J. CHEM. SOC. PERKIN TRANS. I1 1983 The Mechanism of Thermal Elimination. Part 14.’ Pyrolysis of Diacetamide, 2=Acetoxypyridine, Diacetyl Sulphide, and Thioacetic Acid : Possible Involvement of Enol Forms in Gas-phase Eliminations Roger Taylor School of Chemistry and Molecular Sciences, University of Sussex, Brighton BN 1 9QJ,Sussex In the gas phase, diacetamide undergoes unimolecular first-order elimination to give keten and acetamide with log A 12.42 s-l and €, 158.23 kJ m0I-l so that the reaction proceeds via a cyclic six-membered transition state. At 600 K the reaction is 6.7-fold slower than the comparable pyrolysis of acetic anhydride, but 4.4-fold faster than pyrolysis of diacetyl sulphide. This shows that in contrast to ester pyrolysis, heterolysis of the Ca-X bond in these compounds is relatively unimportant, the major rate-determining step being nucleophilic attack of the carbonyl group upon the P-hydrogen; the nucleophilicity of this carbonyl group is probably enhanced by resonance with the lone pair of the group X.The very high reactivity of the diacetyl compounds relative to their ester equivalents suggests that they pyrolyse via the enol forms. Pyrolysis of diacetyl sulphide is accompanied by first-order decomposition of thioacetic acid (one of the primary reaction products) by a number of pathways which give rise to keten, hydrogen sulphide, carbon oxysulphide, methanethiol, and methyl thioacetate, the latter probably arising from a combination of methanethiol with unchanged thioacetic acid which also accounts for the low stoicheio- metry (1.8) of the reaction. The rate of elimination of thioacetic acid is governed by the parameters log A 12.5 s-l and E, 175.7 kJ mol-l ;slightly different values are obtained from the acid produced by de- composition of diacetyl sulphide viz.12.5 and 173.1, respectively, probably due to the differing initial proportions of the thiono- and thiolo-forms. Decomposition is much faster than the Corresponding reactions of acetic acid and diacetamide, probably because the nucleophilicity of the thiol group is not lowered by resonance to the extent that operates for the corresponding nucleophiles in acetic acid and acetamide. 2-Acetoxypyridine also undergoes therma I elimination to 2-pyridone and keten, but satisfactory kinetics could not be obtained, due possibly to surface effects and an equilibrium between the N-acetyl and 0-acetyl tautomers.The bulk of the compound is much more stable than diacetamide because it exists largely in the 0-acetyl form. Acetic anhydride (I) was reported by Szwarc and Murawski 0 to undergo thermal elimination into keten and acetic acid ‘C-CHz o=c =CH, with log A 12.0s-l and Ea 34.5 kcal mo1-I. ?[Statistical analysis of their rate coefficients gives in fact log A 11.884 s-l and E, 0’) <‘H t\n 34.01 kcal mol-I (r 0.998 19).] This compound is prone to c=o CH3COOH surface-catalysed decomposition and if their minimum rate CHj coefficients only are used, this gives log A 12.136 s-’, E, 34.763 kcal mo1-l (145.49 kJ mol-’) and an improved cor- relation coefficient of 0.999 67.We believe these latter para- CH2-q-G CH,= CH,meters are more representative especially since theoretical H -calculations predict log A 12.3 s-l. Three points of interest 0’) -b emerge from these data. ‘C =G C H3COOH (i) The elimination rate coefficient at 600 K of 296 x CHjs-l is very much higher than for elimination from ethyl (111 acetate (11) for which k (600 K) = 9.81 x s-l. The reason for this rate difference of 30 170-fold, i.e. 15 085-fold per P-hydrogen, is not at all clear because polarisation of the 0 C,-O bond should be much easier for the acetate than for ‘C-CH3 Me2C -CH3 Me$ -CH,/the anhydride. One contributing factor, to which we have NH/ 0’ NH previously drawn attention,’ could be the activating effects of ‘c =o Lo ‘c =o the adjacent carbonyl group upon breaking of the P-C-H / / / bond.A P-acetyl group gives an acceleration of 388-fold per CH3 CH3 CH3 f3-hydrogen,6 whilst a P-COOMe group accelerates 144-fold (111) (IY) (Y)per P-hydrogen, so for acetic anhydride a value in betweep might be expected.$ However, this still leaves a factor of ca. 60-fold unaccounted for and in this paper we suggest an alternative explanation (see Discussion section). ate uia a cyclic six-membered transition state and this series is (ii) What would be the effect of replacing 0 by NH to give aimed towards measuring as accurately as possible all the diacetamide (111)? There are many compounds which elimin- factors which determine their reactivity. Changing the C,-0 moiety to Ca-NH gives in principle a wide range of t 1 cal = 4.184 J.compounds, but only the effect between acetates (IV) and $Because the oxygen lone pair is conjugated with two carbonyl amides (V) has been measured, and only for the t-butyl groups and will not be so effective in reducing the electron with- corn pound^.^^^ The respective rate coefficients at 600 K are drawal by the carbonyl group. 3.25 x and 4.73 x s-’ giving a relative rate of Table 1. Kinetic data for pyrolysis of diacetamide (MeCO)*NH TI K 1O3 kiss' log Als-' 546.8 1.975 12.420 556.4 3.725 567.1 6.70 575.7 11.9 580.3 15.6 593.1 29.8 595.1 34.8 600.6 43.8 n (or enol form) 68 700.This very large factor derives from the fact that breaking of the C-0 or C-NH bond is the principal rate- determining step in the elimination. If the same is true of the pyrolysis of acetic anhydride and diacetamide, a similarly large factor might be expected. (iii) 3-Ethoxypyridine has been shown to pyrolyse via a six-membered transition state (VI) to give ethene and 2- pyridone, this reaction being analogous to pyrolysis of ethyl acetate (11). Thus by analogy to pyrolysis of acetic anhydride (I) we could expect 2-acetoxypyridine to pyrolyse to keten and 2-pyridone (VII), and presumably this elimination would be much faster than that of (VI). Accordingly we have pre- pared 2-acetoxypyridine and determined its thermal stability. Results and Discussion (1) Diacetamide.-The pyrolysis of diacetamide was very well behaved kinetically, giving very reproducible rate coefficients with no sign of any surface catalysis, i.e.there were no deviant runs and all rate coefficients fell precisely on the Arrhenius line. The derived log A value is that predicted for the elimination and in our experience this is the most reliable indicator of the absence or otherwise of surface catalysis; in the presence of the latter, low values are always obtained. The stoicheiometry of the reaction was 2.0 and a five-fold change in the amount of compound taken for a run gave no significant change in the rate coefficients so that the reaction is first order. Only keten and acetamide were detected as reaction products.The rate data are given in Table 1 and show that like the pyrolysis of acetic anhydride a cyclic six- membered transition state must be involved. The rate coefficient at 600 K is 43.7 x s-I so that the elimination takes place 6.7 times more slowly than that from acetic anhydride. While this result is qualitatively in agreement with that for t-butyl acetate, versus N-t-butylacetamide, it is J. CHEM. SOC. PERKlN TRANS. I1 1983 Correlation EJkJ mol coefficient r k/s-' at 600 K 158.23 0.999 58 43.7 x 10-3 CHZ-CHEt C6) <'H L2 r'CH=CH, O+ C-CH;! / 4\H O\' /--c =CH, H0' quantitatively very different, which suggests that breaking of the C,-0 or C,-NH bond is not the most rate-determining step. Now it is generally true that for heterolytic gas-phase eliminations to occur there must be a polar bond in the transition state and breaking of this bond will be the principal rate-determining step." When this is not the C,-X bond it is usually the P-X-H bond, e.g.the a-0-H bond in P-hydroxy- alkenes and P-hydroxy-ketones,12 and for compounds of this type the nucleophilicity t of the group attacking the p-hydrogen obviously becomes important. For both these classes of compounds, alkyl substitution at the a-carbon produces only a small change in rate because of the lack of polarity of the C-X bond (in this case the C-CH2 bond). The effect of polarity in the P-X-H bond is clear from the enormously greater reactivity of e.g. 4-hydroxybutene [(VIII); k (600 K) 1.78 x s-'1 compared to hept-1-ene l3 [(IX); k (600 K) 1.34 x lo-' s-'1.With this result in mind it is possible to consider four feasible transition states for acetic anhydride pyrolysis (and likewise for diacetamide pyrolysis), viz. (I) and (X)-(XII). Transition state (I) we have already shown to be difficult to reconcile with the rate data. Transition state (X) differs from it in that the attacking ' base ' is C=CH2 instead of GO. The relative reactivities of compounds containing one or other of these groups can vary very considerably for reasons we shall describe elsewhere. When attack is on a CH, group, as in ethyl acetate and ethyl vinyl ether, the reactivities are com- parable,' so that there would be no advantage of (X) over (I), except for the presence of the HO group.Unfortunately it is not yet possible to predict the effects of substituents at this site because they vary so much in both magnitude and direc- tion, but we note that in for example By-alkenoic acids,I4 a methoxy-substituent (which would be similar to hydroxy) * It has been customary to refer to the nucleophilicity of the attacking group in papers describing pyrolytic elimination, though strictly speaking it is the basicity that should be considered. J. CHEM. SOC. PERKIN TRANS. I1 1983 0 0 >C-CH3 ‘\C -CH,/ (XI11 1 increases the rate 105-fold at 600 K, which is very similar to the effect we are seeking to explain. On the other hand this would imply that ethyl acetate pyrolysed by the enol form and many kinetic data suggest that this cannot be so (e.g.the almost identical reactivities of acetates and formates 7913915). Therefore we do not favour (X) but cannot entirely dismiss it. Transition state (XI) is more likely. Formation of an incipient vinyl cation will be more favourable than formation of the incipient acyl cation in (I), and the rate factor between (VTII) and (IX) is almost exactly that which we have to explain between acetic anhydride and ethyl acetates. The di(eno1)ic transition state (XI I) seems improbable. There are thus good grounds for supposing that an enolic transition state is involved. However, an additional factor needs to be taken into account. For compounds like acetic anhydride, diacetamide, etc.where nucleophilic attack on the P-hydrogen becomes most important, the nucleophilicity of the carbonyl group can become enhanced by resonance with the ethereal oxygen (XIII). (For acetates this can of course take place to an even greater extent, there being only one carbonyl group, but for these and related compounds, nucleophilic attack is less important.) Evidence to suggest this may be an important factor comes from pyrolysis of diacetyl sulphide, described below. (2) Relative Elimination Rates for Acetic Anhydride, Di- acetyl Sulphide, Diacetamide, and Pentane-2,4-dione.--Before considering these related compounds it was necessary to have an accurate rate coefficient for elimination from diacetyl sulphide at 600 K.From a kinetic study carried out between 459 and 530 K, diacetyl sulphide was reported l6 to give keten and thioacetic acid with E, 131 kJ mol-’ and log A 11.02 s-l. No rate coefficients were reported, and since both Arrhenius data are exceptionally low (the activation energy is about the lowest reported for any homogeneous six-centre elimination), we suspected that the reaction might be surface catalysed, even though this was said l6 not to be so. Accordingly we pyrolysed diacetyl sulphide over the temperature range 529.3-655.4 K. The reaction was accompanied by secondary decomposition of thioacetic acid though no reference to this decomposition was made in the original study; l6 decomposi-tion is clearly evident at temperatures used in that work.Since there has not been (as far as we are aware) a previous study of the thermal decomposition of thioacetic acid, we have determined the Arrhenius data and the reactivity is discussed briefly in the next section. Over the temperature range 583.5-655.4 K, the initial de- composition of diacetyl sulphide is so fast that only the subse- quent pyrolysis of thioacetic acid is observed. Below 583.5 K the kinetic plot could be dissected into two good first-order components. The results are plotted in the Figure, along with data obtained by pyrolysis of separately prepared thioacetic acid. The main features observable are as follows. (i) At the highest temperatures (where surface effects are always minimal) thioacetic acid gives a good Arrhenius line, but at lower temperatures there is curvature typical of that produced by the incursion of surface catalysis.(ii) Over the range where the Arrhenius plot for thioacetic c \\ c -0.6-0 A 0-0 -1.4-10 T -V K -l Arrhenius plot for pyrolysis of diacetyl sulphide and thioacetic acid: A, diacetyl sulphide; 0,thioacetic acid produced by decomposition of diacetyl sulphide ; 0, authentic thioacetic acid BUt But But 0/ S/ \ NH/ \ ‘c=o Me/c=o /Me Me Me a 3.25 x lo-* 1* 2 5 x -5 MeC =O MeC=O MeC=O MeC=O / I0-/ S/ NH/ ‘c=o ‘c=o ‘c=o/ / /Me Me Me tie -2 -729.6~1c2 hl x 10 4.37 x 10’ -8 x 10 Scheme. Rate coefficients (kls-’) for elimination at 600 K. Cal-culated from the published data for the ethyl compound l2assumingthat the t-butyl compound will be only 10 times as reactive beeause of the very low polarity of the Ca-CH2 bond acid is curved, so is that for diacetyl sulphide.This supports the view that over this temperature range, both compounds were subject to surface catalysis, and along with this should be noted the fact that we were never able to observe more than 25% of the decomposition of diacetyl sulphide, presumably due to an initial very fast surface-catalysed decomposition, which diminished as active sites became occupied; we were unable to improve upon this by deactivating the surface in the usual way. (iii) Our rate coefficients for pyrolysis of diacetyl sulphide are much lower than those given in the literature.The rate co-efficients at the two highest temperatures will be the least affected by surface effects, and these data give E, 159 kJ rnol-’ and log A 11.84 which seems reasonable. Since the extrapol- ation to 600 K from these measured rate coefficients is only a very small one the observed rate coefficient of 1 x loe2s-’ is unlikely to be in error by more than 10%; this coefficient is 42 times smaller than that given in the literature.I6 The Scheme 2.7,8,13,17 shows the rate coefficients for elimin- ation of t-butyl acetate and acetic anhydride and their respect- ive analogues at 600 K. Two features are evident. First, all the compounds in the first row are less reactive than those in the 92 second. Secondly, the reactivity order in the second row is dif- ferent from that in the first in that diacetamide is more reactive than diacetyl sulphide whereas one might have expected it to be less reactive.This can however be readily rationalised be- cause whereas, in the first row, breaking of the C,-X bond is the most important step of the reaction so that reactivity parallels the polarity of this bond, in the second row this factor is relatively unimportant, the reactivity being governed mainly by the ease of nucleophilic attack upon the b-hydrogen. The nucleophilicity of the carbonyl group will be aided by process (XIII) (and analogues) which will be more difficult for the sulphur compound than for the oxygen and nitrogen com- pounds. The overall reactivity is thus a balance of the ease of both C,-X bond breakage, and nucleophilic attack upon the P-hydrogen.The much lower reactivity of pentane-2,bdione relative to the 0, S, and NH analogues reflects the much poorer con- jugative electron release by the CH, group (via hyperconjug- ation) coupled with the low polarity of the C-CH2 bond, cleavage of which is difficult as shown by the data in the first row. [The rate data for these carbon-containing compounds should not be taken too literally because the Arrhenius data l3 from which they are determined are only estimated values based on a single rate coefficient (no tests for homogeneity) and a theoretical log A value.] (3) Pyrolysis of Thioacetic Acid.-Because thioacetic acid, a by-product of the thermal decomposition of diacetyl sulphide, itself underwent elimination, and since as far as we are aware there are no kinetic data for this compound, we have made a brief study of its elimination.As this was not the main thrust of our work and because the elimination is evidently very complicated this study is incomplete; we do not plan to do further work on this compound. The main features of the elimination are four-fold. (i) The kinetics are essentially first order. (ii) The products detected from the elimination are keten, hydrogen sulphide, carbon oxysulphide, methanethiol, and a substantial quantity of methyl thioacetate; no thioketen was detected. (iii) The stoicheiometry of the reaction was ca. 1.8. (iv) Thioacetic acid is much less stable than acetic acid.Bearing in mind the known behaviour of acetic acid," three possible mechanisms for decomposition of this acetic acid can be envisaged. We take into account the fact that it exists mainly in the thiolo- rather than the thiono-form. The proportion of the latter was said, on the basis of i.r. spectra obtained at 25, 70, and 100 "C,to increase with temperature.'' However in our view the evidence was rather inconclusive because of the dif- fering concentrations of the samples used at each temperature. Moreover, by using n.m.r. between -60 and +60 "C we were unable to detect any change in the amount of the thiono- tautomer. The product from pyrolysis of diacetyl sulphide is the thiono-form which we assume will rapidly equilibriate to the thiolo-form at high temperature.This latter we expect for reasons given below to be the more reactive, and this view is supported by the fact that the fragmentation pattern from the unchanged acid (ca. 80% thiolo-form) uiz: m/e 76 (minor, parent ion), 61 (minor, loss of CH3), and 43 (major, loss of SH) was different from recovered unchanged acid trapped at -78 "C, uiz. 76 (major), 62 (large, loss of CH2), 61 (minor), 47 (large, loss of CHO), and 43 (minor). (i) Second-order loss of hydrogen sulphide (XIV). No second-order process was detected in our work and this is consistent with the lower electrophilicity of the acyl carbon in thioacetic acid compared to acetic acid. The equivalent process involving two molecules of thiono-acid would give water, keten, and dithioacetic acid (which would decompose faster).J. CHEM. SOC. PERKIN TRANS. 11 1983 H keten + MeCOSH (XIV) or f Me Me /I I I H (XV) (XY1) CH, + COS S CH -C4 CH, 4-COS (XYIII)371 -HLO (ii) For acetic acid the equivalent of (XIV) disappears at higher temperature and is replaced by a four-centre process. For thioacetic acid this would be either (XV) or (XVI). The former is consistent with the lower stability of thioacetic acid, the C-S bond being weaker than the C-0 bond, and sulphur is more nucleophilic than oxygen. Likewise acetamide under- goes the corresponding reaction more easily than does acetic acid,20 but much less readily than does thioacetic acid.Although NH2 should be more nucleophilic than SH, it will resonate much more strongly with the carbonyl group than will SH, so the observed result is a logical one. In contrast to (XV) process (XVI) will if anything be more difficult than for acetic acid and would give thioketen (an unknown compound) which we did not detect. (iii) Loss of COS (XVII) or (XVIII). It is more difficult to decide whether (XVII) will be faster than for the acetic acid analogue, but it is fairly safe to assume on the basis of bond strengths that (XVIJ) would be faster than (XVIII). We detected carbon oxysulphide but not methane which would have escaped our analytical method. An additional elimination is involved with thioacetic acid, namely the formation of methanethiol, presumably via the three-centre process (XIX) comparable with that for decar- bonylation of acetyl bromide,21 and of formamide (which J.CHEM. SOC. PERKIN TRANS. I1 1983 0 ( XIX) gives ammonia).2z Again the high nucleophilicity of the SH group compared to OH should make this reaction faster than for acetic acid. No methanol was found so that the thiono- acid does not undergo the equivalent reaction. Only a trace of methanethiol was detected in the products, but this is to be expected since it should react with thioacetic acid to give methyl thioacetate, a substantial quantity of which was obtained. This removal of thioacetic acid also accounts nicely for the stoicheiometry being less than the required 2.0. The kinetics of the elimination were essentially first order, though below 631 K the rate coefficients during a run in- creased during the reaction before becoming perfectly con- stant. The latter portions of the kinetic runs gave reproducible rate coefficients, and the proportion of the run needed before first-order conditions were obtained increased as the tem- perature of the run decreased.This first-order behaviour was thus obtained immediately at 653 K but only after ca. 60% reaction at 591 K. There was some secondary decomposition (perhaps of keten) at 653 K but this decreased with decreasing temperature more rapidly than did the primary elimination. Rate coefficients obtained from the separately pyrolysed thio- acetic acid were ca. 60% of those obtained from the acid pro- duced by pyrolysis of diacetyl sulphide.In view of the dif- ferent initial concentrations of the tautomers, and the com- peting eliminations together with removal of the thiolo-form by reaction with methanethiol, this is not altogether un- expected. From the Arrhenius data (Table 2 and Figure) the elimination rate coefficient at 600 K is ca. 1.6 x s-', i.e. 10' times faster than elimination from acetic acid.18 (4) 2-Acetoxypyridine.-2-Acetoxypyridine does indeed undergo the expected elimination to 2-pyridone and kcten, but satisfactory kinetic data could not be obtained because the compound displayed very unusual kinetic behaviour, quite unlike that for any other compound we have studied. The abnormal features were as follows.(i) The stoicheiometry was only ca. 1.6. (ii) The kinetics were not first order, the rate coefficient decreasing with time during a given run. This discounts the possibility of polymerisation of either of the reaction pro- ducts since this would give an increasing apparent rate, and moreover this is not a problem in reactions in which products are (separately) produced. (iii) The reaction appears to be surface catalysed because in some runs an initial very fast elimination was observed. One explanation of the cessation of this fast decomposition is the gradual filling of sites on the surface, although it could arise from other reasons described below. (iv) In runs at higher temperatures, after the initial pressure increase obtained on injecting the compound into the reactor, there followed a pressure decrease, and subsequently a slow normal pressure increase.(v) Over the temperature range 514-728 K the approximate first-order rate coefficients showed little variation and were of the order 10-5-10-4s-' . There are too many complicationk here to permit other than a tentative explanation of the results. The low stoicheiometry derives in part from both the tendency of 2-acetoxypyridine to give some 2-pyridone on fractional distillation (presumably via surface-catalysed elimination), and its extreme sensitivity to moisture. (A sample left in an open vessel is completely hydrolysed to 2-pyridone and acetic acid in a few days. The Table 2.Kinetic data for pyrolysis of diacetyl sulphide and thio- acetic acid lo3k1s-l lo3 kls-' TIK I(MeC0)S1 (MeCOSH) 655.4 51.2 652.8 26.6 a 642.1 16.8 639.1 22.5 631.4 9.55 a 623.5 10.2 621.2 5.29 a 608.5 4.32 606.0 2.22 a 598.1 2.64 590.9 0.94 583.7 4.13 0.99 569.4 1.8 0.30 555.7 0.88 0.13 542.2 0.51 0.069 529.3 0.315 0.03 Obtained from separately prepared thioacetic acid; the other data in this column are for thioacetic acid obtained by decomposition of diacetyl sulphide. The former set give E. 175.7 kJ mol-I, log A1s-l 12.50, r 0.999 68, and the corresponding values for the latter are 173.1, 12.51, and 0.999 57. odour of acetic acid is very evident on opening a vessel which has been sealed for a few days, due, presumably, to hydrolysis by water vapour in the flask, though it could be due to therm- ally produced keten which subsequently hydrolyses.) Thus the presence initially of some 2-pyridone in the kinetic samples could make a small contribution to the low stoicheiometry.Another possibility is that there is an initial and very rapid elimination which is too fast to be observed. (Part of a rapid elimination was observed with some runs.) This could be due either to an initial surface catalysis, or to the fact that both N-acetyl and 0-acetyl tautomers are present. If one of these eliminates much more rapidly (or slowly) than the tautomer giving rise to the bulk of the observable elimination, then the low stoicheiometry could follow, providing that both tauto- mers are not readily interconvertible at the elimination tem- perature.N.m.r. studies showed that ca. 10% of the N-acetyl derivative is present but within experimental error this was almost invariant between -35 and +50 "Cso no inform- ation concerning the relative stabilities at different temper- atures could be reliable deduced from this. In the case of 2-ethoxypyridine pyrolysis, no elimination takes place via the N-alkyl tautomer (XX),4323 because the C-N bond, the breaking of which is of primary importance here, is insufficiently polar (cf. the low reactivity of amides compared to esters); elimination therefore takes place via the 0-alkyl tautomer (VI). Nucleophilic attack on the P-hydrogen is relatively unimportant for these compounds so the ad- vantage which (XX) would have over (VI) in this respect does not apply.For 2-acetoxypyridine however the situation is quite different. For compounds of this type nucleophilic attack on the P-hydrogen is the most important step of the reaction, and polarity of the a-C-X bond relatively unimport- ant (as shown by the closely similar reactivities of acetic anhydride and diacetamide). Thus the N-acetyl derivative (XXI) should pyrolyse faster than the 0-acetyl tautomer (VII) and at a comparable rate to that of diacetamide (approx- imately half in fact due to statistical effects). Tn product runs the proportion of the N-acetyl isomer in recovered unchanged ester appeared to diminish relative to the 0-acetyl isomer, also suggesting that it is more reactive. This (XX) O=C =CH, (or enol tautomer) (XXI1 could explain the curious kinetic behaviour.The initial fast elimination could be due to pyrolysis of the N-acetyl tautomer and this is followed by slow pyrolysis of the 0-acetyl tautomer or the N-acetyl tautomer produced by slow interconversion from the 0-tautomer. If the N-tautomer, which is thermo- dynamically less stable, is more difficult to form at higher temperatures, then this could account for the relatively small increase in apparent reaction rate with increasing temperature. It is also possible to account for the pressure decrease noted in some runs shortly after injection of the ester into the reactor.The N-acetyl tautomer will pyrolyse to keten and the hydroxy-form of 2-pyridine, and these may recombine to give the 0-acetyl tautomers which then slowly pyrolyses to the keto-form of 2-pyridone and keten. This explanation requires that the keto and enol forms of 2-pyridone are not rapidly interconvertible at high temperatures, and also makes it difficult to account for the relative insensitivity of reaction rate to temperature. Further speculation on the cause of these anomalies is not appropriate at this time, but one can conclude that whereas 2-ethoxypyridine and ethyl acetate have similar reactivities because nucleophilic attack on the (3-hydrogen is not kinetic- ally very important, in the case of acetic anhydride and 2- acetoxypyridine, nucleophilic attack is important so that the 0-acetyl tautomer of the latter is very much less reactive than is acetic anhydride.Experimental Diacetamide.-This was a comniercial sample, recrystallised before use. Diacetyf Sulphide.-This was prepared according to the literature method.24 Acetyl chloride (54.7 g) was added to thioacetic acid (25.6 g) and the mixture heated under reflux during 5 h. Fractional distillation of the product at ca. 175 mmHg gave a fraction, b.p. 100-106 "C, and a main fraction, b.p. 106-1 11 "C. This latter was slightly pink and had nD20 1.4597, much lower than the literature value of 1.4810, and we concluded that acetyl chloride must have codistilled with the required product, despite the column having ca. 7 theor- etical plates.The main fraction was redistilled at low pressure and a pure fraction collected, b.p. 38 "C at 5 mmHg, nD20 1.4798. Thioacetic Acid.-A commercial sample was purified by fractional distillation, the product, b.p. 90 "C, being used for kinetic studies. It should be noted, because of the ease with which this compound undergoes elimination, the ampoules J. CHEM. SOC. PERKIN TRANS. 11 1983 in which it is supplied can be under substantial pressure and may explode on scoring the glass if not thoroughly cooled beforehand; the temporary social results can be distressing, T (CDC13) 4.26 (s, SH), 7.62 (s, CH,CO), and 7.96 (s, CH,CS); the peak height ratios of these latter two peaks measured at various temperatures between -60 and +60 "C indicated the presence of 19 f1% of the thiono-form.2-Acetoxypyridine.-Sodium 2-pyridone was prepared by mixing stoicheiometric quantities of sodium hydroxide and 2- pyridone in water. The water was largely removed using a rotary evaporator (with a final vacuum of 0.01 mmHg) during several days. This gave a product which could be ground into a fine powder which was then dried further under high vacuum during several days. Acetyl chloride (8.1 g, 0.103 mol) was added to sodium 2-pyridone (12 g, 0.103 mol) in dry diethyl ether at a rate suffi- cient to maintain reflux. The suspension of sodium chloride was removed, and the filtrate fractionally distilled to give 2-acetoxypyIidine (38%, 5.4 g), T (CDC13) 1.56 (1 H, m, J 5 and 2 Hz, 6-ArH), 2.19 (1 H, tn, J 8 and 2 Hz, 4-ArH), 2.71 (1 H, m, J 5 and 2 Hz, 5-ArH), 2.90 (1 H, d, J 8 Hz, 3-ArH), 7.20 (0.3 H, s, NCOCH3), and 7.69 (2.7 H, s, OCOCH3).N.m.r. studies carried out between -35 and +50 "Csuggested that the amount of the N-acetyl tautomer decreased with in-creasing temperature. Because of the tendency of the peak due to this compound to merge with the side bands of the 0-acetyl peak as the temperature was lowered, studies outside the temperature range given were not possible. Product Studies.-A solution of diacetamide in chloro-benzene was carried in a nitrogen stream down a column of heated helices at 650 K such that the residence time was ca. 20 s. The products were passed through a glass coil sur- rounded by dry ice and the effluent passed into sodium hydroxide.The contents of the cold trap were analysed by both g.1.c. and n.m.r. which indicated the presence only of dichlorobenzene and acetamide. The contents of the aqueous trap were acidified with hydrochloric acid, extracted with deuteriochloroform, and analysed by g.1.c. and n.m.r. The only product was acetic acid, produced by hydrolysis of keten. The products from pyrolysis of thioacetic acid (under kinetic conditions) were condensed in a trap at -78 "C (and also in one experiment at -196 "C). The products were analysed by g.c.-m.s. (Kratos MS 25). The gaseous products could not be resolved from argon which is the carrier balancing gas in our kinetic system, but by monitoring for single ions in the total ion current in this region we were able to infer the presence of HzO, CH2C0, and COS from the m/e values of 34, 42, and 60, respectively; this latter was not due to acetic acid which is a trace impurity in thioacetic acid and appeared at a different point in the chromatograph.Analysis in the same way of the liquid products showed the presence of MeSH, MeCOSH (different from the starting material, see Discus- sion section), and MeCOSMe, the m/e values being 48, 76, and 90, respectively. Product studies of 2-acetoxypyridine pyrolysis were carried out in a similar manner to that for diacetamide except that no chlorobenzene was needed and temperatures up to 740 K were used. Keten was detected by mass spectrometry of the gaseous products and 2-pyridone was obtained along with unreacted ester and was identified byits n.m.r., T (CDCI3) -3.52 (1 H, s, NH), 2.51 (1 H, t, J 10 Hz, 4-ArH), 2.54 (1 H, d, J 6.5 Hz, 6-ArH), 3.39 (1 H, d, J 10 Hz, 3-ArH), and 3.69 (1 H, t, J 6.5 Hz, 5-ArH).The peak in the n.m.r. of the unchanged ester, due to the N-acetyl group, tended to be smaller in the recovered material, indicating that this tautomer is more reactive than the 0-acetyl form. J. CHEM. SOC. PERKIN TRANS. 11 1983 Kinetic Studies.-These were carried out in the manner and apparatus previously described.25 Diacetamide was injected into the reactor as a solution in chlorobenzene. The kinetic data are given in Tables 1 and 2 and displayed in the Figure.The Arrhenius parameters for decomposition of thioacetic acid produced from diacetyl sulphide are cal-culated only from those points which are evidently on a linear plot and therefore free from surface catalytic effects. Acknowledgements We thank Mr. R. A. August for preparation of a sample of 2-acetoxypyridine and Mr. A. M. Greenaway for the mass spectral analysis. References 1 Part 13, C. Eaborn, F. M. S. Mahmoud, and R. Taylor, J. Chem. SOC.,Perkin Trans. 2, 1982, 13 13. 2 M. Szwarc and J. Murawski, Trans. Faraday SOC.,1951, 47, 269. 3 W. H. Richardson and H. E. O’Neal, ‘ Comprehensive Chemical Kinetics,’ Elsevier, Amsterdam, 1972, vol. 5, p. 381. 4 N. Al-Awadi, J. Ballam, P. R. Hemblade, and R. Taylor, J. Chem. SOC.,Perkin Trans. 2, 1982, 1175. 5 R. Taylor, ‘The Chemistry of the Functional Groups,’ Sup- plementary Volume B. Acid Derivatives,’ ed. s. Patai, Wiley, London, 1979, p. 871. 6 R. Taylor, J. Chem. Res., 1981, (S) 250. 7 R. Taylor, J. Chem. SOC.,Perkin Trans. 2, 1975, 1025. 8 A. Maccoll and S. S. Nagra, J. Chem. SOC.,Faraday Trans. I, 1973, 69, 1108. 9 R. Taylor, J. Chem. SOC.,Chem. Commun., 1978, 732. 10 R. Taylor, unpublished work. 11 G. G. Smith and B. L. Yates, J. Chem. SOC.,1965,7242. 12 B. L. Yates and J. Quijano, J. Org. Chem., 1969, 34, 2506. 13 A. T. Blades and H. S. Sandhu, Znt. J. Chem. Kinet., 1971, 3, 187. 14 D. B. Bigley and A. Al-Borno, J. Chem. SOC.,Chem. Commun., 1978, 1025. 15 R. Taylor, J. Chem. SOC.,Perkin Trans. 2, 1978, 1255. 16 P. G. Blake and A. Speis, J. Chem. SOC.,Perkin Trans. 2, 1974, 1879. 17 P. C. Oele, A. Tinkelberg, and R. Louw, Tetrahedron Lett., 1972,2375. 18 Ref. 5, p. 864. 19 R. Mecke and H. Spiesecke, Chem. Ber., 1956,89, 1110. 20 Ref. 5, p. 868. 21 Ref. 5, p. 867. 22 Ref. 5, p. 869. 23 L. B. Kasunic, I. L. Evoy, and C. N. Sukenik, J. Org. Chenz., 1981, 46, 1970. 24 W. A. Bonner, J. Am. Chem. SOC.,1950,72,4271. 25 R. Taylor, J. Chem. SOC.B, 1971, 2382. Received 29th October 1981;Paper 1/1692