ENQUERY CWQERNENG WE FEMPERAWRE EEPENDENCE 0E BufiCQNQAW @EEEFERWM ESQTOPE EWEQ‘FS: $©LVOLY$E§ 0F ACE'WLi'.’ CHLORIDE “Mavis gem firm Dogma; 0E ’pk'D. MEEHEGAN STATE UN’EVERSI'EY Thomas AsMfieEvam V > 1968 Jimmy“ Michigan State University This is to certify that the thesis entitled Inquiry Concerning the Temperature Dependence of B-Secondary Deuterium Isotope Effects: Solvolysis of Acetyl-g3 Chloride presented by Thomas Ashlie Evans has been accepted towards fulfillment of the requirements for Ph.D. degree in_Ch_emi_§’t_:_ry ((_ f [[ULQ/ng G. J. Karabatsos Major professor Ikne September 3, 1968 0—169 ABSTRACT INQUIRY CONCERNING THE TEMPERATURE DEPENDENCE OF B-SECONDARY DEUTERIUM ISOTOPE EFFECTS: SOLVOLYSIS OF ACETYL-g3 CHLORIDE BY Thomas Ashlie Evans The temperature dependence of B—secondary deuterium isotope effects can be described by the expression ln kH/kD = —AAH*/RT + AAS*/R Experimentally, isotope effects are observed for which AAF* 2’AAS*, as well as the result predicted by theory, AAF* 2’AAH* (1). The fi-isotope effect for the solvolysis of acetyl—c_i_3 chloride has been determined over a range of solvents and temperatures. The results of Bender and Feng (2), kH/kD ' 1.51 i .07 (90% acetone-water), 1.62 i .08 (80% acetone- water), have not been reproduced. The values obtained are much lower, kH/kD = 1.059 r .002 (90% acetone-water), 1.106 r .003 (80% acetone-water)(3). The difference between the two sets of results appears in the rate of solvolysis of the hydrogen compound. The temperature dependence of the fi—isotope effect has been determined in 80%, 85%, 90%, and 95% acetone—water. In Thomas Ashlie Evans 80% and 85% acetone-water, the influence of temperature is not detectable beyond experimental error. In 90% and 95% acetone—water, the isotope effect is observed to increase with increasing temperature, kH/kD(90%) = 1.026 r .010 (-33.68°), 1.070 r .010 (-9.54°); kH/kD (95%) = 1.004 r .004 (—25.47°), 1.030 r .007 (- .200). The isotope effect was found to increase with increas- ing polarity of the solvent: /kD = 1.004 i .004 (95% kH acetone-water, -25.47°), 1.135 r .020 (75% acetone-water, —31.010). The observed influence of solvent and temperature is consistent with an activation process in which kObs — k . + k lim nucl' CD3 k . + _ ,0 >c=o —-l—l—n—‘-> CD3C=0 + c > c03-c< Cl OH kH kD > 1 0H CD3 k , o \C=O 493% CD3-C-OH > CD3-C// CI/ " \OH Cl kH kD < 1 The sensitivity of the activation process is believed to be the result of changes in the relative contributions of klim and k . nucl Activation parameters have been determined for the solvolysis of acetyl chloride, propionyl chloride, iso— butyryl chloride, and pivaloyl chloride. They do not lead Thomas Ashlie Evans to an unequivocal definition of the activation process. It is concluded that because of the ambiguity in de— fining the transition state of solvolysis, no decision con— cerning the origin of the isotope effect can be made. An examination of the temperature dependence of the fi—isotope effect may be useful in characterizing the behavior of other systems where borderline Snl—Sn2 reactivity is ob— served. (1) L. Hakka, A. Queen, and R. E. Robertson, J. Am. Chem. Soc., s1, 161, (1965). (2) M. L. Bender and M. S. Feng, ibid., 82/ 6318 (1960). (3) C. G. Papaioannou, private communication. INQUIRY CONCERNING THE TEMPERATURE DEPENDENCE OF B-SECONDARY DEUTERIUM ISOTOPE EFFECTS: SOLVOLYSIS OF ACETYL-Q3 CHLORIDE BY Thomas Ashlie Evans A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1968 657043 3'4?“- é? To Laura ii ACKNOWLEDGMENTS The author wishes to express his appreciation to Pro- fessor G. J. Karabatsos for his patient guidance during the course of this investigation. The financial assistance provided by the National Science Foundation, the Petroleum Research Foundation, and a Dow Chemical Summer Fellowship is gratefully acknowledged. The author also wishes to thank Professor Harold Hart for arranging a Special Graduate Research Assistantship during the summer of 1964. The author wishes to thank Dr. C. G.IE$Bioannou for his initial contribution to the work reported here, Dr. G. C. Sonnichsen for computer programs RATE, AKTIV, and HANDS and many valuable discussions, and Mr. Richard Nipe for the use of his personal library. The use of the facilities of the Michigan State University Computer Laboratory is gratefully acknowledged. Finally, the author would like to thank his parents, whose sacrifice and interest in his education have been a source of strength over the last eight years. iii INTRODUCTION . . . . . . . . . . . . . . . . . . EXPERIMENTAL . . . . . . . . . . . . . . . . . Kinetics . . . . . . . . . . . . . . . . . . . A. Preparation of Solvents . . . . . . B. Conductance Apparatus . . . . . . . . C. Conductance Cell . . . . . . . . . . D. Measurement of Time . . . . . . . . . E. Constant Temperature Bath . . . . . . . F. Calibration of the Beckmann Thermometer G. Rate Determinations . . . . . . . . . H. Treatment of Data . . . . . . . . Preparation of Acid Chlorides . . . . . . . . RESULTS AND DISCUSSION . . . . . . . . . . . . . A. Mechanism of Acid Chloride Solvolysis . B. Solvent Influence on the Isotope Effect . . . . . . . . . . . . . . . . C. -Relationship of Activation Parameters and B-isotope Effects . . . . . . . . D. Origin of the fi—isotope Effect . . . . E. Isokinetic Temperatures in B-isotope Effects . . . . . . . . . . . . . . . F. Activation Parameters for the Solvolysis of Other Acid Chlorides . . . . . . . G. Heat Capacities of Activation . . . . . H. Conclusion . . . . . . . . . . . . . . REFERENCES . . . . . . . . . . . . . . . TABLE OF CONTENTS iv Page 30 30 30 31 32 32 32 33 33 34 41 44 58 60 68 80 85 87 93 98 99 Table II. III. IV. VI. VII. VIII. IX. XI. XII. XIII. LIST OF TABLES Page Solvolytic data for t—butyl chloride in 50:50 ethanol—water . . . . . . . . . . . . 5 Rates and isotope effects for l—phenylethyl chlorides corrected to 50% ethanol-water, 500 9 Kinetic isotope effects in some uni—molecular solvolyses of deuterated compounds . . . . 13 Isotope effects as an indication of charge development . . . . . . . . . . . . . . . . 14 Temperature dependence of the B-deuterium isotope effect for the solvolysis of iso- propyl compounds in water . . . . . . . . . 21 Contributions to the secondary deuterium isotope effects for the solbolysis of iso- propyl compounds in water from internal rotation effects . . . . . . . . . . . . . 23 Effect of deuterium substitution on barriers to internal rotation . . . . . . . . . . . 24 Binary solvents . . . . . . . . . . . . . . 31 Run no. 249, acetyl chloride, 75% acetone— water . . . . . . . . . . . . . . . . . . . 38 Run no. 220, acetyl chloride, 95% acetone— water . . . . . . . . . . . . . . . . . . . 39 Acid chlorides and their constants . . . . 42 Rates of solvolysis of acetyl chloride in 90% acetone—water, -20° . . . . . . . . . . 43 B—isotope effect for the solvolysis of acetyl chloride and acetyl—g3 chloride . . . . . . 45 LIST OF TABLES — Continued. Table XIV. XV. XVI. XVII. XVIII. XIX. XXI. XXII. XXIII. XXIV. XXVI. XXVII. Temperature dependence of the B—secondary deuterium isotope effect in the solvolysis of acetyl chloride and acetyl—d3 chloride Theoretical temperature dependence of the B—isotope effect assuming AAF* = AAH* . Rates of solvolysis of acetyl chloride and acetyljgs chloride in 80% acetone-water . . Rates of solvolysis of acetyl chloride and acetyl-d3 chloride in 85% acetone—water . Rates of solvolysis of acetyl chloride and acetyl—g3 chloride in 90% acetone—water Rates of solvolysis of acetyl chloride and acetylfid3 chloride in 95% acetone-water Rates of solvolysis of acetyl chloride and acetyl-g3 chloride in 75% acetone—water Solvent dependence of the B—isotope effect for the solvolysis of acetyl-d3 chloride Influence of solvent on the relative rates of reaction of halogen—substiuted acetyl chloride .. . . . . . . . . . . . . . . Influence of solvent on the activation pa— rameters of acetyl chloride solvolysis, -200 Arrhenius and transition state theory pa— rameters for the solvolysis of acetyl chloride and acetyl—d3 chloride ... . . Transition state theory parameters for the solvolysis of acetyl chloride and acetyl—d3 chloride . . . . . . . . . . . . . . . . . Activation parameters determined from the temperature dependence of the B—isotope ef— fect of acetylfiQS chloride solvolysis Influence of solvent on the activation parameters of solvolysis reactions . . . . vi Page 47 48 50 52 53 56 57 61 67 69 70 71 73 75 LIST OF TABLES - Continued Table XXVIII. XXIX. XXXI. XXXII. XXXIII. XXXIV. Solvent independent rate constants char— acterizing the temperature dependence of benzoyl chloride solvolysis . . . . . . . . Temperature dependence of the isotope ef— fect in the solvolysis of 8—methylfig3—1- naphthoyl chloride . . . . . . . . . . . . Activation parameters for the solvolysis of some aliphatic acid chlorides, -25>. . . Arrhenius and transition state theory pa— rameters for the solvolysis of propionyl chloride, iggfbutyryl chloride, and pivaloyl chloride . . . . . . . . . . . . . . . . . Rates of solvolysis of propionyl chloride, iso-butyryl chloride, and pivaloyl chloride in 80% acetone-water . . . . . . . . . . . Activation parameters for the dissociation of protonated aliphatic esters and acids in HSOSF_SbF5 o o o o o o o o o o o o o o o 0 Activation parameters for the solvolysis of chloroformate esters in water . . . . . . . vii Page 79 82 89 90 91 92 96 LIST OF FIGURES Figure 1 Vibrational energy potential function for HX and DX . . . . . . . . . . . . . . . . . 2 Relationship of force constant changes to zero point energy differences . . . . 3 Theoretical temperature dependence of the B— isotope effect . . . . . . . . . . . . . . . 4 Run no. 249, acetyl chloride, 75% acetone— water . . . . . . . . . . . . . . 5 Run no. 220, acetyl chloride, 95% acetone— water . . . . . . . . . . . . . . . . . . . . 6 Correlation of the rate of solvolysis of some acid chlorides with solvent . . . . . 7 Correlation of the rate of solvolysis of acetyl chloride with solvent . . . . . . viii Page 20 36 37 63 65 INTRODUCTION Kinetic secondary isotope effects arising from B—sub— stitution of deuterium are believed to be the result of vibrational force constant differences between the initial state and transition state of a reaction. There is little agreement about the characteristics which determine the force constant difference. In fact, such is the state of the art of rationalizing the origin of these isotope ef- fects that if the mechanistic implications of such studies are not pleasing, they can be dismissed (1). Theoretical calculations by Wolfsberg and Stern (2) and others (3) have demonstrated that reasonable force constant changes in model reactions lead to isotope effects of the magnitude experimentally observed. The influence of isotopic substitution on the vibrational energy of a mole— cule is illustrated in Figure 1. Isotopic substitution does not alter the potential energy surface, i.e., the force constant for X—H and X—D bonds is the same. Two important conclusions can be drawn from Figure 1: (1) the zero point vibrational energy (ZPE) of X-H is greater than the ZPE of X-D and (2) when the vibration is described by an anharmonic potential function, the mean vibrational amplitude of X-H is greater than the mean Vibrational amplitude of X-D. 1 T E I. “1 X-D 5...... w .ZPE “ I'a H ' H .ZPED J Din r. . . interatomic distance Figure 1. Vibrational energy potential function for HX and DX. ‘ 3 Figure 2 illustrates how the force constant changes are related to the isotopic rate ratio kH/kD. As the force constant increases, the potential function becomes steeper, AAZPE > 0, and kH/kD < 1 (inverse). For a force constant decrease, the potential function becomes shallower, AAZPE < 0, and kH/kD > 1 (normal) . From experimental data it is pos— sible to determine the direction of the force constant change; the problem has been one of rationalizing the direc- tion of the change in the vocabulary of structure-reactivity relationships. Important confirmation of the significance of vibration— al force constant changes has been obtained from vibrational analysis of tfbutyl chloride and tfbutyl—QB chloride, and their corresponding carbonium ions (4). A complete vibra— tional analysis of the chlorides was combined with theoreti— cal calculations for the carbonium ions to obtain an equi- librium isotopic ratio which agreed closely with the experi- mentally observed kinetic isotope effect (5,6). Results from a study of the temperature dependence of the isotope effect in the tfbutyl chloride case also sup— port the assumption that force constant changes are the de? termining factor. The temperature dependence results are shown in Table I. The enthalpy of activation difference (AAH*) is approximately equal to the free energy of activa— tion difference (AAF*), and the entropy of activation dif— ference (AAS*) is negligible. This is expected if the rate ratios are the result of ZPE differences. AAZPE = AZPE - AZPEi > O t kH/kD < 1 X-H AZPE t X-D transition state E AZPEi initial state reaction coordinate Case I. Force constant increase AAZPE < 0 ' 1 kH/kD > E transition state AZ PE 1 ""‘1L—. xan initial state reaction coordinate Case II. Force constant decrease Figure 2. Relationship of force constant changes to zero point energy differences. Table I. Solvolytic dagab ethanol-water ’ . for tfbutyl chloride in 50:50 Temp. k1 x 104 sec_1 k /k 00 k k H D H D 5.686 0.3069 i 0.0005 0.1207 i 0.0005 2.542 10.550 0.6112 i 0.0009 0.2440 i 0.0007 2.505 15.020 1.119 1 0.004 0.4539 i 0.0021 2.465 20.054 2.153 i 0.010 0.8900 i 0.0011 2.424 24.999 4.001 i 0.008 1.676 i 0.002 2.387 30.014 7.310 i 0.001 3.117 t 0.004 2.345 AH*(cal/mole) AS* (eu) AC; (eu) H9 21,220 -2.9 -34 09 21,780 —2.7 —31 AAH* = -570 AAS* = -0.2 aRef. 6 b *= *_ it. *2 *— AAH AHH AHD, AAS ASH AS 6 The role of the solvent has not been defined, but re- sults by Frisone and Thornton (7) indicate that solvent does not affect the isotope effect in the solvolysis of tfbutyl chloride, tfbutyljdg chloride. Robertson and coworkers (6) have also concluded that solvent does not play a role in determining the isotope effect in the Efbutyljdg chloride solvolysis. Assuming that AAZPE E'AAH* S'AAF*, workers have devoted their efforts to rationalizing kH/kD in terms of the ori- gin of the force constant change. Three factors are most often invoked: hyperconjugation, inductive effects, and steric effects. Of these, hyperconjugation is the most widely applied criterion. It is ironic that hyperconjuga— tion, a concept about which there is wide debate, should have such a central role in the rationalization of B—sec— ondary isotope effects (8). Shiner and coworkers are responsible for some of the best work on secondary isotope effects and the contribution of hyperconjugation. Their approach and that of other workers has been to examine the correlation between the pre— dicted contribution of hyperconjugation and the experimental isotope effect. Three characteristics have been examined: (1) variation of the isotope effect as a function of the dihedral angle between the isotopic bond and the reacting bond, (2) transmission of the isotope effect through a F- system, and (3) variation of the isotope effect as a function 7 of electron demand from the reaction center. All three are considered to be unique features of hyperconjugation (9). Shiner and Humphrey demonstrated the angular dependence of the isotOpe effect by solvolysis of compounds Ia - c (10). The isotopic substitution in Ib is at bonds nearly parallel to the developing positive charge, hyperconjugation is ex— pected to be significant, and kH/kD will be large. In compound Ic, isotopic substitution is at a bond in the nodal plane of the developing positive charge, and hyper— conjugation is reduced, which should be reflected in kH/kD. Experimentally, for Ib, kH/kD = 1.14 i .01 and for Ic, kH/kD = 0.986 f .01, supporting the premise that hypercon— jugative interactions are the origin of the force constant change. Ia Ib Ic Transmission of the isotope effect through a v—system has been demonstrated by a number of workers (11). Shiner and Kriz (12) studied the solvolysis of 4—chloro—4—methyl- 2—pentyne (Ila) and its 1,1,1—d3 (IIb) and 4—methylegé- 5,5,5—g3 (IIc) analogs, and observed transmission of the isotope effect through the acetylenic linkage. The isotope 8 effect of IIb, kH/kD = 1.092 f .001, was found to be in- dependent of solvent and can only be due to a hyperconjuga- tive interaction. The isotope effect of IIc, kH/kD = 1.655 f .004, was much larger, as expected, although con- clusions concerning the role of hyperconjugation cannot be drawn. CH3 CH3 CD3 I I I ' CH3-C-CEC-CH3 CH3-C—CEC-CD3 CD3-C-CEC-CH3 I t I Cl ' C1 C1 ' IIa IIb IIC The variation of the isotope effect as a function of electron demand from the reaction site has been used not only as evidence for the contribution of hyperconjugation, but also as a mechanistic probe of the transition state of reactions. Shiner's recent paper (13) illustrates the evi— dence and arguments. The solvolysis of a series of m7 and pfsubstituted a— phenylethyl chlorides was examined. The mechanism of the solvolyses was limiting (Snl) in most cases,' since the large a—secondary deuterium isotope effect, kH/kD 2’1.14, is characteristic of limiting reactions. While the a-effect remained fairly constant over a range of substituents, the B—effect varied significantly. The variation was correlated successfully with the anticipated electron demand of the reaction site. Electron releasing substitutents reduced the positive charge at the reaction site, reducing its electron demand, and lowering kH/kD. Some of Shiner's results are summarized in Table II. Table II. Rates and isotope effects for l-phenylethyl chlorides correcteda to 50% ethanol, 25° Substituent Reiztive a kH/kD 6 kH/kD prethoxy 11.4-7.6 x 105 1.157 1.113 prhenoxy 3.0—2.0 x 103 1.157 1.164 prethyl 59.0 1 157 1.200 pfFluoro 3.0 1.152 1.211 ngethyl 2.0 1 151 1.222 None 1.0 1.153 1.224 meromo 6.6 x 10‘3 1.133 1.221 EfNitro 2.9 x 10"5 1.098 1.151 aSee Ref. 13. The same argument was applied to the unusual solvent dependence of kH/kD of the pfphenoxy compound. In 80% acetone kH/kD = 1.159 while in 93% acetone kH/kD = 1.184. It was argued that in the water-rich solvent, the charge— separated resonance form IIIa contributes substantially be- cause it is stabilized by the polar solvent. In a less polar solvent, IIIb, a tight ion pair, becomes more important, increasing the charge localization and increasing kH/kD. + + C6H5‘O©=CH_CD3 C6H5_O_ Q ‘CH-CD3 Cl" I 01' IIIa IIIb 10 Electron demand considerations have also been used to explain an unusual result in the decomposition of bisulfite addition products (14). In the following example, AAF*/n(-CD3) 3'-20 cal/mole, but AAF*/n(-CD2-) 3’0 cal/mole. The authors argue that ground state hyperconjugation of the methylene with the phenyl group (the electron demand of the phenyl group) blocks hyperconjugation with the reaction site. Given the uncertainty of the contribution of the ground state hyperconjugation effects (8,9), their explanation is somewhat tenuous. 0 0 003 \ ,OH 5; 003 \ 0H “ ;< /\ X —> /\ dDCIHZ SOSH Cp—CHZ CD3 ¢CD2 SO3H ¢CD2 D3 kH/kD 1 . 14 kH/kD 1 .12 The variation of isotope effect with electron demand has been used as a criterion for the existence of non- classical carbonium ions. Two systems in which non-clas- sical ion behavior has been proposed, the gxg-norbornyl and cyclopropylcarbinyl carbonium ions, have been the subject of isotope effect studies. Schaefer and-coworkers (15) reported the results of the solvolysis of norbornyl—3,3—dq bromides IVa and IVb. For the eggg isomer, IVa, a B—isotope effect kH/kD = 1.16 was obtained. This value, which is larger than any pre— viously reported isotopic ratio for a secondary bromide, and the large a-effect, kH/kD ? 1.28, imply that the 11 reaction is limiting, with considerable charge development at the 2-position. The gxg_isomer, IVb, gave kfi/kD = 1.04 by polarimetry and kH/kD = 1.02 by titrimetric analysis. The decrease in kH/kD relative to the endo isomer was IVa IVb attributed to distribution of the charge by o-participation, . i.e., a reduction in electron demand, which reduced the hyperconjugative contribution from the 3—position. Solvolysis of a tertiary norbornyl cation leads to a different conclusion. Sunko and coworkers (16) obtained the results given below from the solvolysis of the chlorides in water. The magnitude of the B-effect is the same for both tertiary ions. There is no evidence for unusual charge distribution in the norbornyl system, as might have been expected. 01 CH3 kH/kD .024 35.000 1.21 t .07 .017 25.000 1.25 i .03 12 Results from similar isotOpe studies in the cyclopropyl- carbinyl system suggest that the cyclopropyl group is cap- able of significantly delocalizing the positive charge of the transition state (17). Table III summarizes the solvol- ysis results for the cyclopropylmethylcarbinyl and methyl— cyclobutyl chlorides. Isotopic ratios are also given for some compounds in which charge delocalization is not exten- sive. The values of kCH3/kCD3 are found to be significant— ly lower in the cyclopropylcarbinyl system, implying de— localization of the positive charge by the cyclopropyl func- tion. While these results may not be conclusive (18), they seem to indicate that o—participation in the norbornyl sys— tem is less effective in stabilizing a positive charge than a methyl group, and both are less effective than the cyclo- propyl function. The charge development criterion has been used to char— acterize the transition state of ketone reduction by sodium borohydride. Lazslo and Welvart (19) determined the extent of ketone reorganization in the transition state of reduc- tion by comparing the B—isotope effect for reduction with the B—isotope effect of tosylate Solvolysis. The data which they provided for comparison are given in the first two columns of Table IV. In terms of hydrization change, tosylate solvolysis (sp3 —> spz) is the reverse of ketone reduction (Sp2 ~> sp3). The authors anticipated that if charge develop— ment was similar in the two processes, kH/kD (solvolysis) 13 Table III. Kinetic isotope effects in some unimolecular solvolyses of deuterated compounds. k /k . . . H D H D Compound Reaction conditions k /k per atom D (corrected) CD3 Cl a D< 50% EtOH, 500 1.09 1.029 CD3?(M83)2 60% EtOH, 250 1.33b 1.100 c1 (CD3)2?‘CH3 60% EtOH. 25o 1.71b 1.102 c1 (003)30-01 60% EtOH, 25° 2.33b 1.103 Me EfBu-CH2-C-CD3 80% EtOH, 25° 1.40C 1.139 c1 ETBu"'CD2'?(Me92 80% EtOH, 25° 1.03‘3'd 1.046 Cl EtCDZCH-CD3 HCOOH, 24.9° 1.73d 1.136 3 OTs [:>_CH~003 96% EtOH, 40o 1.18a 1.057 c1 aRef. 17 bV. J. Shiner, Jr., B. L. Murr, and G. Heineman, J. Amer. Chem. Soc., 35, 2413 (1963). cV. J. Shiner, Jr., and J. G. Jewett, J. Amer. Chem. Soc., §§, 945 (1964). dRef. 29. ,14 .Amomfiv mmmm .mepumq QOHUmQMHumB .mpmfimq .0 tan mummcmw .mm .mH .mmm U .Ammmfiv wwma .Mw ..00m .8050 .Hmaé .h .mmacflm .E .M tam mnmwcsmm .m .30 .AwmmHv mama me 4.00m oEGQU .EAN ..O .HMSNSW om USN .xflmgmm .0 cm HBOOMHJ om om ...HH. .HGflHQsPUHmHflm .4 Q .mom .m .H-H> mHnma .H> .mmno. .wmmfi ..%.z .xuow 302 .mmon pamcom 0&9..mfimflqmn002 mflmmwo>aom .QODCHOSB .m .md m on . o .m h“ +//, \\ m.NHH moH.H 0mH.H ox\mx oun-u\\ 4: o +/ \ / o m 6mm Al. + m . o . E Q ll \ l\a H.mH H 69o H emo H x\ x m 0 mo/ uo/Ilm m/// .m . Q m \ \ em m oo.m mm.H x\ 3 rue 4| 0 O D... m +/ meo\\\ //,m :«oomAmmovaoon tumouAmmovmoou moo\mmo .m\m , -ucwEmoam>mU omnmno mo coHumoHonH cm mm muowmmm mmouomH .>H manna 15 should be the inverse of kH/kD (reduction). Since this is not the case, the authors concluded that there was little bond making or bond breaking in the transition state of re- duction. R R R'/CH OTS > R/C-H lac—0+3?" >R\‘H0/ Recently Geneste and Lamaty (20) have rejected the suitability of the representation of the ketone and carbon- ium ion as similar electronic states. They argued that the carbonyl oxygen would accommodate much of the charge in the transition state, lowering the isotope effect. The sol- volysis of ketals was proposed as a more suitable model. On this basis it was argued that the transition state geom- etry for ketone reduction approaches tetrahedral (Table IV). R\C’I:,’I'I\\\\B/H R/ \‘0"' \H The internal consistency of the arguments and evidence for the participation of hyperconjugation in determining B—isotope effects is quite compelling. It should be pointed out that inductive effects and strain effects can often be applied with equal success and are not without their advo~ cates. 16 For most solvolysis reactions it is impossible to separate the contributions of an inductive effect from that of hyperconjugation. The combination of more electroposi- tive deuterium and greater electron density of the c—D bond is expected to make —CD3 a better electron donor than -CH§ (21). Dipole moment studies (22) and nmr results (23) pro- vide evidence for the enhanced -1 effect of deuterium sub— stitution. The strongest chemical evidence comes from the effect of deuterium substitution on acids and bases. In the dissociation of C6H5-CD2-C02H, kH/kD = 1.12, as ex- pected, if -CD2— has a greater -I effect than -CH2— (24). In the same way C6H5-CD2—NH2 has been shown to be a stronger base (kH/kD = 1.13) than its protium analogue. Values for the dissociation of CD3C02H of kH/kD = 1.06 (24) and 1.03 (25) have been reported. Inductive effects are not widely applied in rationalizing isotope effects. Bartell has proposed that isotope effects may be due to a steric effect of isotopic substitution (26). The C-D bond is considered to be shorter than the C—H bond and thus has a smaller steric requirement. This characteristic is expected to determine the isotope effect. Bartell's quanti- tative argument has recently received qualitative support from Brown (27). In an examination of the isotope effect in the quaternization of methylpyridines and their tri- deuteromethylpyridine analogues, an inverse B—effect was obtained in the case of 2—trideuteromethylpyridine. The values depended upon the iodide used: kH/kD = 0.97 with 17 methyl idoide, kH/kD = 0.96 with ethyl iodide, and kH/kD = 0.935 with isopropyl iodide. For 2,6—bis(trideu- teromethyl)pyridine, kH/kD = 0.92 in its reaction with BF3. -Analysis of 3-trideuteromethylpyridine and 4—tri- deuteromethylpyridine yielded no evidence of an isotope effect. This led Brown to conclude that the isotope ef— fect was steric in origin. In the same discussion Brown pointed out that steric arguments can be applied to the rationalization of B—iso— tope effects with a high degree of internal consistency. Because of this high degree of consistency, he proposed that steric effects were the primary source of secondary isotope effects. Previous work in these laboratories (28) has been de— voted to analysis of steric effects and their contribution to the isotopic rate ratio. The 1,8-disubstituted naphthalene system was chosen because of the significant pgri interac- tions between the 1 and 8 positions and because induc- tive and hyperconjugative interactions between the 1 and 8 positions were expected to be minimal. Bartell's pro- cedure for calculating isotope effects on the basis of non- bonded interactions was used to obtain theoretical isotope effects for the systems studied. Shown below are the sys— tems studied and their assumed transition states. 18 O x a u HO -C -OE t > VI VII The calculated results for the naphthalene system over- estimated the observed isotOpe effects. When these calcu- lations were applied to systems*in which hyperconjugation was possible, they underestimated the observed isotope ef- fect. It was concluded that steric factors were not of major importance in determining the isotOpic rate ratio when hyperconjugation was possible. Brown's proposal notwithstanding, hyperconjugative interactions are probably the major source of force constant changes. However, some of the results from studies of the temperature dependence of the isotope effect are not con- sistent with this view. 19 Expressions describing the temperature dependence of the isotope effect can be obtained from either Arrhenius or transition state theory. From Arrhenius theory: In k = - Ea/RT + In A In kH/kD = - AEa/RT + 1n AH/AD From transition state theory: ln k = KkT/h - AH*/RT + AS*/R ln kH/kD = - AAH*/RT + AAS*/R The temperature dependence is determined by AAH* (E AEa). Graph A of Figure 3 describes the behavior ex- pected of a normal isotope effect. This behavior has been observed in the temperature dependence of the isotope ef- fect in the solvolysis of tfbutyl chloride and Efbutyl- “d9 chloride (6). It has also been observed much less elegantly in the acetolysis of 2—pentyl tosylate by Lewis and Boozer (29). In both results, there is no apparent con- tribution from AAS*. - In other cases, however, AAS* has been observed to have a significant, if not dominant, role in determining the isotope effect. The extreme case is described in graph B of Figure 3, in which AAH* = O and the isotope effect, now independent of temperature, is determined entirely by AAS*.A This behavior was observed by Leffek, Robertson, and Sugamori (30) for the solvolysis of some isopropyl compounds. Their results are summarized in Table V. 20 Figure 3. Theoretical temperature dependence of the B-isotope effect. kH/kD > 1 AAF* E’AAH* m = —AAH*/R l T Graph A kH/kn I 1 AAF* Q'AAS* + m : ~AAH*/R Q i m 1 .540 T a H Graph B 21 Table V. Temperature dependence of the B—deuterium isotope effect for the solvolysis of isopropyl compounds in water. Compound: Isopropyl Isopropyl Isopropyl Methanesulfonate Toluenesulfonate Bromide ngp. kH/kD ngp. kH/kD Tgmp. kH/kD C 30.001 1.547 30.017 1.545 69.994 1.324 25.000 1.551 30.001 1.548 69.993 1.317 20.002 1.551 25.005 1.539 65.002 1.322 12.514 1.547 24.999 1.542 60.000 1.318 12.501 1.540 20.005 1.543 59.933 1.315 5.000 1.555 15.003 1.545 55.005 1.313 10.003 1.537 40.005 1.312 6.008 1.542 40.005 1.317 *_ * *_ * Compound (AHD AHH) (ASD ASH) i std. error i std. error cal/mole cal/mole deg. Isopropyl methanesulfonate 7 i 28 -0.84 i 0.1 Isopropyl‘p—toluenesulfonate -21 f 14 —0.93 i 0.05 Isopropyl bromide -35 f 15 —0.65 i 0.05 aRef. 30. There have been two explanations proposed for the temperature independence of this isotope effect, one by the authors and one by Halevi (21). satisfying. Neither is particularly 22 The authors proposed that the temperature behavior was the result of the influence of deuterium on the rotational barriers of the substrate. Their analysis depends upon the validity of two assumptions: (1) there is a decrease in rotational barrier on going from a tetrahedral ground state to a carbonium ion transition state, and (2) the barrier to rotation of the methyl groups is higher for -CH3 than for -CD3. -Assumption (1) is reasonable. It is well known that"_ six—fold barriers are generally lower than threeefold bar- riers (31). The second assumption, even more important to the analysis, is much less acceptable. In Table VI are listed the results of calculations (32) used to justify the AAS* term as a rotational effect. Note' that the influence of reducing the barrier in the transi- tion state is to introduce a AAH* term opposite in sign to that expected from hyperconjugation. Only when a ground state difference of 600 cal/mole (20%) is added does a sig- nificant entropy difference appear. V Is a value of 600 cal/mole justified? Table VII is a compilation of barriers of rotation in which the influence of deuterium substitution has beenidetermined. No values - for rotation about an sp3—sp3 bond have been obtained (33). So little is known about the source of rotational barriers that it is difficult to see how the small barriers obtained in the sp3-sp2 cases can be extrapolated to the 600 cal/mole required of an sp3—sp3 bond with any real confidence. 23 .om .m0m m om.ol 0v: ooNH oowm ooom om.on Hm: ooo oovm ooom oN.o| mm: o oowm ooom vo.o mHI ooma ooom ooom no.0 om: ooo ooom ooom oo;o mm: o ooom ooom A00Hm0o 0HoE\Hmov A0HOE\H00V .momo zuom .Umo Q .Umu m msoum H>£u0E H0O -msoum Hh£u0e H0O 0umum COHDHmeHB 00000 HmHuHQH AWmQ I mmdv Ammd I mmflv A0HOE\Hmov COHumDOH H>£u0fi on H0HHHmm m.muo0mm0 :oHumuou Hmcu0DCH Eonm H0003 CH mocsomaoo ammonmomfl mo mHmmHo>Hom 0:0 How muo0mm0 0mouomH EDHH0DS0U mumocoomm 0:0 00 mCOHDDQHHucoo . H> 0HQMB 24 Table VII. Effect of deuterium substitution on barriers to internal rotation. Compound Reference Barrier to Methyl Rotation (cal/mole) 4&0 CH3CH2C\\ 2,254 i 30 H a ¢O CH3CD2‘C\\ 2,317 i 30 H 0 003—065 1,143 i 30 H b 0 cna—cf: 1,151 i 30 ‘D O / . CH3-C<: 1,162 i 30 H b //0 003-c<: 1,143 i 30 H 0 CHa-C-Cfia 778 i 30 c CDs-C-CDa 726 i 30 0 as. S. Butcher and E. B. Wilson, Jr., J. Chem. Phys., 40, 1671 (1961) ’“‘ bR. W. Kilb, C. C. Linn, and E. B. Wilson, Jr., J. Chem. Phys., 26, 1695 (1957). CR. Nelson and L. Pierce, J. Mol. Spect., 18/ 344 (1965). 25 Leffek and coworkers rejected the possible role of sol— vent on the basis of results of the solvolysis of some ethyl and isopropyl compounds (34). Their results indicated that there was no significant solvent reorganization on changing the substrate from ethyl to isopropyl. They also failed to observe a significant change in solvent isotope effect as a function of substrate. It was difficult to justify a sol- vent effect resulting from deuterium substitution which could not be observed from methyl substitution. It is Halevi's opinion (21), however, that differential solvation effects can be used to rationalize the unusual temperature behavior described above. He proposes that electron delocalization is more extensive in the deuterium compound than in the protium compound. With this enabling assumption he applies generally accepted solvation theory in a straightforward fashion. Solvation increases with increasing concentration of charge. Hence in polar solvents the hydrogen compound is more effectively solvated in the transition state than the deuterium compound. This introduces a AAH* term which augments the effect of hyperconjugation. The AAS* would favor the less-solvated transition state (deuterium). In water-rich solvents the gain in solvation energy is offset by the entropy gain in breaking the quasi—crystalline state of water. Halevi argues that this is what is occurring in the case of the isopropyl compounds. The energy required to. 26 break the additional hydrogen bonds to solvate the protium compound Opposes the effect of hyperconjugation, with re— sulting AAH* 2'0. Since there are more hydrogen bonds brdken in solvating the protium compound, AAS* favors the protium compound. Certainly Halevi's rationalization coincides with the experimental results. However, there is little reason to support his feeling that the charge distribution in the pro— tium and deuterium compounds will be sufficiently different to give the result he proposes. Halevi's results from an examination of the temperature dependence of kH/kD for the basic hydrolysis of ethyl ace- tate and ethyl trideutero acetate (35) also argue against an interpretation in terms of zero point energy differences. The results, which are given below, suggest that the value of kH/kD at 25° describes a minimum in temperature dependence behavior. Halevi argues that since (E: — E3) = —RT2(dln kH/kD/dT), (E: - E2) must be nearly zero at 25° and the observed isotope effect is determined by an entropy term due to differential solvation. Temp. 0° 25° 35° 65° kH/kD 1.00:0.01 0.90i0.01 0.93i0.01 1.15:0.09 Some time ago Shiner (36) reported the solvolysis of some tertiary halides and concluded that solvent interactions were responsible for some of the isotope effect. Examination of B—isotope effect in the solvOlysis of 2,3—dimethyl—34g72- chlorobutane let to a free energy difference of-147 cal/mole but an activation enthalpy difference of’580 cal/mole. 27 Since elimination products were observed, Shiner pro— posed that part of the isotope effect arose from specific interaction with the B-isotopic bond. He seemed unwilling to suggest that the entire weakening of the B—C—H (-D) bond could be due to intramolecular effects. Work since then has shown that an elimination—type solvent interaction is not necessary. Lewis and Coppinger (11a) examined the solvolysis of some Efsubstituted phenyhmdiwl carbinyl chlorides in which elimination products were not observed. The isotopic ef— fects obtained were comparable to those observed by Shiner. The free energy of activation difference was found to vary with temperature, indicating a contribution from the entropy term. An entropy contribution was also observed in the con- formational kinetic isotope effect reported by Mislow and coworkers (37). The isotopic free energy difference was found to be determined by the entropy term. Although AAH* for the racemization of VIII was large, about 0.2 kcal/mole, it was attenuated by an opposite entr0py effect which re— duced AAF* to one third to one half of the value of AAH*. It was proposed that ca. 50 cal/deuterium could result from steric effects. The authors concluded in a qualitative way that steric factors did not contribute a major portion of the isotOpe effect in solvolysis reactions. 28 CH3 CH3 VIII The entropy again makes a substantial contribution to the isotope effect in the solvolysis of pfmethyl and pr methyl—g3—benzhydryl chloride (11c). It was found that AAEW was only about 10% of AEa. Although these results seem to invalidate the ZPE approach, this is not necessarily the case. Wolfsberg and Stern (2) have shown that it is possible to rationalize temperature independent isotope effects in terms of force constant changes. Their work has shown that if force con— stants giving rise to small frequencies (torsional) in the initial state increase in the transition state and force constants which give rise to large frequencies (stretching) decrease on going to the transition state, no temperature dependence will be observed. It would appear that the correct interpretation of B—isotope effects rests not only on a correct rationaliza— tion of the origin of force constant change, but also on ac— curate identification of the number of force constants changing, i.e. an accurate definition of the activation process. 29 The work presented here is a study of the temperature dependence of the B—secondary deuterium isotope effect in the solvolysis of acetyl chloride and acetyl-d3 chloride. It represents a continuation of efforts in these labora— tories to elucidate the origin of B-isotope effects. EXPERIMENTAL I. .Kinetics A. Preparation of Solvents Conductivity7Water: Conductivity water was prepared by passing distilled water through a 5 x 80 cm column con- taining alternate layers of "Baker Analyzed" reagent Dowex 1-X8 (anion exchange resin) and "Baker Analyzed" reagent uDowex 50W-X8 (cation exchange resin). The Dowex 1-X8 was converted to the hydroxide form by first treating it with concentrated potassium hydroxide solution, followed by several washings with distilled water. *Water treated in this manner had a specific conductance 6 of about 2 x 10- mho/cm. The quality of the conductance water was checked before each preparation of mixed solvent. Conductivity Acetone: The Conant-Kirner method (38) was used to prepare solvent acetone. 'Acetone (Fisher Certi— fied A-18) was distilled from basic potassium permanganate as described by Papaioannou (28c). The acetone obtained in 8 this manner had a specific conductance of about 1 x 10- mho/cm. Mixed Solvents: Mixtures of acetone and water were prepared according to the weight ratios determined from the densities of acetone and water and the desired volume/vol- ume ratios. The densities and ratios used are given in Table VIII. 30 31 Table VIII. Binary solvents v/v (% acetone— 95% 90% 85% 80% 75% water w/w 93.72 87.73 81.73 76.35 70.24 Density of acetone 25°, 0.7844 g/mla Density of water 25° , 0.977044 g/mlb aInternational Critical Tables, E. W. Washburn, ed., McGraw— Hill, New York, 1928, Vol. 3, p. 33. bHandbook of Chemistry and Physics, 42nd ed., Chemical Rubber Publishing Co., Cleveland, Ohio, p. 2142. A Torbal balance (Model PL-12, Torsion Balance Co.) with a capacity of 2000 g and a stated accuracy of 0.1 g was used. It was found that the accuracy of the balance needed to be checked at intervals. Solvents were prepared in batches of slightly less than 2 liters. The water portion was always weighed first. For the preparation of 93.72% acetone—water, the water portion was weighed on a Mettler balance (Model BS, Mettler Instru- ment Co.) accurate to i 0.1 mg. Solvents prepared in this manner gave individual rates which were reproducible to 1% in all cases except 93.72% acetone—water, where the deviation was slightly larger. The reproducibility of the isotope effects was generally better than 1% in all solvents. All solvent preparations were recorded in a log book designed specifically for that purpose. B. Conductance Apparatus A Wayne-Kerr conductance bridge (Model B221, Wayne—Kerr Co., Ltd.) was used for all conductance measurements. Readings 32 of both conductance and capacitance were accurate to 0.1%. For all runs No. 239 and above the bridge was equipped with a WayneeKerr Autobalance Adapter (Model AA221) with a P8109 power supply unit. vAddition of the automatic balancing function allowed the determination of faster rates and was compatible with the future use of digital voltmeters and recorders. See Wayne-Kerr Publication TP 214. C. Conductance Cell The conductance cell used is described by Papaioannou - (28c). The cell was stored with spent reaction solvent to avoid adsorption of ions during a kinetic run. The cell was rinsed four or five times with conductance water and twice with conductance acetone before use. D. Measurement of Time A Precision Scientific electric digital timer (no. 69237, Precision Scientific Co.) accurate to 1/100th of a minute was used. E. Constant Temperature Bath The bath used is described by Papaioannou (28c). A new electronic relay was used (RHOtron Co.). A Beckmann Dif— ferential Thermometer (Arthur H. Thomas Co.) was used to monitor the bath temperature. Temperatures constant to 10.0030 were obtained over a temperature range from -35° to 0°. 33 The bath was equipped with a submersible magnetic stir- rer (Model 700, Henry Troemner, Inc.) which was mounted to allow its raising and lowering underneath the cell. F. Calibration of the Beckmann Thermometer The Beckmann differential thermometer was calibrated by means of a platinum resistance thermometer (No. 8163, Leeds & Northrup Co.). The measuring circuit consisted of a bat— tery power supply (2mA), a Mueller resistance bridge (No. 8076, Type G—1, Leeds & Northrup Co.), and a Pye galvanometer (No. 29—115, Ealing Corp.). The bridge accomodates three of the four thermometer leads. Eight resistance readings were taken over a range of at least 0.0500, the leads were reversed, and eight more readings were taken. The average resistance reading obtained in this way was converted to temperature by solution of the Callendar—Van Dusen equation (39) using the Newton—Raphson method (40). The precision of the calibration was i 0.002). How— ever, the accuracy of the method is not considered to be better than i 0.01° (41). All calibrations were entered in a log book designed for that purpose. G. Rate Determinations All rate determinations were numbered consecutively from 100 and were recorded in a log book designed for that purpose. Hydrogen and deuterium runs were performed alter— nately. 34 Approximately 200 ml of solvent was added to the rinsed conductance cell containing a Teflon stirring bar. The cell was placed in the constant temperature bath in an upright position and allowed to equilibrate with stirring. .Solvent conductance became constant after 15 to 30 minutes, but an hour was allowed for complete equilibration. The solvent conductance, which was generally less than 0.1 x 10-6 mho, was recorded. It did not remain constant from run to run, but was found to vary somewhat. No correla— tion between the initial solvent conductance and the rate was observed. A drop of the appropriate acid chloride was added, the timer was started, and stirring was continued for at least .30 minutes to assure homogeneity. After mixing was complete, the stirring motor was lowered and the cell was tilted so that solvent completely filled the electrode chamber. If this was not done, considerable noise was observed in the measuring circuit. The initial reading was taken when the capacitance had become fairly constant, 22° 1.1 x 10"10 F., after about one- quarter reaction. From.18-25 conductance readings were taken over the first two half-lives. The infinite value was obtained after 13 half-lives. 6 mho. The range of conductance read was 25 to 100 x 10- H. Treatment of Data First order rate constants were obtained by a least 35 squares solution of the integrated first order rate expres— sion using PROGRAM RATE. ln (Ca) - Ct) = —kt + lnCOO Co) = conductance at infinite time Ct = conductance time at time t. Conductance was taken as directly proportional to the con— centration of hydrochloric acid formed. This is not entirely accurate. A noticeable curvature was observed in all sol— vents. The resulting deviation from limiting behavior ranged from 0.06% in 75% acetone—water to 0.8% in 95% ace— tone—water and were not temperature dependent. Since we are interested in the temperature dependence of the rate constant, it was felt that the observed devia— tions would not lead to appreciable error as long as the reaction was observed over the same region at all tempera- tures. Typical runs are graphed in Figures 4 and 5. The out— put from RATE for the same runs is listed in Tables IX and X. Average rates are reported as the mean of at least three independent rate determinations. The uncertainty indicated is the standard error 0. /2 Q ll _ 1 ; [(Xi - Xi>/n] l X. 1 observed rate Xi: mean. Rate ratios, kH/kD, are the ratios of the corresponding means. The uncertainty indicated is the standard error 0 36 - t, min. Figure 4. Run No. 249 acetyl chloride, 75% acetone-water. 37 J3- J74 4.) U I .6 cut (3 U U) 0 H \Np \ \\° \ o)... 95% acetone—water. Figure 5. Run No. 220 acetyl chloride 38 Table IX. Run No. 249 -- Acetyl chloride -—75% acetone/water Time Conductance Calculated K R Conductance at Infinity = 6.46740—005 2.27 2.67000-005 3.71356-003 -7.18726—003 2.86 3.15000-005 3.73498-003 -5.38018—003 3.30 3.47000-005 3.74929-003 -3.37528-003 3.79 3.79000-005 3.76103-003 —1.20641-003 4.29 4,08000-005 3.76806—003 4.44855-004 4.86 4.37000-005 3.77025—003 1.14208-003 5.34 4.59000—005 3.77720—003 3.48292—003 5.84 4.79000—005 3.77528-003 3.13579-003 6.61 5.06000-005 3.77801-003 4.63213-003 7.14 5.22000-005 3.77928-003 5.54538-003 7.79 5.39000-005 3.77739-003 5.16903-003 8.41 5.53000-005 3.77477-003 4.35736-003 9.02 5.65000-005 3.77260—003 3.39107—003 9.60 5.75000-005 3.77123—003 2.81768—003 10.14 5.83000—005 3.76473-003 -9.75361-004 10.73 5.91000-005 3.76604—003 —1.89274-004 11.63 6.01000-005 3.75795—003 —5.84644-003 12.27 6.07000-005 3.7594-003 -9.85869-003 Rate Constant = 3.76633—033 Std. Dev. of K = 5.82886—006 39 Table X. Run No. 220 —— Acetyl chloride --95% acetone/water Time Conductance Calculated K R Conductance at Infinity = 1.09130-004 3.57 3.13000-005 1.04686-003 ~2.37454—002 4.07 3.44000-005 1.08470—003 —1.78317-002 4.78 3.86000—005 1.12627-003 -9.30713-003 5.53 4.26000-005 1.14862—003 —3.01940-003 6.33 4.65000-005 1.16251-003 1.81857—003 6.95 4.93000—005 1.16848-003 4.48853-003 7.79 5.29000—005 1.17526-003 8.19618-003 8.53 5.58000-005 1.17676-003 9.74476-003 9.26 5.85000—005 1.17750—003 1.09913—002 10.07 6.13000-005 1.17695-003 1.16173-002 10.80 6.37000-005 1.17684v003 1.23895—002 11.53 6.59000-005 1.17408-003 1.13193—002 12.44 6.85000—005 1.17130—003 1.01357—002 13.16 7.04000-005 1.16787-003 8.01437-003 13.80 7.20000—005 1.16466—003 5.74711—003 14.58 7.39000—005 1.16240-003 4.09304-003 15.28 7.55000—005 1.15985-003 T1.94822—003 16.03 7.71000-005 1.15626-003 —1.40367-003 16.87 7.88000—005 1.15257—003 —5.21717-003 17.50 8.00000-005 1.14952-003 —8.61028-003 18.32 8.15000-005 1.14616-003 —1.27037-002 19.32 8.32000—005 1.14162-003 -1.86654-002 Rate Constant = 1.15772-003 Std. Dev. of K = 7.99750-006 40 obtained from the expression (42) 1/2 2 2 + 2 kH OD/ks] 2 OkH/kD = [OH/kD Three computer programs were used to analyze the temper— ature dependence of the rate constant. The most SOphisti— cated and the one whose values are used in the Discussion was ACTENG, written by Prof. Delos DeTar and modified to run on the CDC 3600 (43). ACTENG calculates both Arrhenius parameters (ln k.y§ 1/T) and transition state theory param- eters (ln k/T y§_1/T) by an iterative least squares method. ACTENG was also used to obtain solutions of the relationship In kH/kD = —AAH*/RT + AAS*/R . The parameters AAH* and AAS* were obtained from the "Arrhenius" parameters by the following treatment: AAH* = AEobs obs Ecalculated A : AH/AD = exp (AAS*/4.576). calculated AKTIV calculates activation of parameters by a single least squares solution of ln k T'yg l/T. HANDS calculates AAH* and AAS* by a single least squares solution of In kH/kD y§_1/T. AKTIV, HANDS, and RATE were written by Dr. George Sonnichsen. Activation parameters for a particular temperature were calculated from the equations: 41 *= _4 AH Ea RT AS* 4.576log (A/T) - 49.203. All manual calculations were performed on a Friden electronic calculator (Model 130). II. Preparation of Acid Chlorides All aliphatic acid chlorides were prepared by the method of Brown (44). The general procedure is given below. To a round bottom flask equipped with a vigeaux fraction— ating column and standard distillation head was added 0.15 mole of the acid and 0.225 mole of benzoyl chloride. The reaction mixture was heated and the acid chloride was dis— tilled into an ice—cooled receiver as quickly as possible. The acid chloride boiled somewhat low due to dissolved hydrochloric acid. The distillate was refluxed for 30 minutes to purge the hydrochloric acid. It was distilled through an 8 cm vacuum— jacketed glass spiral column. The distillation was repeated a second time. Overall yield was 60—707. The purity of the acid chloride was checked by compari- son of the infrared spettrum with that appearing in the Sadtler Index (45) and by nmr. Nmr was used to detect the presence of benzoyl chloride, which appears as a multiplet at 1.9 and 2.4T. The acid chlorides used are listed in Table XI. The source of the acid chloride did not influence the rate as shown by the results in Table XII. Distillation of 42 . . 0 mm m mo .wv .60 .m0m .COHummeonumo0o m9 «Ammouvmno Eoum o0nmm0um0 .mooH .XHOM 302 .mm0Hm wuflmu0>HQD Unowxo ..U0 .Q0>0Dm .m tam Mooaaom .4 .m .o .mocdomfiou UHcmmHO mo NHMGOHDUHQ Q .56 .Homm .Hdm um0£mH£ wmmvm e.we Ame ocnmoH ceoH Hson>Ho .nocmHm oHom oHHm>Ho c.8e 1H.oomv . . . .DH00 UHom e I mmmfiH E w Am UV Qon oom H%H>DDQ OmH .Hmfimflm UHHhDSQIOmH H.ne AN.UV .DH0U neon econ m.we Am no necm ewe HscoHooHo .HocmHm oHconoHo fil III: «In . . t 0H00 6 oon ch o Hmuooe oncsoz m m 0H000< ud0000H oHom Uflumom omoc n.8e Amv ammuHh eHh Hmuooa .Hoana HcHomHo .m.H Aum0cv .m.n .m.Q 0UHH0H£O 00H50w teem 0H0Huomm MEG .DHH oflom Dn0umm .mucmumcoo HH0£D 0cm m0©HHoH£0 Uflom. .Hx 0HQMB 43 acetyl chloride from N,N—dimethylaniline or degassing acetyl chloride with dry nitrogen did not affect the rate of sol- volysis. All acid chlorides were found to be as kinetically pure as acetyl chloride. Table XII. Rates of solvolysis of acetyl chloride in 90% aqueous acetone at —200 .Nufiggr k X 103 i X 105 Commercial acetyl chloride 277 2.342 2.362i.021 280 2.383 Acetyl chloride prep. 1 3;: 3:32: 2.367i.008 Acetyl chloride prep. 2 3:3 g:ggg 2.3723.013 The extent of deuteration in CD3C02D was determined by integration of the nmr methyl resonance against the resonance of a known amount of hexamethylbenzene. It was determined to be 98% deuterated. The acid chlorides were stored in glass vials with polyethylene caps which were sealed with "Parafilm" plastic. RESULTS AND DISCUSSION The B-secondary deuterium isotope effect for the sol— volysis of acetyl chloride and acetylfig3 chloride was re- ported by Bender and Feng (46) in 1960. ATheir results are given in Table XIII. The large normal isotope effect was interpreted as arising from force constant changes due to hyperconjugative interactions in an acylium ion intermediate. Papaioannou (49), in initial work in these laboratories, observed a much smaller normal isotope effect. His results are also given in Table XIII. The rate of solvolysis of the hydrogen compound accounts for the difference in the two sets of data. Bender and Feng report using commercial reagent grade acetyl chloride with- out further purification. Perhaps this influenced the rate of solvolysis of the hydrogen compound in some way. Papaioannou's results are more in agreement with the anticipated B-isotope effect. The influence of deuterium substitution should be less in an acylium ion, in which charge localization has been reduced by resonance (IX <—> X), than in a tertiary carbonium ion, where delocalization does not occur. + CD3—C=0 <——> CD3-CEO IX X 44 45 .mv .M0m 3 0H ommmm H Mom I moo.oaooH.H mH.ono.o¢ oH.oHvo.wv ow Ho.NNI Qsocconmmmm H How I moo.OHomo.H ooo.ohmam.o HHo.oHoow.o oo Ho.NNI Qsoccmonmmm mmaovml wo.oumo.a o.NHo.om o.mHo.mo ow o.NNI mmc0m out H0on0m mNHmoNI no.oHHm.H m.onw.o m.o«o.OH oo o.NNI 0mC0m out H0oc0m m m , no mo A0HOE\Hmov Q m M. M AH0003 m<< M\ M I0cou0om RV Uo * I000 voH x M DG0>Hom .QE0B 00Hdom H .mmramu0om 0:0 0UHHOH£0 Hmu0om mo mHm>Ho>Hom .ocHHoHno 0:0 How 000HH0 moouomHIm .HHHx 0Hnme 46 Ifirom Bender's work, AAF*/n = —70-80 cal/mole, which is stenitially larger than the value observed for similar stiinition for a tertiary carbonium ion, AAF*/n = — 50-60 ,hnole. The work reported here indicates AAF*/n = ~7—18 /mole. The temperature dependence of the B—isotope effect is sual. The results summarized in Table XIV show the iso- Ie effect to be constant within experimental error in 80% . 85%;acetone-water. However, in 90% and 95% acetone- ;er the isotope effect increases with increasing tempera— :e. The calculated temperature dependence of the B-isotope Eect, based on the assumption that AAF* E'AAH*, is given r each solvent in Table XV. The experimental results in % acetone—water are significantly different from the cal- lated isotope effect. The validity of the results in 95% etone—water is less clear. It seems reasonable to con— der the isotOpe effect to be increasing with increasing mperature. Theory based on zero point energy differences predicts at the temperature dependence should fit the relationship ln kH/kD = - AAH*/RT + AAS*/R maentropy term, AAS*, is expected to have little influence i'umaobserved isotope effect. This requires that a normal adage effect (kH/kD > 1) should be inversely proportional D'Ummerature, i.e., it should decrease with increasing emperature . 47 .e XIV. Temperature dependence of the B-secondary deu— terium isotope effect in the solvolysis of acetyl chloride and acetylfg3-chloridef1 I Solventb Temp. k /k AAFIC. Icetone—water) °C _H D (cal/mole) 80 —21.18 1.119i0.017 -56i8 —22.01d 1.106i0.006 —50i3 -26.37d 1.114i0.006 ~53i3 -28.88d 1.105i0.043 —48i27 —31.92d 1.130i0.021 -58i13 —34.92 1.108i0.008 -48i3 85 -21.18 1.101i0.004 —48i2 -25.47 1.109i0.010 ~51i8 —31.00 1.112i0.013 -51i6 90 - 9.54 1.070i0.01 —35i5 —15.72d 1.072i0.015 —35i7 —22.01d 1.059i0.002 -28i1 -28.88d 1.044i0.009 —21i5 -33.68d 1.026i0.010 ~12i5 95 - .20 1.030i0.007 -16i4 - .20 1.026i0.011 —14i5 — 5.42 1.016i0.006 - 8i3 — 9.55 1.018i0.004 - 9i2 —15.51 1.018i0.004 — 912 -22.62 1.008i0.001 — 4:1 -25.47 1.004i0.004 — 2i2 r___._. kD ratios are the ratios of the average rates obtained from the individual rates listed in Tables XVI, XVII, —XVIII,'XIX. vents were prepared as the W/W ratios, but are reported the corresponding V/V ratio, see Table VIII. orrected, acetyl-g3 chloride contained 98% deuterium. ues determined by C. G. Papaioannou. 48 .ox\mx ooHemcmom.mI n *moom mHH.H on.H who.H woo.oaon.H cHo.onHH.H oHo.oHomo.H oHH.H moH.H who.H eHo.oHoHH.H ooo.oaHoH.H oHotoHoso.H mm- on. 8.6m- ow mm oo o m eHo.H Aonov x\ x Q m 0H5D0H0QE0B 30A eoo.ohcoo.H Amnov x\ x m hHo.H Aonov ox\ x Q m 0H5D0H0QE0B LmHm soo.oHomo.H Amnov x\ x . I A0HOE\H00V N w *mQQ AH0D03I0GOD000 RV mo . un0>aom 3342:3000 JDD44D DLDJDD4IQ DHHJ ...) DJHHDHJHHDLD.) D43JU4DL§DJ 4UD4JD4DDHHH m *moq u *aoq 0 >4 ”4.2.0-4.. 49 In the discussion which follows, we will examine the dence available for defining the mode of reaction of d chlorides and its relationship to the isotope effect its temperature dependence. 50 oow.H ON NON.H oH OHo.H wH OOO.HmOb.H mmb.H SH Nmo.H OH moo.H mH mwo.H vH ONO.H MH bmflwwl mvo.HmOH.H ww.mNI OHO.HHmo.H who.H NH Nmm.m HH Noo.Homm.m mum.m OH boo.m o Nmo.m w «HomI ooo.HwHH.H hm.ONI mHo.HHmo.N mmo.m b mwo.v o omo.v m mHO.Hmoo.v vmo.v v «Hm.w m bow.v N mHomI OOO.HOOH.H HO.NNI OHO.H¢ov.v Hom.v H mmv.v oom va.v vom ooo.H>h¢.w me.w Non OHO.m mom, vmo.m mom wHomI bH.OHOHH.H wH.HmI wHo.Hooo.m moo.v Hom 8no «mo 0HOE\HmU QM\mM 0o HI00m M M H0QESZ *m<< , .mEOB M"OH x M 000 M"OH x M cum. 51 Hme. 5mm OOHw. mom omHo. mom. boom. omm woo.HOOH.H No.6ml vHHo. vmm Nom.H OHm «mm.H_ me mom.H mHm omm.H me NNm.H OHm memI OOO.HNNH.H No.HmI boo.HHmm.H mHm.H va mmN.H om me.H om ONO.HOON.H ooN.H mm omw.H 5N oo¢.H on mHmeI HNO.HomH.H Nm.HmI boo.Hmmv.H mmv.H mm moo.H 6N wow.H mm wNm.H mm HHw.H Hm 8no «mo 0HoE\Hmo QM\mM 0o HIo0m M M H0QE§Z *m<< .mE0B «OH x w HIo0m «OH x M com 52 wmwo. omm omvo. mum ONOO.Hova. mmvo. wmm finch. bNm womb. mum QHHmI MHO.HNHH.H oo.HmI owo.HN®Hb. meb. Nmm me.H omm bvN.H mom moo.HHmN.H HmN.H mmm owm.H 5mm .. _ vwm.H omm wHHmI OHO.H®0H.H bv.mml boo.wam.H ©®M.H vmm omo.N NHm wNO.N OHm bOO.H¢NO.N mHo.N mom. mNN.N HHm NmN.N mom NHwVI VWO.HHOH.H wH.HNI NNO.H®NN.N bNN.N how 000 «no «mo 0HOE\Hmo QM\mM 0o HI x M M HOQESZ *m<< .mE0B «OH w HIo0m «OH x M com new cH ooHHoHro .H0u03I0cou0om «WIHhu0om new 0OHHOHSU Hmu0om mo «HmMHo>Hom mo mmumm . HH>N .0HQMB 53 mNm.o NH mom.o NH ooo.HmHm.o HHm.O HH HHo.o OH ONOuo o HHwNI Noo.Hono.H HO.NNI HHO.Howo.o www.o w oo.NH b oo.NH o OHO.HHO.mH NN.MH m Hew.mH H. mw.mH m . HH.¢H N NHOMI mHo.HNhO.H Nb.mHI OHO.Hvo.mH wo.mH H oo.vN OON _ No.mN NON ONO.Hvo.wN mH.VN OON wo.mN NON mb.mN HON mHmmI OHO.HOSO.H mm.o I Hmo.HNh.mN. mw.mN ooH m .. m 0HOE\H00 QM mM 0o HIU0m GUM MUM H0 52 *h<< \ .QEOB vOH x M GM” I HI00m H.OH x M RD» EH 00HHD4ZD “jail-INJDWJU 1144.0 Hui/.14.).1n...) JIM-J”: I1) liIIN- llllll II .H0D0310cou000 'lll-" 54 wom.H ON OO®.H mm Noo.Hbmm.H mmm.H VN ®V®.H MH ®H®.H NN mHNHI OHO.H®NO.H w®.mMI ®H0.Hmm®.H wm®.H HN mmmom ON ONm.N mH HHO.H®N®.N vwm.N wH mmo.m NH who.m OH mwo.m mH mHHNI moo.Hv¢o.H ww.wml ¢N0.meo.m mmo.m wH 0HOE\H00 _ , ._ Q. MM 0o Hlo0m «GUM «MOM *hfld vn\ @808 CH N I M HwnHEHHZ v I HI00m H.OH x M com \ .1133); .. llf 55 th.m owH NOH.m va mHO.Hme.m OOH.m NwH HbN.m me NbN.O me NHO I voo.HwHo.H mm.OI boo.HOSN.m ONN.m HwH wmm.b OHN HNm.b mHN mNm.h HHN mNo.Howm.h omm.b OHN woo.h vHN vo>.b NHN woo.b NON me I OOO.HOHO.H Nv.ml omo.Hwoo.b mNo.h NON mo.OH NNN mo.HH ONN oo.Hvo.HH OH.HH NNN mm.HH ONN vm.HH vNN meHI HHO.HONO.H ON.OI mO.Hmm.HH ON.HH NNN Nm.HH HNN Nm.HH oHN mo.va.HH om.HH bHN mm.HH ONN OO.HH wHN ¢HOHI bOO.HomO.H ON.OI wo.Hwo.HH bh.HH OHN 06330 n , mM oc Huoom MuABM emoM *mdd MK .QE0B «OH x M H0HESZ I HI000 H.OH N M cam - .H0u0310c0u000 Rom GH 0OHHOH£U nmeru0om out 0OHH0H£U Hmu0om mo mHmmHo>Hom mo m0umm .MHN 0HQMB 56 bHH.H Own ONH.H mom «OO.H>HH.H OHH.H Nvm OHH.H Ovm ONH.H va NHN I voo.H«OO.H >«.ONI OOO.HNNH.H ONH.H own omm.H OOH N«O.H OOH NOO.HH«O.H O«O.H «OH OOO.H OOH OOO.H. SOH Hfiw I Hoo.Hwoo.H NO.NNI NOO.H«OO.H NOO.H OOH Hoo.N OSH OHO.N SSH bHo.N ObH OOO.H>OO.N bow.N «SH «OO.N OBH OOO.N ObH NHO I woo.HwHo.H HO.OHI OOO.HOOO.N bwo.N mbH 000 «no, «mo Uo _ HI M M. 0HoE H00 Q m . HOHEUZ *mwd M\ M @809 «OH x w HIo0m «OH x M cum H.0coov .xHx 0Hnoe 57 NOo.m MOO HO0.0 Hmm HOO.HHO0.0 HO0.0 ONO «H0.0 NOO OH0.0 omm OHOOI OOO.HOOH.H HO.HOI HNO.HOO¢.O OOw.m ONO, OO0.0 SON Hom.m OON me.m «ON ovm.m HON SOO.HOO0.0 OH0.0 OON vww.m OON OOS.O NON SOS.O OON «HHOOI ONO.HOOH.H ON.OOI OOO.HVO0.0. ON0.0. SON . 006, «no «mo 0 HI M M 0HOE Hmo Q m . o _ H0HESZ *m«< M\ M QEOB «OH x w HIo0m «OH x M csm H0003I0cop000 ROS CH 0OHHOH£U nmerp0om Odo 0UHHOH£0 Hmu000 O0 mHmSHo>Hom mo m0pmm .XM 0HQ09 58 A. .Mechanism of Acid Chloride Solvolysis Unlike most reactions to which B—isotope effects are applied, the solvolysis of acid chlorides remains mechanis— tically ambiguous. -The reactivity of acid chlorides re- sembles that of methyl chloromethyl ether (50) or allyl chloride (51) in;fis apparent duality. All three compounds can react by either an Snl (limiting) mechanism or an Sn2 (nucleophilic) mechanism. This accounts for much of the difficulty in establishing a mechanism for acid chloride solvolysis. Of the three common types of acid chlorides available, benzoyl chloride and its derivatives, chloroformate esters, and acid chlorides derived from aliphatic carboxylic acids, the former two have been examined most extensively. This condition reflects the suitability of available techniques for obtaining kinetic data rather than unusual interest in these systems. Johnson (52) and Minato (53) have recently reviewed the solvolysis of aromatic acid chlorides. Johnson's con— clusion, that the solvolysis of acid chlorides is best represented by two simultaneous reactions, represents the majority View. It is proposed that solvolysis proceeds either by a limiting reaction (XI) XI 59 or a nucleophilic reaction (XII) OH R o‘ \\ r.d. . ' ./C=° ::——§—e R-C—OH -———> R-Cff XII c1 H20 él OH The observed isotope effect in the solvolysis of acetyl chloride,.kH/kD > 1, requires limiting reaction XI, in which an acylium ion is formed. .Minato has proposed that the entire solvolysis reaction proceeds through a tetrahedral intermediate, as shown in the mechanism which he has proposed. 5+ OH OH R I 26" k I k + OH . O \c=0 .1, R-CHO —3> R-C-OH 3> R-c” > R-cff CI < ' ' _ \\0H 0H Cl Cl c1 XIII XIV The observed isotope effect requires that the ioniza- tion step k3 be at least similar in magnitude to k1. Given the extensive reactivity ofJGJL it is difficult to justify its role in the rate determining step. At high temperatures k1 becomes the rate determining step, and the isotope ef— fect should decrease, rather than increase as observed. In other situations where "borderline“ reactivity is observed, attempts have been made to rationalize the acti— vation process in terms of a single reaction pathway (54). A similar explanation is possible in the reaction of acid chlorides. 60 It is proposed that as the activation process begins, the electron density of the entering nucleophile is trans— ferred first to the chloride, as in XV o /’ H20 + cna-c<: Cl _, 1- F— OH + C +. 2 0 0H2 OH + 6 O l _ ' c03c=o <—— 003-—¢=o <——> cn3—cH00 -—> CD3-C-0H . 8 ‘ I c1 c1 _ 5' °1 _ XVIII XV XVI XVII In highly polar solvents in which chloride departure is facile, the transition state goes to an acylium ion inter- mediate XVIII. In less polar solvents, the halogen depar- ture requires more energy and bond making increases to the extent that excess charge leaks onto the carbonyl oxygen, as in XVI. Proton transfer in XVI leads to XVII. 'Note that the electron density at the carbonyl carbon increases as XV goes to XVI. The observed isotope effect suggests that the transition state in water—rich solvents resembles XV. B. -Solvent Influence on the Isotope Effect The influence of solvent on the isotope effect in the solvolysis of acetyl chloride is indicated in Table XXI. The isotope effect decreases with a decrease in water con- tent of the solvent, i.e., a decrease in polarity. These results are consistent with a dual mechanism in which the contribution of the nucleophilic mechanism 61 .MM MOSOHMD H>N m0HQmB ECHO G0Mmam o m voo.Hvoo.H ooo.vao.H mHo.HNHH.H OHO.HNNH.H ONO.HOOH.H M\ M e«.«mn ««.«mu Ho.H«I mo.H«I Ho.H«I Aoev .osoa AH0pm3I0cou0um RV Oo oo Ow ow OS Dc0>Hom mo mHmmHo>Hom 030 MOO m0OHHOH£0 nmIHmu0om uo0mm0 0mouomHIO 05D mo 00C0OG0Q0O DG0>Hom .HMN 0Hnme 62 increases as the solvent becomes less polar. -In terms of the unified mechanism, the transition State shifts from XV to XVI as halogen departure becomes less favored. An increase in nucleophilicity is not required by the results. It is possible that the change in solvent reduces the charge development at the carbonyl oxygen without any change in the extent of bond making. The importance of the nucleophilic reaction in acid chloride solvolysis suggests that the former explanation is correct. (Ugi (55) has observed that amine bases catal— yze the solvolysis of both aliphatic and aromatic acid chlorides. For example, addition of pyridine to the sol- volysis of acetyl chloride or benzoyl chloride in 89.9% acetone—water increases the rate by more than 105. Numerous attempts have been made to characterize the role of water in the solvolysis of acetyl chloride (56). The work consists of the determination of the apparent order of water from pseudo-first order rate constants. Termolecular and bimolecular mechanistic proposals have re— sulted, which do not appear to be consistent with a limit- ing mechanism as required by the isotope effect. Correla— tion of the rate of solvolysis with dielectric constant has led to the proposal of a highly polar transition state which includes water (56e). ‘ Ugi (55) has obtained the solvolysis rates of a number of acid chlorides as a function of the GrunWaldHWinstein func- tion Y (57). These results are graphed in Figure 6. Note _-3-0 -2.0 -1.0 0.0 +1.0 +1 i 1 I I 4 I -4 -1 _ 11- II —5 J .—— .-M 0 0.2 0.4 0.6 0.8 1.0 mole fraction H20 ..ql1f.Acetyl chloride (-20°C) -o-o- Acetyl iodide (-40°c) +0— Oxalyl chloride (-50°c) 4434}-Diisopropyl carbamyl chloride (+20°c) Hawk-Benzoyl chloride (+20°C) JVAF-Z, 4, 6— —Trimethylbenzoyl chloride (-20°C) Figure 6. Correlation of the solvent parameter Y with the rate of solvolysis of some acid chlorides.a aSee Ref. 4. 64 that the rate is also linear with respect to mole fraction of water (57). In Figure 7 the results from the work.re— ported here are plotted. In both cases a linear relation- ship is observed in which the slope indicates that acetyl chloride solvolysis is as "solvent sensitive" as the sol— volysis of Efbutyl chloride. The correlation of rate with either Y or the mole fraction of water is limited by the requirement that the solvent interactions in the solvolysis of acetyl chloride be the same as those in the model reaction, in this case, solvolysis of Efbutyl chloride. The presence of the car- bonyl oxygen and the observation of base catalysis suggest that this need not be the case. From Figure 6 it can be seen that of the acid chlorides studied, only mesitoyl chloride does not give a linear correlation with Y. From other investigations (58), it has been concluded that mesitoyl chloride reacts by-a limiting mechanism. It would seem appropriate to consider mesitoyl chloride as the model compound for solvent correlations in acid chloride solvolysis. If this is done, then all of the other acid chlorides must give a nonlinear relationship, which is consistent with the complex mechanism proposed. A recent and excellent discussion of solvent parameters and their application has been published by Kosower (59). Gold and coworkers (60) have concluded that solvent influences the extent of acylium ion formation in the 65 0 -3.0 -2.0 Y —1 0 0.0 I I ' I —1.0 I / n —2.0 / log k / _3 0 /O 1 / 1 o’ 1 -4.0 ... 0 .2 4 6 mol fraction H20 Figure 7. Correlation of the rate of solvolysis of acetyl chloride with solvent a (-15°). aRate constants calculated from parameters in Table XXIV. 66 solvolysis of benzoyl chloride. In the solvolysis of benz— oyl chloride in the presence of N,N-dimethylaniline, in— creased anilide formation was observed as the polarity of the solvent was increased. This suggests that kobs _ klim + knucl' which is consistent with the isotope effect observed. Similar studies have not been carried out in aliphatic sys- tems, where the resulting acylium ion is not stabilized by resonance. The order of reactivity F-< Cl < Br < I suggests that bond breaking can be involved in the rate determining step in aliphatic acyl halides (61). Solvent can also determine the importance of bond breaking. The results of Ugi (55) and Briody and Satchell (62) for the reaction of chloro— acetylchlorides in various solvents are given in Table XXII. In protic solvents, the relative rates depend upon the sub— strate's ability to disperse negative charge, which implies extensive bond making in the transition state. In non—polar aprotic solvents, the order of reactivity is reversed and bond breaking becomes the dominant influence. These results demonstrate the variability of the acti- vation process, but do not require that kobs = klim + knucl 67 Table XXII. Influence of solvent on the relative rates of reaction of halogen—substituted acetyl chlorides. Compound Solvent Rel. Rate Rate (k x 104) CH3COC1a Hydroxylic 1 CCl4 2 ClCHZCOClé Hydroxylic 2.2 CCl4 1.3 C12CHCOCla Hydroxylic 3.3 CCl4 1 CH3COClb 89.1% acetone-water 10.9 ClCHZCOClb 89.1% acetone—water 203 ClZCHCOClb 89.1% acetone—water 31,000 C13CCOC1b 89.1% acetone-water >100,000 a Ref. 62 OH ’0 ’,O O’ 1’ I + RC\ > R b Cl Ref. 55. 68 No clear definition of the activation process can be obtained from the influence of the solvent on reactivity and the isotope effect. The approach contains too many variables to define the transition state with the precision required to rationalize the isotope effect. C. Relationship of Activation Parameters and B-Isotope Effects Activation parameters for the solvolysis of acetyl chloride and acetylj§3 chloride have been determined in 80%, 85%, 90%, and 95% acetone-water. The results are given in Table XXIII. Two trends appear: the enthalpy and entropy of activation decrease as the solvent polarity decreases, and, in those cases in which the isotope effect is con- sidered to be increasing with increasing temperature (90% and 95% acetone—water), AH; > AHE. The magnitude of the activation parameters depends some- what upon the method of calculation. Calculation by ACTENG, in which the rate constants are weighted not only by their deviation from the “best" line but also by the experimental uncertainty of the rate constant, leads to the results in Tables XXIII and XXIV. Calculation by ACTENG in which only deviation of the point from the "best" line is used to weight the data leads to essentially the same values as those of AKTIV in which no weighting procedures are used. These values are given in Table XXV. 69 Table XXIII. Influence of solvent on the activation param- eters of acetyl chloride solvolysisatb(-20°). Solvent ISO_ AH* 88* 88H*C 888*C (%w::::one- tope (cal/mole) (caééggle- (cal/mole) (:31:— deg.) 80 CH3 14,226i48 -12.3i.2 -58i67 0i.4 003 14,284r46 —12.3r.2 85 CH3 13,373r132 —17.3r.6 —303r144 —1i.9 003 13,676i92 -16.3r.4 90 CH3 13,882i162 -17.4i.6 102i185 .5i.9 CD3 13,78or92 -17.9i.4 95 CH3 11,8111120 —28.5i.4 124i170 .5i.7 CD3 11,694i120 -29.0i.4 aCalculated from the Arrhenius parameters of Table XXIV. 8H =-E - RT ; 88* = 4.576 log (A/T) — 49.203. 88H = AH — 8H* ; 888* = As; — 8s * H D D Table XXIV. 70 Arrhenius and transition state parameters for the solvolysis of acetyl chlofide and acetyl—d3 chloridea. ysolvent Ea , AS* AH* acetone- A a1 ole— (éwater) (cal/mole) (C dég) (cal/mole) 80 2.94:.28 x 101° 14,729148 -12.2f.2 14,242i48 2.97:.28 x 1010 14,787r46 -12.2r.2 14,300r46 85 2.40i1.06 x 109 13,876i132 —17.2i.6 13,382i132 4.00:.7 x 109 14,179192 -16.2i.4 13,689f92 90 2.24f.24 x 109 14,385i162 —17.4i.6 13,887i162 1.72:.32 x 109 14,282r92 —17.9i.4 13,790192 95 8.57i1.8 x 106 12,321i120 -28.5i.4 11,809i120 6.61i1.0 x 106 12,196i120 —29.0i.4 11,681i120 aCalculated from ACTENGI std. dev. T 0.010, std. dev. k = % 0. 71 Table-XXV. Transition state theory parameters for.the—sol- volysis of acetyl chloride and acetyl—g3 chloride. Solvent *a 88*a *b 88*b <%w:g:;§ne- $23; (caiim01.> (cagégile‘ (caiim.1.> (caéégfile‘ 80 CH3- 14,247i66 -12.2i.3 14,240i50 —12.2i.2 CD3— 14,309i41 —12.1i.2 14,298i60 -12.21.2 85 CH3— 13,525i92 -16.6i.4 13,544i105 —16.6i.4 CD3- 13,645i72 -16.4i.3 13,664i105 —16.3i.4 90 CH3- 13,830i124 -17.6i.5 13,832i150 —17.6i.4 CD3— 13,608i93 —18.6i.4 13,6101120 —18.6i1.4 95 CH3- 11,924i98 -28.0f.4 11,927i120 -28.1i.5 CD3- 11,8111113 -28.5i.4 11,819i120 —28.5i.5 aCalculated from AKTIV. bCalculated from ACTENG std. dev. T = 0.010, std. dev. k = .5%. 72 The differences arising from different weighting pro- cedures are taken as an indication of the quality of the data. ~Weighting by the standard deviation of each point may not be justified since each point is the average of at least three independent rate determinations. The variation in parameters with method of calculation does suggest that the data cannot be analyzed too closely. There are two options for determining AAH* and AAS*. One is to take the difference of the individually deter- mined activation parameters. ~In the present case, this leads to differences which are generally smaller than the combined experimental error of the two rate determinations (TableXXIII). The second approach is solution of the equation in kH/kD = —88H*/RT + 888*/R . These results are given in Table XXVI. Note that the experi— mental uncertainty is greatly reduced. rThe results calcu- lated by HANDS, a single least square solution, differed from those of-ACTENG only in one instance. For solvolysis in 80% acetone—water, HANDS gave AAH* = 13 i 59 cal/mole, 888* = 0.26 i .24, whereas ACTENG gave AAH* = —56 i 60 cal/ mole, AAS* 3 0 i 0.3 . Weighting factors included the deviation of each rate constant. It was concluded that the differences in 85%, 90%, and 95% acetone-water are real in sign, but cannot be inter— preted in terms of magnitude. The situation.in 80% acetone— Table XXVI. 73 Activation parameters determined from the tem— perature dependence of the B—isotope effect of acetyl-g3 chloride solvolysisa *b AA3*C Solvent *b AAs *c 0’5 gggggge- (Hi/moi.) (“$433“— (‘éfii/m.) Wiggle- 95 108i22 0.410.1 110:20 0.45i0.1 90 225i55 0.9:0.4 225133 1.0ro.1 85 —139i50 -0.4iO.5 —119i24 —0.3i0.1 80 — 56:57 0 i0.3 13i59 0.3:0.2 aCalculated from the data in Tables XVI through XX. b kH/kD = Calculated from ACTENG, std. dev. T = 0.01, std. dev. % 0 CCalculated from HANDS. 74 -water is ambiguous, but it seems reasonable to conclude that AAH* is negative. The two trends remain: the activation parameters de- crease with decreasing polarity of solvent and AAH* is positive in 90% and 95% acetone—water. The activation parameters in 90% acetone—water are somewhat high. No ex- planation for this can be offered. The trend in activation parameters is consistent with a dual mechanism. In solvolysis reactions, both activation enthalpies and entropies are found to decrease as a reac- tion becomes more nucleophilic (63). If in the solvolysis of acetyl chloride, the nucleophilic portion of the observed rate increases as the solvent becomes less polar, the acti- vation parameters would be expected to decrease. This behavior of the activation parameters is not limited to those cases in which there is mechanistic ambigu- ity. Similar results were obtained for the solvolysis of 8—methyl-1-naphthoyl chloride (Table XXIX). ,In‘Table XXVII are listed the results of similar studies in which the mechanism is better characterized. The influence of sol- vent on activation parameters has been extensively examined (64). The results are complex and difficult to generalize upon. The result that AH; > AH; in those solvents in which the isotope effect is observed to increase is also consistent 75 Table XXVII. .Influence of solvent on the activation param- eters of solvolysis reactions. Compound Solvent * A885 (% a:::::§‘ (kcal/mole) (caéé3319- 'E—Butyl chloridea 90 21.85 -16.8 80 22.00 -11.0 70 21.16 -10.0 50 20.78 - 5.4 Benzhydryl chlorideb 90 18.8 —20.0 80 20.4 - 9.0 Benzyl chloridec 70 21.30 —23.96 50 20.60 -22.41 aA. H. Fainberg and S. Winstein, J. Am. Chem. Soc., 22x 1602 (1957). bRef. 64b. cG. R. Cowie, H. J. M. Fitches, and G. Kohnstam, J. Chem. Soc., 1585 (1963). 76 with a dual mechanism. This can be demonstrated by examin- ing the thermodynamics of a model system. Reasonable values for the activation parameters for k . and k nu were assigned. The isotope effect was con— lim cl sidered to arise solely from AAH* (ZPE differences). Hydrogen -)(- _ * :— AHlim - 17,000 cal/mole 88lim 5 cal/mole deg * _ * _ AHnucl - 10,000 Asnucl -35 Deuterium * ._ -X- _ AHlim 17,200 Aslim 5 * = * :— AHnucl 9,900 Asnucl 35 It was assumed that kObs = aklim + Bknucl' The co- efficients a and B were determined from the expression ..x. nucl *- AFlim - 8F = -2.3026 RTlog d/B. The free energy of activation, AF:, was determined from the expression * -)(- 'X' * _ * AFi — O‘AHlim + BAHnucl _ (QASlim + fiASnucl)T For T1 = 233° a = 0.505, B = 0.495. AF; = —8H* — T88* = 13535 + 4625 = 18160 cal/mole AF; = 13635 + 4625 = 18261 888* = 88H* 2 -101. For T2 = 2530 a = 0.78, B = 0.22. 8F = AH* - T8s* 2 15460 + 2935 18395 cal/mole = 15594 + 2935 18529 AF U*III* 88F* = 88H* = -134. 77 Rate constants determined from AF: according to the equation log ki = log kT/h - AF:/2.3026 RT gave the following rate ratios: 4.38 x 10'5/3.53 x10"5 1.24 For T1 kH/kD For T2 kH/kD The isotope effect increases with increasing temperature 6.64 x 10‘475.09 x 10‘4 1.30. as required by the experimental results. Values for AH: and 88: were determined from the rate constants by the equations * AH = 2 3026 R T1T2/(T2 - I.) log szl/lez); 45* = 2-3026 R109 k/h + 8H*/T + 2.3026 Rlog kl/Tl. The results are as follows -)(- -X- = 15438 cal/mole 8sH = —14.4 cal/mole deg *_ *__ 8HD _ 15145 88D _ 16.1 88H* = + 284 888* = + 1.7 . Note that AAH* and AAS* obtained from the temperature dependence of the isotope effect are not related to the ZPE difference which is the origin of the isotope effect. The deviation is in the expected direction. If a dual mechanism, or any model which includes a change in the potential energy of the activation process with temperature, accurately de— scribes the activation process, the parameters AAH* and AAS* do not have a significance which can be related to the origin of the isotope effect. Evidence suggests that the solvolysis of acid chlorides is sufficiently flexible to meet this criterion, which 78 requires that the activation process be a function of tem- perature. Cason and Kraus (65) observed that alcoholysis of hindered acid chloride XIX with carefully dried methanol gave 25% of the corresponding half-ester, presumably by a nucleophilic mechanism. The formation of the half—ester was suppressed by carrying out the reaction at high tempera- tures. Similar behavior was not observed for acid chloride XX. $2H5 $2H5 CH3002HCHZCH2-c—cggl CESCCHZCHZ-C-COZCH3 C4H9 C4H9 XIX XX Kelly and Watson (66) have examined the solvolysis of benzoyl chloride Over a wide range of concentrations of aqueous acetone. At high concentrations of water, the linear dependence of ln kobs/[HZO] y§_ln[HZO] observed at low concentrations disappears. This was taken as evidence for two mechanisms. The authors resolved the observed rate into two sol— vent-independent rate constants, k1 and kg, for the limiting and nucleophilic reactions, respectively. The observed rate was found to fit the equation 0°0092 — 1-14 k = k2[H20]( T ) 8 0138 + k1[H20] to withinilO% in all cases. The values of k1 and k2 are given in Table XXVIII. They demonstrate the dominance of 79 Table XXVIII. Solvent independent rate constants character- izing the temperature dependence of benzoyl chloride solvolysisa T§$p° 106 x kzb 1016 x klb 0 1.81 6.90 15 5.32 38.9 25 10.3 112 35 19.0 300 45 33.4 880 55 56.6 2100 aRef. 66. bkl and k2 are solvent—independent rate constants which fit the relationship 0°0092T - 1.14 kObS = k2[H20]( ) + k1[H20]8 80 k1 as the temperature is increased. The narrowness Of the range of solvents used for the solvolysis of acetyl chloride and the need for better than i10% accuracy precludes a simi- lar treatment for acetyl chloride. Even though the evidence is somewhat conflicting, the most reasonable description of the activation process is a dual mechanism. 'The alternative, a nucleophilic reaction in which an entropy effect is the major factor, is without precedent. There is sufficient evidence for the complexity of the activation process to justify its influence on the unusual temperature dependence of the B-isotope effect. D. Origin of theyB—isotgpe Effect The B—isotope effect has been rationalized in terms of a dual mechanism, which changes rapidly in 90% and 95% acetone-water and appears to be predominantly limiting in 75%, 80%, and 85% acetone-water. It has been demonstrated that if this is the correct description, the activation parameters, AAH* and AAS*, do not have a simple meaning. Since AAH* and AAS* cannot be attributed to a single transition state, only AAF* can be used to evaluate ZPE dif— ferences. In the Introduction, a number of examples were given of isotope effects in which an apparent entropy con- tribution was present. Two explanations are possible: (1) the eXperimental error is too large to justify inter— pretation of the magnitude of AAH*, or (2) the activation process is a function of temperature. 81 This latter alternative meets the criterion of Wolfs— berg and Stern (2) for those cases in which the temperature dependence is unusual, namely, that two force constants be changing in opposite directions. The data from Robertson's group (6) is very precise and should be taken as.a standard for evaluating the other studies. As discussed previously, the solvolysis of t- butyl chloride and tfbutyljgg chloride shows no entropy effect, AAF* E’AAH*. This is observed in the solvolysis of most tertiary carbonium ions (67). The temperature-independent isotope effect observed in the solvolysis of the isopropyl compounds may be similar to the behavior of acetyl chloride. Since isopropyl compounds may proceed by either limiting or nucleophilic mechanisms (68), it is possible that the charge development in the transition state is temperature dependent. r .+ ‘ - ~ +' OH. CH3 + X _CH3\ 5/I70H2 CH3\. 2 CH3\ ‘>£*H <—— //C:H <""—> l/CiH ‘—+> //CHOH CH3 _ CH3 \\X_ CH3 . ' _ CH3 The fact that AAH* 2’0 rather than a positive value as observed for acetyl chloride may be due to the greater flexibility of the acid chloride with its carbonyl oxygen. -Karabatsos and Ennioannou (28b,c) have recently re- ported results from the solvolysis of 8-methyl—g3—1-naphthoyl chloride which were similar to those obtained in the solvol- ysis of acetyl—g3 chloride. Their results are given in Table XXIX. 82 .ewm .mmmo .QM\mM moHI u Bm\*m<< Eonm mHoE\Hmo mm: H *mdq mcflESmmm nx\mx Umumasoamon .nmm .mmmm . .n . , . moo www.maoe\amo a +5 om- n *mq maoa\amo swaemo «H n *m< m www.maos\amu m.aw.om- u *mq maos\amu wmmflmoo.wa u mwmq mHHw I mmo.flmao.a Hmfi.wvnw.m mmH.HHum.a mmm\vm mama- mHo.Homo.H amo.wvvm.w mmo.flmso.w uwv.vm «Hm . HHo.HmHo.H HHo.HmHo.m omo.Homo.m ¢HH.mH oemm .n . n . mno www.maoa\amu H +m w- u *mq maoa\amo ma+mom 5H u *m< maos\amo moHHwom- n *m44 «mo www.maoa\amo v.Hm.m- u *m< mHoe\Hmo mmanmm.mfi u *m4 mmH.H wave- moo.wmoa.fl mmoo.HwHow.o mmoo.amooo.fl moo.vmn smH.H mama- wHo.HbmH.H moo.an¢o.m mwo.Howm.m mam.mmu omH.H HHHHm- nmo.«oma.a mHo.Hmwm.o 5H.Ham.s mmw.om- mnu.vm\mm.m> Q m. U. Hmum3 x\ x m”no «mo. ago «ca x «mo . o umcoumom A3 3V n.0Hmu *m<< x\. x aOH x x x mama ucm>aom mofiuoano HmonunmmCIHImmramznmEIw Ucm wUHHoHno HmonuAQMCIH Iamnumfilm mo mHmMHo>Hom on» CH uommmw onuomH mnu mo mocmocmmmo mudumnmmEmB .xHNx magma 83 It was observed that both kH/kD and the activation par rameters of solvolysis decreased with a decrease in solvent * polarity. This was attributed to a contribution from.AAS*, rather than to a dual mechanism. It was argued that solvolysis of 8—methy1—1—naphthoyl chloride was limiting in both "95%" and "75%" acetone-water because of the steric hindrance to formation ofia tetra— hedral intermediate and because it solvolyzed 8 times faster than naphthoyl chloride. The corresponding esters show the opposite reactivity relationship. -It is not inconceivable, however, that there could be some bond-making in the transi— tion state in water-poor solvents similar to that proposed for acetyl chloride. Using a unified mechanism as a model, the following sequence can be envisioned: 1 CD3Q§C/C XXIV V XXI XXII XXIII Although it is not expected that the transition state XXII will contribute significantly, it is possible that some bond-making can occur in weakly polar solvents where 84 chloride departure is less favored. This would lead to a reduction in the isotope effect. Karabatsos and Papaioannou have interpreted the enthalpy of activation difference, AAH* =‘308i100 cal/mole, in "75%" acetone-water as being significantly different from the free energy difference, AAF* = —60—70 cal/mole. This re— quires that a considerable entropy effect be present. The experimental reality of the situation is that an isotope effect which behaves "normally" (AAF* E'AAH*) can— not be determined beyond experimental error. This is demonstrated by the calculations in Table XXIX in which AAF* = —65 cal/mole. The deviation from the best straight line gives a mis- leading indication of the actual error involved. A better criterion for the significance of AAH* and AAS* would be the requirement that a "normal” isotope effect be observ- able outside experimental error before the relative magni— tudes of AAF* and AAH* can be interpreted. It is concluded that the free energy difference, AAF*, remains the best indication of ZPE differences. Comparison of the steric isotope effect in 8—methyl—g3-1—naphthoyl chloride with the isotope effect due to hyperconjugation in acetyl chloride solvolysis is obscured not only by the mechanistic uncertainty of solvolysis but also by a lack of knowledge of the charge density distribution in the acylium ion, i.e., the extent to which resonance form X contributes. 85 + + CD3C=O < > onscso IX X The calculated isotope effect (28d), AAE = -38 cal/mole, for the limiting solvolysis of acetylega chloride is less than the observed free energy difference, AAF* = —60i14 cal/mole, for solvolysis in 75% acetone-water at —30° (Table xx). For solvolysis of 8-methyl-g3—l—naphthoyl chloride, the calculated isotope effect, AAE = -200i100 cal/mole, is larger than the experimentally observed free energy dif— ference, AAF* = —65i10 cal/mole. Previous observations are confirmed (28). The calcu— lated isotope effect underestimates the observed isotope effect in cases in which hyperconjugation is possible and overestimates the observed isotope effect in cases in which steric effects alone influence the isotopic rate ratio. E. .Isokinetic Temperatures in B—isotope Effects The work of Karabatsos and Papaioannou (28b) and the apparent contribution of entropy effects in the B—isotope effect of acetylfids chloride solvolysis suggest the exist— ence of an isokinetic relationship similar to that proposed by Leffler (69). The interpretation of the solvolysis re- action in terms of a dual mechanism precludes the applica— tion of an isokinetic relationship. The existence of an 86 isokinetic relationship is taken as an indication of similar- ity in reaction mechanism, rather than a diVergence. The existence of the isokinetic relationship AAH* = BAAS* has received extensive scrutiny since its proposal. Ritchie has recently reviewed the objections, his own and those of other workers, to the isokinetic relationship (70). He concluded that a relationship exists between AAH* and AAS*, but that it depends upon the reaction temperature as well as the reaction. Peterson, Margraf and Ross (71) have demonstrated that a linear relationship of AH* and AS* can be obtained with— out the intemmction of the rates required by the isokinetic relationship. Errors in the determination of AH* and AS* have been shown to lead to a linear relationship. ‘Numerous derivations of the linear relationship have appeared (70,72). There is little evidence for the source of a linear relationship. Ritchie noted that the commonly used generalization that solvent effects which lower the enthalpy of activation require a corresponding restriction of motion and loss of entropy of activation requires neither a linear relationship nor defines the magnitude of the changes involved. There may exist a relationship between AAH* and AAS* for fi-isotope effects. Neither previous studies nor the work reported here justifies its inclusion in the long list of factors determining B—isotope effects. 87 F. Activation Parameters for the Solvolysis of Other Aliphatic Acid Chlorides Throughout the discussion of the fi—isotope effect for the solvolysis of acetyl-g3 chloride, the approach has been similar to that of Brown (1). Those results which do not seem consistent are discounted and those which lend strength to the arguments are weighted more heavily. .Such an ap— proach is required by the subtlety of B-isotope effects, the complexity of the activation process, and the magnitude of the experimental uncertainty. Escape from this subjective and essentially circular argument is difficult. .One possible route is the examina— tion of B and y isotope effects in the solvolysis of other aliphatic acids. A determination of the activation param— eters for the solvolysis of propionyl chloride, isobutyryl chloride, and pivaloyl chloride suggests that this route cannot provide an unequivocal rationalization of the B—iso- tope effect. The relative rates of the aliphatic acid chlorides, Rf , \Cl suggests that the reaction becomes more nucleophilic as are CH3“ > CH3CH2- > (CH3)2CH‘ >> (CH3)3C"'I WhiCh branching in R increases. This may not be the case. Olah (73) has observed a similar order of reactivity for the dissociation of proton— ated aliphatic acids and esters to acylium ions. 88 The activation parameters for the solvolysis of the acid chlorides are given in Tables XXX and XXXI. ’The re— sults of Olah and coworkers are given in Table XXXIII. The rate sequence appears to be determined by.solvation effects rather than by substrate electronic effects. Little is known about the solvation effects in HSO3F-SbF5. The overall result is similar to that encountered in mixed solvents since AS* is negative for an ion-forming transition state. The 6 and y-isotope effects cannot be predicted a_priori from the activation parameters. iln fact the parameters can be made to agree with any isotope effect observed. For ex- ample, an inverse B-isotope effect is consistent with the lower activation parameters, indicating increased nucleo- phility of the reaction for propionyl chloride and Egg- butyryl chloride. Steric hindrance to solvation of the acylium ion can be invoked to explain the decreased partici— pation of the acylium ion. A normal isotope effect would be consistent with the implications of Olah's work, and with the increased steric hindrance to nucleophilic attack. Certainly the results will be useful, but the quality of the argument will not be much improved over that which has been presented here to rationalize the B—isotope effect for the solvolysis of acetylegs chloride. It will be im— possible to exclude an entropy contribution from the B—iso— tope effect because of the uncertainty of the activation process. 89 Table XXX. Activation parameters for the solvolysis of some raliphatic acid chlorides (-25°).a Y —“"—. Acid Chloride AH*(;:{é)y . ffAs*(e.u.) Acetyl chlorideb 14,236i48 -12.3:.2 Propionyl chloride 13,746i45 -15.9i.2 igngutyryl chloride 12,645i45 —21.5:.2 Pivaloyl chloride 13,548i180 —21.4i.6 aCalculated from Ea and A of Table XXXII AH* = Ea — RT. AS* = 4.576 log A/T - 49.203. bCalculated from Table XXIV. 90 Table.XXXI. Arrhenius and transition state theory parameters for the solvolysis of propionyl chloride, iso- butyryl chloride, and pivaloyl chloride in 80% acetone-watera Lg- ’ W” *(aéI/g . .- R A E AS*(e.u.) AH* mole) CH3CH2- 4.59:.4 x 109 14,239i40 —15.95:.2 13,745:40 CH3 :bH- 2.77:.3 x 108 13,138i45 —21.5:.2 12,645i45 CH3 (CH3)3C- 2.99:1.0 x 108 14,041i180 -21.4i.6 13,534i180 aCalculated from the solvolysis data of Table XXXII by ACTENG, std. dev. T = 0.01, std. dev. k = % o. nvma. mom mmwm. mom Nbvv. Hum NvNH. mmm mmvm. wmm momv. Ohm wHoo.OHHmNH. ommfi. 0mm vooo.OmevN. bmvm. 5mm vHoo.OHHmwv. vmww. mmm omv.mmu n mama owm.mH- u mama ems.mau u mama mUHHOHQU Hmon>Hm mmmw. rpm who». mmm va.H mmm mmmw. wbm ume». Nmm OHM.H mom omoo.owmvmv. ovmv. mum mvH0.0HHmHh. mvmb. Hmm vH0.0Hmom.H Ham.H Hon omH.om- u mama cmw.mmu u mama ems.mHu u mama mofluoano HNH%usmlomH wbmb. chm H®N.H wfim wov.w mom I vab. mum NQN.H mwm mmv.m mom smoo.OHmva. Hmmb. mum HO0.0HH®N.H HwN.H wwm mHo.OH®vv.N va.m vmm oofi.om- u mama cs¢.mm- n mama oou.mH- n mama wUHHOHLU HmCOHmoum alomm HIomm. Hon alomm Hloom Hon Hloom alomm Hon «OH x m «CH xxx 1852 «CH x w nOH x x IESZ «OH x w mOH x M I852 . qsm cum :sm .Hmum31mcoumom Row CH mUHHoHSU Hmoam>flm Usm .mUflHoHao Hmumpsnlomfl .wwfluoano H>GOHQOHQ mo mHmmHo>Hom mo mmumm .HHxxN manme 92 .m . 0 mu m MU M .mom.mw . Ae\fluo¢ .HHHXXN manme 93 G. Heat Capacities of Activation Free energy relationships (solvent effects, structure- reactivity relationships) and activation parameters have been used to rationalize the fi-isotope effect in acetyl-g3‘Chlor- ide solvolysis. Collectively, they suffer from an inability to define the activation process with enough precision to identify the origin of the isotope effect observed. Recent investigation of the heat capacity of activa- tion, ACE, in solvolysis reactions suggests that similar investigation with acid chlorides can provide a definite conclusion about the origin of the B—isotope effect. Robert— son (63) is responsible for most of the work in water. Kohnstam (74) is the principal worker in mixed solvents. A review by Hulett (75) is also useful. The chief limitations to the determination of heat capacities of activation are the extreme precision which is required of the rate data, better than 0.2%, and the wide temperature range, > 30°, which must be studied before sig— nificant values of the heat capacity can be obtained. Heat capacities are obtained by fitting the experimental rate constants to a three-parameter equation log k = A/T + B log T + C where * * A = —AH;/2.3026; AH; - AHO + ACpT B = ACE/R + 1 c = (AS* — Ac;)/2.3026R + log k/h; A5; = AS* + ACE. 94 i Kohnstam (74b) discussed a least squares method for ob— taining A, B, and C. Heat capacities of activation generally range from —20 to —100 cal/mole-deg and are considered to be a measure of solvent reorganization during the activation process. The mechanism differs somewhat, depending upon whether the sol— vent is water or an aqueous—organic solvent. Robertson (63) has proposed that, in water, initial state solvent-solvent interactions are much greater than solute-solvent interactions. The activation process in- cludes only the destruction of the ground state solvation shell. The heat capacity effect is observed because the solvation shell is in equilibrium with unstructured solvent. As the temperature increases, this equilibrium shifts in favor of the unstructured state, weakening the solvation shell and lowering the energy required to achieve the transi— tion state. Snl reactions are observed to have a more negative ACE than Sn2 reactions. The difference between the two values is attributed to the extent of reorganization of solvent required to achieve the transition state. Robertson (6) has used heat capacity data to examine the influence of solvent on the B-isotope effect in the sol— volysis of'Eybutyl-gs chloride. Solvolysis in water gave a value for Ac; of —83 cal/mole'deg and kH/kD = 2.351 at 30.160. In 50:50 ethanol—water the heat capacity of 95 activation was considerably different, AC; = -34 cal/mole- deg, but kH/kD was unaffected, kH/kD = 2.346 at 30.0140. Despite the significant difference in solvent reorganization, the isotope effect was unaffected. The authors concluded that differential solvent interactions due to isotopic sub— stitution were negligible. Queen (76) has examined Ac; for the solvolysis of a number of chloroformate esters. (His results are given in Table XXXIV. Phenyl and methyl chloroformate solvolyze by a nucleophilic mechanism. Ethyl, propyl, and isopropyl chloroformate gave a less negative ACE, which was inter— preted as indicating not only solvent reorganization but also a change in mechanism as a function of temperature. The extreme negative value of AC; for dimethylcarbanyl chlor— ide was indicative of a limiting mechanism. Kohnstam (74) has observed that the ratio ACE/AS* for solvolysis in aqueous-organic solvents is independent of the substrate and depends only upon the reaction mechanism. As in water, Ac; for a limiting reaction is more negative than AC; for a nucleophilic reaction. Initial state solvent structure is less pronounced in mixed solvents than it is in water. rAs a result of the loosely bonded structure, higher vibrational levels are thermally populated. During the activation process, the solvent interacts strongly with the solute, raising the vibrational energy to a level where the higher vibrational levels can no longer be thermally populated. The energy 96 .os .mmmm m m.mm- H.m- m.m m.m . o.mm- «.mw- Amme.maoa\amuv *o< om.m NH.OH mm.oH| oo.sH: so.mH| ms.mH1 .Ammw.maos\amov *m< mom.om- «OH.vm HmH.sH oso.>fi oHH.oH OHH.wH Amflos\amoxv *m< mofluoano HmEmnumo Hmmmmm ammoum Hanum Hmnumz Hmcmnm m Iamnumsflo Ho/ Hum + Noo + mom A u \onojm aofluommm o m o\ mumumw muMEHomouoaso mo mfimmao>HOm map How muoumfimumm coaum>flu0¢ m.nwum3 CH .>Hxxx magma 97 lost in this way appears as ACE. The activation process has been compared with the freezing of water. ACp for water is 18 cal/mole-deg, ACp for ice is 9 cal/mole-deg. The ACp for the freezing process is —9 cal/mole-deg. rAn examination of AG; for the solvolysis of acetyl chloride will provide conclusive evidence for the nature of the activation process. ‘Heat capacity data allows a distinc— tion between nucleophilic and limiting reactions which is not possible from analysis of AH* and AS*. ’More importantly, such data can identify complex temperature—dependent acti- vation processes such as that proposed for the solvolysis of acetyljg3 chloride. Archer and Hudson (77) have determined heat capacities of activation for the solvolysis of benzoyl chloride. (In 61.05% (W/W) acetone—water, ACE = -21 cal/mole°deg, Ac; = 53 cal/mole~deg for reaction in 70.24% (W/W) acetone-water, Ac; = 61 cal/mole°deg in 81.68% (W/W) acetone—water. The latter two results are indicative not only of solvent reor- ganization but also of changes in the activation process. -Even in cases in which a wide temperature range is not possible, determination of the isotope effect with the ac- curacy required to obtain AC; is desirable. 98 H. Conclusion .The behavior of the B~isotope effect in the solvolysis of acetyl-d3 chloride in 90% and 95% acetone—water has led to the conclusion that the reaction is best described by an activation process in which k0 = k . + k This bs lim nucl' conclusion is based on the assumption that hyperconjugative interactions are responsible for the isotope effect. As a result this work does not permit an evaluation of the origin of B-isotope effects to be based on anything other than the reasonableness of the reaction mechanism proposed. It has been demonstrated that the activation parameters for the solvolysis of acetyl chloride and other aliphatic acid chlorides are of little value in defining the activa- tion process. It is concluded that the determination of heat capacities of activation offers the only method for characterizing the reaction with the precision required to identify the origin of the isotope effect. While these results do not allow an analysis of the origin of the fi—isotope effects, they imply that an examina— tion of the influence of temperature on the isotope effect for the solvolysis of other "mixed" reactions could lead to useful information concerning the mechanism involved. (1) (2) (3) (4) (6) (7) (8) (9) (10) (11) (12) (13) (14) REFERENCES 1 Private communication, H. C. Brown, during the Renaud Lectures, Michigan State University. For a more op- timistic view of the value of B-secondary isotope ef- fects, see E. R. Thornton, Ann. Rev. Phys. Chem., 17, 349 (1966); P. Laszlo and Z. Welvart, Bull. Chim. SSE. France, 2412 (1966). ' M. Wolfsberg and M. J. Stern, Pure and Appl. Chem., 8 225, 235 (1964). a) J. Bigeleisen, J. Phys. Chem., 11, 675 (1949); b) L. Melander, Isotope Effects on Reaction Rates, Ronald Press Co., New York, 1960, p. 15. J. C. Evans and G. Y.—S. Lo, J. Am. Chem. Soc., 88, 2118 (1966). V. J. Shiner, Jr., B. L. Murr, and G. Heinemann, ibid., 82“ 2413 (1963). L. Hakka, A. Queen, and R. E. Robertson, ibid., §Zfl 161 (1965). G. J. Frisone and E. R. Thornton, ibid., 86, 1055 (1964). Indiana University Conference on Hyperconjugation, Tetrahedron, 5, 105-274 (1959). For an excellent critical review of hyperconjugation, see M. J. S.Ikmmr. Hyperconjugation, Ronald Press Co., New York, 1962. V. J. Shiner, Jr., and J. S. Humphrey, Jr., J. Am. Chem. Soc., §§/ 2416 (1963). a) E. S. Lewis and G. M. Coppinger, ibid., 76“ 4495 (1954); b) E. S. Lewis, R. R. Johnson, and G. M. Coppinger, ibid., 81” 3140 (1959); c) V. J. Shiner, Jr., and C. J. Verbanic, ibid., 22/ 373 (1957). V. J. Shiner, Jr., and G. S. Kriz, Jr., ibid., §§/ 2643 (1964). V. J. Shiner, Jr., W. E. Buddenbaum, B. L. Murr, and G. Lamaty, ibid., 92/ 418 (1968). G. Lamaty and J. R. Rogue, Tetrahedron Letters, 5293 (1967). 99 (15) (16) (17) (18) (19) (20) (21) (25) (26) (27) (28) 100 J. P. Schaefer, M. J. Dagani, and P. S. Weinberg, in Am. Chem. Soc., 22/ 6938 (1968). K. L. Servis, S. Borcic, and D. E. Sunko, Tetrahedron, 32/ 1247 (1968). M. Nikolectic, S. Borcic, and D. E. Sunko, Tetrahedron, 23, 649 (1967). It was Schaefer's results which Brown rejected; see Ref. 1. P. Lazslo and Z. Welvart, Bull. Chim. Soc. France, 2412 (1966). P. Geneste and G. Lamaty, ibid., 669 (1968). For a review particularly sensitive to the contribu— tion of inductive effects, see E. A. Halevi, "Secondary Isotope Effects," Progress in Physical Organic Chem- istry, Vol. 1, S. G. Cohen, A. Streitweiser, and R. W. Taft, ed., John Wiley and Sons, New York, 1963, D. R. Lide, J. Chem. Physics, EE/ 1519 (1960) and references therein. a) G. van Dyke Tiers, ibid., 22/ 963 (1958); b) G. van Dyke Tiers, J. Am. Chem. Soc., 22x 5585 (1957). E. A. Halevi and M. Nassim, Bull. Res. Council Isreal, SeCt. A5, 263 (1956); J. Chem. Soc., 866 (1963). M. Paabo, R. G. Bates, and R. A. Robinson, J. Phys. Chem., 22” 540 (1966). L. S. Bartell, J. Am. Chem. Soc., 83“ 3567 (1961). H. C. Brown and G. J. McDonald, J. Am. Chem. Soc., 88, 2514 (1966); H. c. Brown, M. E. Azzaro, J. G. Koelling, and G. J. McDonald, ibid., 88“ 2510 (1966). a) G. J. Karabatsos, G. C. Sonnichsen, C. G. Papaio- annou, S. E. Scheppele, and R. L. Shone, J. Am. Chem. S29,, 82/ 463 (1967); b) G. J. Karabatsos and C. G. Papaioannou, Tetrahedron Letters, 2629 (1968); c) Ph.D. Thesis, C. G. Papaioannou, Michigan State University, 1967; d) Ph.D. Thesis, G. C. Sonnichsen, Michigan State University, 1967. E. S. Lewis and E. E. Boozer, J. Am. Chem. Soc., 16, 791 (1954). (30) (31) (32) (33) (34) (35) (36) (37) (38) (39) (40) (41) (42) (43) (44) (45) (46) (47) (48) 101 K. T. Leffek, R. E. Robertson, and S. E. Sugamori, Can. J. Chem., 39“ 1989 (1961). K. S. Pitzer and D. W. Scott, J. Am. Chem. Soc., 62” 803 (1943). K. S. Pitzer and W. D. Gwinn, J. Chem. Phys., 12“ 428 (1942). D. R. Lide, ibid., 22/ 1426 (1958). K. T. Leffek, R. E. Robertson, S. E. Sugamori, Can. J. Chem., 32/ 1836 (1961). E. A. Halevi and Z. Margolin, Proc. Chem. Soc., 174 (1964). V. J. Shiner, Jr., J. Am. Chem. Soc., 16, 1603 (1954). K. Mislow, R. Graeve, A. J. Gordon, and G. H. Wahl, Jr., ibid., §§/ 1733 (1964). J. B. Conant and W. R. Kirner, ibid., 42” 245 (1924). H. F. Stimson, J. Research, Natl. Bur. Standards, 42“ 209 (1949). D. D. McCracken, A Guide to Fortran IV Programming, John Wiley, New York, I965, p. 31. Leeds & Northrup Co. Publication AI—2212, 1967. For a discussion of error treatment see D. P. Shoemaker and C. W. Garland, Experiments in PhysicalJChemistry, McGraw—Hill, New York, 1962, Chap. 2, p. 17. A Fortran IV copy of ACTENG was obtained from Prof. Donald F. Noyce, University of California at Berkeley. Fortran II copies are available from Quantum Chemistry Program Exchange (QCPE), Indiana University, Blooming- ton, Indiana. H. C. Brown, J. Am. Chem. Soc., 62” 1325 (1938). Sadtler Standard Spectra, Sadtler Research Laboratories Philadelphia. M. L. Bender and M. S. Feng, J. Am. Chem. Soc., 82” 6318 (1960). Ph.D. Thesis, R. A. Mount, Michigan State University, 1967. A. Murray III and D. Lloyd Williams, Organic Synthesis with Isotopes, Interscience, New York, 1958, Vol. 2, p. 1262. 102 (49) Private communication, Dr. C. G. Papaioannou. (50) T. C. Jones and F. R. Thornton, J. Am. Chem. Soc., 89, 4863 (1967). ’”” (51) P. B. D. de la Mare and J. Vernon, J. Chem. Soc., 2504 (1954). (52) S. C. Johnson, "General Base and Nucleophilic Catalysis, .Advances in Physical Organic Chemistry, Vol. 5, V. Gold, ed., Academic Press, London, 1967, p. 23. (53) H. Minato, Bull. Chem. Soc. Japp, 31, 316 (1964). (54) a) E. R. Thornton, J. Am. Chem. Soc., 89,2915 (1967). b) c. G. Swain, L. E. Kaisor, T. E. J. Knee, ibid. 11, 1122 (1955). ““' (55) I. Ugi and F. Beck, Chem. Ber., 94, 1839 (1961). (56) a) C. G. Swain and C. B. Scott, J. Am. Chem. Soc., b) G. Zimmerman and C. Yuan, ibid., 77, 332 (1955); c) B. L. Archer and R. F. Hudson, J. Chem. Soc., 3352 (1953). ) V. Gold, J. Hilton, and E. G. Jefferson, ibid., 2756 (1954). e E. J. Cairns and J. M. Prausnitz, J. Chem. Phys., 32,169 (1960). O; V (57) E. Grunwald and S. Winstein, J. Am. Chem. Soc., 10, 846 (1948). (58) D. A. Brown and R. F. Hudson, J. Chem. Soc., 3352 (1953). (59) E. S. Kosower, Physical Organic Chemistry,-John Wiley, New York, 1968, Part Two, p. 259. (60) V. Gold, J. Hilton and E. G. Jefferson, J. Chem. Soc., 3303 (1955). (61) a) C. A. Bunton and J. H. Fendler, J. Org. Chem., 31, (196 7); b) Ref. 55. (62) J. M. Briody and D. P. N. Satchell, Proc. Chem. Soc., 268 (1965). (63) R. E. Robertson, "Solvolysis in Water", Progress in Physical Organic Chemistry, Vol. 4, A. Streitweiser, Jr., and R. F. Taft, ed., Interscience,.New York, 1967, p. 213. 103 (64) a) J. E. Leffler, J. Org. Chem., 20, 1202 (1955); b) S. Winstein and A. H. FainbergCnJ. Am. Chem. Soc., 22, 5937 (1957); c) J. B. Hyne, R. Wills, and R. E. Wonkka, ibid., 22/ 2914 (1962). (65) J. Cason and K. W. Kraus, J. Org. Chem., 22/ 2624 (1961). (66) M. J. Kelly and G. M. Watson, J. Phys. Chem., 62/ 260 (1958). (67) A. Queen and R. E. Robertson, J. Am. Chem. Soc., 88, 1363 (1966). (68) E. D. Hughes, C. K. Tugold and V. G. Shapiro, J. Chem. Soc., 225 (1936). (69) J. E. Leffler, J. Org. Chem., 22/ 1202 (1955). (70) C. D. Ritchie and W. Sager, "An Examination of Struc- ture—Reactivity Relationships," Progress in Physical Organic Chemistry, Vol. 3, A. Streitweiser, Jr., and R. W. Taft, ed., Interscience, 1965, p. 323. (71) a) R. C. Peterson, J. H. Margraf, and S. D. Ross, J, Am. Chem. Soc., 83, 3819 (1961); b) R. C. Peterson, J. Org. Chem., 29:7 3133 (1964). (72) J. E. Leffler and E. Grunwald, Rates and Equilibria of Organic Reactions, John Wiley, New York, 1963, Chap. 9, p. 315. (73) a) G. A. Olah and A. M. White, J. Am. Chem. Soc., 82“ 3591 (1967); b) G. A. Olah, D. H. O'Brien, and A. M. White, ibid., 82” 5694 (1967). (74) a) G. Kohnstam, "The Interpretation of Entropies and Heat Capacities of Activation in Solvolytic Reactions," The Transition State, Special Publication No. 16, J, Chem. Soc., 1962, p. 179; b) G. Kohnstam, "Heat Capaci— ties of Activation and Their Uses in Mechanistic Studies,” Advances in Physical Organic Chemistry, Vol. 5, V. Gold, Academic Press, New York, 1967, p. 121. (75) J. R. Hulett, Quart. Rev., l§x 227 (1964). (76) A. Queen, Can. J. Chem., fiéx 1622 (1967). (77) B. L. Archer and R. F. Hudson, J. Chem. Soc., 3259 (1950). i‘lllijjll" u ... \9‘ .0. I . y K.“ 9 v g IIIIIII 1mm) 31 A 17.))11110111171111.1171)?!“