanoacrylates

Makromol. Chem. 190, 1613 - 1622 (1989) 1613 Water effects on the zwitterionic polymerization of cyanoacrylates Ighodalo C. Eromosele, David C. Pepper: Bernard Ryan University Chemical Laboratory, Trinity College, Dublin, Ireland (Date of receipt: December 28, 1988) SUMMARY This paper presents a number of qualitative and semi-quantitative observations on the effect of small concentrations of water on the polymerizations of butyl cyanoacrylates by tertiary amines in THE It reports also that, in the absence of other bases, large concentrations (approx. 1 mol/l) of water can cause the polymerization of ethyl cyanoacrylate, in THF, even in presence of normally inhibiting amounts mol/l) of p-toluenesulfonic acid. A formal kinetic scheme of a stationary-state polymerization, initiated by hydroxyl anions, is presented and discussed. Introduction In the ‘classical’ studies of carbanionic polymerization of styrene and other nonpolar monomers, which led to the detailed understanding of the kinetics and mechanisms of ‘living’ polymerizations, consistent results were obtainable only after the most careful exclusion of traces of nucleophilic ‘carbanion killers’, notably water, which caused severe reductions in rate and molecular weight. In contrast, the polymerizations of alkyl cyanoacrylates, also propagated by carbanionic species, show no such sensitivity, and can even be conducted in presence of high concentrations (e. g. ca. 1 mol/l) of water. An explanation that water reacts but causes transfer (since the hydroxyl ion is known to be a rapid initiator) is not adequate since molecular weights are also largely unaffected. This inertness of the cyanoacrylate anion to potential reactants other than its monomer can be given a general explanation as the consequence of a very high stability conferred by the simultaneous presence of two electrophilic substituents, the CN and COOH groups. This insensitivity is, however, insufficient to resist interaction with appreciably acidic molecules, strong acids causing termination and weak acids transfer reactions. With water the interaction is weak enough to have been ignored as insignificant in the usual type of experiment using conventionally ‘dry’ solvents. Nevertheless the potential roles of water deserve consideration if only to clarify the apparent paradox between the belief that water adsorbed on solid surfaces is the likely initiator of the rapid polymerization of the cyanoacrylate adhesives and the knowledge that the adhesive compositions contain up to 200 ppm of dissolved water (plus smaller concentrations of acidic stabilizers). This paper presents a number of so far unpublished observations and partial investigations, and attempts a correlation in terms of a theory of orthodox character (hydroxyl ion-initiated, proton terminated), which is normally suppressed in stabilized, dry systems, but can become apparent at sufficiently high water concentration. 0025- 1 16X/89/$03.00 1614 I. C. Eromosele, D. C. Pepper, B. Ryan Experimental part The methods of purification and handling of reagents, monomers and solvents, etc., have been described in refs. Rates of polymerization were followed by the calorimetric method of refs. 2 , 3 ) , and molecular weights derived from GPC2). Results The effect of water on anion-initiated polymerizations of butyl cyanoacrylate (BCW Butyl lithium was the initiator with which the original observations of insensitivity to water were made5).These were semi-quantitative, in that they amounted to the finding of 'virtually identical' thermograms ( A T vs. reaction time curves) when successive aliquots of monomer were added to a BuLi/THF solution, even though it had been exposed to intervening injections of water-saturated air. Tetrabutylammonium acetate (TBAAc.) and hydroxide (TBAOH) 6 ) . These, and other TBA salts, are not easily dispensed into T H F without also a small concentration of water, since the anhydrous salts are insufficiently soluble at 20°C, and TBAOH is difficult to dehydrate from its normally-supplied 40% aqueous solution. Reassurance that the quantitative effects of these initiators were not significantly affected by approximately equivalent amounts of water was obtained by the following experiments: mol/l in T H F at 20°C, With TBAAc. at 5 , 9 . lo-' mol/l and [BCA] = 5 , 2 . [H20] = [TBAAc.] though uncontrolled within a factor of 2-3, three replicate experiments gave visually identical thermograms, i. e. overall rates. The derived values of the propagation rate constant differed by only 4%, to be compared with up to 50% variation commonly found as a result of imperfections in monomer purification and other experimental errors. To observe a substantial effect (30%reduction in rate) excess water up to mol/l had to be added, - 170 times the concentration of the initiator! In TBAOH solutions, dispensed as the 40% aqueous solution, the ratio of [H,O]/[TBAOH] is 2 2 1 . In a series of reactions in T H F at 20°C, with [BCA] = mol/l, it was found necessmol/l, and [TBAOH] = 0,6 - 4 3 . 0,6 - 1,3 . ary to add extra water up to ca. 10 mol/l (ca. 18000 . [TBAOH]) before appreciable consequences (20% reduction in rate) could be detected. At [H,O]/[TBAOH] > lo6, the rate was still 20% of the 'dry' value, and the polymer molecular weight unaffected. At the highest [H,O], 3 mol/l, water alone caused polymerization. -' Water effects on initiation by tertiary amines Limited results so far suggest different responses from the two cyanoacrylate esters, ethyl and butyl, in that no clear-cut effects have been observed with BCA, but quite striking behaviour with ECA in T H F initiated by triethylamine or by pyridine. Thus Johnston') reported no effect of H,O up to 1 v01.-Vo on the rates of polymerization of BCA in T H F by Et,N or by pyridine. With ECA he found a small increase of rate, with Et,N, as [H,O] was increased up to approximately 1/2[Et,N], but no further effect beyond this ratio. Water effects on the zwitterionic polymerization of cyanoacrylates 1615 With ECA and pyridine, and considerably higher concentrations of water, Ryan’) found a slight accelerating effect on the overall rate but a surprisingly strong influence on the rate of initiation, as judged by a sharp reduction in the inhibition periods (ti) in acid-inhibited reactions. The results are shown in Tab. 1. If interpreted by the theory developed for the ‘dry’ reactions2), they would imply a strong effect of water to increase the rate constant of the composite initiation process by pyridine, and little effect, possibly a small decrease, in the propagation rate constant. Initiation by addition of pyridine to form zwitterion might certainly be expected to be accelerated by water through a purely physical effect, since it must increase the dielectric constant of the medium and favour the necessary charge-separation. But it seems likely that the large, more than fortyfold, increase indicated in Tab. 1 is more than should be expected from this source. A more plausible interpretation may be that there is a shift in the chemistry of the initiation towards the mechanism suggested by Johnston for the water effect in the Et,N/ECA system, i. e. water-cocatalysed initiation by O H -, giving in the present case: Py + H,O =+ PyH+ + OH- M ----+ HOM- Apart from its interest as the anionic analogue of water-cocatalysed cationic polymerization by Friedel-Crafts halides, this suggestion deserved further consideration since it would imply also a change in the propagating species, now a simple carbanion not a zwitterion. The polymer product, after acidification in work-up, would be a neutral macromolecule and not a salt. There would also be expected to be kinetic consequences. These are considered in detail in a later section. The effect of water alone, i. e. in the absence of other bases Information to date is confined to a single observation with BCA and a short series of experiments with ECA both in T H F at 20 OC’). With BCA (containing ca. mol/l TSA as stabilizer), no reaction was observed on addition of ‘traces’ of water, and only a very slow (days) polymerization in presence of 10 vo1.-070 of water giving polymer of molecular weight ca. 20000. ECA at similar high concentrations of water gave measurable rates even in presence of higher acidic concentrations. Quantitative interpretation is however difficult, since under these conditions the systems tend to become inhomogeneous, i. e. a polymer-rich phase separates as the reaction proceeds. This tendency was found to be reduced when a slightly unconventional method of preparation was used, i. e. by the addition of dry acidified THF/monomer to the aqueous T H F mixture. Tab. 2 summarizes the findings. The complication of phase-separation must mean that these results have little quantitative significance, but qualitatively they seem definitely to imply that water can both start and stop polymerization chains. The obvious first postulate to make is that of a hydroxyl ion-initiated polymerization terminated by the conjugate acid H,O . Under neutral conditions, when these are equivalent, one might expect relatively slow initiation and fast terminations, and hence a stationary state polymerization. In + d, ') b, a) ti /s 21,8 12,6 lo5. [TSA] mol . dm - 3 190 190 0 0 - - OS 82 1,o 170 ki mol-2. 2,88 4,42 12 - 1,66 12. i), s-1 a) 0,30 0.34 - q/s-l b, k, 1.8 1.9 - l.mol-' . s - ' 0,083 mol/l, [Pyl0 = 5 - = Initiation constant in Ri = ki [Py] [MI2, where ki = [TSA],/(f, [Pyl0[MI (cf. ref. ')). (, = final gradient of 1st order plot. I k, = propagation rate constant = uf/(ki [Py], [MI2dt). Mp = 'peak' rnol. wt. (relative, i. e. = Mp of poly(BCA) having same GPC elution volume). 0 190 10 190 290 0 H2O Vl-7 0.00 Tab. 1. Effect of water on polymerization of ECA by pyridine in THE [ECAJ, p-toluenesulfonic acid c) mol/l, T 20°C. TSA 0,94 - 2,6 10-6.Mpd) = = 7 3 Y ;a m 0 P 0 1617 Water effects on the zwitterionic polymerization of cyanoacrylates Tab. 2. Polymerization of ECA in THF by water 10’ . [ECA] 10’ . [TSA] m~l.dm-~ m ~ l - d m - ~ Hzo u, Val.-% a) - (B) ECA/THF/TSA added to THF/H20: 8,3 071 10 83 091 15 8,3 0,1 16,2 8,3 091 17,5 8.3 0.1 20 b, ‘ ) Remarks S-’ (A) ECA added to THF/H,O/p-tol. SO,H (TSA): 0 10 v. low 6,5 8,3 0 30 2,M 8,3 0 20 06 8,3 071 20 0.35 8,3 1 9 0 20 0,08 a) Mpb’ v. low 0,02 0,15 0,20 0,24 8 200 95 OOO 22 OOO 50 OOO 43 OOO 39 OOO Polymer separates Polymer separates Polymer separates Homogeneous up to ca. 20% Homogeneous Homogeneous Homogeneous Homogeneous ‘) Homogeneous ‘) U , = ‘final’ gradient of 1st order plot, but gives an approximate half-life of the whole reaction, i.e. t,,2 = 0,7/U,. M p = peak mol. wt. on GPC trace. Till late in reaction. presence of other acidic or basic species, the ionization equilibria and consequential rates will be profoundly affected. The kinetic scheme given below attempts to formulate some of these possibilities. Theory of stationary state polymerization by water Two simplifying assumptions are postulated: 1 ) The concentrations of initiating and terminating ions are determined simply by the water-ionization equilibria. 2) The re-initiation process after transfer is sufficiently fast to permit the stationary concentration of propagating species to be determined simply by the balance of initiation and termination. Kinetic Scheme: Kw Ionization 2H,O H,O+ + O H - [OH-] + = [H,Otl M Initiation OH- Propagation P,- + M P,- Transfer Termination +M P,- + H,O = K F [H,O] k, +P,- - P,- + H,O+ kP P2- k, P,;, kl, __* kl P,H L P,H + OH+ H,O I. C. Eromosele, D. C. Pepper, B. Ryan 1618 A stationary concentration in [P-] would be determined by: d[P-]/df = ki[OH-][M] i.e. [P-I,, = - kt[P-][H30+]= 0 (ki/kt)[Ml[OH-l/[H30+] giving a stationary rate of monomer consumption: -d[M]/dt = (kikp/k,) [M]2[OH-]/[H30+] and ‘instantaneous’ degree of polymerization: Different consequences must be expected when the ionization is influenced by the presence of acid (HA) of base (B), thus: Case 1. Neutral conditions (as above) Here [ O H - ] Rate = [H,O+] = Kk/2 [ H 2 0 ] (kikp/k,)[MI2 = independent of [ H 2 0 ] m, = k, [M]/(klK&’2[H20] + kt, [H20]) (1) (2) The same form of equation for rate could be expected to apply to polymerization in any solvent (S) sufficiently basic to make the preferred ionization process: H20 + S = SH+ + OH- giving [OH-] = [SH’] = K1’2([H20] [Sl)’’’, cancelling in the rate equation, but the equation for DP should have mixed exponents in [H,O] i.e. = kp[M]/(k,K1/2 [S]1/2 [H,0]’/2 + k,,[H20]) Case 2. In presence of strong acid, HA + K,[H2012/[HAl Here [H,O+] = [HA] ; [OH-] giving Rate = ( k i k p / k , [MI2 K , [H2OI2/[HAl2 ) m, = k,[Ml/(ktWAl + kt,[H201) Case 3. Both acid, HA, and base, B, present Here the equilibria must also satisfy: B + H,O Kb & BH+ + O H - (3) Water effects on the zwitterionic polymerization of cyanoacrylates 1619 and the general case will give complicated equations. But under most practical conditions, where [ H 2 0 ] % [B], > [HA], = [BH+], they will reduce to Ploand [Bl z ( P I 0 - [HA],) f Effects o conversion The above equations refer to the instantaneous rates and degrees of polymerization at the concentrations specified, i. e. are appropriate for the initial behaviour in slow reactions but not for the high conversions in the rapid cyanoacrylate polymerizations. Monomer conversion: If all other concentrations remain unaltered, the equations are simply adapted by integration over thefall in [MI giving the fractional conversion, Y = I [M]/[M],, and the integrated DP, over conversion Y, related to its initial value DP,,(O)in the ratio Y/ln(l - Y ) - ' (ref.')), e.g. ~ - at Y = 0,5 09 0,99 m, = 0,721 . 0,390. 0,215. m,(O) m,(O) m,,(O) Acid consumption: Here, since [HA], is normally very small, and yet k , [HA] % k, [MI, the consequence is virtually complete suppression of -d [M]/dt until all [HA] has been consumed, i. e. an inhibition period, after which the kinetics should rapidly converge to those of the neutral condition. The rate of acid consumption during the inhibition period has an unusual inverse form, viz.: -d[HA]/dt k,[HA] [P-I = or in presence of B = k , [MIK, [H2012/[HAl = k i [MI Kb [B], [H20]/[HA] which integrates to - [HA]: = 2kt, where k = ki[MloK,[H2012 or [HA]; = ki [MloKb P I 0 [HzOl giving a definite inhibition period, t i , at which [HA] + 0 I. C. Eromosele, D. C. Pepper, B. Ryan 1620 Note the unusual form, ti proportional to the square of [HA], differing from the inverse first power found when initiation involves addition of base, B, to the monomer ,). Application of experimental results (i) Tab. 1, B = pyridine Here most of the information relates to the effect of water on the inhibition time, t i . As already noted, analysis according to the pyridine-initiation theory of ref. would imply initiation rates at the highest [H,O] perhaps too great to be plausible. Analysed by the present theory, the value of ti can be seen in Fig. 1 to vary approximately with l/[H,O], as required by Eq. (9) above, with a gradient yielding a value for the product kiKb = 1,6 1. mol-' . s - ' . Whether this is a plausible figure can only be guessed, but it can be recalled that ki for initiation by ,Ph,P, expected to be less than for O H - , has been estimated at 100 1 . mol-' . s-' at 20 "C, and K b is quoted at 2 * in aqueous solution. - < 15- c Fig. 1. Polymerization of ECA by pyridirdwater: Effect of [HzO] on inhibition by p-toluenesulfonic acid (10 mol/l). [ECA] = 0,083 mol/l, [Py] = 5 . 1 0 - ~ mol/l, T = 2 0 " ~ -' [H201-'/(l. rnol-'1 For the effect of water on the rate in nominally acid-free condition, Thb. 1 provides only one measurement. The slight increase observed is in qualitative accord with Eq. (6) but can be no real test. An inference that 1% [H,O] is inadequate to make an appreciable change of initiator from Py to OH- could be drawn from the finding that the results in this reaction could still be analysed by the 'slow initiation-no termination' theory of ref. z, to provide a k, value. (ii) Tab. 2, water alone The main importance of these results is the clear qualitative indication that, even in presence of appreciable concentrations of acid, water in sufficient quantities can cause polymerization. The complication of phase separation rules out firm quantitative interpretation, but at a tentative level it might be speculated that the large effect between 10 and 30% H O might result from the squared dependence required by Eq. (4). , The effect on polymer molecular weight described in the second section of Tab. 2 seems systematic enough to warrant a tentative analysis. The degree of polymerization, Water effects on the zwitterionic polymerization of cyanoacrylates 1621 Fig. 2. Polymerization of ECA by water alone: Effect of [H,O] ondegree of polymerization, DP,. [ECA] = 0,083 mol/l, [TSA] = mol/l, T = 20°C calculated crudely from M p/ 125, can be seen from Fig. 2 to lie close to the reciprocal relationship with [H,O] required by Eq. (7) (with one exception, the result at 15% H,O). The gradient would correspond to a value for k , , / k , of 2,4- lo-' from the crude results, or 8,3 . if all samples are (plausibly) taken to have reached the same 95% conversion. Such figures may seem very low, but d o not appear out of the question when compared with k , , / k , = deduced for CH3COOH in the polymerization of BCA by TBAAc '). General conclusions This collection and review of information, fragmentary though it is, permits a clearer, if still tentative, understanding of a number of features of these systems. The demonstration that very large concentrations of water are needed before any strong effect can be seen should help to lay any fears of some investigators that their results may have been invalidated by neglect of the extreme precautions (vacuum handling, etc.) found necessary in the analogous polymerization of non-polar monomers. A second point is that the relative insensitivity of these polymerizations to water is best understood not as an insensitivity of the carbanion but as a consequence of the extreme reactivity of the monomer, which results in propagation being favoured over other reactions to which the carbanion may also be susceptible. It is relevant to note that, while CH,COOH is a quite active transfer agent, concentrations as high as 0,2 mol/l d o not totally suppress the rate of polymerization by TBAAc in THF''). With hind-sight it is obvious that, in any ionic reaction involving water, acid-base equilibria must be important. It seems likely that in the stabilizing action of acid in water-containing monomer samples, its role as a suppressor of O H - initiation is more important than that of inhibitor. How stabilized monomer solutions may nonetheless be polymerized on surfaces still remains unclear, unless it is because in a surface multilayer the local concentration of water is in fact very high. The formal theory here is obvious but may serve to correlate any future systematic studies. These are certainly needed to provide more extensive tests than available from the few experiments considered here. One feature of particular interest to test is the predicted inverse square dependence of the inhibition period in presence of both water and base. There is one unpublished report of such behaviour in the quinoline/ECA system, but further confirmation of such unusual effect is much needed. 1622 I. C. Erornosele, D. C. Pepper, B. Ryan The most direct test of the ideas advanced here would of course be the chemical identification of the polymer end-group, which should change for example, from P y + - to HO- as the water content of reaction mixture was increased. ') *) 3, 4, 6, ' ) ' ) ' ) lo) D. S. Johnston, D. C. Pepper, Makromol. Chem. 182, 393 (1981) D. C. Pepper, B. Ryan, Makrornol. Chem. 184, 383 (1983) J. P. Cronin, D. C. Pepper, B. Ryan, Chem. Ind. (London) 1982, 775 J. P. Cronin, D. C. Pepper, Makromol. Chem. 189, 85 (1988) E. F. Donnely, D. S. Jonston, D. C. Pepper, D. J. Dunn, J. Polym. Sci., Polym. Lett. Ed. 15, 399 (1 977) I. C. Erornosele, Doctoral Thesis, Dublin Univ. 1985 B. Ryan, unpublished work, Dublin 1980 G. V. Schulz, G. Harborth, Makrornol. Chem. 1, 104 (1947) D. C. Pepper, Makromol. Chem. 188, 527 (1987) I. C. Erornosele, unpublished