The aldol condensation of acetone with acetophenone

The aldol condensation of acetone with acetophenone J . PETERGUTHRIE' AND XIAO-PING WANC Department of Cheinistry, Univer.ri!y of Western Ontario, London, Ont., Canada N6A 567

Received March 8, 1991' J . PETERGUTHRIE and XIAO-PINGW A N CCan. . J . Chem. 70, 1055 (1992). The kinetics and equilibria involved in the aldol condensation of acetone, acting as carbon acid, and acetophenone have been studied in aqueous alkaline solution. The enone isolated is the E isomer. The reactions are all first order in hydroxide, with rate and equilibrium constants (defined for E-enone as initial compound) of: k12= (5.55 f 0.17) X M-I s-I, k2, = (8.00 2 0.40) X M-I s -I, K2, = (1.44 + 0.55) (ketol to E-enone), KZ4= 0.160 ? 0.033 0.26) X M-I (acetone plus acetophenone to ketol), k2, = 0.180 i (ketol to Z-enone), Kj2 = (1.89 0.005 M-' s-', k32 = (3.41 i 0.49) X M-' s-'. There is an equilibration of the two enones in base that is faster M-I sC1;k4, = (2.81 i 0.61) X lo-' M-' s-'; K14= 0.112 -t than hydration to the ketol: k14= (3.14 f 0.84) X 0.019. To analyze the behavior of the enone:ketol equililbrium system in acid we simultaneously fitted analytical data for all three species (E-enone, Z-enone, and ketol) to a kinetic model, so that the rate constants were determined by the best fit to all of the data for an experiment.

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 203.206.176.157 on 11/04/15 For personal use only.

*

J . PETER GUTHRIE et XIAO-PINGW A N GCan. . J . Chem. 70, 1055 (1992). OpCrant en solution aqueuse alcaline, on a CtudiC les cinktiques et les tquilibres irnpliquCs dans la condensation aldolique de I'acCtone, agissant comme acide carbont, et 1'acCtophCnone. L'Cnone isolCe est l'isomkre E. Les rCactions sont toutes du premier ordre en ion hydroxyde alors que les constantes de vitesse et d'Cquilibre (definies pour I'Cnone E comme produit initial) sont : k12= ( 5 3 5 ? 0,17) X M-' s-I; KZ1= 1,44 M-' s-I, k2, = (8,00 t 0,40) X 0,55 (cCtol de 1'Cnone E ) , K2, = 0,160 0,033 (cCtol de l'enone Z ) , K3' = (1,89 ? 0,26) X lo-' M-' (acetone plus acCtophCnone conduisant au cCtol), kz3 = 0,180 + 0,005 M-I s-I et k32 = (3,41 -t 0,49) x lo-' M-' s-I. En milieu basique, il existe un Cquilibre qui est plus rapide que l'hydratation en cCtol : k14= (3,14 0,84) X M-I s-I, kJI M-I s-I; K14= 0,112 f 0,019. Dans le but d'analyser le comportement de 1'Cquilibre cCto= (2,81 + 0,61) X Cnolique en milieu acide, on a ajustC les donnCes analytiques obtenues pour les trois espkces (Cnones E et Z et cCtol) B un modkle cinetique; on a ainsi pu determiner les constantes de vitesse qui correspondent le mieux avec l'ensemble des donnCes pour une experience. [Traduit par la rkdaction]

*

I

Introduction We recently reported on the aldol condensation of acetophenone, acting as carbon nucleophile, with acetone (1). This was, in part, background information for ongoing studies of elimination reactions catalyzed by steroidal imidazoles (2, 3). In the course of this investigation we became aware that there was a concurrent condensation in the opposite sense, with acetone, acting as carbon nucleophile, attacking acetophenone. The reactions involved are described in Scheme 1. The kinetic analysis is from the perspective of the retroaldol reaction, since this is the direction that gives the more accurate rate constants. This reaction turned out to be somewhat more difficult to sort out, but we have now completed the kinetic analysis, and wish to report it in turn. This provides another example of a fully analyzed aldol condensation, to add to those we (1, 4-10) and others (1 1-14) have previously reported. To complete the analysis we had to develop computer programs allowing us to fit simultaneously multiple sets of data described by equations with common parameters. Results Preparation When we set out to make 3-hydroxy-3-methyl-1-phenyl1-butanone we found that directed aldol condensation using the preformed lithium enolate of acetophenone gave mixtures consisting mainly of the desired product with small

'Author to whom correspondence may be addressed. 2Revision received October 23, 199 1.

*

amounts of 4-hydroxy-4-phenyl-2-pentanone. These were essentially impossible to separate on a preparative scale, and we had to use an alternative synthesis (1). 4-Hydroxy-4phenyl-2-pentanone could be prepared in pure form by reaction of phenyl magnesium bromide with 2,4-pentanedione (15), and 4-phenyl-3-penten-2-one could be obtained by dehydration of the ketol. The enone so obtained was the E isomer, based on comparison of its 'H NMR spectrum with literature values for the two isomers. The diagnostic signals are those for the CH,-C=C signal, which appears at 2.096 (16a), Z isomer, or 2.356 (16a), 2.476 (16b), E isomer, and the vinyl hydrogen, for which the signals are at 6.946 (16a), Z isomer, or 6.216 (16a), 6.346 (16b), E isomer, all in CCl,. We observed signals at 2.456 and 6.356 in CCl,. Overall kinetics We studied the retroaldol kinetics of both E-enone and ketol in dilute alkaline solution, and discovered that the two compounds reacted on markedly different time scales, with ap arent second-order rate constants of 0.180 2 0.005 M-I s P Pfor disappeaiance of the ketol and (9.99 2 0.25) x M-' s-' for disappearance of the E-enone; see Figs. 1 and 2. The data for the E-enone also showed a small amplitude kinetic phase with an apparent second-order rate constant of (3.12 & 0.68) x 10-' M-' s-'; because of the small amplitude of this phase it is not well defined by the data, but appears to be real. We attribute this behavior to isomerization of the enones prior to conversion to the ketol and retroaldol cleavage. This isomerization will be considered in more detail later. The experimental results are found in Table 1. The observed behavior is interpreted in terms of rapid, essentially complete, breakdown of ketol to acetone and


CAN. J. CHEM. VOL. 70,1992

188888.

20Lwil.

388888.

t

4!a3088.

588888.

88%88&

known from the analysis of the kinetics of equilibration in acid, we M-I s-' may calculate k,, and k,,, as (8.00 + 0.40) x and (7.93 2 2.22) x M-' s-I respectively. We have now completed the kinetic analysis to the point of having all of the rate and equilibrium constants for the ratelimiting processes for the two principal stages of the aldol condensations leading to the E- and 2-enones. The accuracy of the A, values is low, because these parameters are determined by a small amplitude process at short times, but we can extract approximate estimates of kI4or k,,. A2/[OH-] has a weighted average value of (3.12 0.68) x M-I s-I. Since K,, is known from both kinetics in base, K,, = 0.159 + 0.033, and kinetics in acid, K,, = 0.1 12 + 0.019 (values which agree within their estimated standard deviations, we will use the more accurate value from kinetics in acid, and can calculate that k,, = (3.14 ? 0.84) x M-I s-I. lop5M-' s-' and k,, = (2.81 + 0.61) X

*

This result is at fust surprising, but simply means that if two rapidly interconverting species can both go to product, there are in effect two reaction channels, and reaction is twice as fast as if there were only one. Since the observed value for A3 = (9.99 + 0.25) X M-I s-I, and K,, = 0.112 2 0.019, we may now calculate k,, = (9.99 + 0.25) X M-' s-' (1 0.112 k 0.019)/2 = (5.55 2 0.17) x M-I s-I. It follows that k4, = k,,/K,, = (4.96 2 0.93) x M-I S - I .

+

Discussion In the course of the work described in this paper we have developed computer programs written in FORTRAN that


1064 I

I

I

I

2-

I

~

--

6i

~I

II

I

I

--

Keto 1

I

I

I

I

E-Enone

-----


a

(D

6i

8

-

5

6i 4

I I

I I

I I

I I

I I

I I

I I

I I

[I

Q

I

I I

I I

I I

I I

I I

I l

I I

I I

~

-

Z-Enone

----

-.

--

-

-

-

+ +

. I

I

I I

t

I I

I I

I I

I I

I I

--

-

-

--

-

--

--

-

--

--

-

a, 6i-

1: Y

I

-

1:

,m m

r

~

-.

--

N

6i

~

+

-PI

I

----

6i-

Tt

a

-.

, + + I

I

100.

I

300.

I

I

I

500.

I

I

700.

I

I

I

900.

I

I

I

300. 500.

I

700.

I

I

I

900.

I

I

300.

I

I

500.

I

I

700.

900.

time

FIG. 5 . Equilibration in acid; concentration of the ketol, E-enone, and Z-enone as a function of time, starting with ketol or E-enone. The lines are calculated using the parameters given in Table 2 .

TABLE3. Equilibrium in the aldol condensation of acetone and acetophenone by H P L C ~ Species analyzed

[AcPh]

[Acetone]

[OH]

a,

a2

103a3

a4

106a5

1.86 k 0.43 1.86*0.43

-238 k 17 -21 i z 17

( a ) Experiment starting with no enone present:

[E-Enone] [Z-Enone]

0.01 0.01

1.086 1.086

0.887 0.887

221 29

k

Equilibrium constant = [Enone2],/[A~Ph]~[Acetone]~ = 3.09

( b )Experiment starting with [E-enone] [E-Enone] 0.01 1.086 [Z-Enone] 0.01 1.086

10

*9 X

27 k 30 -16*29

6.38 6.38

k k

0.43 0.43

6.38 6.38

k

0.43

M-'

M:

= 1.0 X

0.887 0.887

246 k 15 5k12

Equilibrium constant = [Enone2If /[A~Ph]~[Acetone]~ = 2.36 Average value of the equilibrium constant = (2.73 0.37) X

*

X

221 2 24 -143k22

*

1.86 0.43 1.86k0.43

677 390

k

21

* 19

* 0.43

M-' M-'

"In aqueous solution at 2S°C, followed by HPLC analysis monitoring absorbance at 280 nm. T o obtain convenient peak heights the volume injected and the absorbance scales were adjusted; for ease of comparison the peak heights have been corrected to a common basis, even though this corresponds to peaks over a meter high in some cases. Peak heights vs. time data were fitted to peak height = a , a, (exp(-a,t) + a, exp(-a$), with a , , a ~ , a 4specific to the particular data set, while a, and a, were the same for all data sets, and all four data sets were fitted simultaneously.

+

allow us to fit several sets of data simultaneously with some parameters common to all sets and some specific to a particular set. Two variations on this theme were used. In the first we fitted the exact integrated rate law, using theoreti-

cal expressions for both the exponential terms (apparent rate constants) and the preexponential terms. In favorable cases, including the present example of fitting equilibration data for the two isomeric enones and the ketol in acidic solution, this


GUTHRIE AND WANG

!

1065

t i m e (ks)

FIG.6. Equilibration in base: the concentration of E-enone as a function of time (as measured by HPLC analysis), starting with either no enone present or 1.0 X lod4M enone present. In both cases the solution contained 1.086 M acetone and 0.010 M acetophenone, and 0.887 M NaOH.

allows direct calculation of the relevant rate constants. In the second, used where the data were less precise and we knew there were experimental problems precluding detailed interpretation, we simply fitted all sets of peak heights vs. time data to the same form of integrated rate law, with common apparent rate constants and individual preexponential terms. This allows satisfactory fitting to all the data, even when an individual data set would not give a satisfactory fit because the least-squares program could not converge. It is of course essential to be sure that all data sets used do in fact belong to a single reaction system, so that it is legitimate to apply this procedure. If this precondition is not met, then fitting in this way will provide a new way to mislead oneself. With the ready availability of HPLC apparatus, it will become increasingly common to use this technique for re) action kinetics. The advantage is that one follows individ/ ual species (assuming satisfactory analytical conditions with good resolution); the disadvantage is that the precision is less than can be obtained by UV spectrophotometric kinetics. By treating data for all species involved in a reaction, and for sets of reaction that must have the same values for pseudo first-order rate constants, we can obtain the best possible precision, as well as having the greater accuracy inherent in having a more specific analytical technique.

The second mechanism, shown in Scheme 3, assumes that only the original isomer of the enone undergoes hydration to ketol, while the second isomer does not. This seems peculiar at first glance, but can be rationalized in terms of the conformations of the cations obtained by protonation at the carbonyl oxygen. The E isomer can be planar, as is required for maximum conjugative stabilization of the positive charge, although this incurs some destabilizing steric interactions. The Z isomer cannot be planar, and may plausibly be assumed to have the phenyl ring nearly orthogonal to the plane of the protonated enone. This avoids steric interactions of the ortho hydrogens with the cis acetyl group, but imposes interference by these ortho hydrogens to the attack of water on the C=C double bond. Thus acid-catalyzed interconversion of the geometrical isomers by way of the protonated enone can lead to addition of water only to the protonated E enone. Further logical implications of the above mechanisms are that acid-catalyzed rotation about the C3-C, bond occurs because the bond order is decreased in the cation, even though the phenyl group can provide little or no conjugative stabilization for the Z isomer of the cation. This seems to require that rotation from the Z cation to the E cation entails only a modest energy barrier because loss of conjugation to


1066


J

fast

the OH can be compensated by gain of conjugation to the rotation occurs. phenyl, which rotates as C,-C, It is striking that in base, where the form of the overall kinetics followed by UV demands that the ketol undergo retroaldol much faster than dehydration to enone, equilibration of the two enones nonetheless occurs at rates competitive with retroaldol cleavage. Since our principal interest was in determining rate constants for the aldol condensation, we have not settled the mechanism of this isomerization. If it occurs by addition-elimination, then it must follow that the enolate, which is the immediate product of hydroxide addition to either enone, can rotate and expel hydroxide faster than it undergoes reprotonation. If isomerization occurs by deprotonation-reprotonation, then this must be fast relative to ketol formation. The kinetics of equilibration were fitted to a double exponential integrated rate law. In the present case, since we did not have a sample of the Z-enone, and could not apply mass balance constraints, we could not determine a calibration constant for the Z-enone. Instead of the more elaborate

procedure used for the equilibrations in acid, we simply imposed the requirement that the same apparent rate constants, for both exponentials, apply to all of the species whose reaction could be followed for a given hydroxide concentration. We could then fit four sets of peak height vs. time data (for the two isomeric enones, starting from solutions with identical hydroxide, acetone, and acetophenone concentrations, but initially zero or finite E-enone content) and fit simultaneously to obtain the best values of the apparent rate constants. The values so obtained are found in Table 3. Naturally, in these experiments we also observed the formation of the alternative aldol condensation product, 3methyl- 1-phenyl-2-buten- 1-one. The data in Table 4 show that there is a linear free energy relationship between the equilibrium constant for aldol addition reactions with acetone as carbon nucleophile and the equilibrium constant for the addition of water. This is illustrated in Fig. 8. Although the correlation is not exact, it is enough to suggest that there is a useful way to obtain approximate predictions. Perhaps unsurprisingly there is much less pattern discernible in the equilibrium constants for the dehydration step. Clearly a phenyl group on the @-positionof the enone favors dehydration relative to a hydrogen, as does a methyl group, and a phenyl is much more effective (6 log K = +2.75 for phenyl, +0.52 for methyl). A methyl forced to be cis to the acetyl group in the enone is detrimental, although the is not uniform, 6 log K being - 1.22 when effect on log K7_, the first @ substituent is phenyl, and -0.35 when the first @ substituent is methyl. As significant steric effects develop, the magnitude of the effect becomes specific to the particular compound, and is not susceptible to back-of-the-envelope calculations. Molecular mechanics calculations may in time be the best way to predict these effects. When we compare the two modes of reaction possible for the crossed aldol of acetone and acetophenone we find that the reaction studied in the present work is considerably more favorable than might have been expected. As we have just shown, there is generally a correlation between the equilibrium constant for an aldol addition and that for the addition of water. The two nucleophiles, acetone and acetophenone, have very similar inherent tendencies to add to carbonyls, as measured by their y values (4, 22), and the equilibrium constants for hydration differ by a factor of 208, favoring acetone. Thus it is disconcerting to find that the equilibrium constants for the two alternative modes of crossed aldol addition differ only by a factor of 2, favoring addition of acetophenone to acetone. Because we had expected a factor more like 200 to disfavor addition of acetone to acetophenone, we d were surprised when the synthetic e ~ ~ e r i m e n t s s h o w ethat the two reactions were competitive. After the encouraging signs of predictable behavior that we had found, it is perhaps salutary to be reminded that the aldol reaction still has surprises.

Experimental Materials 4-Hydroxy-4-phenyl-2-pentanone Brornobenzene ( 1 1.8 g, 75 rnrnol) in 30 rnL dry THF was dropped into magnesium turnings (1.8 g, 75 rnrnol) in 20 rnL dry

THF under nitrogen in a 100 rnL round-bottomed flask with a reflux condenser over a period of 1 h and left to react at room ternperature for an additional 0.5 h. A solution of 2.5 g (25 rnrnol) of


FIG.7. Reaction coordinate diagram for equilibration of ketol and enones, using the mechanism of Scheme 4. There is an ambiguity because we do not know whether the activation barrier for conversion of the enolate of the ketol to the E- or Z-enone is higher; consequently we have drawn them both the same, equal to the observed value. As before the en01 content of the ketol was assumed the same as for acetone, and a pK, of 10 was assumed for this enol. Proton transfers to and from oxygen were assumed to be diffusion controlled in the thermodynamically favored direction.

2,4-pentanedione in 20 mL dry THF was added dropwise to the stirred solution of the Grignard reagent. After all the diketone solution was added, the reaction mixture was stirred overnight (13 h), and poured gradually into 100 mL of a stirred, ice-cold, saturated solution of ammonium chloride in water. The organic layer was separated, dried over anhydrous magnesium sulfate, filtered, and evaporated to dryness, yielding 14 g. The crude product, 1 g, from above, was chromatographed on a column of silica G , 40 g, eluted with a linear gradient (0-30%) of ethyl acetate in hexane. The eluates were pooled according to TLC analysis and evaporated separately to dryness. The fraction that had the same retention time as authentic PhC(OH)(CH3)CH2COCH3was pumped for a half hour to remove residual solvents, yielding 0.7 g. The proton NMR indicated that it was mainly PhC(OH)(CH3)CH,COCH3, with some PhC(CH3)=CHCOCH3 and other impurities.

Pure 4-hydroxy-4-phenyl-2-pentanone was obtained by crystallizing relatively purified ketol at 0°C from 1 : 1 diethyl ether/hexane and washing with cold hexane. UV (MeOH), A,,:258 nm, E,, 187, €280 5 1. NMR (CDC13, 200 MHz), 6: 1.51 (s, 3H), 2.06 (s, 3H), 2.83, 3.19 (dd, J = 17.1 Hz, 2H), 4.54 (s, IH), 7.22-7.45 (m, 5H); lit. (23) (CDCl,, 6 0 MHz) 6: 8.66-7.30 (m, C6H5)r 4.73 (s, OH), 3.43, 2.43 (dd, J = 18 Hz, 2H), 2.16 (s, CH3-C-0), 1.60 (s, CH,). Exact Mass calcd.: 178.0994; found: 178.0993 0.00009.

4-Phenyl-3-penten-2-one The crude product (mainly PhC(OH)(CH,)CH2COCH3), 0.5 g, from above was added to a crucible containing a small amount of silica gel G. Three drops of 6 N HCI and a small amount of methanol were added. The entire mixture of solid and liquid was mixed

1068


TABLE4. Equilibrium constants for aldol addition and hydration of various carbonyl compounds" Carbonyl acceptor CH20 CH-jCHO PhCHO CH3COCH3 PhC0CH-j

log KHIO

log K32

log Kzl

log K3,

1.6

6.67 1.59 1.07 -1.41 -2.72

- 1.37 -0.85 1.38 -1.20 0.16

5.30 0.74 2.45 -2.61 -2.56

- 1.72

-2.83 -4.60 -6.92


"All in aqueous solution at 25°C. Equilibrium constants for addition of water taken from ref. 4; equilibrium constants for aldol addition, aldol dehydration, and aldol condensation from refs. 4 and 25 and this work.

by HPLC analysis were carried out in suitable vessels, thermostatted in a water bath, and samples were taken at intervals. For reactions in base, the samples were quenched with 1 M H3P0,; for reactions in acid the samples were quenched with 1 M Na2HP04. At first quenched samples were stored and analyzed together, but it was discovered that slow equilibration of the enones was occurring in the quenched samples. For the experiments in Table 2, analysis was carried out immediately after quenching. For the experiments in Table 3, analysis was delayed, and thus no detailed interpretation of enone concentrations can be meaningful except at the end of the equilibration. The solvents used were different mixtures of methanol and water flowing at 2.0 mL/min through a C18 Radial Pak column. Stock solutions were prepared by weighing purified compounds into volumetric flasks and diluting with solvents (spectrophotometric grade methanol or acetonitrile) to the mark. Sodium hydroxide solutions were titrated against standardized hydrochloride solution using two indicators (phenolphthalein and methyl red) and correcting for carbonate. The extinction coefficients of various compounds were measured by adding measured amounts (Hamilton syringe) of a stock solution in methanol to 25 mL of a solvent (either methanol, ethanol, water, or aqueous potassium chloride solution).

Acknowledgments We thank the Natural Sciences and Engineering Council of Canada for financial support of this work.

FIG. 8. Correlation between log K,I,,I for the aldol addition of acetone as nucleophile reacting with formaldehyde, acetaldehyde, benzaldehyde, acetone, and acetophenone, and log KH20for the addition of water. Both equilibrium constants are expressed with units of M-I. The correlation line, fitted by least squares, is log KddOl 4.42 2 0.94 + 1.09 2 0.15 log KH20.

completely and heated on a steam bath for 1 h. The reaction mixture was washed through filter paper, using solvent ethyl acetate. The filtrate was evaporated to dryness and loaded on a column of silica G , 40 g, eluted with a linear gradient (0-30%) of ethyl acetate in hexane. The eluates were evaporated separately to dryness according to TLC analysis. Pure 4-phenyl-3-penten-2-one was obtained by crystallizing relatively purified enone at O°C from 1 : 1 diethyl ether/hexane and washing with cold hexane. UV( MeOH), A,,:280 nm, E,, 18 300; lit. (24) (EtOH), A,,: 280 nm, E,, 19 000. NMR (CDCl,, 200 MHz), 6: 2.30 (s, 3H), 2.54 (d, J = 1.3 Hz, 3H), 6.5 1 (m, J = 1.2 Hz, lH), 7.26-7.50 (m, 5H); lit. (16) (CCl,, 60 MHz), 6: 2.06 (s, 3H), 2.35 (d, J = 1 Hz, 3H), 6.21 (m, IH), 7.10 (br s, 5H). Exact Mass calcd.: 160.0888; found: 160.091 1 2 0.0030. Methods Kinetics The procedure for kinetics was as described (16). Reactions followed by UV spectroscopy were carried out in 10-cm cells holding 25 mL of solution at 25OC. Reactions that were followed

1. J. P. Guthrie and X.-P. Wang. Can. J. Chem. 69, 339 (1991). 2. J. P. Guthrie and S. O'Leary. Can. J. Chem. 53, 2150 (1975). 3. J. P. Guthrie, J. Cossar, and B. A. Dawson. Can. J. Chem. 64, 2456 (1986). 4. J. P. Guthrie. Can. J. Chem. 56, 962 (1978). 5. J. P. Guthrie. Can. J. Chem. 52, 2037 (1974). 6. J. P. Guthrie. Can. J. Chem. 59, 45 (1981). 7. J. P. Guthrie and B. A. Dawson. Can. J. Chem. 61, 171 (1983). 8. J. P. Guthrie, J. Cossar, P. A . Cullimore, N. M. Kamkar, and K. F. Taylor. Can. J . Chem. 61, 2621 (1983). 9. J. P. Guthrie, K. J . Cooper, J. Cossar, B. A. Dawson, and K. F. Taylor. Can. J. Chem. 62, 1441 (1984). 10. J. P. Guthrie, J. Cossar, and K. F. Taylor. Can. J . Chem. 62, 1958 (1984). 11. A. de Blic and P. Maroni. Bull. Soc. Chim. Fr. 512 (1975). 12. D. S . Noyce and W. L. Reed. J. Am. Chem. Soc. 81, 624 (1959). 13. C. C. French. J. Am. Chem. Soc. 51, 3215 (1929). 14. K. Koelichen. Z. Phys. Chem. 33, 129 (1900). 15. T . J. Leitereg and D. J. Cram. J. Am. Chem. Soc. 90, 4019 (1968). 16. ( a ) J. Klein and R. Levene. J. Chem. Soc. Perkin Trans. 2, 1971 (1973); ( b ) K. Takai, M . Sato, K. Oshina, and H. Nozaki. Bull. Chem. Soc. Jpn. 57, 108 (1984). 17. D. S . Noyce, W . A. Pryor, and P. A. King. J. Am. Chem. SOC.81, 5423 (1959). 18. J. L. Jensen and D. J. Carre. J. Org. Chem. 39,2103 (1974). 19. Y. K . Kim and J. D. Hatfield. J. Chem. Eng. Data, 30, 149 (1985). 20. A. Ikai and C. Tanford. J. Mol. Biol. 73, 145 (1973). 2 1. P. R. Bevington. Data reduction and error analysis for the physical sciences. McGraw-Hill, New York. 1969. 22. E. G. Sander and W. P. Jencks. J. Am. Chem. Soc. 90, 6154 (1968). 23. I. Stahl, B. Schramm, R. Manske, and J. Gosselck. Liebigs Ann. Chem. 1158 (1982). 24. J. P. Phillips and F. C. Nachod (Editors). Organic electronic spectral data. Vol. IV. Wiley, New York. 1963. p. 340. 25. J. P. Guthrie. J. Am. Chem. Soc. 113, 7249 (1991).