How does organic structure determine organic reactivity? Nucleophilic substitution and alkene-forming elimination reactions of α-carbonyl and α-thiocarbonyl substituted benzyl derivatives

Article abstract of DOI:10.1021/ja9629064

The effect of α-(N,N-dimethylcarbamoyl) and α-(N,N-dimethylthiocarbamoyl) substituents on the rate constants for partitioning of α-substituted 1-(4-methoxyphenyl)ethyl carbocations between nucleophilic addition of 50:50 (v:v) MeOH/H₂0 (k(s), s^-1) and deprotonation by this solvent (k(e), s^-1) have been investigated. The data show that these α-amide and α-thioamide substituents result in 80-fold and ≤ 30000-fold decreases, respectively, in k(s) for capture of the 4-methoxybenzyl carbocation by solvent, but that they lead to much smaller changes in k(e), for deprotonation of the corresponding α-substituted 1-(4-methoxyphenyl)ethyl carbocations by solvent. The large effect of the α-thioamide substituent on the partitioning of α-substituted 1-phenylethyl carbocations between formation of the products of solvolysis and elimination is therefore due primarily to the effect of this α-substituent on k(s), for capture of the carbocation by solvent. The results of experimental and computational studies are consistent with the conclusion that the relative magnitude of the rate constants k(s) and k(e) for partitioning of α-substituted 1-phenylethyl carbocations is strongly controlled by the relative thermodynamic stabilities of the neutral products of these reactions.

Full text of DOI:10.1021/ja9629064

J. Am. Chem. Soc. 1996, 118, 12603-12613
12603
How Does Organic Structure Determine Organic Reactivity?
Nucleophilic Substitution and Alkene-Forming Elimination
Reactions of R-Carbonyl and R-Thiocarbonyl Substituted
Benzyl Derivatives
John P. Richard,*,† Shrong-Shi Lin,† Jeanne M. Buccigross,‡ and Tina L. Amyes†
Contribution from the Department of Chemistry, UniVersity at Buffalo, SUNY,
Buffalo, New York 14260-3000, and the Department of Chemistry, College of Mount St. Joseph,
Cincinnati, Ohio 45233-1670
ReceiVed August 19, 1996X
Abstract: The effect of R-(N,N-dimethylcarbamoyl) and R-(N,N-dimethylthiocarbamoyl) substituents on the rate
constants for partitioning of R-substituted 1-(4-methoxyphenyl)ethyl carbocations between nucleophilic addition of
50:50 (v:v) MeOH/H2O (ks, s-1) and deprotonation by this solvent (ke, s-1) have been investigated. The data show
that these R-amide and R-thioamide substituents result in 80-fold and g30 000-fold decreases, respectively, in ks for
capture of the 4-methoxybenzyl carbocation by solvent, but that they lead to much smaller changes in ke for
deprotonation of the corresponding R-substituted 1-(4-methoxyphenyl)ethyl carbocations by solvent. The large effect
of the R-thioamide substituent on the partitioning of R-substituted 1-phenylethyl carbocations between formation of
the products of solvolysis and elimination is therefore due primarily to the effect of this R-substituent on ks for
capture of the carbocation by solvent. The results of experimental and computational studies are consistent with the
conclusion that the relative magnitude of the rate constants ks and ke for partitioning of R-substituted 1-phenylethyl
carbocations is strongly controlled by the relative thermodynamic stabilities of the neutral products of these reactions.
The reactions of the liberated carbocation intermediates of
the stepwise reactions of simple R-methylbenzyl derivatives in
nucleophilic solvents give mainly the products of solvolysis and
less than 1% of the corresponding alkene product of an
elimination reaction.1,2 We were therefore surprised to learn
that the substitution of an R-thioamide group for an R-methyl
group at ring-substituted cumyl esters leads to a dramatic change
in the product distribution for the reaction of these substrates
in nucleophilic solvents, from mainly the solvent adducts, to
the exclusive formation of the alkene product of an elimination
reaction (Scheme 1).3 This shows that the R-thioamide group
causes a profound change in the partitioning of ring-substituted
cumyl carbocations, from the predominant nucleophilic addition
of solvent (ke/ks , 1, Scheme 1), to the virtually exclusive loss
of a proton to form the R-substituted styrene (ke/ks . 1, Scheme
1).
Experimental Section
Materials. Unless noted otherwise, inorganic salts and organic
chemicals were reagent grade from commercial sources and were used
without further purification. Tetrahydrofuran was distilled from sodium/
benzophenone and dichloromethane was distilled from calcium hydride.
The water used for kinetic studies and HPLC analyses was distilled
and then passed through a Milli-Q water purification system.
Synthesis. R-(Ethoxycarbonyl)-4-methoxybenzyl alcohol (EtOC-
(O)-1-OH)4 and its pentafluorobenzoate ester (EtOC(O)-1-(pentafluo-
robenzoate)),4 2-(4-methoxyphenyl)-2-propyl 4-nitrobenzoate (CH3-2-
(4-nitrobenzoate)),4 and 1-(N,N-dimethylthiocarbamoyl)-1-(4-methoxy-
phenyl)ethanol (Me2NC(S)-2-OH)3 were prepared by literature proce-
dures. The procedures for the synthesis of the following compounds,
We report here experiments and calculations which were
aimed at characterization of the effect of R-thiocarbonyl and
R-carbonyl substituents on the following: (1) the rate constant
ks (s-1) for the addition of solvent to R-1+ and R-2+, (2) the
rate constant ke (s-1) for deprotonation of R-2+ to form the
R-substituted 4-methoxystyrene R-3, and (3) the relative ther-
modynamic stabilities of the alkene R-3 and the corresponding
water adduct (alcohol) R-2-OH.
Scheme 1
† University at Buffalo, SUNY.
‡ College of Mount St. Joseph.
X Abstract published in AdVance ACS Abstracts, November 15, 1996.
(1) Richard, J. P.; Rothenberg, M. E.; Jencks, W. P. J. Am. Chem. Soc.
1984, 106, 1361-1372.
(2) Richard, J. P.; Jagannadham, V.; Amyes, T. L.; Mishima, M.; Tsuno,
Y. J. Am. Chem. Soc. 1994, 116, 6706-6712.
(3) Creary, X.; Hatoum, H. N.; Barton, A.; Aldridge, T. E. J. Org. Chem.
1992, 57, 1887-1897.
(4) Amyes, T. L.; Stevens, I. W.; Richard, J. P. J. Org. Chem. 1993, 58,
6057-6066.
S0002-7863(96)02906-X CCC: $12.00 © 1996 American Chemical Society

12604 J. Am. Chem. Soc., Vol. 118, No. 50, 1996
Richard et al.
MeOH/H2O was determined using authentic materials. A ratio of ꢀalkene
along with spectroscopic and analytical data for these compounds, are
given in the Supporting Information: Me2NC(S)-1-OH; Me2NC(O)-
2-OH; Me2NC(O)-1-OH; EtOC(O)-2-OH; Me2NC(S)-1-N3; Me2NC-
(S)-2-N3; Me2NC(S)-2-(4-nitrobenzoate); Me2NC(S)-2-(4-methoxyben-
zoate); Me2NC(O)-2-(4-nitrobenzoate); Me2NC(O)-2-(pentafluoro-
benzoate); EtOC(O)-2-(pentafluorobenzoate); Me2NC(S)-1-(4-nitroben-
zoate); Me2NC(S)-1-(3,5-dinitrobenzoate); Me2NC(S)-1-(pentafluo-
robenzoate); Me2NC(O)-1-(pentafluorobenzoate); Me2NC(S)-3; Me2-
NC(O)-3; EtOC(O)-3.
HPLC Analyses. The products of the reactions of R-1-Y and R-2-Y
were separated by HPLC as described previously,1,2,5,6 except that peak
detection was by a Waters 996 diode array detector. The products were
detected by their UV absorbance at the following wavelengths, which
are λmax for the corresponding alcohols: Me2NC(S)-1-Y, 271 nm; Me2-
NC(O)-1-Y, 275 nm; EtOC(O)-1-Y, 275 nm; Me2NC(O)-2-Y, 275 nm;
EtOC(O)-2-Y, 274 nm; CH3-2-Y, 273 nm. The products of the
reactions of Me2NC(S)-2-Y were detected at 266 nm, which is λmax
for Me2NC(S)-3.
/
ꢀRN ) 2.7 at 266 nm for Me2NC(S)-3 and Me2NC(S)-2-N3 was
3
calculated using eq 2, where ∆Aalkene and ∆ARN3 are the changes in the
HPLC peak areas for Me2NC(S)-3 and Me2NC(S)-2-N3, respectively,
observed for reaction of a constant amount of Me2NC(S)-2-(4-
nitrobenzoate) in the presence of increasing concentrations of azide
ion. The reproducibility of the extinction coefficient ratios determined
in different experiments was better than (10%. The relative extinction
coefficients of the alkene EtOC(O)-3 and the corresponding nucleophile
adducts EtOC(O)-2-Y were not determined.
Calculation of Rate Constant Ratios. Dimensionless rate constant
ratios for the reactions of nucleophiles Nu1 and Nu2 with R-1-Y and
R-2-Y were calculated from the product yields using eq 3. The rate
kNu /kNu ) [RNu1][Nu2]/[RNu2][Nu1]
(3)
1
2
constant ratios ke/ks for reaction of solvent as a Brønsted base to form
the elimination product (ke) and as a nucleophile to form the solvolysis
products (ks) were calculated directly as the ratio of the yields of these
products. Rate constant ratios kaz/ks (M-1) for partitioning of R-1-Y
and R-2-Y between reaction with azide ion and solvent were calculated
using eq 4. The estimated error in these rate constant ratios is (10%,
the uncertainty in A1/A2 (eq 1), plus any uncertainty in the value of
ꢀP /ꢀP (see above).
The products of the reaction of solvent and azide ion with R-1-Y
and R-2-Y were identified as described in earlier work.1,2,6 The azide
ion adducts Me2NC(S)-1-N3 and Me2NC(S)-2-N3 were also isolated
and characterized by NMR and IR spectroscopy (see Supporting
Information). The alkenes R-3 formed from the reactions of R-2-Y
were identified by comparison of their HPLC retention times with those
for authentic materials.
2
1
-
kaz/ks (M-1) ) [RN3]/ [ROSolv][N3
]
(4)
Procedures for Product Studies. Product studies were carried out
at room temperature (22 ( 2 °C). Control experiments showed that
the product ratios obtained at room temperature and at 25 °C are
identical. Aqueous solutions of sodium azide were adjusted to pH ≈
7 with concentrated HClO4 before use. Reactions were initiated by
making a 100-fold dilution of a solution of substrate (10-2-10-3 M)
in acetonitrile into 50:50 (v:v) TFE/H2O or 50:50 (v:v) MeOH/H2O at
I ) 0.50 (NaClO4). The reactions of Me2NC(S)-2-(4-nitrobenzoate)
in 50:50 (v:v) MeOH/H2O containing both azide and acetate ions were
buffered with 1 or 5 mM sodium phosphate. These solutions were
prepared by dilution of 20 mM sodium phosphate buffer in water at
pH ) 7.0 to the appropriate final concentration in 50:50 (v:v) MeOH/
H2O (I ) 0.50, NaClO4). The dehydration of Me2NC(S)-2-OH to give
Me2NC(S)-3 was carried out in 50:50 (v:v) TFE/H2O containing 0.05-
0.50 M perchloric acid (I ) 0.50, NaClO4).
∑
Rate constant ratios for partitioning of the carbocation intermediates
R-2+ of the reactions of R-2-Y in the presence of azide ion alone, or
in the presence of both azide and acetate ions, were determined from
the nonlinear least squares fit of the product data to the appropriate
equation (see Discussion), using SigmaPlot from Jandel Scientific.
Kinetic Methods. Kinetic studies were carried out in 50:50 (v:v)
TFE/H2O or 50:50 (v:v) MeOH/H2O at 25 °C and I ) 0.50 (NaClO4).
Aqueous solutions of sodium azide were adjusted to pH ≈ 7 with
concentrated HClO4 before use. The reactions were initiated by making
a 100-fold dilution of a solution of substrate (10-2-10-3 M) in
acetonitrile into the appropriate reaction mixture. For reactions that
were monitored by HPLC analysis, the reaction mixture also contained
ca. 10-5 M of 9-hydroxy-9-methylfluorene or 9-methoxyfluorene as
an internal standard to correct for small variations in the HPLC injection
volume.
The yields of the products of the reaction of Me2NC(S)-2-(4-
methoxybenzoate) with azide ion in 50:50 (v:v) MeOH/H2O were
determined during the first 3 h of the reaction, because at later times
there was significant conversion of the azide ion adduct to the alkene
Me2NC(S)-3. For all other R-1-Y and R-2-Y, the products were shown
to be stable for at least three halftimes of the reaction of the substrate.
Ratios of product yields ([P1]/[P2]) were calculated using eq 1, where
A1/A2 and ꢀP /ꢀP are the ratios of the peak areas from HPLC analysis
The reactions of substituted benzoate esters were monitored by
following either (a) the change in UV absorbance at the following wave-
lengths: Me2NC(S)-1-(pentafluorobenzoate), 275 nm; Me2NC(S)-2-
(4-nitrobenzoate), 274 nm; or (b) the disappearance of the substrate
by HPLC at the following wavelengths: Me2NC(S)-2-(4-methoxyben-
zoate), 266 nm; Me2NC(S)-1-(4-nitrobenzoate), 269 nm; Me2NC(S)-
1-(3,5-dinitrobenzoate), 271 nm; Me2NC(O)-2-(4-nitrobenzoate), 262
nm; Me2NC(O)-2-(pentafluorobenzoate), 273 nm; Me2NC(O)-1-(pen-
tafluorobenzoate), 273 nm; EtOC(O)-2-(pentafluorobenzoate), 273 nm;
EtOC(O)-1-(pentafluorobenzoate), 273 nm.
The progress of the dehydration of Me2NC(S)-2-OH to give Me2-
NC(S)-3 in 50:50 (v:v) TFE/H2O containing 0.05 - 0.50 M perchloric
acid (I ) 0.50, NaClO4) was monitored by HPLC analysis at 271 nm.
For all reactions, good first-order kinetics were observed over at
least three half-lifes. First-order rate constants were determined from
the slopes of linear semilogarithmic plots of reaction progress against
time. The second-order rate constant for the acid-catalyzed dehydration
of Me2NC(S)-2-OH was determined as the slope of the linear plot of
2
1
and the extinction coefficients of the two products at the wavelength
of the analysis, respectively. The reproducibility of the product ratios
[P1]/[P2] ) (A1/A2)(ꢀP /ꢀP )
(1)
(2)
2
1
ꢀalkene/ꢀRN ) ∆Aalkene/∆ARN
3
3
from HPLC analysis was (10%. For R-1-Y, EtOC(O)-2-Y, and CH3-
2-Y, the extinction coefficients of the alcohols and the corresponding
methanol and azide ion adducts were shown to be identical by
solvolyzing the corresponding benzoate esters in the presence of
increasing concentrations of methanol or azide ion and showing, by
HPLC analysis of a constant amount of total product, that the decrease
in the HPLC peak area for the alcohol product is equal to the increase
in the HPLC peak area for the methanol or azide ion adduct. The
extinction coefficients of Me2NC(O)-2-N3 and Me2NC(O)-2-OH at 275
nm were assumed to be identical because the absorbance of these
compounds at this wavelength is due primarily to the 4-methoxyphenyl
k
obsd against [HClO4]. Rate constants determined spectrophotometrically
were reproducible to (5%, and rate constants determined by HPLC
analysis were reproducible to (10%.
Ab Initio Calculations. The potential energy surfaces of R-sub-
stituted 1-phenylethanols and the corresponding R-substituted styrenes
were examined with the semiempirical AM1 Hamiltonian7 using
MOPAC 4.08 on an IBM ES9121-400 computer at Miami University,
group. A ratio of ꢀalkene/ꢀROH ) (12 600 M-1 cm-1)/(1800 M-1 cm-1
)
(7) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. G.; Stewart, J. P. J. Am.
Chem. Soc. 1985, 107, 3902-3909.
(8) Stewart, J. J. P.; Seiler, F. J., MOPAC 4.0 (QCPE 455), Quantum
Chemistry Program Exchange: Department of Chemistry, Indiana Univer-
sity.
) 7.0 at 275 nm for Me2NC(O)-3 and Me2NC(O)-2-OH in 50:50 (v:v)
(5) Amyes, T. L.; Richard, J. P. J. Am. Chem. Soc. 1990, 112, 9507-
9512.
(6) Richard, J. P. J. Am. Chem. Soc. 1989, 111, 1455-1465.

How Does Organic Structure Determine Organic ReactiVity
J. Am. Chem. Soc., Vol. 118, No. 50, 1996 12605
Table 1. Observed First-Order Rate Constants for the Reactions of
R-1-Y and R-2-Y in 50:50 (v:v) Methanol/Water and 50:50 (v:v)
Trifluoroethanol/Water at 25 °C and I ) 0.50 (NaClO4)a
Table 2. Rate Constant Ratios for Partitioning of R-Substituted
4-Methoxybenzyl Carbocations R-1+ between Reaction with
Nucleophilic Reagents in 50:50 (v:v) Methanol/Watera
R-1-Y
kobsd (s-1
)
50:50 (v:v)
MeOH/H2O
50:50 (v:v)
TFE/H2O
b
R
kMeOH/kHOH
kaz/ksc (M-1
)
R
leaving group
Me2NC(S)d
Me2NC(O)f
EtOC(O)h
5
22
11
9.0
7 × 104 e
CH3
4-nitrobenzoate
pentafluorobenzoate
4-nitrobenzoate
3,5-dinitrobenzoate
pentafluorobenzoated
pentafluorobenzoate
pentafluorobenzoate
1.3 × 10-4 b
1.5 × 10-2 c
3.7 × 10-5
3.1 × 10-4
4.2 × 10-3
2.3 × 10-6
5.5 × 10-7 b
210g
44
7.2
Me2NC(S)
3.6 × 10-5
2.7 × 10-4
5.3 × 10-3
1.9 × 10-6
7.0 × 10-7
i
CH3
a At 22 ( 2 °C and I ) 0.50 (NaClO4) and determined from product
analysis by HPLC. b Dimensionless ratio of second-order rate constants
for reaction of the carbocation with methanol and with water (eq 3).
c Rate constant ratio for partitioning of the carbocation between reaction
with azide ion and with solvent, calculated using a first-order rate
constant for the reaction of solvent (eq 4). d Identical results were
obtained using Me2NC(S)-1-(4-nitrobenzoate), Me2NC(S)-1-(3,5-dini-
trobenzoate), or Me2NC(S)-1-(pentafluorobenzoate) as the precursor
to the carbocation. e Average ((10%) of four determinations in the
range [N3-] ) 0.5-1.0 mM. f Determined using Me2NC(O)-1-(pen-
tafluorobenzoate) as the precursor to the carbocation. g Average ((10%)
of five determinations in the range [N3-] ) 1.0-10 mM. h Determined
using EtOC(O)-1-(pentafluorobenzoate) as the precursor to the car-
bocation. i Data from ref 1.
Me2NC(O)
EtOC(O)
R-2-Y
kobsd (s-1
)
50:50 (v:v) 50:50 (v:v) R-Me/
R
leaving group
MeOH/H2O TFE/H2O
0.090b
R-He
690g
CH3
4-nitrobenzoate
Me2NC(S) 4-methoxybenzoate 3.7 × 10-5 2.0 × 10-5
4-nitrobenzoated
1.3 × 10-3 7.1 × 10-4
1.4 × 10-6 8.5 × 10-7
36f
19g
Me2NC(O) 4-nitrobenzoate
pentafluorobenzoate 1.6 × 10-4 9.4 × 10-5
pentafluorobenzoate 4.0 × 10-5 3.5 × 10-5
84f
41g
57f
64g
reaction progress by UV spectroscopy or HPLC analysis, as
described in the Experimental Section. There was no significant
change in kobsd ((10%) for reaction of any of these substrates
when the concentration of azide ion was increased from zero
to 0.50 M at I ) 0.50 (NaClO4). At [N3-] ) 0.50 M, the yields
of the corresponding azide ion adducts from these reactions were
at least 40%.
EtOC(O)
a Unless noted otherwise, rate constants were determined by fol-
lowing the disappearance of the substrate by HPLC as described in
the Experimental Section. b Data from ref 4. c Estimated from kobsd
)
1.3 × 10-4 s-1 for CH3-1-(4-nitrobenzoate) and a rate constant ratio of
114 for reaction of Me2NC(S)-1-(pentafluorobenzoate) and Me2NC(S)-
1-(4-nitrobenzoate) in 50:50 (v:v) trifluoroethanol/water (this work).
d Determined spectrophotometrically as described in the Experimental
Section. e Effect of the addition of an R-methyl group on kobsd for
reaction of R-1-Y, calculated as the ratio of rate constants for reaction
of R-2-Y and R-1-Y. f In 50:50 (v:v) methanol/water. g In 50:50 (v:v)
trifluoroethanol/water.
The yields of the products of the reactions of R-1-Y and
R-2-Y in 50:50 (v:v) MeOH/H2O (I ) 0.50, NaClO4) in the
presence of increasing concentrations of azide ion were deter-
mined by HPLC analyses. Tables 2 and 3 give the dimension-
less product rate constant ratios kMeOH/kHOH for partitioning of
R-1-Y and R-2-Y between reaction with methanol and with
water in 50:50 (v:v) MeOH/H2O, which were calculated from
the yields of the methyl ethers and the corresponding alcohols
or MOPAC 939 on a DEC 5000 workstation at the University of New
England, Australia. After the initial optimization of geometry, the bonds
between the benzylic carbon and the phenyl ring and the benzylic carbon
and the R-substituent were systematically rotated and the energy
recalculated, in order to ensure that a global minimum had been located.
The AM1 global minimum was used as the starting geometry for
the ab initio calculations, which were performed using Gaussian 9210
on a Cray Y-MP8/864 computer at the Ohio Supercomputer Center.
Complete geometry optimization was performed by restricted Hartree-
Fock calculations with the split-valence 3-21G basis set.11,12 The
calculations were performed without symmetry restrictions and all
geometric parameters were fully optimized. Single-point calculations
with the polarization basis set 6-31G*13 were performed on each
minimum energy structure that was located with the 3-21G basis set.
using eq 3. Tables 2 and 3 also give the values of kaz/ks (M-1
)
for the reactions of R-1-Y and R-2-Y with azide ion and solvent,
which were calculated from the yields of the azide ion adduct
and the solvent adducts using eq 4. There was no detectable
formation of solvent adducts from the reactions of Me2NC(S)-
2-Y.
The rate constant ratio ks/ke ) 1.2 (Table 3) for the
partitioning of Me2NC(O)-2-(pentafluorobenzoate) between
solvolysis (ks) and elimination (ke) in 50:50 (v:v) MeOH/H2O
was calculated directly as the ratio of the yields of the products
of these reactions. The observed ratio of the HPLC peak areas
for the solvolysis and elimination products of the reaction of
EtOC(O)-2-(pentafluorobenzoate) in 50:50 (v:v) MeOH/H2O
was used to calculate ks/ke ≈ 50 (Table 3). This probably
underestimates this ratio because the molar absorbtivity of the
conjugated alkene EtOC(O)-3 at 274 nm (the wavelength used
for HPLC analysis) is expected to be larger than that of the
solvent adducts EtOC(O)-2-OSolv (ꢀalkene > ꢀROSolv, eq 1).
There was no detectable formation of solvent adducts from the
reactions of Me2NC(S)-2-Y in 50:50 (v:v) MeOH/H2O, and the
yield of the alkene CH3-3 from the reaction of CH3-2-(4-
nitrobenzoate) in this solvent is less than 1%.14 Therefore, the
values of ks/ke (Table 3) for the reactions of Me2NC(S)-2-Y
and CH3-2-(4-nitrobenzoate) are upper and lower limits, re-
Results
Table 1 gives the observed first-order rate constants, kobsd
(s-1), for the reactions of R-1-Y and R-2-Y in 50:50 (v:v)
MeOH/H2O (I ) 0.50, NaClO4) and 50:50 (v:v) TFE/H2O (I )
0.50, NaClO4), which were determined by monitoring the
(9) Stewart, J. J. P. MOPAC 93, Fujitsu (c), 1993.
(10) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Wong, M. W.; Foresman, J. B.; Robb, M. A.; Head-Gordon,
M.; Replogle, E. S.; Gomperts, R.; Andres, J. L.; Raghavachari, K.; Binkley,
J. S.; Gonzalez, C.; Martin, R. L.; Fox, D. J.; Defrees, D. J.; Baker, J.;
Stewart, J. J. P.; Pople, J. A. Gaussian 92/DFT, ReVision G.3; Gaussian
Inc.: Pittsburgh, PA, 1993.
(11) Binkley, J. S.; Pople, J. A.; Hehre, W. J. J. Am. Chem. Soc. 1980,
102, 939-947.
(12) Gordon, M. S.; Binkley, J. S.; Pople, J. A.; Pietro, W. J.; Hehre,
W. J. J. Am. Chem. Soc. 1982, 104, 2797-2803.
(14) It was not possible to demonstrate that CH3-2-(4-nitrobenzoate) was
completely free of the alkene CH3-3, so that the observed formation of less
than 1% of CH3-3 may represent an impurity present in the substrate.
(13) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213-
222.

12606 J. Am. Chem. Soc., Vol. 118, No. 50, 1996
Richard et al.
Table 3. Rate Constant Ratios for Partitioning of R-Substituted
1-(4-Methoxyphenyl)ethyl Carbocations R-2+ between Nucleophilic
Addition and Base-Catalyzed Deprotonation in 50:50 (v:v)
Methanol/Watera
kMeOH
kHOH
/
kaz/ksd
(M-1
kaz/ke
b
b
b
R
ks/kec
)
(M-1
)
kaz/kB kB/kAc
Me2NC(S)e
<0.01f
1.2
65g 0.59h
4.5i
Me2NC(O)j 31
1400
1700k 10h
EtOC(O)l
14
≈50m
84 ≈4200k large
n
CH3
9.0 >100o
18 >1800k large
a At 22 ( 2 °C and I ) 0.50 (NaClO4) and determined from product
analysis by HPLC. b Dimensionless ratio of second-order rate constants.
c Dimensionless ratio of first-order rate constants, calculated as the ratio
of the yields of the solvent adducts and the alkene. d Rate constant ratio
for partitioning of the carbocation between reaction with azide ion and
with solvent and an average ((10%) of values determined at five or
more concentrations of azide ion (eq 4). e Identical results were obtained
using Me2NC(S)-2-(4-methoxybenzoate) or Me2NC(S)-2-(4-nitroben-
zoate) as the precursor to the carbocation. f Upper limit calculated with
the assumption that a 1% yield of the solvent adducts could have been
detected by HPLC analysis. g Determined from the nonlinear least
squares fit of the product data in Figure 1A to eqs 6 and 8 (see text).
h The ratio of the constant limiting yields of the azide ion adduct and
the alkene at high [N3-] (see Figure 1). i Calculated from kB/ke ) 110
M-1 (see text) and kAc/ke ) 24.5 M-1 that was determined from the
nonlinear least squares fit of the product data in Figure 2 to eqs 9 and
10 (see text). j Determined using Me2NC(O)-2-(pentafluorobenzoate)
as the precursor to the carbocation. k Calculated as (ks/ke)(kaz/ks).
l Determined using EtOC(O)-2-(pentafluorobenzoate) as the precursor
to the carbocation. m Calculated from the ratio of the observed HPLC
peak areas for the solvent adducts and the alkene with the assumption
that ꢀalkene ≈ ꢀROSolv (eq 1). n Determined using CH3-2-(4-nitrobenzoate)
as the precursor to the carbocation. o The yield of the alkene is less
than 1% and the upper limit was calculated with the assumption that a
1% yield of the alkene could have been detected by HPLC analysis.
Figure 1. (A) Effect of azide ion on the fractional yields of the azide
ion adduct Me2NC(S)-2-N3 and the alkene Me2NC(S)-3 from the
reactions of Me2NC(S)-2-Y in 50:50 (v:v) methanol/water at 22 ( 2
°C and I ) 0.50 (NaClO4). Key: b and 1, Y ) 4-nitrobenzoate; 2
and 9, Y ) 4-methoxybenzoate. (B) Effect of azide ion on the fractional
yields of the azide ion adduct Me2NC(O)-2-N3, the solvent adducts
Me2NC(O)-2-OSolv, and the alkene Me2NC(O)-3 from the reaction of
Me2NC(O)-2-(pentafluorobenzoate) in 50:50 (v:v) methanol/water at
22 ( 2 °C and I ) 0.50 (NaClO4).
Table 4. Ab Initio Energies of the Geometry Optimized Structures
of R-Substituted 1-Phenylethanols R-9-OH and the Corresponding
R-Substituted Styrenes R-10a
energy (hartrees)b
3-21G//3-21G
6-31G*//3-21G
R
OH
CH3
-420.321749
-606.880681
-625.963840
-947.075646
CH3
-422.659327
-610.262997
-629.452363
-952.077792
MeOC(O)
Me2NC(O)
Me2NC(S)
Figure 2. Effect of acetate ion on the fractional yields of the alkene
Me2NC(S)-3 and the azide ion adduct Me2NC(S)-2-N3 from the reaction
of Me2NC(S)-2-(4-nitrobenzoate) in 50:50 (v:v) methanol/water at 22
( 2 °C and I ) 0.50 (NaClO4) in the presence of 5 mM sodium
phosphate buffer (pH ) 7 in water) and a constant concentration of
azide ion: (A) [N3-] ) 5 mM; (B) [N3-] ) 10 mM.
R
CH2
CH3
-344.692880
-531.250113
-550.337389
-871.449629
-346.621141
-534.224382
-553.416676
-876.047629
MeOC(O)
Me2NC(O)
Me2NC(S)
Figure 2 shows the effect of increasing concentrations of
acetate ion on the fractional yields of the products of the reaction
of Me2NC(S)-2-(4-nitrobenzoate) in 50:50 (v:v) MeOH/H2O (I
) 0.50, NaClO4) in the presence of 5 mM sodium phosphate
buffer (pH ) 7.0 in water) and 5 mM (Figure 2A) or 10 mM
(Figure 2B) azide ion. The dilute phosphate buffer was added
in order to minimize the reactions of hydroxide and methoxide
ions as specific bases. It was shown that the product yields are
unaffected by a decrease from 5 to 1 mM phosphate buffer.
The progress of the dehydration of Me2NC(S)-2-OH to give
the alkene Me2NC(S)-3 in 50:50 (v:v) TFE/H2O containing
0.05-0.50 M perchloric acid (I ) 0.50, NaClO4) was followed
by HPLC. The plot of kobsd against [HClO4] is linear over this
range of acid concentration, which shows that there is no
a Geometries were fully optimized by standard RHF procedures at
the 3-21G level (see the Experimental Section) and the results of these
calculations are given in the Supporting Information. b 1 hartree )
627.51 kcal/mol.
spectively, that were calculated with the assumption that 1%
yields of Me2NC(S)-2-OSolv and CH3-3 could have been
detected by HPLC analysis.
Figure 1A shows the effect of increasing concentrations of
azide ion on the fractional yields of the products of the reactions
of Me2NC(S)-2-(4-nitrobenzoate) and Me2NC(S)-2-(4-meth-
oxybenzoate) in 50:50 (v:v) MeOH/H2O (I ) 0.50, NaClO4),
and Figure 1B shows the corresponding data for the reaction
of Me2NC(O)-2-(pentafluorobenzoate).

How Does Organic Structure Determine Organic ReactiVity
J. Am. Chem. Soc., Vol. 118, No. 50, 1996 12607
detectable protonation of the R-thioamide group of the substrate
by 0.50 M HClO4. The second-order rate constant for the acid-
catalyzed dehydration of Me2NC(S)-2-OH, determined from the
slope of this plot, is kH ) 7.2 × 10-5 M-1 s-1. After ten
reaction halftimes no detectable Me2NC(S)-2-OH (< 1%)
remained, which shows that Kalk ) [alkene]eq/[alcohol]eq > 100.
It was not possible to monitor the dehydration of Me2NC(O)-
2-OH or EtOC(O)-2-OH to the corresponding alkenes, because
the acid-catalyzed hydrolysis of the R-amide or R-ester group
of these substrates is significantly faster than the dehydration
reaction.
Calculations. The minimum-energy conformations of R-sub-
stituted 1-phenylethanols and the corresponding R-substituted
styrenes, and the energies of these conformers were determined
using ab initio methods, as described in the Experimental
Section. The results of these calculations are summarized in
Table 4. The geometries of the optimized structures of these
compounds are given in the Supporting Information.
fluoroacetate,3,18 CH3-1-Y,1,19 CH3-2-Y,2,20 and EtOC(O)-1-
(pentafluorobenzoate)4,21 react by a stepwise mechanism through
carbocation intermediates.
Substituent Effects on Carbocation Formation. The
observed first-order rate constants, kobsd (s-1, Table 1), for the
stepwise reactions of the R-thioamide substituted benzyl deriva-
tives Me2NC(S)-1-Y and Me2NC(S)-2-Y are much larger than
those for reaction of their R-amide and R-ester substituted
counterparts, but they are smaller than those for reaction of the
corresponding R-methyl substituted derivatives. This shows that
the interaction of the R-thioamide group with the developing
positive charge in the transition state for ionization of Me2NC-
(S)-1-Y and Me2NC(S)-2-Y is less destabilizing than the
corresponding interactions of R-amide and R-ester groups. This
is due to the smaller destabilization of the developing positive
charge by the polar electron-withdrawing effect, and its larger
stabilization by the resonance electron-donating effect of the
R-thioamide group compared with the R-amide group.3,18,22
These results and conclusions are in good agreement with those
of earlier studies of the effects of R-carbonyl and R-thiocarbonyl
substituents on the rate constants for solvolysis reactions.3,18,23
Discussion
The reaction of Me2NC(S)-1-(pentafluorobenzoate) (1 × 10-4
M) in 50:50 (v:v) TFE/H2O (I ) 0.50, NaClO4) gives the
following products: 44% of the solvent adducts Me2NC(S)-1-
OH and Me2NC(S)-1-OCH2CF3; 5% of 2-dimethylamino-6-
methoxybenzothiophene, 4; and 52% of the dimeric product 5.15
The formation of 4 and 5 is consistent with a stepwise
The addition of an R-methyl group leads to a 690-fold
increase in kobsd (s-1) for the stepwise reaction of CH3-1-(4-
nitrobenzoate), but to significantly smaller increases of ca. 30-
fold, ca. 60-fold, and 60-fold in kobsd for the stepwise reactions
of Me2NC(S)-1-(4-nitrobenzoate), Me2NC(O)-1-(pentafluo-
robenzoate), and EtOC(O)-1-(pentafluorobenzoate), respectively
(Table 1). The addition of the strongly electron-withdrawing
R-amide and R-ester groups to the 4-methoxybenzyl carbocation
results in an increase in the delocalization of positive charge
away from the benzylic carbon and onto the 4-methoxyphenyl
ring,24 which serves to minimize the destabilizing polar electron-
withdrawing effects of these substituents;25 there may also be
some direct resonance delocalization of positive charge onto
these groups.26 Therefore, the relatively small effect of the
addition of an R-methyl group on kobsd for reaction of the
R-thiocarbonyl and R-carbonyl substituted benzyl derivatives
is probably due to the tendency of these electron-withdrawing
groups to effect delocalization of the developing positive charge
in the transition state away from the benzylic carbon. This in
turn will reduce the magnitude of the stabilizing hyperconju-
gative interaction between positive charge at this center and an
added R-methyl group. An R-thioamide group is less polar
electron-withdrawing3,18 but more resonance electron-donating
than an R-amide group.3,18,22,26 These differences will lead to
larger and smaller, respectively, effects of the addition of an
R-methyl group effect on kobsd for Me2NC(S)-1-Y than for Me2-
NC(O)-1-Y, with the net result being a 2-fold smaller effect
for the former (Table 1).
mechanism and a rapid intramolecular cyclization of the
carbocation intermediate Me2NC(S)-1+ to give the ben-
zothiophene 4.16 However, this activated alkene does not
accumulate to any large extent, but rather it undergoes nucleo-
philic addition to a second molecule of Me2NC(S)-1+swhich
is formed continuously during the reaction of Me2NC(S)-1-
(pentafluorobenzoate)sto give 5.15 In order to minimize the
formation of 4 and 5, the product studies for the reaction of
Me2NC(S)-1-Y described in this work were carried out in the
more nucleophilic solvent of 50:50 (v:v) MeOH/H2O. In this
solvent, the trapping of Me2NC(S)-1+ by solvent competes
effectively with its intramolecular cyclization, and the yields
of 4 and 5 from the reaction of Me2NC(S)-1-(pentafluoroben-
zoate) are zero and 2%, respectively.15
Reaction Mechanism. For all the R-1-Y and R-2-Y in Table
1, reaction in 50:50 (v:v) MeOH/H2O in the presence of 0.50
M NaN3 gives at least a 40% yield of the product of nucleophilic
substitution by azide ion, in a reaction that is zero-order in the
concentration of this nucleophile. These observations show that
these substrates react by a stepwise DN + AN (SN1)17 mechanism
through the liberated carbocation intermediates R-1+ and R-2+
which can be trapped by azide ion. It has been shown in earlier
work that 1-(N,N-dimethylthiocarbamoyl)-1-phenylethyl tri-
Rate and Equilibrium Constants for the Formation and
Reaction of r-Substituted 4-Methoxybenzyl Carbocations.
Table 5 gives the absolute rate constants ks (s-1) for reaction
of R-1+ with a solvent of 50:50 (v:v) MeOH/H2O that were
(19) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1984, 106, 1373-
1383.
(20) Richard, J. P.; Amyes, T. L.; Vontor, T. J. Am. Chem. Soc. 1991,
113, 5871-5873.
(21) Richard, J. P.; Amyes, T. L.; Stevens, I. W. Tetrahedron Lett. 1991,
32, 4255-4258.
(22) Creary, X.; Aldridge, T. J. Org. Chem. 1991, 56, 4280-4285.
(23) Creary, X. Acc. Chem. Res. 1985, 18, 3-8.
(24) Lee, I.; Chung, D. S.; Jung, H. J. Tetrahedron 1994, 50, 7981-
7986.
(25) Richard, J. P. Tetrahedron 1995, 51, 1535-1573.
(26) Lien, M. H.; Hopkinson, A. C. J. Am. Chem. Soc. 1988, 110, 3788-
3792.
(15) Richard, J. P.; Lin, S.-S.; Williams, K. J. Org. Chem., in press.
(16) Ablenas, F. J.; George, B. E.; Maleki, M.; Jain, R.; Hopkinson, A.
C.; Lee-Ruff, E. Can. J. Chem. 1987, 65, 1800-1803.
(17) Guthrie, R. D.; Jencks, W. P. Acc. Chem. Res. 1989, 22, 343-349.
(18) Creary, X.; Aldridge, T. E. J. Org. Chem. 1988, 53, 3888-3890.

12608 J. Am. Chem. Soc., Vol. 118, No. 50, 1996
Richard et al.
Table 5. Equilibrium Constants for Formation of R-Substituted
4-Methoxybenzyl Carbocations R-1+ from the Corresponding
Neutral Azide Ion Adducts and Rate Constants for Reaction of the
Carbocations with Solvent (Scheme 2)a
R
Kazb (M)
ksc (s-1
)
Me2NC(S)
Me2NC(O)
EtOC(O)
CH3
g1.7 × 10-17 d,e
6.1 × 10-21 d
1.8 × 10-21 f
2.6 × 10-17 f
7.4 × 10-20 f
e7 × 104 e
2.4 × 107
1.1 × 108
6.9 × 108
2.0 × 109 g
H
a In 50:50 (v:v) methanol/water at 25 °C and I ) 0.50 (NaClO4).
b Equilibrium constant for formation of the carbocation R-1+ from the
azide ion adduct R-1-N3, calculated as Kaz ) ksolv/kaz, where ksolv is the
4
rate constant for the stepwise reaction of R-1-N3 and kaz ) 5 × 109
M-1 s-1 for the diffusion-limited reaction of azide ion with R-1+, unless
noted otherwise. c First-order rate constant for reaction of the carboca-
tion with solvent, calculated from kaz/ks (M-1, Table 2) and kaz ) 5 ×
109 M-1 s-1 for the diffusion-limited reaction of azide ion with R-1+,
d
unless noted otherwise. ksolv for the reaction of R-1-N3 was estimated
from kobsd for the reaction of R-1-(pentafluorobenzoate) (Table 1) and
the 6.2 × 104-fold larger value of kobsd for the stepwise reactions of
R-1-(pentafluorobenzoate) than of R-1-N3 that was calculated from the
data in ref 4. e This is a limit because the rate constant for the reaction
of azide ion with Me2NC(S)-1+ may be smaller than that for a diffusion-
limited reaction. f Calculated using ksolv for the reaction of R-1-N3 in
50:50 (v:v) trifluoroethanol/water (ref 4), because the rate constants
for the reaction of other R-1-Y in 50:50 (v:v) methanol/water differ
from those for reaction in 50:50 (v:v) trifluoroethanol/water by less
than 2-fold (Table 1). g Calculated from kaz/ks ) 2.5 M-1 that was
estimated from kaz/ks ) 25 M-1 in 50:50 (v:v) trifluoroethanol/water
(ref 4) and the 8-fold and 15-fold smaller values of kaz/ks for partitioning
of EtOC(O)-1+ and CH3-1+, respectively, in 50:50 (v:v) methanol/water
(data from Table 2) than in 50:50 (v:v) trifluoroethanol/water (data
from ref 4).
Figure 3. Logarithmic correlation between the rate constants for the
reaction of R-substituted 4-methoxybenzyl carbocations R-1+ with
solvent (ks, s-1) and the equilibrium constants for their formation from
the corresponding neutral azide ion adducts (Kaz, M) in 50:50 (v:v)
methanol/water at 25 °C and I ) 0.50 (NaClO4) (Scheme 2). The figure
was constructed using the data in Table 5. The open circle denotes
that the values of ks and Kaz for Me2NC(S)-1+ are upper and lower
limits, respectively.
thermodynamic instability, the R-ester substituted carbocation
EtOC(O)-1+ has a relatively long lifetime (1/ks), and it is one
of several R-substituted 4-methoxybenzyl carbocations whose
lifetimes are nearly independent of their thermodynamic stabil-
ity.4,21,28 The data obtained in this work show that the R-amide
group is yet another example of a strongly electron-withdrawing
R-substituent that results in thermodynamic destabilization of
the 4-methoxybenzyl carbocation relative to the neutral azide
ion adduct (Kaz), but to a small decrease, rather than the expected
increase, in the rate constant for reaction of the carbocation with
a nucleophilic solvent (ks) (Figure 3 and Table 5). The detailed
descriptions and explanations of these R-substituent effects have
been presented in earlier work.4,25,28
Figure 3 shows that ks (s-1) for reaction of the R-thioamide
substituted carbocation Me2NC(S)-1+ with 50:50 (v:v) MeOH/
H2O is anomalously small in comparison with the rate constants
for the addition of this solvent to other R-1+. For example,
thermodynamically, Me2NC(S)-1+ is at least as stable as the
R-methyl carbocation CH3-1+, but it is also at least 104-fold
less reactive towards solvent (Table 5). This relatively large
kinetic barrier to the capture of Me2NC(S)-1+ by solvent reflects
the stabilizing interactions between the R-thioamide group and
the benzylic cationic center, and the requirement for the
relatively large fractional loss of these stabilizing interactions
on proceeding from the ground to the transition state for reaction
of the carbocation.29,30 There are two stabilizing interactions
between the R-thioamide group and the positive charge at the
benzylic carbon that may contribute to the relatively large kinetic
barrier to the reaction of Me2NC(S)-1+ with solvent.
Scheme 2
calculated from the product rate constant ratios kaz/ks (M-1
,
Table 2) and kaz ) kd ) 5 × 109 M-1 s-1 for the diffusion-
limited reaction of azide ion with unstable benzylic carboca-
tions.1,4,6,19,20,27 With the exception of kaz/ks ) 7 × 104 M-1
for the partitioning of Me2NC(S)-1+, the azide ion selectivities
for partitioning of all the R-1+ in Table 2 lie well below kaz/ks
) 104 M-1, which is the approximate threshold for the change
from a diffusion- to an activation-limited reaction of azide ion
with benzylic carbocations.25,27a Therefore, the value of ks for
the reaction of Me2NC(S)-1+ with solvent (Table 5) is an upper
limit that was calculated using kaz e 5 × 109 M-1 s-1 45
.
Table
5 also gives equilibrium constants for formation of the car-
bocations R-1+ from the corresponding neutral azide ion adducts
in 50:50 (v:v) MeOH/H2O, which were calculated as Kaz ) ksolv
/
kaz (M), where ksolv is the experimentally-determined, or
estimated first-order rate constant for the stepwise reaction of
R-1-N3,4 and kaz is the second-order rate constant for the reverse
reaction of the carbocation with azide ion (Scheme 2).
Effects of r-Thiocarbonyl and R-Carbonyl Substituents
at R-1+. Figure 3, constructed using the data in Table 5,
illustrates the effects of R-thioamide, R-amide, R-ester, and
R-methyl groups on the equilibrium stability (log Kaz) and the
kinetic reactivity towards solvent (log ks) of the 4-methoxy-
benzyl carbocation in 50:50 (v:v) MeOH/H2O. Despite its great
(1) Me2NC(S)-1+ may be strongly stabilized by resonance
delocalization of the positive charge at the benzylic carbon onto
the weakly electronegative and strongly polarizable sulfur atom
(A, R′ ) 4-MeOC6H4, Scheme 3). The requirement for the
(27) (a) McClelland, R. A.; Kanagasabapathy, V. M.; Banait, N. S.;
Steenken, S. J. Am. Chem. Soc. 1991, 113, 1009-1014. (b) McClelland,
R. A.; Cozens, F. L.; Steenken, S.; Amyes, T. L.; Richard, J. P. J. Chem.
Soc., Perkin Trans. 2 1993, 1717-1722.
(28) Schepp, N. P.; Wirz, J. J. Am. Chem. Soc. 1994, 116, 11749-11753.
(29) Bernasconi, C. F. AdV. Phys. Org. Chem. 1992, 27, 119-238.
(30) Bernasconi, C. F. Acc. Chem. Res. 1992, 25, 9-16.

How Does Organic Structure Determine Organic ReactiVity
J. Am. Chem. Soc., Vol. 118, No. 50, 1996 12609
Scheme 3
relatively large fractional loss of such resonance stabilization
of carbocations in the transition state for their capture by
nucleophiles has been shown to result in the unusually low
reactivities of several R-substituted benzyl carbocations toward
solvent.4,25,31
the reactive intermediate is the closed carbocation, but that
strongly electron-donating ring substituents favor formation of
the products of attack on the open carbocation.
(2) Me2NC(S)-1+ may also be stabilized by the nucleophilic
addition of sulfur to the benzylic carbon to give the “closed”
carbocation (B, R′ ) 4-MeOC6H4, Scheme 3). However, if
solvent reacts exclusively with an open carbocation that is the
minor species in a mixture of open and closed carbocations,
then sulfur bridging would result in a particularly large decrease
in ks for the addition of solvent to Me2NC(S)-1+, because this
stabilizing interaction will be completely lost in the transition
state.29,30
The results of ab initio calculations at the Hartree-Fock
6-31G* level show that there is extensive delocalization of
positive charge onto the R-thioamide group at the open form
of the R-thioamide substituted methyl carbocation (A, R′ ) H,
Scheme 3) and that this substituent stabilizes the methyl
carbocation by 35.7 kcal/mol (MP2/6-31G**) relative to the
hydride ion adducts.26 However, the conversion of the open
carbocation to its closed form (B, R′ ) H, Scheme 3) is
stabilizing by a further 47.2 kcal/mol (MP2/6-31G**).26 This
suggests that both resonance delocalization and sulfur bridging
to give a closed carbocation may contribute to the relatively
low reactivity of Me2NC(S)-1+ toward solvent.
The following results are consistent with the conclusion that
in strongly ionizing solvents Me2NC(S)-1-Y reacts to form a
significant concentration of the open R-thioamide substituted
carbocation.
(1) The acetolysis of optically active Me2NC(S)-1-O2CCF3
proceeds with partial loss of optical activity.18 This requires
formation of a symmetrical open carbocation intermediate for
the reaction of this substrate in acetic acid.
(2) Neighboring group participation by the R-thioamide group
would reduce the development of localized positive charge at
the benzylic carbon in the transition state for ionization of Me2-
NC(S)-1-Y, and this should lead to a reduction in the hyper-
conjugative stabilization of this transition state by an additional
R-methyl group. However, the data in Table 1 show that the
effects of the addition of an R-methyl group on kobsd for the
reactions of Me2NC(S)-1-Y, Me2NC(O)-1-Y, and EtOC(O)-1-Y
are similar (see above), which is consistent with the formation
of similar, open, carbocation intermediates for all of these
reactions.
Reactions of Me2NC(S)-2+. Figure 1A shows that the
reactions of Me2NC(S)-2-Y (Y ) 4-nitrobenzoate or 4-meth-
oxybenzoate) in 50:50 (v:v) MeOH/H2O in the presence of
increasing concentrations of azide ion lead to the formation of
good yields of the azide ion adduct Me2NC(S)-2-N3, and a
decrease in the yield of the alkene Me2NC(S)-3 from 100% in
the absence of azide ion, to a constant limiting yield of 63% in
the presence of large concentrations of azide ion. One
consequence of this unusual result is that the observed rate
constant ratios (kaz/ke)obsd (M-1) that are calculated from the
yields of the azide ion adduct and the alkene at high concentra-
tions of azide ion using eq 5, are inversely proportional to [N3-].
-
(kaz/ke)obsd ) [RN3]/[alkene][N3 ]
(5)
This is inconsistent with the simple partitioning of a carbocation
intermediate between deprotonation by solvent and nucleophilic
capture by azide ion, because this would require that the yield
of the alkene approach zero at high [N3-], and that the
partitioning ratio (kaz/ke)obsd be independent of [N3-]. There
are several possible explanations for the constant limiting yield
of Me2NC(S)-3 from the reactions of Me2NC(S)-2-Y in the
presence of high concentrations of azide ion.
Explanation 1: The Elimination Reaction is Concerted.
The constant limiting yield of Me2NC(S)-3 may be the result
of a concerted elimination reaction that bypasses the carbocation
intermediate of the stepwise mechanism, with an observed rate
constant that is independent of [N3-]. However, this possibility
can be excluded because the overall KIE of k(R-CH3)/k(R-CD3)
) 1.03 for the reaction of Me2NC(S)-2-(4-nitrobenzoate) in 80%
ethanol in water3 is very close to unity, so that it effectively
rules out concerted proton transfer and cleavage of the bond to
leaving group in the transition state.
Explanation 2: The Carbocation Undergoes Deprotona-
tion by the Leaving Group within an Ion Pair Intermediate.
The carbocation-leaving group ion pair intermediate Me2NC-
(S)-2+‚Y- is expected to be too short-lived to be trapped by
azide ion, because of its fast breakdown by diffusional separation
to give the free ions.32,33 Therefore, the constant limiting yield
of Me2NC(S)-3 might result from very fast proton transfer from
the carbocation to the leaving group within this ion pair.
However, this is unlikely because the reactions of Me2NC(S)-
2-(4-nitrobenzoate) and Me2NC(S)-2-(4-methoxybenzoate) in
the presence of azide ion give identical yields of the alkene
and the azide ion adduct (Figure 1A).
Explanation 3: Azide Ion Reacts with the Carbocation
as a Brønsted Base. Azide ion normally reacts with R-methyl
substituted carbocations exclusively as a nucleophile (Lewis
base).1,20 However, Me2NC(S)-2+ is an unusual R-methyl
carbocation, because it exhibits an extremely large tendency to
undergo loss of a proton to solvent to give the alkene Me2NC-
(3) The UV-visible and 13C NMR spectra of the cationic
species generated by ionization of 9-[N,N-dimethyl(thioforma-
midyl)]-9-fluorenyl chloride under stable ion conditions and by
laser flash photolysis are consistent with an open structure for
this carbocation in trifluoroethanol and methylene chloride.46
(4) The acetolysis of 6 gives an 86% yield of the rearranged
product 7 which results from the nucleophilic addition of solvent
to the closed carbocation, followed by rearrangement of the
initial nucleophile adduct.18 However, the change to the strongly
electron-donating 4-methoxy ring substituent results in the
exclusive formation of Me2NC(S)-1-O2CCH3 from the addition
of solvent to the benzylic carbon.18 This shows that when the
aromatic ring substituent(s) are strongly electron-withdrawing,
(32) Richard, J. P. J. Org. Chem. 1992, 57, 625-629.
(33) Amyes, T. L.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 7900-
7909.
(31) McClelland, R. A. Tetrahedron 1996, 52, 6823-6858.

12610 J. Am. Chem. Soc., Vol. 118, No. 50, 1996
Richard et al.
Scheme 4
-
(S)-3 (ks/ke < 0.01, Table 3). If Me2NC(S)-2+ exhibits the same
tendency to undergo deprotonation by azide ion (kB, Scheme
4), then the limiting ratio of the yields of the alkene and the
azide ion adduct from the stepwise reaction of Me2NC(S)-2-Y
at high [N3-] will represent the rate constant ratio kB/kaz for
reaction of azide ion as a Brønsted base and as a nucleophile,
as shown in Scheme 4. The solid lines in Figure 1A show that
there is a good fit of the product data to eqs 6 and 8 (ke . ks),
which were derived for the mechanism shown in Scheme 4.
The nonlinear least squares fits were obtained using the
relationship kB ) 1.7kaz (kaz/kB ) [RN3]lim/[alkene]lim ) 0.59),
to give kaz/ke ) 65 ( 2 M-1 as the single unknown parameter
in eqs 6 and 8 (the error is the standard deviation).
1 + (kB/ke)[N3 ]
falk
)
(6)
-
1 + (ks/ke) + (kB/ke + kaz/ke)[N3 ]
(ks/ke)
fROS
)
(7)
(8)
-
1 + (ks/ke) + (kB/ke + kaz/ke)[N3 ]
-
(kaz/ke)[N3 ]
fRN
)
-
3
1 + (ks/ke) + (kB/ke + kaz/ke)[N3 ]
-
1 + (kB/ke)[N3 ] + (kAc/ke)[AcO-]
1 + (kB/ke + kaz/ke)[N3 ] + (kAc/ke)[AcO-]
falk
)
(9)
-
We now consider whether it is reasonable that azide ion is
sufficiently reactive as a Brønsted base to abstract a proton from
Me2NC(S)-2+ to give the alkene Me2NC(S)-3 (Scheme 4). The
rate constant ratios kaz/kB ) 0.59 and kaz/ke ) 65 M-1 obtained
from the fit of the product data in Figure 1A to eqs 6 and 8 can
be combined to give kB/ke ) 110 M-1, so that the mechanism
in Scheme 4 requires that 1.0 M azide ion be 110-fold more
reactive than solvent as a Brønsted base for deprotonation of
Me2NC(S)-2+. The pKas of acetic acid (4.76) and hydrazoic
acid (4.72) are similar,34 so that acetate and azide ions should
show similar reactivities as Brønsted bases for deprotonation
of Me2NC(S)-2+ if the mechanism shown in Scheme 4 is
correct. The relative rates of deprotonation of Me2NC(S)-2+
by acetate and azide ions were determined from the effect of
increasing concentrations of acetate ion on the product distribu-
tion for the reaction of Me2NC(S)-2-(4-nitrobenzoate) in 50:50
(v:v) MeOH/H2O in the presence of fixed concentrations of
0.005 M (Figure 2A) or 0.010 M (Figure 2B) azide ion. Figure
2 shows that the addition of acetate ion results in an increase
in the yield of the alkene Me2NC(S)-3 at the expense of the
azide ion adduct; there was no detectable formation of the
acetate ion adduct in these experiments. The solid lines in
Figure 2 show that there is a good fit of the product data to eqs
9 and 10, which were derived for the mechanism shown in
Scheme 4 (ke . ks). The nonlinear least squares fits were
obtained using kB/ke ) 110 M-1 and kaz/ke ) 65 M-1 determined
in the absence of acetate ion (Figure 1A), to give kAc/ke ) 23
and 26 M-1 for [N3-] ) 0.005 and 0.010 M, respectively, as
the single unknown parameter in eqs 9 and 10. The average
value of kAc/ke ) 24.5 ( 1.5 M-1 can be combined with kB/ke
) 110 M-1 to give kB/kAc ) 4.5 as the rate constant ratio for
deprotonation of Me2NC(S)-2+ by azide and acetate ions.
The data in Figure 2 for the effect of acetate ion are fully
consistent with the mechanism shown in Scheme 4. The 4.5-
fold greater reactivity of azide than of acetate ion as a Brønsted
-
(kaz/ke)[N3 ]
fRN
)
(10)
-
1 + (kB/ke + kaz/ke)[N3 ] + (kAc/ke)[AcO-]
3
base for deprotonation of Me2NC(S)-2+ is not unusually large
for bases of similar pKa.35 It may result from either (a) a smaller
steric barrier to reaction of the linear azide ion or (b) a smaller
barrier to formation of the carbocation-anion contact ion pair
for azide than for acetate ion. Similarly, a difference in the
barriers to ion pair formation has been proposed to explain the
10-fold larger rate constant for the encounter-limited nucleo-
philic reaction of azide than of acetate ion with the 1-(4-
methylphenyl)ethyl carbocation (8).19
Table 3 summarizes the rate constant ratios for partitioning
of Me2NC(S)-2+ that were obtained from the fits of the data in
Figures 1A and 2 to eqs 6, 8, 9, and 10. Me2NC(S)-2+ should
be even more stable than its counterpart without an R-methyl
group, Me2NC(S)-1+, for which kaz e 5 × 109 M-1 s-1 (see
above). Therefore, the rate constant ratios kaz/ke (M-1) and kaz/
kB were combined with (kaz + kB) e 5 × 109 M-1 s-1 for the
reaction of azide ion with Me2NC(S)-2+ as a nucleophile and
as a Brønsted base to give the absolute rate constants ke (s-1
)
and kB (M-1 s-1) for the deprotonation of Me2NC(S)-2+ by
solvent and azide ion given in Table 6. An upper limit on ks
(s-1) for the reaction of Me2NC(S)-2+ with a solvent of 50:50
(34) 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.
(35) See, for example, the deviations from the Brønsted correlation of
rate constants for general-base-catalyzed deprotonation of triosephos-
phates: Richard, J. P. J. Am. Chem. Soc. 1984, 106, 4926-4936.

How Does Organic Structure Determine Organic ReactiVity
J. Am. Chem. Soc., Vol. 118, No. 50, 1996 12611
partitioning of Me2NC(S)-2+ between bimolecular deprotonation
by acetate ion and a unimolecular reaction is even more
faVorable than for the partitioning of 8, and that there is no
efficient intramolecular pathway for the conversion of Me2NC-
(S)-2+ to Me2NC(S)-3.
Table 6. Rate Constants for Reaction of R-Substituted
1-(4-Methoxyphenyl)ethyl Carbocations R-2+ with Solvent and for
Base-Catalyzed Deprotonation of the Carbocations by Solvent and
Azide Ion (Scheme 4)a
R
ksb (s-1
)
kec (s-1
)
kBd (M-1 s-1
)
Me2NC(S)
Me2NC(O)
EtOC(O)
CH3
e1 × 104 e
3.6 × 106
6.0 × 107
2.8 × 108
e2.9 × 107 f
≈1.2 × 106 j
< 2.8 × 106 j
e3.1 × 109 g
Reactions of Me2NC(O)-2+. Figure 1B shows that the
reaction of Me2NC(O)-2-(pentafluorobenzoate) in the presence
of increasing concentrations of azide ion in 50:50 (v:v) MeOH/
H2O results in formation of good yields of the azide ion adduct
Me2NC(O)-2-N3 at the expense of the solvent adducts Me2NC-
(O)-2-OSolv and the alkene Me2NC(O)-3. At high [N3-], the
yield of the solvent adducts decreases to zero, while the yield
of the alkene decreases to a constant limiting yield of 9%. As
discussed above for the reactions of Me2NC(S)-2+, these data
are consistent with a mechanism in which azide ion acts as a
Brønsted base to abstract a proton from Me2NC(O)-2+ to give
the alkene Me2NC(O)-3 (kB, Scheme 4). Table 3 summarizes
the rate constant ratios for partitioning of Me2NC(O)-2+ that
were evaluated from the product data for the reaction of Me2-
NC(O)-2-(pentafluorobenzoate) as follows:
2.9 × 106 h
4.5 × 108 i
a In 50:50 (v:v) methanol/water at 22 ( 2 °C and I ) 0.50 (NaClO4).
b First-order rate constant for reaction of the carbocation with solvent,
calculated from kaz/ks (M-1, Table 3) and kaz ) 5 × 109 M-1 s-1 for
the diffusion-limited reaction of azide ion with R-2+, unless noted
otherwise. c First-order rate constant for deprotonation of R-2+ by
solvent. d Second-order rate constant for deprotonation of R-2+ by azide
ion. e Upper limit calculated from ks e7 × 104 s-1 for the reaction of
Me2NC(S)-1+ with 50:50 (v:v) methanol/water (Table 5) and the 7-fold
effect of the addition of an R-methyl group on ks (s-1) for the reaction
of this solvent with Me2NC(O)-1+ (Tables 5 and 6). f Calculated from
kaz/ke (M-1) and kaz/kB (Table 3) and (kaz + kB) e5 × 109 M-1 s-1 for
the reaction of azide ion with the carbocation as both a nucleophile
and a Brønsted base. g Calculated from kaz/kB (Table 3) and (kaz + kB)
e5 × 109 M-1 s-1
.
h Calculated from kaz/ke (M-1, Table 3) and kaz
i Calculated from kaz/kB (Table 3) and (kaz + kB) )
.
j Calculated from ks (s-1) and the value of ks/ke (Table
)
(1) The rate constant ratio kaz/ks ) 1400 M-1 for partitioning
of Me2NC(O)-2+ between nucleophilic addition of azide ion
and solvent (Scheme 4) was determined from the ratio of the
yields of the azide ion adduct and the solvent adducts using eq
4.
5 × 109 M-1 s-1
.
5 × 109 M-1 s-1
3).
Scheme 5
(2) The rate constant ratio ks/ke ) 1.2 for partitioning of Me2-
NC(O)-2+ between nucleophilic addition of solvent and depro-
tonation by solvent was determined as the ratio of the yields of
the solvent adducts (54%) and the alkene (46%) in the absence
of azide ion (Figure 1B).
(3) The rate constant ratio kaz/kB ) 10 for partitioning of Me2-
NC(O)-2+ between nucleophilic addition of azide ion and
deprotonation by azide ion acting as a Brønsted base was
calculated as [RN3]lim/[alkene]lim ) 10, which is the ratio of
the limiting yields of the azide ion adduct (91%) and the alkene
(9%) at high [N3-] (Figure 1B).
The solid lines in Figure 1B show the fits of the product data
to eqs 6-8, which were calculated using ks/ke ) 1.2, kaz/ke )
(ks/ke)(kaz/ks) ) 1700 M-1, and kB/ke ) (kaz/ke)(kB/kaz) ) 170
M-1. The absolute agreement between the observed product
yields and those expected for the mechanism shown in Scheme
4 is better than (1%, so that this mechanism provides an entirely
consistent and satisfactory explanation for the effect of azide
ion on the product distribution for the reaction of Me2NC(O)-
2-(pentafluorobenzoate). The rate constant ratios for partitioning
of Me2NC(O)-2+ were combined with (kaz + kB) ) kd ) 5 ×
109 M-1 s-1 for the diffusion-limited27 reaction of azide ion
with Me2NC(O)-2+ as a nucleophile and as a Brønsted base, to
give the absolute rate constants ks (s-1), ke (s-1), and kB (M-1
s-1) for the reactions of this carbocation given in Table 6.
Reactions of EtOC(O)-2+ and CH3-2+. Table 3 gives the
rate constant ratios for partitioning of EtOC(O)-2+ and CH3-
2+ between nucleophilic addition of azide ion and solvent that
were calculated from the yields of the azide ion adduct and the
solvent adducts using eq 4. The effect of azide ion on the low
(≈2%, Table 3) yield of the alkene EtOC(O)-3 from the reaction
of EtOC(O)-2-(pentafluorobenzoate) in 50:50 (v:v) MeOH/H2O
was not determined.
(v:v) MeOH/H2O was obtained in two ways: (1) A limit of ks
< 2.9 × 105 s-1 was calculated from ke e 2.9 × 107 s-1 (Table
6) and ks/ke < 0.01 (Table 3). (2) A smaller limit of ks e 1 ×
104 s-1 (Table 6) was calculated from ks e 7 × 104 s-1 for the
reaction of Me2NC(S)-1+ with 50:50 (v:v) MeOH/H2O (Table
5), with the assumption that the addition of an R-methyl group
leads to the same 7-fold decrease in ks as that observed for the
reaction of Me2NC(O)-1+ (Tables 5 and 6).
Explanation 4: Efficient Intramolecular Proton Transfer.
The quantitative yield of alkene obtained from the reaction of
Me2NC(S)-2+ in the absence of azide ion might be a conse-
quence of a high “effective molarity”36 of the R-thioamide group
as an intramolecular proton or hydride acceptor (Scheme 5).
However, the observation that good yields of Me2NC(S)-3 are
obtained from intermolecular general base catalysis of the
deprotonation of Me2NC(S)-2+ by acetate and azide ions
(Figures 1A and 2) effectively excludes the possibility of an
efficient intramolecular pathway for formation of the alkene.
The value of kAc/ke ) 24.5 M-1 for partitioning of Me2NC(S)-
2+ between bimolecular and unimolecular elimination in 50:50
(v:v) MeOH/H2O is even larger than kAc/ke ) 2.0 M-1 for the
partitioning of 8 in 50:50 (v:v) TFE/H2O,37 for which intramo-
lecular proton transfer is not possible. We conclude that the
Table 6 gives absolute rate constants for the reactions of
EtOC(O)-2+ and CH3-2+ that were calculated from the rate
constant ratios in Table 3 and kaz ) kd ) 5 × 109 M-1 s-1 for
(37) Calculated from kAc/ks ) 1.43 × 10-3 M-1 and ke/ks ) 7 × 10-4
for partitioning of the 1-(4-methylphenyl)ethyl carbocation between depro-
tonation by acetate ion or solvent and the nucleophilic addition of solvent
in 50:50 (v:v) TFE/H2O.19
(36) Kirby, A. J. AdV. Phys. Org. Chem. 1980, 17, 183-278.

12612 J. Am. Chem. Soc., Vol. 118, No. 50, 1996
Richard et al.
Table 7. Effects of R-Carbonyl and R-Thiocarbonyl Substituents
on the Relative Stabilities of R-Substituted 1-Phenylethyl Alcohols
and the Corresponding R-Substituted Styrenes (Scheme 6)a
∆H (kcal/mol)b
∆G (kcal/mol)c
R-substituent R 3-21G//3-21G 6-31G*//3-21G expl
ke/ksd
CH3
0
1.07
-1.52
-1.79
0
0.27
-1.57
-5.04
0
nd
nd
<0.01
≈0.02e
0.83
MeOC(O)
Me2NC(O)
Me2NC(S)
<-3.9f >100
a Energy changes are for the isodesmic reaction shown in Scheme
6, unless noted otherwise. b Calculated from the ab initio energies of
R-9-OH and R-10 given in Table 4. c For compounds with a 4-methoxy
substituent at the aromatic ring. d Rate constant ratio for partitioning
of R-2+ between deprotonation by solvent to give R-3 and nucleophilic
addition of solvent to give R-2-OSolv in 50:50 (v:v) methanol/water
(Table 3). e For R ) EtOC(O). f Calculated from Kalk ) 0.13 and >100,
respectively, for the dehydration of CH3-2-OH to give CH3-3 (ref 2)
and of Me2NC(S)-2-OH to give Me2NC(S)-3 (this work) in 50:50 (v:
v) trifluoroethanol/water.
the diffusion-limited reaction of azide ion27 as a nucleophile
with these unstable carbocations.
Substituent Effects on the Reactivity of R-2+. The effects
of the change from an R-methyl substituent to an R-amide or
an R-ester substituent on ks (s-1) for the reactions of CH3-1+
(Table 5) and CH3-2+ (Table 6) with a solvent of 50:50 (v:v)
MeOH/H2O are similar. However, there is no observable
nucleophilic addition of solvent to the R-thioamide substituted
carbocation Me2NC(S)-2+, and solvent reacts with this car-
bocation exclusively as a Brønsted base to abstract a proton
from the R-methyl group to give Me2NC(S)-3 (ks/ke < 0.01,
Table 3). Me2NC(S)-2+ also shows an unusual resistance to
the addition of nucleophilic anions, which react preferentially
as Brønsted bases: there is no detectable nucleophilic addition
of acetate ion to Me2NC(S)-2+ (see above), and the potent
nucleophile azide ion is 1.7-fold less reactive as a nucleophile
than as a Brønsted base toward Me2NC(S)-2+ (Table 3).
Table 3 shows that there is a progressive increase in the rate
constant ratio ke/ks for the partitioning of R-2+ between loss of
a proton from the R-methyl group and the nucleophilic addition
of solvent as the R-substituent R is changed along the series
R-methyl, R-ester, R-amide, R-thioamide. The absolute rate
constants for these reactions (Table 6) show that these changes
in ke/ks are a consequence of the relatively large decreases in ks
along the series CH3-2+, EtOC(O)-2+, Me2NC(O)-2+, Me2NC-
(S)-2+, relative to ke, which shows a much smaller sensitivity
to changes in the R-substituent at R-2+.
The changes in the rate constant ratio ke/ks for partitioning
of R-2+ between loss of a proton and the nucleophilic addition
of solvent may be due to interactions which develop in the two
different transition states for these reactions, and which are then
lost on proceeding to the products; or they may simply be the
consequence of changes in the overall thermodynamic driving
force for formation of the two different products, which are
partially expressed in the respective transition states for their
formation (Figure 4). The change from an R-methyl group to
an R-thioamide group results in a large increase in the
equilibrium constant for the dehydration of R-2-OH to give the
alkene R-3, from Kalk ) [alkene]eq/[alcohol]eq ) 0.13 for the
dehydration of CH3-2-OH to give CH3-3 in 50:50 (v:v) TFE/
H2O,2 to Kalk > 100 for the dehydration of Me2NC(S)-2-OH to
give Me2NC(S)-3 (this work), which corresponds to a change
in ∆G for this reaction of more than 3.9 kcal/mol (Table 7).
This shows that R-3 is strongly stabilized relative to R-2-OH
by the change in R from an R-methyl to an R-thioamide group
(Figure 4). This large ground state effect is partly or entirely
responsible for the dramatic effect of the change from an
Figure 4. Hypothetical free energy profiles for the partitioning of CH3-
2+ and Me2NC(S)-2+ between nucleophilic addition of solvent to give
the solvent adducts (ks) and deprotonation to give the alkene (ke) in
50:50 (v:v) methanol/water. The arrows indicate the effect of a change
from an R-methyl to an R-thioamide substituent at R-2+ on the relative
stabilities of the two products and of the respective transition states
for their formation.
Scheme 6
R-methyl to an R-thioamide group on ke/ks for partitioning of
R-2+ between loss of a proton and nucleophilic capture by
solvent (Figure 4 and Table 7).
We were unable to determine the equilibrium constants for
the acid-catalyzed dehydration of Me2NC(O)-2-OH and EtOC-
(O)-2-OH, because these compounds undergo competing acid-
catalyzed hydrolysis of the R-amide and R-ester groups. The
energies of R-thioamide, R-amide, R-ester and R-methyl sub-
stituted 1-phenylethyl alcohols (R-9-OH) and the corresponding
R-substituted styrenes (R-10) were therefore examined using
ab initio methods (Table 4), and the results were used to
calculate the enthalpy changes for the isodesmic reactions shown
in Scheme 6 (∆H, Table 7). The experimental value of ∆G <
-3.9 kcal/mol for the reaction of Me2NC(S)-2-OH with CH3-3
to give CH3-2-OH and Me2NC(S)-3 (cf. Scheme 6) in 50:50
(v:v) TFE/H2O is close to ∆H ) -5.04 kcal/mol for the
corresponding reaction of Me2NC(S)-9-OH (Scheme 6) deter-
mined from ab initio calculations at the 6-31G*//3-21G level
(Table 7). This provides good evidence that the values of ∆H
for the reactions shown in Scheme 6 determined by ab initio
calculation provide reasonable estimates of ∆G for the corre-
sponding reactions of the 4-methoxy compounds in TFE/H2O.
The data in Table 7 show that there is good correlation
between the increase in the rate constant ratio ke/ks for

How Does Organic Structure Determine Organic ReactiVity
J. Am. Chem. Soc., Vol. 118, No. 50, 1996 12613
partitioning of R-2+ between loss of a proton to give the alkene
R-3 and nucleophilic capture by solvent to give R-2-OSolv and
the change to a more negative value of ∆H for the reaction
shown in Scheme 6. This supports the conclusion that the
relative rate constants for loss of a proton from R-2+ (ke) and
its nucleophilic capture by solvent (ks) are strongly controlled
by the relative stabilities of the alkene and alcohol products of
these reactions. This provides yet another example of a
substituent effect on rate constants for the interconversion of
stable neutral and reactive charged species that is dominated
by the substituent effect on the relative stabilities of neutral
compounds.38-41
Our experiments and calculations do not allow for a detailed
explanation of the effects of R-thiocarbonyl and R-carbonyl
substituents on the observed changes in ∆G and the calculated
changes in ∆H for the reaction shown in Scheme 6. However,
we suggest that these changes are controlled largely by the
balance between two opposing interactions of these R-substit-
uents at the alkenes R-3 or R-10.
the electron-deficient alkene double bond at R-1043 which
decreases in magnitude on proceeding from the strongly
electron-withdrawing R-ester substituent to the more weakly
electron-withdrawing R-amide and R-thioamide substituents.44
The data in Table 7 suggest that these opposing effects
described in 1 and 2 above cancel at MeOC(O)-10, which is of
similar stability to CH3-10, relative to the corresponding alcohols
R-9-OH, but that the stabilizing interaction between adjacent
sp2-hybridized centers is responsible for the much greater
stability of Me2NC(S)-10.
Acknowledgment. The ab initio calculations were made
possible by a grant to J.M.B. of computer time on the Cray
Y-MP8/864 computer at the Ohio Supercomputer Center
(PSJ000-1). The experimental work was made possible by a
grant to J.P.R. from the National Institutes of Health (Grant
GM 39754). We are grateful to Xavier Creary for providing
compounds used during the initial stages of this study.
(1) A stabilizing interaction between the adjacent sp2-
hybridized centers, which is of roughly equal magnitude for
the R-thioamide, R-amide, and R-ester groups. This is not likely
to be a simple resonance interaction, because the adjacent
π-centers are not coplanar in the ground states of R-10 optimized
at the 3-21G level, but are twisted away from coplanarity by
70°, 60°, and 10° for Me2NC(S)-10, Me2NC(O)-10 and MeOC-
(O)-10, respectively. Rather, the stabilization of the alkenes
R-10 by R-thiocarbonyl and R-carbonyl substituents (Table 4)
is likely a consequence of the σ-bond between the sp2-hybridized
carbon of these R-substituents and that of the alkene double
bond, which is expected to be stronger than the corresponding
σ-bond between these R-substituents and the benzylic carbon
at R-9-OH.42,43
Supporting Information Available: Text giving details of
procedures for the synthesis of the ring-substituted alcohols,
azides, and benzoate derivatives R-1-Y and R-2-Y and the
R-substituted styrenes R-3 and spectroscopic and analytical data
for these compounds and tables of 3-21G geometries of the
optimized structures of Me2NC(S)-9-OH, Me2NC(O)-9-OH,
MeOC(O)-9-OH, CH3-9-OH, Me2NC(S)-10, Me2NC(O)-10,
MeOC(O)-10 and CH3-10 (16 pages). See any current masthead
page for ordering and Internet access instructions.
JA9629064
(43) Hine, J. Structural Effects on Equilibria in Organic Chemistry;
Wiley-Interscience: New York, 1975; pp 270-276.
(44) The change in the geminal interactions between the R-substituent
R and the hydroxyl group at R-9-OH with changing R has not been
determined. The destabilizing polar interactions between the electron-
withdrawing hydroxyl group and the geminal R substituent should decrease
as R is changed from an R-ester to the more weakly electron-withdrawing
R-amide and R-thioamide substituents. However, this would result in the
stabilization of the alcohol Me2NC(S)-9-OH relative to the alkene Me2-
NC(S)-10, rather than the observed stabilization of the alkene R-10 by the
R-thioamide substituent (Table 7).
(2) A destabilizing dipole-dipole interaction between the
electron-deficient R-thiocarbonyl or R-carbonyl substituent and
(38) Apeloig, Y.; Karni, M. J. Chem. Soc., Perkin Trans. 2 1988, 625-
636.
(39) Apeloig, Y.; Biton, R.; Abu-Freih, A. J. Am. Chem. Soc. 1993, 115,
2522-2523.
(40) Richard, J. P.; Amyes, T. L.; Rice, D. J. J. Am. Chem. Soc. 1993,
115, 2523-2524.
(45) The recently reported value of kaz ) 1.5 × 109 M-1 s-1 for reaction
of the 9-[N,N-dimethyl(thioformamidyl)]-9-fluorenyl carbocation with azide
ion in a solvent of trifluoroethanol lies below the diffusion-controlled limit.46
(46) Lew, C. S. Q.; Wong, D. F.; Johnston, L. J.; Bertone, M.; Hopkinson,
A. C.; Lee-Ruff, E. J. Org. Chem. 1996, 61, 6805-6808.
(41) Jagannadham, V.; Amyes, T. L.; Richard, J. P. J. Am. Chem. Soc.
1993, 115, 8465-8466.
(42) Dewar, M. J. S.; Schmeising, H. N. Tetrahedron 1960, 11, 96-
120.