Progress in physical organic chemistry, volume 4

Reaction Pathways The first step in acetal, ketal, or ortho ester hydrolysis in which the making or breaking of covalent bonds to carbon is involved may be visualized as occurring via o

ORGANIC CHEMISTRY

ANDREW STREITWIESER, JR., Department of Chemistry

University of California, Berkeley, California

ROBERT W TAFT, Department of Chemistry

University of California, Irvine, California

1967

INTERSCIENCE PUBLISHERS

a diuision, of *John Wiley 4 Sons New York London Sydney

Copyright @ 1967 hy John Wiley and Sons, Inc

All rights reserved

Library of Congress Catalog Card Number 63-19364

PRINTED IN THE UNITED STATES OF AMERICA

Introduction to the Series

Physical organic chemistry is a relatively modern field with deep roots in chemistry The subject is concerned with investigations of organic chemistry by quantitative and mathematical methods The wedding of physical and organic chemistry has provided a remarkable source of inspiration for both of these classical areas of chemical

cndeavor Further, the potential for new developments resulting

from this union appears to be still greater A closening of ties with all aspects of molecular structure and spectroscopy is clearly antici- pated The field provides the proving ground for the development of

basic tools for investigations in the areas of molecular biology and biophysics The subject has an inherent association with phenomena

in the condensed phase and thereby with the theories of this state of matter

The chief directions of the field are: (a) the effects of structure and environment on reaction rates and equilibria; (b) mechanism of re-

actions; and (c) applications of statistical and quantum mechanics

to organic compounds arid reactions Taken broadly, of course, much

of chemistry lies within these confines The dominant theme that characterizes this field is the emphasis on interpretation and under- standing which permits the effective practice of organic chemistry The field gains its momentum from the application of basic theories

arid methods of physical chemistry to the broad areas of knowledge

of organic reactions and organic structural theory The nearly in- exhaustible diversity of organic structures permits detailed and syste- matic investigations which have no peer The reactions of complex natural products have contributed to the development of theories of physical organic chemistry, and, in turn, these theories have ulti- mately provided great aid in the elucidation of structures of natural products

Fundamental advances are offered by the knowledge of energy

states and their electronic distributions in organic compounds and the relationship of these to reaction mechanisms The development, for example, of even an empirical and approximate general scheme

V

vi INTRODUCTION TO THE SERIES

for the estimation of activation energies would indeed be most notable The complexity of even the simplest organic compounds in terms of physical theory well endows the field of physical organic chemistry with the frustrations of approximations The quantitative correla- tions employed in this field vary from purely empirical operational formulations to the approach of applying physical principles to a workable model The most common procedures have involved the application of approximate theories to approximate models Critical assessment of the scope and limitations of these approximate applica- tions of theory leads to further development and understanding Although he may wish to be a disclaimer, the physical organic chem- ist attempts to compensate his lack of physical rigor by the vigor of his efforts There has indeed been recently a great outpouring of work in this field We believe that a forum for exchange of views and for critical and authoritative reviews of topics is an esseritial need of this field It is our hope that the projected periodical series of volumes under this title will help serve this need The general organization and character of the scholarly presentations of our series will corre- spond to that of the several prototypes, e.g., Advances in Enzymology, Advances in Chemical Physics, and Progress in Inorganic Chemistry

We have encouraged the authors to review topics in a style that is not only somewhat more speculative in character but which is also more detailed than presentations normally found in textbooks Ap- propriate to this quantitative aspect of organic chemistry, authors have also been encouraged in the citation of numerical data It is intended that these volumes will find wide use among graduate stu- deuts as well as practicing organic chemists who are not necessarily expert in thc field of these special topics Aside from these rather obvious considerations, the emphasis in each chapter is the personal ideas of the author We wish to express our gratitude to the authors for the excellence of their individual presentations

We greatly welcome comments and suggestions on m y aspect of these volumes

ANDREW STREITWIESER, JR

ROBERT W TAFT

A Streitwieser and R W Taft regret very much that Saul G

Cohen has considered it necessary to withdraw as Co-editor on this and subsequent volumes We are greatly indebted for his contribu-

tions to Volumes 1-3

Indiana University, Rloomington, Indiana

Metcalj Chemical Laboratories, Brown University, Providence, Rhode Island

Cornell Universitg, Ithaca, New York

National Research Council Laboratories, Ottawa, Canada

Ionic Reactions in Acetonitrile

Nucleophilic Displacements on Peroxide Oxygen and Related

Conformation and Structure as Studied by Electron Spin

Mechanism and Catalysis for the Hydrolysis of Acetals, Ketals, and Ortho Esters

BY E H CORDES

Indiana University, Bloomington, Indiana

C O N T E N T S

I Introduction

11 Reaction Pathways

A The Site of Ca Cleavage

B The Question of Solvent Participation as Nucleophilic Reagent C The Intermediates

D Some Dissenting Suggestions E Summation

111 The Rate-Determining Step A Some General Conside

B Kinetic Studies Employing Proton Magnetic Resonance Spectros- copy

C Further Considerations Concerning Structure-Reactivity Correla- tions

IV Catalysis

A General Acid Catalysis

B Catalysis by Detergents

V Some Related Reactions

References

I Introduction

1

2

3

5

21

23

24

24

25

25

29

32

32

36

38

41

Studies concerned with mechanisms and catalysis for the hydrolysis

of acetals, ketals, and ortho esters have been seminal in the develop-

ment of a general understanding of these topics for reactions in aque-

ous solution Indeed, pioneering studies on general acid-base ca- talysis, solvent deuterium isotope effects, reaction kinetics in strongly acidic media, and structure-reactivity correlations have employed these substances as substrates Such early studies, together with significant recent developments, have clearly established the principal mechanistic and catalytic features of these hydrolytic processes These are summarized in this review

In acidic aqueous solutions, acetals, ketals, and ortho esters hydro- lyze according to the overall stoichiometry indicated in eq (1)

1

Edited by Andrew Streitwieser, Jr Robert W Taft

Copyright 0 1967 by John Wiley & Sons, Inc

2 E H CORDES

These renctions occur with the rupture of two covalent bonds to car- bon and involve at least two proton transfer reactions Hence the overall reaction must be multistep Our first concern in this review

is the nature of the intermediates formed in such multistep processes; i.e., the reaction pathway Subsequently, attention is directed to identification of the rate-determining step, to catalytic, mechanisms, and, in general, to a precise definition of transition state structures

II Reaction Pathways

The first step in acetal, ketal, or ortho ester hydrolysis in which the making or breaking of covalent bonds to carbon is involved may be visualized as occurring via one of the four transition states shown in structures 1-4 Each of these transition states is pictured, for the sake of clarity, as having arisen from the conjugate acid of the sub-

strate Kinetic studies indicate the presence of a proton or the kinetic equivalent in the transition state but leave uncertain the question of timing of proton transfer relative to cleavage of the C-0

picture these hydrolyses We return to this point below as ocrnrring via unimolecular decomposition

of the conjugate acids of the substrat,es with cleavage of the carhonyl

carbon-oxygen and alcohol carbon-oxygen bonds, respectively (A-

1 reactions) The rorresponding carhonium ions are the immediate

ACETALS, KETALY, AND OllTHO ESTERS 3

products Transition states 3 and 4 include the participation of

water as nucleophilic reagent with carbon-oxygen bond cleavage at the sites indicated (A-2 reactions) The immediate products are identical to those formed from addition of one molecule of water to the carbonium ions generated from transition states 1 and 2 Dis-

tinction between these transition states involves (a) localization of

the site of C-0 bond cleavage and (b) identification of the immediate product of C-0 cleavage as a carbonium ion or its hydrate We consider these topics in sequence

A THE SITE OF CARBON-OXYGEN BOND CLEAVAGE

Several lines of evidence conclusively establish that, for most cases

at least, the hydrolysis of acetals proceeds with cleavage of the car- bony1 carbon-oxygen bond The earliest convincing evidence for this point of view is the important work of Lucas and his associates

on the hydrolysis of acetals derived from optically active alcohols For example, hydrolysis of the D( +)-Zoctanol acetal of acetaldehyde

in dilute aqueous phosphoric acid yields 2-octanol having the same optical rotation as the original alcohol from which the acetal was syn- thesized (1) This finding excludes formation of the alkyl carbonium ion (transition state 2), in which case substantial or complete race-

niization of the alcohol would be expected, and an A-2 reaction in- volving nucleophilic attack of solvent on the alcohol (transition state

4), in which case optical inversion of the alcohol would be expected

Similarly, the formal, acetal, and carbonate derived from D( -)-2,3- butanediol and the acetal derived from ~(+)-2-butanol undergo acid-catalyzed hydrolysis with complete retention of configuration at the carbinol carbon of the alcohol (2,3)

Drumheller and Andrews have investigated the possibility that certain acetals, prepared from alcohols capable of forming relatively stable carbonium ions, might hydrolyze by the alkyl carbonium ion pathway (transition state 2) (4) The parent alcohols chosen for study were (-) a-phenethyl alcohol, methylvinyl carbinol, and phenylviny1 carbinol Derivatives of each of these alcohols are well known to readily undergo SN1-type displacement reactions Hy- drolysis of the acetal prepared from (-)a-phenethyl alcohol (in dilute sulfuric acid solution) produced alcohol with optical properties iden- tical to those of the original alcohol, as in the cases described above Similarly, hydrolysis of the methylvinylcarbinyl acetal yielded only

4 E H COBDES

methylviuylcarbinol, and hydrolysis of the phenylvinylcarbinol ac- eta1 yielded phenylvinylcarbinol as the immediate reaction product Thus, the latter two hydrolyses proceed without the allylic rearrange- ments (yielding crotyl alcohol and cinnamyl alcohol) characteristic

of the corresponding carbonium ions (5,6) Finally, the possibility that hydrolysis of these substrates occurred via transition state 4

(nucleophilic attack of solvent at the carbinol carbon atom) was ex- plicitly excluded through the observat,ion that methanolysis of a phen- ethyl alcohol-derived acetal yielded phenethyl alcohol arid not the corresponding methyl ether Thus, it is safe to conclude that even

in these cases, deliberately chosen to accentuate the possibility of al- cohol carbon-oxygen bond cleavage, acetal hydrolysis occurs with carbonyl carbon-oxygen bond cleavage

Bourns et al have strongly corroborated the above conclusiou in an isotope tracer study of acetal formation and hydrolysis (7) The condensation of benzaldehyde arid n-butyraldehyde, enriched in

180, with n-butyl and ally1 alcohols yielded acetsls of normal isotopic abundance and 180-enriched water [eq (2)] In a like fashion, hy-

drolysis of benzaldehyde di-n-butyl acetal and n-butyraldehyde di-n-

butyl acetal in lsO-enriched water yielded alcohols of normal iso-

topic content [the reverse of eq (2) 1 Thus, these rcactions clearly proceed with carboiiyl carbon-oxygen bond cleavage (or formation) Less experimental work on the site of carbon-oxygen bond cleavage has been reported for the cases of ketal and ortho ester hydrolysis One would expect that these substrates behave in a fashion similar to that of acetals Some very early work on hydrolysis of ketals tends

to bear out this supposition The acetone ketals of the cis 1,Bdiols

of tetrahydronaphthalene, hydrindene, and 1-phenyl cyclohexane yield the cis diols almost exclusively on hydrolysis; a result consistent

only with carbonyl carbon-oxygen bond cleavage (8-10) Recently, Taft has studied the hydrolysis of methyl orthocarbonate in HPO

While most of the lSO does appear in the carbonyl group of dimethyl

carbonate as expected, there is appreciable formation of CHPOH

and CH3-O-CH3, i.e., methylation of the nucleophiles water and methanol either by the orthocarbonate or the corresponding car-

ACETATAS, KWTALR, ANT) ORTHO ESTERS 5

boniuni ion (11) These products almost certainly arise in bimo- lecular reactions involving alcohol carbon-oxygen bond cleavage

(transition state 4 or a variant thereof)

The early suggestion of Hammett, based on the relative rates of

hydrolysis of several formals (12), that acetal hydrolysis occurs via

formation of the alcohol-derived carbonium ion must be abandoned

in light of the above considerations (13) The data available to

Hammett is also consistent with formation of the carbonium ion ac- cording to transition state 1 A sizable amount of subsequent work

on structurereactivity corre1at)ions for acetal and ketal hydrolysis

provides strong support for the latter alternative (14)

In summary, the data indicated above and reasonable extrapola- tions thereof strongly suggest that, in the preponderant majority of cases, acid-catalyzed hydrolysis of acetals, ketals, and ortho esters occurs with cIeavage of the carbonyl carbon-oxygen bond We now turn to a consideration of the distinction between the two transition states, 1 and 3, which involve bond cleavage of this type

Several independent lines of evidence strongly suggest that the acid-catalyzed hydrolysis of acetals, ketals, and ortho esters proceeds

by a reaction pathway not involving solvent as nucleophilic reagent, i.e., that 1 describes the transition state for the initial reaction in which covalent bonds to carbon are broken [eq (3)] These lines of

cleophilic reagents, and (8) solvent eff ccts We consider the results

of these studies sequentially

G E H CORDES

1 Reaction Kinetics

Thc hydrolysis of the substrates in question is almost invariably

Thal is, the rate laws for reactions dependent upon acid catalysis

in dilute aqueous solution have the form

h t , S = k ~ t ( H + ) + Z ~ H A % ( H A ) ~ (4)

in which the terms in the summation on the righthand side of the

equation are frequently negligible (see p 32) The completc de-

pendence of these reactions on acid catalysis suggests that water does not participate as a nucleophilic reagent If water were able tx, expel alcohol from the protonated substrates in nucleophilic reactions, then one might expect that hydroxide ion (or other nucleophilic rcagent) would expel alkoxide ion from the corresponding free hases Since the latter reactions are not observed, one suspects that the former rc- actions do not occur either This is, of course, a naive argument and provides only weak evidence against nucleophilic participation by water

The hydrolyses of 2-phenyl-1,3-dioxanes possessing 0- or p-phenolic substitumt,s do exhibit pH-independent, as well as acid-catalyzed, reactions (44) However, solvent deuterium isotope effects suggest that the pH-independent reaction is, in fact, the hydronium ion- catalyzed hydrolysis of the phenolic form of the substrates This is,

of course, a kinetically indistinguishable alternative to a formulation involving an uncatalyzed (or solvent-catalyzed) reaction pathway

In some instances of acetal hydrolysis, nucleophilic reaction paths

do seem to be important I n each of these cases, the nucleophilic re- action is intramolecular, not intermolecular The cleavage of certain glycosides, such as phenyl-P-D-glucoside, is subject to catalysis by hydroxide ion (116) These reactions are not properly regarded as hydrolyses, however, since the oxygen at C-2 participates in the ex- pulsion of phenoxide ion with formation of the 1,6-anhydro sugar Similar comments apply to the recent conclusion of Capon and Thaclier that acid-catalyzed ring closure (not hydrolysis) of dimethyl scetals of glucose and galactose involves nucleophilic attack syn-

Th' 1s con- rlusion is based on the observations (1) that the configuration at

carbon 4 of the sugar moiety influences the rate of furanoside forma-

tion and (2) that the rate of these ring closures is 30 to 340 times more

rapid than that predicted from data on related substrates

chronous with rupture of the acetal bond [eq (.5)] (1.5)

ACETATAS, KETALS, ANT) ORTHO ESTERS 7

the reaction pathway for methylthioacetaldehyde diethyl acetal, then formation of the cyclic sulfonium ion must be rate-determining since both molecules of ethanol liberated in the reaction appear simultane- ously

ACETALS, KETALM, AND ORTHO ESTERS 9

2 Xtructure-Reactivity Correlations

At this point, attention is directed to those structure-reactivity correlations which exist within individual reaction series (i.e., relative hydrolysis rates for methyl acetals or alkyl aldehydes) Interseries comparisons are deferred for the moment (cf p 29)

In several instances, second-order rate constants for reactions of interest here are correlated by one or more linear free-energy relation-

u t0.5 (u+-u)

Fig 1 Plots of the logarithms of second-order rate constants for hydrolysis of (0) substituted benaaldehyde diethyl acetals and ( 0 ) 2-(substituted pheriyl)-1,3- dioxolanes against u + 0.5( u + - u) The values on the left ordinate refer to the benzaldehyde acetals and those 011 the right to the dioxolanes Constructed from data of Fife and Jao (16)

10 E H COltDES

ships These cases are collected in Table I Second-order rate con- stants for acetal and ketal hydrolysis are very sensitive to structural alterations in both the aldehyde and alcohol moieties Such rate eonstants for hydrolysis of a series of m-substituted diethyl acetals of benzaldehyde are correlated by the Hammett u constants and a p

value of -3.35 (lG) This p value is consistent with and support for rate-determining carbonium ion formation since, in this case, electron doriatioil from a polar substituent will both favor preequilibrium sub- strate protonation arid stabilize the carbonium ion developing in the transition state For compounds substituted in the para position with groups capable of clectrori donation by resonance, second-order ratc constants fall somewhat above the line established by the m-sub- stituted compounds when plotted against the u constants and some- what below a corresponding line when plotted against the u* con- stants Data of this type may be treated according to the considera- tions of Yukawa ant1 Tsuno, who have suggested a linear free-energy correlation of the form (21)

The sccond-order rate constants for hydrolysis of p- and m-substituted hcnzaldchyde diethyl acetals are well correlated by eq (7) and values

of p and r of -3.35 and 0.5, respectively, as illustrated in Figure 1

A very similar situation exists for hydrolysis of 2-(p-substituted

phenyl-)-l,3-dioxolanes (Table I and Fig 1) (16) The fact that these reaction rates are correlated by a set of substituent constants intermediate between u and u+ is fully consistent with rate-determin-

ing carbonium ion formation

I<rccvoy and Taft have found that second-order rate constants for the hydrolysis of 24 diethyl acetstls and ketals of nonconjugated alde- hydes and ketones are well correlated, with one exception, by the linear free-energy relationship

Iog(k/Lo) = ( 2 ~ * ) p * + (An)h

in which u* is the sum of the appropriate polar substituent constants (19), An is the difference in the total number of a-hydrogen atoms in the carbonyl moiety and the six in the standard of comparison, diethyl aoetonal, and h is an empirical constant measuring the facilitating effect of a single hydrogen on the rate (18) This structure-reactivity correlation is illustrated ill Figure 2 for the case h = 0.54 Both the

ACETALS, KETALS, AND ORTHO ESTERS 11

magnitude of the value of p*, -3.60, and the necessity of including a hyperconjugation term suggest rate-determining carbonium ion formation, not rate-determining solvent attack The abnormal re-

activity of the methyl neopentyl ketal (Fig 2), the exception noted above, may be accounted for in terms of relief of steric strain as the tetrahedral carbon atom approaches the trigonal configuration in the transition state The latter point has been further pursued by Kree- voy et al in a study of hydrolysis rates for bulky and cyclic ketals

(22) In all cases, the observed rate constants are consistent with the hypothesis that the transition state has made considerable prog-

12 E H CORDES

ress toward carbonium ion geometry This conclusion is in full accord with those resulting from related studies on the hydrolysis of cyclic acetals and 1ret:tls (30,118,119)

The marked effects of suhstituents on rates for acetal and ketal hydrolysis are not reflected in the overall equilibrium coilstants for their formation Hartung and Adkins observed only modest effects on

the equilibrium constant for a series of saturated I1 groups of similar

steric requirements (23)

The early studies of Skrabal and Eger on the acidic hydrolysis of

symmetrical formals have revealed a marked sensitivity of the rates to

changes in substituerits (Table I, entries Ti and 6) (12,19,20) This

TABLE I1

Comparison of Substituent, Effects on the Relative Rates of Hydrolysis of

Acetals and Ketals to Those on the Relative Rates of Hydrolysis of Ortho

Esters (modified from ref 25)

Relative hydrolysis

sensitivity is presumably the primary consequence of polar effects 011

the preequilihrium protonation of the substrates These findings are fully corroborated by a recent and extensive study of formal hydroly-

ACETALS, KETALS, AND OBTHO ESTERS 13 sis by Salomaa (24) In addition, this worker has developed methods for sorting out the relative contributions of the two C-0 fission re- actions which occur in the hydrolysis of unsymmetrical acetals arid ketals

Substituent effects on rates of ortho ester hydrolysis are much smaller than the corresponding effects on acetal and ketal hydrolysis

(25) Furthermore, the rate constants for ortho ester hydrolysis do

not increase uniformly with increasing electron-donating power of the substituent A quantitative comparison of substituent effects in

the two reaction series is presented in Table 11 The detailed inter- pretation of these substituent effects is deferred to a later section,

Suffice it to say a t this point that these effects are consistent with the intermediacy of carbonium ions in ortho ester hydrolysis The systematic study of substituent effects in benzaldehyde ortho ester

hydrolysis (17) (Table I, entry 7) is badly clouded by an unfortunate

choice of solvent (see discussion below), and the value of p obtained cannot be firmly relied upon

3 Entropies of Activation

The use of entropies of activation as a criterion of mechanism for

acid-catalyzed reactions in aqueous solution has been reviewed by Schaleger and Long (38) Briefly stated, experience indicates that reactions proceeding with unimolecular decomposition of the pro- tonated substrate (A-1) usually exhibit entropies of activation near zero or somewhat positive while, in contrast, those proceeding with

nucleophilic attack of solvent on the protonated substrate (A-2)

usually exhibit corresponding values which are large and negative That bimolecular reactions should exhibit more negative entropies of activation than unimolecular reactions is reasonable in view of the loss of rotational and translational freedom of the water molecule in

the transition state However, variability in the A S accompanying

the protonation reaction may cloud the picture, and differences in entropies of activation are not always large enough to permit un- ambiguous conclusions A compilation of data relevant to the sub- strates under consideration is presented in Table 111 This data, in

light of the above generalization, lends additional support to the concept of carbonium ion intermediates for acetal, ketal, and ortho ester hydrolysis

15 Methoxymethyl acnetate hydrolysis

16 Methoxymethyl formnte hydrolysis

niolecular are in the range of -2 to +6 cm.3/mole while those for reactions considered to be bimolecular (with nucleophilic participation

of solvent) are in the range -6 to -10 ~rn.~/mole (39) This result

is intuitively reasonable since, in the unimolecular case, some loosen- ing of a covalent bond will have occurred in the transition state with

an attendant overall increase in volume of the reacting species, while

in the bimolecular case, the partial formation of a covalent bond be- tween the substrate and water in the transition state may result in an overall decrease in volume of the reacting species Volumes of

activation are probably a more reliable guide to mechanism than the corresponding entropies in that the volumes changes accompanying

ACETALS, KETALS, ANT, ORTHO ESTERS 15

TABLE I V Volumes of Activation for Certain Acid-Catalyzed Hydrolytic Reactions

the preequilibrium protonation seem less susceptible to variation than

do the entropy changes In Table IV, volumes of activation for several reactions of interest are recorded I n each case, the value falls into the range typical of reactions involving unimolecular de- composition of the protonated species

5 Isotope Ffects

Solvent deuterium isotope effects on the rate of hydrolysis of certain acetals and ortho esters are collected in Table V Most of these values fall in the range JcD,O+/kHsO+ = 2 to 3 Such solvent deuterium

isotope effects probably primarily reflect the isotope effect on the pre-

equilibrium protonation reaction (47) Rate increases of two- to

threefold in DzO compared to HzO are typical of acid-catalyzed re- actions considered to be unimolecular (A-1) and are similar to those

predicted theoretically For example, Bunton and Shiner have calculated a deuterium solvent isotope effect for acetal hydrolysis of

2.5 (48) Of particular note are the isotope effects on the hydrolysis

of ethyl orthocarbonate With the hydrated proton as catalyst, the isotope effect is small, and with acetic acid as catalyst, the isotope ef- fect is actually less than unity These results suggest the involve- ment of proton transfer in the transition state (cf discussion, p 33)

Kilpatrick has carried out a very careful study of the effect of temperature on the solvent deuterium isotope effect for 1,l-di- methoxyethane and 2-methyl-1 ,&dioxolane hydrolysis in the tem- perature range 0 to 40" (45) I n both cases, the isotope effect was

observed to decrease with increasing temperature The temperature dependence of the isotope effect is given by kDBO+/kHsO+ = 1.166

TABLE V Kinetic Solvent I)erit,eriiim Isotope lSffec+s for Acid-Cntnlyxed Acetnl,

Ketal, and Ortho Eskr FTydrolysis*

5Oy0 dioxane Wat er Water

10% aczeto-

nitrile W:ti er Water Wat>er Water Water Water Water Water

25

25 2.5

exp (521/RT) in the former case and by kD80+/kH30+ = 1.191 exp

(AOlIRT) in the latter

Shiner and Cross have measured the secondary deuterium isotope effect resulting from a-deuteration in the carbonyl component on the rates of hydroIysis of the diethyl ketals of acetone, methyl ethyl

ketone, methyl isopropyl ketone, and phenoxyacetone (49) For the

fully a-deuterated substrates, a value of krr/ko of 1.1 to 1.25 was ob-

tained in each case These results may be attributed to either the greater relative inductive electron donating power of H compared to

D or to the greater relative hyperconjugative electron-donating power

of H compared to D or to both (50) Regardless of the precise ex-

planation, these results substantiate the earlier conclusion that elec- tron donation accelerates ketal hydrolysis, as expected in terms of mte-determining carhonium ion formation

fi Corrdnlions of Rates with Acidit?] Functions

The hydrolysis of dimethoxymethane (53), diethoxymethane (as),

and 1,1-dimrthoxy-2-chloroethane (51) in moderately concentrated

solutions of mineral acids is characterized by rate constants which are

ACETALS, KETALS, AND OHTHO ESTERS 17

correlated by the Hammett acidity function, ho, and not by the molar concentration of acid This finding, according to the Zucker-Ham- mett hypothesis, suggests a unimolecular reaction path (54,55)

So many exceptions to this hypothesis have been recognized the cor- relations of rate constants with ha cannot be relied upon as an indica- tion of mechanism (56,57) Nevertheless, the above findings are consistent with and limited support for a unimolecular reaction path for the hydrolysis of acetals at least Kreevoy has extended these observations to 50% aqueous dioxane solutions, a medium in which

ha is not strictly defined, for the hydrolysis of four ketals and acetals

(58) With varying coricentrations of perchloric acid in this solvent,

the rates of hydrolysis are correlated with the proton-donating power

of the solvent as measured by the extent of protonation of 2-1iitro-4- chloroaniline

A related criterion of mechanism developed by Bunnett is COII-

siderably more rigorous and is based on the correlation (w values) of rate constants with the activity of water for reactions run in moder- ately or strongly acidic media (59) Correlation of rate constants with water activity for acetal hydrolysis yields values of w which are characteristic of reactions thought to occur by unimolecular reaction paths (59)

7 Experiments with Added Nucleophilic Reagents

Each of the criteria indicated above provides support for the thesis that the initial step involving carbon-oxygen bond cleavage for the hydrolysis of acetals, ketals, and ortho esters involves unimolecular decomposition of the protonated substrates rather than a bimolecular reaction involving solvent as the nucleophilic reagent Taken to- gether, these criteria constitute a strong case for this conclusion Considerable further support is provided by studies of the hydrolysis

of methyl orthobenzoate and ethyl orthocarbonate in the presence of added nucleophilic reagents

The first-order rate constants for the decomposition of methyl orthobenzoate in slightly acidic aqueous solution are independent of the concentration of added hydroxylamine and semicarbazide under conditions in which an appreciable fraction of the ortho ester yields

amine addition products rather than methyl benzoate (32) This

result is illustrated in Figure 3 For example, in the presence of

0.9M hydroxylamine at pH 5.45, approximately 85% of the methyl

positionof methylorthobenzoate a t 25" arid ionic strength 0.50 plotted against the concentrution of (circles) liydroxylainiiie aiid (triangles) semicarbazide In addi-

tion, the fraction of methyl orthobenzoate yielding ester product (open figures) is plotted against the concentration of these amines

orthobenzoate yields a hydroxylamine addition product, probably N-hydroxymethyl benzimidate, yet the first-order rate constant (0.00038 sec.-l) is not appreciably different from that measured in the absence of hydroxylamine (0.00035 sec.-l> Methyl benzoate does

riot react with hydroxylamine under the conditioris of these reactions

at an appreciable rate Furthermore, the fraction of methyl ortho- benzoate yielding methyl benzoate as the product may be accurately calculated, assuming that the conjugate acid of the ortho ester under- goes a uniinolecular decomposition yielding an intermediate car- boniuni ion which is then rapidly partitioned between water, yielding methyl benzoate, and amine, yielding amine addition product On this basis, the free base of hydroxylamine is calculated to be 2000-fold and the free base form of semicarbazide 275-fold more reactive toward the carbonium ion than water These results strongIy suggest that solvent does not participate as a riucleophilic reagent in the rate- determining step of acid-catalyzed methyl orthobenzoate hydrolysis The above experiments arc closely related to the rate-product criterion originally employed by Ingold and his co-workers for the identifica- tion of uniniolecular solvolysis reactions (60)

ACETALS, KETALS, AND ORTHO ESTERS 19

Related experiments have been performed by Kresge and Preto for the case of ethyl orthocarbonate hydrolysis, a reaction subject to

modest general acid catalysis (61) Thcsc workers observed that thc

rate of hydrolysis of this substmtc is thc samc in thc prcsence of

sodium iodide as in the prcsence of sodium perchlorate Employing

the Swain-Scott equation, a reasonable calculation reveals that, were the hydrolysis in fact bimolecular, the concentration of iodide ion employed should have more than doubled the rate through a direct nucleophilic displacement reaction Since no rate increase was ob- served, one concludes that the hydrolysis reaction is, in fact, not bi- molecular Thus, this study fully corroborates that performed em- ploying methyl orthobenzoate as substrate

8 Solvent Effects

Solvent effects on the rate of hydrolysis of substances of interest here have been rather little studied and do not shed much additional light on the reaction mechanisms

Wolford has observed that the second-order rate constants for acid- catalyzed hydrolysis of 1,l-diethoxyethane pass through a minimum

in an approximately equimolar mixture of water and acetone (62)

This behavior mimics that for the hydrolysis of certain esters of car-

boxylic acids (63) despite the fact that the latter reactions almost cer-

tainly involve the nucleophilic participation of solvent In a related study, the first-order rate constants for methyl orthobenzoate hydroly- sis in 0.001M HCI have been determined as a function of solvent composition for several organic solvent-water systems and the re- sults are indicated in Figure 4 (64) Kwart and Price have obtained data similar to that for methanol-water mixtures illustrated in this figure (17) It is interesting to note that methanol and ethanol, molecules which are capable of reaction with an intermediate car- bonium ion to regenerate ortho ester, are considerably more effective

at inhibiting ortho ester hydrolysis than is acetone, which has approx- imately the same dielectric constant as these alcohols but is relatively nonnucleophilic This result suggests that at high concentrations of these alcohols, the intermediate carbonium ion will be trapped by the alcohols, regenerating starting materials, more frequently than by water, thereby yielding the hydrolysis products Surprisingly, in view of these data, DeWolfe and Jensen have reported modest in- creases in rate constants for the hydrolysis of ethyl orthoacetate with

We discuss them briefly

20 B 11 CORDES

20 4 0 60 80 100

.01

PERCENT ORGANIC SOLVENT

Fig 4 Logarithms of first-order rate constants for methyl orthobenxoate hy- drolysis plotted against tlhe percentage concentration of organic solvent for several water-organic solvent mixtures Reactions were carried out at 2.5' in the pres- ence of 9.6 X 10-4M hydrochloric acid

increasing concentrations of dioxane in dioxane-water mixtures Wolford and Ba1,es have studied the hydrolysis of 1,1-diethoxy- ethane in N-methylpropionamide-water and N,N-dimethylformam-

idc-water solvcnts (65) The quantitative aspects of this investigation are clouded by the fact that rates were measured at a variety of acid concentrations for the different solvent mixtures and compared by

calculating second-order rate constants from the expression: kz =

basic character seems not to have been established and, indeed, the first-order rate constants for the hydrolysis of 2,&dimethoxypropane

in 30% DMF-700/, water are not linearly related to the stoichiometric concentration of added HC1 (66) Nevertheless, the conclusion of these authors that the abrupt drop in rate with increasing concentra- tion of the amides [a drop which has been fully corroborated in experi- ments in which the total acid concentration was maintained constant

(27)

ACETALS, KETALS, AND OltTHO ESTERS 21

finding that the rate constants for hydrolysis of 2,2-dimethoxypropane

in DnIlLwatcr mixtures coritainirig 0.01M HC1 parallel quite closely the protoriatirig power of these solvents as measured spectrophoto- metrically employing p-chloroaniline as indicator (66) Thus, the observed rate effects reflect priricipally the fact that less acetal is con- verted to the conjugate acid at equilibrium as the concentration of the more basic component of the solvent system is increased

The tnininzum reaction scheme for acetal, ketal, or ortho ester hy- drolysis involving uriiniolecular decomposition of the protoriated substrates is indicated in eq (3) Thus, at least two intermediates,

an alkoxy carbonium ion and a tetrahedral species, are formed during thc overall hydrolytic processes The tetrahedral intermediate may

be either a hemiacetal, hemiketal, dialkyl ortho ester, or trialkyl orthocarbonate, depending on the nature of the substrate

By abstraction of hydride ion from a variety of acetals and ortho esters, Meerwcin and his collaborators have prepared several salts of

dialkoxy- and trialkoxycarbonium ions (67,68) In addition, such carbonium ions may be prepared directly from similar substrates in acidic solutions For example, Taft and Ramsey have provided

convincing evidence for the generation of trialkoxycarbonium ions

from orthocarbonates in sulfuric acid solution [eq (lo)] (69)

(IiO)4c + 2HzSO4 + (11O)&+ + ROSOaH- + HoO+ + 2HS04- (10)

The validity of eq (10) is based 011 the following observations:

( I ) Ethyl orthocarbonate in sulfuric acid gives an i factor (70) of

5.1 f 0.1

(2) Proton magnetic resonance spectra of solutions of orthocarbon- ates have integrated absorption intensities identical to those predicted

on the basis of eq (10)

(3) The infrared spectra of sulfuric acid solutions of ethyl ortho- carbonate are similar to those recorded for the salts prepared by Meer- wein and his group In a similar fashion, Taft and his co-workers have prepared alkoxy- and dialkoxycarbonium ions by the addition of ketals and ortho esters to 25% S03-75% HzS04 and 96% HzS04, re-

spectively

22 E H CORDES

In addition to direct demonstrations of the existence of the postu-

lated carboniurn ions, Taft and Ramsey have observcd that dilute solutions of niethyl orthocarbonate in anhydrous methanol exhibit only a single inethyl proton resonance absorption in the presence of

0.5-1.OM toluenesulfonic acid (69) The coalescence of the methyl

peaks of orllio ester and methanol is most reasonably interpreted as

reflecting the rapid and reversible formation of the trimethyl ortho- carbonate carbonium ion The lifetime of this carbonium ion in

anhydrous methanol solutions containing unit concentrations of ortho ester and acid is estimated to be approximately 0.001 see

Direct evidence is also available for the existence of compounds closely related to the implicated tetrahedral intermediates Bender has observed the addition of methoxide ion to acyl-activated esters

in n-dibutyl ether solutions with formation of salts of dialkoxy ortho

esters, the pertinent inlermediate for ortho ester hydrolysis [cy

(1 1) ] (7 1) The structural assignment indicated in ey (I 1) was nude

R- 8 -0Et + CHiO- + 1%- 1- -0Et (11)

on the basis of the disappearance of the characteristic infrared ab- sorption of the carbonyl function upon the addition of methoxide and the extreme lability of these compounds to nucleophilic reagents Furthermore, the reaction between ethylene glycol and chloral, the reverse of cyclic acetal hydrolysis, has been demonstrated to take place in two steps The intermediate, a hemiacetal, is stable and may be prepared in crystalline form (72) Addition of sulfuric acid

to this intermediatc yields the cyclic acetal [eq (12)] In two cases,

(73,74)

ACETALS, KETALS, AND ORl’HO ESTERS 23

Research 011 acetal, ketal, and ortho ester hydrolysis has been characterized over the last two decades by a surprising near-una- riimity on the question of mechanism In particular, only two serious suggestions have been advanced to the effect that these reactions may proceed with nucleophilic participation of solvent First, Kwart and Price, on the basis of correlation of rate constants with an em- pirical solvent composition-acidity function (75,76), have proposed that the hydrolysis of methyl p-nitroorthobenzoate occurs by an

the solvent was varied by altering the methanol content in methanol- water mixtures, an unfortunate choice of solvent system as originally realized by I h e v o y and Taft (18) As pointed out by DeWolfe and Jenseri (27): “Regardless of the mechanism, the rate of solvolysis

of a methyl ester in aqueous methanol is not determiried by the con- centration of protonated ester alone since reaction with methanol regenerates starting material while reaction with water yields hy- drolysis product The departure of the observed rates from propor- tionality to concentration of ortho ester conjugate acid may be due entirely to competition between methanol and water in the product- forming step.” This conclusion receives mild support from the sol- vent effects on hydrolysis of this substrate cited above (Fig 4)

In addition, investigations discussed later reveal that water arid meth- anol are about equally reactive toward intermediate carbonium ions derived from ortho esters, demonstrating rather directly that the sug- gestion of DeWolfe and Jensen is correct Consequently, the results

of Kwart and Price cannot be taken as evidence for or against solvent participation as nucleophilic reagent for ortho ester hydrolysis The above points also render uncertain the interpretation of structure- reactivity relationships investigated by these authors for the hydroly- sis of substituted methyl orthobenzoates (Table I) inasmuch as the reactions were carried out in 70% methanol A reinvestigation of structure-reactivity relationships for these substrates would be of interest

Second, kinetic studies of ICactling and Andrcws on the hydrolysis

of the diethyl :mtd of p-nitrobeiizoi)hcrioiie have yielded results con- sistent with cither an A-1 or A-2 reaction path, the latter alternative being favored by these authors (77) These reactions, like those dis- cussed just above, were carried out in concentrated alcoholic solutions

by carbonyl carbon-oxygen bond cleavagc and is unimolrcular, that is, this transition state is closcly related to 1 (or a variant thcrcof) Second, eq ( 3 ) is the minimum reaction path required to account for these roactions A closer examination of these points is presented

in the following section

111 The Rate-Determining Step

Several tinics in thc foregoing discussion it has hccn pointed out that cwtairi findings (e.g., entropies and volumes of avtivation) are evidciicc riot only for a reaction path involving the formation of a carbonium ion but for rate-determining unimolecular formation of a cwboninm ion It does not follow, howcvcr, from this coilelusion that formation of the cwbonium ion, the first reaction in eq (3), is neocssarily rate detcrmining This is true siricc thc decomposition of

the tetrtthcdral intermediate, the terminal step in eq (3) is also almost

rcrtainly uriiniolecular and yields a species posscssirig carbonium ion

H

OH +B tH’_ \& -3 \&)- - \ C=O (13b)

/

- -H’ / \+ -B H /

’ ‘OR L;R

character This decomposition may be corisidcrcd to yield the (son-

jugate mid of a carboniyl compound [eq (13a) ] or the c*arbonyl voni-

pound itself [cq (13b)l Thus, there exist two closely related species,

ACETALS, KETALS, AND ORTHO ESTERS 25

both formed in unimolecular processes on the reaction pathway, and t]he formatlion of either may he rate-determining Rlechanisms in- volving rate-determining formatioii of either the cwbonium ion or the product (or its conjugate acid) are (#orisistent with substantially all

of the experimental iriforniation detailed above In view of thcse con- siderations, it is somewhat surprising that the initial step in eq (3) has been nearly unanimously agreed on in the literature as the rate- determining step

A SOME GENERAL CONSIDERATIONS The conclusion that formation of the carboiiium ion is rate-deter- mining is in accord with expectations based on chemical arguments

In solutions containing little alcohol, rate-determining reaction of this carbonium ion with solvent is extremely unlikely since this requires that alcohol react with the carbonium ion, regenerating starting ma- terial, more rapidly than water reacts with the carbonium ion, yield- ing products Since the rate constants for reaction of alcohol and water

are almost certainly about the same (see p 28), the rate for the latter reaction must be greater than that for the former A similar

argument suggests that tetrahedral intermediate decomposition, by either route (13a) or (13b), is not rate-determining Since the overall equilibrium constant for interconversion of, for example, methyl ortho- benzoate and dimethyl orthobenzoate should be about unity, the latter would be present in much greater concentration than the former were equilibrium established, due to the high concentration of water relative to methanol Since the rate constants for decomposition of these species should be about equal (or that for dimethyl orthoben- zoate may actually be appreciably larger if reaction occurs by route (13b) as a similar route is unavailable to methyl orthobenzoate), the rate for dimethyl orthobenzoate decomposition should be greater than the corresponding quantity for methyl orthobenzoate These con- clusions are fully corroborated in kinetic studies of ketal and ortho ester hydrolysis conducted in the presence of deuterated alcohols which will now be described

B KINETIC STUDIES EMPLOYING PROTON MAGNETIC

RESONANCE SPECTROSCOPY

A study of the kinetics and product composition for the hydrolysis

of methyl ketals and methyl ortho esters in met~hanol-d4-deuterium

26 E H CORDES

oxide mixturw [eq (14)], employing proton magnetic resonancc

spectroscopy as an analytical tool, has provided a simple and straight- forward experimental distinction between several of the possible rate- determining steps for these reactions (78,79) Thc experiment a1

quantities determined by this method in the study of, for example, methyl orthobenzoate hydrolysis, include the first-order rate con- stants for the disappearance of the methoxy protons of the ortho ester, kortho for the appearance of the methyl protons of methariol,

k M e O W , and for the appearance of the methoxy protons of the car- boxylic ester, lcester Since the proton resonance singlets for each of these groups are well separated, the rate constants can be determined simultaneously I n addition, the ratio of integrated proton intensities

at infinite time of the products to some internal, time-independent standard provides n quant,itative measure of produet composition for many wibstrates

Both the product composition and the relative magnitudes of the various rate constants are functions of the nature of the rate-detw- mining step If cnrbonium ion formation were rapid and reversible (i.e., cnrboniiim ion formation not rate-determining), the melhoxy

Fig 5 Initial, iritermediate, and final proton magnetic resonance spectra for

the hydrolysis of met,hyl orthobenzoate in an eqiiimolar mixture of deuterium oxide

and methanol-& Methoxy protons of ort,ho ester appear a t 3.0 p.p.m., those of

methyl benzoate a t 3.8 p.p.m., and those of methanol at 3.25 p.p.m

groups of the starting material would be rapidly exchanged for deu-

teriomethoxy groups through reaction of the carbonium ion with

solvent deuteriomethanol The ortho ester would be converted more

slowly to carboxylic estcr produvt Thus, kortho ester and ~ M ~ O H

would be considerably larger than kester Furthermore, little or no

carboxylic ester product containing methoxy protons would be

formed since virtually all of the ortho ester would have been con-

verted into the corresponding deuterated material in the preequi-

librium exchange reactions In contrast, if carbonium ion formation

were rate-determining, methanol would not be exchanged for deuterio-

methanol in a preequilibrium reavtion; hence, kortho ester, ~ M ~ O H , and

kester would be nearly identical In addition, only protio carboxylic

ester woulcl be produrcd as reaction product,

Studies of this type have been performed employing 2,2-dimethoxy-

propane, 6,6,6-trimethoxyhexanonitrile, methyl orthobenzoate, arid

methyl orthocarbonate as substrates (79) The course of the hy-

28 E H COHDEM

drolysis of methyl orthoberizoate as a function of time is indicated

in Figure 5 As may be judged qualitatively from this figure, the rate of disappearance of the methoxy protons of the ortho estcr is comparablc to thc rate of appearance of the corresponding protons

of methyl benzoate Furthermorc, a rather substantial amount of methyl beiizoatc, as opposed to deuteriomethyl benzoate, is formed

in tlic reaction as judged from thc intensity of the appropriate signal

in the infinite time spectrum These results are just those predicted

011 thc basis of rate-determini ng carboniurn ion formation The

TABLli: VI

Pirsl-Order Rate Constants for the Hydrolysis of Ketals and 01 tho Ir:sters in

Deuteriomethanol-Deuterium Oxide Solutions

VI In each ( m e , the characteristic rate constants are similar A

quantitative treatment of the kinetics of reactions of this type reveals that the relative magnitudes of the various rate constants and the product composition pattcrris are those expected for rate-determining carbonium ion formation provided that dcuterium oxidc and deuterio- methaiiol arc about cqudly reactive toward the cnrbonium ion The lattcr conclusion is in accord with the relative rcactivities of water and methanol toward, for example, the t-butyl and benzhydryl

carbonium ions (80-83)

A C E T A M , XETALS, AND ORTHO E S T E R S 29

We consider that the above results clearly establish that the first step indkated in eq (3) is rate determining for a t least those sub-

strates explicitly studied, and that they provide strong evidence for similar hehavior in the ('ttses of related substrates It is iniportarit

to recognize, however, that the first step in eq (3) is, in fwt, com- posed of two steps: protoriation and decomposition The data cited above do not distinguish between rate-determining protoriation and preequilibrium protonation It is also possible, of course, that pro- tonation and carbon-oxygen bond caleavage may be concerted proc- esses

C FIJRTHER CONSInERATIONS CONCERNING STRUCTTJRE-

REACTIVITY CORRELATIONS The notion of rate-determining decomposition of the protonated substrates for acetal and ketal hydrolysis, A-1 reaction paths, is fully consistent with all of the known facts concerning these reactions in- cluding the strong accelerating eff ecat of electron-donating polar sub- stituents (Table I) Thus, those substrates yielding the most stable carbonium ions react most rapidly The situation is not quite so straightforward for the hydrolysis of ortho esters in the regard that rates of hydrolysis do not parallel the expected stabilities of the cor-

responding carbonium ions As noted in Table 11, ethyl orthocar- bonate is less reactive than ethyl orthobenzoate which is, in turn, less reactive than ethyl orthoacetate Thus, for this limited set of substrates, the rates are actually inversely related to carbonium ion stabilities Neglecting the now excluded possibility that these reac-

tions are A-2, several possible explanations remain for the failure of

substituents capable of electron donation by resonance to accelerate the rate of ortho ester hydrolysis

First, DeWolfe and Jensen have noted that rates for ortho ester hydrolysis follow the inductive effect of the substituents and, on this basis, have suggested that the transition state for these reactions is reached so early along the reaction coordinate that the central carbon

is essentially tetrahedral therein (27) If this were the case, electron

donation by resonance would be of little importance and the rates would be expected to follow the inductive effects alone We do not find this hypothesis appealing particularly in view of the large Bron- sted alpha value for general acid catalysis of orthocarbonate hydrol-

30 E H CORnES

a Roiinded values of rehtive stabilization energy of cation based upon the as-

siimption that sitbstitiient effect,s in the neiitral substrate are negligible hy com-

parison to those in the cation

ysis (see p 32) This point of view seems to have been abandoned

by a t least one of the above authors (25)

That is, it

is possible that the transition state for a diallioxy carbonium ion is so stabilized by these functions that the remaining aryl, alkyl, or alkoxy substituent does not lend appreciable further stabilization The appearance potential measurements of Martin, Taft, and Lampe indicate that a saturation effect does indeed occur in the gas phase,

at least (84) In Table VII, appearance potentials and the derived relative stabilization energies are collected for the formation of several cwboriiurn ions by loss of OCHa The appearance potential results parallel, qualitatively at least, the acid-catalyzed hydrolysis rates of the appropriate suhstrates It is, of course, not possible to extrap- olate these data directly to the situation in aqueous solution in which so1v:ition cnergies are certainly important Furthermore, satura-

have not previously been noted in carbonium ion reactions

in solution as evidenced by the fact that the relative rates of solvol- ysis of CeH6CH2C1, (C6H&CHC1, and (CsH5),CC1 are approximately 10": lo5: lo9 (85) Nevertheless, it does scem quite liltely that a satu- ration effcvt does owur, to a modest extent at least, in the caw of ortho ester hydrolysis and that this saturation effect arrounts, in part, for the obscrved order of reactivities

Third, Him has suggested that the relative rates of hydrolysis of ethers, ltctnls, ortho esters, and orthocarbonates can be accounted for

A second possibility is that a saturation effect occurs

ACETALS, KETALS, AND ORTHO Esrmis 31

on the basis of the summation of substituent effects upon starting material and transition state employing the concept of double bond-

no bond resonance (86) Consider, for example, the cases of ethyl

orthocarbonate and ethyl orthoacetate With the former substrate,

a total of 12 double bond-no bond resonance structures may be writ-

ten while only six will contribute to the structure of the latter sub-

strate Since stabilization by double bond-no bond resonance will

be of lesser importance in the transition state (and will probably com- pletely disappear in the carbonium ions themselves), this factor will tend to decrease the reactivity of the orthocarbonate relative to the orthoacetate This argument may be extended to account for the relative reactivities of ethyl orthobenzoate and ethyl orthoacetate as well, since substituents which donate electrons by resonance will themselves be able to participate in double bond-no bond resonance

Thus, one would expect structure 5 to make a greater contribution to

the stability of methyl orthobenzoate than structure 6 to the stability

of ethyl orthoacetate The force of these arguments is weakened

somewhat by the lack of thoroughly reliable information as to the quantitative importance of double bond-no bond resonance for sys- tems involving oxygen-oxygen internc.tions Virtually no informa- tion is available concerning the importance of such stabilizations of systems involving carbon-oxygen interactions

Finally, one may argue that ortho ester hydrolysis does not proceed via an A-1 reaction path but involves proton transfer in the rate-

32 E H CORDES

determitiing step (25) Such reactions may be regarded as electro-

philic displacmierits 011 oxygcn and are designated S,2 re:wtions The postulittioii of :in S,2 reaction nieehanism does not, in itself, account for the ohserved structure-reactivity rclationshipb It does introduce riovcl c8otisitlcr:itions into the evaluation of such effects and provides a new framework into which the factors cnumcrtllccl above may be p1:tced We cwisidrr the arguments suggesting that these reactions

arr, in fact, Sl2 in character in some detail later

IV Catalysis

The hydrolysis of acetals, kelals, and ortho esters is subject to catal-

Certairi of these sub-

These in-

ysis by the hydrated proton, as noted earlier

strates arc subject to catalysis by other species as well

clude acids in general and certain detergents

Ortho ester hydrolysis was the first reaction demonstrated to be wbject l o the phenonienon of general acid catalysis In 1929,

Broristed arid Wyiiiie-Jones found that the first-order rate constantb for hydrolysis of ethyl orthocarbonate, ethyl orthoacetate, and ethyl orthopropionate in aqueous solution increased with increasing buffer roncentration at constant pH (87) Thc hydrolyses of ethyl ortho-

formate mid methyl orthobenzoate are, iri addition, subject to general wid cat:tlysis in aqueous dioxane and aqueous methanol solutions, respectively (17,88) The Bronstcd a value for general avid catalyzed elhyl orthoc*arbonate hydrolysis is near 0.70 (61,87) :LS is that for

methyl orthobcnzoatc hydrolysi.; (17) Although rather few system-

atic studies have been carried out, it seems likely that tlic hydroly '

of all ortho esters will be characterized by large v:tlues of a

In coritra5t to the ( m e for ortho ester hydrolysis, there seem to bc

no reports of general acid cataly4s for aretal and ketal hydrolysis

A rather careful searcah for such catalysis eniploying 2,2-dimcthoxy-

propane arid the diethyl ketal of aretophenone as substrates and car- boxylic arid buffers as potential catalysts revealed only slight de-

likely that this rate inhibition results from trapping of the carbonium ion by the cnrboxylate ions with formation of acylals The lattcr substrates probably hydrolyze slightly more slowly than the ketals urider the conditions of these experiments

ACETALS, KETALS, AND ORTHO ESTERS 33

The finding of general acid catalysis for ortho ester hydrolysis im- riiediatcly raises the question of tho mcchhanism of the hydroriium ion catalyzed rcaction The following possibilities exist First, sub- stratc protonation is :L precquilihrium reaction and the subsequent

decomposition of the conjugate acid of the substrate is rate-determin-

ing (transition state 7) Second, substrate protonation is the rate-

determining step and the subsequent decomposition is rapid (transi- tion state 8) Third, substrate protonation and decomposition are independent reactions and occur with approximately the same rate

(transition states 7 and 8 both important) Fourth, proton transfer

to the substrate and substrate decomposition are concerted processes (transition state 9) The last possibility itself may be refined into

subcategories which are considered below It is to be noted that each

of the above transition states is related to transition state 1 originally

proposed and whose basic features are correct for acetal and ketal hydrolysis, a t least Interesting calculations recently reported by

by the aliphatic substituent constants of Taft, o*, with a p* value of approximately 3 (89), the basicities of ortho esters may be obtained

by use of the appropriate substituent constants, the above assump- tion, and the values of pK, of -3.6 for diethyl ether and -3.8 for dimethyl ether (90) The estimated values of pK, are -5.4 for 1,l-

dimethoxyethane, - 7 for ethyl orthoformate, - 7.6 for ethyl ortho- benzoate, and -8.5 for ethyl orthocarbonate (83) If we formulate ortho ester, ketal, and acetal hydrolysis as