摘要:

J . Chem. SOC., Faraday Trans. I , 1982, 78, 1615-1627 Arrhenius Parameters of Elementary Reactions Involved in the Oxidation of Neopentane BY ROBERT R. BALDWIN, MOHAMED W. M. HISHAM AND RAYMOND W. WALKER* Department of Chemistry, The University, Hull HU6 7RX Receiued 27th July, 1981 The reactions of neopentyl radicals in an oxidising environment have been studied by adding neopentane to slowly reacting mixtures of H, + 0, over the temperature range 380-520 OC. Over a wide range of mixture composition, the only detectable initial products at these temperatures are 3,3-dimethyloxetan (DMO), acetone, i-butene, methane, and formaldehyde. A relatively simple mechanism involving the formation of neopentylhydroperoxide (QOOH) radicals gives a quantitative interpretation of the product yields.Although the major source of i-butene is the C-C homolysis of neopentyl radicals, a significant proportion is formed in reaction (6) (6) From measurements of the product ratios ([acetone] + [DMO])/[i-butene] and [acetone]/[DMO] at each temperature used, Arrhenius parameters have been determined for a number of the elementary steps. The recommended value of k , = 1.20 x lo', exp (- 120 kJ mol-'/RT) s-' for the 1,5p H-atom transfer in neopentylperoxy radicals is compared with previously determined parameters for the 1,4p H-atom transfer in ethylperoxy radicals. With the Arrhenius expressions for these two transfers as a basis, thermochemical calculations are used to obtain a self-consistent set of Arrhenius parameters from values of rate constants at 480 O C for primary, secondary and tertiary H-atom transfers in alkylperoxy radicals involving ring sizes in the transition state varying from 4 to 8 (CH,),C(CH,OOH)CH, + (CH,),C=CH, + HCHO+OH.(CH,),CCH,O, + (CH,),C(CH,OOH)CH,. (3) The urgent need for accurate rate data for elementary reactions involved in hydrocarbon oxidation has been accentuated recently by attempts to use computer modelling to examine many practical combustion problems. Although the position has improved markedly over the last decade, rate constants and Arrhenius parameters for many elementary steps are still unknown.'* This situation applies particularly to the important reactions responsible for the formation of 0-heterocyclic and other oxygenated products, which are frequently formed in high yield in the temperature range 350-550 OC.Earlier studies with ethane,3 n-b~tane,~ i-b~tane,~ n-pentane6 and neopentane' have established that these products are formed in primary processes involving the following general mechanism : R+ 0, $ RO, RO, + QOOH (H atom transfer) QOOH -+ 0-heterocycle +OH QOOH + 0, + oxygenated products. The most extensive investigation of these reactions has been made by studies' of the addition of small amounts (1 %) of neopentane to slowly reacting mixtures of H, + 0, in aged boric-acid-coated Pyrex vessels at 480 "C. Two particular features 16151616 ELEMENTARY STEPS IN THE OXIDATION OF NEOPENTANE of neopentane facilitate a relatively simple interpretation of results. (i) All the C-H bonds are identical so that only one species of alkyl radical is involved.(ii) Unlike the majority of simple alkyl radicals, where reaction with 0, to form the conjugate alkene is the dominating process over the temperature range 30O-55O0C, such a reaction is structurally impossible for the neopentyl radical, so that reactions normally of minor importance with other radicals are dominant. In the H, + 0, addition system, neopentyl radicals are formed by radical attack on the neopentane, and at 480 O C the only initial products of importance are i-butene, 3,3 -dime t hyloxe tan (DM 0), acetone, methane and formaldehyde, which collectively account for at least 95% of the neopentane con~umed.~ Product formation was fully explained by the following reaction scheme. CH4 HCHO In confirmation of this mechanism, the relative yields of the products were found to fit eqn (i) and (ii) over a wide range of mixture composition:7 d([acetone] + [DMO])/d[i-butene] = K , k,[O,]/k, d [ ace t one] /d[ DMO] = k , [ 0 ,] / k , .(ii) K , is the equilibrium constant for reaction (2). In deriving eqn (i), the justifiable assumption k , << k-, has been made. Rate constants for a number of the elementary steps were obtained at 480 O C . With the advent of more sophisticated and sensitive gas chromatography, sufficient accuracy in analysis of products is available to justify an attempt to determine Arrhenius parameters for the elementary steps. EXPERIMENTAL The apparatus and general procedure have been described el~ewhere.~? * Product analysis was carried out using a Pye 104 gas chromatograph, the output of which was fed via an interface unit to the Perkin-Elmer Sigma 10 microprocessor unit which acts as a sensitive integrater.Phillips research grade neopentane (ca. 99.9%) was used without further purification, other than degassing and trap-to-trap distillation under vacuum. No impurities were then detectable by gas chromatography. Pressures below 50Torr were measured by use of a Southern Electronics S.E.180 pressure transducer, the output of which was fed to a 1 mV recorder, maximum sensitivity being ca. 1 Torr for full-scale deflection. An aged boric-acid-coated vessel, 52 mm in diameter, was used. As previously, a standard mixture of high N, content was selected containing 140, 70 and 290 Torr of H,, 0, and N,,R. R. BALDWIN, M.W. M. H I S H A M A N D R. W. WALKER 1617 respectively. Independent variation of H, and 0, was achieved by interchanging these gases with N,. As slight changes in the relative concentrations of H, 0, OH and HO, when an additive is present are not important for analytical work,' 1% of neopentane was used to facilitate the accurate measurement of products, particularly in the early stages of reaction. RESULTS AND DISCUSION Detailed product analyses were carried out using 5 Torr of neopentane, 140 Torr of H,, and variable pressures of 0, at a fixed total pressure of 500 Torr by use of added N,. As in the previous study at 480 OC, the only major initial products between 380 and 520 OC are methane, formaldehyde, 3,3-dimethyloxetan (DMO), acetone and i-butene, and all are clearly formed in primary processes. Variation in the pressure of neopentane between 1 and 10Torr and in the total pressure between 250 and 600 Torr had no observable effect on the product distribution. Increase in the pressure of H, between 15 and 300 Torr increased the [CH,]/[HCHO] ratio, but did not alter the distribution of the other products. These observations are completely consistent with the suggested mechanism and, in particular, confirm that reaction (1) is in its high-pressure range under the conditions used, in agreement with Furimsky and Laidler's9 observations. The 0, pressure was normally varied over a wide range at each temperature used between 380 and 520 OC.Maximum pressures of 0, of 100 and 70 Torr at 380 and 520 OC, respectively, were necessary because of the low yields of i-butene at low temperatures and the very fast rates at 520 OC.The minor products, isobutyraldehyde, 2-methylprop-2-en- 1 -al, 2,2-dimethyloxiran, 2-methylprop-2-en- 1-01 and propene, were also detected. At all temperatures and concentrations used, the total molar con- centration of the minor products accounted for < 5 % of the neopentane consumed at 10% reaction. Further, the relative yield of these products increased noticeably as the neopentane was consumed in the early stages of reaction, suggesting that they were mostly formed in secondary processes. It is considered unlikely that these minor products account for > 1-2% of the neopentane consumed initially, and they are ignored in later discussion. Owing to consumption and to formation in secondary processes, the relative concentrations of the major products varied with extent of reaction.Initial values of R, = ([DMO] + [acetone])/[i-butene] the ratios and R, = [acetone]/[DMO] were obtained from plots of the observed values of the ratios against [i-butene] and [DMO], respectively, with extrapolation to zero reaction in each case. Fig. 1 and 2 show typical plots for the range of 0, pressures used at 440 "C; similar plots were obtained at all temperatures used. R, varies with [O,] in a closely linear manner as predicted by eqn (ii) and, as fig. 3 shows, R,/[O,] decreases by < 10% over a tenfold increase in [O,], except possibly at 500 OC where the acetone yield at low [O,] is so small that reliable measurement is difficult.However, although the value of R, increases markedly with [O,] as predicted by equation (i), fig. 4 shows that the value of R,/[O,] falls noticeably as [O,] is increased. The simplest explanation for the fall-off is that additional i-butene is formed from the (CH,),C(CH,OOH)CH, radical by reaction (6) in competition with the formation of DMO and acetone by reactions (4) and (9, respectively. Formation of alkene from alkylhydroperoxide (QOOH) radicals has been (CH,),C(CH,OOH)CH, -+ (CH,),C=CH, + HCHO + OH (6)1618 ELEMENTARY STEPS I N THE OXIDATION OF NEOPENTANE I I 0 0.1 0.2 0.3 [ i-butene I /Torr FIG. 1 .-Plots of ([acetone] + [DMO])/[i-butene] against [i-butene] at 440 O C . p(H,) = 140, p(neopentane) = 5, total pressure = 500 Torr. A, p ( 0 , ) = 30; 0, p ( 0 , ) = 70; x , p ( 0 , ) = 140; 0, p ( 0 , ) = 210; V, p ( 0 , ) = 300 Torr.1.6 - 0 1 . 2 E E 2 a, 0 + 2 0.8 Y 0 . 4 0 I I 1 0.1 0.2 0.3 [ DMO l/Torr FIG. 2.-Plots of [acetone]/[DMO] against [DMO] at 440 "C. p(H,) = 140, p(neopentane) = 5, total pressure = 500 Torr. A, p ( 0 , ) = 30; 0, p ( 0 , ) = 70; x , p ( 0 , ) = 140; 0, p ( 0 , ) = 210; V, p ( 0 , ) = 300 Torr.R. R. BALDWIN, M. W. M. HISHAM AND R. W. WALKER 1619 l1 although experimental evidence was not suggested previously by several available. Application of stationary-state treatment now gives eqn (iii) (iii) where a = k , / k , and jl = k,/k,. Further, the intercepts I in fig. 4 are given by eqn 1.5 1 . 1 3 I = K2k3/kl(l +p). I 1 r V - T-L V D D --% 1 I I 80 160 240 [0,1 /Ton Frci.3.-Plots of R 2 / [ 0 2 ] against [O,] at different temperatures. p(H,) = 140, p(neopentane) = 5, total pressure = 500 Torr. 0, 380; 0, 400; x , 440; V, 480; A, 500 OC. Values of a may be obtained directly from the results in fig. 3, and the values of jl and K , k3/k1 at each temperature may be selected to give the best fit to the experimental points shown in fig. 4. The values of K , k , / k l are determined largely by extrapolation through the experimental points, but the accuracy is increased by use of the correct value of jl. As shown, the fall-off in the observed ratio R1/[O,] is most marked at low [O,], and this is consistent with the predicted behaviour from eqn (iii). The ‘best’ fits (full lines) are shown in fig. 4, and the values of jl and K2k3/kl used are given in table 1, together with the corresponding values of a from fig.3. No simple explanation of the slight fall in R,/[O,] (= a) is available; a is taken as the intercept in fig. 3. Arrhenius plots are given for a and p in fig. 5 and for K2k3/kl in fig. 6. The closely linear relationships provide strong support for the view that the same basic mechanism operates over the whole temperature range used. An excellent straight line may be drawn through the points for a and that for is extremely satisfactory in view of the1620 ELEMENTARY STEPS I N THE OXIDATION OF NEOPENTANE uncertaintyofca. lO%inthevaluesofB.ValuesofA, = (1.67f0.35) x dm3 mol-l, E, = - 81 f 1.5 kJ m o t 1 and (with double weighting of the points at 400, 440 and 480 "C) Ap = 225 If: 90, Ep = 43 f 3 kJ mol-l are obtained.A mean line through the points in fig. 6 gives the overall parameters associated with K,k,/k, of Aobs = (2.8 f 1.5) x dm3 mol-l, Eobs = - 105 f 7 kJ mol-l. The points at 380 and 0.2 0 80 160 2 40 3 20 FIG. 4.-Plots of R,/[O,] against [O,] at different temperatures. p(H,) = 140, p(neopentane) = 5, total pressure = 500 Torr. 0 , 3 8 0 ; 0 , 4 0 0 ; x ,440; V, 480; A, 500; 0 , 5 2 0 OC. Points, experimental; full lines, calculated. [O,IITO~ TABLE 1 .-VARIATION OF a, P AND K , k3/kl WITH TEMPERATURE T/OC a/dm3 mol-l P ( K , k3/k,)/dm3 mol-l 380 400 440 480 500 520a 493 319 150 69 48 0.07 (0) 0.11 0.16 0.2 1 0.30 - 5660 3080 1345 497 316 196 a a and P could not be determined accurately at 520 OC, and since K 2 k 3 / k , can only be obtained by assuming linear extrapolation, the value is less accurate than at other temperatures.400 O C in fig. 6 are ca. 25% low relative to a straight line drawn through the other points. Two causes of possible inaccuracy may be suggested for the low-temperature results. (i) The amount of i-butene formed at 380 and 400 O C is extremely low, particularly at high pressures of 0,; at 380 O C detectable amounts of i-butene were obtained only over the lower range of 0, pressure. (ii) The mechanism may be more complicated at the lower temperatures, and in particular the possibility of surface reactions cannot be ignored, as observed by Knox.12 As the importance of surface reactions undoubtedly decreases at higher temperatures,13 it is recommended that the most reliable Arrhenius parameters for K , k3/k1 are those obtained from the straight line drawn through the points between 440 and 52OoC, from which Aobs = (7.2 2.0) x dm3 mo1-l and Eobs = - 1 13 f 2 kJ mol-l.R.R. BALDWIN, M. W. M. HISHAM AND R. W. WALKER 1621 - 0.5 -0.7 Qa 0, 5 -0.9 -1.1 I 1 I 1.3 1.4 1.5 103 KIT FIG. 5.-Plot of log a and of log /3 against 1/T. Left-hand scale: x , log 8. Right-hand scale: 0, log a. 3 . 8 - I 3 . 4 E "E '0 . 2 - Do - 2.6 2.2 I I I I I 1 1 1.3 1.4 1.5 FIG. 6.-Plot of log (K2k3/kl) against 1/T. 103 KIT ARRHENIUS PARAMETERS FOR THE ELEMENTARY REACTIONS REACTIONS (I), (2) AND (3) By the use of the simple Arrhenius expression k = Ae-E/RT, which is certainly of sufficient precision over the temperature range concerned, the suggested mechanism gives A2A,/A-,A1 = 7.2 x lops dm3 mol-1 and E,+E3- E-,-E, = - 113 kJ mol-l.The only experimentally determined Arrhenius parameters for reaction (1) are those1622 ELEMENTARY STEPS I N THE OXIDATION OF NEOPENTANE determined by Furimsky and la idle^-,^ who studied the mercury-photosensitised decomposition of neopentane and give the high-pressure values A , = 2.5 x lo1, s-l and El = 121.3 kJ mol-1 over the temperature range 230-335 OC. In addition, Anderson and Bensonl, give k , = 103.3*0.3 s-l at 762 K from a study of the pyrolysis of neopentane in the presence and absence of HC1. Although their value has been increased to 5.5 x lo3 s-l by the use of more recent kinetic data,' it is considerably below the value of 1.21 x lo5 at 762 K calculated from the parameters of Furimsky and Laidler.Unfortunately, Furimsky and Laidler use a fairly complex interpretation, and their parameters are directly dependent on rate constants for the reaction between CH, radicals and neopentane and for the combination of CH, radicals with t-C,H, radicals, for which they assume lolo., dm3 mol-1 s-l, which is almost certainly too high.15 Although their parameters for reaction (1) are not unreasonable, it may be argued on thermochemical grounds that both A and E are slightly low. It has been suggestedl6? l7 for the C-C homolysis of alkyl radicals that A S # z 4 J K-l mol-l, so that absolute rate theory gives A, = 1013.65 s-l at the mean temperature used by Furimsky and Laidler. El may be estimated from the value of E-, and of AU,. Based on the recently revisedls values of the heats of formation of alkyl radicals, AHf~[(CH,),CCH,],,, = 34.5 & 3 kJ mol-1 is taken with standard values for CH, and i-butene to give AU, = 86.5 f 5 kJ mol-1 at the mean temperature used by Furimsky and Laidler.Kerr and Parsonagel9 have reviewed the data on the addition of CH, radicals to i-butene, which occurs mainly at the terminal position, and recommend Llu = 28.9 kJ mol-l but no value for E-, CH, + (CH,),C=CH, -+ (CH,),CCH,CH,. ( - 1 4 Recently, Canosa and Marshal120 have obtained AE = 16.6 f 2.6 kJ mol-l over the temperature range 20-328 OC for the difference in the activation energy between non-terminal and terminal H-atom addition to i-butene. As reaction (- 1) is only 2 1 kJ mol-1 less exothermic than (- 1 a), compared with a difference of 3 1 kJ mol-l for the analogous additions of H atoms, a value of E-, - Llu = 12 & 3 kJ mol-1 is suggested. Combining the thermochemical values then gives El = 127.4 & 6 kJ mol-l.As it is impossible to analyse the data of Furimsky and Laidler critically, the Arrhenius parameters obtained for reaction (1) by thermochemical methods, which are more consistent with data on other alkyl radicals,l8 are considered the more reliable so that the expression k 1 - - 1013.65*0.2 exp(- 127.4+6 kJ mol-l/RT) s-l is recommended. of 10 kJ mol-l for the t-C,H, group gives Use of Benson's additivity data2, with an assumed internal rotational energy barrier K2 = 2.7 x exp( + 105.1 kJ mol-l/ RT) dm3 mol-l over the temperature range 350-550 OC.Combination of the parameters for reactions (1) and (2) with the observed values of A,A,/A-,A, = 7.2 x dm3 mol-1 and E,+E,-E-,-E, = - 113 kJ mol-l gives A , = 1.2 x lo1, s-l and E3 = 120 kJ mol-l. Use of the experimental values of Furimsky and Laidler for reaction (1) would give A , = 6.7 x 10l2 s-l and E3 = 113.5 kJ mol-l. The only other experimental parameters available for a similar process are those for the 1,4p H-atom transfer in reaction (3e) C,H,OO -+ C,H,OOH (3 4R. R. BALDWIN, M. W. M. HISHAM A N D R. W. WALKER 1623 for which the values A,, = 1013.3*0.6 s-l and E,, = 143.7f 10 kJ mo1-l have been obtained, between 400 and 540 OC. Based on transition-state theory, the A factors for the H-atom transfers should decrease with increase in the ring size in the transition state owing to increasing loss of entropy as additional internal rotations are restricted.It is likely that the loss of entropy is ca. 17 J K-' mol-l per internal rotation restricted in the transition state, so that A(1,4)/A(1,5) = 7.7. In determining the A factor ratio for reactions (3) and (3e), allowance should be made for the different path degeneracies of 9 and 3, respectively, so that on a per C-H bond basis the value of A,,/A, is 8.9 using the experimental values of Furimsky and Laidler for reaction (1) and 5.1 using the thermochemically modified data. In view of the possible errors in the quantities, both of these values are in good agreement with the predicted ratio of 7.7, and it is certainly not possible to use the comparison to distinguish between the two values of the A factor suggested for reaction (3).The value of E3,-E3 of ca. 30 kJ mol-1 (k, expression for Furimsky and Laidler) or 25 kJ mol-l (k, expression thermochemically) may be compared with a value of 25 kJ mol-l suggested by Benson22 for the difference in strain energy involved in the formation of the two ring transition states, which will be the only important factor in determining the difference in the activation energies. REACTIONS (4) AND (6) No independent experimental values of k , or k , are available. However, on thermochemical grounds it is likely that reaction (6) involves initial C-C homolysis into i-butene and the CH,OOH radical, so that its Arrhenius parameters will be closely related to those of reaction (1). From thermochemical calculations, it is suggested that A , = A, and, based on data for the addition of CH, and substituted methyl radicals, particularly CCl,, that the replacement of one of the CH, groups in the neopentyl radical by the powerfully electron-withdrawing CC1, group lowers the bond dissociation energy of the adjacent C-C bond by ca.24 kJ mol-l. By comparison it is suggested that the presence of the less effective OOH group lowers the bond dissociation energy of the adjacent C-C bond in (CH,),C(CH,OOH)CH, by 12 kJ mol-l, which may be taken as the value of El-E,. Based on the expression of Furimsky and Laidler fork, then A6 = 2.5 x s-l, E6 = 109 kJ mol-', and based on the thermochemically modified expression for k,, then A6 = 4.5 x loi3 s-', E6 = 116 kJ mol-l.Combination with the experimental values of Eg = 43 kJ mol-l and AD = 225 gives the two sets of parameters for reaction (4), A , = 1.1 1 x lo1, s-l, E4 = 66 kJ mol-l and A , = 2.0 x loll s-l, E, = 73 kJ mol-l, respectively. These parameters for reaction (4) may be compared with Benson's2, estimate of A z 10" s-l and E z 63 kJ m o t 1 for the decomposition of alkylhydroperoxide radicals (formed by 1,5 transitions) into oxetans, and with Mill'sz4 experimental determination of A = s-l, E = 59 8 kJ mol-l for the liquid-phase decomposition into 2,2,4,4-tetramethyloxetan in the following reaction: H2C-c(CH3)2 -+ I I +OH. CH3 I CH3 I I CH3C-CH,-C-CH, (CH,),C-0 OOH REACTION (5) Combination of the experimental values of A , = 1.67 x dm3 mol-l, E, = - 81 kJ mol-1 with the Arrhenius parameters given above for reaction (4) gives1624 ELEMENTARY STEPS I N THE OXIDATION OF NEOPENTANE A , = 1.85 x lo7 dm3 mo1-1 s-l, E, = - 14.7 kJ mol-l or A , = 3.3 x lo7 dm3 mol-l s-l, E, = -8 kJ mol-l.The overall reaction ( 5 ) probably occurs in at least two stages: (CH,),C(CH,OOH)CH, + 0, -+ (CH3),C(CH,00H)CH202 (7) (8) Evidence presented previously7 suggests that reaction (7) is effectively equilibrated, and this would give a simple explanation of the negative activation energy associated with overall step (5). Accepting that reaction (7) is equilibrated and that K , = K,, on thermochemical grounds, then from k, = K7k,, values of A , = 6.9 x loll s-l, 4, = 90 kJ mol-1 or A , = 1.25 x lo1, s-l, E, = 96.5 kJ mol-l are obtained. These parameters are not unreasonable for a step which may involve cyclisation and possible formation of a bi-radical. The alternative possibility that k, 9 k-, would give k , = k,.It is likely that k(QO0H +O, -+ QOOHO,) w k(R+0, -+ RO,), for which Ben~on,~ has suggested a value of ca. lo9 dm3 mol-1 s-l F ith zero temperature coefficient, so that combination with the values of A , and E, gives A , w 6 x 10l2 s-l and E4 z 81 kJ mol-l. These values differ considerably from those obtained earlier and are almost certainly too high. (CH3),C(CH,00H)CH,0, 4 CH3COCH3 + 2HCHO +OH. TABLE 2.-sUMMARY OF RATE DATA based on based on k, = 2.5 x 1013 exp( - 121 000/RT) s-l k , = 4.45 x 1013 exp( - 127 500/RT) s-l reaction A/s-l E/kJ mol-l A/s-l E/kJ mol-l (3) 6.7 x 10l2 113.5 1.20 x 1013 120 (4) 1 .1 1 x 10" 66 2.0 x 10" 73 (6) 2.5 x 1013 109 4.5 x 1013 116 (8) 6.9 x 10" 90 1.25 x 10l2 96.5 The Arrhenius parameters determined in the present study are summarised in table 2. Now that relatively precise Arrhenius parameters are available for reactions (3) and ( 3 e ) , the 1,4 H-atom transfer in C,H,OO radicals, parameters for other H-atom transfers may be estimated from values of rate constants determined only at 480 O C . As a basis for this, an A factor of 1.00 x 10l2 s-l per C-H bond will be taken for 1,5 H-atom transfers, being virtually the mean of the two values obtained earlier in the paper. Based on transition-state theory, as discussed earlier, the calculated A factor per C-H bond available for the transfer is decreased by a factor of 8 each time the size of the ring transition state is increased by one atom.The calculations are summarised in table 3 for H-atom transfers where rate data are available. Several points concerning the table may be made. (i) Where two determinations of the rate constant at 480 O C are available for a particular type of transition there is good agreement, particularly on a per C-H bond basis. All values depend ultimately on the rate constant for the appropriate radical- cracking reaction (which in turn depends on a radical-radical recombination rate constant) and on a calculated equilibrium constant for the appropriate R + 0, + RO, reaction. There may also be slight differences in the bond dissociation energies even for the same type of C-H bond (for example primary) in the RO, radicals.(ii) From studies, of the addition of i-butane to slowly reacting mixtures of H, + 0,TABLE ~.-ARRHENIUS PARAMETERS FOR H-ATOM TRANSFER REACTIONS IN RO, RADICALS reaction E/kJ mol-l k1s-l A/s-' ref. type kls-1 (per C-H) (per C-H)u ElkJ mol-l (for type) CH3CH2cH(OO)CH3 CH,CH,CH(OOH)CH, CH,CH,OO -+ CH,CH,OOHb (CH,),CCH,OO -+ (CH3),C(CH,00H)CH,b (CH,),C(H)CH,OO -+ CH,C(H)(CH,OOH)CH, CH,CH,CH,CH,OO -+ CH,CH,CH,CH,OOH CH,CH,CH,CH,CH,OO -+ CH,CH,CH,CH,CH,OOH CH3CH(OO)CH,CH3 -+ CH,CH(OOH)CHCH, CH,CH(OO)CH,CH,CH, -+ CH,CH(OOH)CHCH,CH, CH,CH,CH,CH,OO -+ CH,CHCH,CH,OOH CH,CH(OO)CH,CH,CH, -+ CH,CH(OOH)CH,CHCH, CH,CH,CH,CH,CH,OO 4 CH,CHCH,CH,CH,OOH CH,CH(OO)CH,CH, --+ CH3C(OOH)CH,CH3 (CH,),CHCH,OO -+ (CH,),CCH,OO 4 L4P 3 L4P present 1,5p present 1,5p 5 175P 4 M P 6 L7P (1 73s) 4 1,4s 6 1,4s 4 1,5s 6 1,5s 6 1,6s (1 97s) 4 1,3t 5 1,4t (1 3 ) (1,6t) (L7t) 8 x 10l2 2.2 x 103 7.3 x 10, 8 x 10l2 9.2 x 104 1.02 x 104 1.0 x 10l2 115.2 5 .7 ~ 104 6.3 x 103 1.0 x 10l2 118.2 4.45 x 104 7.4 x 103 1.0 x 10l2 117.2 6.0 x 104 2.0 x 104 1.25 x 1011 97.9 9.3 x 104 3.1 x lo4 1.55 x 1Olo 82.1 (6.4 x 1013) (168)c 2 . 4 ~ 104 1.2x 104 8 x 10l2 2.0x 105 1.0 x 105 1.0 x 10'2 3 . o ~ 105 1.5 x 105 1.0 x 10'2 6 . 6 ~ lo5 3.3 x 105 1.25 x 1011 80.4 (1.55 x lolo) (64)c 6.4 x 1013 153.3 1.83 x 105 1.83 x 105 8 x 10l2 110.2 (1.0 x lo1,) (84)c (1.25 x 10") (64)c (1.55 x 1 Ole) (48)c 145.5 144*8 1 1.95 x 103 6 . 5 ~ lo2 I 8 x 10l2 123.0 27*2 1 98.4 1 4 . 7 ~ 104 2 . 3 5 ~ 104 1.5 x 103 1 . 5 ~ 103 145 117 98 82 125 100 80 153 110 All A factors, except for the two reactions marked (b), have been calculated from the experimental values for these two reactions using transition theory, which itself is consistent with the two experimental A factors.Experimentally determined Arrhenius parameters for these reactions are discussed in the text. Activation energies given in brackets are calculated on the assumption (discussed in the text) that the differences in the activation energies for the transfer of a particular type of H atom (say primary) are due to differences in the strain energies in the ring transition *f states. Thus, for example, E( 1,3s) = E( 1,4s) + E( 1,3t) - E( 1,4t). r !a - ? ? W L o\ r4 wl1626 ELEMENTARY STEPS IN THE OXIDATION OF NEOPENTANE at 480 OC, the rate constant ratio for the 1,4t and 1,5p H-atom transfers in the (CH,),CHCH,OO radical has been determined directly and is not dependent on the values of any other rate constant.The value of k(1,4t)/k(1,5p) = 24.7 may be compared with values of 29 and 18 for the ratio using the two values for k( 1,5p) from the present work together with the absolute value for k(1,4t) from the i-butane s t u d i e ~ . ~ The good agreement provides strong support for the absolute values of the rate constants given in table 3. (iii) The activation energy for an H-atom transfer such as reaction (3) may be considered as the sum of the activation energy of the intermolecular abstraction reaction between RO, and RH and of the strain energy involved in the formation of the ring transition state. On this basis, the change in the observed activation energy with ring size in the transition state for a particular type of H atom transferred should be a measure of the difference in the ring strain energies and should be the same regardless of whether the comparison is made for primary, secondary or tertiary H atoms.This is supported by the results in table 3 as shown by E( 1,4p) - E( 1,5p) = 28 compared with E( 1,4s) - E( 13s) = 25 kJ mol-1 and E( 1,5p) - E( 1,6p) = 19 compared with E( 13s) - E( 1,6s) = 20 kJ mol-l. These differences may be compared with those estimated by Fish,lo who assumed strain energies in the ring transition states identical to those found in ground-state cycloalkane systems, as follows : eight-membered (1,7 transfer), 42 kJ mol-l; seven-membered (1,6 transfer), 27; six-membered ( 1 3 transfer), 2.5; five-membered (1,4 transfer), 27; four-membered (1,3 transfer), 109.Although the values for E( 1,4) - E( 1,5) are consistent with the values suggested by Fish, the other figures in table 3 do not support the view that there is minimum strain energy in the six-membered transition state but suggest a continuous decrease down to the eight-membered ring. To achieve a minimum in the six-membered ring, the A factor would have to increase markedly in the 1,6 and 1,7 transitions, and this is unacceptable theoretically. (iv) The residual strain involved in the 1,7 transfers is likely to be so small that the activation energies suggested in table 3 for the 1,7p, 1,7s and 1,7t transfers of 82, 64 and 48 kJ mol-l, respectively, should approximate to the activation energies for the intermolecular abstraction of primary, secondary and tertiary H atoms from alkanes by RO, radicals.No experimental values are available for these reactions of RO, radicals, but on thermochemical grounds their activation energies should be virtually identical to those of the equivalent reactions of HO, radicals. Recently a value of 81.7 k 8 kJ mol-l has been obtained for the abstraction of H atoms from tetramethyl- butane (all primary C-H) by HO, radicals,25 in excellent agreement with E(1,7p) = 82 kJ mol-l, if there is no strain energy involved. The lower values for abstraction at secondary and tertiary C-H positions are consistent with the decrease in endothermicity of these reactions. R. W. Walker, in Reaction Kinetics (Specialist Periodical Report, The Chemical Society, London, 1975), vol.1, p. 161. R. W. Walker, in Gas Kinetics and Energy Transfer (Specialist Periodical Report, The Chemical Society, London, 1977), vol. 2, p. 296. R. R. Baldwin, I. A. Pickering and R. W. Walker, J . Chem. SOC., Faraday Trans. 1, 1980, 76, 2374. R. R. Baker, R. R. Baldwin, A. R. Fuller and R. W. Walker, J . Chem. SOC., Faraday Trans. 1, 1975, 71, 736, 756. R. R. Baker, R. R. Baldwin and R. W. Walker, J . Chem. SOC., Faraday Trans. 1, 1978, 74, 2229. R. R. Baldwin, J. P. Bennett and R. W. Walker, J . Chem. SOC., Faraday Trans. 1, 1980, 76, 1075. R. R. Baker, R. R. Baldwin, C. J. Everett and R. W. Walker, Combust. Flame, 1975, 25, 285. R. R. Baldwin, D. E. Hopkins and R. W. Walker, Trans. Faraday SOC., 1970, 66, 189. E. Furimsky and K. J. Laidler, Can. J. Chem., 1972, 50, 1 1 15, 1123, 1129.R. R . BALDWIN, M. W. M. HISHAM AND R. W. WALKER 1627 l o A. Fish, in Organic Peroxides, ed. D. Swern (Wiley, New York, 1970). vol. 1, p. 141. R. T. Pollard. in Comprehensizle Chemical Kinetics, ed. C. H. Bamford and C. F. H. Tipper (Elsevier, Amsterdam. 1977). vol. 17, p. 249. J. Hay. J. H. Knox and J. M. C. Turner, Proc. 10th Int. Combustion Symp. (Combustion Institute, Pittsburgh, 1965). p. 33 1. l 3 R. R. Baldwin and R. W. Walker, Proc. 14th Int. Combustion Symp. (Combustion Institute, Pitts- burgh. 1973). p. 241. l 4 K. H. Anderson and S. W. Benson, J. Chem. Phys., 1964, 43, 3747. l 5 J. H. Purnell, in Frontiers of Free Radical Chemistry, ed. W. A. Pryor (Academic Press, New York, 1980), p. 93. Ifi R. R. Baldwin, R. W. Walker and Robert W. Walker, J. Chem. Soc., Faraday Trans. 1, 1981,77,2157. S . W. Benson and H. E. O'Neal. Kinetic Data on Gas Phase Unimolecular Reactions. NSRDS-NBS21 (U.S. Government Printing Office, Washington, D.C., 1970). l H R. R. Baldwin. R. W. Walker and Robert W. Walker, J. Chem. Soc., Faraday Trans. I , 1980,76,825. l 9 J. A. Kerr and M. J. Parsonage, Eraluated Kinetic Data on Gas-phase Addition Reactions (Butter- 2n C. E. Canosa and R. M. Marshall, J. Chem. Soc., Faraday Trans. I , 1980, 76, 846. worths. London. 1972). S. W. Benson and R. Shaw, in Organic Peroxides, ed. D. Swern (Wiley, New York, 1970), vol. 1, p. 105. 22 S. W. Benson, J. Am. Chem. Soc., 1965, 87, 972. 23 S. W. Benson. The Mechanisms of Pyrolysis, Oxidation and Burning of Organic Compounds, N.B.C. Special Publication 357 (U.S. Department of Commerce, Washington, D.C., 1972), p. 121. 24 T. Mill. Proc. 13th Inf. Combustion Symp. (Combustion Institute, Pittsburgh, 1973). p. 237. 25 R. R. Baldwin, M. W. M. Hisham and R. W. Walker, .I. Chem. Soc., Faraday Trans. I , in press. (PAPER 1/1192) 53-2

ISSN:0300-9599    

DOI:10.1039/F19827801615

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. I, 1982, 78, 1629-1631 The Self-diffusion Coefficient of Sulphuric Acid BY KENNETH R. HARRIS^ Diffusion Research Unit, Research School of Physical Sciences, Australian National University, Canberra, A.C.T. 2600, Australia Received 3rd August, 1981 The self-diffusion coefficient, D, of sulphuric acid has been measured at 15.1 and 25 OC to a maximum pressure of 194 MPa using the n.m.r. steady-gradient spin+cho method. The pressure dependence of D is similar to that of methanol, (aD/i3p)T being negative, and dissimilar to that of water, where (aD/dp)T passes from positive to negative values at low temperatures. The purpose of this note is to report measurements of the pressure dependence of the self-diffusion coefficient, D, of sulphuric acid at temperatures in the region of the freezing point.The molecules of sulphuric acid are believed to be extensively associated by hydrogen bonds and the low diffusion coefficients observed are consistent with this. Of other hydrogen-bonded liquids, only water and methanol have been intensively studied: the pressure dependence of D for sulphuric acid is similar to that of methanol, but differs from that of water. EXPERIMENTAL The diffusion coefficients were measured using the n.m.r. steady-gradient spin-echo method. The apparatus used and the technique followed have been described previously.'* The precision and accuracy of the results are estimated to be k 1.5 and f 3%, respectively. The sulphuric acid sample was confined in a glass, Kovar and 3 16 stainless-steel bellows cell, sealed with a cap made from half a Swagelok union and a ball bearing, also of 316 stainless steel.This cell and the combined gradient coil and r.f. coil probe were contained in the pressure vessel. Temperatures and pressures were measured with an accuracy of k0.03 K and k 0.3 MPa, respectively. The Pt resistance sensors and Heise bourdon tube pressure gauge had been calibrated at the National Measurement Laboratory, CSIRO, Sydney. The sulphuric acid was prepared from concentrated (98 %) aqueous sulphuric acid solution (Ajax Chemicals, Sydney: AnalaR grade). Sulphur trioxide was distilled into a sample of the acid to form a dilute oleum solution which was titrated against more of the concentrated acid until the electrical conductivity was a minimum.This titration was done in a mixing chamber attached to one side of the conductance cell to which the burettes were fitted. The conductance cell had been calibrated at 25 OC with a 0.1 demal aqueous KCl solution prepared from fused, recrystallized KCl and conductivity water. The cell electrodes were platinized and the frequency dependence in the measurement range 0.3-2 kHz was negligible. Conductivities were measured with a Leeds and Northrup Jones-Dike bridge. The temperature of the oil thermostat was controlled within the limits k 3 mK. The cell constant was 35.77 f 0.01 cm-l. The minimum conductivity was found to be (1.0428 & 0.0005) x 1 0-2 S cm-' at 25 O C . Gillespie et aZ.3 have reported a value of (1.0432 k 0.0005) x lop2 S cm-l for this quantity, which corresponds to a water composition of 0.002 mol kg(so1ution)-'.A sample of acid of this conductivity was taken for the diffusion measurements, though it undoubtedly absorbed moisture during the time taken to rinse, fill and seal the n.m.r. cell. From measurements of t Present Address: Chemistry Department, Chelsea College, Manresa Road, London SW3 6LX. 16291630 SE LF-D I FFU S I ON COEFFICIENT 0 F SU L P HU R I C A c I D the rate of absorption of acid samples exposed to the air, the maximum amount of water contamination was estimated to be a mole fraction of 0.0002. The results are listed in table 1 and are shown plotted against pressure in fig. 1. TABLE 1 .--SELF-DIFFUSION COEFFICIENTS OF SULPHURIC ACID 15.10 0.09 23.8 51.0 98.8 149.5 24.6 50.7 100.9 151.1 193.7 25.00 0.09 45.4 41.6 39.3 33.8 29.2 65.7 61.9 59.4 50.3 43.8 41.5 20 1 1 I I 1 100 200 p/MPa FIG.1.-Self-diffusion coefficient of sulphuric acid as a function of pressure. DISCUSSION The properties of liquid sulphuric acid indicate that it is highly associated. Like water, it has a high boiling point (317 "C), dielectric constant (100 at 25 "C), and surface ten~ion.~ The entropy change on freezing (4.54 R)5 is similar to that of many alcohols, and the viscosity (24.54 mPa s at 25 "C)6 is also high. On the basis of X-ray diffraction measurements' of the solid structure, which is monoclinic, it has been suggested that each molecule is linked by hydrogen bonds to four others, there being two 0-0 distances, 0.264 and 0.287 nm. The bonded molecules form layers within the crystal.This is consistent with the high melting point, 10.371 0C.4K. R. HARRIS 1631 It has been observed that some salts, e.g. NH,HSO,, have negative Jones-Dole viscosity B coefficients in sulphuric acid solution.6 Sulphuric acid as a solvent is similar to water, glycol and other polyhydric alcohols in this respect, but dissimilar to methanol and other monohydric alcohols.* If this is interpreted to mean that such ions distort or break up the hydrogen bonding between solvent molecules, thereby increasing the fluidity, then it might be expected that a similar effect could be produced by the application of pressure. Such an increase in the fluidity, and therefore the self-diffusion coefficient, is well documented in the case of water9-13 close to and below the melting temperature (i.e.< 30 "C). On the other hand no such effect is observed in the case of methanol,l49 l5 where the lowest temperature investigated was 29 K above the freezing point,15 but (aD/@), was still clearly negative along this isotherm (ca. - m2 s-l MPa-l). The measurements reported here show no anomaly in the pressure dependence of D for sulphuric acid at the temperatures investigated. Due to freezing of the sample on the application of pressure, it was not possible to approach the freezing temperature more closely than 4.7 K. The freezing pressure at 15.1 "C was ca. 50 MPa and the higher pressure points lie in the supercooled region. The results suggest that the structure of sulphuric acid is quite different from that of water, though both are highly associated liquids. It may be that the liquid retains some of the planar association present in the solid.In this respect it is interesting that it is H30+ and NH,+ ions which have been observed to produce negative viscosity B coefficients. These are of course ions which can hydrogen bond with solvent molecules and so interrupt linkages between them. Further interpretation of these diffusion measurements requires high-pressure density data which are not yet available. K. R. Harris, R. Mills, P. J. Back and D. S. Webster, J. Magn. Reson., 1978, 29, 473. K. R. Harris, Physica (Utrecht), 1978, 93A, 593; 1978, 94A, 448. R. J. Gillespie, J. V. Oubridge and C. Solomons, J. Chem. SOC., 1957, 1804. R. J. Gillespie, Rev. Pure Appl. Chem., 1959, 9, 1. J. E. Kunzler and W. F. Giauque, J. Am. Chem. Soc., 1952, 74, 3472. R. J. Gillespie and S. Wasif, J. Chem. Soc., 1953, 215. J. Padova, Nonaqueous Electrolyte Solutions, in Water and Aqueous Solutions, ed. R. A. Home (Wiley-Interscience, New York, 1972), chap. 4. L. A. Woolf, J . Chem. SOC., Faraday Trans. 1, 1975, 71, 784. ' R. Pascard, C.R. Acad. Sci., 1955, 240, 2162. lo C. A. Angell, E. D. Finch, L. A. Woolf and P. J. Back, J. Chem. Phys., 1976, 65, 3063. l1 T. DeFries and J. Jonas, J. Chem. Phys., 1977, 66, 5393. l 2 K. Krynicki, C. D. Green and D. W. Sawyer, Faraday Discuss. Chem. Soc., 1979, 66, 199. l3 K. R. Harris and L. A. Woolf, J . Chem. SOC., Faraday Trans. I , 1980, 76, 377. l4 J. Jonas and J. A. Akai, J . Chem. Phys., 1977, 66, 4946. R. L. Hurle, Thesis (Australian National University, Canberra, 1981). (PAPER 1/1218)

ISSN:0300-9599    

DOI:10.1039/F19827801629

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1982, 78, 1633-1639 Critical Comments on the Interpretation of Heat Capacities of Activation and on the Experimental Evidence for other Recent Mechanistic Proposals for Solvolyses of t-Butyl Halides in Water and in Binary Aqueous Mixtures BY T. WILLIAM BENTLEY* AND GILLIAN E. CARTER Department of Chemistry, University College of Swansea, Singleton Park, Swansea SA2 8PP Receiued 3rd August, 1981 Kinetic data for solvolyses of t-butyl bromide in homogeneous solutions in water and in aqueous binary mixtures with acetone, methanol and ethanol at 25 O C are reported, illustrating the development of simple conductometric techniques for studying relatively fast solvolytic reactions of sparingly soluble solutes. Solvolyses of t-butyl chloride in water and in aqueous ethanol are re-examined.Experimental errors due to (a) equilibration of solute between liquid and vapour phases and (b) formation of hydrophobic aggregates which can be disrupted by ultrasonic irradiation are discussed. Four recent mechanistic proposals are critically reassessed. It is shown that: (1) the induction period observed at low solute concentrations for solvolyses of t-butyl chloride in water is due to the buffering action of absorbed carbon dioxide; (2) recent kinetic data on the effects of ultrasonic irradiation are unreliable; (3) deviations from first-order kinetics for solvolyses of t-butyl chloride in 40% ethanol + water mixtures are due to equilibration of solute between liquid and vapour phases. It is also argued that (4) the most recent interpretations of heat capacities of activation, suggesting partially reversible formation of an intermediate in hydrolysis of t-butyl chloride, do not take adequate account of alternative interpretations.Solvolyses of t-butyl halides, (CH,),CX, provide one of the cornerstones of homogeneous solution kinetics.' The reaction is first order and is believed to occur by rate-determining heterolysis of the carbon-halogen bond to give a contact ion pair. The role of solvent as nucleophile cannot be assessed from the first-order rate law, and it is likely that these reactions proceed with some bimolecular (S,2) character by rearside nucleophilic solvation of the developing positive charge on the a-carbon atom and/or on the P-hydrogen atom.2 Alternatively such reactions could be described as mainly dissociative with weak associative ~haracter.~ More complicated kinetics could arise if there were reaction between the strong acid (HX) produced during solvolysis and another component of the reaction mixture, e.g.isobutene (product) or alcohol (co~olvent)~ may react with HX, or HX may catalyse further reaction of t-butyl halide or of isobutene. Consequently it may be necessary to add a buffer (e.g. 2,6-lutidine, a sterically hindered weak base) to prevent these side reactions. Recently there have been four novel mechanistic proposals, which we now wish to evaluate critically: (1) that an induction period observed at low concentrations of t-butyl chloride in water is due to a requirement for acid catalysis;5 (2) that the effects of ultrasonic irradiation on solvolysis of t-butyl chloride in water and in aqueous ethanol are due to disruption of solvent structure;6 (3) that solvolyses of t-butyl chloride in 40 % ethanol + water mixtures show mechanistically significant deviations 16331634 SOLVOLYSIS OF t-BUTYL HALIDES from first-order kinetics;6 (4) that the negative heat capacities of activation observed for hydrolysis of t-butyl chloride are due to partially reversible formation of contact ion pairs (or of some other intermediate).' Our reassessment of the first three proposals is based largely on reinvestigation of the experimental evidence: accurate kinetic studies of organic solutes in water are made difficult by their low solubility and/or high reactivity.Also, t-butyl chloride is partitioned between liquid and vapour phases, so it is necessary to minimise the vapour space above the solution.8 In this paper we report kinetic data for t-butyl chloride and bromide in water and in aqueous binary mixtures.These studies require a combination of techniques for assessing the reliability of kinetic data for reactions of sparingly soluble solutes, because it is necessary to check whether truly homogeneous solutions are formed rather than aggregates of small number of molecules. EXPERIMENTAL t-Butyl halides were distilled from anhydroLj sodium sulphate and their purity was checked by n.m.r. and refractive index. Acetone, ethanol and methanol were purified as described previouslyg and binary mixtures were prepared as stock solutions by weight or in smaller quantities using calibrated pipettes.Conductance techniques for rapid reactions were as described previously'* with the following modifications of method A. The conductance cell (capacity ca. 20 cm3) was fitted with a side-arm used for rapid injection of < 20 mm3 of a 0.25-5 % solution of the alkyl halide in acetone. Either a nitrogen driven stirrer was used throughout the kinetic run or the cell was shaken vigorously using a B and T shaker before recording the conductance data. Servoscibe Is or Phillips PM8041 recorders (maximum chart speeds 1 cm s-' and 2 cm s-l, respectively) were used, the latter with millimetre graph paper. After using the mechanical shaker the conductance reading was not steady for ca. 1 s, and the fastest reactions (t-butyl bromide in water at 25 "C) were followed from 80-90% reaction onwards.Low-temperature measurements (< 20 "C) were carried out using a Grant refrigerated thermostatic bath ( f 0.05 "C) and other measurements using Grant SP50 thermostatic baths. Slower kinetic runs (ti > 120 s) were monitored using a Wayne Kerr model B33 1 autobalance conductance bridge. RESULTS Included in this paper are kinetic data for solvolyses of t-butyl chloride in aqueous ethanol at 0.5 OC (table l), of t-butyl chloride and bromide in aqueous ethanol at 25 OC (table 2) and of t-butyl bromide in aqueous methanol and aqueous acetone at 25 O C (table 3). These results were obtained conductimetrically using < mol dm-3 solutions, assuming that the change in conductance is proportional to the extent of reaction." Where direct comparisons can be made, our results agree well with those obtained titrimetrically.8* l2 By combining our techniques for sparingly soluble solutess with those for relatively fast solvolytic reactions,lo we examined a wider range of solvolyses of t-butyl bromide than was possible titrimetri~al1y.l~ Our result for the fastest reaction examined at 25 O C , t-butyl bromide in water, is in agreement with the extrapolated value previously obtained conductometrically.14 Satisfactory bromide/ chloride (Br/Cl) rate ratios (tables 2 and 3) provide further evidence for the reliability of the results.Additional internal checks of the reliability of the kinetic runs were carried out; it was established that: (i) the rate constant calculated for 50% reaction was in satisfactory agreement with that for the full run (> 85% reaction); (ii) the precision of fit to first-order kinetics was high (usually better than 0.3%); (iii) there was generally insignificant 'drift' of the conductance of the solution even after ten half-lives.T.W. BENTLEY AND G. E. CARTER 1635 TABLE l.-sOLVOLYTIC RATE CONSTANTS ( k ) FOR t-BUTYL CHLORIDE IN AQUEOUS ETHANOL AT 0.5 "C solvent rate constant k/ lo4 s-' mole ethanol (%)" fraction ref. (6)b ref. (12)C3d this workbTe 10 0.033 6.06 5.15 5.2 k0.2f 20 0.072 4.05 3.44 3.49 f 0.069 30 0.117 2.43 1.99 1.96 f 0.019 40 0.171 1.24 0.62 0.62 f 0.019 a % v/v ethanol + water; determined conductometrically ; calculated from data at 0 and 25 "C; minimum vapour space above the solution; conductivity cell unstirred; f mean of four kinetic runs; one additional kinetic run with continuous stirring gave k = 5.0 x determined titrimetrically ; s-'.9 Mean of two kinetic runs, errors shown are average deviations. TABLE 2.-sOLVOLYTIC RATE CONSTANTS ( k ) FOR t-BUTYL BROMIDE AND CHLORIDE IN AQUEOUS ETHANOL AT 25 O C solvent rate constant k/s-' rate constant k/s-l ethanol (%)" (CH313CBr (CH3),C Cle (n)b Br/C1 (water)d (7.0 f0.2) x lo-' (5) (2.8 k 0 . l ) x (10) 25 10 (4.9 f 0.2) x lo-' ( 6 ) (1.83k0.04) x (4) 27 20 (2.7 k0.2) x lo-' (6) (1.10f0.04) x (7) 25 30 (1.26f0.05) x lo-' (5) (4.8 f0.1) x 10-3 (3) 26 40 (4.0 k0.2) x (6) (1.46f0.03) x (21) 27 a v/v ethanol+water. Number of kinetic runs; errors shown are average deviations.In Data at other temperatures: k/s-'; -0.08 "C; (1.67 t 0.08) x lO-l, 14.46 "C; errors shown are average devia- satisfactory agreement with data reported in ref. (8). (2.08 & 0.07) x tions of duplicate runs; activation parameters A S = 93 kJ mol-l, A S = + 62 J mol-l K-l. TABLE 3.-%LVOLYTIC RATE CONSTANTS (k) FOR t-BUTYL BROMIDE IN AQUEOUS METHANOL AND AQUEOUS ACETONE AT 25 "C rate constant k / s l rate constant k/s-l solvent (%)" me thanolb Br/CIC ace toneb Br/Clc ~ ~ ~~ ~ 10 (4.7 f0.2) x 10-1 27 (3.6 f0.3) x 10-l 23 20 (2.57f0.07) x lo-' 26 (2.01 f0.06) x 10-1 27 30 (1.34 f 0.05) x 10-1 26 (8.6 f0.2) x 31 50 (2.69 f 0.02) x 31 40 (6.51 f0.06) x 29 (3.07 f 0.05) x 35 80 (8.8 fO.l) x 39d " Volume % of organic solvent+water. Determined condtktometrically at least in duplicate; k calculated from data at an accurately determined temperature close ( f 0.2 "C) to 25 "C.Rate constant for bromide [ref. (13)] is 8.7 x Bromide/chloride rate ratios; kinetic data for chloride from ref. (8). s-l, in agreement with this work.1636 SOLVOLYSIS OF t-BUTYL HALIDES Anomalous results were initially obtained for solvolyses of ca. 0.0005 mol dm-3 t-butyl bromide in 80% (v/v) methanol + water solutions. Although t-butyl bromide dissolved rapidly at 25 OC, there was little change in the conductance of the solution for ca. 4 min after injection of the bromide. The induction period remained the same irrespective of how long or how vigorously the mixture was shaken, but disappeared when solutions free from carbon dioxide were used.Similarly we observed the previously reported5 induction period of up to 30 s for hydrolysis of t-butyl chloride (injected as an acetone solution) in water at 25 OC, but no induction period when freshly distilled water was used. DISCUSSION The induction period for solvolyses of low concentrations of t-butyl halides in aqueous media appears to be due to the buffering effect of dissolved carbon dioxide, which is known to affect the conductance of dilute solutions of HCl.15 Consequently there is no reliable evidence supporting the proposal5 that hydrolysis of t-butyl chloride is catalysed by HCl. Solvolyses of t-butyl chIoride in water at 25 OC did not give satisfactory infinity conductance values, unless the vapour space above the solution was minimised [see also ref.(S)]. The second mechanistic proposal is based on rate accelerations observed on ultrasonic irradiation of solutions of t-butyl chloride in aqueous ethanol at 0.5 0C.6 Unfortunately, the kinetic data reported6 in the absence of ultrasonic irradiation are not consistent with literature values,12 nor with our measurements (table 1). t-Butyl chloride does not dissolve readily in aqueous ethanol at 0 OC, so we prepared the solutions at ca. 10 "C before equilibrating to 0.5 OC. Consequently ca. 80% of the total reaction was studied (table 1). We observed lower rate constants if the vapour space above the solution was not minimised, and lower rate constants are usually obtained if molecular aggregates instead of truly homogeneous solutions are f ~ r m e d . ~ Therefore these sources of error do not explain why the recently reported6 rate constants in the absence of ultrasonic irradiation are too high (table I).A trivial explanation such as incorrect temperature or solvent composition appears to be required. We have found that ultrasonic irradiation disrupts the molecular aggregates of sparingly soluble solutes in aqueous media.9 This effect could explain the 50-100% increase in rate constants on ultrasonic irradiation of solutions of t-butyl chloride in aqueous ethanol. Alternatively these effects may be due to inadequate temperature control-we were unable to obtain satisfactory thermostatic control inside an unstirred conductivity cell placed in an ultrasonic bath containing iced water, as described by earlier workers.6 Despite the unreliability of the experimental evidence it is conceivable that there is a small residual rate enhancement (i.e.not explicable by the effects discussed above) on ultrasonic irradiation of solvolyses of t-butyl chloride in aqueous ethanol. For hydrolyses of various esters it has been proposed that intense localized pressure increases in collapsing bubbles account for the observed rate enhancements;16 it is known that solvolytic reactions are accelerated by increases in pressure. l7 Consequently, recent proposals6 that ultrasonics provides evidence for disruption of solvent structure around a reacting solute in low concentration, and 'new insights' into the mechanisms of direct displacement reactions at carbon seem highly questionable. The third mechanistic proposal is based on deviations from first-order kinetics observed beyond 70% reaction for solvolyses of t-butyl chloride in 40% ethanol + water solution at 25 0C.6 We have observed similar results on numerous occasions, but reproduceable first-order rate constants, in agreement with the literature valueT.W. BENTLEY AND G. E. CARTER 1637 obtained titrimetrically,s can be obtained from the first 70% of reaction. t-Butyl chloride dissolves readily in 40% ethanol c water solution at 25 O C , so poor solubility does not appear to be the cause of these deviations. We have found that high-precision results from < 10% to > 99% reaction can be obtained if the vapour space above the solution is minimised; duplicate runs gave k = (1.45 0.01) x l 0-3 s-l with correlation coefficients > 0.99997 and a precision of fit to first-order kinetics better than 0.1 %.Consequently the reported deviations from first-order kinetics are not mechanistically significant. I I I 1 I 1 0 0 e 0 0 e 0 I I I I I I 1 I I 1 -5 - 4 - 3 - 2 log k (CH,), CCl FIG. 1.-Plot of logarithms of solvolysis rates of t-butyl bromide against t-butyl chloride for water (e), ethanol +water (0; slope: 0.94; correlation coefficient: 0.9999), methanol + water (a; slope 0.94; correlation coefficient: 0.9998), and acetone + water (0 ; slope 0.90; correlation coefficient: 0.9999). Kinetic data from tables 2 and 3, and ref. (8), (1 3) and (18). The kinetic results for t-butyl bromide (tables 2 and 3) complete three series of data: ethanol + water; methanol +water and acetone + water; kinetic data for the less aqueous media have been reported by l8 These results give good linear plots (fig.1) for logarithms of rate constants for t-butyl bromide against t-butyl chloride (or Y value8), with a small dispersion of lines for different solvent pairs. A consistent pattern of solvent effects on reactivity and mechanism has emerged from many studies of solvent mixtures using free energy plots,1a* 2, 9 3 lo, 1 2 ~ l3 e.g. fig. 1 using log k or AGI, and this treatment can be extended to solvents of widely varying ionizing power and nucleophilicity.la9 lo Despite the complexity of solvation effects these studies suggest that there is an underlying simplicity in the mechanism of aliphatic nucleophilic substitution,la with accumulating gas-phase data indicating S , 1 reactivity under solvent-free conditiom2 Unfortunately dissection of AGt for solution reactions into Am, AS$, AC; and even dAC$/dT provides a much more complex view.l2? 1 9 7 2o Such studies have been described as a ‘fetish’;21 they may represent unnecessary attention to detail, given the1638 SOLVOLYSIS OF t-BUTYL HALIDES present lack of understanding of the activation process in solution and the tendency for changes in Am and TAS to compensate each other.However, as there is considerable current interest in mechanistic proposal (4), we comment here on interpretations of ACi. This differs from the previous three proposals (above) in that the experimental rate data are well-established.l’* 22 For most, if not all solvolyses, a plot of log k against 1 / T is non-linear. Such curved Arrhenius plots may imply a dependence of activation energy with temperature. Of the various ways of treating the experimental data, the simplest is a direct plot of A S (obtained by differential methods22) against T.The results fit eqn (1) Am = AH,S+ TACb (1) where ACJ is the heat capacity of a c t i v a t i ~ n . l ~ * ~ ~ Alternatively a large amount of data has been fitted to the empirical eqn (2) log k = A/T+B log T+C (2) B = (ACb/R)+ 1 . (3) from which eqn (3) can be obtained using transition-state theory if dACb/dT = 0 2 0 Accurate kinetic data and alternative equations now indicate that AC: is temperature dependent,’? 22 as anticipated by Kohnstam19 who included a fourth term (D7) in eqn (2).This additional complexity has led to proposals for major revisions7’ 23 of previous interpretations (see later), when minor refinement may have been more readily justifiable. Until recently it was argued that ACL [from eqn (2) and (3)] probed solvent reorganization,ll* 1 9 9 2o rather than the mechanism of reaction of solute. This is consistent with the large variation in A d p on addition of cosolvent to 24 when only minor mechanistic changes would occur. Also A b p is almost constant for solvolyses of arenesulphonates of widely varying showing insensitivity to mechanism. Furthermore the parallels between A c p + and Winstein-Grunwald rn value show the relationship between solvent reorganization and the extent of charge dispersion or charge development in the transition These arguments were undermined by the observation of AC,f values that appeared to be too large for the equilibrium (required by transition-state theory) between covalent solute and partially ionic transition state.26 However, Jencks has recently suggested that some reactive intermediates in solution may not be solvent eq~ilibrated,~’ and this proposal could be extended to transition states.The suggested major revisions7* 23 in the above interpretation are based on an earlier ‘hypothetical’ proposal that AC$ is ‘spurious’, and is due entirely to mechanistic complexity-a two-stage mechanism in which formation of an intermediate [e.g. an ion pair-see eqn (4)Iz2 is only partially rate-limiting. The revised two-stage inter- pretation leads to a four-parameter equation, consistent with the experimental data and superior to eqn (2):227 23 (4) k, k, k-i RX $ [intermediate] -+ product.However, the fit of the data to the extended (four-parameter) version of eqn (2) is also good.22 Even if the general argument for eqn (4) is accepted, consideration of specific examples leads to further difficulties. Anomalous Arrhenius behaviour is observed for hydrolysis of ethyl for which an ion-pair intermediate is firmly excluded.’? 28 Also there is little support for the implied requirement7$ 22 that a t-butyl ion pair undergoes return in water, and we have recently argued against ion-pair returnT. W. BENTLEY AND G. E. CARTER 1639 even for solvolyses in very weakly nucleophilic alcohols. lo Therefore it seems unlikely that ion pairs are the intermediates that would be required by eqn (4).Alternatively the intermediate may be a preassociation complex,27 but this proposal would be difficult to verify. Further serious doubts about the two-stage mechanism arise from the derived values of kJk2 [eqn (4)]. Numerous studies (see above discussion fig. 1) have established that solvolysis rate constants vary uniformly with change in solvent composition but, for hydrolysis of t-butyl chloride, calculated values of k_Jkz show no regular trends with addition of organic cosolvents. We conclude that deductions about non-linear Arrhenius behaviour based on eqn (4) remain ‘hypothetical’ for solvolyses of aliphatic substrates. We thank the S.R.C. for financial support, and M. J. Blandamer, J.Burgess, R. H. Davies, R. E. Robertson and J. M. W. Scott for helpful comments. For recent reviews of the background to this work see: (a) T. W. Bentley and P. v. R. Schleyer, Adv. Phys. Org. Chem., 1977, 14, 1; (b) W. J. Albery, Annu. Rev. Phys. Chem., 1980, 31, 227. T. W. Bentley, C. T. Bowen, W. Parker and C. I. F. Watt, J. Am. Chem. SOC., 1979, 101, 2486. J. 0. Edwards, Inorganic Reaction Mechanisms (Benjamin, New York, 1964), p. 100. P. v. R. Schleyer and R. D. Nicholas, J. Am. Chem. Soc., 1961, 83, 2700. P. A. Adams, J. G. Sheppard and E. R. Swart, J. Chem. SOC., Chem. Commun., 1973, 663. J. P. Lorimer and T. J. Mason, J. Chem. SOC., Chem. Commun., 1980, 1135. Commun., 1981, 13; J. Chem. Soc., Faraday Trans. 1, 1981, 77, 1999. T. W. Bentley, C. T. Bowen, H.C. Brown and F. J. Chloupek, J. Org. Chem., 1981, 46, 38. ’ M. J. Blandamer, J. Burgess, P. P. Duce, R. E. Robertson and J. M. W. Scott, J. Chem. SOC., Chem. * A. H. Fainberg and S. Winstein, J. Am. Chem. Soc., 1956, 78, 2770. lo T. W. Bentley, C. T. Bowen, W. Parker and C. I. F. Watt, J. Chem. SOC., Perkin Trans. 2, 1980, 1244. l 1 E. A. Moelwyn-Hughes, R. E. Robertson and S. Sugamori, J. Chem. Soc., 1965, 1965. l 2 S. Winstein and A. H. Fainberg, J. Am. Chem. SOC., 1957, 79, 5937. l3 A. H. Fainberg and S. Winstein, J. Am. Chem. SOC., 1957, 79, 1602. l 4 E. A. Moelwyn-Hughes, J. Chem. SOC., 1962, 4301. l5 R. H. Stokes, J. Phys. Chem., 1961, 65, 1242. l6 D. S. Kristol, H. Klotz and R. C. Parker, Tetrahedron Left., 1981, 907, and references cited therein. l7 T. Asano and W. J. LeNoble, Chem. Rev., 1978, 78, 407. l9 G. Kohnstam, Adv. Phys. Org. Chem., 1967, 5, 121. 2o R. E. Robertson, Prog. Phys. Org. Chem., 1967, 4, 213. 21 M. J. S. Dewar, The Molecular Orbital Theory of Organic Chemistry (McGraw Hill, New York, 1969), 22 W. J. Albery and B. H. Robinson, Trans. Faraday SOC., 1969, 65, 980. 23 M. J. Blandamer, R. E. Robertson, P. D. Golding, J. M. MacNeil and J. M. W. Scott, J. Am. Chem. 24 K. M. Koshy, R. K. Mohanty and R. E. Robertson, Can. J. Chem., 1977, 55, 1314. 25 R. K. Mohanty and R. E. Robertson, Can. J. Chem., 1977, 55, 1319. 26 M. J. Blandamer, R. E. Robertson, J. M. W. Scott and A. Vrielink, J. Am. Chem. SOC., 1980, 102, 27 W. P. Jencks, Ace. Chem. Res., 1980, 13, 161. 28 M. H. Abraham, J. Chem. SOC., Perkin Trans. 2, 1973, 1893. E. Tommila, kf. Tiilikainen and A. Voipio, Ann. Acad. Scient. Fenn., Ser. A2, 1955, 65, 3. p. 28?. SOC., 1981, 103, 2415. 2585. (PAPER 1/1221)

ISSN:0300-9599    

DOI:10.1039/F19827801633

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1982,78, 1641-1665 Aqueous Solutions Containing Amino Acids and Peptides Part 13.-Enthalpy of Dilution and Osmotic Coefficients of some N-Acetyl Amino Acid Amides and some N-Acetyl Peptide Amides at 298.15 K BY G. MICHAEL BLACKBURN, TERENCE H. LILLEY* AND ELIZABETH WALMSLEY Chemistry Department, The University, Sheffield S3 7HF Received 17th August, 1981 The energetics of the interactions occurring between some N-acetyl amino acid amides and some N-acetyl peptide amides in aqueous solutions at 298.15 K have been investigated. Osmotic coefficients of solutions containing N-acetylglycinamide (G), N-acetyl-L-alaninamide (A) and N-acetyl-L-leucinamide (L) and equimolal solutions of G + A, G + L and A + L have been obtained using the isopiestic vapour pressure technique.Enthalpies of dilution of N-acetylglycylglycinamide (G,), N-acetyl-L-alanyl-L-alaninamide (A2), N-acetylglycylglycylglycinamide (G,), N-acetyl-L-alanyl-L-alanyl-L-alaninamide (A,) and N-acetyl- L-alanylglycinamide (AG) and equimolal solutions of G + G,, G + G, and A +A, were obtained using microcalorimetry. The results obtained were used to calculate the pairwise free energy and enthalpy parameters for like-like and like-unlike solute interactions. The effects of molecular structure and substitution on these parameters are considered and the efficiency of a group interaction approach is investigated. The group interaction idea works well for both the rather limited free energy data set considered and the more extensive enthalpy data set, with the exception of the most hydrophobic molecule A,.It is suggested that the results for A, indicate a degree of intramolecular folding which perturbs the intermolecular interactions. There is at present considerable general interest in the so-called ' non-bonding' interactions which occur between atoms and molecules. Much of this interest stems from an appreciation of the important role which such interactions play in biological The work here is a continuation of earlier work5 on the non-bonding interactions between some substituted peptides and amino acids in aqueous solutions and has been partly described in a preliminary communication.6 The systems investigated embody certain features which must contribute, in some measure at least, to the behaviour of oligopeptides and proteins in aqueous environments, and our hope is that such studies will give insight into the factors which affect the conformational stability of proteins and enzyme-substrate interactions.The available evidence indicates that globular protein and enzymes are stabilised in a relatively narrow distribution of conformations in aqueous s y ~ t e m s ~ - ~ by various intramolecular solute-solute interactions and by a range of solute-solvent interactions, and it would seem that these manifold interactions are individually energetically weak but collectively of great importance with regard to conformational stability. The nature and arrangement of the amino acid side-chain along the protein backbone are responsible for the individual characteristics of the macromolecule and it has been recognised for some time that all of the information pertaining to the protein is implicit in the amino acid s e q ~ e n c e .l ~ - ~ ~ Because of the difficulty of studying protein systems per se it is of particular benefit to investigate the properties of systems containing small molecules16 which incorporate 16411642 AQUEOUS SOLUTIONS CONTAINING AMINO ACIDS AND PEPTIDES functional groups present in proteins. Indeed, most of our, albeit sketchy, knowledge of the magnitudes of non-bonding interactions has been obtained from studies on well-defined small molecules.17 The objective of the present work is to study the energetics of the interactions occurring in solutions of amino acid and peptide derivatives which combine both biological relevance and structural simplicity.For the moment we have directed our attention to a study directed at substituted amino acids and peptides containing only glycine, L-alanine, L-valine and L-leucine. This particular set was chosen to demonstrate the effect of increasing side-chain hydrophobicity and is representative of the amino acids most commonly found in hydrophobic areas within proteins. EXPERIMENTAL The experimental procedures used for obtaining the heats of dilution5 and the osmotic coefficients18 have been described previously. PREPARATION AND PURIFICATION OF MATERIALS The preparation of the N-acetyl amino acid amides of glycine, L-alanine, L-leucine and L-valine has been described previ~usly.~ N - ACETY L - L - A L A N Y L - L - ALANIN A M I DE The ethyl ester was obtained as follows.N-Benzyloxycarbonyl-L-alanyl-L-alanine ethyl ester (3.2 g, 10 mmol dm-,) and p-toluenesulphonic acid monohydrate (1.9 g, 10 mmol drnT3) were dissolved in dimethyl formamide (DMF) (20 cm3). After purging with nitrogen, 10% Pd/C catalyst (0.5 g) was added and hydrogen bubbled through at atmospheric pressure overnight. The catalyst was filtered off and the solvent was lyophyllised to give an oil which was dissolved in dry pyridine at 0 OC and acetic anhydride (20 mmol drn-,, 5 cm3) added. After one hour stirring at 0 OC, excess solvent was lyophillised and the residue dissolved in ice/water. Excess Amberlite MB-3 ion-exchange resin was added and the solution stirred for 15 min at 0 OC. It was filtered, the solvent evaporated and the residue crystallised from ethyl acetate +petrol to give the product, m.p.126-127 "C (7473, [a]2,2 - 54.7 O (c 0.5, EtOH). (Found: C, 52.24; H, 7.76; N, 12.14; C,,H,,N,O, requires C, 52.16; H, 7.88; N, 12.16%.) The ester was dissolved in anhydrous ethanol saturated with ammonia and set aside for one day at 20 OC. Solvent was removed in U ~ C U O and the product repeatedly crystallised from ethanol + ether to constant melting point, 250-251 OC, [a]E-42O (c 1, MeOH). (Found: C, 47.81; H, 7.57; N, 20.82; C,H,,N,O, requires C, 47.75; H, 7.51; N, 20.88%.) d(D,O) 3.69-3.55 (2H, m, 2 aCH), 1.35 (3H, s, CH,CO), 0.73 (3H, d, J 5 Hz, CH,). N-ACETY LGLYCY LGLYCINAMIDE N-Benzyloxycarbonylglycylglycine ethyl ester (2.9 g, 10 mmol drn-,) and p-toluenesulphonic acid (1.9 g, 10 mmol dm-,) were dissolved in dry DMF (25 cm3) and the solution purged with nitrogen and hydrogenated in the presence of 10% Pd/C catalyst (0.5 g) to yield the dipeptide ester as its p-toluenesulphonic acid salt.Solvent was removed in U ~ C U O and the oily solid triturated with dry ether. Without further purification the residue was dissolved at 0°C in pyridine (60 cm3) and an excess (1.5 equiv.) of acetic anhydride added. The acetylation reaction was followed closely by t.1.c. and when judged to be complete the solvent was evaporated and the residue dissolved in ice/water. Excess Amberlite MB-3 ion-exchange resin was added. The solution was stirred for 15 min at 0 OC, then filtered, the solvent was evaporated and the residue crystallised from ethanol+ether, m.p.148-149 OC, (68%), [lit.19 m.p. 146-148 "C], R, = 0.39 in 9: 1 CHC1,:i-PrOH. (Found: C, 47.7; H, 6.98; N, 13.9; C,H,,O,N, requires C, 47.5; H, 6.98; N, 13.86%.) 6 (CDCl,) 6.95 br (lH, m, NH), 6.63 br (lH, m, NH), 4.21 (2H, q, J 7 Hz, OCH,), 4.02 (2H, d, J 5 Hz, aCH,), 3.98 (2H, d, J 6 Hz, aCH,), 2.03 (3H, s, CH,CO), 1.27 (3H, t, J 7 Hz, CH,). This ethyl ester was dissolved in anhydrous ethanol saturated with ammonia and the solution allowed to stand in a pressure vessel at 20 "C for one day. Solvent was removed in U ~ C U O andG. M. BLACKBURN, T. H. LILLEY A N D E. WALMSLEY 1643 the product repeatedly crystallised from ethanol +ether, m.p. 206-208 OC (82%). (Found: C, 41.8; H, 6.4; N, 24.3; C,H,,N,O, requires C, 41.6; H, 6.38; N, 24.26%.) N - A CETY LG LY CY LGLY CY LG LY c I N AMIDE N-Acetylglycylglycylglycine ethyl ester20 was dissolved in a large excess of anhydrous ethanol saturated with ammonia and the solution was allowed to stand at 20 OC for two days.Solvent was removed under reduced pressure and the product repeatedly crystallised from water+methanol, to constant melting point 253-255 OC. (Found: C, 41.66; H, 6.09; N, 24.40; C,H,,N,O, requires C, 41.74; H, 6.13; N, 24.34%.) G(D,O) 3.22 (2H, s, aCH,), 3.30 (2H, s, aCH,), 3.26 (2H, s, aCH,), 1.38 (3H, s, CH,). N - ACETY L - L - A LANY L - L - A L A N Y L - L - A LANIN AMIDE N-Benzyloxycarbonyl-L-alanyl-L-alanyl-L-alanine ethyl ester21 (3.93 g, 10 mmol drnp3) was hydrogenated in dry DMF (20 cm3) in the presence of p-toluenesulphonic acid (1.9 g, 10 mmol drn-,) and Pd/C catalyst (0.6 g, 5% w/w) using the procedure described above.The tripeptide salt was then dissolved in an ice-cold mixture of pyridine (10 cm-,) and excess acetic anhydride (2cm3). After stirring 1 h at OOC, solvent was evaporated and the residue dissolved in ice/water. Excess ion-excha,ige resin was added and the solution stirred at 0 OC for 15 min. After filtration, the residue was crystallised twice from ethyl acetate+ether to give the pure product, m.p. 246-246.5 OC (71 %), [a]g - 73.2 O (c 0.5, EtOH), R, = 0.33 in 9: 1 CHCl,: i-PrOH. (Found: C, 51.86; H, 7.58; N, 13.90; C,,H,,N,O, requires C, 51.82; H, 7.69; N, 13.94%.) This ester was dissolved in anhydrous ethanol saturated with ammonia and the solution allowed to stand at 20 OC for 36 h.Solvent was removed under reduced pressure and the product repeatedly crystallised from methanol ether to give the pure product, constant melting point 295-300°C (64%), [a]2,2--61.5 O, (c 0.75, MeOH). (Found: C, 48.54; H, 7.5; N, 20.65; Cl,H,oN,O, requires C, 48.52; H, 7.40; N, 20.57%.) G(D,O) 3.77-3.60 (3H, m, 3 aCH), 1.42 (3H, s, CH3CO), 0.83 (3H, d, J 6 Hz, CH,), 0.79 (3H, d, J 7 Hz, CH,), 0.76 (3H, d, J 7 Hz, CH,). N - A CE TY L - L - A LAN Y LG LY c I N A MID E N-Benzyloxycarbonyl-L-alanylglycine ethyl ester2, was dissolved in dry DMF containing one equivalent ofp-toluenesulphonic acid and hydrogenated overnight in the presence of 10 % Pd/C catalyst. The solution was filtered and solvent evaporated in U ~ C U O to give crude the dipeptide salt as an oil which without further purification was dissolved in ice-cold pyridine and acetic anhydride and stirred for 30 min at 0 "C.The solvent was evaporated and the residue dissolved in ice/water. Excess mixed-bed ion-exchange resin was added and the solution stirred at 0 OC for 15 min. After filtration the solvent was evaporated and the residue crystallised several times from ethyl acetate+petrol (60:80) to yield the product, m.p. 137-138 OC (63%). (Found: C, 50.36; H, 7.39; N, 12.99; C,Hl,N20, requires C, 49.99; H, 7.46; N, 12.96%.) G(CDC1,) 7.29 (lH, t, J 5 Hz, NHCH,), 6.77 (IH, d, J 7 Hz, NH-CH), 4.65-4.59 (lH, m, aCH), 4.19 (2H, q, J 7 Hz, OCH,),3.99 (2H, d, J 5 Hz, aCH,), 2.00 (3H, s, CH,CO), 1.38 (3H, d, J 7 Hz, side chain CH,), 1.26 (3H, t, J 7 Hz, CH,).N-Acetyl-L-alanylglycine ethyl ester was dissolved in anhydrous ethanol saturated with ammonia and allowed to stand at 20 "C for 1 day. Solvent was removed in U ~ C U O and the product repeatedly crystallised from ethanol +ether to constant melting point 148-149 OC. (Found: C, 44.83; H, 7.18; N, 22.43; C,H,,N,O, requires C, 44.90; H, 7.00; N, 22.45%.)6(D20)3.81 (1 H, q, J7Hz, CH), 3.42(2H, s, CH,), 1.54(3H, s, CH,CO), 0.88 (3H, d, J 7 Hz, CH,). RESULTS The excess Gibbs free energy per kilogram of solvent may be representedl8 as a power series in solute molalities1644 AQUEOUS SOLUTIONS CONTAINING AMINO A C I D S AND PEPTIDES where the coefficients g,,, gijk etc. are taken to represent interactions between the subscripted species. Using the relationship^^^ and Gkx = [a(GeX/m)/am-l],, GgX = RTm(1-4) we obtain from eqn (1) the following expression for the osmotic coefficient If we consider the application of eqn (4) to single-solute systems containing the solutes A and B, respectively, we obtain: d A = + (gAA mA +gAAA mi + ' * * )IRT ( 5 ) 4 B = l + ( g , , m B + g B , B m ~ + .. . ) / R T . (6) TABLE 1 .-ISOPIESTIC MOLALITIES FOR THE SYSTEMS INVESTIGATED m(solute)/mol kg-' m(urea) - /mol kg-' 4(urea) Ga A L G+Ab G+Lb A+Lb 0.5400 0.5522 0.5769 0.6466 0.6987 1.0986 1.1436 1.1863 1.227 1 1.2892 0.978 1 0.9777 0.9768 0.9742 0.9723 0.9588 0.8577 0.9559 0.9547 0.9528 0.5416 0.5512 0.5746 0.6422 0.6957 1.0950 1.1425 1.1782 1.2214 1.278 1 0.5441 0.5556 0.5798 0.6479 0.7023 1.1045 1.1513 1.1875 1.2803 1.2917 0.6176 0.6348 0.668 1 0.7610 0.8356 1.4562 1.555 1 1.6691 1.7652 1.8936 0.546 1 0.5800 0.5786 0.6480 0.7004 1.0959 1.1418 1.1788 1.2228 1.2870 0.5563 0.575 1 0.6047 0.6809 0.7388 1.1898 1.2266 1.2884 1.3402 1.4109 0.5631 0.5812 0.61 13 0.6882 0.7467 1.2041 1.2609 1.3101 1.3635 1.4320 a Abbreviations used: G = N-acetylglycinamide, A = N-acetyl-L-alaninamide, L = N- acetyl-L-leucinamide.The binary solute systems were equimolal within O. 1 per cent. The corresponding expression for the osmotic coefficient for a binary solute system containing the solutes A and B in equimolar amounts bA = yB = 0.5) is 4 A B = +[(gAA+2gA,+gB,)m/4+(g,AA+3g,A,+3gA,*+gB,B)m2/4+ * a (7) (8) All three eqn (5)-(7) can be written in the form 4 = 1 +(g,m+g,m2+. . . ) / R T where g , now embraces all pairwise interactions and g3 all triplet interactions.table 1. The osmotic coefficients of the various solutions were obtained from The experimental data obtained from the isopiestic experiments are presented in # = [m(urea) $(urea)/m] (9) the osmotic coefficients for urea being obtained from Ellerton and Dunlop's very precise measurements. 24 Each of the data sets (4, m) was fitted to a polynomial of the above form [eqn (8)] and the appropriate coefficients were determined by a least-squares procedure.G . M. BLACKBURN, T. H. LILLEY AND E. WALMSLEY 1645 In the program used, if the 95 % confidence limit of a particular coefficient was found to be more than 100% the data were reanalysed with that coefficient excluded from the fit. This procedure continued until the coefficients remaining all had errors less than 100% within a 95% confidence limit.Thus any coefficient excluded had an effectively zero value. Hence, only the minimum number of parameters was retained. The g , and higher-order interaction parameters yielded by linear regression for the systems studied are given in table 2. TABLE 2.-EXCESS FREE ENERGY PARAMETERS FOR THE PEPTIDE SYSTEMS INVESTIGATED solute g2 g3 g4 A B /J kg m o P /J kg2 m ~ l - ~ /J kg3 m ~ l - ~ 1030b - - G G -82.9 (4.2)a 2.2 A A - 144.2 (1 1.7) +38.5 (10.5) -_ G A - 149.5 (36.8) +53.5 (33.5) - G L -312.4 (35.4) +63.5 (29.1) - A L - 358.8 (1 7.4) + 78.9 (14.0) - 1.3 4.2 4.1 2.0 L L -731.9 (62.9) +279.5 (106.8) -73.1 (41.4) 3.8 a The number in parentheses represents the 95% confidence range of the coefficient.o is the standard error of the least-squares fit. TABLE 3 .-PAIRWISE FREE ENERGY COEFFICIENTS FOR THE PEPTIDE SYSTEMS INVESTIGATED solute gAB A B /J kg mo1P G G -82.9 (4.2)" G A -185.5 (81.5) G L -217.3 (103.9) A A -144.2 (11.7) A L -279.6 (71.8) L L -731.9 (63.0) a Bracketed term represents the 95% confidence limit. The expansions all converge rapidly in the sense that lg,l > lg31 > 1g41; this in itself implies that the systems are not strongly interacting. Indeed for most systems the first two terms describe the behaviour quite adequately. Only the solution containing N-acetyl-L-leucyl amide requires a quartet term. Comparing eqn ( 5 ) and (6) and (8) for a single-solute system, the g2 term corresponds exactly to the pairwise interaction coefficient for like species.For binary-solute systems, the g, term embraces like and unlike solute pair interactions. These may be separated using eqn (7) in the form and since we have values for the like terms g A A and gBB, the cross-interaction term gAB can also be evaluated. The pairwise like and unlike coefficients are presented in table 3. The corresponding1646 AQUEOUS SOLUTIONS CONTAINING AMINO ACIDS AND PEPTIDES TABLE 4.-EXPERIMENTAL ENTHALPIES OF DILUTION AT 298.15 K m/mol kg-I n/IOp3 mol m’/mol kg-’ q/10-’ J A/lO-* J N-acetyl-L-alaninamide + N-acetyl-L-alany I-L-alaninamide 0.5077 0.8763 0.2538 - 1.2571 0 0.5077 0.1457 0.3752 - 1.0527 + 10 0.5077 1.6196 0.3158 - 1.7363 -6 0.1765 0.9076 0.0892 -0.1743 + 3 0.1765 0.1536 0.0437 -0.1 180 +6 0.1765 0.1341 0.0304 -0.1198 0 N-acetylglycylglycinamide + N-ace t ylglycylgl ycylgl ycinamide 0.6057 0.6057 0.6057 0.6057 0.6057 0.2685 0.2685 0.2685 0.3221 0.3221 0.3221 0.3221 0.1730 0.1730 0.1730 0.1730 0.4686 0.4686 0.4686 0.4686 0.2404 0.2404 0.2404 0.2404 0.2404 0.2404 0.4542 0.4542 0.4542 0.4542 0.4542 0.2191 0.2191 0.2191 0.2191 1.1569 0.5105 2.2350 I .0297 0.648 1 0.5048 0.2629 0.9263 0.2889 0.1066 0.47 18 0.1809 0.1569 0.1316 0.0664 0.2094 1.1547 0.8462 0.9055 1.3736 0.0540 0.2349 0.1971 0.1732 N-ace t yl-L-alanyl-L-alaninamide 1.1685 0.6054 0.9177 0.2691 0.3425 0.5553 0.1765 0.2457 0.2576 0.1523 0.2168 0.0659 0.0863 0.1275 0.0463 0.0541 -0.7072 - 0.9750 - 0.9025 - 0.6482 - 0.2759 -0.2333 -0.2241 - 0.2730 N-acet yl-L-alan y lglycinamide 0.9067 0.2252 -0.6348 1.6626 0.3619 - 0.5 162 0.4272 0.1057 - 0.4244 0.8548 0.1469 -0.7807 0.4820 0.1171 - 0.1808 0.2232 0.0452 - 0.121 8 0.8842 0.1829 -0.1305 0.2546 0.0601 -0.1225 0.4765 0.1 128 -0.1614 0.2758 0.0625 - 0.1308 N-acetylglycylglycinamide 0.8716 0.2142 0.4323 0.1070 0.423 1 0.0845 1.7096 0.3502 1.1078 0.3040 0.4275 0.1071 0.2351 0.0554 0.841 1 0.1733 0.1645 0.0357 1.0961 0.8280 0.8664 0.9009 0.8545 0.29 18 0.2181 0.2248 0.1743 - 25 + 26 -61 - 35 + 12 + 12 + 26 -6 0 -6 +4 0 + 3 +3 + 2 +2 -7 - 12 + 16 + 1 - 12 +2 + 14 +7 0 +8 - 10 +6 +4 +3 + 1 + 10 - 12 +2 -7G.M. BLACKBURN, T. H. LILLEY A N D E. WALMSLEY 1647 TABLE 4.-(continued) 0.1109 0.1109 0.1109 0.0558 0.0558 0.0558 0.0558 0.1145 0.1145 0.1446 0.1446 0.1446 0.0760 0.0760 0.0760 0.0760 0.0304 0.0304 0.0304 0.0304 0.0304 0.0304 m/mol kg-l n/ lop3 mol m'/mol kg-l q / 10-l J A/ 10- J N-acet y lglycinamide + N-acetylgl ycy lgl ycylglycinamide 0.2048 0.0544 0.0794 0.4284 0.0862 0.0764 0.2782 0.0783 0.0642 0.1031 0.0264 0.0202 0.1565 0.0402 0.0243 0.0507 0.0107 0.0634 0.0368 0.0084 0.0870 N-ace tylglyc ylglycy lgl ycinamide 0.1152 0.0287 0.1520 0.2061 0.0371 0.2336 0.2545 0.0459 0.3614 0.4205 0.1096 0.2191 0.1540 0.0493 0.2428 0.0790 0.0160 0.0772 0.0806 0.020 i 0.0666 0.0425 0.0097 0.0444 0.1360 0.0240 0.1 174 N-ace tyl-L-alany 1-L-alanyl-L-alaninamide 0.0562 0.0147 -0.0173 0.0535 0.0093 - 0.0224 0.01 18 0.0204 - 0.0191 0.0328 0.0066 -0.0207 0.1125 0.0239 - 0.0087 0.03 19 0.0078 - 0.0 185 -2 +2 + 1 - 1 +7 -9 -3 +4 -6 - 15 -1 + 23 +7 -1 +2 + 11 0 + 3 -1 - 2 + 1 -1 expression to eqn (1) for the excess enthalpy per kilogram of solvent is He" = h i j m , m j + ~ x Z hijkmimjmk-k... (1 1) i j i j k where the coefficients h, etc. are the appropriate enthalpy parameters for interaction between the subscripted species. These are related to the free energy parameters by, (12) for example, If we consider a solution containing one solute species, A or B, then the expressions for the excess enthalpies are hij = [ a ( g i j / V / a ~ l l p * HZ" = mi(hAA + hAAA mA -k . . .) (13) Hk" = m&(hBB + hBBB + . . .). (14) The expression for the excess enthalpy of a solution containing solutes A and B in equimolar amounts is &% = m2[(hAA + 2 h ~ ~ + h ~ ~ ) / 4 + (A,,, + ~ ~ A A B + ~ ~ A B B + h ~ ~ ~ ) m / 4 + . . .I. (1 5 ) Consequently as with the osmotic coefficients we can write a generalised expression to include eqn (1 3)-( 15) uiz.Hex = m2(h2+mh,+. . .) (16)1648 AQUEOUS SOLUTIONS CONTAINING AMINO ACIDS AND PEPTIDES where the h, and h, terms represent all pairwise and all triplet interactions in the system under consideration. The enthalpy coefficients were determined in the present investigation using dilution experiments. The experimental heat change, q, associated with a dilution is given by q = n(m’-m) [h,+(m’+m) h,+ . . .I. (17) The calorimetric data systems for the various systems (see table 4) were fitted to the above equation. The program used was a simple modification of that used to analyse the free energy data. TABLE 5.-ENTHALPY OF INTERACTION PARAMETERS’ soluteb h2 h3 103 a d A B /J kgmol-2 /J kg2 m0F3 /J mol-l - 646.0 (26.6)‘ 175.4 (39.1) 0.8 0.3 1.2 4880.5 (1767.6) -65.902 (3210) 0.2 5.5 0.4 0.7 1 .o 939.5 (4.5) - -321.8 (29.1) - -701.7 (16.6) - 622.7 (50.7) - 284.3 (5.0) - - - 1499.1 (65.7) a In the binary solute systems, the solutes were equimolal within 0.1 %.Abbreviations: G2 = N-acetylglycylglycinamide ; AG = N-acetyl-L-alanylglycinamide ; A, = N-acetyl-L-alanyl- L-alaninamide ; G3 = N-acetylglycylglycylglycinamide ; A, = N-acetyl-L-alanyl-L-alanyl-L- alaninamide. a is the standard error of the least-squares fit. Bracketed term represents the 95% confidence limit. Values for the parameters h, and h, determined from the least-squares analyses are given in table 5. The systems seem to be well described by rapidly converging series and lh,l > lh,l for most systems.Frequently only the leading term is necessary to describe the concentration dependence. As can be seen from table 5 the coefficients vary considerably in both sign and magnitude. A notable exception to this general pattern of behaviour is A,. For this solute the pairwise interaction coefficient is very large and lh31 > Ih21. Comparison of these values with those for G, where the solute is approximately the same size suggests that the self-interaction behaviour of A, is qualitatively different. This will be commented on later. The pairwise enthalpy interaction parameters are tabulated in table 6. For a single-solute system, h, corresponds to the pairwise interaction constant whereas for a binary-solute system with solutes present in equimolal quantities we have [see eqn (1 5 ) and (1 6)] Using the relationship gAEi = hAB- TsAB [see eqn (1 2)] the entropy parameters for the solutions containing amino acid amides were determined.These values, together with the corresponding free energy and enthalpy parameters, are given in table 7.1649 G. M. BLACKBURN, T. H. LILLEY A N D E. WALMSLEY TABLE 6.-PAIRWISE ENTHALPY COEFFICIENTS FOR THE PEPTIDE SYSTEMS INVESTIGATED solute h,, A B /J kg mol-2 G2 G2 A2 A2 G2 G3 A2 A3 G2 G G3 G A2 A AG AG - 646.0 (26.6)a 9 3 9 . 5 (4.5) - 1499.1 (65.7) - 2 1 0 . 6 (16.5) - 543.7 (70.3) 4880 (1768) 6 4 1 . 4 ( 1 10.5) 284.3 (5.0) a The parenthetical figure denotes the 95% confidence limit. The terms for the binary solute were calculated from the 95 % confidence limits in h, (see text) and the component single-solute terms.TABLE 7.-FREE ENERGY, ENTHALPY AND ENTROPY PAIRWISE INTERACTION COEFFICIENTS solute g A B h.4, %I3 A B /J kgmol-2 /J kgmol-2 /J kgrnoF2 G G - 83 - 220 - 137 G A - 186 86 272 G L -217 547 764 A A - 1 4 4 269 41 3 A L - 280 899 1 1 7 9 L L - 732 1714 2446 DISCUSSION The excess thermodynamic functions have been analysed using a molality expansion to yield appropriate solute pairwise interaction coefficients. Although these coefficients are thermodynamic quantities and as such relate to properties of the solutions, the relationship between them and molecular events is complicated. A link between the free energy coefficients and solute interactions in solution can be made in an exact way using the M~Millan-Mayer~~ theory of solutions. It is possible to t r a n ~ p o s e ~ ~ ? ~ ~ suchexperimental interactioncoefficients to the McMillan-Mayer state via a knowledge of the partial molar volumes of the solutes and the isothermal compressibility of the pure solvent.Whilst the latter is known the partial molar volumes of the solutes investigated are not. The transposition of the enthalpy coefficients to the McMillan- Mayer state is even more demanding of primary experimental data. In particular the temperature dependence of the solute partial molar volumes is required. Consequently, given the lack of volumetric data for the present systems it has not been possible to pursue deconvolutions like those previously carried out and we have taken recourse to more qualitative and semi-empirical interpretations of the results.We begin with a qualitative survey of the results obtained.1650 AQUEOUS SOLUTIONS CONTAINING AMINO ACIDS A N D PEPTIDES h - - I 1 I 1 I -500 I AMINO ACID AMIDES We now have available (see table 7) the pairwise interaction coefficients for free energy, enthalpy and entropy for the amino acid derivatives G, A and L. The negative value of the pairwise excess free energy parameters indicates that there is positive association or net attraction between these solutes in aqueous solution. The amino acid derivatives in this set all share the same general formula in which the ‘backbone’ of the molecule is constant, whereas the substituent on the a-carbon varies from one amino acid to another: R I CH,-CONH-CH-CONH, R = H, CH,, CH,CH(CH,), I I I 1 1 1 The pairwise interaction coefficients become increasingly more negative as the molecular size increases.This observed increase would therefore seem to be a result of greater association of the molecules through their side chains. The g A B values are the result of large but opposing h A B and sAB values. This compensatory effect is a widely-observed phenomenon28 and the specific linear relationship between entropy and enthalpy change which is observed in a variety of processes involving small molecules in aqueous solution has become known as Lumry’s law. This linear relationship is shown in fig. 1 for the present systems. In fig. 2 g,, is plotted against hAB; this plot is qualitatively useful in that it reveals an approximately linear relationship between the pairwise free energy and enthalpy parameters for the systems studied.Although empirical, this could be useful in a predictive sense to evaluate order of magnitude values for the g A B terms from a knowledge of the corresponding h A B terms. As observed in this study the enthalpy terms are more easily determined experimentally than the corresponding free energy terms. From such a limited data set, however, it is not possible to judge fully the usefulness of this approach.G . M. BLACKBURN, T. H. LILLEY A N D E. WALMSLEY 1651 For molecular pair interactions which rely on solute hydrogen bonding or strong dipolar effects the enthalpy 30 should be negative. From table 7 we see that the only amino acid derivative from the set for which this is so is N-acetylglycinamide.The main functional groups of this molecule are the two amide residues in the backbone and unlike the other amino acids there is no alkyl substituent on the a-carbon atom. The molecule is predominantly polar and we can surmise from its solution - 500 0 200 400 6 00 1 expt gAB /J kg mol-2 10 FIG. 2.-Plot of the experimental pairwise free energy coefficient (giTt) against h i y t for the N-acetyl amino acid amides. characteristics that the weak association of the solute in water is primarily through dipole interactions between the amide groups. The association of G is accompanied by an exothermic heat change. In general, the nature of the enthalpic change must depend on the relative enthalpic favourability of solute-solute and solvent-solvent interactions over the corresponding solute-solvent interactions.Since both water and acetylglycinamide have hydrogen-bonding characteristics the heat change can be rationalised30 by assuming that breaking amide-water hydrogen bonds and forming amide-amide and water-water hydrogen bonds is a net exothermic process. Spectroscopic studies on model amides3l9 32 in various solvents indicate that the strength of the intermolecular -C=O - - - H-N- bond varies inversely with the hydrogen-bonding potential of the solvent. Model compound studies, however, have been less than unanimous in their evaluation of the amide-amide interaction in aqueous solution, although it is generally agreed that the bond strength is weak.1652 AQUEOUS SOLUTIONS CONTAINING AMINO ACIDS AND PEPTIDES estimate^^^.^^ of the enthalpy of rupture of a peptide-peptide hydrogen bond range from 0 to 8 kJ mol-l.If we assume that the association of G is brought about entirely by intermolecular hydrogen bonding between amide groups and that the simple equilibrium is set up: G+G 2 G - * *G then it can be that hG-G = AH$?, KG-G (19) gG-G = -RTKG-G (20) and AGO = - RT In KG-G (21) where AHgG is the standard enthalpy of association, KGG is the pairwise association constant of G with itself and AGg-G is the standard free energy of association. Combining the above equations we get and substituting values from table 7 gives AHg-, = -6.6 kJ mol-l AGgG = 8.4 kg mol-l. KG-G = 0.033 kg mol-1 TABLE 8.-THERMODYNAMIC ASSOCIATION CONSTANTS AND ASSOCIATED PARAMETERS FOR THE SYSTEMS INVESTIGATED solute K A G e AH0 Ah93 A B /mol-l kg /kJ mol-l /kJ mol-1 /J K-l mol-l G G 0.033 8.4 - 6.6 - 50 G A 0.075 6.4 1.2 - 1 8 G L 0.088 6.0 6.3 1 A A 0.058 7.1 4.6 - 8 A L 0 .1 1 3 5.4 8:O +9 L L 0.295 3.0 5.8 +9 Assuming further that on average four amide-amide hydrogen bonds are formed per dimer leads us to an estimate of - 1.6 kJ mol-1 for the enthalpy of formation of an amide hydrogen bond. This value would clearly increase if less than four hydrogen bonds were formed and, although approximate, it is compatible with earlier estimates. Without justification we have completely ignored the possibility that methyl and methylene group effects are involved in the interaction; clearly if these effects are significant the estimate would need further adjustment.Similar calculations can be performed for the other systems investigated and the results of such calculations are given in table 8. Note that both dipole-dipole and hydrogen-bonding mechanisms are contributing in some way to the heat effects even though the enthalpy is dominated by the heat associated with the breaking of hydrophobic interactions. If we can assume that the contribution from polar effects is similar for all solutes in the peptide series, i.e.G. M. B L A C K B U R N , T. H. LILLEY A N D E. WALMSLEY 1653 equivalent to AHEG, it allows us to assign values to the alkyl group contributions by subtracting the polar (i.e. AHg,) contribution from the above values. Contributions from the methyl and isopropyl groups to the standard enthalpy of association of the peptides are thus evaluated as AHgeMe = + 1 1 .1 kJ mol-l AHg-pr = + 12.3 kJ mol-l. Although this type of group contribution approach is frequently adopted, especially with low molecular weight species,33 the accuracy of the derived group values is difficult to assess. Several questionable assumptions have been made, in particular that the amide group contribution is invariable. Studies3'+ 38 on the affinities of salts to various alkyl-substituted amides and on the interaction of salts with a-amino acids indicate that the salt-dipole interaction is modulated by the alkyl groups adjacent to the amide dipole. Thus while methyl or, in general, apolar groups do not participate directly in the dipole interaction they are able to exert an influence through their hydration spheres.Part of the trend observed here might therefore be due to the varying influence of the hydrocarbon side chains on the amide interactions. It would therefore seem inappropriate to place much emphasis on the numerical values obtained from thi; analysis. Of particular note from the present set of results is the significant difference in behaviour exhibited by G and A. These species differ by the addition of one methyl group to the backbone structure. The effect on the entropy of association is large and the hydrophobic effect of the methyl group rapidly outweighs the effect of the two polar amide groups. One feature of hydrophobic interactions which is of particular relevance to its biological role is cooperativity.This is also very marked in aggregation processes such as micelli~ation~~ but can be shown to contribute to the interactions of small alkyl derivatives such as alcohols and carboxylic 40 The triplet interactions included in table 2 demonstrate the cooperative element in the free energy of the peptide interaction. Another way of highlighting this cooperative effect is to compare contributions of pair and triplet effects at constant molality as a homologous series is ascended. A comparison of this sort is made in table 9. TABLE g.-COMPARISON OF PAIR AND TRIPLET CONTRIBUTIONS TO THE TOTAL SOLUTE INTERACTION AT 1 mol kg-l G A L lOOg,/g, 0 26 38 D I P E P T I D E AMIDES The peptide amides used in this study have the general formula R R I I CH3-CONH-CH-CONH-CH-CONH, where R represents either a methyl group or a hydrogen atom.Because of solubility limitations only enthalpy measurements were made. Comparison of the pair enthalpy1654 AQUEOUS SOLUTIONS CONTAINING AMINO ACIDS AND PEPTIDES interaction coefficients for the dipeptides with the results from the corresponding monomer is shown in fig. 3. The trend in the coefficients is more pronounced than in the corresponding monomer series but follows the same general pattern. G, contains three amide groups in the backbone, which should result in the molecule being more hydrophilic than G. This is indeed reflected in the negative enthalpy 1000/ average R.M.M. of interacting species/kg mol-I corresponding di-amino acid amides (a). FIG. 3.-Plot of against relative molar mass for the N-acetyl mono-amino acid amides (0) and the coefficient.If, however, the amide contribution was strictly additive the enthalpy of the dipeptide interaction would be 1.5 times that of the monomer; the experimental enthalpy term is actually nearer three times that of the monomer. The pairwise enthalpy coefficient for the alanine dipeptide has a large positive value. While free energy data are not available it is reasonable to assume that the dipeptide is more strongly associated than the monomer and that hydrophobic interactions make large contributions to this. It would seem that the effect of additional methyl group far outweighs the effect of the additional amide group. The increased interaction over that of the monomer is larger than one would have predicted from the size alone.The self-interaction of AG as expected from its constitution is intermediate in value between the interaction values for A, and G,. Thus while the pairwise enthalpy coefficients for the dipeptides are consistent with the corresponding values for the amino acid derivatives they could not have been extrapolated correctly from the monomer results.G. M. BLACKBURN, T. H. LILLEY A N D E. WALMSLEY 1655 TRENDS I N ENTHALPY COEFFICIENTS By discussing the amino acid and peptide derivatives separately we were able to compare molecules with the same backbone structure differing only in side-chain functionality. If we now compare the pairwise enthalpy coefficients for the series amino acid, dipeptide, tripeptide, the combined effects on the solute interaction of increasing chain length and side-chain functionality may be considered.Within these series of model compounds the structures have been varied system- atically in an attempt to correlate the thermodynamic values obtained with the structural details of the molecules. These correlations are purely empirical and must be discussed with caution. In the series G, G, and G, each glycine residue contributes the unit [NH-CH,-CO] to the molecule. Throughout the series the pairwise enthalpy coefficients are negative and become increasingly so as the relative molar mass increases. The data are consistent with strong dipolar association and we have surmised (see above) that the attraction is essentially through the amide groups. The change in enthalpy is not, however, linear with the number of amide groups.Thus while the polar interaction is clearly important the total molecular interaction must be considered to be a more complex process. In the series A, A, and A, each alanine contributes the unit [NH-CH(CH,)-CO] to the molecules. The effect of replacing the hydrogen atom on the a-carbon by a methyl group has a most striking effect. Contrasting the solution behaviour of the glycine and alanine peptides highlights the importance of the side-chain on peptide association. The enthalpy coefficients all take positive values and these increase rapidly with increasing chain length. As in the glycine series, the trend is not monotonic and as the ratio of non-polar to polar groups is increased the solution behaviour of the solutes changes to that of a ‘typically non-aqueous’ The hydrophobic effect of the methyl groups seems to rapidly outweigh the hydrophilic effect of the backbone amide groups, i.e.the numerical values of the pairwise enthalpy terms increase more rapidly in the alanine than in the glycine series. This difference in behaviour is most noticeable if we compare G, with A,. Both solutes are similar in that they are approximately the same size and have the same backbone structure. While the pairwise enthalpy coefficient for A, is large and positive the triplet term is numerically even larger. If the free energy term is proportionately large then the interaction between these solutes is very strong and would indicate cooperative behaviour which is consistent with the relatively dominant hydrophobic character of the species.It is apparent that the aqueous solution behaviour is determined by several features of the solute molecule, and as we anticipated the excess enthalpy of interaction is not a simple function of amide and hydrocarbon contributions or chain length. A complete description of the interaction must take into account the optimisation of all inter- molecular effects, but without resort to a more sophisticated procedure this is not possible. G R O U P ADDITIVITY The description of molecular interactions in terms of group interactions is poten- tially a useful approach. Whereas there is a vast number of chemical compounds of interest the number of functional groups which constitute these compounds is much smaller.If we presume that the physical properties of such compounds can be described in terms of the sum of contributions made by the molecules’ functional groups, we obtain a method for correlating the properties of a large number of systems using a much smaller number of parameters. Each parameter would characterise some1656 AQUEOUS SOLUTIONS CONTAINING AMINO ACIDS A N D PEPTIDES property of an individual group. Any group contribution method is necessarily approximate since the contribution of a given group in one molecule is not necessarily the same in another molecule. It is similar in this regard to the concept of bond energies. In order to implement a group contribution approach several assumptions must be made. The most fundamental of these is additivity: the contribution made by one group is assumed to be independent of that made by all other groups on the same molecule.Intermolecular forces acting on a group or whole molecule are thus determined by the average group composition of the system, i.e. are independent of how the groups are arranged in the molecule. This treatment is partially justified in that for non-electrolytic molecules, intermolecular forces are short range. The effect of distant groups in a molecule will be small unless they are brought together by conformational effects. However, nearest neighbour and steric effects are usually not insignificant and an approach which ignores molecular shape must remain approximate. Moreover, the local group composition adjacent to a particular molecule.will not necessarily be equal to the average group composition of the whole solution. Random arrangement of the molecules is only strictly true for ideal solutions. Group additivity would thus seem to operate best within a group of molecules which are of related structure. There is a growing collection of experimental data available for hydrophobic, hydrophilic and ‘mixed’ solutes in water and many attempts have been made at estimating additive contributions to thermodynamic quantities of certain functional groups. Schrier and S ~ h r i e r ~ ~ have used a simple additivity scheme to describe the salting out behaviour of amides; the non-polar and polar residues on the amides were assumed to make independent contributions to the overall salting-out effect on the molecule.An alternative a p p r o a ~ h ~ ~ - ~ ~ using a ‘solution of groups’ model has been developed for estimating equilibrium properties of non-electrolyte mixtures. Functional groups are assumed to interact in solution in proportion to their mole fractions and are further assumed to be completely independent of how the groups are bonded together. Using similar concepts to these Wood and Savage46 developed a group additivity scheme which they applied to pair interactions between non-electrolytes in dilute aqueous solution. The approach was successfully applied to the description of the solution behaviour of over sixty solute species, including several amides. In view of this success, we felt it worthwhile to see how well such a procedure would represent the present results.The group additivity approach assumes that when two solute molecules interact all groups on molecule A interact with all groups on molecule B and that each of these group interactions has a characteristic effect on the free energy or enthalpy, which is independent of the positions of the functional groups. The total pairwise interaction is then the sum of all the various group interactions that are present, e.g. for the enthalpy (23) h,, = Z nfn? Hij i, i where np is the number of groups of type i on molecule A, n? is the number of groups of type j on molecule B and Hij is the characteristic contribution to the enthalpy of one i group interacting with one j group. The summation is taken over all groups i on molecule A and all groups j on molecule (B).An analogous equation can be used for the free energy term. If the magnitude of the interactions is small compared with the thermal energy of the molecules then all of the functional groups on one moleculeG . M. BLACKBURN, T. H. LILLEY A N D E. WALMSLEY 1657 are able to interact with each of the groups on the other molecule. Neglecting steric and neighbouring group effects, the total free energy of interaction would consequently be group additive. In applying this type of model it is necessary to assign a set of functional groups which when added together make up the set of molecules in the data set. The choice is arbitrary and the accuracy of any correlation improves with increasing distinction of groups. In the limit as more and more distinctions are made we recover the molecule itself, and in that event the advantage of the group additivity method is lost.Thus for practical purposes a compromise must be reached. The number of distinct groups must remain small but not so small as to neglect significant effects in molecular structure which affect the physical properties. we considered that the peptide and amino acid derivatives studied here could be adequately described by two functional groups : the hydrocarbon and the peptide group. The hydrocarbon function is based on the methylene group and as before it is further assumed that one CH, is equivalent to 1.5 CH, groups and that a CH group is equivalent to 0.5 CH, groups. By ‘peptide group’ we refer to the units -CONH or -CONH,, i.e. no distinction is made between the peptide group proper and an unsubstituted amide group.These two primary groups give rise to three types of group interactions representing peptide-peptide, hydrocarbon-hydrocarbon and peptide- hydrocarbon contributions. Using the terminology more usually applied to describe peptide interactions, these correspond at least approximately to backbone-backbone, side-chain-side-chain and backbone-side-chain interactions. In terms of the above group definitions, the pairwise enthalpy coefficients for the interaction between molecular species A and B can be expressed as Following the previous where, for example, ngH2 represents the number of CH, groups in species A and HPep-Pep is a term representing the enthalpy of interaction of one peptide group with another.If an expression of the above type is used to fit a set of pairwise enthalpy data the group interactions can be obtained. Initially5 the group enthalpy interaction parameters generated by Savage and Wood were used to calculate the molecular interaction parameters, hAB, for the present systems. The calculated values agree in sign and approximate magnitude with the experiment ally determined en t halpies. However, notwithstanding the generally good agreement, the experimental results are usually greater than the calculated values and in particular the calculated coefficients for the solute A, and A, were considerably different from the experimental values. In a later paper, Wood36 applied the same additivity scheme to evaluate excess free energy group parameters. In adapting the approach to free energy data it is necessary to choose a concentration scale with which to define an ideal solution.On the molal scale M , RT and the mole fraction scale g A B = -_+En? n?Gii. i j Wood found that there was a better fit between experimental and calculated values if the mole fraction scale was used. Using his group interaction terms and eqn (26) 54 FAR 11658 AQUEOUS SOLUTIONS CONTAINING AMINO ACIDS AND PEPTIDES the interaction terms were calculated for the various solutes and are compared with experimentally determined pair free energy terms in fig. 4. The predicted values for the free energy terms do not agree will with the experimentally determined values. Although thecorrect sign is predicted the magnitudes of the values are not at all well reproduced.1500 - 1000 - 0 500 1000 1E 0 FIG. 4.-Plot of gygt against giy. The calculated values were obtained using the earlier36 parameters for group interactions. The breakdown in group additivity for the free energy parameters could be a result of incorrect assumptions inherent in the model but it is also probable that the group interaction parameters are in error and need reassessment. In particular doubt are the interactions involving the peptide groups since in the original data set no molecule contained more than one such group. In a group additivity approach it is assumed that the physical properties of a group may be transferred from one molecule environment to another and suffer relatively little change. This assumption is most likely to be valid if the group environments are similar in both molecules.Savage and Wood46 used a very diverse set of non-electrolyte data to evaluate group interaction parameters. The data set included alcohols, sugars, amides and ureas. No distinction was made between unsubstituted, mono- or di- substituted amides and ureas. Since the long-term aim of this work is to relate these model solute interactions to peptide and polypeptide systems it is important that the model solutes are good analogues for peptide residues. It is obvious considering the particular molecules investigated here that without reference to possible conformational effects the molecular environments for the functional groups are similar. Group interaction parameters derived from these model systems should therefore transpose well.G.M. BLACKBURN, T. H. LILLEY A N D E. WALMSLEY 1659 We therefore determined to select a more consistent data base by combining our substituted peptide and amino acid data with data on unsubstituted and monoalkyl- substituted amides from which we could extract a set of refined group interaction parameters which would better describe peptide interactions. Since few pairwise free energy coefficients have been reported only a limited data set could be assembled. The TABLE IO.-REFINED GROUP FREE ENERGY PARAMETERS, Gij(J kg mol-2) GCH2-CH2 GCHz--Pep GPep-Pep molal scale -29.8 +29.7 -59.1 mole fraction scale -29.1 +27.8 -48.5 fit -gAB/J kg rnol-' FIG. 5.-Plot of ged;,Pt against gB& for amide systems. The fitted values were obtained using the mole fraction group additivity parameters given in table 10.amides selected for this data set were : N-acetylglycinamide, N-acetyl-L-alaninamide, N-acetyl-L-leucinamide, formamide, acetamide and pr~pionamide.~~ By using a data set of greater coherence we avoid some of the unsatisfactory approximations used previously. Refined group interaction parameters were evaluated using eqn (25) and (26) using a linear regression analysis. Pairwise free energy coefficients evaluated on the molal and mole fraction scales were analysed separately to test which gave the better correlation. The standard deviation of the fit on the mole fraction scale (53.2 J kg moP2) was slightly better than that on the molal scale (60.4 J kg mol-I) but the values of the group parameters are not significantly different bearing in mind the crudity of the approximations made (see table 10).The agreement between experimental values and values predicted using the mole fraction group parameters is shown in fig. 5. The experimental data are reasonably reproduced by the refined group parameters. 54-21660 AQUEOUS SOLUTIONS CONTAINING AMINO ACIDS AND PEPTIDES The standard deviation is slightly higher than in the original correlation, largely as a result of the much smaller data base. However, we feel that the values are likely to be more realistic in the context of amide and peptide interactions. To generate group enthalpy parameters, data from the present study were combined with results on unsubstituted and monoalkyl-substituted amides and all were fitted to eqn (26).When A, was included in the data set the fit was extremely poor with the standard deviation rising to 521.7 J kg moP2 as compared with a standard TABLE 1 1 .-FUNCTIONAL GROUP PAIRWISE ENTHALPY OF INTERACTION PARAMETERS, Hij(J kg mol-2) Hc H *-c H 2 Savage and Wood +40 +41 (34) -252 (113) 220 refined parameters + 14 (13) +95 (29) -311 (57) 140 parameters a The numbers in parentheses represent the 95% confidence limits. CT is the standard deviation of the fit in J kg molt2. I 1 I !OOO 1000 0 1000 - h E /J kg mol' FIG. &--Plot of hyzt against hgk. The fitted values were obtained using the group additivity parameters given in table I I . deviation of 220 J mo1F2 found in the original Savage and Wood46 correlation. We felt justified in omitting this system from the parameterisation since it is our belief that additional specific effects are influencing the solution behaviour of this molecule.With the omission of this system from the data base, enthalpy parameters were generated which gave a much improved correlation between experimental and calculated h,, values. The new parameters are shown in table 11 as are the earlier36 parameters. The results are also presented in a different form in fig. 6 where good agreement, both in sign and magnitude, is apparent between the fitted and experimental en t halp y coefficients.G. M. BLACKBURN, T. H. LILLEY A N D E. WALMSLEY 1661 The refined parameters while of similar sign are different in magnitude to those obtained earlier. The present data base includes molecules with one, two, three and four amide groups, whereas the previous data set was limited to molecules with only one amide group.We are therefore confident that the PepPep and PepCH, interactions are better characterised by our refined parameters and that they will be more satisfactory for predicting peptide and amide interactions. Using the above revised group free energy and enthalpy parameters we can evaluate the group entropy values from A list of all of the group parameters is presented in table 12. Gij = Hij - TS,, . (27) TABLE 12.-FREE ENERGIES, ENTHALPIES AND ENTROPIES OF INTERACTION OF FUNCTIONAL GROUPS. UNITS ARE J kg rnoP CH, Gij -29 Hii +14 TSii +43 Pep Gij +28 Gij -48 Hij +95 Hii -311 TSij +67 TSij -263 The CH,-CH, pair interaction is qualitatively consistent with the established view39,47,49,50 of hydrocarbon interactions.A negative Gij term is usually interpreted as a net attraction between the groups which arises from a large positive term. We also observe that I TSCH,-CHP I > JHCHoCH2J, which is a recognised characteristic of the hydrophobic interaction. The hydrophobic interaction has been widely s t ~ d i e d ~ , ~ ' 55 both theoretically and experimentally, but unfortunately there are few published data directly comparable with the present work. It has been the custom to calculate hydrophobic side-chain contributions to protein conformation by considering the t r a n ~ f e r ~ ~ - ~ * of amino acids and peptides from water to a non-aqueous solvent medium which might simulate the interior of the folded protein.One is thus able to calculate the difference in say the free energy between a side-chain completely surrounded by water and that of the side-chain surrounded by apolar groups without any contact with water molecules. The thermodynamic parameters thus estimated50 do not correspond to the formation of pairwise hydrophobic bonds as described by our model. We have assumed a weak, non-specific association between hydrocarbon groups in an aqueous medium and anticipate only a partial shielding from water on association. Our pairwise association parameters are thus more in keeping with Nemethy and Scheraga's minimum hydrophobic bond strength parameters. Comparison of the present results for the CH,-CH, interaction with other quantitative estimates of the hydrophobic interaction is also difficult because of the different models ~ ~ e d .~ ~ - ~ ~ In our model systems the hydrocarbon parts are flanked by amide groups. We thus recognise that the polar group may interfere with the establishment of water structure around the non-polar group which would modify the hydrocarbon interaction. Our results are therefore most appropriately referred to peptide and amide systems and may only be extrapolated to alternative systems with caution.1662 AQUEOUS SOLUTIONS CONTAINING AMINO ACIDS AND PEPTIDES Various values33-36q 59 have been attributed to the strength of the peptide-peptide bond and it is not clear to what extent these interactions affects the stability of protein structures. The present results indicate that the peptide group interactions make a large contribution to the overall peptide interaction and the magnitude of the PepPep term shows that it has a larger effect than that for the CH,-CH, interaction.The negative pairwise enthalpy term corresponds to an association dominated by hydrogen bonding or dipole effects. Using eqn (22) and values for GPepPPep and HPep-Pep (from table 12) we estimate the standard enthalpy of association of two amide groups as AHpeppPep = - 16.1 kJ mol-l The present group estimate is much more negative than our preliminary estimate from the Gly data (see earlier). In this initial estimate it was assumed that N-acetylglycinamide self-interactions were dominated by the two amide groups and methyl and methylene group effects were ignored.Klotz and F r a n ~ e n ~ ~ reported a zero value for the enthalpy of rupture of the CO - - - HN bond from studies on N-methylacetamide but they too ignored the methyl group contribution. However, the positive enthalpy of formation of the hydrophobic bond will in part cancel the negative enthalpy contribution made by the hydrogen bond. Clearly if this destabilising contribution is ignored the amide interaction will be underestimated by an equivalent amount. Thus oversimplification of the model can lead to misleading conclusions. The large decrease in entropy associated with PepPep interaction indicates less freedom of motion when the two groups interact. This is consistent with the formation of a strong hydrogen bond or bonds which would tend to reduce the entropy of the dimer.Model indicate that amide interactions probably do involve a hydrogen-bonding mechanism. In previous studies on model compounds much emphasis was placed on evaluating the hydrocarbon and amide self-interactions and usually only passing reference40 is made to the effect that polar and non-polar groups exert on each other. The results from this work indicate that these latter effects are by no means insignificant. The association of a peptide and a hydrocarbon group produces a strong destabilising effect which is approximately equal to the stabilising effect of the CH; = .CH, interaction. Since there is no compatible mode of bonding between these two groups the enthalpic term is repulsive, but is compensated in part by the positive entropy term. This may be rationalised if it is assumed that the hydration spheres surrounding the polar and apolar groups are incompatible and that as the groups approach each other the hydration spheres overlap and the water structure is in part broken down.This effect is consistent with the view37~38~60 63 that group effects can be modulated by adjacent groups via the solvent structure. While interpretation of model studies using a site-binding model would underestimate such cross-interaction terms, the present group additivity possibly overemphasises them. However, it seems likely that these interactions are real and their importance should be acknowledged. CONCLUDING COMMENTS The group additivity model has allowed us to evaluate free energy, enthalpy and entropy contributions for the peptide and hydrocarbon groups.In principle we could use these values to predict the molecular pair interaction coefficients for any molecule composed only of these groups. This model is recognised to be a first approximation and several questionable assumptions are inherent in this scheme. Although the group expansion are empirical they can be given a rough interpretation. The molecular interaction parameters g A , are related to the osmotic second virial coefficient via the integral of [exp (- W,,/RT)- 13 where W,, is the potential ofG. M. BLACKBURN, T. H. LILLEY A N D E. WALMSLEY 1663 mean force for pairs of molecules. If WAB is to be expressed in terms of additive group contributions, this implies that Gij (the interaction between each pair of groups) must not depend on the presence of neighbouring groups on either molecule, i.e. all steric effects can be ignored, and that the effects of solvation are made up of additive contributions.Also the sum of all the group interactions must be small compared with kT to allow the factor [exp (WA,/kT)- 11 to be expanded into a linear sum of interaction terms. These assumptions, while providing a useful working model, represent a fairly extreme case and will not necessarily be applicable to all molecular interactions. The model assumes that each functional group on one molecule is able to interact freely with each functional group on another molecule, independently of their positions. Steric effects could reduce interaction by preventing free access to all groups or conversely could enhance interactions.While such effects must cause deviations from the simple additivity rule, it should be possible to use these deviations to investigate the nature of such anomalous behaviour. As an example of this, we found that the additivity treatment was inadequate for solutions containing A,. This discrepancy does not appear to arise from the behaviour of the peptide backbone since correspondence between experiment and calculation for G, is good. It thus almost certainly arise from intramolecular cooperativity of the alanine methyl groups, and we believe this intramolecular effect modifies the intermolecular behaviour. While thermodynamic analysis in itself cannot reveal molecular features, it does establish constraints within which interpretation of molecular behaviour must operate.Structural molecular detail can be disclosed by a variety of spectroscopic probes and although such studies have not been reported for the solutes studied in this work, related results are enlightening. It has been shown that in aqueous solution small peptides adopt extended conformations which are not completely random.64 Amino acid side-chains considerably effect the molecular conformation and free rotation is restricted for all peptide linkages except those involving glycine. Thus while glycine peptides are almost fully extended in aqueous solution, trialanine approaches a simple helix. There is also evidence which supports the existence of folding in short-chain alanine peptides and subsequent side-by-side aggregation of the folded 66 We have used space-filling models of A, to investigate the possibility of folded structures in this species.Several conformations are accessible, in one of which the three methyl groups are drawn into a cluster arrangement giving the molecule one hydrophobic and one hydrophilic face with a consequent reduction in the overall extent of alkyl group’s surface in contact with solvent. If this conformation was energetically favoured in aqueous solution it would be possible for the molecules to associate through their hydrophobic faces. Although unsubstantiated, this idea would explain the behaviour of A, and would be in accord with the enthalpy data. In this study we have restricted our attention to molecules comprised of amino acids with aliphatic side-chains only.The relative simplicity of these molecules and the study of homologous series has helped to clarify the analysis. However, if the approach is to have predictive power for aqueous protein systems the group parameter analysis must be expanded to embrace other functions. If the analysis could be developed to handle longer peptides it might well highlight further examples of the A, type and would enter the zone in which intramolecular peptide interactions compete with intermolecular forces. This information should be particularly relevant to confor- mational features such as thep-turn and identification of polypeptide folding nucleation centres in proteins. One other area which might also be usefully explored using group additivity approaches is that of enzyme-substrate interactions. These and other areas are currently being explored in this laboratory.1664 AQUEOUS SOLUTIONS CONTAINING AMINO ACIDS AND PEPTIDES We acknowledge support from the A.R.C.for equipment and the S.R.C. for the award of a Research Studentship to E. W. G. Nemethy and H. A. Scheraga, Q. Rev. Biophys., 1977, 10, 239. E. Clementi, Lecture Notes in Chemistry (Springer-Verlag, Berlin, to be published). See, e.g. Water, A Comprehensive Treatise, ed. F. Franks (Plenum Press, New York, 1973 and 1975), vol. 2 and 4. G. NCmethy, W. J. Peer and H. A. Scheraga, Annu. Rev. Biophys. Bioeng., 1981, 10, 459. G. M. Blackburn, T. H. Lilley and E. Walmsley, J . Chem. Soc., Faraday Trans. I , 1980, 76, 915. G. M.Blackburn, T. H. Lilley and E. Walmsley, J . Chem. SOC., Chem. Commun., 1980, 1091. F. Franks and D. Eagland, CRC Crit. Rev. Biochem., 1975, 3, 165. P. L. Pnvalov, Adu. Protein Chem., 1980, 33, 167. T. H. Lilley and I. R. Tasker, to be published. lo C. J. Epstein, R. F. Goldberger and C. B. Anfinsen, Cold Spring Harbor Symp. Quant. Biol., 1963, 28, 439. l1 C. B. Anfinsen and H. A. Scheraga, Adv. Protein Chem., 1975, 29, 205. D. B. Wetlaufer and S. Ristow, Annu. Rev. Biochem., 1973, 42, 135. l 3 R. R. Hantgan, G. G. Hammes and H. A. Scheraga, Biochemistry, 1974, 13, 3421. l4 P. N. Lewis, F. A. Monamy and H. A. Scheraga, Proc. Natl. Acad. Sci. USA, 1971, 68, 2293. l 5 E. Ralston and J-L. De Coen, J. Mol. Biol., 1974, 83, 393. l6 M. N. Jones and H. A. Skinner, Annu.Rep. Chem. SOC., 1979, 76C, 253. l8 T. H. Lilley and R. P. Scott, J. Chem. SOC., Faraday Trans. I , 1976, 72, 184. l9 P. K. Nandi and D. R. Robinson, J. Am. Chem. SOC., 1972, 94, 1299. 2o M. Bodansky, J. T. Sheehan, M. A. Ondetti and S. Lande, J . Am. Chem. SOC., 1963, 85, 991. 21 M. Goodman, R. Rupp and F. Naider, Bioorg. Chem., 1971, 1, 294. 22 P. G. Katsoyannis, K. Fukuda and A. Tometski, J . Am, Chem. SOC., 1963, 85, 1681. 23 H. L. Friedman, Ionic Solution Theory (Interscience, New York, 1962). 24 H. D. Ellerton and P. J. Dunlop, J. Phys. Chem., 1966, 70, 1831. 25 W. G. McMillan and J. E. Mayer, J . Chem. Phys., 1945, 13, 276. 26 See, e.g. B. P. Kelley and T. H. Lilley, J. Chem. SOC., Faraday Trans. I , 74, 2771. 27 R. H. Wood, T. H. Lilley and P. T. Thompson, J. Chem. SOC., Faraday Trans I , 1978, 74, 1990. 28 R. Lumry and S. Rajender, Biopolymers, 1970, 9, 1125. 29 R. B. Cassel and R. H. Wood, J . Solution Chem., 1973, 2, 119. 30 F. Franks, M. Pedley and D. S. Reid, J. Chem. SOC., Faraday Trans. I , 1975, 72, 359. 31 G. Rialdi and J. Hermans, J. Am. Chem. SOC., 1966,88, 5719. 32 I. M. Klotz, J . Colloid Interface Sci., 1968, 27, 84. 33 I. M. Klotz and J. S. Franzen, J . Am. Chem. SOC., 1962, 84, 3461. 34 S. J. Gill and L. Noll, J. Phys. Chem., 1972, 76, 3065. 35 G. C. Kresheck, J. Phys. Chem., 1969, 73, 2441. 36 B. Y. Okamoto, R. H. Wood and P. T. Thompson, J. Chem. SOC., Faraday Trans. I , 1978,74, 1890. 37 A. Hambata and P. H. von Hippel, Biochemistry, 1973, 12, 1264. 38 A. Hambata, S. Chang and P. H. von Hippel, Biochemistry, 1973, 12, 1271. 39 C. Tanford, The Hydrophobic Effect (John Wiley, New York, 1973). 41 J. S. Rowlinson, Liquids and Liquid Mixtures (Butterworths, London, 1959). 4 2 E. E. Schrier and E. B. Schrier, J . Phys. Chem., 1967, 71, 1851. 43 G. M. Wilson and C. H. Deal, Ind. Eng. Chem., Fundam., 1962, 1, 20. 44 G. A. Ratcliff and K. C. Chao, Can. J . Chem. Eng., 1969,47, 148. 45 A. Fredenslund, R. L. Jones and J. M. Prausnitz, AZChE J., 1975, 21, 1086. 46 J. J. Savage and R. H. Wood, J . Solution Chem., 1976, 5, 733. 47 C. Tanford, J . Phys. Chem., 1962, 84, 4240. 48 T. H. Lilley and R. H. Wood, J. Chem. SOC., Faraday Trans. 1, 1980, 76, 901. 49 G. Nemethy and H. A. Scheraga, J . Phys. Chem., 1962, 66, 1773. 50 W. Kauzmann, Adv. Protein Chem., 1959, 14, 1 . 51 H. A. Scheraga, Ann. N.Y. Acad. Sci., 1977, 2, 303. 52 H. A. Scheraga, J . Am. Chem. SOC., 1979, 12, 7. 53 E. E. Schrier, M. Pottle and H. A. Scheraga, J . Am. Chem. SOC., 1964, 86, 3448. 54 H. Schneider, G. C. Kresheck and H. A. Scheraga, J . Phys. Chem., 1965, 69, 1310. 55 D. G. Oakenfull and D. E. Fenwick, Aust. J . Chem., 1973, 26, 2649. 56 C. H. Spink and M. Auker, J . Phys. Chem., 1970. 74, 1742. F. Franks, in Biochemical Thermodynamics, ed. M. N. Jones (Elsevier, Amsterdam, 1979). J. Konicek and I. Wadso, Acta. Chem. Scand., 1971, 25, 1571.G. M. BLACKBURN, T. H. LILLEY AND E. WALMSLEY 1665 ’’ P. K. Nandi, Int. J. Pept. Protein Res., 1976, 255, 253. 58 G. C. Kresheck, H. Schneider and H. A. Scheraga, J. Phys. Chem., 1965, 69, 3132. 5s H. Sasi, S . M. Timasheff and J. S . Ardi, J. Biol. Chem., 1964, 239, 3051. 6o J. E. Desnoyers, G. Perron, L. Avedikian and J-P. Morel, J. Solution Chem., 1976, 5, 631. 62 G. NCmethy, I. Z. Steinberg and H. A. Scheraga, J. Am. Chem. SOC., 1963, 85, 3866. 63 T. H. Lilley and I. R. Tasker, J. Chem. Soc., Faraday Trans. I , 1982, 78, 1. 64 B. A. Levine and R. J. P. Williams, Jerusalem Symp. Quantum Chem. Biochem., 1976, 95. 65 M. Goodman, N. Ueyama and F. Naider, Biopolymers, 1975, 14, 901. M. J. Mastroianni, M. J. Pika1 and S . Lindenbaum, J. Phys. Chem., 1972, 76, 3050. M. Goodman, N. Ueyama, F. Naider and C . Gilon, Biopolymers, 1975, 14, 915. (PAPER 1/1324)

ISSN:0300-9599    

DOI:10.1039/F19827801641

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J . Chem. Soc., Faraday Trans. I , 1982,78, 1667-1675 Ion-Molecule Reactions in Gaseous Hydrogen + Pentane Mixtures BY M ICHAEL NEUM ANN-S PA LLART Institut de Chimie Physique, Ecole Polytechnique Federale de Lausanne, CH-1015 Lausanne, Switzerland Received 17th August, 1981 The reactions of H: and D: with n-C,D,, and n-C5HI2, respectively, have been studied. The protonating species have been generated from hydrogen by 6oCo y-rays at 10Torr or by electron impact in an ion-cyclotron-resonance apparatus at 1 OP4 Torr. Using both techniques the relative probability of fragmentation reactions of the pentonium ion was measured and the reaction mechanism is discussed. The transferred proton is shown to be retained to a different extent in the fragment alkyl ions. The pentyl ion isomerizes to the tertiary structure and yields mainly 2-methylbut-2-ene, 2-methylbut- 1 -ene, and isobutene after being neutralized.Ion-molecule reactions of hydrocarbons in the gas phase have been the subject of many investigations carried out in the field of mass spectrometry as well as radiation chemistry over the last few decades.l Chemical ionization has been shown to be a useful tool in deriving mechanistic and kinetic knowledge about the often rather complicated reaction schemes found in hydrocarbon systems. More recently ion-cyclotron-resonance (i.c.r.) studies have helped to define the role of the various ionic precursors and fregmentation channels.,? Information about ionic structures, reactions at elevated pressure and neutralization reactions has mostly been derived from steady-state y- and pulse-radiolysis investigations.Chemical-ionization reactant ions such as CH:, C,H,+ and HZ generated from methane or hydrogen, respectively, are chosen since they easily transfer protons to or abstract hydride ions from hydrocarbons. In the direct radiolysis of pentane, C5HT2 is formed. Most of the parent ions fragment, the fragment ions undergoing ion-molecule reactions. Some parent ions become unreactive and are neutralized by electron^.^ In chemical ionization C5Ht3 is formed by protonation and all C5H:3 ions undergo fragmentation, as will be shown below. The chemical ionization of perdeuterated n-pentane with HZ has already been investigated at atmospheric pres~ure.~ This investigation showed that a proton is quantitatively transferred to pentane.Fragmentation reactions of the C5D12H+ ion thus formed were discussed on the basis of the radiolytic end-products, assuming that the transferred proton is entirely incorporated into the neutral fragment. Yet it has been found in the chemical ionization of n-hexane with methane that there is some retention of labelling in the butyl and propyl ions. Thus, we became interested in seeing if the same effect could be found in n-pentane when the hydronium ion is taken as the reactant. In the present study we therefore reinvestigated the hydrogen + pentane system. Radiolysis experiments of H, + C5D1, mixtures were carried out at low pressure (10 Torr H,) and pentane concentrations of 0.5 mol%. In parallel, i.c.r. experiments were performed with D, + C5H12 mixtures.Another object of this study was the fate of the pentyl ion. This ion is supposed I6671668 ION-MOLECULE REACTIONS I N H,+C,H,, to be formed in high yield via the hydride abstraction from pentane of all fragment alkyl ions. Pentyl ions are obviously unreactive in the system. Thus neutralization is the most probable reaction path; indeed typical products of this were found in our radiolysis experiments. In some experiments oxygen was added as a scavenger to exclude contributions from radical reactions. EXPERIMENTAL MATERIALS The purity of the hydrogen used was 99.995% and that of the deuterium (L'Air Liquide) 99.4%. Perdeuterated pentane was prepared catalytically.6 It contained 10% C,D,,H and, after distillation, 0.1 % isopentane as main impurities.It was dried over molecular sieve (4A) and air was removed by repeated freeze-pumpthaw cycles. For y-irradiations spherical Pyrex vessels (500 and 4000 cm3) connected via metal-glass joints to metal valves equipped with indium seals were used. Each vessel was checked for leaks with a helium leak detector. After pimping out to IO-'Torr they were filled with pentane and hydrogen at least 10 h before iiradiation to allow mixing of the gases to take place. IRRADIATION A N D ANALYSIS Radiolysis was carried out in a Sulzer irradiation chamber (50 kCi yGOCo). The dose rate was determined with an ethylene dosimeter taking G(H,) = 1.36.' The dose rate in hydrogen was calculated by correcting for the stopping-power ratio of hydrogen and ethylene for electrons from Pyrex.8 Values of 1.35 x lozo and 9.83 x IOl9 eV g-' h-' for the 500 and 4000 cm3 vessels, respectively, were obtained.The absorbed dose was varied between 2 x 1919 and 9 x loz0 eV g-l. After irradiation the vessels were immersed in liquid nitrogen for 2 h and hydrogen was then pumped off. Subsequently the radiolytic products were allowed to condense into a short injection loop (20 cm) at liquid-nitrogen temperature for I h. Instead of condensation aliquots were taken by expanding the products into a relatively long injection loop (4m) for the determination of methane, ethane and ethylene, since these substances could not be condensed completely at 76 K or had already been partially pumped off with the hydrogen. The sample loop was then attached to a gas-chromatography-mass-spectrometry system or to a dual column gas-chromatography unit.The latter has already been de~cribed.~ However, the sample introduction system was modified by connecting it to a high-vacuum line. Chromatogramms were run for 80 min at 300 K and then temperature programmed at 4 K min-' up to 336 K. Peak areas were integrated with an Autolab computing integrator (Spectra Physics). The gas-chromatography-mass-spectrometry system consisted of a Perkin-Elmer F- 1 1 chromatograph and a Durapak-Squalane combined single column (resolution 4000 theoretical plates for both injection loops) as described above. The exit of the column was connected to a flame ionization detector (FID) and via a silicone-membrane separator (1 in. diameter, 2/1000 in.thick) held at 353 K (Beta Peripherals) to a QMG 31 1 quadrupole mass filter (Bakers). Signals from the electron multiplier were amplified with a Keithley electrometer and averaged with a Nicolet signal averager. Signals from the FID were treated by a Hewlett-Packard laboratory data system. Substances were identified by their retention times and mass spectra. Absolute amounts were calculated from calibration runs. Principally, hydrocarbons from C, to C,, could be detected by the double-column system. The i.c.r. measurement has alreay been de~cribed.~ RESULTS The amounts of different products formed in steady-state y-radiolysis were very low in most of the experiments (partial pressure 10+ Torr for some products and total quantities down to 10-lo mol); thus carrying out calibration and analysis was difficult.However, in spite of this relatively good reproducibility was obtained ( f 5 %).M. NEU M A NN-S P A L LA RT 1669 TABLE 1 .-G-VALUES OF y-RADIOLYSIS PRODUCTS OF H, + C,D,, MIXTURES AT 10 Torr (i) 7.6 x lo1, eV g-l, 0.5 mol% C,D,,; (ii) 7.6 x lo1, eV g-l, 0.5 mol% C5D12, 0.1 mol% 0,; (iii) 7.6 x lo1@ eV g-l, 0.06 mol% C5D12; (iv)1.2 x loz1 eV g-l, 0.06 mol% C5D12; (v) 7.5 x lozo eV g-l, 0.06 mol% C5D12, 0.1 mol% 0,. methane ethane ethene propane propene isobutane butane isobu tene isopentane 2-methylbut-1 -ene 2-me thy1 but-2-ene 2.0 0.8 0.5 0.9 0.18 0.17 1 .o 0.4 (0.1)" 0.31 (0.7)a 0.72 (1 .O)a - 1.8 0.7 0.6 0.9 0.18 0.06 1.1 0.1 0.4 1 .o - 0.5 0.15 0.4 0.06 0.07 0.9 0.14 (0.4)a 0.3 (0.1)" 0.31 (0.7)a 0.71 (1.0)" ~~ - 1.7 0.6 0.3 0.04 0.07 0.5 0.3 0.02 0.03 0.16 0.01 0.9 0.7 0.02 0.7 0.01 0.04 0.03 0.11 0.17 - a 1.8 x lo1, eV g-l.TABLE 2.-ISOTOPIC COMPOSITION OF 7-RADIOLYSIS PRODUCTS (%) OF H, + C,Dl, MIXTURES AT 10 Torr (VALUES NOT CORRECTED FOR THE ISOTOPIC IMPURITY OF THE STARTING MATERIAL) (1) 5.5 x 1019 eV g-l, 0.5 mol% C,D,,; (ii) 8.8 x 1020 eV g-l, 0.5 mol% C6D12; (iii) 8.8 x 1020 eV g-l, 0.5 mol% C5D12, 0.1 mol%O,; (iv) 8.8 x 1020 eV g-l, 0.06 mol% C5D1,; (v) 8.8 x 1020 eV g-l, 0.06 mol% C5D1,, 0.1 mol% 0,. product 0) (ii) (iii) (iv) ( 4 CD3 H CD4 c2 D, H c2 D, c3 D, H C4 D, H '2 D4 C2D5H '3 D8 '4 D I O i-C, D, H3 i-C, D,, H, i-C, D,, i-C, D,, i-C, D, H,-en- 1 i-C, D, H-en- 1 i-C, Dl0-en- 1 i-C, D, H,-en-2 i-C, D, H-en-2 i-C, D,,-en-2 55 45 22 78 12 88 3 23 74 0 21 79 28 72 61 39 26 74 13 87 12 32 30 26 6 30 64 6 27 67 46 54 16 84 9 91 3 34 63 0 27 73 80 20 28 72 71 29 31 69 13 87 17 40 28 13 9 37 54 5 29 66 67 33 22 78 9 91 4 34 62 0 27 731670 ION-MOLECULE REACTIONS I N H,+C,H,, The G values under various conditions are listed in table 1.Assuming a G( - pentane) value of 5 the conversion of pentane was 2.6% at a dose of 8.8 x 1019 eV g-l. Traces of compounds in the C, to C,, region were found and were ascribed to the direct radiolysis of pentane. The observed decrease in yield of some lighter products at lower pentane concentrations can only partially be explained by contributions from the direct radiolysis of pentanelO at 0.5 mol% pentane. The isotopic compositions of the radiolysis products were calculated as follows.Mass spectra obtained from calibration runs for each non-deuterated compound were used to calculate the mass spectra of the partially deuterated compounds assuming no isotope effect on fragmentation. These spectra were used to estimate the isotopic compositions given in table 2. The spectra of some of the products listed in table 1 could not be measured because of the intensities of the M+, ( M - I)+, etc. ions were too low. The G values and isotopic distributions presented are mean values of several determinations. In the i.c.r. experiments mixtures of D, + n-C,H,, were used [p(D,) = 4 x lo-, Torr, p(C5H12) = 1 x lo-, Torr]. Relative fragmentation probabilities of C,H,,D+ formed by deuteron transfer from D: were monitored by modulated ejection of D3+.Their values are given in tables 3 and 4 together with the values for the corresponding products (see Discussion) from y-radiolysis. (G values are converted to M/N+ using M/N+ = G W/ 100. W, the energy required fbr the formation of one ion pair is 36.3 eV for electrons in hydrogenll and CM/N+ = 1 for ionic products.) TABLE 3.-LABEL RETENTION IN FRAGMENT ALKYL IONS (I.C.R. OF D, + c5 Hl, MIXTURE) AND CORRESPONDING STABLE END-PRODUCTS (7-RADIOLYSIS OF H, + C5Dl, MIXTURES AT 10 Torr, VALUES CORRECTED FOR ISOTOPIC IMPURITIES OF THE STARTING MATERIAL) ion relative y-radiolysis relative (i.c.r.) intensity product abundance c3 H;: c4 H,+ C3 H, D+ C, H, D+ 82 18 96 4 TABLE 4.-FRAGMENTATION OF C,X:3 FORMED FROM D, + C5H12 MIXTURES (I.C.R. EXPERIMENT) AND H, + C5 D,, MIXGURES (~-RADIOLYSIS EXPERIMENT) heat of ion product reaction/ (i.c.r.) Ii/ZIi (y-radiolysis) Mi/N+ kcal mol-I '5 'Tl c4x; + c3x; 0.20 x2 (0.22)" 0.32 '4 x10 + '3 x6 0.44 0.48 c,x, 0 .3 4 0.21 c4x,o 0.37 0.1 1 C3X6 0.7 1.9 8.9 13.7 a Value calculated using M(C,Xtl)/N+ = 1 - [M(C,X,+)/N+ + M(C,X,f)/N+ + M(C,X;)/N+].M. N EU M AN N-S P ALL ART 1671 DISCUSSION When hydrogen is irradiated with y-rays or electrons, it is ionized and transfers a proton to another hydrogen molecule: H l + H, + H3+ + H AH = - 38 kcal mol-l. (1) The reactant ion thus generated will transfer a proton to the added hydrocarbon in an exothermic process: (2) The heat of reaction (2) is given by the difference in proton affinities (A,) of hydrogen and pentane (see Appendix for all the heats of reactions) and may differ by several kcal mol-1 according to the assumed structure of the C5H;t, ion.This ion therefore carries a large amount of internal energy. According to ref. (12) the transferred proton is localized in a three-centre bond when a hydrocarbon is protonated: H i + C, D,, + H, + C , D,, H+ AH = - 56 kcal mol-l. Subsequent scission of such a bond leads to the formation of an alkyl ion and a neutral molecule, as is also indicated by the experimental results: Reaction enthalpies are estimated from the heats of formation of the non-deuterated compounds (see the Appendix). The formation of P-C,X;~ (X = H,D) is clearly not favoured because of its high endothermicity (AH = 21 kcal mol-l) as it was demonstrated in an i.c.r. experiment on the chemical ionization of [ 1 ,5-D6]pentane with CH, where the pentyl ion formed via CHZ as precursor contained 99% C5H5D,+.13 Fragment alkyl ions were detected in the i.c.r.experiments with mixtures of D, and C5H12 (see tables 3 and 4). The postulated reaction5* l4 C,Xt3 + C,X$ + C,X, AH = 20 kcal mol-l (9) does not play an important role in the reaction scheme : only very weak signals ( < 1 % total ionization current) of m / z 29 and 30 were observed. Alkyl ions from reactions (3)-(8) are known to isomerize rapidly to the secondary structure1 and to abstract H- from pentane (rate constant z cm3 molecule-l s-l):15 (10) (1 1) (12) (13) AH = - 1.8 kcal mol-l AH = - 5.9 kcal mol-l. I s - C ~ D i + C5 D1, + C4 D1, + s - C ~ D;i S-C4 D, H+ + C5D12 + C, D, H + s-C, Dt1 s-C~DT + C5D12 + C3 D i + s - C ~ Dt1 S-CaDBH++ C5 D12 -+ C, D, H + s-C, Dl11672 ION-MOLECULE REACTIONS I N H,+C5Hl, The G value of butane and propane and the corresponding products (methane and ethane, respectively) of a given isotopic composition indicate the probabilities of fragmentation channels (3)-(8).According to the proposed reactions [reactions (3)-(8) and (10)-(13)] all ions are ultimately converted to pentyl ions (the so-called funnel principle). Their reactions will be discussed below. Butane originating from ion-molecule reactions possesses an unbranched structure, as is shown from the stable end-product analysis of y-radiolysis experiments. The formation of isobutane was surpressed when oxygen was added as a radical scavenger and is thus interpreted as a radical process.Accordingly, butyl ions maintain their initial unbranched structure under radiolysis conditions within the nanosecond time-scale. This was recently confirmed by Shold and Ausloos.16 It is shown (table 3) that the incoming proton is almost entirely incorporated in the neutral fragment formed by methane loss, in agreement with experiments on the methane + n-hexane system. The formation of C,Hz probably occurs by further fragmentation of butyl ions which have isomerized to the tertiary structure: (14) It was shown that all carbon atoms are equivalent in this process.17 Consequently the tertiary structure is ascribed to butyl ions fragmenting. Hydride abstraction is also thought to be the main reaction for C,X; ions formed in reaction (14): c, x,+ -+ c, x; + c x , .C,X,++C,X,, -+ C,X,+C,X[~. For the propyl ion high retention was found: C,H,D+/C,X; = 0.20 (i.c.r.) and C,D, H/C,X, = 0.18 (y-radiolysis). These values are lower limits since some C, H;f (C, D8) arises from reaction (1 6): S-C,X;~ -+ n-C,X:+C,X, AH = 44.5 kcal mol-l. (16) The same process needs only 28.5 kcal mol-l when s-C, X; is formed in a rearrangement reaction, as was found for hexyl radicals (formed by H- abstraction from hexane by C,X;) losing ethylene., Such a process might well be favoured because of the higher exothermicity of the initial proton-transfer reaction when H: or D3+ is taken as a reagent. In pentyl ions low label retention was found in the i.c.r. experiment: C, Hl,D+/C,X~l = 0.04. Since pentyl ions arise from different secondary reactions, no conclusions about reactions (3) and (4) could be drawn from y-radiolysis experiments, as it was not possible to determine the radiolytic hydrogen yield in the large amount of the reacting gas, hydrogen.Generally, label retentions in y-radiolysis are also reflected (as a mirror-image of those of fragment alkyl ions) by the isotopic composition of the neutral products of fragmentation reactions (3)-(8). Also, the G values of the corresponding products should be the same. For ethane (G = 0.8) and propane (G = 0.9) [table 1, column (i)] formed in reactions (7) and (8) and (12) and (13), respectively, this is in fact the case. Combining these G values with the isotopic abundances (table 2) reasonable agreement was observed for the corresponding products [G(C, H, D) = 0.43, G(C,D,) = 0.73 and G(C,D,) = 0.35, G(C,D,H) = 0.21)], taking into consideration that at least a part of C,D, originates from the fragmentation reaction (16).Comparing the products from reactions (5) and (6) and from subsequent reactions, reaching (lo), (1 l), (14) and (1 5), in the same way, G(butane) + G(propene) = G(buty1 ion) [reactions (5) and (6)] and G(methane) (after substraction of G(CD,) [formed in reaction (1 4)], which is equal to G(C, D6) = 0.1 8} should be equal. Yet the comparison [G(butane) + G(propene) = 1.18, G(methane) - G(propene) = 1.71 suggests an addi- tional path to methane formation, as will be discussed below.M. NEU M ANN-S PA L LAR T 1673 Summing up the i.c.r. intensities of C,X; and C,X; and comparing them with the intensities of C,X: and C,Xrl (table 4) the distribution does not correspond to the sequence of reaction enthalpies as it does in the CH, + C,H,, system (except for the hexyl ion)., One reason is the further fragmentation of the pentyl ion [reaction (16)], another might be the difference in the heats of formation of pentonium ions of different structures. Relatively large differerlces are to be expected between C-H and C-C protonated forms, which possibly may provide an explanation for the unexpectedly low reaction probability leading to the parent alkyl ions, as was found for both the CH,+C,H,, and CH,+C,H,, systems.It also follows from table 4 that secondary fragmentations play a much greater role in the low-pressure i.c.r. experiments, e.g.comparing the ratio of propane to butane and propyl ion to butyl ion. It is therefore difficult to estimate the relative amounts of initially formed fragment ions by y-radiolysis. Yet it is shown that at 10 Torr the ratio of propane to butane follows the order of reaction enthalpies for reactions (5)-(S), in contrast to the i.c.r. case. The very low retention in the butyl ion reflects the same trend, namely that methane loss, which demands less energy, is a faster process than ethane loss leading to propyl ions. The reactions forming pentyl ions [reactions (3) and (4)] have to be treated separately (see above). The estimated radiolytic yield (M(penty1 ion)/N+ = 1 - [M(propane)/ N+ + M(butane)/N+ + M(propene)/N+] = 0.22) is close to the value found in i.c.r. experiments (tables 3 and 4).A further comparison of the distribution of pentonium-ion fragments in i.c.r. and y-radiolysis shows that the most endothermic reaction, the formation of propyl ion, is relatively faster in the i.c.r. case. This can be explained by the fact that the protonating ions, Hi, are internally excited under i.c.r. conditions at lo-, Torr, as was demonstrated by Bowers et aZ.lS A relatively lower extent of proply-ion formation accompanied by a higher degree of label retention (33%) was measured by Ausloos in the y-radiolysis of a hydrogen+pentane mixture at 760 Torr,, if we interpret the data given in this reference in the manner given above; i.e. reaction (9) does not take place. The effect can be explained by de-excitation of H i and/or C,X;t, at high pressure.From reactions (3)-(S), (10)-(13) and (15) it follows that all ions initially formed should ultimately yield C,Drl with a theoretical M/N+ value of 1 (corresponding to G = 2.76). C,D;, plays the typical role of an unreactive ion in the system. It has been shownlo? l8 to undergo rearrangement to the tertiary structure. Neutralization must be considered as a predominant reaction for t-C,Dfl ions t-C, Drl + e- (t-C, D,,)* (17) forming highly excited pentyl radicals. In the y-radiolysis of n-pentanelO and the pulse radiolysis of neopentane,19 isopentenes and isobutene were found, inferring an ionic precursor of branched structure, namely the tertiary pentyl ion. Probable reactions of excited pentyl radicals are fragmentations : (t-C, Dll)* --* (CD,), C=CD CD, + D (1 8) (19) -+ CD=C(CD,)CD, CD, + D + (CD,), C=CD, + CD,.The products of reactions (1 8)-(20) were indeed found in our system, and no linear pentenes were detected. Thus it is proved that all pentyl ions have time to isomerize to the tertiary structure before neutralisation under the conditions given here. CD, radicals formed in reaction (20) are thought to combine with H atoms, present at high concentration in the system, thus explaining the relatively high G value of CD,H discussed above.1674 ION-MOLECULE REACTIONS I N H,+C,H,, Table 1 shows that the G values of all unsaturated products decrease strongly with increasing dose accompanied by an increase of G values of some radical products like isopentane, which can be suppressed almost completely with oxygen. On the other hand, oxygen, when added, increases the yields of unsaturated products by the protection of double bonds from radical attack.The isotopic distribution of 2- methylbut-1 -ene and 2-methylbut-2-ene shifts towards higher concentrations of higher deuterated compounds at lower conversions (table 2). The same effect is seen at high conversion when oxygen is added as a free-radical scavenger. This may be interpreted as the consequence of hydrogen-radical attack at the double bond forming isopentyl radicals, the disproportionation of which would again yield branched pentenes and isopentane. Their recombination product with hydrogen atoms is also isopentane. Indeed an increase of G(isopentane) with dose was observed. The same trends and patterns of isotopic distribution of isobutane (data not given) indicate that similar processes take place when isobutene is attacked by hydrogen atoms.The sum of G values of 2-methylbut-1 -ene, 2-methylbut-2-ene, isobutene and isobutane ( = 2.2) at the lowest dose used (table 1) accounts for 80% of the pentyl ions expected theoretically when all ions ultimately react to give these ions. The author thanks Drs R. Houriet and J. Dawson for help with the i.c.r. experiments, Ammanz Ruf for operating the gas-chromatography-mass-spectrometry system, Dr S. Lukac for many discussions and Prof. T. Gaumann for encouraging this work, which was supported by the Swiss National Foundation. APPENDIX The following values were used for the calculations of heats of reaction, AH: A,(H,) = 101 kcal mo1-l A,(C5 HI,) = 157 kcal mol-l AH,(s-C, HTl) = 176 kcal mol-1 AH,(p-C, Htl) = 195 kcal mol-I (estimated by extrapolation from known proton affinities of the lower hydrocarbons).’ and were calculated from AH,(s-C,H,,) = 8 kcal mol-l, AHf(p-C,Hll) = 11 kcal mol-l, ionization potential (s-C,H1,) = 168 kcal mol-l, ionization potential (p-C, Hll) = 184 kcal mol-1 [all estimated by extrapolation from data of ref.(20)]. AH,(s-C,H;) = 183 kcal mo1-l AH,(s-C,H;) = 192 kcal mol-I AHf(Cz H$) = 119 kcal mol-’ were taken from ref. (20). S. G. Lias and P. Ausloos, Ion-Molecule Reactions-Their Role in Radiution Chemistry (American Chemical Society, Washington, D.C., 1975). R. Houriet, G. Parisod and T. Gaumann, J. Am. Chem. SOC., 1977,99, 3599. R. Houriet and T. Gaumann, Int. J. Mass Spectrom. Ion Phys., 1978, 28, 93. M. Neumann-Spallart and S. LukaC, Radiat. Phys. Chem., 1980, 15, 723. P. Ausloos and S..G. Lias, Discuss, Faraday Soc., 1965, 39, 36. T. Gaumann, H. Oz and 0. Piringer, Helu. Chim. Acta, 1978, 61, 258. I. Janovsky, J. Radiat. Phys. Chem., 1976, 8, 396. D. W. Huyton and T. W. Woodward, Radiat. Res. Rev., 1970, 2, 205.M. N E U M A NN-S PALL A R T 1675 S . LukaE, Chromatographia, 1979, 12, 17. lo S . LukaE, Radiat. Phys. Chem. 1980, 15, 713. l1 A. Henglein, W. Schnabel and J. Wendenburg, Einfiihrung in die Strahlenchemie (Verlag Chemie, Weinheim, 1969). K. Hiraoka and P. Kebarle, J. Am. Chem. Soc., 1976, 98, 61 19. Spectroscopy, ed. H. Hartmann and K. P. Wanczek (Springer, Berlin, 1978). 13 P. Houriet and T. Gaumann, Lecture Notes in Chemistry, vol. 7, Ion Cyclotron Resonance l 4 P. Ausloos, S. G. Lias and R. Gordon Jr, J . Phys. Chem. 1963, 39, 3341. l 5 L. W. Sieck and S. G. Lias, J. Phys. Chem. Ref. Data, 1976, 5, 1123. I6 D. M. Shold and P. Ausloos, J. Am. Chem. SOC., 1978, 100, 7915. 1’ Hei-Wun Leung, Chun Wai Tsang and A. G. Harrison, Org. Mass Spectrom., 1976, 11, 664. T. Su and M. T. Bowers, J. Am. Chem. SOC., 1973,95, 761 1. l 9 R. E. Rebbert and P. Ausloos, J. Res. Natl. Bur. Stand. Sect. A, 1972, 76, 329. 2o F. P. Lossing and G. P. Semeluk, Can. J. Chem., 1970, 48, 995. (PAPER 1 / 1325)

ISSN:0300-9599    

DOI:10.1039/F19827801667

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1 , 1982, 78, 1677-1687 Thermodynamics of Transfer of Sorbitol and Mannitol from Water to Aqueous Solutions of Urea, Guanidine Hydrochloride and Sodium Chloride BY RAKSH V. JASRA AND JAGDISH C. AHLUWALIA* Department of Chemistry, Indian Institute of Technology, New Delhi 110 016, India Received 2nd September, 198 1 Integral enthalpies of solution at 298.15 and 308.15 K and densities at 298.15 K for sorbitol and mannitol in aqueous solutions of urea (2,4 and 6 mol kg-'), guanidine hydrochloride (2 and 4 mol kg-') and sodium chloride (2 and 4 mol kg-I) have been detelxined. These data have been used to derive thermodynamic functions (uiz. enthalpy, heat capacity and partial molal volumes) for transfer from water to aqueous solutions of urea, guanidine hydrochloride and sodium chloride.The thermodynamic data have been rationalized in terms of polyol-cosolute interactions. The difference in the behaviour of sorbitol and mannitol has been explained in terms of a specific hydration model. The effect of sugars and polyols on the solvent properties of water and the interactions of such solutes with electrolytes and non-electrolytes is of biological and thermodynamic importance. A few studie~l-~ have shown that sorbitol and mannitol, despite their similar structures, have different hydration properties. For example, it has been reported' that these compounds show a real difference in the osmotic pressure coefficients in measurements up to 1 mol dm-3. Wilson and Wen,2 from a study of enthalpies of transfer of sodium chloride from water to aqueous solutions of polyols, inferred that sorbitol is a relatively stronger disruptor of the solvent structure than mannitol.Heat capacity and volume s t u d i e ~ ~ - ~ have also shown that mannitol behaves differently from sorbitol as far as hydration is concerned. It will be interesting to see how this difference in the hydration behaviour of sorbitol and mannitol is modified in the presence of an electrolyte or non-electrolyte cosolute in the solution. It is well known6-10 that the extent of denaturation of certain proteins by urea-type denaturants is reduced in the presence of sugars or polyhydroxy alcohols. However, whether this renaturating effect on protein structures is due to the direct binding of urea with the polyhydroxy compound/protein molecules or is due to the alteration of water structure is not known. It was thought that the transfer of polyhy&oxy alcohols from water to aqueous urea solution might shed some light on urea- polyhydroxy alcohol interactions. With these objectives, a thermodynamic study of the transfer of sorbitol and mannitol from water to aqueous solutions of urea and guanidine hydrochloride was taken up.To compare the effect of urea and guanidine hydrochloride with that of an electrolyte, transfer studies were also carried out in aqueous sodium chloride solutions. Thermodynamic parameters determined include enthalpy of transfer, heat capacity of transfer and volume of transfer. 16771678 THERMODYNAMICS OF TRANSFER EXPERIMENTAL The isoperibol submarine calorimeter used for measuring the precise values of integral enthalpies of solution at very low concentration of the solute was essentially similar to that described earlier." The calorimeter consisted of a glass Dewar-type flask (capacity 550 cm3) immersed in a water bath maintained at a constant temperature (to L 0.005 K) by a Tronac PTC 40 proportional temperature controller.A 3 kSZ thermistor (Y.S.I., Ohio, U.S.A.) acting as a temperature-sensing probe formed one arm of the Wheatstone bridge, constructed from precision decade resistors. Output voltage from the bridge was amplified by a Keithley model 140 nanovolt d.c. amplifier and then recorded on a Bryans 28000 potentiometric strip-chart recorder (Bryans Southern Instruments Ltd, U.K.). The calorimeter was calibrated electrically immediately before and after each experiment by passing a constant (ca. 50 mA) current through a non-inductively wound manganin calibration heater. A Keithley model 227 constant current source coupled with a digital timer (model DET 203, Thermadyne, New Delhi) was used as a power source for the calibration heater.Resolution of the temperature measurement system was 1 x lop5 K over the whole range. The operation of the calorimeter was checked by measuring the enthalpies of solution of KC1 in water and THAM (tris-hydroxyaminomethane) in 0.1 mol dm-3 HCl. The AH? value for KC1 at 298.15 K agrees to within 1 % with the reportedl23 l3 values and the AH? value of THAM in 0.1 mol dm-3 HC1 agrees to within & 0.2% with reported13 values under similar conditions.Densities were measured with a vibrating tube digital density meter model DMA 602/60 (Anton Paar, Austria). Water from a constant temperature bath was circulated through the metallic block of the density meter at a flow rate of 3 dm3 min-'. Thermal stability, maintained by Tronac PTC 40 proportional temperature controller, was better than & 1 x lop3 K. The density meter was calibrated with water and dry air. The density of water at 298.15 K was taken as 0.997 047 g cmP3 from Kell's data.14 The reproducibility of the measurements was better than 3 x Sorbitol and mannitol of highest purity grade (Sigma Chemicals) were used without further purification. AnalaR grade sodium chloride and urea (B.D.H.) were used after drying for 48 h (urea at 333 K and sodium chloride at 373 K).Practical grade guanidine hydrochloride (Sigma Chemicals) was recrystallized from water + hydrochloric acid using the method described by Nozaki.15 The recrystallized sample was dried at 333 K for two days before use. Deionized water, obtained by passing distilled water through a Barnstead mixed-bed ion-exchange resin column, was used for enthalpy measure- ments. The water used for density determinations was conductivity water prepared by distilling deionized water in alkaline potassium permanganate. The urea, sodium chloride and guanidine hydrochloride solutions were made up by weight. g cm-3 for dilute solutions. All the solutions were made up by weight. RESULTS The integral enthalpies of solution, AH,, of sorbitol and mannitol were measured in aqueous solutions of urea, guanidine hydrochloride and sodium chloride at 298.15 and 308.15 K. The concentrations of sorbitol and mannitol ranged from to lop3 mol kg-l.The concentrations of cosolutes were: 2, 4 and 6 mol kg-l for urea, 2 and 4 mol kg-l for guanidine hydrochloride and 2 and 4 mol kg-l for sodium chloride. Since the measurements of integral enthalpies, AH,, were performed with very dilute concentrations of sorbitol and mannitol, the effect of concentration, if any, is masked by experimental error. The values of the integral enthalpies of solution at infinite dilution, AH?, for both sorbitol and mannitol, were taken as the average of the AHs values. The AH: values for sorbitol and mannitol in aqueous solutions of urea, guanidine hydrochloride and sodium chloride, at various concentrations, are given in tables 1 and 2, respectively.The uncertainties are computed as 95 % confidence limits.R. V. JASRA A N D J. C. AHLUWALIA 1679 TABLE 1 .-STANDARD THERMODYNAMIC PARAMETERS FOR SORBITOL IN AQUEOUS SOLUTIONS OF SODIUM CHLORIDE, GUANIDINE HYDROCHLORIDE AND UREA AH,"/kJ mol-l ACF/J K-' mol-l Vp a/cm3 mol-l cosolute concentration 298.15 K 308.15 K 303.15 K 298.15 K water 2 mol kg-l sodium chloride 4 mol kg-' sodium chloride 2 mol kg-' guanidine hydrochloride 4 mol kg-' guanidine hydrochloride 2 mol kg-l urea 4.25 mol kg-l urea 6 mol kg-l urea 16.86f0.07 16.56 f0.09 18.80f0.05 18.46 k 0.10 194f9 190k 13 119.26 121.05 16.60 f 0.08 18.54 f 0.03 194f8 121.94 16.24 k 0.05 18.21 fO.11 198 f 12 121.16 16.00 f 0.09 18.00 f 0.09 200f 12 121.64 15.77f0.09 15.38 f 0.15 14.89f0.12 17.89 k 0.12 17.39 f 0.07 17.04 f 0.05 212f 14 201 f 15 215f 13 120.99 120.71 121.1 1 a For volume measurements concentrations of cosolute are: sodium chloride 2 and 3.9 mol kg-l; guanidine hydrochloride 2 and 3:9 mol kg-l; urea 2, 4.5 and 5.9 mol kg-l.TABLE 2.-sTANDARD THERMODYNAMIC PARAMETERS FOR MANNITOL IN AQUEOUS SOLUTIONS OF SODIUM CHLORIDE, GUANIDINE HYDROCHLORIDE AND UREA AH,"/kJ mol-l AC,"/J K-l mol-l i7pa/cm3 mol-l cosolute concentration 298.15 K 308.15 K 303.15 K 298.15 K water 2 mol kg-l sodium chloride 4 mol kg-' sodium chloride 2 mol kg-l guanidine hydrochloride 4 mol kg-l guanidine hydrochloride 2 mol kg-l urea 4.25 mol kg-l urea 6 mol kg-l urea 21.54f0.09 22.01 k0.07 23.86 f 0.04 24.79 f 0.10 232 9 278 f 13 119.53 121.34 22.55f0.10 25.23 f 0.05 268f11 123.37 22.01 f 0.03 24.38 f 0.13 236f 13 121.43 22.07 k 0.09 24.51 f0.08 244f 12 121.85 21.53k0.07 20.99 f 0.13 20.91 f 0.1 1 23.79 f 0.05 23.32 f 0.1 1 23.40 f 0.07 226 f 10 232f 17 249f 13 120.60 120.55 120.95 a For volume measurements concentrations of cosolute are: sodium chloride 2 and 3.9 mol kg-l; guanidine hydrochloride 2 and 3.9 rnol kg-l; urea 2, 4.5 and 5.9 mol kg-l.The enthalpies of transfer, AH,, [AH,, = AH? (cosolute +water) - AH? (water)], for sorbitol and mannitol from water to aqueous solutions of urea, guanidine hydrochloride and sodium chloride at 298.15 and 308.15 K at various concentrations of these cosolutes are reported in tables 3 and 4 and are plotted as a function of cosolute concentration in fig.1 (298.15 K) and 2 (308.15 K). The integral enthalpies of solution in water for sorbitol and mannitol are taken from the literat~re.~1680 THERMODYNAMICS OF TRANSFER TABLE 3 .-THERMODYNAMIC PROPERTIES ASSOCIATED WITH THE TRANSFER OF SORBITOL FROM WATER TO AQUEOUS SOLUTIONS OF SODIUM CHLORIDE, GUANIDINE HYDROCHLORIDE AND UREA AHJkJ mol-l ACPtr/J K-l mol-l A Kra/cm3 mol-I cosolute concentration 298.15 K 308.15 K 303.15 K 298.15 K 2 mol kg-l sodium -0.30 - 0.34 -4 4 mol kg-l sodium - 0.26 - 0.26 0 2 mol kg-' guanidine -0.62 - 0.59 4 chloride chloride hydrochloride hydrochloride 4 mol kg-l guanidine -0.86 -0.80 6 2 mol kg-l urea - 1.09 -0.91 18 4.25 mol kg-l urea - 1.48 - 1.41 7 6 mol kg-l urea - 1.97 - 1.76 21 1.79 2.68 1.90 2.38 1.73 1.45 1.85 a For volume measurements concentrations of cosolute are: sodium chloride 2 and 3.9 mol kg-l; guanidine hydrochloride 2 and 3.9 mol kg-l; urea 2, 4.5 and 5.9 mol kg-l.TABLE 4.-THERMODYNAMIC PROPERTIES ASSOCIATED WITH THE TRANSFER OF MANNITOL FROM WATER TO AQUEOUS SOLUTIONS OF SODIUM CHLORIDE, GUANIDINE HYDROCHLORIDE AND UREA AHJkJ mol-l ACPtr/J K-l mol-l Aytra/cm3 mol-l cosolute concentration 298.15 K 308.15 K 303.15 K 298.15 K 2 mol kg-l sodium 0.47 0.93 46 4 mol kg-l sodium 1.01 1.37 36 chloride chloride hydrochloride hydrochloride 2 mol kg-l guanidine 0.47 0.52 4 4 mol kg-l guanidine 0.53 0.65 12 4.25 mol kg-l urea -0.55 -0.54 0 2 mol kg-l urea - 0.0 1 - 0.07 6 6 rnol kg-' urea - 0.63 - 0.42 17 1.81 3.84 1.90 2.32 1.07 1.02 1.42 a For volume measurements concentrations of cosolute are: sodium chloride 2 and 3.9 mol kg-l; guanidine hydrochloride 2 and 3.9 mol kg-l; urea 2, 4 .5 and 5.9 mol kg-l. The heat capacities of dissolution at infinite dilution, AC;, of sorbitol and mannitol in aqueous solutions of urea, guanidine hydrochloride and sodium chloride at 303.15 K, calculated using the integral heat method,ll* 16-19 are given in tables 1 and 2, respectively. The heat capacities of transfer, ACptr, for sorbitol and mannitol at 303.15 K from water to aqueous solutions of urea, guanidine hydrochloride and sodium chloride are reported in tables 3 and 4, respectively. The heat capacities of dissolution in water at 303.15 K were taken from the literat~re.~R. V. JASRA A N D J.C. AHLUWALIA 1681 The apparent molal volumes are calculated from density data using the relation M 1000 (d-do) 9 v = d - dd,m where M is molecular weight of the solute, m is the molality of the solute, d is the density of solution and do is density of pure water. The partial molal volumes for sorbitol and mannitol were obtained by least-squares fitting of the apparent molal volume data and are given in tables 1 and 2. 2 L 6 [cosolutel/mol kg-' FIG. 1.-Enthalpies of transfer of sorbitol (closed symbols) and mannitol (open symbols) from water to various mixed aqueous solutions at 298.15 K. 0, @, water + water+NaCl; 0, ., water + water+ guanidine hydrochloride; A, A, water + water + urea. The volumes of transfer, A I&, values are reported for sorbitol and mannitol in tables 3 and 4, respectively.The values of the partial molal volumes for sorbitol and mannitol in water were taken from the literat~re.~ The volumes of transfer for sorbitol and mannitol are plotted against concentration of cosolute in fig. 3 and 4. DISCUSSION ENTHALPIES OF TRANSFER Both sorbitol and mannitol show negative enthalpies of transfer at all concentrations of urea (tables 3 and 4, fig. 1 and 2) and the negative enthalpy of transfer increases in magnitude with increasing urea concentration. Note also that the enthalpy of transfer for sorbitol is more negative than that of mannitol at all concentrations of urea.1682 THERMODYNAMICS OF TRANSFER No direct calorimetric study of the transfer of a hydrophilic non-electrolyte from water to aqueous urea solution could be traced in the literature.However, the enthalpies of transfer of a peptide unit (-CONH) or peptide backbone (-CH,CONH-), which have potential hydrogen-bonding sites similar to the hydroxyl groups of sorbitol and mannitol, have been reported by some worker^.^^-^^ In these groups negative enthalpies of transfer, which become more negative at higher concentrations of urea, were also observed. I I I 2 4 6 [ cosolute] /mol kg-’ FIG. 2.-Enthalpies of transfer of sorbitol (closed symbols) and mannitol (open symbols) from water to various mixed aqueous solutions at 308.15 K. See fig. 1 for key. -2.01 0 The enthalpy of transfer behaviour for sorbitol and mannitol from water to aqueous urea solution is seen to contrast with that of hydrophobic solutes20* 24-28 (which show positive enthalpies of transfer) but is similar to that of ionic, hydrophilic 30 (which show negative enthalpies of transfer).The positive enthalpies of transfer from water to aqueous urea solution for hydrophobic solutes and negative enthalpies of transfer for ionic hydrophilic solutes have been rationalized in terms of the structure-breaking effect of urea molecules. However, it appears that the behaviour of hydrophilic non-electrolytes can be best accounted for in terms of solutexosolute interactions. The negative enthalpies of transfer for sorbitol and mannitol seem to result from interactions between urea and polyol molecules. The most probable interaction appears to be formation of a hydrogen-bonding complex, since both urea and polyols possess potential sites for hydrogen bonding. According to this view, the magnitude of exothermicity should increase with increasing urea concentration, which is indeed the case.The idea ofR. V. JASRA AND J. C. AHLUWALIA 1683 [ cosolute]/mol kg-I FIG. 3.-Volumes of transfer of mannitol from water to aqueous solutions of urea (A), guanidine hydrochloride (a) and sodium chloride (0) at 298.15 K. FIG. 4.-Volumes of transfer of sorbitol from water to aqueous solutions of urea (A), guanidine hydrochloride (0) and sodium chloride (0) at 298. 15 K.1684 THERMODYNAMICS OF TRANSFER complex formation between a carboxylic/peptide group and urea molecules to explain the exothermic enthalpies of transfer for some amino acid/peptide groups has been invoked b e f ~ r e .~ ~ ? ~ ~ The difference in magnitude of AH,, for sorbitol and mannitol can be explained in terms of the difference in their compatibility with the tetrahedral structural order in water as reflected5 in their different heat capacity values in water. Mannitol, due to its planar zigzag conformation in aqueous solution, is expected to fit better into the tetrahedral water structure than sorbitol with its sickle-bent s t r ~ c t u r e . ~ ~ ~ 34 Consequently, the availability of mannitol molecules for hydrogen bonding to urea is less than that of sorbitol molecules. Thus the larger exothermicity observed for sorbitol may result from more extensive interactions of sorbitol and urea molecules. Tables 3 and 4 show that sorbitol has a negative enthalpy of transfer from water to aqueous sodium chloride solution.On the other hand, mannitol has positive values. Furthermore, for sorbitol as well as mannitol, it is seen that AH,, values in aqueous sodium chloride solution are more endothermic/less exothermic than those in aqueous urea or aqueous guanidine hydrochloride solution at the same concentration of cosolute. These observations are consistent with the compatibility concept used to explain the difference in hydration properties of sorbitol and mannitol. Planar mannitol is more compatible with the water structure than non-planar sorbitol. As a result, the hydroxyl groups of mannitol are less free for interactions with cosolute sodium chloride molecules. If the enthalpy of transfer is taken as a combination of the contributions of non-polar -CH, and polar -OH groups, it appears that for mannitol the contribution of non-polar groups24 (positive AH,,) outweighs the contribution of polar hydroxyl groups (negative AH,,), thus resulting in positive enthalpies of transfer.Tables 3 and 4 show that, whereas sorbitol exhibits exothermic behaviour on transferring from water to aqueous guanidine hydrochloride, mannitol exhibits endothermicity. Negative enthalpies of transfer from water to aqueous guanidine hydrochloride are also reported for a polar backbone unit (-CH,CONH-) by Lapanje et a1.21*35 and Stimson and S ~ h r i e r . ~ ~ Fig. 1 and 2 show that the behaviour of polyols in guanidine hydrochloride is intermediate between that of urea and sodium chloride. This is not surprising, since guanidine hydrochloride, in addition to having some of the structural features of urea, also possesses the ionic character of an electrolyte like sodium chloride.The difference in enthalpies of transfer of sorbitol and mannitol can also be explained in terms of their interactions with guanidine hydrochloride molecules and their compatibility with the water structure. HEAT CAPACITIES OF TRANSFER Our results on heat capacities of transfer from water to aqueous urea solutions (tables 3 and 4) show positive ACPtr, both for sorbitol and mannitol. The heat capacity changes accompanying the transfer of various groups (viz. -CONH and -CH,CONH-) having potential hydrogen-bonding sites from water to aqueous urea solution have also been shownz2 to be positive. Ionic solutes like alkali-metal halides also show3’ increased heat capacity values when transformed from water to aqueous urea solutions.However, it has been o b s e r ~ e d ~ ~ . ~ ~ that the heat capacities of hydrophobic solutes decrease on transferring from water to aqueous urea solutions. Polyols behave in a manner similar to ionic hydrophilic groups as far as their heat capacity of transfer is concerned. But positive AC,,, values in the case of sorbitol and mannitol might result from complex formation between urea and polyols, probablyR. V. JASRA A N D J. C. AHLUWALIA 1685 through hydrogen bonding. Similar conclusions were reached in our enthalpy studies. Other workers20>22 have also used the idea of complex formation between urea and -COOH/-CONH groups to explain the positive AC,,, from water to aqueous urea solution.However, heat capacities observed in polyols are much smaller in magnitude than those observed in the case of -COOH or -CONH groups. This reflects the weaker or smaller interactions in the case of polyols. Polyols, according to the specific hydration are strongly/extensively bound with water molecules. Thus their interactions with urea molecules are expected to be smaller than peptide or amino acids which might not have any specific interactions with water. Tables 3 and 4 show that ACptr of sorbitol and mannitol from water to aqueous sodium chloride solutions is almost zero for sorbitol, whereas mannitol exhibits a significantly positive ACptr value. These observations are in agreement with the view that in mannitol the non-polar CH, groups' contribution towards AC,,, outweighs the contribution of the OH groups.However, in sorbitol the situation is different because of the comparatively free availability of hydroxyl groups. In the case of guanidine hydrochloride solutions, due to the small changes it is difficult to draw conclusions from the heat capacity of transfer data alone. However, it is certain that unlike hydrophobic heat capacities do not decrease on transferring polyols from water to aqueous guanidine hydrochloride solutions. This suggests the possibility of the existence of interactions between polyols and guanidine hydrochloride molecules. VOLUMES OF TRANSFER Transfer from water to aqueous urea solution both for sorbitol and mannitol is accompanied by positive volume changes.Sangster et al.39 have also reported an increase in the partial molal volume of sucrose when it is transferred from water to aqueous urea solutions. It has been shown that the transfer of hydroph~bic~~ and ionic solutesgl~ 42 from water to aqueous urea solution also exhibits positive volumes of transfer. The structure-breaking effect of urea in aqueous solution has been used to explain the positive volumes of transfer from water to aqueous urea solutions for hydrophobic as well as ionic solutes. In the case of polyhydroxy compounds like sorbitol and mannitol, it appears, especially in the light of heat capacity and enthalpy data, that interactions of these molecules with urea also contribute to volume changes. It has been pointed out by Franks et al.43 that the partial molal volume of a non-electrolyte is a combination of two factors, uiz.the intrinsic volume of the solute and the volume due to its interactions with solvent. Some investigatorsg4* 45 have approximated the intrinsic volume as where V,, is the van der Waals volumeg6 and Vvoid is the associated void or empty volume. 47 Shahidi et al.44 have modified this equation to find the contribution of one moIecule towards the partial molal volume of a hydrophilic non-electrolyte solute as f'intrinsic = V,w+ Vvoid iP = u,, + vVoid -no, where oS is the shrinkage in volume caused by interactions of a hydrogen-bonding group with water molecules and n is the number of potential hydrogen-bonding sites in a molecule. Thus the partial molal volume of a sugar or polyol molecule can be represented as - V = K w + 6 o i d - Vshrinkage.1686 THERMODYNAMICS OF TRANSFER If one assumes that J&, and Koid have the same magnitudes in water and aqueous urea solution, the positive volume change accompanying the transfer of sorbitol and mannitol might arise from the decrease in Khrinkage in urea solution.Because of the interactions of the polyol hydroxyl groups with urea molecules, the effect of hydroxyl groups on water structure is decreased, thus causing a decrease in Vshrinkage. Furthermore, the interactions of urea with polyols also result in urea having a reduced structure-breaking effect on water. In other words, more water is released as bulk water in the presence of polyol. Since bulk water has higher volume4* contributions than structure-broken water, this factor may also be contributing to the positive volume changes (tables 3 and 4, fig.3 and 4) observed in our study. The higher magnitude of AKr observed in sorbitol might be a reflection of the stronger interaction of sorbitol with urea. Volumes of transfer from water to aqueous sodium chloride/aqueous guanidine hydrochloride solution for sorbitol and mannitol also show positive values. It appears that in these cases, too, the positive AKr results from the decreased effect of cosolute and solute on water structure which arises due to solute-cosolute interactions as mentioned earlier. CONCLUSION The thermodynamic parameters (viz. enthalpy, heat capacity and volume) accom- panying the transfer of sorbitol and mannitol from water to aqueous urea, guanidine hydrochloride and sodium chloride point toward the possibility of interactions between hydroxyl groups or polyols and cosolute molecules.The difference in the magnitude of the thermodynamic transfer functions supporis the view that the planar conformation of mannitol is more compatible with the tetrahedral water structure than the non-planar conformation of sorbitol. F. Franks, D. S. Reid and A. Suggett, J. Solution Chem., 1973, 2, 99. D. P. Wilson and W. Y. Wen, J. Phys. Chem., 1976, 80,431. G. Dipaola and B. Belleau, Can. J. Chem., 1977, 55, 3825. 0. D. Bonner and P. J. Cerutti, J. Chem. Thermodyn., 1976, 8, 105. R. V. Jasra and J. C. Ahluwalia, J. Solution Chem., in press. C. D. Ball, D. T. Hardt and W.J. Duddles, J. Biol. Chem., 1943, 151, 163. ' R. B. Simpson and W. Kauzmann, J. Am. Chem. SOC., 1953, 75, 5139. S. Y. Gerlsma and E. R. Stuur, Int. J. Pept. Protein Res., 1972, 4, 377. J. W. Danovan, J. Sci. Food. Agric., 1977, 28, 571. lo H. Uedaira and H. Uedaira, Bull. Chem. SOC. Jpn, 1980, 53, 2451. l1 S. Subramanian and J. C. Ahluwalia, J . Phys. Chem., 1968, 72, 2525. l2 V. B. Parker, Thermal Properties of Aqueous Uni-Univalent Electrolytes, NSRDS-NBS-2 (U.S. l 3 S. R. Gunn, J. Phys. Chem., 1955, 69, 2902. l4 G. S. Kell, J. Chem. Eng. Data, 1970, 15, 119. l5 Y. Nozaki, in Methods in Enzymology, ed. C. H. W. Hirs and S. N. Timasheff (Academic Press, New l6 J. C. Ahluwalia and J. W. Cobble, J. Am. Chem. SOC., 1964, 86, 5401. l7 T. S. Sarma, R. K. Mohanty and J.C. Ahluwalia, Trans. Faraday SOC., 1969, 65, 2333. l8 C. M. Criss and J. W. Cobble, J. Am. Chem. SOC., 1961, 83, 3233. lB J. C. Ahluwalia and J. W. Cobble, J. Am. Chem. SOC., 1964, 86, 5377. 2o K. P. Prasad and J. C. Ahluwalia, Biopolymers, 1980, 19, 273. 22 D. R. Robinson and W. P. Jencks, J. Am. Chem. SOC., 1965, 87, 2482, 2472. 23 P. K. Nandi and D. R. Robinson, J. Am. Chem. SOC., 1972,94, 1299, 1308. 24 B. Chawla and J. C. Ahluwalia, J. Chem. SOC., Faraday Trans. I , 1973, 69, 434. 25 T. S. Sarma and J. C. Ahluwalia, J. Phys. Chem., 1972, 76, 1366. Government Printing Office, Washington, D.C. 1965). York, 1972), p. 43. S. Lapanje, J. Gkerjank, S. C. Levnik and S. Zibret, J. Chem. Thermodyn., 1978, 10, 425.R. V. JASRA A N D J. C. AHLUWALIA 1687 26 F. Franks and D. L. Clarke, J. Phys. Chem., 1967, 71, 1155. 27 G. C. Kresheck and L. Benjamin, J. Phys. Chem., 1968, 68, 2476. 28 N. Roychoudhury and J. C. Ahluwalia, J. Chem. SOC., Faraday Trans. 1, 1981,77, 3119. 29 J. H. Stern and J. D. Kulluk, J. Phys. Chem., 1969, 73, 2795. 30 Y. Pointud and J. Juillard, J. Chem. SOC., Faraday Trans. 1, 1977, 73, 1048. 31 J. A. A. Ketelaar and B. 0. Loopstra, Rev. Trav. Chem., 1955, 74, 11 3. 32 C. Tanford, J. Am. Chem. SOC., 1962, 84, 4240. 33 D. Horton and M. J. Miller, J. Org. Chem., 1965, 30, 2457. 34 S. A. Angyl and R. Lefur, Carbohydr. Res., 1980, 84, 201. 35 S. Lapanje, B. Crear and G. Trtnik, J. Chem. Eng. Data, 1980, 25, 328. 36 E. R. Stimson and E. E. Schrier, Biopolymers, 1975, 14, 509. 37 N. Desrosiers, G. Perron, J. G. Mathieson, B. E. Conway and J. E. Desnoyers, J. Solution Chem., 38 M. J. Tait, A. Suggett, F. Franks, S. Ablett and P. A. Quickenden, J. Solution Chem., 1972, 1, 131. 39 J. Sangster, T. T. Teng and F. Lenzi, J. Solution Chem., 1980, 5, 575. 40 W. A. Hargraves and G. C. Kresheck, J. Phys. Chem., 1969, 73, 3249. 41 H. L. Wells, M. J. Kustek and R. E. Lindstrom, J. Solution Chem., 1975, 4, 391. 42 H. L. Wells and R. E. Lindstorm, J. Solution Chem., 1976, 5, 155. 43 F. Franks, M. A. Quickenden, D. S. Reid and B. Watson, Trans. Faraday SOC., 1970, 66, 582. 44 F. Shahidi, P. G. Ferrell and J. T. Edwards, J. Solution Chem., 1978, 5, 807. 45 S. Terasawa, H. Itsuki and S. Arakawa, J. Phys. Chem., 1975, 79, 2345. A. Bondi, J. Phys. Chem., 1964, 68, 441. 4 7 A. Bondi, J. Phys. Chem., 1959, 58, 929. 48 J. L. Fortier, P. R. Philip and J. E. Desnoyers, J. Solution Chem., 1974, 3, 523. 1974, 3, 789. (PAPER 1 / 1380)

ISSN:0300-9599    

DOI:10.1039/F19827801677

出版商:RSC    

年代:1982

数据来源: RSC