J . Chem. SOC., Faraday Tians. I , 1982,78, 75-88 Kinetics and Mechanism of the Quaternization of Poly(4-vinyl pyridine) with Alkyl and Arylalkyl Bromides in Sulpholane BY ERNEST A. BOUCHER* AND CHRISTOPHER C. MOLLETT School of Molecular Sciences, University of Sussex, Brighton BNl 9QJ Received 6th November, 1980 Quaternization reactions of poly(4-vinyl pyridine) with several organic bromides show retardation of reaction in excess of that predicted by second-order kinetics. Kinetic expressions from a neighbouring-group model, in which a reaction'is characterized by the rate constants ki for reaction of a pyridyl group having i = 0, 1 or 2 already reacted neighbours, have been applied to the experimental results. Quantities obtained include k,, K = k,/k,, L = k,/k,, activation energies E,( = E, = E2) and pre-exponential factors A,.The ratios K and L for the bromides range from 0.95 and 0.54, respectively, for ethyl bromide to 0.55 and 0.28 for 1-bromo-3-phenylpropane. On average K = 2L. The dependence of k , values on the nature of the organic group of the bromides is discussed in terms of steric hindrance in the transition state; similar reasoning can be applied to molecules with a single pyridyl group. Linear empirical relationships have been found between K and L and the extended lengths of the alkyl and arylalkyl groups of the bromides. The deviation from second-order kinetics ( K < 1, L < 1) is explained in terms of steric hindrance, involving the transition state of a reacting pyridyl group and the alkyl group on a neighbouring already reacted pyridyl group.Electrostatic effects are not thought to influence the reaction kinetics. In 1955, Coleman and FUOSS~ reported that the quaternization of poly(4-vinyl pyridine), P4VP, with n-butyl bromide in sulpholane departed from overall second- order kinetic behaviour. They found retardation of the reaction which was attributed to a neighbouring-group effect with the reactivity of a pyridyl group depending on whether or not its neighbours have reacted. The quaternization of P4VP by a variety of alkyl and arylalkyl halides is a good example of a polymer-transformation reaction which yields a polyelectrolyte. In a wider context, polymer-transformation reactions have been reviewed2 with emphasis on kinetic and statistical treatments.Several groups of workers3-* have previously studied some of the many possible quaternization reactions of P4VP, but for the reasons given in the following summary few of these studies seemed to be reliable for elucidating the cause of the retardation as reaction proceeds. The general consensus of opinion found in previous reports was that retardation was due to electrostatic effects, although one study attributed retardation to steric hindrance (on unreliable grounds). Fuoss et aL3 assumed that the reactivity of a group with one reacted nearest neighbour was the same as that for a group with neither neighbour reacted. They ascribed all of the retardation to the lower reactivity, due to electrostatic effects, of a group having both neighbours already quaternized.Morcellet-Sauvage and Loucheux4~ also analysed kinetic measurements for n-butyl bromide in sulpholane by the approximate methods of Fuoss and coworkers, and thought that electrostatic effects were involved. Arends9 analysed some of the data of Coleman using more appropriate mathematical expressions, but again attributed the effect to groups having both neighbours already reacted. He formed the opinion that steric hindrance was involved because the activation energy (based on experiments at three temperatures) 7576 QUATERNIZATION OF POLY(4-VINYL PYRIDINE) for reaction of groups with both neighbours reacted was reckoned to be the same as that for a group with two unreacted neighbours. However, the prior assumptions that, in the notation used herein, k, = k, and that k,/k, dQes not depend on temperature ensures that the activation energies are equal: E, = El = E,. Boucher and MollettlO and Boucher et d.ll have analysed experimental data for the quaternization of P4VP with several alkyl bromides in sulpholane using rigorously derived kinetic expressions based on the neighbouring-group model involving rate constants k,, k, and k,.These studies showed that the existence of one reacted neighbour affected the reactivity of a group, i.e. k , # k,. Furthermore, for reasons which are elaborated on below, it was concluded that the retardation is due mainly to steric hindrance and that the magnitude of the effect increases with increase in the size of the organic bromide. In a complementary study,12 where viscosity changes of the reacting systems were measured as a function of the extent of reaction, it was found that the addition of N-ethyl pyridinium bromide suppressed expansion of the polymer molecules, but did not usually affect the kinetics.This provided a major piece of evidence against the importance of electrostatic effects in causing retardation. We have extended the kinetic study to a wider range of alkyl and arylalkyl bromides and are able to quantify the neighbouring-group effect. The use of different organic bromides also throws more light on the fundamental nature of these examples of the Menschutkin reaction. Many of the older studies of the quaternization reactions of small molecules were not particularly thorough, but the essentially S,2 character was no doubt correctly deduced.A mystery regarding many quaternization reactions has been the nature of the colours sometimes found as quaternization proceeds (reported, for example, by Hantzsch13). A separate account will show that the products of the quaternization reactions involving polymer and small pyridyl molecules are probably charge-transfer complexes, the electronic absorption properties of which depend on the species of pyridyl molecules, the halide species and the reaction conditions (e.g. temperature and solvent species). NEIGHBOURING-GROUP MODEL The development of neighbouring-group models has recently been reviewed in detail and the most important kinetic expressions summarized.2 However, it is appropriate to give a brief account leading to the method of data analysis used in this study.The pyridyl groups along a polymer chain are represented by circles, where 0 denotes an unreacted group and 0 denotes a reacted group. A portion of the chain, . . . 0 0 0 0 0 0 0 0 0 0 0 0 . - t t t k2 kl k0 signifies that the probability of reaction of a group in the time interval t to t+dt is k, dt, k, dt or k , dt, depending on whether none, one or two neighbours of the group have already reacted at time t. The fractional extent of reaction of the pyridyl groups for large molecules (strictly for the number of monomer units rn -+ a) is given by2 5 = 1 - 2KaL J: (1 - exp [ - 2( 1 - K ) (1 - u)] du + 2aAJ: (1 - u) ~ 2 ~ - ~ exp [ - 2( 1 - K ) (1 - u)] du -(2-a) a2Kexp[-2(1 -K)(l--a)].E. A. BOUCHER AND C . C . MOLLETT 77 Integration can either be carried out by numerical computation or analytically, as here, after the exponentials have been expanded.The time variable a in eqn (1) is given r rt 1 by and K = k,/k,, L = k,/k,. (3) In eqn (2), c, and c: are the instantaneous and initial values of the concentration of the bromide reagent. When c, = c: is constant, a = exp ( - k , t), and eqn (1) would apply to a unimolecular reaction or to a pseudo-first-order bimolecular reaction. In the present reactions, eqn (2) allows for the depletion of reagent as reaction proceeds. An overall second-order bimolecular reaction corresponds to K = L = 1. The pro- cedure used to obtain a consisted of plotting c,/c: against actual time t on a large scale and using this for integration over suitable time intervals, with k, having been obtained separately in the limit t -+ 0 from second-order plots.Plots according to the usual form f = (c: - c;)-l In (c: c,/c:c,) = k, t (4) are expected to curve towards the time axis when there is retardation over and above that due to reducing concentrations cp of polymer groups and c, of reagent. We estimate the limits of precision on individual k, values to be ca. & 2%. However, values quoted for a given halide species under standard conditions are obtained by interpolation of values at several temperatures and are more precise (table 1). When eqn (1) is expanded and integrated, K(l -aa+l)+[K(b-2)+ 1](1 -aa+2) a + 2 < = 1 -2aLe-b -(2-a)a2Kexp[-b(l-a)]; a = 2K-L-1, b = 2(1-K). (5) This series converges sufficiently rapidly for only the first eight terms in a to be required.Having obtained a suitable range of values of a for a particular reaction, the next step in comparing theory with experiment is to evaluate r from eqn ( 5 ) for several pairs of K and L values. To do this, large-scale plots were made of robs against tabs, and the theoretical plots of 5 against t superimposed. The ' goodness of fit' was judged by eye and tested by evaluating the deviation of the experimental points from the theoretical plots. We have paid particular attention to testing the reliability of K and L values by systematically varying k,, K and L from best-fit values. It is not possible to illustrate the changes on small-scale reproductions, but it clearly emerges that if k , has an unreliable value, then changes in K and L will not compensate for this and restore the goodness of fit.Furthermore, K dominates the shape of the curves in the approximate range r = 0.3-0.7, and L is dominant over the range 0.7-1 ; Kand L values are not strongly correlated. To test the above procedure, deviations between the fitted curve and experimental values of < have been expressed as the standard deviation CT for a few temperatures. Values of o confirmed that best-fit curves were obtained; typical values were for 2-bromoethylbenzene at 344.1 K. = 7 x for i-butyl bromide at 338.9 K and 6.8 xro C TABLE 1 .-SUMMARY OF KINETIC RESULTS FOR QUATERNIZATION OF POLY(4-VINYL PYRIDINE) IN SULPHOLANE 4 lo5 k, range /dm3 mo1-I s-' K L 107 A , EO temp. no. of 4 range r-t range key / K expts /mol dm-3 reagent letter /mol dmP3 at 340 K (f0.03) (k0.03) /dm3 rno1-I s-' /kJ mol-l 306.2-343.2 307.9-348.2 308.6-353.2 324.2-349.1 306.2-353.2 304.3-323.7 308.5-349.0 323.7-348.0 306.2-328.2 308.4-349.4 3 13.0-348.9 3 30.0-348.9 9 9 9 6 4 8 8 7 4 9 6 6 0.049-0.227 0.107-0.159 0.105-0.21 4 0.087-0.107 0.17 1-0.195 0.034-0.036 0.105-0.107 0.105-0.106 0.042-0.044 0.105-0.108 0.104-0.106 0.104-0.106 ethyl bromidea n-propyl bromide n-butyl bromideb n-pentyl bromide n-hexyl bromidea ally1 bromide i-butyl bromide bromometh ylcyclohexane benzyl bromidea 2-bromoethylbenzene 1 -bromo-3-phenylpropane 1 -bromo-2-phenylpropane A B D F G C E H I J M N 0.200-1.869 0.383-0.635 0.215-0.428 0.380-0.394 0.195-0.378 0.I09 0.406 0.383 0.392 0.382 0.382 0.042-0.066 91.02 k 0.90 0.95 0.54 31.57f0.31 0.80 0.38 27.47 _+ 0.26 0.70 0.3 1 31.28k0.42 0.65 0.30 33.13 k 0.32 0.55 0.27 3736k5.2 0.80 0.40 1.721 fO.O1 0.60 0.37 1.22k0.016 0.65 0.33 5730 k 50 0.75 0.34 16.01 f0.15 0.65 0.32 29.16f0.23 0.55 0.28 0.602 & 0.01 0.60 0.32 3.25 k 2.52 0.45 & 0.30 0.34f 0.24 0.42 k 0.3 1 0.44k0.32 2.84 1.30 0.028 k 0.01 0.01 7 k 0.010 0.67 k 0.52 0.16 f 0.08 0.29 0.19 0.0086 & 0.004 68.8 & 3.3 66.2 4.3 65.8 k 4.0 66.1 & 1.1 66.0 &- 3.5 57.9 4 1.2 66.6 4 0.9 66.1 k3.1 52.6 k 4.9 65.3 k 2.3 65.2 & 2.3 66.2 4 2.7 a Analysis of Groves' data; analysis of Fletcher's data.=3 0 z 0, v 0 rE. A. BOUCHER AND C. C. MOLLETT 79 EXPERIMENTAL The basic techniques of polymer preparation, and of making kinetic measurements, based on bromide ion determinations by Volhard’s method, have been described previously.There are, however, new reagents involved, and some aspects of technique are emphasized. The concentration ranges of the polymer and reagent (organic bromide) species, the reaction temperature ranges and the number of kinetic runs are given in table 1. The purification techniques are mainly based on the monograph by Perrin et all4 Ally1 bromide (Aldrich) was purified by washing with aqueous sodium bicarbonate (10% w/v) and distilled water, dried over CaC1, and distilled (298 K at 130 Pa). n-Pentyl bromide (Aldrich) was washed with concentrated H,SO,, water, aqueous sodium carbonate (10% w/v) and water again, then it was dried (CaCl,) and distilled (305 K at 130 Pa). Neopentyl bromide (Fluka) was used without further purification.Bromomethylcyclohexane (Aldrich), was dried (CaCl,) and distilled (303 K at 13 Pa), as were 2-bromoethylbenzene (Aldrich, 307 K, 13 Pa), 1 -bromo-3-phenylpropane (Aldrich, 308 K, 13 Pa) and 1 -bromo-2-phenylpropane (Aldrich, 310 K, 13 Pa). All reagents were stored in the dark. The basic method of making kinetic measurementslo’ll was modified for the very rapid quaternization reactions of poly(4-vinyl pyridine) with allyl bromide. A known weight of P4VP was dissolved under dry N, in a known volume of sulpholane. The reaction vessel was equilibrated in a thermostatted water bath for 1.8 ks (30 min), and reagent added from a calibrated syringe. At intervals 2 cm3 aliquots of reaction mixture were removed with a syringe (calibrated at the reaction temperature) and the bromide ion concentration determined by Volhard’s method.The samples of bulk free-radical polymer were not dissolved in t-butyl alcohol during purification, as has been customary in previous studies.lo9 l1 Khosravi-Babadi16 has shown that polymer recovered from t-butyl alcohol contains entrapped solvent. The polymer used in the present study was extruded into toluene (2 dm3) and then thoroughly washed with fresh toluene. Elemental analysis of the polymer samples has been found to be unreliable, attributed to incomplete combustion of the samples. However, reaction involving the samples of P4VP have consistently reached fractional extents ( of reaction of 0.99 kO.01, which is taken, together with u.v., i.r. and n.m.r. spectra, as a good indication of polymer purity.The viscosity-average molecular weight of the polymer was determined in absolute ethanol at 298 K using Mark-Houwink constants,16 K = 2.5 x cm3 g-l, a = 0.68. The molecular weight was (576 & 14) x lo3 for the polymer recovered by extrusion into toluene, compared with (301 f 6) x lo3 for samples involving recovery from t-butyl alcohol. RESULTS Table 1 shows the kinetic results for the quaternization of P4VP with several alkyl and arylalkyl bromides. The time dependence of the fractional extent of reaction for four typical systems differing in reaction rate constants and activation energies are shown in fig. 1-4. The curves were obtained from the theoretical analysis summarized above. Fig. 5 shows a comparison of the time dependence of the extent of reaction of P4VP with n-pentyl bromide in sulpholane at 339.6 K ( K = 0.65, L = 0.30) with that for a true second-order reaction ( K = 1, L = 1).A typical example of a second-order plot is given in fig. 6 for the quaternization of P4VP with allyl bromide in sulpholane at 305.6 K. The initial rate constant k, is given by the t -+ 0 limiting tangent to the curve. The filled-in circles and the curve were obtained from the theoretical time dependence of 5. SUMMARY OF KINETIC RESULTS Table 1 shows the ratios K and L defined by eqn (5) for all the systems studied. The fractional error on L is approximately twice that on K. Fig. 7 shows the80 QUATERNIZATION OF POLY(4-VINYL PYRIDINE) FIG. 1 .-Time dependence of the fractional extent 5: of reaction for P4VP + ally1 bromide in sulpholane at: (a) 323.7, (b) 315.7 and (c) 305.6 K.All curves in fig. 1-5 are from the theory. f I I 1 I I 0 5 10 15 20 25 tlks FIG, 2.-Dependence of < on t for PllVP+i-butyl bromide in sulpholane at: (a) 349.2, (b) 339.2 and (c) 325.2 K.E. A. BOUCHER AND C. C. MOLLETT 81 FIG. 3.-Dependence of 5 on t for P4VP+2-bromoethylbenzene in sulpholane at: (a) 323.7, (b) 317.8 and (c) 308.4 K. tlks FIG. 4.-Dependence of < on t for P4W+ I-bromo-3-phenylpropane in sulpholane at: (a) 348.9, (b) 340.6 and (c) 333.2 K.82 QUA TE RNI Z A T I ON OF P 0 L Y(4-V IN Y L P Y R ID I N E) FIG. 5-Dependence of 5 on t for P4VP+n-pentyl bromide in sulpholane at 339.6 K for K = L = 1. and theoretical curve FIG. 6.-Second-order plot for P4VP+allyl bromide in sulpholane at 305.6 K. The open circles are experimental and the filled circles are from the theory. The straight-line limiting, t/ks -+ 0, tangent gives ko.dependence of K and L on n, the number of carbon atoms in the organic group attached to the bromine atom of the reagent. There is a surprising degree of systematic variation for such a widespread set of molecules, with the alkyl species clearly distinct from the others. The values of k, given in table 1 are all for 340.2 K and were obtained by interpolation of Arrhenius plots based on the temperature ranges shown, withE. A. BOUCHER AND C. C. MOLLETT 0 / (0) 83 O.* t O * * l 0 2 I 6 8 10 n FIG. 7.-Dependence of the ratios K , sets (a), and L, sets (b), on the number n of carbon atoms in the organic bromides.comparable but not exactly identical initial concentrations of reagent and polymer. Some of the kinetic results have previously been givenll for 319.2 K. Assuming that there are three activation energies E,, El and E2 corresponding to k,, k, and k, in the neighbouring-group model, we can write ki = Aiexp(-E,/RT), i = 0, 1, 2. ( 6 ) The values of A , and E, and the associated experimental errors were given by least-squares determinations of straight lines fitting Ink, as a function of 103K/T. Similar plots of In k, and In k , showed that, within experimental error, El and E2 are identical to E,, i.e. K and L are not themselves temperature dependent. Table 2 gives results for quaternization with n-propyl bromide and benzyl bromide in sulpholane, propylene carbonate and 2,4-dimethyl sulpholane.The values of K and L are not dependent on the solvent species, although k, can be expected to depend on solvent for the reasons already discussed. l2 TABLE 2.--KINETIC QUANTITIES FOR QUATERNIZATION OF P4vP 104k, reagent solvent T/K /dm3 mol-l s-l K L n-propyl bromide sulpholane 333.1 17.39 0.80 0.38 n-propyl bromide propylene 333.5 11.35 0.80 0.36 carbonate sulpholane n-propyl bromide 2,4-dimethyl 333.2 6.74 - - benzyl bromide sulpholane 306.2 691.0 0.75 0.34 benzyl bromide 2,4-dimethyl 306.2 2 5 4 . 0 0.75 0.36 sulpholane84 QUATERNIZATION OF POLY(4-VINYL PYRIDINE) The initial rate constants k , in table 1 show that values for the C, to c6 n-alkyl bromides are reasonably constant, but that k, for ethyl bromide is ca.3 times that for the others. The value for 1 -bromo-3-phenylpropane is comparable with those for the C, to c6 alkyl molecules, and that for 2-bromoethylbenzene is about one half the value. Rate constants for i-butyl bromide, bromomethylcyclohexane and 1 -bromo- 2-phenylpropane, on the other hand, are very approximately 1/30 of those for the C,-C, molecules, and those for allyl bromide and benzyl bromide are ca. 100 times greater. The following discussion deals with the magnitude of the rate constant for a particular organic bromide, as well as with the values of the ratios K and L. The activation energies E, are the same within experimental error, except for values for allyl bromide and benzyl bromide. It is more difficult to explain the pre-exponential A , values because the errors on them are relatively large: a frequently occurring difficulty with this quantity.DISCUSSION THIS STUDY The quaternization reactions have been found to go to completion, with most experimental limits being = 0.99+0.01. There is no evidence that these reactions are reversible. If they were they would be expected to reach an equilibrium limit 5 < 1 . The mechanism of quaternization, involving nucleophilic substitution (S,2) of the bromide on the a-carbon of the reagent by the pyridyl base, is thought to be exactly analogous to quaternization reactions of small pyridyl species.17 The rate-determining step is the formation of the transition state. This is represented schematically in fig. 8, where C, and C, are the a-carbon and a-carbon of the organic bromide, A, B and C are substituents on C,, Y is the leaving bromine and X is the entering pyridyl species.FIG. 8.--Schematic representation without implied scale of the transition state for a quaternization reaction. See text for explanation. We first explain why ethyl bromide reacts more rapidly than n-propyl bromide (and the C,-C6 bromides: table 1). The activation energies are constant indicating that the transition state for n-propyl bromide is not distorted for the C,Br compared with the C,Br,18 e.g. were C,-X and C,-Y distances larger for C,Br to reduce steric interactions with C, substituents, E, would increase. Magat and coworker^^^^ 2o pointed out that if the methyl group on C, cannot be rotated so that A, B and C canE.A. BOUCHER AND C. C. MOLLETT 85 equally well occupy all three of the positions shown in fig. 8, then the entropy of activation, and k,, should be smaller compared with a transition state where all three positions are equally probable. The enthalpy AH and entropy A S of activation2’ can be related to Ei and 2, by AH= Eo-RT A , = (kT/h)exp(AS/R) (7) where k is Boltzmann’s constant and k is Planck’s constant. A H is associated with binding forces, and A S with the freedom of movement of the atoms. Clearly, therefore, variations in A , and A S will be important in all the reactions for which E, is constant. We suppose therefore that in the transition state, the substituents on Cp of the n-propyl group are restricted to a unique conformation (position), whereas with the ethyl group, three equally probable positions exist, and the ratio of the rate constants is expected to be 1 : 3, compared with 3 1.6 : 9 1 .O found experimentally (table 1).Similar arguments apply to the rate constants for the n-C,-C, bromides. The ratios of the rate constants for i-butyl and n-butyl bromides to ethyl bromide, 0.02 and 0.40, respectively, with Eo virtually identical, suggest that there is severe steric hindrance involving the two methyl groups on C, and the X and Y groups. When there are three methyl groups on C, (neopentyl bromide), the steric hindrance completely inhibits reaction (no bromide ions could be detected after 42 days). With ally1 bromide and benzyl bromide the large rate constants are in part associated with lower E, values than for the n-alkyl bromides. The transition state is more readily formed because of n-bond overlap between X and Y and the organic part of the halide molecule (cf.similarly large rate constants compared with that for ethyl bromide found by Clarke and Rothwel122 with the quaternization of 4-ethyl pyridine). By comparison, the rate constants for 2-bromoethylbenzene and 1 -bromo- 3-phenyl propane relative to that for ethyl bromide are 0.18 and 0.36, respectively, but E, values are comparable for all three bromides. It is concluded that when one or more carbon atoms exist between a benzene ring and C,, there is no enhanced rate constant due to a favourable transition state, i.e. k, values are similar to those for n-C, to C , bromides. With bromomethylcyclohexane the low rate of reaction is due to increased steric hindrance in the transition state compared with the n-alkyl bromides, whilst there is no decrease in the activation energy as observed with benzyl bromide.The low reactivity of 1 -bromo-2-phenylpropane is not unexpected and is analogous with that of i-butyl bromide in the aliphatic series of bromides. The systematic decrease of K and L ratios with increasing size of organic group attached to the bromide is in itself evidence that steric hindrance plays a part, and could be the sole factor in the neighbouring-group effect. An electrostatic effect would be expected to be virtually independent of the nature of the organic group. We realize that errors on pre-exponential factors are relatively large, but if it is accepted that Eo = El = E,, then retardation of reaction due to the neighbouring-group effect must arise for the condition A , > A , > A,.Therefore, for a particular organic bromide, as the reaction proceeds there is increasing steric hindrance restricting the number of configurations which can be adopted by the organic group in the transition state due to the presence of already quaternized nearest neighbours. It seems reasonable that the ‘reach’, or the longest distance that the organic group can influence, is a measure of its ability to be involved sterically. Fig, 9 shows K and L plotted against the extended lengths of the molecules relative to ethyl bromide as obtained from space-filling scale models. The distinction between the n-alkyl bromides and the other molecules is no longer evident, and both sets of results (for K and L)86 QUATERNIZATION OF POLY(4-VINYL PYRIDINE) 0 4 1 I Q- I Q- 0.4- 1 (b) 8 1 *w.@\ 0.2 - 0 I I 1 1.5 2 .o 2.5 I I 1 length ratio FIG. 9.-Dependence of K (a) and L (b) on the length of the organic groups of the bromide reagents in table 1. fall on reasonably straight lines. It is also noted that the benzene ring of benzyl bromide ( K = 0.75) provides less steric hindrance than the ring in bromomethyl- cyclohexane ( K = 0.65) which can exist in chair or boat forms. The values of K (0.75) are the same for n-propyl and allyl bromides, i.e. there are unlikely to be inductive effects since any reduction in the rate of reaction of a pyridyl group having a reacted neighbour would be expected to be less for reaction with allyl bromide compared with n-propyl bromide.Comparison of the K and L values for these quaternization reactions with values obtained by other workers is of limited value. Previous analyses of kinetic data have in the main used less rigorous model equations and have assumed that K = 1, which disagrees with our own findings. Necessarily, the values of L obtained on this basis disagree with the values given in table 1. The mean value of K / L for the twelve reagents used is within the limits of precision 2.0. This is taken as support for a mechanism of retardation involving steric hindrance of the alkyl group attached to the a-carbon in the transition state. If the organic group attached to an adjacent reacted pyridyl group is capable of preventing the organic group of a reagent molecule from occupying certain positions in the transition state, then it is to be expected that the presence of a second reacted neighbour will exert an approximately equal and additive effect.RELATED WORK Previous studies on polymer quaternization have been summarized in the introduc- tion and in a recent review.2 Generally direct comparison of rate constants is not possible due to differing reaction conditions or methods of data analysis. Comparison has already been made1' between rate constants for polymer quaternization and those for the quaternization of 4-ethyl pyridine.E. A. BOUCHER AND C. C. MOLLETT 87 TABLE 3.-cOMPARISON OF RATE CONSTANTS RELATIVE TO CORRESPONDING ETHYL HALIDE VALUE KI + P4VP + organic organic chlorides reagent bromides [ref.(23) group (this study) and (24)] n-propy 1 0.38 0.41 ally1 41 .O 31.6 n-butyl 0.32 0.40 n-pentyl 0.34 0.56 n-hexyl 0.36 0.52 benzyl 63 80 ethyl benzene 0.18 0.48 4-Ethyl pyridine + n-propyl and n-butyl bromides give 0.35 and 0.30, respectively. Comparison can also be made with a few analogous reactions,23! 24 especially regarding ratios of reaction rates for organic halides, as shown in table 3 for the halide exchange between KI and organic chlorides. There is a broad measure of agreement among values relative to those for reaction involving ethyl bromide and ethyl chloride, but ratios for organic chlorides exceed those for the corresponding bromides. Hammett and coworker^^^^^^ found that the ratio of the rate constants for the reaction of n-propyl halides with sodium thiosulphate in ethanol S,O;-+ C3H7Br -+ C3H7S,0; + Br- Bunte salt relative to the rate constant for the corresponding ethyl halide was ca.0.3, and that the activation energies were approximately the same for a particular halide. They explained these findings in terms of steric hindrance using a model similar to that used above for the polymer reactions. However, they found that an increase in the activation energy, associated with an increase in the halide to a-carbon bond distance, was responsible for low reactivity of i-butyl halides; this has not been found for quaternization with i-butyl bromide. Dostrovsky and Hughesz7 found that ethyl, i-butyl and neopentyl bromides reacted with sodium ethoxide (giving the corresponding ether + sodium bromide) in the ratio of rate constants of 1 : 0.04: which is comparable with our ratio for quaternization of the polymer, namely, 1 : 0.02: 0.CONCLUSIONS The wide range of systems studied has greatly increased the amount of quantitative information on quaternization reactions. There is evidence that differences in reactivity of an isolated pyridyl group towards the organic bromides depends significantly on the degree of steric hindrance in the transition state. The retardation of the overall quaternization reaction as the reaction proceeds is explained in terms of steric hindrance between the organic group on the reacting group and that on an already reacted neighbour. There is no support in this study for an electrostatic effect being responsible for the retardation.88 QUATERNIZATION OF POLY(4-VINYL PYRIDINE) C .C . M. acknowledges with gratitude a studentship from the States of Guernsey Education Department. B. D. Coleman and R. M. Fuoss, J. Am. Chem. Soc., 1955,77, 5472. E. A. Boucher, Prog. Polym. Sci., 1978, 6, 63. R. M. Fuoss, M. Watanabe and B. D. Coleman, J. Polym. Sci., 1960, 48, 5 . J. Morcellet-Sauvage and C. Loucheux, C.R. Acad. Sci., Ser. C, 1972, 275, 873, J. Morcellett-Sauvage and C. Loucheux, Makromol. Chem., 1975, 176, 31 5. E. Tsuchida and I. Irie, J. Polym. Sci., Polym. Chem., 1973, 11, 789. C. L. Arcus and W. A. Hall, J. Chem. SOC., Suppl., 1964, 2, 5995. N. A. PlatC, Pure Appl. Chem., 1976, 46, 49. C. B. Arends, J. Chem. Phys., 1963, 39, 1903. lo E. A. Boucher and C. C. Mollett, J. Polym. Sci., Polym. Phys. Ed., 1977, 15, 283. l1 E. A. Boucher, J. A. Groves, C. C. Mollett and P. W. Fletcher, J. Chem. Soc., Faraday Trans. I , 1977, l2 E. A. Boucher, E. Khosravi-Babadi and C. C. Mollett, J. Chem. SOC., Faraday Trans. 1, 1979, 75, l 3 A. Hantzsch, Berichte, 1909, 42, 81. l4 D. D. Perrin, W. L. F. Armarego and D. R. Perrin, Purification ofLaboratory Chemicals (Pergamon, l5 E. Khosravi-Babadi, 1979, unpublished results. l6 J. B. Berkowitz, M. Yamin and R. M. Fuoss, J. Polym. Sci., 1958, 28, 69. l7 E. A. Boucher, E. Khosravi-Babadi and C. C. Mollett, J. Chem. Soc., Faraday Trans. 1,1978,74,427. lo N. Ivanoff and M. Magat, J. Chim. Phys., 1950, 47, 914. 2o E. Bauer and M. Magat, J. Chim. Phys., 1950, 47, 922. 21 A. Streitwieser, Soluolytic Displacement Reactions (McGraw-Hill, New York, 1962). 22 K. Clarke and K. Rothwell, J. Chem. Soc., 1960, 1885. 23 J. B. Conant and W. R. Kirner, J. Am. Chem. Soc., 1924, 46, 232. 24 J. B. Conant and R. E. Hussey, J. Am. Chem. Soc., 1925,47, 476. 25 T. I. Crowell and L. P. Hammett, J. Am. Chem. SOC., 1948, 70, 3444. 26 P. M. Dunbar and L. P. Hammett, J. Am. Chem. Soc., 1950, 72, 109. *' I. Dostrovsky and E. D. Hughes, J. Chem. Soc., 1946, 157. 73, 1629. 1728. Oxford, 1966). J. P. Hine, Physical Organic Chemistry (McGraw-Hill, Tokyo, 2nd edn., 1962). (PAPER O / 1698)