J. Chem. SOC., Faraday Trans. I , 1982, 78, 1 165-1 176 The Decomposition of 2,2,3,3-Tetramethylbutane in KC1- and B,O,-coated Vessels in the Presence of Oxygen BY ROY R. BALDWIN,* MOHAMED W. M. HISHAM, ALAN KEEN AND RAYMOND W. WALKER Department of Chemistry, The University, Hull HU6 7RX Received 18th May, 1981 The decomposition of 2,2,3,3-tetramethylbutane (TMB) in the presence of oxygen has been studied in both KCI-coated and aged boric-acid-coated vessels. The values for k, obtained with the two types of vessel surface are in close agreement, and combination of the values over the range 400-542OC gives A, = 1.04 x 1017 s-l, El = 294.7 f 3 kJ mol-l, effectively identical with previous values using KC1-coated vessels, and thus confirming the thermochemistry suggested for the t-butyl radical : (CH,),C--C(CH,), --f 2(CH,),C.(1) The observed rate constant, kobs, defined by the equation -d [TMB]/dt = kobs[TMB] increases with increasing TMB concentration with both types of vessel surface and this variation has been used to evaluate the ratio k,/k$. The results with both types of vessel surface are in agreement, giving A,/& = 4.4 x lo5 (dm3 mol-' s-l)+, E4-8E7 = 81.7f 8 kJ mol-l over the range 400-520 OC: HO2 + (CH,),C-C(CH,), --* H20, +(CH,),C + C(CH,)zCHz (4) (7) HO, + HO, -+ H,O, + 0,. Previous papers' on the decomposition of 2,2,3,3-tetramethylbutane (TMB) in the presence of 0, have concentrated on the evaluation of the Arrhenius parameters for reaction (1) and on the use of the system to study the reactions of t-butyl radicals' with H,, D, and 0,, and to evaluate their entropy and enthalpy of formation., In KC1-coated vessels, where HO, radicals are efficiently destroyed at the vessel surface, the basic mechanism involves reactions (1)-(3), and gives the rate expression (i) - d [TMB]/dt = kl [TMB].(9 However, the observed rate constant increases2 both with increasing concentration of TMB and with addition of inert gas. This is caused by two contributions to a very short chain, from HO, radicals through reaction (4), which competes with reaction (3), and reaction (5), and from OH radicals through reactions (8), (9) and (5), the H,O, being formed by reactions (4) and (7) and being predominantly destroyed by the surface process (6). Reaction (10) also occurs to a very minor extent with k , , / k , = 0.01; Arrhenius parameters for this reaction have been given.(CH3),C-C(CH3), + 2t-C,H9 t-C,H, + 0, + i-C,H, + HO, surface HO, + $H,O+20, 11651166 D E C 0 M P 0 SIT I ON OF 2,2,3,3-TE T R A M E T H Y L BUT A NE (CH,),CC(CH,),CH, + i-C,H, + t-C,H, surface H,O, + H,O +p, HO, + HO, -P H,O, + 0, (7) The present paper uses the variation of the observed rate constant in eqn (i) with TMB concentration in both KC1-coated and aged boric-acid-coated vessels to evaluate the rate constant k,. EXPERIMENTAL Reactions were carried out in cylindrical Pyrex vessels, 20 cm in length and 5.1 cm diameter. KClcoatings were renewed every week, as previous studies2 have shown that the efficiency of the surface for destruction of HO, and of H,O, decreased markedly with older coatings. Aged boric-acid coatings were prepared as previously de~cribed.~ Studies were made over the range 400-520 "C in boric-acid-coated vessels but restricted to 420-470 OC in KC1-coated vessels.A wide range of mixture compositions (0.1-4.0 Torr TMB) was examined. The reaction was followed by using gas chromatography to measure the isobutene formed. Electromagnetic valves (opening and closing time < 0.1 s) were used with reaction times below 100 s to admit the gases from the mixing bulb into the reaction vessel, and to admit the reactants into the sampling bulb after a pre-determined time interval. To avoid complications resulting from reactions of isobutene, which become noticeable beyond 10% reaction, consumption of TMB was restricted to a maximum of 5% for those points used to elucidate rate constants.The time for 5% consumption varied from ca. 10 s at 520 O C to cu. 3000 s at 400 OC. RESULTS AND DISCUSSION KCl-COATED VESSELS If H,O, is always destroyed by reaction (6), and if the minor reaction (10) is ignored, the rate of reaction is given by eqn (ii) if reaction (3) is the sole termination process, and by eqn (iii) if reaction (7) is the sole termination process: - d [TMB]/dt = k, [TMB] + 2k1 k, [TMB]'/k3 (ii) - d [TMB]/dt = kl [TMB] + k4 (kl/k7)a[TMB]). (iii) If reaction (3) is the sole termination process, the chain length increases with increasing temperature, since E4 x 80 kJ mol-l, with increasing TMB concentration, and with inert gas addition (since k, is diffusion-controlled). If reaction (7) is the sole termina- tion process, the chain length increases with TMB concentration, is independent of inert gas and decreases as the temperature increases, since El (ca.290 kJ mol-l) > 2E4. Increase in temperature, in TMB concentration or addition of inert gas increases the HO, concentration and thus favours reaction (7) relative to reactions (3) and (4). Use of preliminary values, of k, and k,/k74 gives the percentage reaction due to the HO, chain for a mixture containing 2 Torr TMB+30 Torr 0, at pressures of 60 and 500 Torr (N, addition) shown in table 1 for a range of temperature. At 60 Torr, the percentage increases and then decreases at high. temperatures. This is because the second-order radical-radical reaction (7) increases in importance relative to theR.R. BALDWIN, M. w . M. HISHAM, A. KEEN AND R. w. WALKER 1167 first-order processes (3) and (4) as the temperature increases, due to the increased HO, concentration resulting from the increased rate of dissociation of TMB. Table 1 shows that the HO, chain contribution is small, and has its highest value in the range 400-470 OC. A further difficulty in using this small chain contribution to evaluate k, is that the OH chain becomes increasingly important as the temperature increases above ca. 450 OC; 2% of the total OH chain is due to reaction (lo), with k,,/k, 0.01, the remainder being due to the formation and decomposition of H,O,. TABLE VARIATION OF CHAIN LENGTH WITH PRESSURE AND TEMPERATURE chain reaction (%) HO, chain OH chain T/OC at 60 Torr at 500 Torr at 60 Torr at 500 Torr 400 7 31 2 2 440 12 29 2 6 470 16 24 3 16 500 12 12 5 31 540 9 5 14 60 To evaluate k , from the isobutene (IBE) against time curves, a computer treatment was used.It is convenient to express the HO, radical concentration in terms of the parameter G = k$[HO,], since it is the ratio R, = k,/k$ that is actually involved in the computer treatment. Similarly k, enters as the ratio R, = k,/k,k The differential equations for TMB, H,O,, O,, IBE, isobutene oxide (IBO) and H,O are given by eqn (iv)-(ix), and the stationary-state equations for HO,, OH and t-C,H, radicals are given by eqn (x)-(xii), respectively : - d [TMB]/dt = k, [TMB] + R, G[TMB] + k, [OH] [TMB] (iv) (v) d [H,O,]/dt = G2 + R4 G[TMB] - k6 [H20,] - k8 [MI [H202] - d [O,]/dt = (k, + klo) (t-C,H,] [O,] - 0.75R3 G - 0 3 , [H,O,] - G2 d [IBE]/dt = k, [t-C,H,] [O,] + R, G[TMB] + k, [OH] [TMB] d [IBO]/dt = klo[t-C4H,] [O,] d[H,O]/dt = k6 [H,0,] + 0.5R3G + k, [OH] [TMB] 2k, [TMB] + 2k, [MI [H,O,] = 2G2 + R, G 2ks [MI [H20214- kio [t-C,H,I Lo21 = kg [OH] [TMBI (vi) (vii) (viii) (ix) (x) (xi) 2k, [TMB] + R, G[TMB] + k , [OH] [TMB] = (k, + k1J [t-C,H,] [O,].(xii) G is obtained from eqn (x), and eqn (xi) and (xii) then give [t-C,H,] and [OH]. Eqn (iv)-(ix) are solved by the Kutta-Runge numerical integration method. H,O, rapidly reaches a quasi-stationary value, so that computer time can be considerably reduced by using a short time step until the H,O, is close to its quasi-stationary value, after which a stationary state is written for H,O, with a much bigger time step.Operation of the computer program requires values of the parameters k,, klo/k,, R,, R,, k, and k,. The yield of IBE at various times is calculated for a range of mixtures1168 DECOMPOSITION OF 2,2,3,3-TET R A ME THY LBU TAN E and compared with the experimental values. An optimisation procedure adjusts the parameters k , and R, so as to obtain minimum r.m.s. deviation. Reaction (10) plays only a very minor part, and the ratio k,,/k, = 0.01 obtained from previous studies, of the IBO yield is sufficiently accurate. k , is known from direct studies5', of the homogeneous decomposition of H,O,. The values of k , and k, depend on the efficiency of the surface for destruction of HO, and H,O,, respectively. In general, the surface destruction constant k, for a cylindrical vessel is given7 by eqn (xiii) k, = aA,/P( 1 + 4aB/P) (xiii) where A, = 32D,A/d2, B = 8D,A/~cd.D,A is the diffusion constant of the species in gas A at unit pressure, E is the surface efficiency for destruction of the species, c the average velocity of the species and d the vessel diameter. a is a coefficient expressing the variation of diffusion coefficient with mixture composition, and is given7? * by eqn (xiv) (xiv) where xA, xB and xc are the mole fractions of the components. The value of R, = k,/k,l has been obtained experimentally from a studys of the oxidation of HCHO in KCl-coated vessels, where the HO, concentration is given by eqn (x), but with 2k1, [HCHO] [O,] replacing 2k1 [TMB] : (1 1) HCHO + 0, + HCO + HO,. k,, was obtained by studies at very low HCHO concentrations, where the chain length is effectively zero.Since k, is the unknown parameters determining the rate of HCHO oxidation are R,, k, and k,,/k$ (12) A , [ = A , in eqn (xiii)] has been calculated9 for H,O, using the method given by Hirschfelder et aZ.;l0 use of 4N,) = 0.375 nm, a(H,O,) = 0.40 nm and a(0,) = 0.354 nm gives A , = 850 Torr s-' at 500 OC for N, in a 5.1 cm diameter vessel, and a ( 0 , ) = 0.989 relative to a(N,) = 1. Use of a(TMB) = 0.745 nm gives a(TMB) = 3.23. The value of A , varies with temperature, but the values of a are effectively constant. A value of kl,/k$ has been obtained,, by studying the effect of HCHO on the induction period of the H, + 0, reaction at 500 OC in 5.1 cm diameter aged boric-acid-coated vessels.The only unknowns are thus R, and B,, which relates k, to A , via eqn (xiii). Optimisation of the agreement between the observed and calculated value of the CO yield at various times for various mixtures at 500 OC gave R, = 4.56 (Torr3 s-l); and B, = 255 Torr (M = N,). The value of R, is very close to the value of 4.28 calculated from A , on the assumption that the diffusion constants for HO, are the same as those for H,O,. The difference of ca. 7% could be due to small errors in any of the quantities k,,lk,$ obtained from the H,+O, induction periods in the presence of HCHO, with an estimated error of f lo%, R, obtained from the HCHO+O, system with an estimated error of & lo%, the diffusion coefficient or the value (2.0 x lo9) taken fork,.Consequently, the experimental value of R, = 4.56 (Torr3 s-l)i at 500 OC has been taken, together with the calculated value of A, for H,O,. Values at other temperatures have been calculated from the change in diffusion coefficient with temperature given by the treatment of Hirschfelder et aZ.l0 The optimum value of B, [for H,O, in eqn (xiii)], and to a much lesser extent the value of R,, differ from those quoted earliers of 161 and 4.64, respectively. This is because the earlier results of HCHO+O, have been treated to allow for the interaction7 between the uniform profile of chain centres throughout the vessel, HO, + HCHO + H,O, + HCO.R. R. BALDWIN, M. w. M. HISHAM, A. KEEN AND R. w . WALKER 1169 resulting from homogeneous termination processes, and the diffusion profile created by an efficient surface termination.In the case of a homogeneous initiation process, 0, such as the decomposition of TMB, which is uniform throughout the vessel, this interaction has been studied' by solving the differential eqn (xv) by numerical integration methods : d2n dn Dy+(D/x)--6n2+0 dx dx = 0. n is the chain-centre concentration, x the vessel coordinate, 6 the mutual homogeneous destruction constant and 0 the initiation rate. Both with linear (first-order) homo- geneous termination, where algebraic solution is po~sible,~? l2 and with mutual (second- order) termination, the interaction may reduce the radical concentration by up to 20% below the value calculated assuming no interaction. Both in the HCHO+O, system and in the TMB+O, system, the interaction is further complicated by the fact that the dissociation of H,O, is a secondary source of initiation; since H,O, is formed, by reactions (4) and (7), from HO, which is not at a uniform concentration throughout the vessel, and since H,O, is predominantly destroyed at the surface [although by a process that is only moderately efficient at most pressures if B, = 255 Torr (M = N,)], the total initiation process is not uniform.This difficulty was overcome by solving the two stationary-state simultaneous differential eqn (xvi) and (xvii) d2n dn dt2 dx D,-+ (D,/x) -- Sn2 + 2kI [TMB] + 2k,[M]p = 0 d2P dP D,@+ (D,/x)z- k , [M]p+ 0.5 6n2 + k,n[TMB] = 0 (xvi) (xvii) where n = [HO,], p = [H,O,], and D,, D, are the corresponding diffusion coefficients.These equations were solved numerically by the method used to solve7 the single eqn (xv). With eqn (xv), it is necessary to guess no, the value of n at the centre of the vessel, to start the integration procedure, and the program varies this starting value until satisfactory convergence of successive solutions has been achieved. With a high ratio of homogeneous to surface termination, the initial value of no has to be located with increasing accuracy, otherwise unreal solutions are obtained. With eqn (xvi) and (xvii), initial values of both no and po have to be provided, and accurate estimates are required, particularly at high ratios of homogeneous termination to surface termination, so that several trials are often necessary. Since incorporation of this procedure into the main program, together with optimisation procedures, would make the running time excessively long, eqn (xvi) and (xvii) were solved separately, the volume average concentrations of HO, and H,O, were compared with the concentrations obtained on the assumption of no interaction, and corrections obtained which could be inserted into the main program.Because of the presence of mutual termination it was also necessary to obtain corrections to the volume average of [HO,],. The corrections to [HO,] varied from 0 to -2O%, to [HO,], from -25 to +25%, and to [H,O,] from - 15 to + 15%. However, although the accurate evaluation of the role of H,O, is complex, interpretation is not particularly sensitive to the parameters used for the surface destruction of H,O, since, under the conditions used to evaluate HO, + TMB, the contribution of the OH chain is reasonably small.Results obtained in this way are summarised in table 2. As table 1 shows, the contribution of the HO, chain is only some 20-30% even under the most favourable conditions. While the r.m.s. deviation between observed and calculated yields of IBE at 440 O C implies an accuracy of ca. 5% in k,/k,i, the1170 DECOMPOSITION OF 2,2,3,3-TETRAMETHYLBUTANE TABLE 2.-vALUES OF k4/k,i AT VARIOUS TEMPERATURES IN KCl-COATED VESSELSa r.m.s. no. of range of k4lk,$ dev. T/OC mixtures [TMB]/Torr /(dm3 mot1 s-l)$ (%) data ~ ~~~ 440 7 0.50-4.0 0.436 2.4 Evans2 41 8 8 0.25-4.0 0.257 8.0 Keen13 437.5 9 0.25-4.0 0.455 2.8 Keen13 466 10 0.25-4.0 0.674 6.2 Keen13 a In all cases the range of pressure was 60-500 Torr.results are less satisfactory at 418 and 466 O C . Moreover, Nalbandyanl* has reported the detection of HO, radicals when H,O, is decomposed on a KC1-coated surface. Such a possibility would increase the concentration of HO, radicals above the value calculated from the present mechanism, although it would be unimportant at low temperatures and pressures, since, very little H,O, is formed because HO, radicals predominantly undergo reaction (3). However, at higher temperatures and pressures reaction (7) becomes the dominant reaction, and formation of HO, radicals by surface decomposition of H,O, might become important. No indication appears available from Nalbandyan's work of the yield of HO, radicals produced in this way.BORIC-ACID-COATED VESSELS In aged boric-acid-coated vessels, the surface is extremely inefficients both for the destruction of HO, and of H,O,. To avoid the difficult treatment of the interaction of homogeneous and heterogeneous termination, as well as the possible complication of surface decomposition of H,O, to give HO, radicals, studies were made in aged boric-acid-coated vessels to see if results consistent with those in KC1-coated vessels could be obtained. Fig. 1 shows the [IBE] against time curves at 440 O C in a 5.1 cm diameter aged boric-acid-coated vessel for a mixture containing 1 Torr of TMB, 3 Torr of 0, and at various total pressures from 15 to 200 Torr, obtained by addition of N,. The curves are significantly autocatalytic and the extent of autocatalysis increases with pressure because of the M term in reaction (8).This contrasts with the effectively linear [IBE] against time curve over the first 5% of reaction in KC1-coated vessels, where the very low quasi-stationary concentration of H,O, is reached in a time that is usually small compared with the earliest sampling time, so that the initial autocatalysis is not detected. In aged boric-acid-coated vessels, the autocatalysis persists far into the reaction, well beyond the 5% consumption of TMB that was normally set as a limit to avoid secondary reactions of IBE. In boric-acid-coated vessels, as shown in table 3, the chain due to HO, radicals is significantly greater than in KC1-coated vessels at lower temperatures, although at higher temperatures the main termination process is reaction (7) with both types of surface.However, as table 3 also shows, the chain length due to OH is also greater because of the much higher yields of H,O,. Since the H,O, increased continuously with time in the early stages, values corresponding to ca. 2% reaction are given in table 3. Operation at low pressure, by reducing the value of M in reaction (8), reduces the OH chain. A pressure of 15 Torr of 0, + TMB was thus used, except at the lowestR. R. BALDWIN, M. w. M. HISHAM, A. KEEN AND R. w. WALKER 1171 time/min FIG. 1.-Effect of inert gas addition on [IBE] against time relationship. 1 Torr TMB, 3 Torr 0,, +N,, 440 OC. 0, N, = 11 Torr; x , N, = 26 Torr; A, N, = 56 Torr; 0, N, = 146 Torr; V, N, = 196 Torr. TABLE 3.-vARIATION OF CHAIN LENGTH WITH PRESSURE AND TEMPERATURE IN BORIC-ACID- COATED VESSELS For each condition, sample time selected corresponds to ca.2% reaction. TMB = 2 Torr. chain reaction (%) HO, chain OH chain at at at at at at T/OC 15Torr 60Torr 500Torr 15 Torr 60Torr 500Torr ~~ ~~~~ ~ ~ 400 45 37 25 18 35 61 440 35 32 18 8 19 62 470 21 19 11 8 18 61 500 15 14 10 6 13 511172 DE C OM POS I TI 0 N OF 2,2,3,3-TE TR A ME THY LB U T ANE temperature of 400 O C , when the partial pressure of 0, was 30 Torr and the total pressure was made up to 60 Torr by N, addition. An algebraic method of allowing for the OH chain can be devised since, in boric-acid-coated vessels, k, and k, are zero, and in the absence of an HO, chain, the combination of eqn (v) and (x) simplifies to eqn (xviii) -- fH2021 - k, [TMB].dt (xviii) For small consumption, this can be integrated to give eqn (xix) [H,O,] = k,[TMB] t . (xix) With the same assumption, and neglecting the 1 % contribution from reaction (lo), substitution of this expression for [H,O,] in eqn (xii) and eliminating [OH] between eqn (iv) and (xii) gives eqn (xx): - d [TMB]/dt = k , [TMB] + 2k8 [M]k, [TMB] t. (xx) 6.0 3.01 1 0 2 4 6 8 time/ min FIG. 2,-Variation of kobs with time in aged boric-acid-coated vessels. 440 O C , kobs = [IBE]/2[TMBlo t . 0, [TMB], = 4, [OJ = 11 TOIT; A, [TMB], = 2, [O,] = 13 TOIT; 0, [TMB], = 1, [O,] = 14 Torn; V, [TMB], = 0.5 [O,] = 14.5 Torr; x , [TMB], = 0.25, [O,] = 14.75 Torr; 0, [TMB], = 0.1, [O,] = 14.9 Torr.R. R. BALDWIN, M. w. M. HISHAM, A. KEEN AND R.w. WALKER 1173 Integration of eqn (xx) for small consumptions gives eqn (xxi): A[IBE]/t = 2k, [TMB] + 2k, k , [TMB] [MI t. (xxi) Defining kobs = A[IBE]/2[TMB] t kobs = k1+ k1 k,[M] t. The intercept of the plot kobs = A[IBE]/2[TMB] t against t should thus give kzbs = k,, the value when the OH chain is eliminated. Fig. 2 shows that the plots of kobs against z are linear. However, the intercept varies systematically with [TMB]. This arises because of the residual HO, chain, so that the rate equation is given, not by the simple expression (i), but by expression (iii). Thus, kzbs is given by eqn (xxii): kzbs = k, + k, (k,/k,)+ [TMBII. (xxii) A plot of the extrapolated values of kEbs from fig. 2 plotted against [TMBIf gives a reasonably linear relationship, from which k , = 2.7 x s-l, k,/k$ = 0.52 (dm3 mol-l s-l)+ at 440 OC (fig.3). However, the extrapolation of the kobs against t plot is sensitive to experimental error in measuring the small yields of IBE in the early stages of reaction, since at other temperatures the lines through the kobs against t points do not always show the expected increase in gradient with increasing [TMB]. Greater accuracy in the value of k , and particularly of k,/k$ can be obtained by using a computer treatment which, in effect, puts a mean line of the correct gradient for the particular mixture through each set of (kobs, t) points. Numerical integration of the differential eqn (iv)-(ix) is ([TMB 1 /Torr)j FIG. 3.-Plot of k& against [TMBf. 440 O C , data from intercepts of fig. 2.1174 DE C 0 M PO S I TI 0 N OF 2,2,3,3-TE T R A M E THY L B U T A NE carried out as for KCl-coated vessels, except that a differential equation is always written for H,O,.Six TMB + 0, mixtures were chosen, with concentrations ranging from 0.1 to 4.0 Torr at a total pressure of 15 Torr, except at the lowest temperature of 400 O C where 30 Torr of 0, was used with the total pressure adjusted to 60 Torr with N,. For each mixture, the yield of IBE was measured at four times where the consumption of TMB was c 5%; under these conditions secondary reactions of IBE are negligible. As with KCl-coated vessels, an optimisation procedure adjusted values of the parameters k , and k,/k$ so as to give minimum r.m.s. deviation between observed and calculated IBE yields. The only significant parameters determining the calculated [IBE] against t curve are k,, k,/k$ and the parameters for decomposition of H,O,.k, is accurately known, with M = H,, but the value of M,, given by eqn (xxiii), depends on the value taken for TMB relative to H, in reaction (8). Because of the low pressure of 0, and N,, TMB can make major contributions to M, [the value of M in reaction (S)]: M, = 0.35[0,] +O.43W2] + a(TMB) [TMB]. (xxiii) At each temperature, optimum values of k, and k,/k$ were obtained for values of a(TMB) = 1, 2.5, 5.0 and 10 relative to H, = 1. (The values of 0.35 and 0.43 for 0, and N, have been obtained in previous studies?) Unfortunately, the minimum r.m.s. 0.6 X x \ x “E I W Y P .E? i . 0 - n n n L 1.24 1.28 1.32 1.36 1.40 1.44 1.&8 103 KIT FIG.4.-Plot of log k,/k+ against 1/T. Aged B,O,coated vessels: x , u(TMB) = 1 ; 0, a(TMB) = 2.5; 0, u(TMB) = 5; A, u(TMB) = 10. KCl-coated vessels: 0, a(TMB) = 2.5. deviation for each value of a(TMB) did not vary sufficiently to eiable the optimum value of a(TMB) to be determined. However, increasing a(TMB) had relatively little effect on k,, whereas the value of k,/k$ decreased significantly, as shown in fig. 4. Fig. 4 also gives the values obtained for k,/k,f over the range 41 8-466 O C in KCl-coated vessels. In KCl-coated vessels, where a value of a(TMB) = 2.5 was used, the value for k,/k$ is fairly insensitive to a(TMB), partly because the contribution from H,O, dissociation is smaller, and partly because higher pressures of 0, and N, were used, so that the contribution of TMB to M, is less important; variation of a(TMB) from 1 to 5 had negligible effect on k,, and only altered k,/k,f by 1%.Fig. 4 shows that inconsistent variation of k,/kf with temperature is obtained when a(TMB) is givenR. R. BALDWIN, M. w. M. HISHAM, A. KEEN AND R. w. WALKER 1175 the unlikely high value of 10, and that a value of 2.5 gives the best consistency with the results in KCl-coated vessels. The best line through all the results, both in B203- and KC1-coated vessels, gives A,/A7i = 4.4 x lo5 (dm3 mol-l s-')!, E4 -4E7 = 81.7 f 8 kJ mol-l. If the present value15 of k, = 2 x lo9 dm3 mol-1 s-l, independent of temperature, is accepted, A , = 1.97 x 1Olo dm3 mol-1 s-l, E4 = 81.7f 8 kJ mol-l. The value obtained for k , is insensitive to a(TMB), the variation being an increase of 20% at 4OOOC and an increase of 4% at 52OOC as a(TMB) is increased from 1 to 10.The results, together with the r.m.s. deviation and the corresponding values of k , obtained in KCl-coated vessels, are summarised in table 4. TABLE 4.-vALUES OF k , IN B,03- AND KCI-COATED VESSELS _____~ ~ aged B,03coated vessel KC1-coated vessel r.m.s. dev. T/"C k,/s-l (%) k,/s-' - 1.45 x 15.8 400 420 6.17 x 5.7 440 2.77 x 10-5 3.6 2.56 x 10-5" 470 1.70 x 10-4 3.5 2.01 x 10-4" 500 1.37 x 10-3 3.7 1.33 x 10-3" 520 4.05 x 10-3 3.8 3.98 x 10-3a - 1.29 x - - 542 a These values differ slightly from those quoted in ref. (2) because of allowance for the interaction between homogeneous and heterogeneous termination. The high r.m.s. deviation at 400 OC reflects the very small amount of reaction at the lowest TMB concentrations even with reaction times of 2000 s.Measurements at 400 and 420 OC were included to examine the possible effect of an inefficient surface destruction of H,O,. However, both at 400 and 420 OC, introduction of values of k , in the range 0.001-0.008 s-l increased rather than reduced the values of the r.m.s. deviation. The values obtained for k , only changed by ca. 2% over the range k, = 0-0.008 s-l, whereas k4/k7i changed by ca. 15%. Little error is thus likely in the values of k , and k,/k7i due to surface destruction of H202. The best straight line through the values of k , in B203-coated vessels (omitting the value at 470 "C) gives A , = 1.26 x 1017 s-l, El = 295.8 kJ mol-l. Combination of the results for the B203- and KC1-coated vessels in table 4 gives A , = 1.04 x 1017 s-l, El = 294.7 kJ mol-l.These values may be compared with A , = 1.08 x 1017 s-l, El = 295.1 kJ mol-l obtained from the studies in a KC1-coated vessel. The differences are within experimental error. Examination of the scatter of the [IBE] against time data and the log k , against l/Tdata suggests that individual values of k , are accurate to within f. 4%. A 4% error in k, at the extremes of the temperature range would give an error of 1 % in El. The estimated accuracy of El is thus within + 3 kJ mol-l, consistent with the statistical accuracy of ca. 1%. The very close agreement with previously published values2 ( A , = 1.20 x 1017 s-l, El = 295.4 kJ mol-lj confirms the validity of the suggested thermodynamic data2q l6 for the t-butyl radical.It has been suggested', that the decomposition of H20, on a KCl-coated surface1176 DECOMPOSITION OF 2,2,3,3-TETRAMETHY LBUTANE may produce HO, radicals, although the fraction of H,O, giving HO, radicals is not stated. The agreement of the values of k, with the two vessel coatings indicates that the previously published value of El, from which theenthalpy and entropy of the t-C,H, radical have been calculated, has not been distorted by such production. Moreover, the agreement of the two sets of values of k,/k,g suggests that in the range 420-470 OC such production of HO, radicals does not significantly increase the concentration of HO, in KCl-coated vessels. This work was supported by the United States Office of Scientific Research through the European Office of Aerospace Research, United States Air Force. A grant from the S.R.C. for provision of gas-chromatographic facilities is gratefully acknowledged. G. A. Evans and R. W. Walker, J. Chem. SOC., Faraday Trans. 1, 1979, 75, 1458. G. M. Atri, R. R. Baldwin, G. A. Evans and R. W. Walker, J. Chem. SOC., Faraday Trans. 1, 1978, 74, 366. R. R. Baldwin, I. A. Pickering and R. W. Walker, J. Chem. SOC., Faraday Trans. I , 1980, 76, 2374. R. R. Baldwin and L. Mayor, 7th Int. Symp. Combustion (Butterworths, London, 1958), p. 8. R. R. Baldwin and D. Brattan, 8th Znt. Symp. Combustion (Williams and Wilkins, Baltimore, 1962), R. R. Baldwin, D. Jackson, R. W. Walker and S. J. Webster, Trans. Faraday SOC., 1967, 63, 1676. R. R. Baldwin and J. Howarth, J. Chem. SOC., Faraday Trans. I , 1982, 78, 451. R. R. Baldwin, A. R. Fuller, D. Longthorn and R. W. Walker, J. Chem. SOC., Faraday Trans. 1,1974, 70, 1257. R. R. Baldwin, M. J. Matchan and R. W. Walker, Trans. Faraday SOC., 1971, 67, 3521. York, 1967). 68, 1362. p. 110. lo J. 0. Hirschfelder, C. F. Curtiss and R. B. Bird, Molecular Theory of Gases and Liquids (Wiley, New l1 R. R. Baldwin, A. R. Fuller, D. Longthorn and R. W. Walker, J. Chem. SOC., Faraday Trans. 1, 1972, l2 D. R. Blackmore, J . Chem. SOC., Faraday Trans. 1, 1978, 74, 765. l 3 R. R. Baldwin, A. Keen and R. W. Walker, unpublished work. l4 A. B. Nalbandyan, 17th Int. Symp. Combustion (The Combustion Institute, Pittsburgh, 1978), p. 533. l5 R. W. Walker, in Reaction Kinetics (Spec. Period. Rep., The Chemical Society, London, 1975), l6 R. R. Baldwin, R. W. Walker and Robert W. Walker, J. Chem. SOC., Faraday Trans. 1, 1980,76,825. vol. 1, p. 161. (PAPER 1/799)