AN AUTOMATED ANALYSIS OF SECONDARY ISOTOPE EFFECTS IN AQUEOUS ALKALINE ESTER HYDROLYSIS Thesis for the Degree of Ph. D. MICHiGAN STATE UNWERSITY WILLIAM CHARLES SASS 1970 0-169 This is to certify that the thesis entitled An Automated Analysis of Secondary Isotope Effects in Aqueous Alkaline Ester Hydrolysis presented by William Charles Sass has been accepted towards fulfillment of the requirements for _EIL.D_.__ degree inshemistry (1 Aside Major professor Dam Aug. 24, 1970 "DAB & SDNS' 800K BINDERY INC. LIBRARY BINDERS . plums}. mm” ABSTRACT An Automated Analysis of Secondary Isotope Effects In Aqueous Alkaline Ester Hydrolysis By William Charles Sass The B-secondary kinetic isotope effect observed in the aqueous alkaline hydrolysis of ethyl acetate—d3 at 25° has been reported by Bender and Feng (l). Incorporation of this value into the results of a temperature study of the kinetic isotope effect conducted by Halevi and Margolin (2) revealed an unprecedented temperature dependence of the isotope effect. The values reported are: Temp 0° 25° 35° 65° kH/kD 1.00:0.0l 0.90i0.0l 0.93f0.0l l.15i0.09 Both kinetic studies were based on titration techniques. In this study the design of a family of automated instruments capable of a continuously following ester hydrolysis and other reactions in which acid or base is generated is discussed. Two of the instruments were built and found to perform well. The speed and accuracy of one of the instruments made it ideally suited to verify the important tempera- ture dependence reported by Halevi. Temperature studies on ethyl pro- pionate and ethyl isobutyrate hydrolyses were also performed. The results obtained, which displayed little temperature dependence, did William Charles Sass not agree With those of Halev1. The observed kunlabeled /kheavy ratios are: Temp 5.00° l5.00° 20.05° k/kacetate-d3 0.959:0.0l0 0.968:0.0l2 0.958:0.005 Temp 25.00° 29.80° 35.00° k/kacetate-d3 0.9623i0.006 0.974:0.005 0.970:0.004 Temp 45.0l° 50.09° 60.00° k/kacetate-d3 0.96710.008 0.970:0.0l3 0.988:0.0l7 Temp 20.02° 30.00° 40.02° 50.02° k/kpropionate-d2 .936:0.024 O.946:0.0ll 0.949i0.006 0.982i0.0l6 k/kpropionate-g3 .Ol6:0.0l8 l.OlS:0.0l4 l.034:0.005 l.OZl:0.0l5 Temp 30.02° 40.03° 50.02° 60.0l° k/kisobutyrate-dl .Ol8:0.022 l.0l4:0.022 l.02010.0l0 l.006:0.0l4 k/k .94l:0.0l8 O.907:0.020 O.928t0.008 0.923:0.0l2 isobutyrate-d5 William Charles Sass Activation parameters for each ester were calculated. The decrease in rate with substitution resulted from AS*, not AH*. Explanations for the observed rate ratios based on a threefold barrier to rotation have been offered. The normal isotope effect observed for ethyl isobutyrate-g} hydrolysis is taken as evidence for a twofold barrier. Torsional effects have also been considered. (l) H. L. Bender and M. S. Feng, J. Am. Chem. Soc., §g, 63l8 (1960). (2) E. A. Halevi and (Mrs.) Z. Margolin, Proc. Chem. Soc., 174 (l964). AN AUTOMATED ANALYSIS OF SECONDARY ISOTOPE EFFECTS IN AQUEOUS ALKALINE ESTER HYDROLYSIS By William Charles Sass A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry T970 To my old and new families ACKNOWLEDGMENTS The author wishes to express his appreciation to Professor Gerasimos J. Karabatsos for his guidance and for his assistance in interpreting these data. The financial support provided by the National Science Foundation and the Department of Chemistry, Michigan State Univer- sity, is gratefully acknowledged. Appreciation is also expressed to the author's wife, Evelyn, for her help in the preparation of this manuscript. iv TABLE OF CONTENTS LIST OF TABLES . LIST OF FIGURES HISTORICAL AND INTRODUCTION RESULTS AND DISCUSSION . Selection of the System . Analytical Methods Temperature Dependence Interpretation of Isotope Effects . Beta Effects Gamma Isotope Effects Solvation Effects . . Conclusion EXPERIMENTAL . Ester Preparation . . . . . . . . . . . . . . Ethyl Acetate- -d3 and Unlabeled Ethyl Acetate . Ethyl Propionate- 2, 2- d 3, 3, 3, d ; and Unlabeled. Ethyl Propionate ..... Ethyl 2- ~Methylpropanoate; Labeled and Unlabeled Procedure . . . Base Preparation . pH Monitoring . Constant Temperature Bath. Temperature Control Quartz Thermometer . . Rate Determinations Treatment of Data BIBLIOGRAPHY . . Page vi . 21 . 26 . 54 . 54 . . 54 . 58 . 59 . 6O . 62 . 62 . 62 . . 62 . 63 . 64 . 64 I 65 . 66 . 66 . 67 . 67 . 73 Table II. III. IV. VI. VII. VIII. IX. XI. XII. XIII. LIST OF TABLES Acidity Reduction in Deuterated Acids . Activation Differences Between Labeled and Unlabeled ISOpropyl Derivatives. Calculated Acitvation Differences Between Labeled and Unlabeled IsoprOpyl Derivatives. Substituent Effects on the Hydrolysis to Exchange Ratio. . . . . . . . . Substituent Effects on Alkaline Ester Hydrolysis Rates. . . . . . . . . . . . . . . . . Ethyl Acetate Base Hydrolysis . Ethyl Propionate Base Hydrolysis. Ethyl Isobutyrate Base Hydrolysis . IsotOpe Effects at 40°. Activation Energy Differences . Activation Parameters; Ethyl Acetate. Activation Parameters; Ethyl Propionate . Activation Parameters; Ethyl Isobutyrate. vi Page . l4 . l5 . 24 . 25 . 4O . 44 . 47 . 49 . 50 . 5T . 52 . 53 Figure l. 10. ll. 12. l3. l4. LIST OF FIGURES (a) Decreased Force Constants Resulting in Decreased Vibrational Differences. . (b) Increased Force Constants Resulting in Increased Vibrational Differences. Interference to Out-ofPlane C-H Bending Motions. Reported Temperature Dependence of k k Ethyl Acetate/ Ethyl Acetate-d3 . The Significance of Protonation Kinetics on the Hydrolysis . 23 to Exchange Ratio . Integrating pH Stat Integrating pH Stat Performance. (a) Calibration of pH . (b) Linear Calibration of PH Typical Linearity of Data. Typical Computer Output Steric Interactions in Ethyl Isobutyrate Hydrolysis . Second Order Rate Program 'Sec R' , , Temperature Dependence of Ethyl Acetate. Temperature Dependence of Ethyl Propionate . Temperature Dependence of Ethyl Isobutyrate. . vii Page . l8 . 28 . 3O . 34 . 34 . 36 . 37 . 58 . 68 . 39 . 43 . 46 HISTORICAL AND INTRODUCTION It is generally accepted that kinetic secondary isotope effects are the result of force constant changes occurring between the ground state and transition state (1). Figure 1 demonstrates the effect of force constant changes on the activation energy of unlabeled and deuterated compounds. A decrease in the transition state force constant, which corresponds to a widening of the potential well, is accompanied by a decrease in vibrational differences between the two compounds. Although by virtue of its higher vibrational frequency, the energy of the lighter compound still exceeds that of the heavier, the diminished difference will lead to a smaller activation energy and higher reaction rate for the unlabeled compound. This is the "normal kinetic isotope effect." When initial or zero point energy (ZPE) differences are exaggerated in the transition state by an increased force constant, the heavier compound reacts faster. This is referred to as an "inverse kinetic isotOpe effect." Since Lewis and Boozer (2) reported the first observed secondary isotope effect in l952, an intensive search for the origin of the force constant changes has been conducted. In order to make quantative pre- dictions it is necessary to have an accurate representation of both the ground state and transition state structures. While the former is generally obtainable by conventional spectrosconic measurements, the latter is not. ENERGY ENERGY Figure l (a). Decreased Force Constants Resulting in Decreased Vibrational Differences. Ea R-H Ea R-D Iff//fl,flfi,ullllu, R30 4- r 76m. ifieigloflft .. \\\\\\ ///I ’ H D ‘..../ Initial State REACTION COORDINATES Figure l (b). Increased Force Constants Resulting in Increased Vibrational Differences. Ea R-H) Ea R-D Transition State R'” : , Ea R-H> Ea R-D MK /’ }AZPEI Therefore. kH/kD < l \J Initial State REACTION COORDINATES 3 An empirical solution was presented by Streitwieser and coworkers (3). Force constants of a ”carbonium ion like" transition state were assumed to resemble those of aldehydes. The trigonal sp2 geometry of the aldehyde, along with its partial positive charge resulting from inductive effects of the oxygen, support the model. The principal frequency change in going to the assumed transition 1 decrease in the out-of-plane carbon-hydrogen state was a 550 cm‘ bending vibration. Since this corresponds to a kH/kD of 1.38, as compared to a typical observed value of l.l5 for an SNl a-effect, an explanation for the discrepancy must be sought. This explanation is as follows: The transition state is reached with the departing group still in the vicinity of the developing carbonium ion (Figure 2). If its presence were assumed to decrease the frequency difference vH-vfi to about 300 cm'] , the discrepancy between the experimental and the calculated kH/kD value disapears. It should be noted that this explanation is consistent with the absence of a-isotope effects in 8N2 solvolyses. Interference with the C-H bending by both incoming and leaving groups may lead to no frequency change and thus to no isotope effect. / *2 “Q2 (b) (C) H H Figure 2. Interference to out-of—plane C-H bending motions: (a) in aldehydes, (b) in SNl, and (c) in 5N2 transition states. 4 Various theoretical approaches to the transition state structure have been reviewed and compared by Weston (4). Wolfsberg and Stern (1,5) have developed a computerized analysis of kinetic isotope effects based on the assignment of atomic weights, molecular dimensions, and arbitrary force constants to the initial and transition states. Their results confirm the assumption that both a and 8 isotope effects must be accompanied by force constant changes at the site of isotOpic substitution. The authors (6) later reported that acceptable solutions may be approximated by ”cutting off” (and replacing with a rigid comparable mass) portions of the molecule more than two bonds removed from the reaction site. Despite general agreement that secondary isotope effects are caused by force constant changes, the origin of the change remains unclear. Extensive research has led to three interacting effects (7a). They are: hyperconjugation, an inductive effect, and a steric effect. The principal parameters affecting hyperconjugation are the relative electronegativities of hydrogen and carbon, reflected in the Coulomb integral, and the tightness of binding of the electrons in the methyl group, reflected in the overlap integral. Although LCAO-MO calculations on systems containing fully formed carbonium ions are domi- nated by the Coulombic term (7b), it is not unreasonable to assume, in view of the shorter range of exchange forces than Coulomb forces, that in a reaction in which the carbonium ion is being formed by rupture of a CX bond, the effect manifests itself earlier on the overlap integral than on the Coulombic integral (76). As Streitwieser and coworkers (3) have indicated, the result of increased hyperconjugation in the transition state is reduced carbon— 5 hydrogen (deuterium) force constants. This reduction (Figure 1) leads to normal isotope effects without recourse to preferential hyper- conjugation by hydrogen. Since the potential well, as illustrated in Figure l, is unsymmet- rical, the hydrogen, with its greater amplitude of vibration, will be characterized by a larger average bond length. The shorter mean CD bond length will result in a larger D...D repulsion. This increased repulsion enlarges the DCD bond angle. Intuitively, this leads to less p character in the C-D bond, thus making methyl-g3 a better electron donor than methyl (7c). It appears safe to generalize: (a) Increased transition state hyperconjugation stabilizes methyl more than methyl-d3. (b) Methyl-93 donates electrons through the bonds better than methyl. (c) Methyl represents a larger steric bulk than methyl-d3. Interpretation diffi- culties arise in part from the contrary nature of (a) and (b). A fundamental paper by Shiner (8) firmly established the importance of hyperconjugation in kinetic isotope effects. The bicyclo compounds 1 and 2_and their unlabeled counterpart contain C-D(H) bonds which are respectively parallel to and orthogonal to the developing vacant orbital in the limiting solvolysis. CH3 CI CH3 1 one go: 1. Z 6 The predicted increased stabilization of the transition state of l_by hydrogen hyperconjuoation is supported by the observed rate ratio for l of kH/kD = 1.14. The geometry of 2 prevents overlap by the labeled C-H bond. Absence of hyperconjugation results in an observed ratio of kH/kD = 0.986, pre- sumably resulting from better inductive stabilization of transition state by deuterium. A more recent paper by the same author (9) reports isotope effects in the acetylenic systems 3 and 4. _§_ (CD3)2 CClCECCH3 _4_ (CH3)2 CClCECCD3 kH/kD=l.655 kH/kD=l.O92 Since inductive and steric forces cannot Operate through such great distances, the effect must result from hyperconjugation acting through the n system of the acetylene. The general concept of hyperconjugation is based on the Baker-Nathan order. It has been observed in many systems (10) that carbonium ion stabilization occurs in the following order: Me > Et > j;Pr > tyBu > H. This series parallels the number of hydrogens available for hyperconjuga- tion. An opposite rationalization for this violation of the accepted inductive trend claims that the larger groups inhibit solvent stabiliza- tion (ll). Serious doubts over this explanation have been raised by Berliner (12). In a review of the literature he showed that in at least 40 systems, fl substituents followed the Baker- Nathan order in electro- philic reactions while substituents in the getg_positions, from which conjugation is not possible, followed the normal inductive order, 7 tyBu > ifPr > Et > Me, despite their closer proximity to the reaction site. Based on the small rate differences observed in the Baker-Nathan series, Taft and Lewis (13) and Brown and coworkers (14,15) have suggested that some stabilization by carbon-carbon hyperconjugation is occurring. The importance relative to carbon-hydrogen hyperconjugation was calculated to be H/C = 1,310.1 and H/C = 1.25, respectively. Perhaps the best view of the controversial question (16) of carbon- carbon hyperconjugation was recently expressed by Jensen and Smart (17). The authors indicate that the only unequivocable evidence for carbon- carbon hyperconjugation occurs in systems containing excessive p charac- ter in the carbon-carbon bonds. Eg:g_substitution by cycloprOpaneincreases the rate of tfcumyl chloride (5) solvolysis by’a factor of 157, while isopropyl substitution increases the rate by only 17.8. ’( H U 5 'CH3 CH3 The rate enhancement, presumed to come from preferred C—C overlap in this strained system, can be substantially diminished by the effects to two added methyl groups (6) which stericaflyprevent the carbon orbital from becoming parallel to the n system. This results in a rate enhancement over the corresponding isopropyl substituted tfcumyl chloride by only a factor of nine. The importance of geometry for strained ring assistance in solvolysis has also been reported by Schleyer (17) in the cyclopropyl adamantyl system. Halevi and coworkers (18), in a careful study of pKa values, have found a decrease in the acidity of deuterated aliphatic acids and 8 benzylammonium ion, as shown below. They attributed this decrease to differences in the electronegativity of hydrogen and deuterium. Table I. Acidity Reductions in Deuterated Acids. Acid 925a CD3C00H 0.026 CH3CD2COOH 0.034 CD3CH2COOH 0.007 PhCDZCOOH 0.048 PhCDZNH: 0.054 Increased electron donation by deuterium over hydrogen was believed to destabilize the anion and cause the resulting decrease. Variation in dipole moment and polarizability of NH3 (19), increased shielding of hydrogen (20) and fluorine (21) in nmr spectra by deuterium substitu- tion, and chemical evidence such as the inverse isotope effect cited above by Shiner (8), are among the more compelling evidences supporting deuterium being more electropositive than hydrogen. Brown and coworkers (22) have shown that in the Lewis acid-base neutralization reaction between boron trifluoride and 2,6-lutidine—d6 (Z), a stronger bond is formed than between the unlabeled species. A‘k 0 CD3 CD3 I N / 7 8 Absence of isotopic effects when 3,5-1utidiene (8) was used, based upon the difference in heat of reaction, tend to discredit inductive and hyperconjugative effects in this system. The authors attributed this isotope effect to non-bonded interactions. Further support for the presumed steric effect was also obtained when it was determined that 9 trimethyl-gs-boron formed a 25% stronger bond with ammonia than did the unlabeled compound in violation of the predicted inductive effect. Mislow and coworkers (23) reported the observation of a purely conformational normal isotope effect of six per cent in the twisted 92 ring system (9). Known not to proceed yia_enolization, the racemi— zation of (9) became the first reported isotope effect not involving bond making or breaking. \ CD3 003 D c3-J l <©> to U 23 Mislow and coworkers (24)—also reported another conformational kinetic isotope effect in the racemization of the stericanyrnon-planar 92 ring system (l_). The larger steric requirements of hydrogen over deuterium are presumed to be responsible for the higher energy required to racemize the hydrogen compound. Another example of nonbonded effects upon racemization kinetics was reported by Melander (25) in the 02 system of 2,2:dibromo-4,4: dicarboxy-6,6:biphenyl-d2 (1 ). Part of the 18% inverse isotope effect Br Br D 4:E::::> COOH HO0C< <:::§E;:E;jE;COOH : ll // 1g D '7‘ Br D observed for this system might have arisen from inductive differences between hydrogen and deuterium upon n orbital overlap. In consideration of this, the 10 I corresponding biphenyl with deuterium in the 5,5 position (l_) was synthesized (26). Racemization yielded no isotope effect outside experimental error (1.0210.02). H. C. Brown (22) has stated that virtually all deuterium isotope effects may be best described in terms of steric differences. L. S. Bartell (27) has gone beyond this and claimed that even such physical phenomena as the shorter bond lengths and the smaller than predicted bond angles in alkenes obtained by electron defraction studies (28) may be explained by nonbonded interactions without recourse to hybridization or hyperconjuga- tion effects. Work in this laboratory by Dr. Sonnichsen (29) compared experimental values with those predicted by Bartell's model (30). The system studied was the family of l-naphthoyl chlorides and methyl esters with IT, D, CH3, and CD3 substituents in the eight position. In this system, which pre- cludeshyperconjugation, satisfactory results between experimental and calculated isotope effects were obtained. Predictions were not as satis- factory,however,in t;butyl-g3 chloride, ethyl acetate-d3, and acetyl-dg chloride,where hyperconjugation can occur. Excluding special conformational factors (8), the reduction in the B-CH force constant in the activation process of limiting solvolysis appears to be related to hyperconjugative interaction with the developing vacant orbital. Despite the arguments already presented supporting a steric interpretation of isotope effects, it should be noted that they generally appear only when hyperconjugation is precluded or highly hindered species (23,24,25) are involved. The solvolyses of 3-pentyl brosylate, 2,4-dimethy1-3-0enty1 brosylate, and their 3 deuterated analogs were examined by Shiner and Stoffer (31). 11 The observation of cumulative effects in the former as one proceeds from the unsymmetrical doubly deuterated to the tetra-deuterated species (1.3171 2 7TT7365) was interpreted as indicating a hyperconjugative effect. The difference in isotope effect resulting from mono and dideuter- ation of the latter compound (1.3671 > /2TTT79), however, indicates that irregularities are occurring, probably resulting from participation. The importance of the leaving group on the magnitude of secondary isotope effects has recently been emphasized by Shiner and coworkers (32) in the solvolyses of metg_and pa:§_substituted l-phenylethyl chlorides and bromides and their e-g and e-d3 analogs. The 15% a isotope effect observed upon chloride solvolysis was unaffected by H, metafmethyl, pgra; fluoro, pgrgfphenoxy, and pgrafmethoxy substituents. It was concluded therefore that all solvolyzed by a limiting mechanism. Progressively smaller 8 effects observed in the same series for compounds more reactive than l-phenylethyl chloride were attributed to a reduction in the demand for hyperconjugation. Despite the similarity of the mechanism for the chlorides and bromides solvolyses, as evidenced by their similar 8 isotope effects, the bromide a effect was three per cent lower than the correspond- ing chloride. The difference is a reflection of weaker ground state bend- ing force constants in the bromide. Trifluoroethanol, with its high ionizing power and low nucleo- philicity, has found acceptance (33) as a solvent favoring a limiting mechanism without posing the difficulties to precision conductance measurements that acetic and formic acid do. In the above paper, Shiner and coworkers also reported the same rate ratios for p-methyl- benzyl-a-g2 chloride in 94% trifluoroethanol-water (1.146 per D) and 70% trifluoroethanol-water (1.140) as had been reported for a—phenyl- ethyl-a—d chloride in various ethanol water mixtures (l l46-l.153). 12 The contrast between the similarity of rate ratios observed in struc- turally different chlorides and the lower isotopic rate ratios (by three per cent) of each of the corresponding bromides led the team composed of Shiner, Rapp, Halevi, and Wolfsberg (34) to postulate that the leaving group, and not the alkyl group, dominates the a-isotope effect in limiting reactions. Their calculations, which were based on transition state theory, assumed comparable bond strengths in the transition state for various halides. The correspondence between experimental and calculated results supports their hypothesis (35) that ground state differences are the source of variation in a kH/kD effects in limiting solvolytic reactions. In a study of incoming and leaving groups, Jackson and Leffek (36) evaluated the rate ratios obtained in thiosulfate displacement reactions on methyl bromide, iodide, methanesulfonate, and p-toluene sulfonate- The rate ratio for methyl p-toluenesulfonate (1.12) is of special sig- nificance since it is the largest reported isotope effect observed to date in nucleophilic displacement reactions. They added their results to the correlation pr0posed by Seltzer and Zavitsas (37). This is an observed linear relationship between the kH/kD values and the difference in pucleophilicity of the incoming and leaving groups. The explanation given for the relationship, based on Hammond's postulate (38), is that a good incoming nucleOphile will make little bond at the transition state. Thus, the decreased steric requirements (or decreased force constants) result in a normal isotope effect. Conversely, a weak incoming nucleophile will be extensively bonded at the transition state and thus lead to an inverse isotope effect. Between these two extremes, 13 incoming and leaving groups of comparable pucleophilicity are predicted by Seltzer to lead to no isotope effect. In an attempt to verify this latter prediction, Seltzer and Zavitsas (37) measured the a isotope effect in the displacement of methyl iodide with isotopic sodium iodide. Their results in methanol (1.05) and in water (1.10) were far from the predicted value of 1.00. In order to rationalize these numbers, they postulated a symmetrical energy profile containing two peaks. The first transition state was reached with little bond making or breaking. As the incoming group moved in closer, stability increased leading to a symmetrical inter- mediate. Stretching of the second bond raised the energy to the second transition state and finally to products. In the preceding discussions,extensive reference has been made to the importance of substrate structure and conformation of the alkyl group on secondary kinetic isotope effects. The significance of enter- ing and leaving groups has also been mentioned. So far, few references have been made to solvatiop effects, even though all the preceding reactions were performed in solvent. The neglect of solvatiop effects is due not to their unimportance, but to the unfortunate lack of quantitative information regarding these effects. Hakka and coworkers (39) have reported that a 25° temperature increase decreases the isotope effect in the solvolysis 11("333011 + OCH;- Figure 4. The Significance of Protonation Kinetics on the Hydrolysis to Exchange Ratio. If the protonation-deprotonation steps are fast, the hydrolysis to exchange ratio reduces to kh/ke = 2k3/k2. The effect of substituents on k3/k2 is expected to be small; the main effect occurs in k3/kh. Both rwates should be increased by electron donating substituents, but k3, Which is one bond closer, should be more strongly affected. Table IV "Eaflects the effects of varying substituents where R = p;XPh—e 24 Table IV. Substituent Effects on the Hydrolysis to Exchange Ratio. -1__ _2__ kh/ke NH2 -0.66 3024 CH3 -0 170 1132 H -0.00 5.7:0.8 c 1 +0.226 6.3tl.10 NO? +0.778 2.810.3 Although the lifetime of the tetrahedral species must be extremely short to allow hydrolysis to compete with rapid proton exchange (k3 2 kg 2 109 sec-1), the occurrence of exchange indicates that the intermediate has a lifetime significantly greater than 10'13 second, the time required for a molecular vibration (56), and as such may be considered an intermediate. It has been reported by Johnson (56) in his review of nucleo- Dhilic catalysis of ester hydrolysis and related reactions that k1 is the slow step in acidic, neutral, and alkaline hydrolysis of oxygen esters, in the acidic hydrolysis of amides and anhydrides, and in the neutral hydrolysis of anhydrides, acid halides, and acetyl— imidazolium ion. Only in the alkaline hydrolysis of amides has k3 become the slow step. The effects of electronegativity and crowding on alkaline hydrolysis rates are indicated in Table V (57). 25 Table V. Substituent Effects on Alkaline Ester MeOAc MeOAc EtOAc Hydrolysis Rates. CH3COOMe CHzClCOOMe CHC12COOMe 1.0 761 16,000 0 l (tonne)2 CH3C00Et CH3—C-C00Et 170,000 0.60 10,000 CH3C00Et CZHSCOOEt (CH3)2CHC00Et (CH3)3CC00Et 1.0 0.47 0.10 0.011 CH3CH20AC (CH3)2CHCH20Ac (CH3)3CCH20AC (C2H5)3CCH20AC 1.0 0.70 0.18 0.031 While it could be argued that some of the rate inhibition by alkyl sub« stitutions on the acid could result from inductive effects of the substituents, the effect of substitution on the alcohol moiety must be steric in nature since the alteration is too far (four bonds removed) from the reaction site. The deuterium isotope effect on alkaline ester hydrolysis is ‘therefore an ideal system for study, because of the preliminary indica- t‘ions of an unprecedented temperature dependence, its ”reverse nature” 26 in comparison to limiting solvolysis, and its irreversible hydrolysis known to proceed via a single mechanism for which the rate determining step has been established. II. Analytical Methods Regrettably, many common techniques are inapplicable to this poten- tially valuable system. As noted earlier, most isotope effects have been determined by conductance measurements. Solvents are generally nonaqueous and initially contain no ionic species. Application of conductance measurements in such solvents as acetic acid are generally avoided because the self—ionization of the solvent introduces a sig- nificantinitial conductance thereby reducing accuracy (33). An examination of the stoichiometry of the ester hydrolysis reveals that it would require not the usual following of the sum of the conduc- tances of the ions generated, but only the difference in conductance between hydroxide and acetate. While this difference is measurable, it represents a substantially smaller change superimposed on larger con- ductances than is generally measured and will result in a correspondingly lower resolution. Ultraviolet Spectroscopy with the commercial availability of thermo- stated cells, has become a convenient technique for following kinetics. Although esters contain a chromophore, the carbonyl must be conjugated with an unsaturated system before the extinction coefficient becomes Significant. The extinction coefficients of acetic acid and ethyl aCetate are 32 and 57 at Amax = 208 and 211 nm, respectively. This CKDntrasts with an extinction coefficient of 13,500 at Amax = 206 pm (58) ff)? gisfcroto ic acid (C=C-C02H) in the same solvent (ethanol). Clearly, 27 ultraviolet spectroscopy is inapplicable to kinetic measurements in this unconjugated system. Olah and Streitwieser (59, 49), among others, have applied nmr spectroscopy to kinetic measurements. Unfortunately, the tendency of the nmr to drift and the difficulty in measuring and maintaining accurate temperatures make this technique less than ideal. Competition studies between labeled and unlabeled species have been widely applied (60,6l). Application to this system would involve the hydrolysis of a known ratio of labeled and unlabeled ester. 4liquots are periodically withdrawn, quenched, and the ester extracted. Mass spectral or nmr analysis of unreacted ester will reveal the relative quantities of each remaining at various times from which the kinetic ratio kH/kD may be calculated. One advantage of this technique is that isotopically pure starting materials are unnecessary. One general disadvantage not applicable to this system is that isotope effects subsequent to the rate determining step may divert unequal portions of labeled and unlabeled material down alternate mechanistic paths leading to other products, thereby invalidat- ing a study based on product ratio analysis. In addition to its tedious nature, competition was ruled out for this study, since only relative and not actual rates are revealed. Having eliminated most of the common techniques by which substrate disappearance could be monitored, one seeks alternate techniques. As the ester hydrolyzes, one mole of base is consumed for each mole of ester that reacts. This formed the basis for the two preceding kinetic studies mentioned (47,48). Both consisted of quenching and titratinq aliquots of reaction mixture. 28 In addition to the extremely tedious nature of titration studies involving the proposed analysis of the eight labeled and unlabeled esters at a total of 17 temperatures with three to six independent determinations each, it was desirable to know as many points as possible along the reaction path. Continuous methods of analysis were therefore considered. Commercially available pH stats are capable of monitoring the kinetics of reactions in which acid (or base) is generated or consumed. Although pH stats may be used to follow kinetics of acid forming or consuming reactions, the technique involves experimental difficulties. The error signal resulting from deviation from the preset point drives a proportional rate pump to restore the balance. This implies that by virtue of the fact that titrant is being added, the solution is necessarily not at the preset pH; the instrument is not a ”pH stat.” It should be noted that errors as small as 0.05 and 0.01 pH units result in 13% and 2.6% errors, respectively. It was thus desirable to design a pH stat which would avoid the consistent deviation from the preset point. The author designed and built an analog integrating servo systeni which avoided this difficulty. The circuit is shown in Figure 5. Summing Amplifier Pulse Titrant Sen51tivity Control Integrator Generator Adder 1 Meg 1 10 K -~’ ' Signal “V '}%1_l 1 Meg 10 K W“ Reference 0” Figure 5. Integrating pH Stat 29 As the signal begins to deviate from the reference, this error signal begins to charge the integrating capacitor. Charge (and thus addition rate) increases with time, until addition is occurring at a rate which will curtail further drift. The error signal now existing, however, continues to charge the integrating capacitor, resulting in ever faster addition until the error signal has been forced back to zero. At this point the charge ceases to grow, but the existing charge overdrives the titration until the negative error signal subtracts enough charge to slow the addition rate to that of the reaction rate. A situation opposite that above now exists and the negative error signal continues to remove charge until the reaction rate exceeds the addition rate and the solution returns to the preset point from the negative. These oscillations are shown in Figure 6 for the case of 35 p1 ethyl acetate in 15 ml 0.005 N KOH (K=O.ll l/mole sec). The result is a 0.015 maximum pH deviation with 0.00 average deviation. It is noteworthy that the amplitude of the maximum deviations (which correspond to the points at which addition rates just compensate for the effects of the reaction) decreases as the reaction slows. The result is a dampened sign wave with an average value coinciding with the preset value. Although the analog system functioned well under test conditions, it contained two defects. First, any small offset voltage will contin- ually be integrated, thus making the analog system prone to drift with time. Second, it is not adequate to have the pH value average the selected setting; it is the [H+] which must average a constant. Since the two are exponentially rather than linearly related, one does not imply the other. The latter deficiency may be corrected by inputting 30 *u.-- -- Y -. T _----,_ ........_...1..... .. _-. ._. 70.1 0.2 0.3 0.4 0,5 0.6 07 0.8 01'9"" 1“. r... ..- 0.0 71 One pH Unit a \:‘——“—‘Time Zero \ f 5 f‘; K\ k" \\ Figure 6. Integrating pH Stat y Performance; pH Versus C Time. \ \ ON \\\\, . oe mo-um~.l flcsmoommhél _ouua¢_mq.n Ahv¢©q¢. :_|uu:«xx. :;.4.:fi re._P . . ~w. mc+wrcmcq. :_uuqmc:_. :;:,.:H, -c. e- . . .~.; mc+mUc__¢. :_Iu¢1>-. ::r~.._ .c. _ ”.6 an. mc+ummm~m. :Hadcn¢n~. ::<%.:. . _ a..« :«i - mc+ucrcmr. c.lm(m_3_. :err.r. z.._p :.m an. mc+ummo®m. a_luc3rafi. cccr.<< c3. . .r.e -oma;-- - oc+u>mn¢m. c.1y_mn._. :cr3.t. ..~ ...A on. mc+m~>mmm. c.su~:<:fl. em.<.,w .ea ..d .: em.. mc+um¢oa~. __uu¢. fl_um~mmtm. ::Jn... ,.._P cu. «n. .H.> 2C~h<3p2mczou Z:HIL 13 dfnflm :-;1.wuhjfl~fl.w. izthNc.f-.g ccc.uruozu» Icr .3 m. 3 _ :ec_: :1 r1ac< ammuaaoa\fl «OH x x maz m~nmh s.\.~.\» -~\.- .s\.\~.a\ 4l qqoo.oH0aom.oumx\mx I. Q m moo.o+omwm.ou x\ x moo.omnmmc.oN Nmmo.oN omqm.o~ qum.oN mwmm.om mwmm.oN ma ooo.qmmfima.mfi mmmmqmw mqmm.ma mwom.ma mmqm.ma moqa.om mm moo.onooo.mmua mqo.qwooe.QH moo.QHoAm.qa wwwqu .mmmqmw qu.qH omn.qa mee.qa Hmm.qfi mom.qa mmq.qa mm mm moo.ommom.mmua I. m ooo.o+m~oa.ounx\ & omo.omammm.aa wmmH.HH mmmN.HH cqqm.aa omom.aa «mmH.HH mo mmo.oH¢mm.oa omnw.oa omnm.oa m0qm.oa Hmow.oH Hmmfi.oa mm moo.omooo.mmua A.pcouv .H> mpnmh .\A\ \- “Ho.qmmmmm.ouox\mx 42 mm.oumoqm.mn ommq.mm mqmo.mm mamq.wm Nwmm.mm wqmm.mm mo N.Amm~oom.mu mammumm. mmma.mm oqm©.mm ommm.mm momm.wm m m oao.owqoo.ooua I. m mao.o+momm.ouox\ x qm.9flm~ofi.mq Nmmo.mq mmom.mq mowm.mq Hmoq.qq m a me.9flm~u~.mq Hook.qq ommm.mq mamm.mq w~mm.~q Nomw.mq mm noo.onmo.omuH Q .I m omoo.o+amm.on x\ x AN.qumomm.nm mamm.nm Nemw.em oawm.em “Acq.mm oqwm.wm ma ooo.ofldao.mqua A.pcoov .H> afipah Ha.oquoflo.om mmmo.om wqqm.om ooao.om Humo.om ooam.cm mm 43 . om.o _ No.0 w \, - e i\\\\\\\\\\\\\\11111 M em.o y \ w m U \x\l\\\\ 5 mod __ \\\\i-\il . mod . Q I l ._ w i x om? 0mm _. omN H P * O _ 111111111111 31111111 will. N3 11 11111 i M .. . co; mo; aw-mpacowaoca _>:pw n x am-apacow00ta _»;pm u o mo.P mucmucmamo mczmemQEm» .mp mczu_d “ o_._ 44 Na mqao.ommao.fiummx\mx HHo.onmqm.oume\:x wac.9floao.uumpx\mx «No.0Hdomm.ou x\mx oa.9flon.aa omw.0fldmm.fifi mom.omqmq.NH .mmmqu moo.oHon.o oma.qmmqm.o mo.oHqu.o .Nwmwww .ummmqmw mmm.HH .mmmqm mamam. .mmmqm qmo.aa mmm.~a Nom.HH mflq.o 0mm.o onq.o qu.HH cam.NH mam.HH Aq~.o mmo.o onm.o nmA.HH qu.~H mno.HH H¢~.o Nae.“ Rom.o mm mm mm mm mm mm oao.oudoo.omue oooo.ommmo.oNuH moHaam me< ummnmfioa\a NOH x x maz mFQme 45 m N mpo.omfimoau Qx\mx ofio.owmmwo.ou mx\:x 00.0Hmom.qm omq.qm om~.qm oom.qm ma m.oHNHn.mm .mmwwmm mam.©m mmo.mm ego.om No omoo.oHAHo.omuH em.qwomo.mm .mqumm qma.mm mafi.mm Ace.qm ooo.mm m: m N I. a m .1 m moo.o+qmo.au x\ x qwoo.0+~omqm.on Qx\ x oqo.oHNHn.mH .mqumw moo.mH www.ma mo Hfi.ommq.am mmmqwm omq.H~ qwe.a~ Na oooo.QHoHo.oeuH 7:08 . :> 2an mo.omnmm.ow qwm.o~ mm~.o~ MNm.ON mm 46 I-" xi 1--———a<—-—~1 0mm r..- u~4u om¢ Pee—m, \J- x em-apac»pDDOmH _%;pw aw-mpmetznomH Fxgpm o we mucmvcmawo mczpmcmcEmH .e_ acamra 0mm ‘_..._______-__'__-,_. -.- V, . (I) O O 47 0 Ho 0 H .I .. .. I. m oNo.o+moa.ou ae\mx Nmo.o+qao.au x\mx wHo.o+qu.ou ax\mx NNo.o+wHo.Hu mx\ x mo.ommmm.o H.0H0qq.© mo.ommam.q oo.ommum.m oo.omoqo.q no.0HmoH.A .mmmum .mmmnm .wmmqm mmmqm .mmmwm Mmmam. «Hm.© mam.© mam.q 0mm.m wmo.q New.“ omfi.o Nom.© Hmm.q NHo.q amo.q mma.o Hmm.© ANm.© Nmm.q 0mm.m onm.m on He mm on an mm omoo.ommmo.oquH omao.omoao.omue moHH< ummlmHoE\H NOH N x mHZHDmOmH qwmem .HHH> m—nmh 48 o H NHO.QHmmm.on Qx\mx qa0.QHooo.Hu Qx\mx .HHHHoaH mmmqmw www.mH mH¢.mH on HH.QHHw.qH omao wNH.mH mam.qa omo.qa Ho .QHNHo.oouH MH.onH.QH com.qa qmm.qa mo©.qH coo.qa m : o H woo.oHHHNm.ou mx\mx OHo.oHo~.Hu mx\mx oo.om¢mo.NH moH.NH HNO.NH mam.aa «wo.ma on A.pcouv .HHH> m—nmh Ho.ommmm.0H mmmqmw mwm.OH Hmm.oH moa.oH Ho oHo.oHHHo.omuH wo.QflHoH.HH mHo.HH «HN.HH qu.HH HNN.HH mm 49 .mwmszcm mmcmzom HmmmH co ummmmm em wmm._ omm.c -mpmcstn0mH me C m . _. ch._ m-mumc>p:LOmH mmmm.c mo.H m-¢pmccwmoca HQN\ANQ.HV mmo.H mmqo.o o\Humwwm wQOHOmH >>mm oma.o mHm.o £¥\x Hmcowcoca .mImHmHooe mace Hm mpoacew mQOHOWH .XH aHan 50 NHHHwnu mm.owoq.c- $1 U-mHmLAH:QOmH .HHH> cmsoccu H> mmHan cw morgue mpg; co ummmmm wmwo.mm- mmwo.wpu< HaHos\Hmov m:<-e:< I A.mmv mHoE\—muv .x wpamh o .1 m<-mm< 51 .HHH> gazetgp H) maHnap cw apes 05H :0 women meoHHmHaonum mo.mH+ o_.m_+ Am_oe\_muxv ewe mmo.owowm.oH+ mm.chq.mm- Hmo.oHNN.HH HH.owmm.H .mImumpmu< chum HH.chmo.oH+ om.cHoH.NN- mmo.ome.HH Ho.owmwm.H mpmHmuq ngpw AmHoe\Hmo¥v A.cmu mHoE\Hmov AmHoe\Hmuxv AN-oHV< *IQ *mq mm mapmpao< erpm mmmme easocep H> maHnaH e? mpme aeH :0 comma meomeHsonum mm.opvoo.cmu mmc.onmm.o— m<_.cwcmm.omu mmo.onwm.c_ @mH.onqm.mmu mwo.owomm.cH A.omn mHoE\Hmuv AmHoe\Hmuxv «ma 8 mmumcowgoca H>cum mamHmzHeu< No.0.mm.e .m-apmeoH00ta H>£Hm om.o,mo.m .m-apacow30ta Hzepm am.o.ee.© apacoraoca Heepm H©-oHV< .HHx mHDmH 53 .HHH> eazocep H> maHneH CH apme asp co mamas meoHHaH:o_mua mm.om+ mm.cm+ «m.om+ AmHoE\Hmuxv *oq m¢.cH¢m©m.m o¢.chmmq.m ow.cwwcmm.m AmHoE\Hmuxv *IQ N¢.me©H.mmu omo.owmom.w mm.HHmmH.nm- omc.chmc.m mq.Hqum.um- cmo.cwqom.w A.omv m_oE\Hmuv Am—os\_muxv ewe mm maHaLAH=DOmH HALHM mempmzHHQ< mH.0wmm.H w_.oHom.H mH.ono.— Am -o_v< em-ame>H3DOmH Heepm .w-mpmc>p=n0mH Heepw mHac>H=QOmH ngpm .HHHx m—an 54 III. Temperature Dependence The observation of a minimum value for the rate ratio as reported by Halevi (47) is clearly inconsistent with the results of this study (Tables VI-IX). The explanation offered by Halevi for the reported minimum was based on assumed differences in energy of solvatiop resulting from isotopic differences in average charge distribution. These varia- tions in solvatiop should be reflected in AAS*, the difference in activa- tion entropy between the labeled and unlabeled esters. No substantial difference has been observed (Table X). Examination of AH* and A5* for the ethyl acetate, propionate, and isobutyrate hydrolyses (Table XI), and their relative contributions to the hydrolysis rates, shows that the decrease in reaction rates within this family of esters is not caused by greater ground state stabilization resulting from the inductive effects of methyl. Indeed, the activation enthalpy favors isobutyrate solvolysis. The effect comes instead from the 10 eu difference in the entropies of activation. IV. Interpretation of Isotgpe Effects A. Beta Effects The ground state conformation of ethyl acetate (1_), in which one hydrogen is eclipsing the carbonyl oxygen, has been inferred from the preferred conformation of acetaldehyde (15), which has been determined by microwave (67) and electron defraction studies (68). The observed threefold barrier to rotation about the spZ-sp3 carbon-carbon bond in Hb 0 ”b 0 H5; 14 : Ha, ; : Ha — 0C2H5 Ha H 55 acetaldehyde is 1162130 cal/mole (67). A similar conformation has been reported for acetic acid based on microwave analysis (69), but with a lower threefold barrier to rotation (483:25 cal). The geometry of the two Ha protons is suitable for orbital overlap with the carbonyl orbital. Hyperconjugation will be more effective in the ground state than in the transition state, and will therefore lead to an inverse kinetic isotope effect (Table IX). Hyperconjugation is expected to overcome the opposing inductive effect of the Hb proton which is in the wrong orientation for hyperconjugation (8). Steric effects in this uncrowded ester are expected to be small (29). Although the microwave analysis of propionic acid has not yet been reported (70), the conformation of ethyl propionate may be inferred from the preferred methyl eclipsing of the carbonyl (1_) observed in aldehydes (71) and the conformations of several compounds containing a single halogen atom alpha to a carbonyl. The AH° for the equilibrium 16_: _Z_ is -560, -500, and O cal/mole for ethyl fluoroacetate, chloroacetate, and bromoacetate, respectively (72); the AH° is -1000 cal/mole for bromoacetyl chloride and -l9OO cal/mole for bromoacetyl bromide (73). For chloroacetaldehyde and bromoacetaldehyde in dimethyl sulfoxide, AH° is -1500 cal/mole and -700 cal/mole, respectively (74). All of these rotational isomers are best described in terms of a threefold barrier to rotation about the spZ-sp3 carbon-carbon bond. The per- centage of ethyl propionate in the conformation with methyl eclipsing the carbonyl oxygen is expected to exceed that observed with halogen 56 eclipsing since dipole-dipole interactions have been reduced. This assumption is supported by the observed AH° of -800 to -900 cal/mole (71) for propionaldehyde. As in the case of ethyl acetate, ethyl propionate (l_) has two Ha protons in a position favorable for hyperconjugation which results in an inverse isotope effect. Steric interactions, while greater than those in ethyl acetate, are not expected to be severe since both the hydroxide and the carbonyl oxygen experience only one methyl interaction in the transition state (19). Hyperconjugation is therefore the likely cause of the observed inverse isotope effect. 0 CH3 OC2H5 13 When one or both substituents on disubstituted acetaldehyde are methyl, methyl eclipsing of the carbonyl oxygen (20) is preferred (71). H "' 20 X "I 2] X "H x H x —— H H 22 H Dihalo substitution in aldehydes results in preferential halogen eclipsing (_2) in polar solvents and hydrogen eclipsing (_l) in non- polar solvents (75). Hydrogen eclipsing in dichloroacetyl chloride (76) is preferred by 200 cal/mole. All of the above are best described by a threefold barrier to rotation about the sp2-sp3 carbon-carbon bond, but all are also characterized by much more positive AH° values for 57 In contrast to mono and disubstituted aldehydes, mono and disub- stituted acid halides, and monosubstituted esters,all of which exhibit threefold barriers to rotation, ethyl difluoroacetate and ethyl dichloro- acetate exhibit twofold barriers to rotation with AHO for 23_: 24 being +25 cal/mole and O cal/mole, respectively (72). H o X 0 x _‘ ‘> 2 X .2_3_ OCZHS H — OCZHS Although the rotation barriers of propionic acid and isobutyric acid have not yet been reported (72), the barrier in acetic acid (which is structurally similar to the ester) is substantially lower (483 cal/ mole) (69) than the barrier in acetaldehyde (1162 cal/mole) (67). Acetaldehyde, monohaloacetaldehyde, and dihaloacetaldehyde all exhibit threefold barriers to rotation, while the change from ethyl monohalo- acetate to ethyl dihaloacetate results in a change from a threefold barrier to a twofold barrier. It is not unreasonable to expect that other disubstituted esters, including ethyl isobutyrate, may also exhibit a twofold barrier to rotation 22.: 24, 58 In both conformations, §§_and gs, the orbital of the methine proton is orthogonal to the orbital of the carbonyl. Since this geometry precludes hyperconjugation, the observed normal isotope effect must be the result of inductive differences between hydrogen and deuterium. B. Gamma Isotope Effects In light of the above structures(l§, 23, and 23), the gamma isotope effects may be readily interpreted. Carbon-carbon hyperconjugation is precluded by the orientation of the methyl in ethyl propionate. As stated earler, steric effects resulting from a single interaction are expected to be only moderate. The remaining inductive difference between hydrogen and deuterium results in the observed normal effect. Although the inductive effect in ethyl isobutyrate-d5 should be substantial, the observation of an inverse isotope effect suggests that it is overcome by the steric effect. The presence of the second methyl group requires the carbonyl oxygen to approach one of the methylsin the more crowded transition state (Figure 10). This conformation also results OC2H5 Figure 10. Steric Interactions in Ethyl Isobutyrate Hydrolysis. 59 in interaction between one of the methyls and the attacking hydroxide. Carbon-carbon hyperconjugation may also be contributing to the large inverse isotope effect. V. Solvation Effects Although the anomalous isotope effects of ethyl isobutyrate have been explained in terms of a twofold barrier to rotation, the results are not necessarily inconsistent with a threefold barrier. It has been noted that the activation entropy is substantially larger for ethyl isobutyrate than for ethyl propionate or ethyl acetate. This difference is undoubtedly due to steric inhibition of solvation in the ground state of ethyl isobutyrate resulting from interactions with methyl. A consequence of this is an increased demand for solvatiop on the side away from the methyl group (2 ). Interference between the CH \30/ / CZHS < Solventl alcohol moiety and solvent will tend to move the alcohol away from its Preferred planer conformation into a position nearer the methyl. The resulting torsional effect on the spZ-sp3 carbon-carbon bond, combined with torsional effects resulting from interaction between the non-bonded Electron pairs of the alcohol oxygen and the methyl, increase the dihedral angle between the carbon-hydrogen (deuterium) bond and the n orbital with which it is to hyperconjugate. This angle, without consideration of 6O torsional effects, is 30°. A 30° angle still allows an overlap with greater than 85% of the effectiveness of parallel orbitals, but it is to the point at which orbital overlap begins to deteriorate rapidly as dictated by the cosine relationship 8 = Bo cos 0 (77). Torsional effects, if indeed they exist, could destroy the orbital overlap which would also eliminate hyperconjugation. Steric effects from this position are expected to be slight, and the remaining indicative effect would lead to the observed normal effect. A small contribution to the rate determining step by the process involving ethoxide expulsion would tend to counteract the isot0pe effects involved in hydroxide attack, the two steps being essentially the reverse of each other. Since a more crowded transition state would lead to increases in both activation barriers,as well as the instability of the tetrahedral intermediate, it is not clear what effect crowding should have on the isotope effects. VI. Conclusion In conclusion, the inverse B isotope effects observed with ethyl acetate and ethyl propionate have been explained by hyperconjugation. The y normal isotope effect in ethyl propionate is believed to be the result of inductive effects while the y inverse effect observed in ethyl isobutyrate is very likely steric in origin in this crowded system, with a possible enhancement by carbon-carbon hyperconjugation. The observation of a normal 3 isotope effect with ethyl isobutyrate is considered to be evidence for a change from a threefold barrier to rotation to a twofold barrier. Torsional effects on a threefold barrier have also been considered. I‘: II) V, n5, 61 The objective of this study was to confirm or deny the unprecedented temperature dependence of the isotope effect reported by Halevi in the alkaline hydrolysis of ethyl acetate-d3 and to extend the temperature study to include the isotope effects of ethyl prooionate and ethyl iso— butyrate. Although the objectives have been fulfilled through the implementation of two rapid and accurate analytical techniques, and explanations for the observed isotope effects have been offered, the implication of the B isotope effect of ethyl isobutyrate-dJ should be considered evidence for rather than proof of a twofold barrier to rotation in ethyl isobutyrate. EXPERIMENTAL I. Ester Preparation A. Ethyl Acetate—g3 and Unlabeled Ethyl Acetate Ethyl acetate-g3 (99%-D) was purchased from Merck Sharp & Dohme Lot No. C-l88. Mass spectral, nmr, and glc on carbowax analyses con- firmed its chemical and isotopic purity. ”Baker Analyzed” reagent ethyl acetate was distilled and used without further purification. B. Ethyl Propionate-2,2-g2; -3,3,3-d3; and Unlabeled Ethyl Propionate All ethyl propionate were prepared by the method of Nolin and Leitch (78). Five and one half gram samples (0.074 moles) of dis- tilled Fisher certified propionic acid, Merck Sharp 8 Dohme propionic-2,2-g2 acid-g_(98%-D) and propionic-3,3,3—d3 acid (98%-D) were cooled, diluted tenfold with watersand neutralized with a cold ten per cent aqueous ammonia to a pH between 7.5 and 8.5. A twenty molar per cent excess of silver nitrate solution was added and the resulting white silver salt filtered, washed first with water and then ethanol, and finally dried over night in a vacuum oven. The yield was generally 98%. To the dry silver salt was added an equal molar amount of distilled ethyl iodide and, with exclusion of water, the two were allowed to reflux over night yielding 86% of the desired ester. Gas chromatographic analysis of the crude distalate indicated an im- purity level of approximately 0.4%. Washing with water, extraction with 62 63 ether, removal of ether under reduced pressure, and finally isolation by preparative glc on an 80 inch, 20% carbowax on chromasorb W column afforded pure ethyl propionates. Mass spectral, g1c on carbowax and FFAP, and nmr analysis confirmed the chemical and isotopic purity of the esters. A11 ethyl propionates exhibited the characteristic nmr ethyl quartet and triplet at 6.26 and 9.13 T. The unlabeled ethyl propionate spectra contained a second methyl triplet at 9.28 T, overlapping the ethyl triplet, and a methylene quartet at 8.10 T. Labeled ethyl propionates exhibited the same spectra except that ethyl propionate-2,2-d2 lacked the methylene triplet,and its second methyl degenerated into a multiplet at 9.27 T. Ethyl propionate—3,3,3-g3 exhibited the methylene at 8.13 r as a multiplet and lacked the second methyl peak. C. Ethyl 2-Methy1propanoate; Labeled and Unlabeled Ethyl 2-methy1-g3 -propanoate-3,3,3-d3, ethyl 2-methy1propanoate- 2-g2,and unlabeled ethyl 2-methy1propanoate were prepared from their parent acids, purified, and analyzed by the same procedure as the ethyl propionates. Symyisobutyric-gfi acid (99.5%-D) was prepared by Adan Effio by the following process: OH Br LiAlHu | PBr3 | CD3C0CD3——————9'CDsCHCD3-———9 CD3CHCD3 Br MgBr Mg A 1) C02 CD CHCD CD HCD CD CHCOOH 3 3‘“* 3 3E7‘fi;fi* ( 312 64 Isobutyric-a-g_acid was prepared by pyrolysing dimethylmalonic acid-d2 at a temperature of 180° for 12 hours. The labeled dimethyl- malonic acid was the product of five exchanges of the unlabeled dimethyl- malonic acid (Aldrich Chemical Co.). with 020 (99%-D) In addition to the ethyl quartet and triplet at 6.22 and 9.06 T present in the nmr spectra of all ethyl isobutyrates, the unlabeled ester Spectrum contained a two methyl double at 9.15 T and a broad methine multiplet at 7.86 T. The ethyl isobutyrate-g5 spectrum lacked the two methyl doublet, and the multiplet had degenerated into a broad singlet at 7.81 T. The Spectrum of ethyl isobutyrate-a-9_lacked the multiplet, and the two methyl doublet reduced to a broad singlet at 9.15 T. II. Procedure A. Base Preparation Carbonate free base was prepared by the method of Albert and Serjeant (79). Fourteen grams of reagent grade potassium hydroxide was dissolved in 1.5 liters of distilled water. A solution containing three grams of barium hydroxide was added, the flask inverted, and the barium carbonate precipitate was allowed to settle overnight. Solution was drawn out of the flask through a tube protruding beyond the precipitate and into a Dowex SON-X8 cation exchange column saturated with potassium ion in order to remove the barium ion present. Dilution of 100 ml of this 0.16 N solution with three liters of freshly boiled conductance water under nitrogen atmosphere afforded carbonate free base of the approximately 0.005 normality desired. 65 Base concentration was determined by titration against seven samples of oven-dried Fisher Primary Standard potassium hydrogen phthalate weighed on a five decimal place Mettler balance using Bromo Thymol Blue indicator. B. pH Monitorimg An Instrumentation Laboratory model 245 scale expanding pH meter was used to continuously monitor the pH. The electrode used was an Instrumentation Laboratory model 14063 high alkaline combination electrode. When not in use, the electrode was stored in dilute hydrochloric acid at the next temperature to be investigated. The output of the pH meter was recorded on a Sargent model SL recorder. All determinations were made in the scale expanded mode. One pH unit represented a 10-inch deflection on both the meter and recorder. No decernable variation between the meter and recorder appeared over this range. The chart speed of the recorder proved to be sufficiently accurate to provide the time base for the readings. C. Constant Temperature Bath Temperature was maintained during the hydrolyses by a 25-liter constant temperature bath. The bath was insulated on its sides and bottom with one inch of styrofoam supported by 5/8 inch plywood. The top of the bath was covered with a tight-fitting 1/4 inch rein- forced plexiglas cover. From the top were suspended a Troemner immersible magnetic stirrer, a Precision Scientific Co. number 62541 thermoregulator (10.005F°), and a Teel number lp622 circulating pump. A sample cell 2.50 inches long blown from 19mm glass tubing was inserted through a tight-fitting hole 66 and came to rest centered on the stirrer with its top flush with the top of the cover. Two additional holes were provided in the top through which additional sample cells could be suspended to allow complete thermal pre-equalibration. 0. Temperature Control The bath was cooled or heated as required. Circulation of tap water from a constant pressure device through three feet of internal 1/4 inch copper tubing provided adequate cooling for temperatures down to 15°. A temperature of" 5° was obtained by pumping ethylene glycol at 0° from a second bath through the cooling coil. In both cases, and at higher temperatures, the temperature was controlled by intermittent heating with a ZOO-watt heating coil, activated by the thermal controller's switching of a relay switch. The amount of heating was limited by a triac clipping circuit until the ”off time" and ”on time" were equal. Both the relay and clipping unit were built by the author. For temperatures above 65°, an additional continuous heater capable of supplying up to 100 watts was added. E. ggartz Thermometer Bath temperatures were measured with a Hewlett-Packard model 2801A Digital Quartz thermometer. Zero point calibration of probes was per- formed in a three liter ice water Dewar flask mounted on a low speed mechanical shaker. Bath temperature variation was found to be less than i0.007° over the range +15 to +65°. 67 F. Rate Determinations At the beginning of each set of measurements, the electrode was rinsed with distilled water, placed in Instrumentation Laboratory number 31060 pH 6.84 buffer (the isoelectric point of the electrode) and the meter balance adjusted. After two more washings, electorde was placed in a Matheson, Coleman & Bell pH 10.00 buffer and the appropriate correction made with the pH/MV slope control. Following two additional washings, 3.88 ml of the standardized potassium hydroxide was dispensed into a sample cell from a 5 m1 burette and transferred under nitrogen atmosohere to the bath. All washings, buffers, and base samples were thermally pre- equalibrated in order to avoid electrode hysteresis upon cooling (79). Six minutes was allowed to elapse while the electrode adjusted to the base. At that time the balance of the pH meter was adjusted to a scale expanded value of 10.95. Fifteen micro liters of ester was injected through septum into the stirred base as the recorder passed a convenient measuring point. Recording was continued until a pH of 10.00 was reached, at which time the electrode was washed, new base added, and the process repeated. The first two runs were always discarded to avoid inconsistencies resulting from incomplete electrode conditioning. Subsequent runs were alternated between unlabeled and the one or two labeled esters. G. Treatment of Data Times read from the recording were taken at 0.05 pH unit intervals from pH 10.90 to 10.05. These values, corresponding to 2.9 half-lines of base, were analyzed by the second order rate program SEC R (Figure 11). 19% 2A 111 +5 at) 146 L)” 68 SFCP FIIHl‘J/Ul 1‘,)