Structure-reactivity relationships and intrinsic reaction barriers for nucleophile additions to a quinone methide: A strongly resonance-stabilized carbocation

Article abstract of DOI:10.1021/ja9937526

Second-order rate constants k(Nu) (M^-1 s^-1) were determined for addition of a wide range of nucleophiles to the simple quinone methide 4- [bis(trifluoromethyl)methylene]cyclohexa-2,5-dienone (1) to give the nucleophile adduct 1-Nu in

Full text of DOI:10.1021/ja9937526

1664
J. Am. Chem. Soc. 2000, 122, 1664-1674
Structure-Reactivity Relationships and Intrinsic Reaction Barriers for
Nucleophile Additions to a Quinone Methide: A Strongly
Resonance-Stabilized Carbocation
John P. Richard,*,†,1a Maria M. Toteva,† and Juan Crugeiras‡,1b
Contribution from the Department of Chemistry, UniVersity at Buffalo, SUNY,
Buffalo New York 14260-3000, and the Departamento de Qu´ımica F´ısica, Facultad de Qu´ımica,
UniVersidad de Santiago, 15706 Santiago de Compostela, Spain
ReceiVed October 19, 1999
Abstract: Second-order rate constants kNu (M-1 s-1) were determined for addition of a wide range of
nucleophiles to the simple quinone methide 4-[bis(trifluoromethyl)methylene]cyclohexa-2,5-dienone (1) to give
the nucleophile adduct 1-Nu in water. Equilibrium constants were determined for the overall addition of HBr
and HI to 1 to give H-1-Nu, and the data were used to calculate equilibrium constants for the addition of Br-
and I- to 1, and to estimate equilibrium constants for the addition of Cl- and AcO-. The values of log kNu
show a linear correlation with the Ritchie nucleophilicity parameter N+ with a slope s ) 0.92 ( 0.10 that is
essentially the same as the electrophile-independent value of 1.0 for highly resonance-stabilized carbocations.
Marcus intrinsic barriers Λ of 12.4, 13.9, 15.4, and 19.8 kcal/mol are reported for the addition of I-, Br-, Cl-,
and AcO- to 1, respectively. The thermodynamic barriers ∆G° and intrinsic barriers Λ for addition of Br-,
Cl-, and AcO- to 1 are 8.4 ( 1.0 and 5.2 ( 0.2 kcal/mol larger, respectively, than the corresponding barriers
for addition of these nucleophiles to the triphenylmethyl carbocation. It is concluded that, by the criterion of
its chemical reactivity, 1 behaves as a highly resonance-stabilized carbocation. Values of N+ ) 4.0, 2.2, 1.2
and 0.60, respectively, are reported for I-, Br-, Cl-, and AcO-, which do not form stable adducts to Ritchie
electrophiles. The slope of 2.0 (r ) 0.98) for the linear correlation between Ritchie (N+) and Swain-Scott (n)
nucleophilicity parameters shows that there is substantially greater bonding between the nucleophile and carbon
at the transition state for nucleophile addition to sp2-hybridized carbon than for addition to sp3-hybridized
carbon. Azide ion and nucleophiles with a nonbonding electron pair(s) at atoms adjacent to the nucleophilic
site (R-effect nucleophiles) exhibit significant positive deviations from this correlation.
Introduction
structure of quinone methides invites comparisons with better
characterized organic functional groups. For example, simple
quinone methides may be thought of as very highly stabilized
p-alkoxy substituted benzylic carbocations (Scheme 1, A) or
as examples of cyclic Michael acceptors (B). However, a
definitive classification is difficult because of the relative lack
of data on the chemical and physical properties of quinone
methides.
Quinone methides are a class of organic compounds with
considerable importance in chemistry and biology.2-14 The
† University at Buffalo, SUNY.
‡ Universidad de Santiago.
(1) (a) Tel: (716) 645 6800 ext 2194. Fax: (716) 645 6963. Email:
jrichard@chem.buffalo.edu. (b) Current Address: Departamento de Qu´ımica
F´ısica, Facultad de Qu´ımica, Universidad de Santiago, 15706 Santiago de
Compostela, Spain.
We have reported methods for generation of the simple
quinone methide 4-[bis(trifluoromethyl)methylene]cyclohexa-
2,5-dienone (1),15 and the results of a study of the uncatalyzed
and specific-acid-catalyzed addition of halide ions to 1 in a
solvent of 50/50 (v/v) trifluoroethanol/water.16 We have now
extended this work and report here the results of a study of the
reaction of 1 with a large number of nucleophilic reagents in
water. These studies were undertaken for the following reasons:
(1) We have treated15,16 quinone methides as members of the
class of strongly resonance-stabilized benzyl carbocations
(Scheme 1, A), and others have also found this classification
to be useful.7,17 A principal goal of this work was to compare
the reactivity of quinone methides and carbocations, to determine
(2) Angle, S. R.; Arnaiz, D. O.; Boyce, J. P.; Frutos, R. P.; Louie, M.
S.; Mattson-Arnaiz, H. L.; Rainier, J. D.; Turnbull, K. D.; Yang, W. J.
Org. Chem. 1994, 59, 6322-6337.
(3) Angle, S. R.; Rainier, J. D.; Woytowicz, C. J. Org. Chem. 1997, 62,
5884-5892.
(4) Boruah, R. C.; Skibo, E. B. J. Org. Chem. 1995, 60, 2232-2243.
(5) Dorrestijn, E.; Pugin, R.; Nogales, M. V. C.; Mulder, P. J. Org. Chem.
1997, 62, 4804-4810.
(6) Gaudiano, G.; Frigerio, M.; Sangsurasak, C.; Bravo, P.; Koch, T. H.
J. Am. Chem. Soc. 1992, 114, 5546-5553.
(7) McCracken, P. G.; Bolton, J. L.; Thatcher, G. R. J. J. Org. Chem.
1997, 62, 1820-1825.
(8) Myers, J. K.; Cohen, J. D.; Widlanski, T. S. J. Am. Chem. Soc. 1995,
117, 11049-11054.
(9) Paul, G. C.; Gajewski, J. J. J. Org. Chem. 1993, 58, 5060-5062.
(10) Pande, P.; Shearer, J.; Yang, J.; Greenberg, W.; Rokita, S. E. J.
Am. Chem. Soc. 1999, 121, 6773-6779.
(11) Tomasz, M.; Das, A.; Tang, K. S.; Ford, M. G. J.; Minnock, A.;
Musser, S. M.; Waring, M. J. J. Am. Chem. Soc. 1998, 120, 11581-11593.
(12) Trahanovsky, W. S.; Chou, C.-H.; Fischer, D. R.; Gerstein, B. C.
J. Am. Chem. Soc. 1988, 110, 6579-6581.
(14) Fischer, M.; Wan, P. J. Am. Chem. Soc. 1998, 120, 2680-2681.
(15) Richard, J. P.; Amyes, T. L.; Bei, L.; Stubblefield, V. J. Am. Chem.
Soc. 1990, 112, 9513-9519.
(16) Richard, J. P. J. Am. Chem. Soc. 1991, 113, 4588-4595.
(17) Hulbert, P. B.; Glover, P. L. Biochem. Biophys. Res. Commun. 1983,
117, 129-134.
(13) Wan, P.; Barker, B.; Diao, L.; Fisher, M.; Shi, Y.; Yang, C. Can.
J. Chem. 1996, 74, 465-475.
10.1021/ja9937526 CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/10/2000

Nucleophile Additions to a Quinone Methide
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1665
Experimental Section
Scheme 1
Materials. Inorganic salts and organic chemicals were reagent grade
from commercial sources and were used without further purification
unless noted otherwise. Amine nucleophiles were purchased in the free
base form, except for hydroxylamine, which was purchased as (NH3-
OH)2SO4 and converted to the basic form using sodium hydroxide.
HPLC-grade methanol was used for all HPLC analyses. Water for
kinetic studies and HPLC analyses was distilled and passed through a
Milli-Q water purification system.
4-MeOC6H4C(CF3)2OTs (Me-1-OTs) was prepared by a published
procedure.25 The quinone methide 1 was generated for use in kinetic
and product studies by making a large (25-150-fold) dilution of Me-
1-OTs (0.15 M - 0.50 M in acetonitrile) into 2/1 (v/v) trifluoroethanol/
water to give a final concentration of Me-1-OTs that ranged from 1 to
20 mM, depending upon the experiment. A 30% yield of 1 forms in
this solvent during a 10 min reaction time.15 These solutions of 1 in
2/1 (v/v) trifluoroethanol/water were used in kinetic experiments within
4 h of preparation, during which time there was less than 20%
conversion of 1 to the corresponding solvent adducts. The concentration
of 1 was determined spectrophotometrically at 283 nm using an
Scheme 2
extinction coefficient of 36 000 M-1 cm-1 15
.
Stock solutions of sodium sulfite and hydrogen peroxide were
prepared daily, and their concentrations were determined immediately
after their use in a kinetic experiment by titration with starch iodine
and KMnO4,26a respectively. The concentrations of thiols were deter-
mined before and after each kinetic run using 5,5′-dithiobis(2-
nitrobenzoic acid).26b In cases where the thiol concentration was
observed to decrease (e7%) during the course of the kinetic run, the
concentration during the kinetic run was obtained by averaging the
initial and final thiol concentrations.
whether, by the criterion of its chemical reactivity, 1 behaves
as a member of the larger class of strongly resonance-stabilized
carbocations.
(2) There have been few measurements of rate and equilib-
rium constants for addition of halide ions to weakly reactive
electrophiles, because these anions do not form stable adducts
to such electrophiles. The adducts of halide ions to 1 (1-Nu,
Scheme 2) are likewise unstable; however, rate and equilibrium
data for their formation can be obtained by coupling nucleophile
addition to protonation of the phenoxide oxygen of 1-Nu to
give H-1-Nu (Scheme 2).16 We report here rate and equilibrium
data for the addition of halide and acetate ions to 1, which allows
for extension to these nucleophiles of the well-known Ritchie
N+ relationship for carbocation-nucleophile addition reac-
tions.18,19
pH Measurements and Buffers. The pH measurements were
performed at 25 °C using a combination electrode in which the salt
bridge filling solution was replaced by 5 M lithium trichloroacetate.27
The pH was determined at the end of reactions monitored by
conventional UV spectroscopy. It was not possible to measure the final
pH of reaction mixtures in the stopped-flow experiments. In these cases,
the pH was determined for control solutions that were prepared to be
identical to the corresponding solutions from the stopped-flow experi-
ments.
In studies of the reactions of ethylamine (pH 10-11), glycylglycine,
(pH 7.6-8.6), and trifluoroethylamine (pH 5.2-6.2) with 1, the
nucleophile also served to buffer the reaction solution. Constant pH
was maintained in studies of other nucleophiles by use of the appropriate
buffer (0.01-0.05 M): dichloroacetate, pH e 2.5; chloroacetate, pH
2.7-3.5; methoxyacetate, pH 3-4.2; acetate, pH 4.2-5.3; 2-(N-
morpholino)ethanesulfonic acid (MES), pH 6-7; 3-(N-morpholino)-
propanesulfonic acid (MOPS), pH 7.5-8; N-tris(hydroxymethyl)methyl-
3-aminopropanesulfonic acid (TAPS), pH 8-9; 3-(cyclohexylamino)-
1-propanesulfonic acid (CAPS), pH 9.5-10.5.
Product Studies. Aqueous solutions (I ) 1.0, NaClO4) at 25 °C
containing 1 and H-1-Br or H-1-I at chemical equilibrium were
prepared by adding measured amounts of HClO4 and sodium bromide
or sodium iodide to solutions that contain a fixed concentration (ca.
10-5 M) of 1 and monitoring the approach to an equilibrium mixture
of 1 and H-1-Nu at 283 nm.16 The ratio [H-1-Nu]eq/[1]eq at equilibrium
was determined from eq 1, where Ao is the absorbance of a solution
that contains only 1, Aeq is the absorbance observed for solutions that
contain an equilibrium mixture of 1 and H-1-Nu, and Amin is the limiting
minimum absorbance observed for a solution that contains, essentially,
only H-1-Nu, determined for the reaction of 1 in the presence of high
concentrations of H+ and Nu-.
(3) We are interested in understanding how Marcus intrinsic
barriers to chemical reactions, and the changes in these barriers
with changing reactant structure, influence structure-reactivity
effects on carbocation-nucleophile addition reactions.20-24
However, reports in the chemical literature of the determination
of these barriers are rare.24 The kinetic and thermodynamic
parameters for addition of halide and acetate ions to 1 allow
for determination of the Marcus intrinsic reaction barriers, and
an analysis of the relationship between these intrinsic barriers
and the effect of changing electrophile reactivity on electrophile
selectivity toward addition of nucleophilic reagents.18,19
(4) A thorough characterization of the reactivity of quinone
methides toward nucleophile addition is one part of the broader
description of these compounds required to explain the relation-
ship between their chemical reactivity and their roles in biology.
(18) Ritchie, C. D. Can. J. Chem. 1986, 64, 2239-2250.
(19) Ritchie, C. D. Acc. Chem. Res. 1972, 5, 348-354.
(20) Richard, J. P.; Amyes, T. L.; Vontor, T. J. Am. Chem. Soc. 1992,
114, 5626-5634.
(21) Richard, J. P.; Amyes, T. L.; Jagannadham, V.; Lee, Y.-G.; Rice,
D. J. J. Am. Chem. Soc. 1995, 117, 5198-5205.
(22) Richard, J. P. Tetrahedron 1995, 51, 1535-1573.
(23) Richard, J. P.; Amyes, T. L.; Williams, K. B. Pure Appl. Chem.
1998, 70, 2007-2014.
(24) Richard, J. P.; Williams, K. B.; Amyes, T. L. J. Am. Chem. Soc.
1999, 121, 8403-8404.
(25) Allen, A. D.; Kanagasabapathy, V. M.; Tidwell, T. T. J. Am. Chem.
Soc. 1986, 108, 3470-3474.
(26) (a) Kolthoff, I. M.; Belcher, R.; Stenger, V. A.; Matsuma, G. In
Volumetric Analysis; Interscience Publishers Ltd.: London, 1957; Vol. III.
(b) Ellman, G. L. Arch. Biochem. Biophys. 1959, 82, 70-77.
(27) Herschlag, D.; Jencks, W. P. J. Am. Chem. Soc. 1986, 108, 7938-
46. Washabaugh, M. W.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 674-
683.

1666 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
Richard et al.
RNH2 at I ) 1.0 (NaClO4), with correction of the values of the
concentration ratio [RNH3+]/[RNH2] for the concentration of hydroxide
ion calculated from the pH. Values of the concentration ratio [HNu]/
[Nu] for a number of other nucleophiles were determined spectropho-
tometrically according to eq 4, where Aobsd is the absorbance of the
test solution and ANu and AHNu are the absorbance values of the solution
when essentially all of the nucleophile is present in the basic and acidic
forms, respectively. The extent of protonation of propanethiolate ion
[H-1-Nu]eq
[1]eq
Ao - Aeq
Aeq - Amin
)
(1)
HPLC product studies were carried out in water at 25 °C and I )
1.0 (NaClO4). Reactions were initiated by making a 100-fold dilution
of a solution of 1 to give a final concentration of (0.2-2.0) × 10-4
M
in an aqueous solution that contains the same volume (2.6%) of
trifluoroethanol that was present in the stopped-flow experiments.
HPLC Analyses. The products of the reactions of acetate, azide,
and propanethiolate anions with 1 were separated by HPLC using
procedures described previously,16,28,29 except that peak detection was
by a Waters 996 diode array detector. Substrate and products were
detected at 268 nm, which is λmax for H-1-OH.15 4-MeOC6H4C(CF3)2-
OH (Me-1-OH), which is formed in ca. 70% yield during the initial
generation of 1 from Me-1-OTs,15 was used as an internal standard to
correct for small variations in the injection volume. Normalized HPLC
peak areas were reproducible to better than (10%. Ratios of product
yields, [H-1-OAc]/[H-1-OH], were calculated using eq 2, where A1/
A2 and ꢀ2/ꢀ1 ) 1.0 are the ratios of the HPLC peak areas and the molar
extinction coefficients at 268 nm, respectively, for the two products.
The value of ꢀ2/ꢀ1 )1.0 is assumed, because it has been shown in earlier
work that the extinction coefficients of 1-(4-methoxyphenyl)-2,2,2-
trifluoroethyl alcohol and 1-(4-methoxyphenyl)-2,2,2-trifluoroethyl
acetates are identical at λmax for the alcohol.20
(0.1-0.2 mM, ꢀRS > ꢀRSH) and peroxide ion (0.02 M, ꢀHOO > ꢀHOOH
)
was determined at 238 and 260 nm, respectively, which are λmax for
the nucleophilic anions. The extent of protonation of azide ion (20 mM,
ꢀHN3 > ꢀN3) was determined at λmax for HN3 (260 nm).
Values of (pKa)ArOH for the phenolic oxygen of H-1-Nu were
estimated using eq 5, where (pKa)PhOH ) 10.0 for phenol,30 F ) 2.2 is
the Hammett reaction constant for ionization of substituted phenols in
water,31 and σeff is the effective Hammett substituent constant for
p-C(CF3)2Nu estimated using eq 6. Equation 6 was derived assuming
additivity of the polar Hammett substituent constants31 for the groups
attached to the benzylic carbon (σp for CF3 and σn for Nu), with an
attenuation factor of 0.40 for the carbon that separates these groups
from the aromatic ring.32 This gives σeff ) 0.54 for C(CF3)2Br and σeff
) 0.55 for C(CF3)2I, which were substituted into eq 5 to give (pKa)ArOH
) 8.8 for both H-1-Br and H-1-I.
(pKa)ArOH ) (pKa)PhOH - Fσeff
σeff ) 0.40(2σCF3 + σNu
(5)
(6)
[P]1/[P]2 ) (A1/A2)(ꢀ2/ꢀ1)
(2)
)
Kinetic Studies. All kinetic studies were carried out at 25 °C and
I ) 1.0 (NaClO4). Reactions in the presence of nucleophilic reagents
employed at least a 10-fold excess of nucleophile over 1. Reactions of
1 with halftimes of less than 5 s were monitored by following the
decrease in absorbance at 283 nm using the SX17.MV stopped-flow
device from Applied Photophysics. The aqueous solution and a solution
of 1 in 2/1 (v/v) trifluoroethanol/water were mixed in a ratio of 25:1
to give a final aqueous reaction mixture containing 2.6% trifluoroethanol
and 1 × 10-5 M 1. First-order rate constants, kobsd, were obtained from
the fit of the absorbance data to a single-exponential function and were
reproducible to (5%. The slower reactions of 1 were monitored using
a conventional UV spectrophotometer and were initiated by making a
100-fold dilution of 1 to give a final concentration of 1 × 10-5 M 1 in
an aqueous solution that contains the same volume (2.6%) of trifluo-
roethanol that was present in the stopped-flow experiments. First-order
rate constants, kobsd, were calculated from the slopes of linear semi-
logarithmic plots of reaction progress against time and were reproduc-
ible to (5%.
Results
By contrast with our previous studies of nucleophile addition
to 1 in 50/50 (v/v) trifluoroethanol/water,16 an aqueous solvent
was used in this work, to avoid protonation of basic nucleophiles
by trifluoroethanol. A value of ks ) 6.4 × 10-4 s-1 for the
reaction of 1 with solvent water at 25 °C and I ) 1.0 (NaClO4)
was determined by following the decrease in absorbance due
to 1 at 283 nm. The products of nucleophilic addition of azide
and propanethiolate ions to 1 were detected by HPLC analysis,
and it was shown for these nucleophiles that conversion of 1 to
the nucleophile adduct is essentially quantitative when [Nu-]
g 1 mM.
First-order rate constants kobsd (s-1) for the disappearance of
1 in the presence of increasing concentrations of nucleophiles
in water at 25 °C and I ) 1.0 (NaClO4) were determined by
monitoring the decrease in absorbance of 1 at 283 nm, either
by conventional or stopped-flow spectrophotometry. Observed
[NuH]
pH ) pKa - log
(3)
(4)
second-order rate constants (kNu obsd
(M-1 s-1) for the reactions
)
[Nu]
of CH3CH2CH2S-, HOO-, SO32-, N3-, ethylamine, trifluoro-
ethylamine, glycylglycine, and hydroxylamine with 1 were
determined as the slopes of linear plots of kobsd (s-1) against
the total concentration of the acidic and basic forms of the
nucleophilic reagent and are reported in Table S1 of the
Supporting Information.
ANu - Aobsd
[NuH]
[Nu]
)
(
)
Aobsd - ANuH
The second-order rate constants (kNu
)
(M-1 s-1) for the reaction
obsd
of nucleophiles with 1 were determined as the least-squares slopes of
linear plots of kobsd against the total concentration of the nucleophile.
Figure 1 shows pH-rate profiles of the observed second-
order rate constants (kNu obsd for the reaction of a variety of
neutral amines and anionic nucleophiles with 1. These correla-
tions have slopes of 1.0 at pH , (pKa)NuH (Scheme 3) and show
a downward break, centered at (pKa)NuH, to a slope of zero in
cases where it was possible to obtain values of (kNu obsd at pH
. (pKa)NuH. Table 1 reports the following: (a) values of kNu
The nonlinear least-squares fit to eq 7 of the pH-rate profile for (kNu obsd
)
)
2-
for the reaction of SO3 (see Results) was obtained using SigmaPlot
from Jandel Scientific.
Determination and Estimation of Acidity Constants. Values of
(pKa)HNu for the conjugate acids of nucleophilic reagents at 25 °C and
I ) 1.0 (NaClO4) were determined from the solution pH and the
concentration ratios [NuH]/[Nu] according to eq 3, using data at 20-
80% protonation of the nucleophile. Except for ethylamine, buffered
amine solutions of known [RNH3+]/[RNH2] were prepared by mixing
solutions containing known concentrations of perchloric acid and RNH2.
The pKa of ethylamine was determined by titration of a solution of
)
(30) Jencks, W. P.; Regenstein, J. In Handbook of Biochemistry and
Molecular Biology, Physical and Chemical Data, 3rd ed.; Fasman, G. D.,
Ed.; CRC Press: Cleveland, OH, 1976; Vol. 1; pp 305-351.
(31) Hine, J. In Structural Effects on Equilibria in Organic Chemistry;
Wiley: New York, 1975.
(28) Richard, J. P.; Rothenberg, M. E.; Jencks, W. P. J. Am. Chem. Soc.
1984, 106, 1361-1372.
(29) Richard, J. P. J. Am. Chem. Soc. 1989, 111, 1455-1465.
(32) Calculated as the average of the ratios of values of σp for p-CH2I,
p-CH2Br, and p-CH2Cl substituents and the corresponding values of σn for
p-I, p-Br, and p-Cl substituents.31

Nucleophile Additions to a Quinone Methide
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1667
We have restricted our analysis of kinetic and product data
for the reactions of acetate ion to concentrations of 1 (e2.0 ×
10-5 M) where the reaction is cleanly first order in [1] and H-1-
OH and H-1-OAc are the only detectable products. Observed
second-order rate constants, (kAcO obsd
(M-1 s-1), for the
)
reactions of 1 in acetate buffers were determined as the slopes
of linear plots of kobsd against the total concentration of acetate
buffer. A value of (kNu + kB) ) 0.049 M-1 s-1 (Scheme 4) was
calculated as the average of the values (kAcO)obsd/fAcO at five
different values of fAcO between 0.2 and 0.9, where fAcO is the
fraction of buffer present as acetate anion. The relative contribu-
tions of the reaction of acetate ion to give the nucleophile (kNu
,
Scheme 4) and the solvent adduct (kB) were determined from
the effect of increasing concentrations of acetate ion on the
product ratio [H-1-OH]/[H-1-OAc] for reactions at [1] ) 1.6
× 10-5 M (Figure 2). The line through the data in Figure 2
shows the linear fit of the product data to eq 8, and the
y-intercept gives kB/kNu ) 0.029 (Scheme 4). This corresponds
to a limiting yield of ca. 3% of the solvent adduct H-1-OH for
the reaction of 1 in the presence of high concentrations of acetate
Figure 1. pH-rate profiles of the second-order rate constants (kNu
)
obsd
(M-1 s-1) for the addition of neutral and anionic nucleophiles to the
quinone methide 1 in water at 25 °C and I ) 1.0 (NaClO4). The lines
through the data show the fits to eq 7 (see text).
(M-1 s-1, Scheme 3) for the reaction of azide ion, sulfite dianion,
and glycylglycine with 1, determined as the average of the
ion. The value of kB/kNu ) 0.029 was combined with (kNu
+
kB) ) 0.049 M-1 s-1 to give kNu ) 0.048 M-1 s-1 for
nucleophilic addition of acetate ion to 1 and kB ) 0.001 M-1
s-1 for reaction of acetate ion as a general base catalyst of the
addition of solvent water.
values of (kNu obsd at high pH where >95% of the reagent is
)
present in the basic form (Figure 1); (b) values of kNu for the
reaction of CH3CH2CH2S-, HOO-, ethylamine, trifluoroethy-
lamine, and hydroxylamine with 1, determined as the average
of the values of (kNu)obsd/fNu, where fNu is fraction of the
kB
ks
[H-1-OH]
[H-1-OAc]
nucleophilic reagent present in the basic form. Unless noted
otherwise, the uncertainty in the values of kNu, estimated from
the range of the values of kNu determined at different pH values,
is ( 10%.
)
+
(8)
kNu[AcO-]
kNu
[H-1-Nu]eq
[1]eq
) [H+][Nu-]KOH
(9)
add
kNu(Ka)NuH
(kNu)obsd ) kNu + kHNu[H+]
(10)
(kNu)obsd
)
(7)
(Ka)NuH + aH
The reaction of 1 in acidic solutions containing Br- or I- to
give H-1-Br or H-1-I (Scheme 5) in water at 25 °C and I )
1.0 (NaClO4) was monitored at 283 nm. These reactions proceed
essentially quantitatively to the nucleophile adduct when
[H+][Nu-] is large, but at smaller values of [H+][Nu-] equi-
librium mixtures of 1 and H-1-Br or H-1-I are obtained. Figure
3 shows the linear dependence of [H-1-Nu]eq/[1]eq at chemical
The solid lines through the data in Figure 1 were calculated
using eq 7 derived for Scheme 3 and the values of kNu (M-1
s-1) and the pKa of the nucleophile given in Table 1. The pKa’s
were determined by direct titration of the nucleophile under our
experimental conditions (see the Experimental Section) except
for (pKa)HNu ) 6.7 for HSO3-, which was obtained from the
nonlinear least-squares fit of the experimental data to eq 7. This
is in good agreement with pKa ) 6.6 for HSO3- determined at
I ) 1.0 (KCl).33 The good fits of the experimental data to eq 7
derived for the mechanism in Scheme 3 show that (a) there are
no detectable reactions of the protonated nucleophiles with 1
in the pH range of these experiments and (b) there is no
detectable reaction of the sulfate ion that was present in the
solutions of hydroxylamine prepared from (NH3OH)2SO4.
The disappearance of 1 in the presence of CH3CO2- followed
good first-order kinetics when [1] e 5 × 10-5 M, but deviations
were observed at [1] > 5 × 10-5 M. Similarly, only H-1-OH
and H-1-OAc were observed by HPLC analysis of the products
of the reaction of 1 with acetate ion when [1] e 2.0 × 10-5 M,
but an additional product with a relatively long HPLC retention
time was detected for this reaction at [1] > 2.0 × 10-5 M. The
fractional yield of this product increased as the pH was
increased. These data are consistent with the conclusion that
the additional product forms by addition of the phenoxide anion
1-OAc to a second molecule of 1 to give a dimeric product
H-1-1-OAc. However, this product was not further character-
ized.
equilibrium on [H+][Nu-], according to eq 9. The slopes of
OH
these plots give the equilibrium constants K ) 1.5 × 104
add
and 1.5 × 105 M-2 for the addition of HBr and HI, respectively,
to 1 to give H-1-Nu.
First-order rate constants, kobsd (s-1), for the essentially
complete reaction of 1 with halide ions in the presence of
hydronium ion to give H-1-Nu were determined in water at 25
°C and I ) 1.0 (NaClO4), and observed second-order rate
constants (kNu)obsd (M-1 s-1) for these reactions were determined
as the slopes of plots of kobsd against [Nu-]. Figure 4 shows the
small increases in (kNu obsd with increasing concentrations of
)
HClO4 for the reactions of iodide, bromide, and chloride ions
with 1. The data were fit to eq 10, derived for the mechanism
in Scheme 5; the intercepts of these plots give kNu (M-1 s-1
,
Table 1) for direct addition of the halide ion to 1, and the slopes
give kHI ) 110 M-2 s-1, kHBr ) 7.4 M-2 s-1, and kHCl ) 1.2
M-2 s-1 as the third-order rate constants for the specific-acid-
catalyzed reactions of iodide, bromide, and chloride ions with
1, respectively.
Discussion
The observed second-order rate constants for nucleophile
addition to 1 (Figure 1) are directly proportional to the fraction
(33) Young, P. R.; Jencks, W. P. J. Am. Chem. Soc. 1977, 99, 8238-
8248.

1668 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
Richard et al.
Scheme 3
Table 1. Rate and Equilibrium Constants for the Uncatalyzed
Addition of Nucleophiles to the Quinone Methide 1 in Watera
nd
kNu (M-1 s-1
4.6 × 106
)
b
c
e
nucleophile
(pKa)NuH
N+
CH3CH2CH2S-
10.3
11.8
8.93
8.52
8.01
7.54
5.28
3.45
4.69
5.05
4.0i
6.95f
5.43 f
5.67
3.92
5.01
4.08
4.59
4.40
4.93
4.02
2.99
HOO-
(2.9 ( 0.6) × 105
1.0 × 105
(4.1 ( 0.5) × 103
SO3
N3
6.7 g
4.6
2-
-
5.5 × 105
CH3CH2NH2
CF3CH2NH2
GlyGly
HONH2
I-
11.0
5.7
8.1
6.2
8.7
430
330
68
1.4
-12h
Br-
-10h
-8h
2.2i
Cl-
1.2i
0.16
AcO-
4.7
0.60i 2.76
4.8 × 10-2
a At 25 °C and I ) 1.0 (NaClO4). b pKa of the conjugate acid of the
nucleophile, determined under the experimental conditions as described
in the Experimental Section, unless noted otherwise. c Ritchie N+ value
taken from ref 18, unless noted otherwise. d Swain-Scott n value for
reactions in water. Values of n for amines were taken from ref 58.
Values of n for anions were taken from ref 56, unless noted otherwise.
e Second-order rate constant for addition of the nucleophile to 1 to give
1-Nu (Scheme 5). Unless indicated, the uncertainty in the values of
Nu is (10%. f Calculated from the value of n for reactions in methanol
Figure 2. Dependence of the product ratio [H-1-OH]/[H-1-OAc] for
partitioning of 1 between addition of solvent and acetate ion on the
concentration of acetate ion in water at 25 °C and I ) 1.0 (NaClO4).
The line through the data shows the fit to eq 8 (see text).
k
(ref 57) and the linear relationship nwater ) 0.69nMeOH + 0.046 for
aliphatic substitution reactions in methanol and in water. g Determined
from the nonlinear least-squares fit to eq 7 of the experimental data
for the addition of sulfite to 1 (Figure 1). h Data from ref 52. i Value
of N+ obtained by extrapolation of the linear correlation (N+ ) (log
kNu + 1.87)/0.92) between log kNu for nucleophile addition to 1 and
N+ shown in Figure 5A (open circles).
Scheme 5
Scheme 4
of the nucleophile present in the reactive basic form. In no case
is there any sign of a break to pH-independent values of (kNu)obsd
at pH , (pKa)NuH that would provide evidence for either addition
of the protonated nucleophile or specific acid catalysis of
addition of the nucleophilic anion. We conclude that direct
nucleophile addition to 1 to form the phenoxide anion 1-Nu,
which is at chemical equilibrium with H-1-Nu (Scheme 5), is
the only significant kinetic pathway for reaction of these
nucleophilic reagents in the pH range shown in Figure 1.
The observation that at [Nu-] ) 1 mM the reactions of azide
and propanethiolate ions with 1 give essentially quantitative
yields of the respective nucleophile adducts H-1-N3 and H-1-
SCH2CH2CH3 shows that the second-order rate constants for
reaction of these anions as nucleophiles are much larger than
those for their reaction as base catalysts of the addition of solvent
water. Our data predict a limiting yield of ca. 3% H-1-OAc for
reaction of 1 in the presence of very high concentrations of
acetate ion (Figure 3). Therefore, nucleophilic addition of acetate
ion to 1 is about 30-fold faster than general base catalysis of
the addition of solvent water by this anion (Scheme 4). There
is no detectable catalysis by acetate ion of the addition of solvent
water to the more reactive 1-(4-methoxyphenyl)ethyl (ks ) 5
× 107 s-1 28
)
and 1-(4-dimethylaminophenyl)ethyl carbocations
)
in 50/50 (v/v) trifluoroethanol/water,34b so that
(ks ) 40 s-1 34a
catalysis by carboxylate ions of the addition of solvent to
benzylic carbocations becomes increasingly important with
decreasing electrophile reactivity.34b
The second-order rate constants kNu (M-1 s-1) for the
reactions of propanethiolate, peroxide ion, sulfite dianion, azide
ion, ethylamine, trifluoroethylamine, glycylglycine, and hy-
droxylamine with 1 are at least 200-fold larger than that for
reaction of acetate ion (Table 1). These differences are due
almost entirely to the greater reactivity of these species as
nucleophiles than as general bases, because general base
catalysis of the addition of water to electrophilic carbon by
strongly nucleophilic reagents is negligible in cases where the
nucleophile adduct is stable.35
(34) (a) Cozens, F. L.; Mathivanan, N.; McClelland, R. A.; Steenken, S.
J. Chem. Soc., Perkin Trans. 2 1992, 2083-2090. (b) Ta-Shma, R.; Jencks,
W. P. J. Am. Chem. Soc. 1986, 108, 8040-8050.

Nucleophile Additions to a Quinone Methide
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1669
ing rate constants for these reactions in 50/50 (v/v) trifluoro-
ethanol/water.16 The larger values of kNu in water than in
trifluoroethanol/water are probably due to stabilization of the
halide ion by hydrogen bonding to the relatively acidic solvent
trifluoroethanol. The smaller effect of this change in solvent
on the values of kHNu is consistent with a compensating greater
activity of the proton in trifluoroethanol/water than in water.
OH
Table 2 gives the values of K (M-2), the overall equilib-
add
rium constant for the addition of HBr and HI to 1 to give H-1-
Nu, determined directly as described in the Experimental Section
(Scheme 5). These experimental equilibrium constants and the
values of kNu (M-1 s-1, Table 1) for halide ions were used to
calculate the following rate and equilibrium constants and
intrinsic reaction barriers reported in Table 2:
(1) Values of Kadd (M-1) for addition of bromide and iodide
ions to 1 to give 1-Br and 1-I were calculated using the
relationship Kadd ) KOH(Ka)ArOH (Scheme 5), where (Ka)ArOH is
Figure 3. Dependence of the ratio of the concentrations of H-1-Nu
and 1 at chemical equilibrium on the product of concentrations
[H+][Nu-] in water at 25 °C and I ) 1.0 (NaClO4): b, data for addition
of HI; 1, data for addition of HBr.
add
the estimated acidity constant for ionization of H-1-Nu to give
1-Nu (see the Experimental Section).
(2) Values of ksolv (s-1) for the expulsion of halide ions from
1-Br and 1-I to give 1 were calculated from the values of Kadd
using the relationship ksolv ) kNu/Kadd (Scheme 5).
(3) Values of ksolv (s-1) for the expulsion of chloride and
acetate ions from 1-Cl and 1-OAc to give 1 were estimated
from the value of ksolv for 1-Br and (ksolv)Br/(ksolv Cl
)
) 14 for
the ratio of rate constants for the DN + AN (SN1) solvolysis of
1-phenylethyl bromide and chloride36 and (ksolv)Br/(ksolv AcO
)
)
107 for the corresponding ratio for solvolysis of 1-phenylethyl
bromide and acetate.36,37
(4) Values of Kadd (M-1) for the addition of chloride and
acetate ions to 1 to give 1-Cl and 1-OAc were calculated from
the values of ksolv using the relationship Kadd ) kNu/ksolv
.
2
1.36 log Kadd
4Λ
1
1.36
log kNu
)
17.44 - Λ 1 -
(11)
(
)
[
]
(5) The Marcus intrinsic barriers Λ (kcal/mol) for the
thermoneutral addition of chloride, bromide, iodide and acetate
ions to 1 to give the respective nucleophile adducts 1-Nu were
calculated from the rate constants kNu and equilibrium constants
Kadd using the Marcus equation (eq 11, derived at 298 K).
Table 2 gives the corresponding values of Kadd (M-1), kNu
(M-1 s-1), and ksolv (s-1) for addition of chloride, bromide, and
acetate ions to the triphenylmethyl carbocation (H3-2) that were
taken from earlier work by McClelland and co-workers.38 The
rate constants kNu and equilibrium constants Kadd were substi-
tuted into eq 11 to give the intrinsic barriers Λ (kcal/mol) for
addition of these nucleophiles to H3-2 (Table 2).
Structure and Reactivity of Quinone Methides. While 1
is formally neutral, the large stabilization associated with
formation of a 6π aromatic system favors a significant contribu-
tion of the zwitterionic valence bond structure of the 4-O--
substituted benzyl carbocation. The following experimental
observations show that, by the criterion of its chemical reactivity
toward nucleophilic reagents, 1 is a member of the class of
Figure 4. Dependence of the second-order rate constant (kNu)
obsd (M-1
s-1) for the addition of halide ions to 1 on the concentration of
hydronium ion in water at 25 °C and I ) 1.0 (NaClO4): b,
acid-catalyzed addition of iodide ion; 9, acid-catalyzed addition of
bromide ion; 1, acid-catalyzed addition of chloride ion.
Reactions of Halide Ions. Two pathways were observed for
the reaction of halide ions with 1. The dominant reaction is the
direct addition of the halide ion to 1 (kNu, Scheme 5) to give
the phenoxide anion 1-Nu, which is in rapid equilibrium with
H-1-Nu. The specific-acid-catalyzed addition of halide ions to
give H-1-Nu directly is also observed (Figure 4) in strongly
acidic solutions (kHNu, Scheme 5).
The values of kNu (M-1 s-1, Table 1) and kHNu (M-2 s-1
,
Results) for the uncatalyzed and specific-acid-catalyzed addition
of halide ions to 1 in water at 25 °C (I ) 1.0, NaClO4) are
6-8-fold and <3-fold larger, respectively, than the correspond-
(36) Noyce, D. S.; Virgilio, J. A. J. Org. Chem. 1972, 37, 2643-2647.
(37) We have chosen to cite the extensive set of data for solvolysis
reactions of ring-substituted 1-phenylethyl derivatives. However, these
relative leaving group abilities are not strongly substrate dependent so that
the uncertainty in these ratios will not affect the interpretation of these
(35) General base catalysis by tertiary amines has been reported for
addition of water to X, Y, Z-2 in cases where the quaternary ammonium
ion adduct is unstable.39,70,71 However, this reaction is much slower than
direct addition of amine nucleophiles. For example, values of kB ) 0.036
M-1 s-1 and kNu ) 7.4 M-1 s-1, respectively, were determined for
quinuclidine-catalyzed addition of water71 and direct addition of n-
propylamine72 to malachite green, (Me2N)2, H-2.
results. For example, values of (ksolv)Br/(ksolv)Cl ) 27 and (ksolv)Br/(ksolv)
AcO
) 108 have been estimated for solvolysis reactions of triphenylmethyl
derivatives.38
(38) McClelland, R. A.; Banait, N.; Steenken, S. J. Am. Chem. Soc. 1986,
108, 7023-7027.

1670 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
Richard et al.
Table 2. Rate and Equilibrium Constants and Intrinsic Reaction Barriers for the Addition of Nucleophiles to the Quinone Methide 1 and the
Triphenylmethyl Carbocation H3-2 in in Water at 25 °C
OH
nucleophile and
electrophile
kNu
K
Kadd
ksolv
d
∆G°
∆∆G°
Λ
∆Λ
anion solvation
add
(M-1 s-1
)
(M-2
)
(M-1
)
(s-1
)
(kcal/mol)e (kcal/mol)f (kcal/mol)g (kcal/mol)h energy (kcal/mol)i
a
b
c
Cl- + 1
0.16
∼4 × 10-5
70k
2.4 × 10-5
6k
2.4 × 10-4
∼8
6 × 107 k
∼4 × 103 j
6.0
15.4
8.5
7.4
5.4
5.0
75
Cl- + H3-2
Br- + 1
2.2 × 106 k
1.4
3 × 104 k
-2.5
10.0
13.9
1.5 × 104
1.5 × 105
6 × 104 j
6.3
70
61
75
Br- + H3-2
I- + 1
AcO- + 1
5 × 106 k
68
8 × 105 k
3 × 105 j
∼0.006j
7 × 10-3 k
-1.1
4.9
8.9
12.4
19.8
0.048
-1.2
9.4
5.2
AcO- + H3-2
4 × 105 k
-10.6
14.6
a Second-order rate constant for uncatalyzed addition of nucleophile to 1 (Scheme 5, data from Table 1) or H3-2 to give the corresponding
nucleophile adduct. b Overall equilibrium constants for addition of HNu to 1 to give H-1-Nu, determined as described in the Experimental Section
(Figure 3 and Scheme 5). c Equilibrium constant for addition of the anionic nucleophile to 1 to give 1-Nu (Scheme 5), see text. d First-order rate
constant for breakdown of the nucleophile adduct by expulsion of the anionic leaving group (Scheme 5). e Gibbs free energy change for addition
of the nucleophile to 1 or H3-2. f Difference in the Gibbs free energy changes for addition of the nucleophile to 1 and H3-2. g Marcus intrinsic
barrier for addition of the nucleophile to 1 or H3-2 calculated from the values of kNu (M-1 s-1) and Kadd given in this table using eq 11. h Difference
in the intrinsic barriers Λ for addition of the nucleophile to 1 or H3-2. i The free energy change for transfer of the anion from the gas phase into
aqueous solution. Data from ref 52. j Calculated from the values of Kadd and kNu using the relationship ksolv ) kNu/Kadd (Scheme 5). k Data from ref
38.
highly resonance-stabilized carbocations, which includes ring-
substituted triarylmethyl carbocations (X,Y,Z-2) and the aryl
tropylium ions (3).
(1) The absolute rate constant for addition of solvent water
to 1 at 25 °C, ks ) 6.4 × 10-4 s-1, lies within the range of rate
Figure 5. (A) Correlation of the second-order rate constants kNu (M-1
constants determined for addition of solvent to ring-substituted
s-1) for the addition of nucleophiles to 1 in water at 25 °C and I ) 1.0
triarylmethyl carbocations. For example, values of ks ) 4.6 ×
(NaClO4) with Ritchie N+ values (data from Table 1). The solid symbols
10-3, 2.1 × 10-4, and 2.0 × 10-5 s-1 at 25 °C have been
are the experimental data that were used to obtain the correlation line
determined for addition of solvent water to Me2N,MeO,H-2,
of slope 0.92 ( 0.10; the open symbols are the data for nucleophiles
(Me2N)2,H-2, and (Me2N)3-2, respectively.39
for which values of N+ have not previously been determined and which
(2) Rate constants for the addition of nucleophiles to
resonance-stabilized carbocations such as 2 and 3 in water show
a good fit to the Ritchie N+ equation (eq 12), where N+ is a
parameter characteristic of nucleophile reactivity.18,19,40 Figure
5A shows that there is a good linear logarithmic correlation
between the rate constants kNu (M-1 s-1) for addition of
nucleophiles to 1 and the Ritchie nucleophilicity parameter N+.
The slope of this correlation is s ) 0.92 ( 0.10,41 which is not
significantly different from the value of 1.0 determined for
addition of nucleophiles to the resonance-stabilized carbocations
2 and 3.
are assumed to follow this correlation. (B) Correlation of the values of
N+ for nucleophile addition to trivalent carbon electrophiles with the
Swain-Scott n values for bimolecular nucleophilic substitution at
aliphatic carbon (data from Table 1).
(Λ(1) - Λ(H3-2) ) 5.2 ( 0.2 kcal/mol) for the addition of
chloride, bromide, and acetate ions to 1 and H3-2 (Table 2).
The ca. 8 kcal/mol more unfavorable change in ∆G° for
nucleophile addition to 1 than to H3-2 shows that resonance
electron donation to the benzylic carbon of 1 is much more
stabilizing than the corresponding electron donation from the
three phenyl rings at H3-2. The ca. 5 kcal/mol larger intrinsic
barrier for nucleophile addition to 1 is consistent with the notion
that the effect of this larger carbocation stabilization by
resonance is to make carbocation-nucleophile addition more
difficult in both a thermodynamic and a kinetic sense.22,23,42 The
almost constant relatiVe values of Λ and ∆G° for the addition
of different nucleophiles to 1 and H3-2 is striking and requires
that Variations in nucleophile structure bring about the same
change in both the transition state and product stability for
nucleophilic addition to these two electrophiles. This provides
log kNu ) N+ + constant
(12)
(3) There are nearly constant differences between the values
of ∆G° (∆G°(1) - ∆G°(H3-2) ) 8.4 ( 1.0 kcal/mol) and of Λ
(39) Ritchie, C. D.; Wright, D. J.; Huang, D. S.; Kamego, A. A. J. Am.
Chem. Soc. 1975, 97, 1163-1170.
(40) Ritchie, C. D.; Wright, D. J. J. Am. Chem. Soc. 1971, 93, 6574-
6577.
(41) The value ks ) 1.3 × 10-2 s-1 for addition of solvent to 1 calculated
using a value of N+ ) 0 for solvent is significantly larger than the observed
value of 6.4 × 10-4 s-1. This deviation is another example of a poor
correlation of the rate constant for addition of solvent with the Ritchie N+
equation.18
(42) (a) Bernasconi, C. F.; Killion, R. B., Jr. J. Am. Chem. Soc. 1988,
110, 7506-7512. (b) Bernasconi, C. F.; Ketner, R. J.; Chen, X.; Rappoport,
Z. J. Am. Chem. Soc. 1998, 120, 7461-7468.

Nucleophile Additions to a Quinone Methide
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1671
good evidence for the development of similar electrophile-
nucleophile bonding interactions at these transition states and
products.
Structure-Reactivity Relationships and Intrinsic Barriers.
There is no simple relationship between the rate and equilibrium
constants for the addition of nucleophiles to 1 or for addition
to the triphenylmethyl carbocation H3-2 (Table 2). For example,
nucleophilic addition of acetate ion to 1 is 6.1 kcal/mol more
favorable than addition of iodide ion, but the rate constant for
addition of iodide ion is 1400-fold larger than that for acetate
ion (Table 2). Such breakdowns in rate-equilibrium relation-
ships are a direct consequence of differences in the intrinsic
barriers for the two nucleophile addition reactions. Thus, the
larger rate constants for addition of iodide ion to 1 than for the
thermodynamically more favorable addition of acetate ion
reflects the 7.4 kcal/mol smaller intrinsic barrier for the iodide
ion reaction (Table 2). Similarly, the observation of decreasing
rate constants kI > kBr > kCl for addition of halide ions to 1, for
reactions that are of similar thermodynamic driVing force,
reflects the increase in the intrinsic reaction barrier along the
series ΛI < ΛBr < ΛCl (Table 2). The tendency of soft
polarizable thiolate anions to show smaller intrinsic barriers than
harder nucleophilic amines42a and alkoxide anions42b toward
addition to R-nitrostilbenes and methoxybenzylidene Meldrum’s
acid, respectively, has been noted in earlier work.
Figure 6. Hypothetical reaction coordinate profiles for the addition
of iodide and chloride ions to electrophilic trivalent carbon. (A) Profiles
for thermoneutral reactions which do not specifically consider desol-
vation of the nucleophile prior to its reaction. (B) Profiles for formally
thermoneutral reactions in which unfavorable desolvation of the
nucleophile (∆Gdes and ∆GIdes) precedes covalent bond formation to
Cl
the electrophile and the actual reaction barriers for reaction of the
desolvated nucleophiles (∆GqCl and ∆Gq ) are significantly smaller than
I
those for reaction of the solvated nucleophiles.
to treat separately the following changes in solvation of the
nucleophile that occur for nucleophilic addition reactions:
(a) Cleavage of a hydrogen bond between water and halide
ion to free an electron pair to react with an electrophile, which
may make a significant contribution to the observed barrier to
the nucleophile addition reactions.47-50 This represents “work”
that needs to be done on reactants before bond formation can
occur. Figure 6 shows that the relative intrinsic barriers for
nucleophile addition may change dramatically depending upon
whether the “work” done in nucleophile desolvation is included
in the overall change in free energy for the reaction. Different
barriers to preequilibrium desolvation for the formally thermo-
neutral nucleophile additions (Figure 6A) would result in
differences in the chemical driVing force for reactions of the
“partly desolvated” nucleophile (Figure 6B). Now, the true
intrinsic barriers calculated for thermoneutral reactions of these
“partly desolvated” halide ions would be smaller than the
intrinsic barriers calculated directly from the experimental rate
and equilibrium data using eq 11, with the decrease being the
largest for the most strongly solvated chloride ion and the
smallest for the most weakly solvated iodide ion. The net result
is to reduce the differences in intrinsic barriers for the reactions
of desolvated halide ions, relative to those calculated using the
experimental data in Table 2.
(b) All other changes in the 60-80 kcal/mol interaction
between solvent and the anionic nucleophile (Table 2) that are
lost upon carbocation-nucleophile bond formation. It is possible
that the intrinsic barriers for gas-phase carbocation-halide ion
addition reactions are similar and that the observed differences
in the barriers for these reactions in water (Table 2) are the
result of the requirement for a larger fractional loss of reactant-
stabilizing halide ion solvation at the transition state relative to
the product-stabilizing bond formation to the nucleophile.
However, the timing between changes in nucleophile solvation
and bond formation to electrophilic carbocations is not well
understood.
These data provide another example of how changes in the
intrinsic kinetic ease of chemical reactions, as measured by
changes in their intrinsic reaction barrier, strongly influence
observed structure-reactivity relationships.22,23 They suggest
that, in developing rationalizations of structure-reactivity
relationships, knowledge of the intrinsic barriers Λ is as
important as knowledge of the Gibbs free energy changes ∆G°,
and that the determination of these intrinsic barriers will often
be essential for a complete characterization of structure-
reactivity relationships.
The principle of nonperfect synchronization provides a useful
framework to explain such differences in intrinsic reaction
barriers.43-45 This principle can be understood by imagining,
for any thermoneutral reaction, that the absence of a thermo-
dynamic driving force reflects the exact balancing of interactions
that tend to destabilize product relative to reactant and interac-
tions that tend to stabilize product relative to reactant. Part or
all of the observed barriers to these reactions may then reflect
the larger fractional expression of product-destabilizing interac-
tions at the transition state, relative to expression of the balancing
product-stabilizing interactions or nonperfect synchronization
in the expression of these interactions. Similarly, changes in
the relative magnitude of the expression of different product-
stabilizing and product-destabilizing interactions will result in
changes in the intrinsic kinetic barrier Λ.43-45
There are at least two interactions that might account for some
or all of the changes in intrinsic barriers with changing
nucleophile that are observed for the addition of halide and
acetate ions to 1 and H3-2:
(1) Solvation of the Reacting Nucleophile. There is a good
correlation between the change in solvation energy for halide
ions and the change in intrinsic barrier for addition of halide
ions to 1 or H3-2 to form a product at which the stabilizing
solvation of the halide ion is largely lost (Table 2).46 It is useful
(2) Bonding Interactions between the Nucleophile and
Electrophile. There is a good correlation between the increasing
(47) McClelland, R. A.; Kanagasabapathy, V. M.; Banait, N. S.; Steenken,
S. J. Am. Chem. Soc. 1992, 114, 1816-1823.
(43) Bernasconi, C. F. AdV. Phys. Org. Chem. 1992, 27, 119-238.
Bernasconi, C. F. Tetrahedron 1985, 41, 3219-3234.
(48) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1984, 106, 1373-
(44) Bernasconi, C. F. Acc. Chem. Res. 1987, 20, 301-308.
(45) Bernasconi, C. F. Acc. Chem. Res. 1992, 25, 9-16.
1383.
(49) Richard, J. P. J. Chem. Soc., Chem. Commun. 1987, 1768-1769.
(50) Berg, U.; Jencks, W. P. J. Am. Chem. Soc. 1991, 113, 6997-7002.
(46) The anion solvation energies were taken from Table 1 in ref 52.

1672 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
Richard et al.
Scheme 6
homolytic bond cleavage to form radical products makes a
significant contribution to the transition state for formal het-
erolytic bond cleavage and that these transition states have a
certain “radical” character.53 This might account for the observed
increase in the stability of transition states for nucleophile
addition to 1 with decreasing nucleophile one electron ionization
potential. These decreasing ionization potentials reflect the
increasing stability of this nucleophile radical, and they may
be manifested in the transition state for nucleophile addition in
proportion to the contribution of the valence-bond configuration
for the nucleophile radical to the overall transition state structure.
Different requirements for nucleophile desolvation, or the
different ease of one-electron transfer, cannot easily account
for the ca. 4.5 kcal/mol smaller intrinsic barrier for the addition
of chloride than of acetate ion to 1 or H3-2, because the solvation
energies (Table 2) and one-electron oxidation potentials52 of
these nucleophiles are similar. We do not know the explanation
for this difference in intrinsic reaction barriers.
The Ritchie N+ Relationship. The values of N+ for halide
ions have not been determined because of the large kinetic and
thermodynamic instability of their adducts to resonance-
stabilized carbocations such as 2 and 3, and only an estimate
of N+ < 2.95 has been reported for acetate ion.54a We have
obtained values of N+ for these nucleophiles (Table 1) by
extrapolation of the linear correlation between log kNu for
nucleophile addition to 1 and N+ shown in Figure 5A (open
circles). The similar values of N+ ) 4.0 determined here for
iodide ion and N+ ) 4.7 for glycylglycine is of particular interest
because of the profoundly different Brønsted basicities of these
nucleophiles and almost certainly reflects the particularly small
intrinsic barrier for reaction of iodide ion.
intrinsic reaction barriers for nucleophile addition to 1 (Table
2) and the decreasing R-deuterium isotope effects kH/kD for
bimolecular nucleophile substitution at N-(methoxymethyl)-N,N-
dimethylanilinium ions (Nu-, kH/kD: I-, 1.18; Br-, 1.16; Cl-,
1.13; AcO-, 1.07).51a These isotope effects are consistent with
an “open” transition state for the bimolecular substitution
reaction of iodide ion, where the hybridization of the R-carbon
is close to that for the free methoxymethyl carbocation, and a
change to a transition state with greater bonding to the R-carbon
on moving along the series I-, Br-, Cl-, AcO-. They provide
evidence that large “soft” polarizable nucleophiles such as iodide
ion can interact from a greater distance to provide electronic
overlap with an electron deficient bonding orbital than can
“hard” more weakly polarizable nucleophiles such as chloride
and acetate ions. A similar trend has been observed for the
inVerse R-deuterium isotope effects on the addition of halide
ions and solvent to diarylmethyl carbocations (Nu-, kH/kD: Br-,
0.96; Cl-, 0.94; H2O, 0.88).51b
There is no general agreement on why essentially constant
nucleophile selectivities with changing electrophile reactivity
are observed for reactions of strongly resonance-stabilized
carbocations that follow the N+ scale (e.g., 2 and 3),19,54b while
sharp changes in selectivity for addition of substituted alkyl
alcohols, alkyl carboxylates, and alkylamines are observed for
changing reactivity of ring-substituted 1-phenylethyl carboca-
tions48 and 1-phenyl-2,2,2-trifluoroethyl carbocations.20 This
difference in reactivity-selectivity behavior requires that the
position of the transition state for nucleophile addition to
strongly resonance-stabilized carbocations remain essentially
constant with changing thermodynamic driving force, but change
relatively sharply with similar changes in thermodynamic driving
force for nucleophile addition to 1-phenylethyl and 1-phenyl-
2,2,2-trifluoroethyl carbocations. We have suggested that the
small shifts in transition state structure for reactions of
resonance-stabilized carbocations result from the large intrinsic
barriers Λ for these reactions (Scheme 7A), while the larger
shifts in transition state structure for reactions of 1-phenylethyl
and 1-phenyl 2,2,2-trifluoroethyl carbocations were proposed
to reflect the relatively small intrinsic reaction barriers (Scheme
7B).20,22
A simple Marcus-type treatment of these data provides a
useful framework for the description of these different structure-
reactivity effects. The first derivative of the Marcus equation
describes the fraction of a particular change in thermodynamic
driving force that is expressed in the reaction transition state,
and the second derivative describes how this fraction changes
with changing driving force (eq 13).20,22,23 A small value of the
second derivative is predicted for reactions with a large intrinsic
barrier that proceed through a transition state of nearly constant
The greater tendency of iodide ion compared with oxygen
nucleophiles to form stabilizing coValent interactions from a
distance may result in a transition state that is “earlier” in the
sense that it occurs at a larger carbon-iodide bond distance.
This may be represented by the free energy profiles shown in
Scheme 6, where cleavage of the bond to iodine at 1-I occurs
along a flatter potential energy surface than that for cleavage
of the bond to oxygen at 1-OAc, due to the (proposed) greater
fractional expression of the bonding interactions as the bond to
the former leaving group is stretched.
There is also a correlation between the decreasing relative
stability of the transition states 4 for the addition of halide ions
along the series I-, Br-, Cl- (ΛI < ΛBr < ΛCl) and the relative
one-electron ionization potentials of halide ions in water, II >
IBr > ICl.52 The bonds between halide ions and 1 or H3-2 have
significant covalent character and involve formal donation of a
single electron from the nucleophile to the reacting carbon.53 It
has been suggested that the valence-bond configuration for
(51) (a) Knier, B. L.; Jencks, W. P. J. Am. Chem. Soc. 1980, 102, 6789-
6798. (b) McClelland, B. Personal communication.
(52) Pearson, R. G. J. Am. Chem. Soc. 1986, 108, 6109-6114.
(53) Pross, A. Acc. Chem. Res. 1985, 18, 212-219. Pross, A. In AdVances
in Physical Organic Chememistry; Academic Press: London, 1985; Vol.
21; pp 99-198.
(54) (a) Ritchie, C. D. J. Am. Chem. Soc. 1975, 97, 1170-1179. (b)
Ritchie, C. D.; Tang, Y. J. Org. Chem. 1986, 51, 3555-3556.

Nucleophile Additions to a Quinone Methide
J. Am. Chem. Soc., Vol. 122, No. 8, 2000 1673
Scheme 7
assumption that the values of Λ for these reactions (Table 2)
remain constant. The slopes of the correlation lines for the
reactions of halide ions are given below each line. These slopes
remain remarkably constant (s ) 0.91 ( 0.02) for a 20 kcal/
mol change in thermodynamic driving force for nucleophile
addition, so that this simple Marcus analysis predicts no more
than small changes in nucleophile selectivities for addition of
halide ions to 1 with changing thermodynamic driving force.
This analysis provides support for the conclusion that the
constant selectivities observed for addition of a variety of
nucleophiles to Ritchie electrophiles (e.g., 2 and 3) are a
consequence of the relatively large intrinsic barriers for these
reactions.20,22 A similar treatment of rate and equilibrium data
for the reactions of the nucleophiles and electrophiles that were
used in the construction of the N+ scale would be required to
provide a more demanding test of this proposal.
The deviations of the rate constant for the reaction of acetate
ion with 1 from the linear correlations for the reactions of halide
ions shown in Figure 7 provide another example of how
variations in intrinsic reaction barriers can lead to breakdowns
in otherwise systematic structure-reactivity correlations.22,23 The
slopes of the correlations that include acetate ion increase from
0.79 to 0.97 (parentheses, Figure 7). However, these changes
do not reflect an apparently anomalous shift to a transition state
with greater covalent bond formation to the nucleophilic reagent
with increasing thermodynamic driving force! Rather, they are
due to the systematic drift in the point for acetate ion (open
symbols) from above to below the correlation line for the
reactions of halide ions with increasing thermodynamic driving
force for nucleophile addition. This drift reflects the different
effects of changes in ∆G° on the activation barrier ∆Gq for the
nucleophilic addition reactions of acetate and halide ions. In
the case of the nearly thermoneutral but intrinsically difficult
(large Λ) reactions of acetate ion, about 50% of the change in
∆G° is expressed at the reaction transition state. By comparison,
sharper changes in ∆Gq with changing thermodynamic driving
force are observed for the reactions of halide ions, because there
is a significantly larger (>50%) expression of the changes in
∆G° in the transition state for these thermodynamically unfavor-
able, but intrinsically easy (small Λ), reactions.
Figure 7. Logarithmic correlations of rate constants kNu (M-1 s-1) for
addition of halide and acetate ions to 1 with the N+ value for the
nucleophile (Table 1). The center correlation is the experimental data
for the addition of halide and ions to 1 taken from Table 1. The upper
and lower two correlations were constructed using hypothetical rate
constants calculated from the Marcus equation (eq 11, derived at 298
K) with incremental 5 kcal/mol decreases and increases, respectively,
in ∆G° for nucleophile addition to 1, with the assumption that the values
of Λ for these reactions (Table 2) remain constant. The numbers below
each line are the slopes of the correlation for halide ions only, and the
slopes of the correlations that include acetate ion are given in
parentheses (see text).
structure (Scheme 7A); an increase in the second derivative
would result from a change to a smaller intrinsic barrier for a
reaction that shows a larger change in transition-state structure
with changing thermodynamic driving force (Scheme 7B).
The Ritchie and Swain-Scott Nucleophilicity Scales. Our
determination of rate constants for the addition of halide and
acetate ions to 1 extends the Ritchie N+ scale to a 108-fold span
of nucleophile reactivity and allows for a broad comparison of
Ritchie N+ nucleophilicity parameters for addition to sp2-
hybridized electrophilic carbon with Swain-Scott n parameters
for bimolecular aliphatic nucleophilic displacement reactions
in water.55-57 Figure 5B shows that there is a good linear
correlation with a slope of 2.0 (r ) 0.98) between the values
of N+ and n for the reactions of all the nucleophiles examined
here, except azide ion and nucleophiles with nonbonding
electron pair(s) at atoms adjacent to the nucleophilic site (R-
effect nucleophiles). There is good agreement between the
∂2∆Gq/∂∆G°2 ) 1/8Λ
(13)
We have used the rate and equilibrium constants and intrinsic
barriers for the addition of nucleophiles to 1 (Table 2) to
calculate hypothetical changes in nucleophile selectivity for
reactions that follow the Marcus equation (eq 11). The results
of these calculations for nucleophilic addition reactions of
acetate and halide ions to 1 are shown in Figure 7 where (a)
the center correlation shows the experimental data for nucleo-
phile addition to 1 (Table 1) and (b) the upper and lower two
correlations were constructed using rate constants calculated
from eq 11 with incremental 5 kcal/mol decreases and increases,
respectively, in ∆G° for nucleophile addition to 1, with the
(55) Swain, C. G.; Scott, C. B. J. Am. Chem. Soc. 1953, 75, 141-147.
(56) Koskikallio, J. Acta Chem. Scand. 1969, 23, 1477-1489.
(57) Pearson, R. G. J. Org. Chem. 1987, 52, 2131-2136.

1674 J. Am. Chem. Soc., Vol. 122, No. 8, 2000
Richard et al.
correlation shown in Figure 5B and an earlier correlation (slope
) 2.1) of Ritchie N+ (103-fold range of nucleophile reactivity)
and Swain-Scott n values for primary and secondary amines.58
The observation that the data for the addition of both anionic
and neutral nucleophiles to charged (carbocations) and neutral
(alkyl halides) electrophiles are correlated by a single line
(Figure 5B) provides further evidence that Coulombic interac-
tions between the electrophile and nucleophile do not have a
significant effect on the relative order of nucleophile reactivity.59
The extended correlation in Figure 5B clearly defines the most
important differences between nucleophilic reactivity toward sp2-
hybridized and sp3-hybridized electrophilic carbon:
(1) The slope of 2.0 for the correlation line in Figure 5B
shows that the rate constants for nucleophile addition to sp2-
hybridized carbon are twice as sensitive to changes in nucleo-
phile reactivity as those for nucleophilic substitution at sp3-
hybridized carbon. This requires that there be a significantly
larger fractional formation of the carbon-nucleophile bond at
the transition state 5A for addition to sp2-hybridized electrophiles
than at transition state 5B for aliphatic nucleophilic substitution.
The rate constants for nucleophilic substitution at the less
selective methyl halides are significantly smaller than those for
nucleophilic addition to the more selective electrophiles which
follow the N+ scale. This is another violation of the reactivity-
selectivity “principle”.60 The observation that the activation
barriers for formation of transition state 5A are significantly
lower than those for formation of transition state 5B in which
there is a smaller fractional bond formation to the reacting
nucleophile shows that there is a much steeper rise in energy
with development of a fifth bond to carbon at the pentavalent
transition state 5B for bimolecular aliphatic substitution, com-
pared with development of a fourth bond to carbon at the
tetravalent transition state 5A. These results are in line with
the expected difficulty of placing 10 bonding electrons at the
central carbon.
(2) There are differences in the relative reactivity of R-effect
nucleophiles toward carbocations and alkyl halides.61 The
deviations for R-effect nucleophiles reflect the well-known larger
effect of nonbonding R-electron pair(s) on nucleophile reactivity
toward carbocations than in bimolecular aliphatic substitution
reactions.61-65 However, our data provide no additional insight
into the origin(s) of the R-effect, which has been the subject of
intense, but not entirely conclusive, discussion.61,66-68
(3) The large positive deviation of the point for azide ion
from the correlation in Figure 5B has been noted in earlier work,
but also is not well understood51a,69
Acknowledgment. We acknowledge the National Institutes
of Health Grant GM 39754 for its generous support of this work
and Tina L. Amyes for helpful discussion.
Supporting Information Available: A table of the observed
second-order rate constants (kNu obsd
(M-1 s-1) for addition of
)
anionic and neutral nucleophiles to 1 as a function of pH (data
plotted in Figure 1). This material is available free of charge
via the Internet at http://pubs.acs.org.
JA9937526
(61) Hoz, S.; Buncel, E. Isr. J. Chem. 1985, 26, 313-319.
(62) Oae, S.; Kadoma, Y.; Yano, Y. Bull. Chem. Soc. Jpn. 1969, 42,
1110-1112. McIsaac, J. E.; Subbaraman, L. R.; Mulhausen, H. A.;
Behrman, E. J. J. Org. Chem. 1972, 37, 1037-1041.
(63) Gregory, M. J.; Bruice, T. C. J. Am. Chem. Soc. 1967, 89, 4400-
4402.
(64) Zoltewicz, J. A.; Deady, L. W. J. Am. Chem. Soc. 1972, 94, 2765-
2769.
(65) Pearson, R. G.; Edgington, D. N. J. Am. Chem. Soc. 1962, 84, 4607-
4608.
(66) Ritchie, C. D.; Minasz, R. J.; Kamego, A. A.; Sawada, M. J. Am.
Chem. Soc. 1977, 99, 3747-3753. Sander, E. G.; Jencks, W. P. J. Am.
Chem. Soc. 1968, 90, 6154-6161. Dixon, J. E.; Bruice, T. C. J. Am. Chem.
Soc. 1971, 93, 3248-3254.
(67) Hoz, S. In Nucleophilicity; Harris, J. M., McManus, S. P., Eds.;
American Chemical Society: Washington, DC, 1987; pp 181-193.
(68) Colthurst, M. J.; Kanagasooriam, A. J. S. S.; Wong, M. S. O.;
Contini, C.; Williams, A. Can. J. Chem. 678, 76, 678-685.
(69) Amyes, T. L.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 7900-
7909.
(70) Gandler, J. R. J. Am. Chem. Soc. 1985, 107, 8218-8223.
(71) Ride, J. N.; Wyatt, P. A. H.; Zochowski, Z. M. J. Chem. Soc., Perkin
Trans. 2 1974, 1188-1189.
(72) Ritchie, C. D.; VanVerth, J. E.; Virtanen, P. O. I. J. Am. Chem.
Soc. 1982, 104, 3491-3497.
(58) Bunting, J. W.; Mason, J. M.; Heo, C. K. M. J. Chem. Soc., Perkin
Trans. 2 1994, 2291-2300.
(59) Ritchie, C. D.; Sawada, M. J. Am. Chem. Soc. 1977, 99, 3754-
3761.
(60) For other examples of “failures” in the reactivity-selectivity
principle see: Young, P. R.; Jencks, W. P. J. Am. Chem. Soc. 1979, 101,
3288-3294.