Kinetic deuterium isotope effects in the reactions of ... - ACS Publications

---

KINETICDEUTERIUM ISOTOPE EFFECTS

427

Kinetic Deuterium Isotope Effects in the Reactions of Methyl Iodide with Azide and Acetate Ions in Aqueous Solution1 by Chong Min Won and Alfred V. Willi*2 College of Pharmaceutical Sciences, Columbia University, New York, New Fork

100$8 (Received August 80, 1971)

Publication coats borne completely by The Journal of Physical Chemistry

The kinetic a-deuterium isotope effect has been determined for the reactions of CH31 (CDaI) in aqueous sohtions with azide ion from 10 to 40” and with acetate ion a t 40’. Kinetic measurements have been carried out in reaction vessels with no gas phase. Rate constants of the reaction of methyl iodide with acetate ion are corrected for simultaneous solvolysis. Results indicate inverse isotope effects: k H / h = 0.907 (CH3I Na-)! and 0.882 (CHJ CH&!OO-) a t 40’. The experimental data for the reaction with N3- are compared with results of model calculations of isotope effects from force constants with an electronic computer. Reasonable choices are made for the transition state force constants ~ C I ~! C N fn, , ~ H C I ,and the force constants not involving reacting bonds. The bending force constant fHCN (in general: ~ H C Y )is adjusted to fit the experimental isotope effect at one temperature. Good agreement then is obtained between calculated and experimental isotope eflects a t all four temperatures. In the series CHZI S Z O P , CN-, N3-, or CH&OO-, a very good correlation exists between the values of k H / k u and AG*, but there is no relationship between AG* (or the nucleophilic power of Y-., respectively) and the adjusted value of the bending force constant, ~ H O Y . The particular role of solvation in the S N transition ~ state is discussed. The models of the isotope effect calculations must be improved to take care of the solvation of the groups X and Y.

+

+

+

This paper is the third in a series3s4referring to experimental studies of secondary Q! isotope effects in SN2 reactions of methyl iodide (CHd in comparison to CD3I) with various nucleophiles. The work is done in comparison with computer calculations of isotope effects from vibrational force constants in order to gain insight into the nature of the transition states and to arrive a t a better understanding of the causes of the isotope effects. The preceding two papers were concerned with determinations of deuterium isotope effects in the reactions of methyl iodide with cyanide iona and thiosulfate This work refers to deuterium isotope effects in the reactions of methyl iodide with azide ion and acetate ion. Results are now available for deuterium isotope effects in reactions of methyl iodide with four different nucleophiles, and it has become possible to draw some general conclusions. The reaction of methyl iodide with azide ion is a simple second order process, and the integrated rate equation l is adequate for the treatment of the experimental data. log (IN3-V [CHd 1) = log ([Na-]o/ [CHaIIo)

+

([Ns-lo - [CHaI]o)kzt/2.303 (1) The reaction with the weakly nucleophilic acetate ion is complicated by simultaneous first order solvolysis with water according to rate eq 2 -d[CHaI]/dt

== kz[CHd] [CH&OO-]

+ kl[CHaI]

(2)

The two parallel reactions may be formdated as CH3I

+ CHaCOO- +CHaCOOCH3 + I-

(kz)

CHsI

+ H2O + CHaCOO- --+ CHaOH + CHaCOOH + I-

(h)

Strong acid formed in the solvolytic reaction is immediately neutralized by acetate ion which is present in excess. Consequently, acetate ion is consumed in both processes which leads to a relatively simple stoichiometric relationship [CHaIIo - [CHJ]

=

[CHaCOO-Io

-

[CH3COO-]

[I-I - [I-lo

==

(3)

Combination of eq 2 and 3 and subsequent integration leads to the integrated rate equation 45*6

with

(1) Taken from the final part of the thesis of Mr. C. M. Won, submitted to the Faculty of Pure Science, Columbia University, New York, N. Y . ,in partial fulfillment of the requirements for the degree of Doctor of Philosophy. (2) Address correspondence to this author a t Kampenweg 22, D 237 Rendsburg, West Germany. (3) A.V. Willi and C. M. Won, J.Amer. Chem. Soc., 90,5999 (1968). (4) A. V. Willi and C. M. Won, Can. J. Chem., 48, 1452 (1970). (6) E. A. Moelwyn-Hughes, Proc. Roy. SOC.,Ser. A , 196,540(1949). (6) Szabo discusses a different case in which the nucleophilic reactant is not neutralized by acid produced in the first order reaction. 2.G. Szabo in “Comprehensive Chemical Kinetics,” Vol. 2,C. N.Bamford and C. F. H. Tipper, Ed., Elsevier, Amsterdam, 1969,pp 45,46. The Journal of Physical Chemistry, Vol. 76,No. 3,1972

428

CHONGMIN WONAND ALFREDV. WILLT

The rate of formation of acid is equal to the rate of the solvolysis reaction

computer with a linear regression program. lcz was calculated from the slope of the line, according to eq 1. It was necessary to apply eq 4 and 6ain the calculation d[acid]/dt = kl [CH31] (5) of the bimolecular rate constant of the reaction of methyl iodide with acetate ion. The ratio Icl/lcz was Equation 6a describes the increase of the acid concencomputed from a pair of values of [acid] and [CHJ] tration as a function of time. It is obtained by combination of eq 2, 3, and 5 and subsequent i n t e g r a t i ~ n . ~ , ~measured a t the same time. This was carried out with a method of successive approximations in which a pre(kilkz) Bo liminary value of kl/kz was inserted at the right hand [acid] - [acidlo = (kl/lcz) In A (ki/kz) Bo - Ao side of eq 6b. An improved value of lcl/kz then was ob(64 kl/kz = [acidl/ln { [ ( k d k z ) Boll

+

+ +

+

Experiments are usually carried out in such a way that [acidlo = 0. As far as the writers know, previous kinetic studies of the reactions of methyl iodide with azide ion or acetate ion have not been published.

Experimental Section General procedures of purification of reactants and solvent and of the kinetic measurements have been described p r e v i ~ u s l y . ~The reaction vessels with no gas phase were variants of those described by Fahim and Moelwyn-Hughes.’ In the measurements of the rates of reaction of methyl iodide with azide ion, 5-ml samples of the kinetic solutions were acidified with 10 ml of 6 N aqueous nitric acid (to prevent precipitation of AgN3) and then titrated with standardized 0.01 N silver nitrate solution at a silver electrode, utilizing a mercurous sulfate reference electrode. The same procedure was applied in the measurements of the rates of reaction of methyl iodide with acetate ion, except that the solution samples were acidified with 1ml rather than 10 ml of 6 N nitric acid. The concentration of acid formed in the experiments with acetate ion was determined at 50-60% conversion of methyl iodide, by titration of 5-ml samples with standardized 0.01 N sodium hydroxide solution, using phenolphthalein as an indicator. The initial concentrations of methyl iodide in the reactions with azide ion were evaluated from iodide ion concentrations in solution samples in which the reaction had gone to 99.9% completion. This procedure was not feasible in the reactions with acetate ion, because of the low rate. However, methyl iodide could be transformed to iodide ion within a reasonable time period by adding a small volume of excess potassium cyanide solution to the solution in the kinetic cell and waiting until the reaction was 99.9% complete. (The reaction time was calculated from known rate constants.3)

Calculation of Rate Constants from Experimental Data For each kinetic run of the reaction with azide ion, the best straight line was fitted to the experimental points of log ( [ni3-]/ [CHJ]) as a function of t, using a The Journa.1 of Physical Chemistry, Vol. 76, N o . 5, 1979

[A

+ ( h / k z ) + Bo - Aol)

(6b)

tained from eq 6b which was utilized a t the right hand side of the equation in the next approximation. This procedure was repeated several times until kl/lcz remained unchanged within 0.1%. Two or three separate determinations of Icl/lc~were carried out in each kinetic run, and the average value was computed. (A value of kl/kz = 0.094 M was obtained for t>hereacting system CHsI CHaCOOHZO at 40°.) In the next step, eq 4 was utilized for the calculation of kz from a series of experimental values of [CHJ] as a function of t. The regression line for the expression at the left, hand side of eq 4 as a linear function of t was calculated in a computer with a linear regression program.

+

+

Results In all experiments, the initial concentrations of methyl iodide or methyl-& iodide are ca. 2 X low2M , those of sodium azide are ca. 5 X lov2M , and those of sodium acetate are ca. 0.25 M . Plots of In ([i\Ta-]/ [CHJI), or In { ( [ C H J I (kdkz) [CH~COO-IO[CH3IIo)/[CHJ J], respectively, as functions of t are linear through 2 half-lives. Consequently, the reverse reactions do not interfere under the conditions of these experiments. Results of second-order rate constants and isotope effects, obtained at different temperatures, are collected in Table I. Each rate constant is the average value of three to five determinations in separate kinetic runs.

+

+

Table I: Rate Constants and Isotope Effects in the Reactions of Methyl Iodide (RI) with Azide Ion and Acetate Ion in Water R = CHa 10~k2, Nucleophile

azide ion azide ion azide ion azide ion acetate ion

mol-1 1.

10Sk2, see-1 mol-] 1.

ICE/~D

0.6819 2.962 11.91 40.45 0.826

0.7550 3.264 13.20 44.68 0.936

0.903 (+0.004) 0.901 (A0.007) 0.903(10.011) 0.907 (10.019) 0.882 (f0.012)

Temp,

880-1

O K

283.2 293.2 303.1 313.2 313.2

R = CDa

(7) R. B. Fahim and E. A. Moelwyn-Hughes, J. Chem. Soc., 1035 (1956).

KINETICDEUTERIUM ISOTOPE EFFECTS Arrhenius parameters of the reaction with azide ion are given as follows: for CH31 N3-, A = 2.48 X 10la (122Oj,)sec-' mol-' l., E, = 24.05 (10.41)kcal; for CD31 N3-, A = 2.82 X lo1*( *22y0) sec-l mol-'l., E, = 24.05 (Lt0.41)kcal; isotope effect, AH/& = 0.879 (10.033), ea^ - E,D = - 3 cal.

429

+

+

0 1-

I.o

kHk

.

Discussion

81

Relationships between Isotope E$ect and Nucleophilic Power. Inverse a-deuterium isotope effects are found in the reactions of methyl iodide with azide ion and acetate ion both of which are relatively weak nucleophiles. Inverse isotope effects have been observed previously in the reactions of methyl iodide with stronger

.so .

\

"'\ CH3COi 0

\

in the reactions of methyl, ethyl, and n-propyl bromides Comparison with Calculations of Isotope Efects f r o m with thiosulfate ion." Normal deuterium isotope efForce Constants. Model calculations of isotope effects fects are found also in the chloride exchange r e a ~ t i o n ' ~ ~ ' ~ from vibrational force constants for the reaction of and in the solvolysis14of benzyl chloride, as well as in methyl iodide with azide ion have been carried out the reaction of benzyl chloride with cyanide prowith the aid of the Wolfsberg-Schacht~chneider'~ Seltzer and ZavitsaslO observed a linear relationship gram. The geometry and the force constants of the between the a-deuterium isotope effect and the differin model of the reactant methyl iodide are the same as ence in nucleophilic powers of the entering and leaving previous cal~ulations.~~4~20 In the model of the transigroups (Edwards' E , constants15). If the isotope eftion state, the central CHagroup is planar, with L HCH fects of the reactions of methyl iodide with thiosulfate = 120°, L H C I = LHCN = 90°, and the following ion4 (20°),cyanide ion3 (20°),azide ion (30°),and acebond lengths: r C H , 1.088; rcI, 2.148 10% = 2.35 tate ion (40")in aqueous solution are plotted us. the 8; r C N , 1.478 4- 10% = 1.62 8. According to the AG values, a straight line is obtained with surprisingly cutoff method suggested by Stern and WolfsberglZ1the good precision (Figure 1). The pbint for the iodide exazide group in the transition state is represented by two change reactionlo is above the line, however. If, on the points, the one being the reacting atom, the other other hand, the isotope effects in the four S N reactions ~ Nz group, with T N N = 1.24A22and LC" = one the studied by the writers are plotted us. E , - El, the ex120". perimental points form a straight line (with some scatTransition state f:rce constants are chose? as follows: tering) 0.04unit below the original line drawn by Seltf" = 9.12 mdyn/AlZ3fc" = 0.1 mdyn A, fHc" = zer and Zavitsas.lo (The point for the iodide exchange reaction of methyl iodide is far above the line.) (8) K. T. Leffek and J. W. McLean, Can. J . Chem., 43, 40 (1965). A correlation of similar appearence is obtained if a(9) J. A . Llewellyn, R. E. Robertson, and J. M. W. Scott, ihid., 38, 222 (1960). deuterium isotope effects in SN reactions of methyl (10) S. Seltzer and A . A. Zavitsas, ibid., 45, 2023 (1967). compounds are plotted us. the difference of the Swain(11) K. T . Leffek, ibid., 42, 851 (1964). ScottlGnucleophilic constants of the entering and leav(12) B. Ostman, J . Amer. Chem. Soc., 87, 3161 (1965). ing groups, n, - nl. It is necessary to use the n values (13) H . Strecker and H. Elias, Radiochim. Acta, 7, 22 (1967). of the halide ions suggested by Petty and Nichols.'' (14) A . V. Willi and Chih-kuo Ho, to be published. The reactions which yield a fairly good linear correla(15) J. 0. Edwards, J . Amer. Chem. Soc., 76, 1540 (1954); 78, 1819 tion between isotope effects and nucleophilic powers are (1956). "downhill" processes, Le., the products are more stable (16) C. G. Swain and C. B. Scott, ihid., 75, 141 (1953). (17) W. L. Petty and P. L. Nichols, ihid., 76, 4385 (1954). than the reactants, no matter whether the kinetic "nu(18) A. V. Willi, 2. Phys. Chem. (Frankfurt a m M a i n ) , 66,317 (1969). cleophilic power" of the entering group is greater or (19) M . Wolfsberg and M. J. Stern, Pure Appl. Chem., 8 , 225 (1964). smaller than that of the leaving group. (It was em(20) A . V. Willi, Can. J. Chem., 44, 1889 (1960); 2. Naturforsch. A , . phasized in a previous paperl8 that the Swain-Scott and 21, 1377, 1385 (1960). Edwards scales refer to rates but not to equilibria.) (21) M. J. Stern and M. Wolfsberg, J . Chem. Phys., 45, 4105 (1966). On the other hand, the equilibrium constant of the (22) The Chemical Society, "Tables of Interatomic Distances and Configurations in Molecules and Ions," Burlington House, London, iodide exchange reaction of methyl iodide must be ap1958. proximately equal to 1, except for a small iodine isotope (23) W. Engler and K. W. F. Kohlrausch, Z. Phys. Chem., Abt. B , effect . 34, 214 (1936).

*

+

The Journal of Physical Chemistry, VoE. 76, N o . 3, 1979

430

CHONGMIN WONAND ALFREDV. WILLI

Table 11: Calculated Isotope Effects in the Reaction of Methyl Iodide with Azide Ion fCN

hz

fRCr

0.20

0.20

1.50 4.43 4.64

0.295 0.295 0.295

1.01 2.85 1.09 2.80 Experimental results 0.

-

Transition state force constants5

,-fCI

---~-DEI/-

7

100

fHCN

0.340 0.320 0.320

0.895 0.896 0.892 0.903

200

30'

0.900 0.900 0.897 0.901

0.905 0.905 0.902 0.903

40"

0,909 0.909 0,905 0,907

Stretching force constants are given in mdyn/b and bending force constants are given in mdyn b.

0.1 mdyn (torsion). The values of the foye constants for the CH stretching ( ~ C H= 5.05 mdynlA in all calculations in this paper), HCH bending, and CH8outof-plane bending motions, as well as the interactions among these motions, are the same as in previous calcul a t i o n for ~ ~ transition ~ ~ ~ ~ ~state models of the general type 1...CHI. .Y. The remaining five force constants are related to reacting bonds. Sample calculations are done with three , f12 (interaction different sets of values for f c N , ~ C I and between the CN and CI stretches). The bendins force constant ~ H C Iis kept constant a t 0.295 mdyn A (the value found for the transition state I - *CH3. * I ) , the bending while for each of the three sets of f c N , ~ C If12 force constant fHCN is adjusted to fit the experimental isotope effect. The three sets of stretching force constants of the reacting bonds have been selected on the following bases: (1) f c N = fcI = 0.2 mdyn/& fit = 1.5 mdyn/d (no particular theoretical model). (2) The force constants are calculated from a semiempirical model of the energy barrier with the aid of a Johnston-type equation.24 The model has been described previously.*S Starting from VI = VZ = 37.2 = AG* = 23.90 kcal, the followkcal, pl = 2.26, V, ing values are obtained: p2 = 2.73, n = 0.477 (GI bond order), $* = 2.53 m d y n / i (force constant describing the curvature of the barrier). (3) The same model is applied, but it is assumed that Vz > VI, as the products are more stable than the reactants. The values of V1 = 37.2 kcal, V , = 45.7 kcal (arbitrary assumption, a well-founded value may be computed from the equilibrium constant which has not yet been measured), pl = 2.26, and V,, = 23.90 kcal lead to p 2 = 2.90, n = 0.486, f* = 2.74 mdyn/& The force constants of sets (2) and (3) are calculated from n and f * with the aid of eq 7, 8, and gZ4 fCN

= FCN(1 ~ C = I

fiZ

= [fCIC2

C

- n>

FCI~

+ + (1 + C2)f*1/2C =

(78)

(7b)

fCN

(8)

(1 - n)/n

(9)

(FCN= 5.45 mdyn/A26 and FCI = 2.25 mdyn/A26 are the values of the stretching force constants in stable molecules.) The Journal o j Phvsieal Chemistry, Vol. 76,No. 5, 1979

Results of isotope effect calculations based on these three sets of transition state stretching force constants are collected in Table 11. The experimental value for ICH/JCD is obtained if the transition state bending force constant ~ H C Nis adjusted to a value in the range 0.320 to 0.340 mdyn A. I n all three examples, calculated and experimental isotope effects agree within experimental error a t all four temperatures. I n Table 111, transition state bending force constants3t4*z0involving one reacting bond are compared with bending force constants for the corresponding bonds in stable molecule^.^^-^^ I n all four examples, the transition state bending force constant is equal to 47-63% of the bending force constant in the stable molecule. There is no relationship between the nucleophilic power of Y- and the transition state bending force constant, fHCY) or the ratio, . f H c x / F H c Y . Table I11 : Transition State Bending Force Constants -Transition

stata-

Stable

mdynA

fHCY mdynA

moleoule FHCY rndynA

0.595" 0.590" O . 695' 0.615g

0.300a 0.2950 O , 400' 0.320g

fHCI fECY

Reaction

CHaI CHJ CH3I CHSI 26.

25.

+ StoaZ+ I+ CN+ Na-

0.52tib 0.55d O. 64'

0.681h

Reference 4. b Reference 27. c Reference 20. Reference 3. Reference 28. This work.

'

fHCY/

FHCY

0.572 0.537 O . 626 0.470 Reference

' Reference

The Role of Solvation in the Mechanism of the S N ~ Reaction. As found recently by Bohme and Young29 applying the flowing afterglow technique, the reactions of methyl chloride with hydroxide or alkoxide ions in the gas phase are fast. Experimental rate constants are of the same order of magnitude as calculated collision rate constants. The same is probably true for (24) H. 8. Johnston, Advan. Chem. Phys., 3, 131 (1961). (25) J. W. Linnett, J. Chem. Phys., 8 , 91 (1940). (26) P. F. Fenlon, F. F. Cleveland, and A. G. Meister, ibid., 19, 1561 (1951). (27) H. Siebert, 2.Anorg. AIEg. Chem., 271, 65 (1952). (28) G. Heraberg, "Infrared and Raman Spectra of Polyatomic Molecules," Van Nostrand, New York, 1945, p 193. (29) D. K. Bohme and L. B. Young, J . Amer. Chem. Soc., 92, 7354 (1970).

431

KINETICDEUTERIUM ISOTOPE EFFECTS many other reactions of methyl halides with chargelocalized anioris in the gas phase. The rate of the gas phase reaction of methyl chloride with alkoxide ions is strongly decreased if the anion is solvated by one alcohol molecule. On the other hand, reactions of methyl halides with anions in solution are slow; they possess activation energies in the approximate range of 15-25 kcal. The rates of these reactions are strongly dependent on the solvation of the anion.a0 It is appropriate to rcopen ~ of the discussion of the mechanism of the S N reaction a methyl halide with a nucleophile in solution with regard to the findings of Bohme and Young. There are three possibilities : (a) The slow step is the desolvation of the anion (and the methyl halide). The free particles react in a fast subsequent step which must be diffusion controlled. (In this case, the rate of combination of the reactants is assumed to be faster than the rate of solvation.) (b) The solution contains very small equilibrium concentrations of the unsolvated reactants. The slow step is the diffusion-controlled combination of the reactants. (The rate of solvation is faster than, or of the same order of magnitude as, the rate of combination of the free reactants.) (c) The concentrations of unsolvatcd reactants in the solution are too low to maintain an appreciable reaction rate, and reaction takes place between solvated particles. If the reactants are solvated by more than one solvent molecule, it may be necessary to remove one solvent molecule from the coordination shells of each of the reactants before the alkyl halide can combine with the anion to form the transition state. (Both reactants still are partially solvated in the transition state.) Activation energy is required in the S N reaction ~ of the solvated particles. A decision among these three possibilities can be made on the basis of observed primary carbon isotope effects in the reactions of methyl halides with water,31 amines,31 h y d r o ~ i d e and , ~ ~cyanide32 ~~~ ions ( k 1 2 ~ / k 1 3 ~ = 1.03-1.07). Bond cleavage between central carbon atom and leaving group must occur in the rate-determining step. Therefore, mechanism (a) may be excluded immediately. Furthermore, mechanism (b) cannot be correct because the carbon-13 isotope effect on the rate of diffusion of the methyl halide must be much lower than 3%. Consequently, mechanism (c) must be correct, Le., the reaction takes place between solvated reactants, and it is slow because a solvated nucleophile is lcss reactive than a free one. There is a wealth of evidence for the bimolecular nature (solvent molecules not counted) of the S N reac~ tion in solution and the occurrence of backside attack by the entering n u ~ l e o p h i l e . ~One ~ or several solvent molecules must be bonded directly to the entering nucleophile in the transition state. According to the principle of microscopic reversibility, the leaving group

in the transition state must be solvated, also. Another way of describing the transition state of the reaction of a methyl halide with an anion would be as follows: the methyl halide and the attacking anion are in the same solvent cage, and the entering and leaving groups are strongly bonded to the walls of the cage. Conclusions Concerning the Transition State Model in Isotope E$ect Calculations. The considerations about the importance of transition state solvation are relevant with respect to the model chosen for the isotope effect calculations. A correlation is observed between the nucleophilic power of the attacking anion and the experimental a-deuterium isotope effect, but there is no clearly visible relationship between the nucleophilic power of Y- and the bending force constant, fHCY, or the sum of the bending force constants, ~ H C I ~ H C Y , respectively. As far as the calculations are concerned, one might fHCY and the iS0expect a relationship between fHCI tope effect, with little dependence on the mass and geometry of Y. Large values of ~ H C I ~ H C Ywould cor. sample calrespond to low values of k ~ / k ~However, culations with the same value of ~ H C I ~ H C Ylead to different results of k ~ / for k ~transition states with different groups Y. For example, a higher value of k ~ / k ~ is calculated for the reaction with CN- than for the reaction with N3-. Therefore, ~ H C I ~ H C Ymust be adjusted to a higher value in order to obtain agreement with the experimental isotope effect in the reaction with Chi-, even though the experimental value of J C H / ~ Dis lower in the reaction with N3- (Table 111). Obviously, the influence of the mass and geometry of Y is not negligible. The isotope effect also depends on the particular way how the motions of H (D) and Y are coupled in some bending vibrational modes of the transition state. The influence of these factors is substantial in comparison to the relatively small range of ob~ served values of k ~ / ink S~ N reactions. On the basis of these considerations and the results of sample calculations, a good correlation between isotope effects and vahes of fmI f H C Y would not be expected for different transition states. The same refers to correlations between isotope effects and nucleophilic powers of Y- which affect reacting bond orders and transition state force constants. A very good correlation between isotope effect and reactivity does exist, however, according to Figure 1. Consequently, the models applied in these calculations cannot be fully adequate. The influence of masses and geometry of entering and leaving groups on the transition state bending frequen-

+

+ +

+

+

+

(30) A. J. Parker, J . Chem. Soc., 1328 (1961); Quart. Rea. (London), 16, 163 (1962); Advan. 0x7. Chem., 5, 1 (1965). (31) M. L. Bender and D. F. Hoeg, J . Amer. Chem. SOC.,79, 5649 (1957). (32) K. R. Lynn and P. E. Yankwich, ibid., 83, 53, 790 (1961). (33),, C. K. Ingold, “Structure and Mechanism in Organic Chemistry, Cornel1 University Press, Ithaca, N. Y., 1953.

The Journal of Physical Chemistry, Vovol. 76, No. 3,1976

R. GOITEINAND THOMAS C. BRUICE

432 cies will be smaller, if it is considered in the model that these groups are solvated by clusters of associated water molecules. Solvation causes some restriction of the motions of X and Y in the transition state. I t is planned to carry out new model calculations of isotope effects with consideration of solvation of the reactant and the transition state.

Acknowledgments. The authors are pleased to acknowledge financial support of this work by the U. S. Atomic Energy Commission through Contract No. AT(30-1)-3796. Furthermore, they wish to thank Professor R. C. Kerber (State University of New York, Stony Brook) for reading the manuscript prior to submission.

Effect of Transfer from Water to 1.0 M Water in Dimethyl Sulfoxide on the Reaction of Nucleophiles with Phenyl Esters

by R. Goiteinl and Thomas C. B r i c e * Department of Chemistry, University of California, Santa Barbara, California 95106 (Received J u l y $6, 1.971) Publication costs assisted by the National Institutes of Health

Solvent S is defined as 1 M Ha0 in DMSO. In solvent SI water structure is eliminated, solvation of charged centers is reduced, the solvation characteristics of a polar aprotic solvent are approached, but water remains an effective nucleophile. For a series of nine oxyacids, a plot of pK, determined in S vs. pK, in water is of slope 2.2 as is a plot of the log of the second-order rate constants for acylate ion displacement on substituted phenyl acetates in S and H20,respectively. These data, together with previous observations of the alteration of AS for ionization and AS* for acylate ion displacement, are suggested to show that departure of phenoxide is important in the critical transition state. On transfer from H2O to S the second-order rate constant for HOattack on p-nitrophenyl acetate is increased 106-fold while the constant for (CH3)IN attack is increased by only 3.3-fold. Comparison of equivalent ionic conductivities suggests that the increase in activity of HO- on transfer from HzO to S is due to a decrease in the activity of HzO which solvates the HO- ion rather than to desolvation of this species.

In transferring from H20t o 1M HzOin DMSO, water structure is eliminated, solvation of polar or charged centers by water is reducedI2the solvation character,~ istics of a polar aprotic solvent are a p p r ~ a c h e dwhile water remains an effective nu~leophile.~Herein are described observations on the change of pK, values accompanying transfer from water to 1 M H20-DMS0 ("solvent S J J )and the values of equivalent conductivities of ions in solvent S. These experimental parameters are employed to provide a better description of the critical transition state for acylate ion attack on phenyl acetate esters and the role of solvation for HO- nucleophilic displacement on these esters. I n solvent S the usual quantitative approach t o acidbase chemistry is possible since pH is determinable4 via procedures originally described by Ritehiens The pK,'s of several phenols and trifluoroethanol have been determined in solvent S by the method of half neutralization employing the apparatus previously de~cribed.~The same apparatus was employed to determine the ionizaThe Journal of Physical Chemistry, Yol. 76, No. 8,1979

tion constants of water (pK,) in solvent S using the method of Harned and Fallon.e I n Figure 1 are plotted the values of the pK,'s of a series of oxyacids in solvent S us. the pK, values in water. The least-squares slope of the line of Figure 1 is 2.2. I n a recent paper by Bruice and Turner4 the rate constants for the nucleophilic attack of acylate ions upon substituted phenyl acetates in water (k11~0)and solvent S (ks) were reported. A plot of log ICs us. log ~ I I * Owas found to be linear and also of slope 2.2. That the slope for the ionization of phenols is identical with the slope (1) Material submitted by R. G. in partial fulfillment of the requirement for the M.S. degree in Chemistry, University of California a t Santa Barbara. (2) C. H. Langford and T. R. Stengle, J . Amer. Chem. Soc., 91,4015 (1969). (3) A. J. Parker, Chem. Reu., 69, 1 (1969). (4) T. C. Bruice and A . Turner, J . Amer. Chem. Soc., 92, 3422 (1970). (5) C. D. Ritchie and R. E. Uschold, ibid., 89, 1721 (1967). (6) H.S. Harned and L. D. Fallon, ibid., 61,2374 (1939).